Quarto trim size: 174mm × 240mm

Industrial Management & Data Systems

Volume 119 Number 7 2019 ISSN 0263-5577
Volume 119 Number 7 2019

Industrial Management & Data Systems

Number 7
Financial technologies: artificial intelligence, blockchain,
Industrial Management
and crowdfunding
Guest Editor: Xiuping Hua
1401 Current practices, new insights, and emerging trends of financial technologies
& Data Systems
Xiuping Hua, Yiping Huang and Yanfeng Zheng Financial technologies: artificial intelligence,
1411 Artificial Intelligence in FinTech: understanding robo-advisors adoption among customers blockchain, and crowdfunding
Daniel Belanche, Luis V. Casaló and Carlos Flavián
1431 Understanding speculative investment behavior in the Bitcoin context from a
Guest Editor: Xiuping Hua
dual-systems perspective

Volume 119 Number 7 2019

Hyun-Sun Ryu and Kwang Sun Ko
1457 Investigating the influence of organizational factors on blockchain adoption: an
innovation theory perspective
Trevor Clohessy and Thomas Acton
1492 Examining the role of narratives in civic crowdfunding: linguistic style and message
substance
Chang Heon Lee, Yiyang Bian, Rajaa Karaouzene and Nasreen Suleiman
1515 Exploring individuals’ behavioral intentions toward donation crowdfunding: evidence
from China
Tao Wang, Yalan Li, Minghui Kang and Haichao Zheng
1535 An empirical analysis of rural farmers’ financing intention of inclusive finance in China:
the moderating role of digital finance and social enterprise embeddedness
Gulizhaer Aisaiti, Luhao Liu, Jiaping Xie and Jun Yang

emeraldpublishing.com

ISBN 978-1-83909-714-0

www.emeraldinsight.com/loi/imds
 The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm

Financial
Current practices, new insights, technologies
and emerging trends of
financial technologies
Xiuping Hua 1401
Centre for Inclusive Finance, Business School,
Received 10 August 2019
University of Nottingham, Ningbo, China Accepted 12 August 2019
Yiping Huang
National School of Development, Peking University, Beijing, China, and
Yanfeng Zheng
Faculty of Business and Economics, University of Hong Kong, Hong Kong

Abstract
Purpose – Financial technologies, also known as “FinTech,” have brought disruptive changes to virtually
every aspect of financial services and are becoming increasingly important in the world economic system.
The purpose of this paper is to proffer a bird view of some recent studies in the key research areas of FinTech,
such as artificial intelligence, blockchain, crowdfunding and then to summarize the key contributions made
by all the six papers in this special issue.
Design/methodology/approach – A literature review approach is adopted, and the summary shows that
most types of FinTech innovations generate positive value to innovators, financial customers and the society.
The current implications and future directions are explored based on theoretical and empirical analyses.
Findings – The benefits from and determinants of FinTech applications vary across different financial
sectors. Together the summary of this special issue suggests that there is substantial value creation in further
exploring the dynamics, mechanisms and social consequences of FinTech.
Originality/value – This study helps to extend knowledge by summarizing the current practices, proffering
new insights and watching out emerging trends of financial technologies, and to shed light on a variety of
subjects of interest to practitioners, academics and policy makers by suggesting for the future research topics.
Keywords Blockchain, Cryptocurrency, Crowdfunding, Artificial intelligence (AI)
Paper type Editorial

1. Introduction
Financial technologies, known as “FinTech,” have brought disruptive changes to virtually
every aspect of financial services, and is currently revolutionizing the whole financial
industry (Goldstein et al., 2019). Driven by start-ups and technology firms, FinTech covers a
wide range of applications from cryptocurrency, mobile payments, marketplace financing,
robo-advisors, to smart contracts and even decentralized autonomous organizations (Böhme
et al., 2015; Yu et al., 2017). FinTech has received significant attention among not only
practitioners, investors, regulators but also consultants and academia who are curious
about how these novel technologies and changes impact organizations and even broader
financial system (Stanko et al., 2017; The Libra Association, 2019; Chong et al., 2019).
In a world that is increasingly focused on integrating digital technology innovations
into normal people’s lives, accustoming users to FinTech is capable of creating
long-lasting and valuable effects and helping firms gain sustainable competitive
advantages. FinTech helps to reduce the costs, to increase the reachability of customers,
and to manage risks in a more efficient way. In this special issue, several key emerging Industrial Management & Data
Systems
Vol. 119 No. 7, 2019
The authors are grateful to Professor Alain Chong, Professor Hing Kai Chan and Dr Zhao Cai for their pp. 1401-1410
support for the special issue. This paper was supported by National Office for Philosophy and Social © Emerald Publishing Limited
0263-5577
Science of China under National Social Science Fund programmes 2019 with Project Code 19BJY252. DOI 10.1108/IMDS-08-2019-0431
IMDS areas of FinTech, such as artificial intelligence (AI), blockchain, crowdfunding, are
119,7 theoretically and empirically examined, and the FinTech’s social consequences on
inclusive finance are also explored.
AI is commonly regarded as a foundational technology that can transform virtually
every corner of financial services, ranging from credit scoring, customer interaction,
robot-advisory, market research and even fraud detection (Belanche et al., 2019; D’Acunto
1402 et al., 2019). It goes beyond conventional algorithms with self-learning mechanisms,
and therefore exhibits tremendous potential in restructuring business operations,
encouraging market development, enabling novel business models and even improving
regulatory effectiveness. D’Acunto et al. (2019) examine a robo-advising program
that delivers AI diversification advice to individual investors, and find that robo-advising
has a variety of impacts on investors’ performance based on different levels of
sophistication, and, at the same time, decreases a set of well-known behavioral biases for
all investors.
Blockchain technology is a novel platform-like technology that enables multiple
applications in economic transactions, fund raising and alike. In the areas of finance,
blockchain is widely used for money transfer, distributed computing and digitalizing
assets (Goldstein et al., 2019). By virtue of its immutability, blockchain holds the
potential of displacing traditional financial intermediaries, and the benefits of blockchain
have been acknowledged across a variety of financial sectors spanning banking
services, trade credit finance, exchanges and regulations and so on (Tapscott and
Tapscott, 2016; Chong et al., 2019).
Crowdfunding has become prevalent in recent years as a novel channel for new venture
financing, especially early-stage, creative and social ventures (Mollick, 2014). Digital
innovations, especially Web 2.0, enable people to pledge for resources through
crowdsourcing like open calls in online social communities (Kleemann et al., 2008), and
support the operation of crowdfunding platforms to facilitate the matching process between
the potential funders and potential entrepreneurs, as well as the financing and investment
processes between entrepreneurs and investors (Burtch et al., 2013).
Applying financial technologies in the practices of inclusive finance to improve the social
welfare and economic growth of a country is a normal practice in many countries, such as
China, India and so on. Understanding the relationship between FinTech and financial
inclusion is very important, but so far only limited academic evidence is provided. The
analysis of how the society and customers approach FinTech as well as mechanisms
through which Fintech influences financial inclusion enhance the knowledge body on the
scope and outcome of FinTech activities (Goldstein et al., 2019).
This special issue on “Financial Technologies: Artificial Intelligence, Blockchain, and
Crowdfunding” called for paper in February 2018, and has accepted all the revised and
resubmitted papers by the middle of July 2019. By summarizing all the accepted papers,
we find that this special issue contributes to the current literature in several ways. It
provides the large-scale theoretical and empirical evidence on the business adoption and
value creation mechanisms of various FinTech tools and innovations. It also helps to
extend knowledge by summarizing the current practices, proffering new insights and
watching out emerging trends of financial technologies. Finally, it sheds light on a variety
of subjects of interest to practitioners, academics and policy makers by suggesting for the
future research topics.
The reminder of this paper is structured as follows. In the next section, we present a bird
view of the most recent literatures focusing on the evolving financial technologies as well as
their business applications. We then discuss the contributions made in this special issue by
summarizing the findings and implications of the papers. Finally, we provide suggestions
for future research.
2. Recent knowledge on emerging FinTech Financial
Despite the rise of FinTech innovation worldwide and the widespread interest in FinTech, technologies
little is currently known about exactly how it will impact existing financial industry and
business models of financial and technology companies (Chen et al., 2019). The knowledge in
the literature is still limited and here we summarize some recent ones to proffer a bird view
of current practices and new insights in the key research areas of FinTech, such as robot-
advisory, blockchain, crowdfunding and so on. 1403
Who are investing in the FinTech industry and drive the radical innovation to date?
Chen et al. (2019) construct a very large and novel data set of published FinTech patent
applications over the 2003–2017 period, and document a number of basic facts about
FinTech innovation. First, their sample reveal that publicly traded companies as a group
have driven only a minority of FinTech innovations to date. Second, in sharp contrast,
private companies and non-firm individuals account for about 62.7 percent of FinTech
patent filings. Finally, they also find that about 57.8 percent of the FinTech corporate filings
are actually from technology firms outside of the financial industry.
The AI’s application in the finance industry has started as a disruptive play by new
entrants, and robo-advising is seen as an opportunity to help customers that human
advisers are not able to serve. For instance, robo-advisers typically use replicable
algorithms based on financial theory, employ technology that in many instances can
simplify and speed up contact with clients, and are much more transparent than human
advisors (D’Acunto et al., 2019). However, robo-advisers have also faced criticism such as the
lack of self-control, inducing overtrading and not being able to translate into better portfolio
performance (D’Acunto et al., 2019).
It is well recorded in the recent literature that blockchain is able to revolutionize the way
how the payment clearing and credit information systems work in banks, and also to promote
the formation of “multi-center-and-weakly-intermediated” scenarios. Cong and He (2019)
examine coordination on blockchain in the landscape for smart contracts and analyze how
decentralization affects consensus effectiveness, and how the quintessential features of
blockchain reshape industrial organization and the landscape of competition. They find that
smart contracts are capable of mitigating informational asymmetry and improving welfare
and consumer surplus through enhanced entry and competition. By analyzing five companies
in mainland China that have rolled out blockchain initiatives, Chong et al. (2019) derive a
typology of five corresponding blockchain-inspired digital business models, add insights into
each business model’s value creation logic and value capturing mechanism. They also
illuminate the challenges associated with pursuing each of these digital business models.
The unexpected success of cryptocurrencies enabled by blockchain technology has
triggered hundreds of business startups to build services such as mobile payment and
international remittance. Since Bitcoin was first created by Nakamoto (2008),
cryptocurrencies have attracted significant attention from investors, regulators and the
media (Böhme et al., 2015; Cheah and Fry, 2015; Foley et al., 2019). The proof-of-work
blockchain protocol, however, is not always capable of offering a new, cost effective way to
record transactions and sometimes it is a coordination game with multiple equilibria
(Biais et al., 2019).
By modeling the proof-of-work blockchain protocol as a stochastic game and analyzing the
equilibrium strategies of rational and strategic miners, Biais et al. (2019) study the dynamics of
the blockchain and prove that the blockchain protocol is not always successful at avoiding
forks. They find that equilibria with forks in public blockchain such as Bitcoin or Ethereum
lead to orphaned blocks and persistent divergence between chains, and show how forks can
be generated by information delays and software upgrades. They also identify negative
externalities and imply that equilibrium investment in computing capacity is excessive. Thus
the pitfalls of blockchain tools in the transactions may not always be avoidable.
IMDS Even worse, Böhme et al. (2015) point out that in the Bitcoin transactions, the early notable
119,7 adopters were businesses that sought features not easily available through other options:
greater anonymity and the absence of rules concerning what might be bought or sold. Foley
et al. (2019) suggest that cryptocurrencies are enabling “black e-commerce” and currently
transforming the way black markets operate. They develop methods for quantifying and
characterizing the illegal trade facilitated by Bitcoin. They find that approximately one-quarter
1404 of Bitcoin users and one-half of Bitcoin transactions are associated with illegal activity.
Although the illegal share of Bitcoin activity declines over time, around $72bn of illegal
activity per year involves Bitcoin, which is rather close to the scale of the European and US
markets for illegal drugs trades. Their research helps to better understand the nature and scale
of the “problem” facing unregulated market of cryptocurrencies and seizes the importance of
further development of regulation framework and technology on FinTech markets in the
future. Griffin and Shams (2018) explores whether another type of digital currency, Tether, has
been manipulated to influence Bitcoin and other cryptocurrency prices during the recent
market boom, and positive empirical results supporting this hypothesis are documented.
Crowdfunding has enriched the tools for new venturing financing, and meanwhile, it also
has contributed to the fluctuations of alternative financing markets. The crowdfunding
investment process combines entrepreneurship and social networking. Calic and
Mosakowski (2016) look into the literature and find that social entrepreneurs, relative to
commercial entrepreneurs, stand at a larger disadvantage at acquiring resources through
traditional financial institutions. They confirm that crowdfunding, as an innovative
institutional form, has emerged to address the needs of social entrepreneurs and other
entrepreneurs with limited access to traditional sources of capital, and thus FinTech such as
crowdfunding generate positive externalities.
This emerging literature has primarily focused on factors that lead to successful
campaigns such as expert opinion, homophily, linguistic style and social capital (Butticè
et al., 2017; Colombo et al., 2015; Greenberg and Mollick, 2017; Parhankangas and Renko,
2017). Interestingly, researchers have also found that pre-existing social network have
significant impact on crowdfunding campaign outcomes. Davis et al. (2017) investigate how
backers’ perceptions of a crowdfunding project’s creativity influence their resource
allocation decisions, and test their conceptual model using a two-stage study design in
which 102 participants in a reward crowdfunding platform. The evidence indicate that
perceived product creativity is positively related to crowdfunding performance via positive
affective reactions of prospective backers. More importantly, their research reveal that
the indirect effect of product creativity is dependent upon the extent to which funders
perceive an entrepreneur to be passionate, such that perceived entrepreneurial urge
increases the positive nature of the indirect effect.
In the technology innovation domain, crowdfunding is now becoming a more and more
commonly used financial channel for innovating entrepreneurs. Ordanini et al. (2011) argue
that nowadays investors in the crowdfunding platforms play a new role of customers who
collectively create a capital pool to invest in crowdfunding efforts. Allison et al. (2017)
describe the reward crowdfunding process as sellers persuading the customers to make the
buying decision from the angle of customer satisfaction theory and elaboration likelihood
model (ELM) of persuasion. Stanko and Henard (2017) argue that backers of crowdfunding
projects play a very active role in the innovation conversation and crowdfunding can be
viewed as one form of open search, namely, actively seeking out ideas from outsiders,
while backers help to generate awareness for the crowdfunded product and thus become the
earliest possible adopters. Therefore, these types of backers/investors are in fact much more
valuable than traditional early adopting consumers.
The failure rates related to crowdfunding are usually high, which is around 60 percent,
there is growing interest exhibited in the literature to better understand the factors related
to a crowdfunding campaign’s success. Kuppuswamy and Bayus (2017) propose that Financial
perceptions of impact matters, and people support crowdfunding projects financially when technologies
they believe that their contribution will make an impact. Parhankangas and Renko (2017)
focus on the linguistic style of crowdfunding pitches and how such a style relates to the
success in raising funds. By analyzing 656 Kickstarter campaigns empirically, they
demonstrate that linguistic styles make the campaigns and their founders more
understandable and relatable to the crowd, and hence help to boost the success of social 1405
campaigns, though hardly matter for commercial campaigns. Roma et al. (2017) investigate
a sample of technology projects launched on Kickstarter. Their analyses demonstrate that
pledging a higher amount of money in crowdfunding is able to flare up professional backers’
interest and thus help to secure subsequent capital, but this positive evidence is in effect
only when complemented by the presence of patents or a large network of social bonds.

3. Contributions in the special issue

In this section, we summarize the key research findings and empirical implications of the six
accepted articles.
It is well known that there is a strong and close link between financial development and
economic growth, and thus the determinants of adoption of digital technology innovation in
the finance industry has been increasingly studied (Ilyina and Samaniego, 2011). Belanche
et al. (2019) propose a research framework to better understand robo-advisor adoption in the
finance industry which predicts that personal and socio-demographic variables, such as
familiarity with robots, age, gender and country, moderate the main relationships. They
collect the data from an internet survey of 765 North American, British and Portuguese
potential users of robo-advisor services and find that consumers’ attitudes toward
robo-advisors, together with mass media and interpersonal subjective norms, are the key
determinants of adoption. The influences of perceived usefulness and attitude are slightly
higher for users with a higher level of familiarity with robots; in turn, subjective norms are
significantly more relevant for users with a lower familiarity and for customers from
Anglo-Saxon countries.
Based on their research findings, Belanche et al. (2019) argue that financial institutions
shall design robo-advisors to be easily used by a wide range of consumers. The financial
customer’s level of familiarity with robots shall be considered in the marketing strategies
and tactics. By identifying the key drivers of robo-advisor adoption as well as the
moderating impacts of personal and social-demographic variables, this paper enhances
the understanding of financial consumers’ perceptions regarding the introduction of AI in
the financial technology industries, and extends the knowledge of how subjective norms
such as social influence affect the adoption of AI among a variety of consumers.
Ryu and Ko (2019) examine the decision-making mechanism of speculative investment
behavior as well as its sequential consequences in the Bitcoin context from a dual-systems
perspective. Original data were collected via a survey of 334 participants with experience in
Bitcoin speculative investment, and they apply the partial least squares method to test the
proposed model. They find that speculative investment behavior in the Bitcoin context is
driven by strong impulse and weak self-control, leading to negative consequences. The
extent of the imbalance between the two cognitive systems is greater with the subjective
norm than without it, thus facilitating speculative investment behavior. Noteworthy
differences in the impulse and self-control effects on Bitcoin speculative investment are
found with differences in Bitcoin objective and subjective knowledge.
The speculative nature of Bitcoin as an asset results in asset bubbles and price
destabilization, and many policy makers and stakeholders have worried about this. To our
knowledge, Ryu and Ko (2019) is one of the first attempts to empirically investigate the
decision-making mechanisms used when speculating in Bitcoin. They have provided an
IMDS insightful starting point for practitioners, researchers and policy makers to understand the
119,7 speculative activities of cryptocurrencies, and therefore raise many more additional
questions on the decision-making processes and behaviors in this market.
Blockchain possesses the potential to disrupt and reshape a plethora of industries in the
coming five to ten years. Clohessy and Acton (2019) conduct a comprehensive literature
review and adopt methods of qualitative content analysis to identify the most important
1406 technology-organization-environment blockchain adoption factors. Employing a
multiple-case study of 20 companies based in Ireland, they investigate how the top three
organizational factors identified from the Blockchain literature impacted affected these
companies decision to adopt or not adopt blockchain. They finally identify three patterns:
Top management support and organizational readiness are enablers for blockchain
adoption and large companies are more likely to adopt blockchain than small to
medium-sized enterprises (SMEs).
Although organizational factors are often viewed as the most significant determinants of
information technology innovation adoption in organizations, there is a lack of systematic
inquiry pertaining to blockchain studies from this perspective. This study fills in the gap in
the literature and elucidates the impact of organizational factors on the adoption of
blockchain as well as the adoption of blockchain in Irish companies. Practical and scientific
contributions are also presented.
Lee et al. (2019) investigate linguistic style and message substance influence persuasion
in civic crowdfunding marketplaces in which written narrative pitch become a vital
communication to attract private contributions to public goods and services. Drawing on the
ELM, they operationalize the linguistic style of the narrative pitch as language power and
message substance as issue-relevant argument quality. The logistic regression models are
estimated to assess the impact of both linguistic style and message substance on
crowdfunding decisions. Empirical results show that funding outcomes can be improved
with psychological language dimensions (i.e. positive affective and perceptual language).
However, extensive use of social language does not help project creators to increase their
chance of funding performance; but instead, such language reduces the likelihood of project
success. Additionally, message substance or issue relevant information such as money and
risk language influences funding outcome.
Crowdfunding is unique in interface, funding pattern, product line formation and pricing,
and makes it easier for social entrepreneurs within emergent categories to acquire resources,
instead of appealing to the established aims, thus changing the distribution of initiation
purposes. However, to the best of our knowledge, very few empirical studies investigated the
differential effects of language style and message substance on funding performance of
crowdfunding campaigns. This paper is a pioneer study that draws upon the dual process of
persuasion as a theoretical base to identify a comprehensive set of linguistic style and message
substance and to examine the role of such features in an emerging civic crowdfunding market.
It advances the application of the dual process in ELM by identifying and examining distinct
persuasive cues originating from linguistics styles and message contents.
Wang et al. (2019) apply the self-determination theory to propose a research model to
predict individuals’ intentions toward donation crowdfunding in China, and employ
structural equation modeling to analyze data collected. Empirical evidence indicates that
self-identity and social identity collectively or separately mediate the effect of sense of
self-worth, face concern, moral obligation, perceived donor effectiveness, social interaction
and referent network size in donation intentions.
The findings of Wang et al. (2019) help to provide academics with a more thorough
understanding of the driving forces of individual behavior intention toward donation
crowdfunding in China, and the results are robust to the self-determination theory and identity
theory in the context of donation crowdfunding. The authors also teach the market by providing
suggestions for service providers on how to improve and perfect the functions, and deepen Financial
insights for donation crowdfunding fundraisers regarding how to increase the success rate. technologies
Employing financial technologies to improve financial inclusion is able to empower the
rural farmers, local vendors and SMEs as well as to accelerate the economic growth of
countryside. Aisaiti et al. (2019) investigate China’s rural farmers financing intention of
inclusive finance in the context of financial technology, and examine related drivers like
knowledge of inclusive finance, perceived benefits and perceived risks of ordering finance. 1407
In their framework, the social enterprise embeddedness and digital finance were integrated
into a conceptual model.
The results suggest that the embeddness of social enterprises can significantly reduce
risk perceptions and improve perceived benefits of ordering finance. Moreover, digital
finance as a modifying factor can significantly strengthen the positive correlation between
perceived benefits of ordering finance and financing intention of inclusive financial services
relating to rural population. This paper is among the first attempts to study the
contributions of FinTech toward financial inclusion in rural China, and supports
the effectiveness of digital finance efforts in achieving an inclusive growth. By identifying
the importance of promoting social enterprise participation and FinTech tools to develop
inclusive finance in rural China, they prove that efficiency of FinTech tools can greatly
reduce the existing transaction cost of farmers.

4. Suggestions for future research

This special issue focuses on the interface mechanisms and models of finance and various
emerging digital technologies such as AI, blockchain and crowdfunding. Nevertheless, one
area of research has fallen into neglect, namely, the regulation of FinTech. While FinTech
provide major benefits to businesses, consumers, and economies, they also raise concerns
regarding data privacy, funding security, and access equality (Zetzsche et al., 2017). The
legal and regulatory communities are expected to tackle these emerging challenges and take
measures to be better prepared for future FinTech developments. Regulators shall work
closely with entrepreneurs and startups to understand new FinTech developments and
upcoming obstacles and to help the financial-related industries to address these challenges.
For example, a digital currency called “Libra,” which is defined as a global, digitally native,
reserve-backed cryptocurrency built on the foundation of blockchain technology, has
attracted public attention worldwide. It aims to provide people with capabilities of sending,
receiving, spending, and securing their money, and thus to enable a more inclusive global
financial system (The Libra Association, 2019). While the literature shows that many
traditional monetary theories and business models continue to matter today, the focus has not
been on their adaptions to the fast changing landscape of FinTech.
What are the theoretical and practical implications of Libra? Will this type of emerging
global digital currencies perform as efficiently or even more than fiat currencies? How will they
compete with state-issued digital currencies? What factors will determine their business
applications and value-creating mechanisms? How has the FinTech development driven
regulatory changes? All these questions remain to be explored in the future. We encourage
more academic efforts to study detailed big data and to look into plausible identification
strategies, influencing channels and social consequences of new FinTech products and services.
Another dimension of future research is the design of robo-advising interventions to
meet the needs of financial customers. Practitioners in AI-related investment markets shall
dig deeper into the optimal design of robo-advising interventions tailored to the needs of
different categories of investors with different personality traits, investment experiences
and relationships with FinTech providers (Belanche et al., 2019; D’Acunto et al., 2019).
Where cultures and informal institutions play a significant role in adoption of AI and
robot-advisory is also a novel research area to explore.
IMDS Apart from businesses’ perspectives, a couple of emerging topics related to
119,7 robo-advising, such as whether and how regulators can better communicate and
therefore regulate those services, are still waiting for academic efforts. After all, recent SEC
letters and penalties issued to leading players such as Wealthfront indicated that they are
not exempt from fraudulent operations. Also technically, the potential systematic risk
created by robo services that are driven by largely similar algorithms has not been predicted
1408 or modeled yet. Another concern to be dealt with in the future research is that the diversity
financial markets strive to maintain may be vanished.
Besides, ecosystem settings raise a new set of issues for both researchers and managers to
consider, and Adner and Kapoor (2010) explicitly link the dynamics of value creation and
their implications for value capturing mechanisms to the structure of interdependence in an
innovative firm’s ecosystem. However, this perspective that they propose and test have not been
exploited in the FinTech firms yet. FinTech firms’ choices regarding both business model design
and organization design need to be further explored in the light of combined value creation.
Finally, it is our great pleasure to thank all authors and reviewers for their hard work
and contributions. Without them, there would be no special issue on “Financial
Technologies: Artificial Intelligence, Blockchain, and Crowdfunding.”

References
Adner, R. and Kapoor, R. (2010), “Value creation in innovation ecosystems: how the structure of
technological interdependence affects firm performance in new technology generations”,
Strategic Management Journal, Vol. 31 No. 3, pp. 306-333.
Aisaiti, G., Liu, L., Xie, J. and Yang, J. (2019), “An empirical analysis of rural farmers’ financing
intention of Inclusive Finance in China: the moderating role of digital finance and social
enterprise embeddedness”, Industrial Management & Data Systems, available at: https://doi.org/
10.1108/IMDS-08-2018-0374 (accessed June 26, 2019).
Allison, T.H., Davis, B.C., Webb, J.W. and Short, J.C. (2017), “Persuasion in crowdfunding: an
elaboration likelihood model of crowdfunding performance”, Journal of Business Venturing,
Vol. 32 No. 6, pp. 707-725.
Belanche, D., Casalo, L. and Flavian, C. (2019), “Artificial intelligence in FinTech: understanding robo-
advisor adoption among customers”, Industrial Management & Data Systems, available at:
https://doi.org/10.1108/IMDS-08-2018-0368 (accessed May 21, 2019).
Biais, B., Bisiere, C., Bouvard, M. and Casamatta, C. (2019), “The blockchain folk theorem”, Review of
Financial Studies, Vol. 32 No. 5, pp. 1662-1715.
Böhme, R., Christin, N., Edelman, B. and Moore, T. (2015), “Bitcoin: economics, technology, and
governance”, Journal of Economic Perspectives, Vol. 29 No. 2, pp. 213-238.
Burtch, G., Ghose, A. and Wattal, S. (2013), “An empirical examination of the antecedents and
consequences of contribution patterns in crowd-funded markets”, Information Systems
Research, Vol. 24 No. 3, pp. 499-519.
Butticè, V., Colombo, M.G. and Wright, M. (2017), “Serial crowdfunding, social capital, and project
success”, Entrepreneurship Theory and Practice, Vol. 41 No. 2, pp. 183-207.
Calic, G. and Mosakowski, E. (2016), “Kicking off social entrepreneurship: how a sustainability orientation
influences crowdfunding success”, Journal of Management Studies, Vol. 53 No. 5, pp. 738-767.
Cheah, E. and Fry, J. (2015), “Speculative bubbles in bitcoin markets? An empirical investigation into
the fundamental value of Bitcoin”, Economics Letters, Vol. 130, pp. 32-36.
Chen, M., Wu, Q. and Yang, B. (2019), “How valuable is FinTech innovation?”, Review of Financial
Studies, Vol. 32 No. 5, pp. 2062-2106.
Chong, A., Lim, E., Hua, X., Zheng, S. and Tan, C. (2019), “Business on chain: a comparative study of
five blockchain-inspired business models”, Journal of Association for Information Systems,
forthcoming.
Clohessy, T. and Acton, T. (2019), “Investigating the influence of organizational factors on blockchain Financial
adoption: an innovation theory perspective”, Industrial Management & Data Systems, available technologies
at: https://doi.org/10.1108/IMDS-08-2018-0365 (accessed June 27, 2019).
Colombo, M.G., Franzoni, C. and Rossi Lamastra, C. (2015), “Internal social capital and the attraction of early
contributions in crowdfunding”, Entrepreneurship Theory and Practice, Vol. 39 No. 1, pp. 75-100.
Cong, L.W. and He, Z.G. (2019), “Blockchain disruption and smart contracts”, Review of Financial
Studies, Vol. 32 No. 5, pp. 1754-1797. 1409
Davis, B.C., Hmieleski, K.W., Webb, J.W. and Coombs, J.E. (2017), “Funders’ positive affective reactions
to entrepreneurs’ crowdfunding pitches: The influence of perceived product creativity and
entrepreneurial passion”, Journal of Business Venturing, Vol. 32 No. 1, pp. 90-106.
D’Acunto, F., Prabhala, N. and Rossi, A.G. (2019), “The promises and pitfalls of robo-advising”, Review
of Financial Studies, Vol. 32 No. 5, pp. 1983-2020.
Foley, S., Karlsen, J.R. and Putnins, T.J. (2019), “Sex, drugs, and Bitcoin: how much illegal activity is
financed through cryptocurrencies?”, Review of Financial Studies, Vol. 32 No. 5, pp. 1798-1853.
Goldstein, I., Jiang, W. and Karolyi, G.A. (2019), “To FinTech and beyond”, Review of Financial Studies,
Vol. 32 No. 5, pp. 1647-1661.
Greenberg, J. and Mollick, E. (2017), “Activist choice homophily and the crowdfunding of female
founders”, Administrative Science Quarterly, Vol. 62 No. 2, pp. 341-374.
Griffin, J.M. and Shams, A. (2018), “Is Bitcoin really un-tethered?”, available at: https://papers.ssrn.com/
sol3/papers.cfm?abstract_id=3195066 (accessed June 25, 2018).
Ilyina, A. and Samaniego, R. (2011), “Technology and financial development”, Journal of Money, Credit
and Banking, Vol. 43 No. 5, pp. 899-921.
Kleemann, F., Voß, G.G. and Rieder, K. (2008), “Un(der) paid innovators: the commercial utilization of
consumer work through crowdsourcing”, Science, Technology & Innovation Studies, Vol. 4 No. 1,
pp. 5-26.
Kuppuswamy, V. and Bayus, B.L. (2017), “Does my contribution to your crowdfunding project
matter?”, Journal of Business Venturing, Vol. 32 No. 1, pp. 72-89.
Lee, C.H., Bian, Y., Karaouzene, R. and Suleiman, N. (2019), “Examining the role of narratives in civic
crowdfunding: linguistic style and message substance”, Industrial Management & Data Systems,
available at: https://doi.org/10.1108/IMDS-08-2018-0370 (accessed June 27, 2019).
Mollick, E. (2014), “The dynamics of crowdfunding: An exploratory study”, Journal of Business
Venturing, Vol. 29 No. 1, pp. 1-16.
Nakamoto, S. (2008), “Bitcoin: a peer-to-peer electronic cash system”, available at: https://bitcoin.org/
bitcoin.pdf (accessed April 3, 2017).
Ordanini, A., Miceli, L., Pizzetti, M. and Parasuraman, A. (2011), “Crowd-funding: transforming
customers into investors through innovative service platforms”, Journal of Service Management,
Vol. 22 No. 4, pp. 443-470.
Parhankangas, A. and Renko, M. (2017), “Linguistic style and crowdfunding success among social and
commercial entrepreneurs”, Journal of Business Venturing, Vol. 32 No. 2, pp. 215-236.
Roma, P., Petruzzelli, A.M. and Perrone, G. (2017), “From the crowd to the market: the role of
reward-based crowdfunding performance in attracting professional investors”, Research Policy,
Vol. 46 No. 9, pp. 1606-1628.
Ryu, H. and Ko, K. (2019), “Understanding speculative investment behavior in the Bitcoin context from
a dual-systems perspective”, Industrial Management & Data Systems, available at: https://doi.
org/10.1108/IMDS-10-2018-0441 (accessed August 22, 2019).
Stanko, M.A. and Henard, D.H. (2017), “Toward a better understanding of crowdfunding, openness and
the consequences for innovation”, Research Policy, Vol. 46 No. 4, pp. 784-798.
Stanko, M.A., Fisher, G.J. and Bogers, M. (2017), “Under the wide umbrella of open innovation”, Journal
of Product Innovation Management, Vol. 34 No. 4, pp. 543-558.
IMDS Tapscott, D. and Tapscott, A. (2016), Blockchain Revolution: How the Technology Behind Bitcoin is
119,7 Changing Money, Business, and the World, Penguin, New York, NY.
The Libra Association (2019), “Libra white paper”, available at: https://libra.org/en-us/whitepaper
(accessed May 15, 2019).
Wang, T., Li, Y., Kang, M. and Zheng, H. (2019), “Exploring individuals’ behavioral intentions towards
donation crowdfunding: evidence from China”, Industrial Management & Data Systems,
forthcoming.
1410
Yu, S., Johnson, S., Lai, C., Cricelli, A. and Fleming, L. (2017), “Crowdfunding and regional
entrepreneurial investment: an application of the CrowdBerkeley database”, Research Policy,
Vol. 46 No. 10, pp. 1723-1737.
Zetzsche, D.A., Buckley, R.P., Arner, D.W. and Barberis, J.N. (2017), “From FinTech to TechFin: the
regulatory challenges of data-driven finance”, New York University Journal of Law & Business
( forthcoming); European Banking Institute Working Paper Series 2017 - No. 6; University of Hong
Kong Faculty of Law Research Paper No. 2017/007; University of Luxembourg Law Working
Paper No. 2017-001, available at: https://ssrn.com/abstract=2959925 (accessed April 25, 2018).

Further reading
Chiu, J. and Koeppl, T.V. (2019), “Blockchain-based settlement for asset trading”, Review of Financial
Studies, Vol. 32 No. 5, pp. 1716-1753.
King, M.R. (2015), “Political bargaining and multinational bank bailouts”, Journal of International
Business Studies, Vol. 46 No. 2, pp. 206-222.
Mollick, E. (2016), “Containing multitudes: the many impacts of Kickstarter funding”, The Wharton
School of the University of Pennsylvania, Philadelphia, available at: https://ssrn.com/abstract=2
808000working paper (accessed July 11, 2016).

About the authors

Xiuping Hua is Professor of Finance at Nottingham University Business School China, Director of
UNNC-NFTZ Blockchain Laboratory, Executive Deputy Director of Centre for New Structural
Economics at University of Nottingham Ningbo China, and Secretary-General for Centre for Inclusive
Finance at People’s Bank of China Ningbo Branch and University of Nottingham Ningbo China. Her
research works have been published in the Journal of International Money and Finance, The European
Journal of Finance, Review of Quantitative Finance and Accounting, Journal of Association for
Information Systems, Business History, International Business Review and so on. Xiuping Hua is the
corresponding author and can be contacted at: xiuping.hua@nottingham.edu.cn
Yiping Huang is Professor of Financial Economics at the National School of Development at Peking
University in China. He has served as Member of the People’s Bank of China’s Monetary Policy
Committee, and currently Member of both the China Finance 40 Forum and China Economist 50
Forum. He has published in The Journal of International Economics, The European Journal of Finance,
Oxford Bulletin of Economics and Statistics, China Economic Review, Journal of Comparative Economics
and so on.
Yanfeng Zheng is Associate Professor in Management and Strategy, Faculty of Business and
Economics, University of Hong Kong, China. His research interests primarily revolve around the nexus
of strategic management and entrepreneurship. His work has been published at leading management
and entrepreneurship journals such as Academy of Management Journal (AMJ), Strategic Management
Journal (SMJ), Journal of Business Venturing ( JBV ), Journal of Management Studies ( JMS), Industrial
and Corporate Change (ICC) and Strategic Entrepreneurship Journal (SEJ). He also serves the editorial
board of AMJ, the most prestigious academic journal in management field.

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
 The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm

Artificial
Artificial Intelligence in FinTech: Intelligence in
understanding robo-advisors FinTech

adoption among customers

Daniel Belanche, Luis V. Casaló and Carlos Flavián 1411
Department of Marketing Management and Marketing Research,
Received 29 August 2018
University of Zaragoza, Zaragoza, Spain Revised 6 March 2019
Accepted 28 March 2019

Abstract
Purpose – Considering the increasing impact of Artificial Intelligence (AI) on financial technology (FinTech),
the purpose of this paper is to propose a research framework to better understand robo-advisor adoption by a
wide range of potential customers. It also predicts that personal and sociodemographic variables ( familiarity
with robots, age, gender and country) moderate the main relationships.
Design/methodology/approach – Data from a web survey of 765 North American, British and Portuguese
potential users of robo-advisor services confirm the validity of the measurement scales and provide the input
for structural equation modeling and multisample analyses of the hypotheses.
Findings – Consumers’ attitudes toward robo-advisors, together with mass media and interpersonal subjective
norms, are found to be the key determinants of adoption. The influences of perceived usefulness and attitude are
slightly higher for users with a higher level of familiarity with robots; in turn, subjective norms are significantly
more relevant for users with a lower familiarity and for customers from Anglo-Saxon countries.
Practical implications – Banks and other firms in the finance industry should design robo-advisors to be
used by a wide spectrum of consumers. Marketing tactics applied should consider the customer’s level of
familiarity with robots.
Originality/value – This research identifies the key drivers of robo-advisor adoption and the moderating
effect of personal and sociodemographic variables. It contributes to understanding consumers’ perceptions
regarding the introduction of AI in FinTech.
Keywords Robo-advisors, Artificial Intelligence, Robots, Finance, Technology adoption
Paper type Research paper

1. Introduction
Robots and Artificial Intelligence (AI) are already transforming all kinds of industries,
from manufacturing, to retail and service provision. This technological revolution is
threatening established principles in economy and labor, since automated technology
penetration has been growing at a rate of 20 percent per year (International Federation of
Robotics, 2017) and may replace almost half of current jobs in the next 20 years (Acemoglu
and Restrepo, 2017).
This is also the case in the finance industry, where financial technology (FinTech) is
being revealed as a key element in the strategy of banks and financial start-up companies
( Jung, Dorner, Weinhardt and Pusmaz, 2018). The concept of FinTech goes beyond
e-banking and consumer digitalization and focuses on the development and successful
introduction of innovative technology instruments to meet users’ financial needs and
demands. In this line, AI represents a clear opportunity to advance the transformation of the
finance industry by providing users with greater value and increasing firms’ revenues
(Park et al., 2016). For instance, more than a million customers of Bank of America use a
chatbot named Erica to answer basic banking questions (Rosman, 2018). Another amazing
example is the case of Nao, a small bank teller humanoid that attends customers side by
Industrial Management & Data
side with employees in some branches of the Bank of Tokyo (Marinova et al., 2017). Systems
However, despite these thought-provoking innovations, the most disruptive phenomenon Vol. 119 No. 7, 2019
pp. 1411-1430
within FinTech has been the automated or assisted management of investments by means © Emerald Publishing Limited
0263-5577
of AI, popularly called robo-advisors. DOI 10.1108/IMDS-08-2018-0368
IMDS In contrast to traditional human advisory services, robo-advisors reduce fees and
119,7 provide 24/7 access to finances (Faubion, 2016; Park et al., 2016). Thus, these autonomous
systems are expected to democratize the use of financial advisor services to a wider range of
customers (Sironi, 2016). As a consequence, banks and other finance companies are
launching robo-advisor services as a source of competitive advantage. Indeed, robo-advisors
are currently managing more than $880,000m in assets, and are growing at an annual rate
1412 over 30 percent (Statista, 2019).
Nonetheless, consumer adoption of robo-advisor services “has been slow so far”
( Jung, Dorner, Weinhardt and Pusmaz, 2018, p. 367). As is typical of disruptive innovations in
initial stages, only a small group of early adopters have been willing to rely on this new system
that is replacing traditional practices in finance management services (Laukkanen and
Pasanen, 2008). That is, after attracting early adopters, in order to achieve greater effectiveness
and value companies have been aiming to introduce the service to a greater public that may be
initially hesitant about the worth of such an innovation (Ryu, 2018). Thus, managers need some
guidance on how to successfully implement robo-advisors in order to help retain existing, and
attract potential, customers.
However, despite the opportunities derived from the launching of FinTech AI applications,
research on robo-advisor introduction is very limited. Most studies in the area have focused on
technical or legal issues (Glaser et al., 2019; Ji, 2017), ignoring the customer perspective, though
this would help to extend these services to a greater number of customers. The scant research
on robo-advisory designs ( Jung, Dorner, Weinhardt and Pusmaz, 2018) has highlighted a need
to increase the usability of these systems to facilitate users’ interaction with them.
Nevertheless, given the potential broad expansion of robo-advisors in the finance industry,
there is a need to develop a comprehensive model that better explains the key perceptions and
motivations driving robo-advisor adoption by a wide range of customers.
To do so, and based on well-established technology adoption theories, the authors
propose a framework wherein perceptions about a robo-advisor’s usefulness and ease of use,
together with consumer attitudes, impact the intention to adopt this service.
Complementarily, acknowledging the disruptive change that robo-advisors entail, the
study’s research model also predicts that subjective norms (i.e. social influence) motivate
customers to start using robo-advisors.
As an additional contribution of this work, some moderating variables are considered
that would be particularly relevant in the adoption of these kinds of AI-driven
innovations. Considering that customers vary in their level of familiarity with AI and
robot-based systems (e.g. some users already have experience with AI through services
such as Alexa and products such as Roomba), the study’s framework proposes that users’
familiarity with such technology may play a moderating role. Specifically, it is suggested
that customers with higher familiarity will tend to assign higher value to their own
attitudes and usefulness perceptions, whereas customers with lower familiarity with
robots will base their decision on subjective norms to a greater extent. Following previous
research on technology adoption (Sun and Zhang, 2006), age and gender are also included
as control variables. Finally, to increase the scope of the research to a broader range of
customers, a post hoc analysis deepens understanding of cultural differences by
contrasting the distinctions between the three countries considered in the empirical study:
Portugal, the UK and the USA.
In sum, in order to adapt to the transformation of the finance industry, firms should seek
to better understand customers’ needs and demands so as to succeed in the introduction of
robo-advisors. Given the lack of empirical work explaining robo-advisor adoption, the
research contributes to the literature by:
• analyzing the relevance of key determinants (i.e. utilitarian and social motivations) of
customers’ decision to adopt robo-advisor systems;
 • evaluating possible differences in the adoption process depending on customers’ Artificial
characteristics (i.e. familiarity with robots, age, gender, culture), which may moderate Intelligence in
the relationships in the framework; and FinTech
• improving understanding and discussing consumers’ perceptions about robo-advisor
services in order to guide successful infusion of AI-driven innovations, which will
benefit both companies and the general public.
The remainder of this paper is structured as follows. The next section presents the research
1413
framework, which is followed by the hypotheses formulation. The international empirical
study, data collection process and methodology are then explained. Then results and main
findings are described, and finally, implications of the research for managers and customers,
together with its limitations, and lines for further research are discussed.

2. Literature review
2.1 The challenge of introducing AI in frontline services
The phenomenon of unprecedented growth of AI and robot-based systems across
industries is having a critical impact on the economic, social and labor domains
(Acemoglu and Restrepo, 2017). From a theoretical perspective, Huang and Rust (2018)
described AI as a major source of innovation that will gradually replace human
jobs in the future. More precisely, they predicted that automated technology will develop
mechanical intelligence first, then analytical capacity (e.g. robo-advisors) and, after
some time, intuitive and even empathetic intelligence; this will require workers to
become specialized in tasks that can be less easily accomplished via automation (Huang
and Rust, 2018).
In a similar vein, an emerging field of research is focusing on the challenge of introducing
service innovations involving robots, droids or AI (Singh et al., 2017; Han and Yang, 2018),
with particular attention to those technologies that directly interact with customers in
frontline operations (e.g. physically, online) (Van Doorn et al., 2017). For instance, Singh et al.
(2017) affirmed that customer interaction with organizations is being profoundly disrupted
by intelligent interfaces. Similarly, Grewal et al. (2017) predicted that AI systems (e.g. Alexa,
Cortana, Siri) would directly impact consumer shopping behaviors. Complementarily,
Van Doorn et al. (2017) indicated that technology infusion within service interactions would
depend on the level of human and automated social presence; that is, on the capacity of
frontline robots to engage customers at a social level.
All in all, there is an increasing awareness of the need for firms to introduce AI advances
to progress their management practices and product offerings (Han and Yang, 2018).
By doing so, companies can achieve competitive advantage to better adapt to a market
transformation taking place in the short and medium term. Nevertheless, these theoretical
insights require empirical evidence to guide both customers and managers in the successful
introduction of AI-related services.

2.2 Previous knowledge on robo-advisors and its adoption by customers

Robo-advisors have been defined as “digital platforms comprising interactive and intelligent
user assistance components that use information technology to guide customers through an
automated investment advisory process” ( Jung, Dorner, Glaser and Morana, 2018, p. 81).
The use of robo-advisors by regular customers is relatively simple. First, this technology-
based service assesses the profile of the customer via an initial questionnaire (i.e. goal, risk,
return expectations). It then starts to make specific recommendations or actions about
investment or portfolio rebalancing, as would a human financial advisor, but it does so
autonomously and based on AI.
IMDS Robo-advisors present numerous advantages compared to traditional human advisors,
119,7 such as improving the temporal and ubiquitous accessibility to financial services,
significantly reducing management fees and providing wider investment options based on
systematic and quantitative analyses without ulterior motivations (Faubion, 2016; Park
et al., 2016). As a result, the amount of assets under management by robo-advisors, which, as
noted above, has already surpassed $880,000m, is expected to obtain an annual growth of
1414 31.45 percent in the period 2019–2022 (Statista, 2019). In other words, banks and financial
firms are launching these services to gain competitive advantage and expand their business
to a wider public.
Despite the increasing interest on this phenomenon, literature on robo-advisor
introduction is currently scarce and has frequently focused on the legal complexities
( Ji, 2017) and risk-management aspects (Glaser et al., 2019). The few recent studies on
robo-advisors from a customer approach have described some of the problems related to
widespread adoption of this service by customers. For instance, Jung, Dorner, Weinhardt
and Pusmaz (2018) pointed out that banks are more enthusiastic about FinTech and robo-
advisors than are customers, who often are reluctant to entrust their money to these
kinds of AI-driven platforms. The results of Jung, Dorner, Weinhardt and Pusmaz’s
(2018), experiment which was conducted in a controlled lab setting, suggest that greater
effort is needed to improve robo-advisor design (especially in terms of layout and
usability) in order to boost adoption of this innovation by customers. From a more
technical perspective, Faloon and Scherer (2017) identified personalization as the
distinctive feature of robo-advisors compared to traditional human advisory services.
In turn, Heinrich and Schwabe (2018) concluded that IT-supported advice-giving
processes increase customers’ learning, enabling them to make more informed decision
related to financial products.
Therefore, due to the novelty of robo-advisory services, there is currently a lack of
knowledge about the key determinants of their adoption by regular customers. As the
innovation attempts to democratize advisory services to a wider public (Sironi, 2016), the
current research seeks to close this gap by proposing a broad, comprehensive model that
includes the main global drivers of customer adoption of robo-advisors.

3. Hypothesis formulation
3.1 An extension of the technology acceptance model to understand robo-advisor adoption
Or proposed model is based on the technology acceptance model (TAM) (Davis, 1989; Davis
et al., 1989). TAM is one of the best frameworks by which to understand users’ reactions
toward technological innovations because it is able to explain, to a great extent, consumer
adoption of many innovations. As a result, this model has been widely used in previous
literature in both finance and online settings (Venkatesh and Davis, 2000; Venkatesh et al.,
2003).
However, the simplicity of TAM has been frequently criticized (e.g. Bagozzi, 2007) for
failing to consider other relevant aspects, such as the influence of social processes, which
suggests that attitudes should be complemented with subjective norms, as proposed by the
theory of reasoned action (Fishbein and Ajzen, 1975). Therefore, and similar to the proposals
of Venkatesh and Davis (2000) regarding software introduction, this study broadens TAM
by integrating the influence of subjective norms (i.e. individuals’ perceptions about the
opinions of others) which could be crucial in the adoption of financial robo-advisors by
potential users.
In addition, to better understand this adoption process, the current study considers the
moderating role of individual characteristics. In particular, although robo-advisors represent a
new service linked to more familiar or experienced AI and robot technology (e.g. Alexa), the
level of familiarity may vary among customers (Young et al., 2009). Thus, it is proposed that
customers’ familiarity with robot technology may moderate the main effects of the study’s Artificial
framework. According to previous literature (Sun and Zhang, 2006), the influence of the Intelligence in
antecedents of behavioral intentions may vary due to heterogeneity across users depending FinTech
on personal characteristics such as age or gender; thus, these two variables are also included
as control variables in the model.
The research model is summarized in Figure 1.
1415
3.2 TAM-related hypotheses
Consistent with TAM formulations (Davis, 1989; Davis et al., 1989), its proposed
relationships are first adapted to the context of adoption of financial robo-advisors in order
to explain customers’ behavioral intentions (i.e. “the strength of a person’s willingness to
perform a certain behavior” Belanche et al., 2012).
Specifically, this model proposes that behavioral intentions mainly depend on attitude
(i.e. “the degree to which a person has a favorable or unfavorable evaluation of the behavior
in question” Ajzen, 1991, p. 188) and perceived usefulness (i.e. “the degree to which a person
believes that using a particular system would enhance his or her performance” Davis, 1989,
p. 320). In turn, following a cost-benefit paradigm, attitude is affected by perceived
usefulness and perceived ease of use (“the degree to which a person believes that using a
particular system would be free of effort” Davis, 1989, p. 320). Finally, perceived ease of use
also plays an instrumental role, and positively influences perceived usefulness as
performance may be increased due to the effort saved (Davis et al., 1989). Adapting these
well-established TAM relationships to the present research context, the following
hypotheses are proposed:
H1a. Perceived ease of use of financial robo-advisors has a positive effect on their
perceived usefulness.
H1b. Perceived ease of use of financial robo-advisors has a positive effect on attitudes
toward them.
H1c. Perceived usefulness of financial robo-advisors has a positive effect on attitudes
toward them.

PERCEIVED H1d: +
USEFULNESS
H1c: +

ATTITUDE INTENTION
H1a: + TO USE
H1e: +

PERCEIVED H1b: +
EASE OF USE

H2: +

INTERPERSONAL
MODERATING
INFLUENCE
VARIABLES
SUBJECTIVE
NORMS
FAMILIARITY (H3)
AGE (control)
EXTERNAL Figure 1.
INFLUENCE SEX (control)
Research model
IMDS H1d. Perceived usefulness of financial robo-advisors has a positive effect on the
119,7 intention to use them.
H1e. Attitude toward financial robo-advisors has a positive effect on the intention to
use them.

1416 3.3 The influence of subjective norms

Normative influences have been described in social psychology as “a person’s perception
that most people who are important to her/him think she/he should or should not perform
the behavior in question” (Fishbein and Ajzen, 1975, p. 302). Literature has clearly
established that these norms rely on social information as a kind of social pressure to
perform a particular behavior based on the personal perception of others’ opinions (Taylor
and Todd, 1995; Fishbein and Ajzen, 1975).
The influence of social norms could be of particular interest with reference to customers
facing a disruptive change with limited information (Taylor and Todd, 1995), an in the case
of robo-advisors. In the absence of their own experience or well-formed beliefs, individuals
are influenced by social norms by means of three psychological mechanisms (Belanche et al.,
2012): internalization of messages from expert sources, need to identify with relevant others
by accepting their opinions and compliance with others to achieve social rewards or avoid
social punishments.
In the specific domain of e-finance, Bhattacherjee (2000) proposed that subjective norms
are based on interpersonal sources (e.g. influence of peers and superiors) and external
sources of information (e.g. influence of mass media). This view agrees with previous
research on technology-based innovations, where the opinion of society affects individuals’
behavior by means of comments and behaviors of people and mass media news and reports
related to the innovation (Belanche et al., 2012). Thus, assuming that consumers need
some basic social confirmation and account for social incentives related to the use of
robo-advisors, the following hypothesis is proposed:
H2. Subjective norms about using financial robo-advisors have a positive effect on the
intention to use them.

3.4 Moderating effects

The use of AI in finance represents a disrupting innovation not only from the internal
firm perspective, but also in the eyes of consumers (Singh et al., 2017). Specifically,
robo-advisors are autonomous intelligent systems that provide advice about finance
management – a task that is clearly replacing the job of human employees and
that could involve interacting with AI systems with pseudo social presence (Van Doorn
et al., 2017). Thus, this transformation suggests that consumers need to adapt to the
innovation in order to be willing to engage with the interactions it entails
for the management of their finances. Indeed, there has been increasing attention
toward robotic and AI systems that provide services (e.g. Roomba cleaner,
Alexa assistant), and there is also a growing number of formative courses and
materials dealing with the programming of robots and AI systems designed to directly
interact with humans. Nevertheless, not all individuals have this knowledge or
experience; thus, customers may differ in their level of familiarity with such innovations
(Young et al., 2009). Consequently, this study suggests that individuals’ differences in
their level of familiarity with robotic systems may play a crucial role in the adoption
of robo-advisors.
In this line, previous research has found that customers with low levels of familiarity
with online banking require auxiliary features (e.g. information about investment topics)
compared to customers with higher familiarity, who tend to focus on utilitarian reasons Artificial
(Mäenpää, et al., 2008). Updating these findings to the current context, it is proposed that Intelligence in
familiarity with AI and robots acts as a moderating variable. Users with a higher level of FinTech
familiarity (i.e. previous interaction or training) with such innovations will have first-hand
knowledge, will value such technology in terms of their usefulness, and will have better
attitudes toward them because they have more knowledge about the practical value of
these systems and have a solid personal predisposition about the targeted behavior 1417
(Castañeda et al., 2007). Conversely, consumers with lower familiarity will be more affected
by subjective norms (i.e. others’ opinions) because they have vaguer and indirect knowledge
about robo-advisors (Venkatesh and Davis, 2000). Consequently, the following moderating
effect of familiarity is proposed:
H3. The effects of perceived usefulness and attitude on intention to use robo-advisors are
strengthened for users with higher familiarity with robots, whereas the effect of
subjective norms on intention to use robo-advisors is strengthened for users with
lower familiarity with robots.
For the sake of completeness, this study also considers the moderating influence of two
sociodemographic characteristics, age and gender. Following Sun and Zhang (2006), these
individual characteristics may help better explain the dynamics of adoption processes.
In this respect, previous literature has suggested that older people have more stable beliefs
and are less susceptible to others’ messages compared to younger people (e.g. Hess, 1994).
Similarly, women are more willing than men to take others’ opinions into account when
deciding to use a new technology (e.g. Sun and Zhang, 2006; Venkatesh et al., 2003). As a
result, the influence of the antecedents of intention to use financial robo-advisors may vary
according to these variables.

4. Method
4.1 Data collection
A web survey was used to collect the data for this study; specifically, participants comprised
765 customers recruited via Prolific, a market research company, which enabled us to obtain
a diverse sample in terms of demographic characteristics such as gender (56.99 percent of
participants are female), age ( o25 years 17.39 percent, 25–34 years 34.25 percent, 35–44
years 26.41 percent, 45–54 years 14.25 percent, 55 or more 7.71 percent), income (o$5,000
5.89 percent, $5,000–10,000 13.48 percent, $10,001–25,000 32.98 percent, $25,001–50,000
31,54 percent, W$50,000 16.10 percent), employment situation ( full-time job 57.89 percent,
part-time job 15.58 percent, student 10.73 percent, unemployed 4.58 percent, retired or other
11.78 percent) and country of origin (28.76 percent of participants come from Portugal, 36.34
percent from the UK and 33.20 percent from the USA). The study was targeted to potential
users of robo-advisors with at least some previous experience with online banking. To
develop the web survey and make the most of this method, the study followed
recommendations by Illum et al. (2010), such as keeping it short and guaranteeing the
anonymity of participants.
Specifically, all participants were invited to read a general description regarding
financial robo-advisors. In order to avoid bias due to brand reputation (MacKenzie et al.,
1986), the robo-advisor was not linked to any specific firm, but the phrase “consider that
you have some money for investment and your bank gives you the possibility to use a
robo-advisor” was added. Next, respondents answered the questionnaire, including a
portion on their perceptions of robo-advisors (i.e. perceived usefulness, perceived ease of
use), subjective norms (i.e. social influence and external influence), attitude toward the
use of robo-advisors, their behavioral intention to use them and previous familiarity
with the use of robots. All scales (see Table AI in the Appendix) were based on
IMDS self-reported measures and used seven-point Likert-type response formats, from 1
119,7 (“completely disagree”) to 7 (“completely agree”).

4.2 Measurement validation

The initial set of items proposed to measure the latent constructs came from an in-depth
review of relevant literature pertaining to online banking, e-commerce and the adoption of
1418 new technologies. The measures were adapted from previous scales assessing perceived
usefulness and perceived ease of use (e.g. Davis et al., 1989; Bhattacherjee, 2000),
interpersonal influence and external influence (e.g. Belanche et al., 2012; Bhattacherjee,
2000), attitude (e.g. Taylor and Todd, 1995; Bhattacherjee, 2000), intention to use
(e.g. Mathieson, 1991; Bhattacherjee, 2000) and familiarity (e.g. Flavián et al., 2006; Casaló
et al., 2008). The extensive review helped to ensure the content validity of the scales.
Following Zaichkowsky (1985), the authors also asked a panel of experts about the degree to
which they judged that the items were clearly representative of the targeted construct, in
order to test for face validity. Items that prompted a high level of consensus among the
experts were retained (Lichtenstein et al., 1990).
To confirm the dimensional structure of the scales, this study used confirmatory
factor analysis and employed the statistical software EQS 6.1. First, the factor loadings of
the confirmatory model were verified regarding whether they were statistically
significant (at 0.01) and higher than 0.5 (Steenkamp and Van Trijp, 1991; Jöreskog and
Sörbom, 1993). No item needed to be eliminated, and acceptable levels of convergence,
R2 values, and model fit were obtained ( χ2 ¼ 404.864, 155 df, p o 0.000; Satorra-Bentler
scaled χ2 ¼ 263.251, 155 df, p ¼ 0.05672; NFI ¼ 0.979; NNFI ¼ 0.989; CFI ¼ 0.991;
IFI ¼ 0.991; RMSEA ¼ 0.030; 90% confidence interval (0.024, 0.036)). To assess
construct reliability, this study also checked that values of the composite reliability
indicator ( Jöreskog, 1971) were above the suggested minimum of 0.65 (Steenkamp and
Geyskens, 2006), as can be seen in Table I. To further ensure convergent validity, it was
verified that average variance extracted (AVE) values were greater than 0.5 (see Table I)
and converged on only one construct (Fornell and Larcker, 1981). Finally, regarding
discriminant validity, Table I shows that each construct shared more variance with its
own measures than with the other constructs in the model (Fornell and Larcker, 1981);
that is, for each construct, the square root of the AVE is greater than correlations
among constructs.

4.3 Multidimensionality of subjective norms

Previous literature has suggested that subjective norms may be formed by different social
influences (e.g. Belanche et al., 2012), and has mainly differentiated between interpersonal
and external influences (e.g. Bhattacherjee, 2000). In order to understand whether these two
influences form a multidimensional structure, two alternative models were compared

Construct (1) (2) (3) (4) (5) (6) CR AVE

Perceived ease of use (1) 0.901 0.945 0.812

Perceived usefulness (2) 0.590 0.906 0.948 0.821
Interpersonal influence (3) 0.410 0.538 0.866 0.900 0.750
External influence (4) 0.340 0.440 0.671 0.868 0.902 0.754
Table I.
Construct reliability, Attitude (5) 0.535 0.701 0.593 0.489 0.934 0.954 0.873
convergent Intention to use (6) 0.600 0.826 0.548 0.456 0.834 0.924 0.946 0.854
validity and Notes: Diagonal elements (italic) are the square root of the AVE (variance shared between the constructs and
discriminant validity their measures). Off-diagonal elements are the correlations among constructs
following a rival model strategy (Anderson and Gerbing, 1988). On the one hand, a first- Artificial
order model in which the dimensions were not differentiated was developed (all items Intelligence in
formed a first-order factor). On the other hand, a second-order model (Steenkamp and Van FinTech
Trijp, 1991) with two dimensions measuring subjective norms (i.e. interpersonal and
external influence) was evaluated. A χ2 difference test with one degree of freedom was
performed to determine which model had a better fit (Bagozzi and Dholakia, 2006).
The result of this test confirmed the multidimensionality of subjective norms, since the 1419
χ2 difference was significant ( χ 2d (1) ¼ 738.359; p o0.01). In addition, the remaining fit
indicators were much higher in the second-order model (see Table II) than in the first-order
model, again suggesting the multidimensionality of the construct subjective norms.

5. Results
5.1 Hypotheses test
To test the TAM-related hypotheses (H1a–H1e) and the influence of subjective norms on the
intention to use financial robo-advisors (H2), a structural equation model was developed.
The model fit showed acceptable values ( χ2 ¼ 697.652, 163 df, p o0.000; Satorra-Bentler
scaled χ2 ¼ 462.847, 163 df, p o0.000; NFI ¼ 0.962; NNFI ¼ 0.971; CFI ¼ 0.975; IFI ¼ 0.975;
RMSEA ¼ 0.049; 90% confidence interval (0.044, 0.054)). First, regarding TAM-related
relationships, it was observed that perceived ease of use has a positive influence on both
perceived usefulness (γ ¼ 0.590, p o0.01) and attitude (γ ¼ 0.177, p o0.01), supporting H1a
and H1b, respectively. In turn, perceived usefulness positively affects attitude (β ¼ 0.721,
p o0.01), which supports H1c, but its direct influence on intention to use is not significant
(β ¼ −0.006, p W 0.1); therefore, H1d is not supported. Finally, intention to use is affected by
attitude (β ¼ 0.800, p o0.01), supporting H1e. In sum, the results confirm all the proposed
TAM-related hypotheses at the 0.01 level, except for H1d (direct influence of usefulness on
intention), which is not supported. Similarly, intention to use is affected by subjective norms
(γ ¼ 0.230, p o0.01), supporting H2.
In addition, the proposed framework implies some indirect effects of perceived ease of
use and usefulness. On the one hand, perceived ease of use exerts significant indirect effects
on attitude (0.425, p o0.01) via usefulness and intention to use (0.478, p o0.01) via
usefulness and attitude. On the other hand, perceived usefulness exerts a significant indirect
effect on intention to use (0.577, p o0.01) via attitude. These relationships can largely
explain the dependent variables, perceived usefulness (R2 ¼ 0.348), attitude (R2 ¼ 0.702) and
intention to use (R2 ¼ 0.684). The values found are considerably high, as TAM models
usually predict between 0.44 and 0.57 of the variance in usage intentions (Venkatesh and
Davis, 2000).
H3 proposes a moderating role of familiarity, such that the effects of perceived
usefulness and attitude on intention to use robo-advisors are strengthened for users with
higher familiarity with robots, whereas the effect of subjective norms on intention to use
robo-advisors is strengthened for users with lower familiarity with robots. In addition,
possible moderating effects of age and gender were checked as the influence of the
antecedents of behavioral intentions may vary due to heterogeneity across users depending
on personal characteristics (Sun and Zhang, 2006). To assess these moderating effects,

χ2 NNFI CFI IFI RMSEA 90% interval

Model ( pW 0.05) (W 0.95) (W 0.95) (W 0.95) (o 0.08) RMSEA
Table II.
First order 755.444 (9 df ), p o0.01 0.777 0.866 0.867 0.231 (0.211; 0.251) Fit indices for the
Second order 17.085 (8 df ), p ¼ 0.029 0.999 0.999 0.999 0.018 (0.000; 0.048) multidimensionality
Note: Recommended values for the fit indices are in brackets (see Hooper et al., 2008 for a review) analysis
IMDS a series of multisample analyses were performed. This method is widely employed to
119,7 evaluate differences among groups formed either by respondents’ perceptions (e.g. García
et al., 2008; Algesheimer et al., 2005; Bagozzi and Dholakia, 2006; Belanche et al., 2019) or by
demographic characteristics such as gender or age (e.g. Hernández et al., 2011; Choi et al.,
2005), facilitating comparison between parameters.
Focusing on the moderating role of familiarity – that is, to distinguish between users with
1420 high vs low levels of familiarity – the total sample was divided according to the arithmetic
mean of the moderating variable, with some cases (±0.5 standard deviation) eliminated
around this mean (García et al., 2008). Following this method, the first group consisted of 262
respondents who had a lower familiarity with robots; the second group comprised 277
participants who reported a higher level of familiarity. After generating an individual
structural solution for each group, a significant difference emerged between the groups at the
0.1 level in the relationship between subjective norms and intention to use (see Table III).
As expected, the influence of subjective norms on intention to use the robo-advisor increases
for users with a lower level of familiarity compared to users with a higher level of familiarity.
That is, when consumers lack familiarity with robots, their first-hand experience is low and thus
their decisions may be based on others’ opinions or information disseminated through mass
media (Bhattacherjee, 2001). However, the influences of perceived usability and attitude on
intention to use are not altered by familiarity. Therefore, H3 is only partially supported.
In addition, as aforementioned, this study controlled for potential differences in the
influence of the antecedents of intention to use robo-advisors due to sociodemographic
characteristics such as age or gender (Sun and Zhang, 2006). According to Sun and Zhang
(2006), these personal characteristics may be relevant moderating factors that could
contribute to a better understanding of the acceptance of a new technology. To assess the
possible moderating role of gender, a second multisample analysis was performed,
comparing between males (n ¼ 329) and females (n ¼ 436). In this case, no significant
differences emerged between the groups (see Table IV ). Therefore, the results suggest that
gender does not alter the influence of antecedents of intention to use, which implies that both
men and women tend to base their adoption decision on a similar basis.
Finally, to assess the moderating role of age, the same procedure was used to perform
another multisample analysis. The sample was divided into two groups: participants aged
less than 35 (n ¼ 395) and those aged 35 years or more (n ¼ 370). The age of 35 was used
as a cutoff point as it represents the difference between the millennial generation (broadly

Estimated coefficients
Constraints Low familiarity High familiarity df χ² difference Probability

Perceived usefulness → Intention to use −0.053 0.058 1 0.475 0.491

Table III. Attitude → Intention to use 0.805* 0.758* 1 0.169 0.681
Multisample analysis: Subjective norms → Intention to use 0.287* 0.124** 1 3.091 0.079
familiarity with robots Notes: *,**Significant at p o 0.01 and p o 0.05, respectively

Estimated coefficients
Constraints Males Females df χ² difference Probability

Perceived usefulness → Intention to use 0.066 −0.053 1 0.002 0.964

Table IV. Attitude → Intention to use 0.707* 0.851* 1 0.626 0.429
Multisample analysis: Subjective norms → Intention to use 0.231* 0.226* 1 0.382 0.536
gender Note: *Significant at p o 0.01
understood as born after 1981) and previous generations (Bolton et al., 2013). However, the Artificial
results in Table V indicate that there is no significant difference between the groups. Again, Intelligence in
the empirical results indicate that age does not alter the influence of antecedents of intention FinTech
to use, meaning that that the determinants of adoption have a similar influence for both
younger and older customers.

5.2 Post hoc analysis 1421

Previous studies have suggested that cultural differences may affect the formation of
consumer intentions (e.g. Belanche, Casaló and Guinalíu, 2015) – an aspect that is crucial in
online services that are being spread globally, as is the case for robo-advisors. The sample
comprised participants from three countries (i.e. Portugal, the UK and the USA), which involve
Latin and Anglo-Saxon countries, as well as American and European ones, in order to
evaluate whether the influence of antecedents of intention to use robo-advisors depends on the
participant’s culture.
To this end, and following previous studies that have evaluated path differences among
samples from different countries (e.g. Sultan et al., 2009), a final multisample analysis was
performed considering three groups: participants from Portugal (n ¼ 220), the UK (n ¼ 278)
and the USA (n ¼ 254). The results of this analysis, depicted in Table VI, reveal interesting
differences between countries. Specifically, even though the relationship between usefulness
and intention to use is not significant in any case, this influence is lower in Portugal than in the
UK ( χ 2d (1) ¼ 4.200; po0.05) and the USA ( χ 2d (1) ¼ 19.528; po0.01). In addition, this influence
is higher in the USA compared to the UK ( χ 2d (1) ¼ 9.759; po0.01). Regarding the relationship
between attitude and intention to use, this influence is always significant and is significantly
higher for Portuguese users than for British ( χ 2d (1) ¼ 2.901; po0.1) and American
users ( χ 2d (1) ¼ 16.759; po0.01), as well for Americans compared to British customers
( χ 2d (1) ¼ 13.885; po0.01). Finally, the influence of subjective norms on intention to use is
significantly lower in Portugal than in the USA ( χ 2d (1) ¼ 14.748; po0.01) or the UK
( χ 2d (1) ¼ 8.054; po0.01); this influence is significant in all cases. No difference arises in the
comparison between the USA and the UK for the influence of subjective norms. In spite of
these differences due to participants’ nationality, the parameters estimated for each country
provide consistent results with the model considering the whole sample, with attitude revealed
as the most relevant predictor of behavioral intentions, followed by subjective norms.

Estimated coefficients
χ²
Constraints Less than 35 years 35 Years or more df difference Probability

Perceived usefulness → Intention to use 0.024 −0.047 1 0.165 0.684

Attitude → Intention to use 0.726* 0.869* 1 0.012 0.913 Table V.
Subjective norms → Intention to use 0.297* 0.190* 1 1.523 0.217 Multisample
Note: *Significant at p o0.01 analysis: age

Estimated coefficients
Constraints Portugal UK USA

Perceived usefulness → Intention to use −0.178 −0.020 0.039

Attitude → Intention to use 0.849* 0.790* 0.823* Table VI.
Subjective norms → Intention to use 0.118** 0.259* 0.259* Multisample
Notes: *,**Significant at p o0.01 and p o0.05, respectively analysis: country
IMDS 6. Discussion
119,7 6.1 Conclusions
Among the advance of innovative FinTech, robo-advisors are of particular interest
because of their differential features. Contrary to other initiatives, robo-advisors rely on
AI systems; that is, automated platforms based on analytical intelligence that are
replacing human advisory services. This service innovation needs to be understood under
1422 the expansion of robotic and AI systems, which are likely to gradually replace many
human jobs in the coming years (particularly mechanical, analytical, intuitive and
empathic tasks Huang and Rust, 2018). More precisely, from a consumer perspective,
users need to adapt to new robotic service providers that play the social role traditionally
attributed to a human employee. Recent literature has highlighted that this advance
represents a disruptive innovation that companies need to carefully understand and
integrate in order to achieve a successful transformation in the medium term (Singh et al.,
2017; Van Doorn et al., 2017). In this line, based on TAM, the contribution of the research
framework to previous literature on robo-advisors’ adoption is twofold. First, this study
identifies the key determinants of customers’ decision to adopt robo-advisor systems
focusing on a wide range of customers and considering both traditional antecedents (such
as those proposed in TAM Davis, 1989; i.e. attitude, perceived usefulness, ease of use) as
well as social influences (i.e. subjective norms). Second, this study evaluates possible
differences in the adoption process depending on the customer’s familiarity with robots (a
crucial aspect taking into account that financial robo-advisors represent a disrupting
service innovation) and sociodemographic characteristics (i.e. age, gender, culture), which
may moderate the relationships proposed in the framework. The inclusion of these
moderating factors helps to better explain this phenomenon from a wider scope.
Regarding the first contribution – the identification of key determinants of customers’
intention to use robo-advisors – the results suggest that attitude is the strongest predictor of
behavioral intention to use financial robo-advisors, followed by subjective norms. Therefore,
as proposed by Fishbein and Ajzen (1975), customers are not only motivated by their
favorable or unfavorable evaluation of the use of these services, but also consider the
expectations of others when making such usage decisions. This may be explained by the
fact that, in a new situation for which there is no obvious course of action – as is probably
the case when using financial robo-advisors – customers may need some basic social
confirmation and may seek information from sources around them to better interpret and
navigate it (Wei and Zhang, 2008). In turn, the influence of perceived usefulness on intention
is non-significant. This is consistent with previous literature, as the sample comprised
individuals that did not have direct previous experience with financial robo-advisors.
Previous studies have suggested that the influence of perceived usefulness on intention to
use is more consistent in post-acceptance stages (i.e. after several months using the system
Bhattacherjee, 2001; Casaló et al., 2010). In addition, attitude is positively influenced by
perceived usefulness and perceived ease of use, wherein the influence of the former is
greater. Again, this is consistent with previous literature suggesting that attitude plays a
central role in the adoption of a new technology-based service (Hernández et al., 2009), as in
the case of robo-advisors. On the one hand, literature has found that perceived usefulness
impacts attitude consistently in both pre-adoption and post-adoption stages; on the other
hand, the influence of perceived ease of use seems to become non-significant in later stages
of adoption (Davis et al., 1989; Karahanna et al., 1999; Bhattacherjee, 2001). Taking into
account that preliminary findings on the robo-advisor field have suggested usability as a
relevant factor for customers ( Jung, Dorner, Weinhardt and Pusmaz, 2018), the results of the
current study also advance this knowledge and support that not only perceived usefulness
but also perceived ease of use play a crucial role in shaping customers’ attitudes toward
robo-advisors in early stages of the adoption process.
 Second, the study contributes to the literature by evaluating possible differences in the robo- Artificial
advisor adoption process. In this respect, familiarity is revealed as a key moderating variable to Intelligence in
understand the intention to use this specific type of service. Specifically, the influence of FinTech
subjective norms was found to be greater when participants have a lower familiarity with the
use of robots than when they have a high familiarity. This is also consistent with previous
findings, as consumers who are more familiar with interacting with robots have been shown to
have more direct and realistic first-hand experience (e.g. Fazio and Zanna, 1981). Thus, users 1423
with higher familiarity will tend to value usefulness and personal attitudes to a greater extent,
instead of relying on subjective norms based on others’ opinions. In turn, when familiarity is
low, first-hand information is more limited, leading to consumers’ behavioral intentions toward
robo-advisors being based on others’ opinions to a greater extent, as previously proposed in the
literature (Venkatesh and Davis, 2000). In turn, age and gender do not moderate the influence of
antecedents of intention to use financial robo-advisors, suggesting that there is no technological
divide related to these demographic characteristics. This finding is not surprising considering
that recent studies have found that these variables may not affect the adoption process of a
new technology when it is targeted to a wide population (e.g. Belanche, Casaló and
Pérez-Rueda, 2015). In the present case, anyone who will potentially make investments during
their life is a potential user of financial robo-advisors, as these systems are specifically designed
to simplify and democratize finance management (Sironi, 2016).
Finally, since the sample comprised participants from three different countries (Portugal,
the UK and the USA), this study also contributes to previous research on robo-advisor
adoption by considering the moderating role of culture on this adoption process. The results
suggest that there are relevant differences in the influence of antecedents of intention to use
depending on participants’ culture. In this respect, the influence of attitude seems to be greater
for Portuguese users. This may be explained by differences in Hofstedes’ (2018) cultural
dimensions; compared to the USA and the UK, in Portugal there is very high uncertainty
avoidance, as well as lower rates of individualism and masculinity. Following Hofstede’s
(2018) description of cultural values, in cultures with higher uncertainty avoidance, attitude
may be a crucial predictor of intention because of individuals need to be secured when making
a decision involving uncertainty. In turn, the influence of usefulness is greater in the USA and
the UK than in Portugal. This may be explained by the fact that the USA and the UK are
countries with higher masculinity values, where people’s motivations are more driven by
aspects such as achievement and success (Hofstede, 2018). Thus, previous literature has found
that in countries with higher masculine values, perceived usefulness (as a functional
aspect linked to perceptions of successful performance) determines consumer behavior to a
higher extent compared to in countries with more feminine values (Belanche, Casaló and
Guinalíu, 2015). Finally, in spite of the fact that the USA and the UK score higher in
individualism (Hofstede, 2018), the influence of subjective norms is greater in these countries
than in Portugal. This represents an interesting finding because it reveals that subjective
norms are crucial even in countries where individualistic or performance-based values are
dominant, in contrast to social values. Perhaps American and British customers perceive
others’ opinions about robo-advisors as a kind of opportunity or invitation to participate in a
competition to achieve greater value for their finances (i.e. convenience, profitability).

6.2 Implications for managers and customers

The successful introduction of robo-advisors in the context of AI and FinTech expansion
represents a challenging issue, but also a source of competitive advantage for many firms in
the finance sector (Park et al., 2016). On the one hand, consumers may be reluctant to switch
to an automated robo-advisor, because of its novel and particularly different features
(e.g. lack of human–employee supervision). On the other hand, the advantages of robo-
advisors have ensured that these systems are coming to manage a growing amount of
IMDS assets, revealing an opportunity to increase the market by spreading the service to a wider
119,7 public ( Jung, Dorner, Weinhardt and Pusmaz, 2018).
Our study provides managers with advice about the introduction of robo-advisors to a
wide spectrum of consumers. A relevant finding of the research is that basic demographic
variables do not alter the relationships within the proposed framework, which suggests that
robo-advisors could be targeted to users independently of their age and gender. However,
1424 the strength of the influence of determinants of intention to use varies among participants
from different countries, which suggests that more precise strategies that could be targeted
to each market can be developed.
As the main finding, the study reveals that consumer attitudes toward these systems are
the key driver of the intention to use them; therefore, the design of an easy to use and useful
platform is fundamental to shape users’ favorable predisposition toward the use of robo-
advisors. In order to increase the number of robo-advisor users, financial companies must be
able to assuage customers’ doubts and enhance their perceptions of these systems. For this
purpose, informing customers about the benefits and profitability of using robo-advisors,
allowing them to interact with the robo-advisor platform (even if they do not invest any
money or the investment is very low) or including supporting conversational systems (e.g.
chatbots or digital assistants that simplify human–machine interactions and improve the
interface’s interactivity SAP, 2018) could be of help. This is especially recommended for
countries with high levels of uncertainty avoidance (like Portugal in the present case)
because, in these countries, customers need to be secure when adopting an innovation that
involves uncertainty. That is, customers need to have a very positive opinion about financial
robo-advisors in order to decide to use them.
In turn, subjective norms also have a remarkable impact on robo-advisor adoption. Both
interpersonal comments and mass media information contribute to increasing consumers’
intentions to start using robo-advisors. In this regard, banks and other companies in the industry
should seek to enhance public opinion by providing information about the benefits of robo-
advisors by means of expert users and professional reports. Interestingly, the effect of subjective
norm is even greater in Anglo-Saxon countries (i.e. the UK and the USA), suggesting that even
putting extra effort into advertising, publicity or lobbying would be highly beneficial in these
countries to boost the social approval of robo-advisors for a higher adoption of such systems.
Finally, the results reveal that effects of the determinants of intention to use vary
depending on users’ familiarity with robotic and AI systems (Young et al., 2009). As
theoretically expected (e.g. Venkatesh and Davis, 2000), others’ opinions (i.e. subjective
norms) have a greater impact on users with lower familiarity than on users with higher
familiarity, who already have their own perceptions about robo-advisors. This finding
suggests a twofold strategy to attract users: marketing campaigns could focus on
consumers who already have experience with robots following cross-selling practices (e.g.
agreements with robotic engineering faculties or associations, alliance with companies in the
AI and robot industry such as iRobot, etc.). Indeed, establishing an ongoing alliance with
robot developers would help finance firms to promote breakthrough innovations (Zheng and
Yang, 2015). Alternatively, and considering the relevance of subjective norms for users with
lower familiarity with robots, firms could address more familiar users to spread the message
to less familiar ones by means of social media campaigns or advertisements in which expert
users explain their favorable experiences with robo-advisors.
Turning to customers, it is strongly recommended that they look for information (e.g. reports,
news, etc.) that explain the characteristics (e.g. profitability, advantages, disadvantages, etc.) of
financial robo-advisors. These customers are encouraged to look for other customers’ opinions
and comments (e.g. on social media or on specialized publications on finance). Since it is difficult
to develop expectations before consuming services, the use of second-hand information from
others may be of help in forming realistic expectations about a given service (Casalo et al., 2015).
Therefore, testimony from peers may be useful in forming more precise expectations about Artificial
financial robo-advisors, reducing the perceived risk of using them. Intelligence in
As well, customers should interact with robo-advisor platforms directly in order to FinTech
evaluate their ease of use and usefulness via first-hand experience, which is considered as
more accurate than second-hand experience (Fazio and Zanna, 1981). In this way, their
beliefs about the robo-advisor and, subsequently, their positive or negative evaluation of
robo-advisors (attitude) will be more stable, helping them in their decision making. This is 1425
especially recommended for customers with low levels of familiarity with robots, due to
their lack of first-hand experience with any kind of AI.

6.3 Further research and limitations

Despite these interesting contributions, this work has some limitations that suggest lines for
further research. First, this study focuses on behavioral intentions as the main dependent
variable. Previous research (Venkatesh and Davis, 2000) has confirmed that intention to use
and actual use are habitually highly correlated in the case of volitional behaviors, as per the
current study. Although the study of intention to use helps to understand initial stages of
the adoption process – afterwards, intentions are continuously refined and modified, which
impacts long-term continuance or discontinuance (e.g. Bhattacherjee, 2001) – a longitudinal
study that collects data about the actual use (Bagozzi, 2007) of financial robo-advisors
would be beneficial to further validate the findings of this research. Second, even though the
study analyzed the moderating role of some individual characteristics, other personality
traits (e.g. technology readiness, need for social interaction, etc.) may also moderate the
relationships of the proposed framework. Third, other variables, such us consumers’
personal experience with human-assisted management of their finances, may also be
relevant in this context. Similarly, a customer’s relationship and familiarity with the
company offering the financial robo-advisor – or other variables related to the company (e.g.
reputation) – may affect the adoption process; therefore, future research could focus on these
aspects. Finally, most participants in this research come from Portugal, the UK and the
USA; since the results show significant differences according to country, future studies
could incorporate other cultures (e.g. Asian, Latin American, Jewish, etc.) to obtain a global
understanding of the adoption of financial robo-advisors.

References
Acemoglu, D. and Restrepo, P. (2017), “Robots and jobs: evidence from US labor markets”, available at:
https://voxeu.org/article/robots-and-jobs-evidence-us (accessed July 29, 2018).
Ajzen, I. (1991), “The theory of planned behavior”, Organizational Behavior and Human Decision
Processes, Vol. 50 No. 2, pp. 179-211.
Algesheimer, R., Dholakia, U.M. and Herrmann, A. (2005), “The social influence of brand communities:
evidence from European car clubs”, Journal of Marketing, Vol. 59 No. 3, pp. 19-34.
Anderson, J. and Gerbing, D. (1988), “Structural equation modeling in practice: a review and
recommended two-step approach”, Psychological Bulletin, Vol. 103 No. 3, pp. 411-423.
Bagozzi, R. (2007), “The legacy of the technology acceptance model and a proposal for a paradigm
shift”, Journal of the Association for Information Systems, Vol. 8 No. 4, pp. 244-254.
Bagozzi, R.P. and Dholakia, U.M. (2006), “Open source software user communities: a study of
participation in Linux user groups”, Management Science, Vol. 52 No. 7, pp. 1099-1115.
Belanche, D., Casaló, L.V. and Flavián, C. (2012), “Understanding the influence of social information
sources on e-government adoption”, Information Research, Vol. 17 No. 3, pp. 1-27.
Belanche, D., Casaló, L.V. and Guinalíu, M. (2015), “The effect of culture in forming e-loyalty intentions:
a cross-cultural analysis between Argentina and Spain”, Business Research Quarterly, Vol. 18
No. 4, pp. 275-292.
IMDS Belanche, D., Casaló, L.V. and Pérez-Rueda, A. (2015), “Determinants of multi-service smartcard
119,7 success for smart cities development: a study based on citizens’ privacy and security
perceptions”, Government Information Quarterly, Vol. 32 No. 2, pp. 154-163, available at:
www.sciencedirect.com/science/article/pii/S0740624X15000222
Belanche, D., Casaló, L.V., Flavián, C. and Guinalíu, M. (2019), “Reciprocity and commitment in online
travel communities”, Industrial Management & Data Systems, Vol. 119 No. 2, pp. 397-411.
1426 Bhattacherjee, A. (2000), “Acceptance of e-commerce services: the case of electronic brokerages”, IEEE
Transactions on Systems, Man, and Cybernetics, Systems and Humans, Vol. 30 No. 4, pp. 411-420.
Bhattacherjee, A. (2001), “Understanding information systems continuance: an expectation-
confirmation model”, MIS Quarterly, Vol. 25 No. 3, pp. 351-370.
Bolton, R.N., Parasuraman, A., Hoefnagels, A., Migchels, N., Kabadayi, S., Gruber, T., Loureiro, Y.K.
and Solnet, D. (2013), “Understanding generation Y and their use of social media: a review and
research agenda”, Journal of Service Management, Vol. 24 No. 3, pp. 245-267.
Casaló, L.V., Flavián, C. and Guinalíu, M. (2008), “The role of perceived usability, reputation,
satisfaction and consumer familiarity on the website loyalty formation process”, Computers in
Human Behavior, Vol. 24 No. 2, pp. 325-345.
Casaló, L.V., Flavián, C. and Guinalíu, M. (2010), “Antecedents and consequences of consumer
participation in on-line communities: the case of the travel sector”, International Journal of
Electronic Commerce, Vol. 15 No. 2, pp. 137-167.
Casalo, L.V., Flavian, C., Guinaliu, M. and Ekinci, Y. (2015), “Do online hotel rating schemes influence
booking behaviors?”, International Journal of Hospitality Management, Vol. 49, pp. 28-36.
Castañeda, J.A., Muñoz-Leiva, F. and Luque, T. (2007), “Web acceptance model (WAM): moderating
effects of user experience”, Information & Management, Vol. 44 No. 4, pp. 384-396.
Choi, K.S., Lee, H., Kim, C. and Lee, S. (2005), “The service quality dimensions and patient satisfaction
relationships in South Korea: comparisons across gender, age and types of service”, Journal of
Services Marketing, Vol. 19 No. 3, pp. 140-149.
Davis, F. (1989), “Perceived usefulness, perceived ease of use, and user acceptance of information
technology”, MIS Quarterly, Vol. 13 No. 3, pp. 319-340.
Davis, F., Bagozzi, R. and Warshaw, P. (1989), “User acceptance of computer technology: a comparison
of two theoretical models”, Management Science, Vol. 35 No. 8, pp. 982-1003.
Faloon, M. and Scherer, B. (2017), “Individualization of robo-advice”, The Journal of Wealth
Management, Vol. 20 No. 1, pp. 30-36.
Faubion, B. (2016), “Effect of automated advising platforms on the financial advising market”,
bachelor’s thesis, Sam M. Walton College of Business, University of Arkansas, Fayetteville, AR,
available at: https://core.ac.uk/download/pdf/72841372.pdf (accessed August 8, 2018).
Fazio, R.H. and Zanna, M.P. (1981), “Direct experience and attitude–behavior consistency”, in Berkowitz, L.
(Ed.), Advances in Experimental Social Psychology, Vol. 6, Academic Press, New York, NY, pp. 161-202.
Fishbein, M. and Ajzen, I. (1975), Belief, Attitude, Intention, and Behavior: An Introduction to Theory
and Research., Addison-Wesley, Reading, MA.
Flavián, C., Guinalíu, M. and Gurrea, R. (2006), “The role played by perceived usability, satisfaction and
consumer trust on website loyalty”, Information & Management, Vol. 43 No. 1, pp. 1-14.
Fornell, C. and Larcker, D. (1981), “Structural equation models with unobserved variables and
measurement error”, Journal of Marketing Research, Vol. 18 No. 3, pp. 39-50.
García, N., Sanzo, M.J. and Trespalacios, J.A. (2008), “Can a good organizational climate compensate for
a lack of top management commitment to new product development?”, Journal of Business
Research, Vol. 61 No. 2, pp. 118-131.
Glaser, F., Iliewa, Z., Jung, D. and Weber, M. (2019), “Towards designing robo-advisors for
unexperienced investors with experience sampling of time-series data”, in Davis, F., Riedl, R.,
vom Brocke, J., Léger, P.M. and Randolph, A. (Eds), Information Systems and Neuroscience.
Lecture Notes in Information Systems and Organisation, Vol. 29, Springer, Cham, pp. 133-138.
Grewal, D., Roggeveen, A.L. and Nordfält, J. (2017), “The future of retailing”, Journal of Retailing, Artificial
Vol. 93 No. 1, pp. 1-6. Intelligence in
Han, S. and Yang, H. (2018), “Understanding adoption of intelligent personal assistants: a parasocial FinTech
relationship perspective”, Industrial Management & Data Systems, Vol. 118 No. 3, pp. 618-636.
Heinrich, P. and Schwabe, G. (2018), “Facilitating informed decision-making in financial service
encounters”, Business & Information Systems Engineering, Vol. 60 No. 4, pp. 317-329.
Hernández, B., Jiménez, J. and Martín, M.J. (2009), “Adoption vs. acceptance of e-commerce: two 1427
different decisions”, European Journal of Marketing, Vol. 43 Nos 9/10, pp. 1232-1245.
Hernández, B., Jiménez, J. and Martín, M.J. (2011), “Age, gender and income: do they really moderate
online shopping behaviour?”, Online Information Review, Vol. 35 No. 1, pp. 113-133.
Hess, T.M. (1994), “Social cognition in adulthood: aging-related changes in knowledge and processing
mechanisms”, Development Review, Vol. 14 No. 4, pp. 373-412.
Hofstede, G. (2018), “Hofstede insights. Compare countries”, available at: www.hofstede-insights.com/
product/compare-countries/ (accessed August 26, 2018).
Hooper, D., Coughlan, J. and Mullen, M. (2008), “Structural equation modelling: guidelines for
determining model fit”, Electronic Journal of Business Research Methods, Vol. 6 No. 1, pp. 53-60.
Huang, M.H. and Rust, R.T. (2018), “Artificial intelligence in service”, Journal of Service Research,
Vol. 21 No. 2, pp. 155-172.
Illum, S.F., Ivanov, S.H. and Liang, Y. (2010), “Using virtual communities in tourism research”, Tourism
Management, Vol. 31 No. 3, pp. 335-340.
International Federation of Robotics (2017), “Executive summary world robotics 2017 service robots”,
available at: https://ifr.org/free-downloads/ (accessed March 29, 2018).
Ji, M. (2017), “Are robots good fiduciaries? Regulating robo-advisors under the investment advisers act
of 1940”, Columbia Law Review, Vol. 117 No. 6, pp. 1543-1583.
Jöreskog, K. (1971), “Statistical analysis of sets of congeneric tests”, Psychometrika, Vol. 36 No. 2,
pp. 109-133.
Jöreskog, K. and Sörbom, D. (1993), LISREL 8: Structural Equation Modeling with the SIMPLIS
Command Language, Scientific Software International, Chicago, IL.
Jung, D., Dorner, V., Glaser, F. and Morana, S. (2018), “Robo-advisory”, Business & Information Systems
Engineering, Vol. 60 No. 1, pp. 81-86.
Jung, D., Dorner, V., Weinhardt, C. and Pusmaz, H. (2018), “Designing a robo-advisor for risk-averse,
low-budget consumers”, Electronic Markets, Vol. 28 No. 3, pp. 367-380.
Karahanna, E., Straub, D.W. and Chervany, N.L. (1999), “Information technology adoption across time:
a cross-sectional comparison of pre-adoption and post-adoption beliefs”, MIS Quarterly, Vol. 23
No. 2, pp. 183-213.
Laukkanen, T. and Pasanen, M. (2008), “Mobile banking innovators and early adopters: how they differ
from other online users?”, Journal of Financial Services Marketing, Vol. 13 No. 2, pp. 86-94.
Lichtenstein, D.R., Netemeyer, R.G. and Burton, S. (1990), “Distinguishing coupon proneness from value
consciousness: an acquisition–transaction utility theory perspective”, Journal of Marketing,
Vol. 54 No. 3, pp. 54-67.
MacKenzie, S.B., Lutz, R.J. and Belch, G.E. (1986), “The role of attitude toward the ad as a mediator of
advertising effectiveness: a test of competing explanations”, Journal of Marketing Research,
Vol. 23 No. 2, pp. 130-143.
Mäenpää, K., Kale, S.H., Kuusela, H. and Mesiranta, N. (2008), “Consumer perceptions of Internet
banking in Finland: the moderating role of familiarity”, Journal of Retailing and Consumer
Services, Vol. 15 No. 4, pp. 266-276.
Marinova, D., de Ruyter, K., Huang, M.H., Meuter, M.L. and Challagalla, G. (2017), “Getting smart:
learning from technology-empowered frontline interactions”, Journal of Service Research, Vol. 20
No. 1, pp. 29-42.
IMDS Mathieson, K. (1991), “Predicting user intentions: comparing the technology acceptance model with the
119,7 theory of planned behavior”, Information System Research, Vol. 2 No. 3, pp. 173-239.
Park, J.Y., Ryu, J.P. and Shin, H.J. (2016), “Robo advisors for portfolio management”, Advanced Science
and Technology Letters, Vol. 141, pp. 104-108.
Rosman, C. (2018), “Mad about Erica: why a million people use bank of America’s chatbot”, American
Banker, New York, NY, available at: www.americanbanker.com/news/mad-about-erica-why-a-
1428 million-people-use-bank-of-americas-chatbot (accessed August 8, 2018).
Ryu, H.S. (2018), “What makes users willing or hesitant to use Fintech?: the moderating effect of user
type”, Industrial Management & Data Systems, Vol. 118 No. 3, pp. 541-569.
SAP (2018), “Trend report 2018. Emerging technology trends”, available at: www.sap.com/
documents/2018/04/02112563-007d-0010-87a3-c30de2ffd8ff.html (accessed February 24, 2019).
Singh, J., Brady, M., Arnold, T. and Brown, T. (2017), “The emergent field of organizational frontlines”,
Journal of Service Research, Vol. 20 No. 1, pp. 3-11.
Sironi, P. (2016), FinTech Innovation: From Robo-Advisors to Goal Based Investing and Gamification,
Wiley, Chichester, available at: https://doi.org/10.1002/9781119227205
Statista (2019), “Robo-advisors worldwide report”, available at: www.statista.com/outlook/337/100/
robo-advisors/worldwide (accessed February 12, 2019).
Steenkamp, J.B. and Geyskens, I. (2006), “How country characteristics affect the perceived value of a
website”, Journal of Marketing, Vol. 70 No. 3, pp. 136-150.
Steenkamp, J.B. and Van Trijp, H.C.M. (1991), “The use of LISREL in validating marketing constructs”,
International Journal of Research in Marketing, Vol. 8 No. 4, pp. 283-299.
Sultan, F., Rohm, A.J. and Gao, T.T. (2009), “Factors influencing consumer acceptance of mobile
marketing: a two-country study of youth markets”, Journal of Interactive Marketing, Vol. 23
No. 4, pp. 308-320.
Sun, H. and Zhang, P. (2006), “The role of moderating factors in user technology acceptance”,
International Journal of Human–Computer Studies, Vol. 64 No. 2, pp. 53-78.
Taylor, S. and Todd, P.A. (1995), “Understanding information technology usage: a test of competing
models”, Information Systems Research, Vol. 6 No. 2, pp. 144-176.
Van Doorn, J., Mende, M., Noble, S.M., Hulland, J., Ostrom, A.L., Grewal, D. and Petersen, J.A. (2017),
“Domo arigato Mr Roboto: emergence of automated social presence in organizational
frontlines and customers’ service experiences”, Journal of Service Research, Vol. 20 No. 1,
pp. 43-58.
Venkatesh, V. and Davis, F. (2000), “A theoretical extension of the technology acceptance model: four
longitudinal field studies”, Management Science, Vol. 46 No. 2, pp. 186-204.
Venkatesh, V., Morris, M., Davis, G. and Davis, F. (2003), “User acceptance of information technology:
toward a unified view”, MIS Quarterly, Vol. 27 No. 3, pp. 425-478.
Wei, L. and Zhang, M. (2008), “The impact of internet knowledge on college students’ intention to
continue to use the internet”, Information Research, Vol. 13 No. 3, pp. 1-21, available at: http://
InformationR.net/ir/13–3/paper348.html (accessed September 6, 2012).
Young, J.E., Hawkins, R., Sharlin, E. and Igarashi, T. (2009), “Toward acceptable domestic robots:
applying insights from social psychology”, International Journal of Social Robotics, Vol. 1 No. 1,
pp. 95-108.
Zaichkowsky, J.L. (1985), “Measuring the involvement construct”, Journal of Consumer Research,
Vol. 12 No. 3, pp. 341-352.
Zheng, Y. and Yang, H. (2015), “Does familiarity foster innovation? The impact of alliance partner
repeatedness on breakthrough innovations”, Journal of Management Studies, Vol. 52 No. 2,
pp. 213-230.
Appendix Artificial
Intelligence in
FinTech
Perceived ease of use
EASE_OF_USE1 Learning to use robo-advisors would be easy for me
EASE_OF_USE2 I would find it easy to manage investments using robo-advisors
EASE_OF_USE3 It would be easy for me to become skillful at using robo-advisors 1429
EASE_OF_USE4 I would find robo-advisors easy to use
Perceived usefulness
USEFUL1 Using robo-advisors would improve my performance in managing investments
USEFUL2 Using robo-advisors would improve my productivity in managing investments
USEFUL3 Using robo-advisors would enhance my effectiveness in managing investments
USEFUL4 I would find robo-advisors useful in managing investments
Interpersonal influence
INT_INF1 My peers/colleagues/friends think that I should use robo-advisors for managing
investments
INT_INF2 People I know think that using robo-advisors is a good idea
INT_INF3 People I know could influence me to try out robo-advisors for managing investments
External influence
EXT_INF1 I have read/seen news reports that using robo-advisors is a good way of managing
investments
EXT_INF2 The popular press depicts a positive sentiment related to using robo-advisors
EXT_INF3 Mass media reports influence me to try out robo-advisors for managing investments
Attitude
ATT1 Using robo-advisors for managing investments seems like a good idea
ATT2 I like the idea of using robo-advisors for managing personal investments
ATT3 Using robo-advisors for implementing my investments seems like a wise idea
Intention to use
INT_USE1 I intend to use robo-advisors for managing investments
INT_USE2 Using robo-advisors for managing investments is something I would do
INT_USE3 My intention is to use robo-advisors rather than any human financial advisor
Familiarity with the use of robots
FAM1 I have worked with or studied robotic artificial intelligence
FAM2 Throughout my life I have had experience interacting with robots Table AI.
FAM3 I am familiar with robots or robot contents (texts, audiovisuals, etc.) Measurement scales

About the authors

Daniel Belanche holds a PhD in Business Administration and is Assistant Professor of Marketing at
the University of Zaragoza, Spain. His research interests include e-commerce and e-government, social
identities, emotions and purchase decision processes. His studies have been published in international
journals such as Journal of Interactive Marketing, Journal of Services Management, Government
Information Quarterly, Information & Management, Industrial Management & Data Systems or
Psychology & Marketing.
Luis V. Casaló holds a PhD in Business Administration and is Associate Professor of Marketing at
the University of Zaragoza, Spain. His research interests include consumer behavior, social networks,
e-commerce and e-government. His studies have been published in international journals such as
International Journal of Electronic Commerce, Journal of Business Research, Information &
Management, Industrial Management & Data Systems, International Journal of Information
Management, Internet Research or Tourism Management.
Carlos Flavián holds a PhD in Business Administration and is a Professor of Marketing at the
University of Zaragoza, Spain. His research has been published in several academic journals,
IMDS specialized in marketing ( Journal of Interactive Marketing, Journal of Business Research, European
119,7 Journal of Marketing, Psychology & Marketing, etc.), tourism (Tourism Management, International
Journal of Contemporary Hospitality Management, International Journal of Hospitality Management,
etc.) and new technologies (Industrial Management & Data Systems, Information & Management,
Internet Research, Computers in Human Behavior, International Journal of Electronic Commerce, etc.).
He is the editor of the Spanish Journal of Marketing-ESIC and Associate Editor of the Services
Industries Journal. Carlos Flavián is the corresponding author and can be contacted at:
1430 cflavian@unizar.es

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
 The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm

Bitcoin
Understanding speculative speculative
investment behavior in the investment
behavior
Bitcoin context from a
dual-systems perspective 1431

Hyun-Sun Ryu Received 10 October 2018

Revised 30 April 2019
College of Software, Accepted 6 July 2019
Sungkyunkwan University, Suwon, The Republic of Korea, and
Kwang Sun Ko
Department of Financial IT, Kim & Chang,
Seoul, The Republic of Korea

Abstract
Purpose – The purpose of this paper is to examine users’ decision-making mechanism of speculative
investment behavior and its sequential consequences in the Bitcoin context from a dual-systems perspective.
Design/methodology/approach – Original data were collected via a survey of 334 participants with
experience in Bitcoin speculative investment. The partial least squares method was used to test the proposed model.
Findings – Speculative investment behavior in the Bitcoin context is driven by strong impulse and weak
self-control, leading to negative consequences. The extent of the imbalance between the two cognitive
systems is greater with the subjective norm than without it, thus facilitating speculative investment behavior.
Noteworthy differences in the impulse and self-control effects on Bitcoin speculative investment are found
with differences in Bitcoin objective and subjective knowledge.
Originality/value – This study is the first attempt to empirically investigate users’ decision-making
mechanism used when speculating in Bitcoin.
Keywords Bitcoin, Speculative investment behaviour, Dual-systems perspective, Subjective norm,
Bitcoin knowledge
Paper type Research paper

1. Introduction
Bitcoin, the most prominent cryptocurrency, has attracted substantial public attention and
been prominently addressed by media, venture capitalists and financial and governmental
institutions. Proposed by Nakamoto (2008), Bitcoin is a peer-to-peer electronic payment
system that can be sent directly from one party to another without central authority or
intermediaries. Although Bitcoin was originally defined as an alternative currency, it has
also been viewed as a speculative asset because of its considerable user demand. With its
price volatility and expected large return, Bitcoin increased in popularity and became
known to the public (Blau, 2017).
Orcutt (2015) pointed out that Bitcoin has not seen broad use in retail transactions
despite its increasing popularity and public attention. Hur et al. (2015) explained Bitcoin’s
incompetence as an alternative currency against the conventional tools of trade. Previous
studies (Baur et al., 2018; Blau, 2017; Cheah and Fry, 2015; Hur et al., 2015) have provided
empirical evidence that Bitcoin is mainly used as a speculative asset and not as an
alternative currency. That is, customer participation in Bitcoin is indeed speculative, and
Bitcoin users perceive Bitcoin only as a speculative investment tool (Baur et al., 2018; Hur Industrial Management & Data
et al., 2015). Nevertheless, the use of Bitcoin as an investment vehicle has been overlooked Systems
Vol. 119 No. 7, 2019
pp. 1431-1456
This work was supported by the Ministry of Education of the Republic of Korea and the National © Emerald Publishing Limited
0263-5577
Research Foundation of Korea (NRF-2019S1A5A8034813). DOI 10.1108/IMDS-10-2018-0441
IMDS because many researchers continue to focus on the value of Bitcoin as an alternative
119,7 currency and medium of exchange.
In late 2017, many people were tempted to speculate in Bitcoin in ways that challenged
their self-control or willpower because of the expected large return derived from Bitcoin
price dynamics. With their expectation of a large return, some Bitcoin users lost their
self-control and became impulsive, ultimately engaging in speculative investment. At that
1432 point, Bitcoin users’ speculative investment behavior might not have been intended or
planned but rather impulsive. Because the previous IS approaches to examine planned
behaviors (e.g. theory of planned behavior (TPB) (Ajzen, 1991) and theory of reasoned action
(Ajzen and Fishbein, 1977)) are inadequate to explain speculative investment behavior in the
Bitcoin context, a new approach is needed.
In addition, Bitcoin is not regulated for risk mitigation or governance requirements,
leading to substantial risks in the market (Bohr and Bashir, 2014; Moore and Christin, 2013).
The speculative nature of Bitcoin as an asset results in asset bubbles and price
destabilization (Shiller, 1981; Stein, 1987), causing harmful effects on individuals and
society. Policy makers and regulators are seriously concerned about the potential risks of
Bitcoin and have thus tried to prohibit speculative trading in Bitcoin to block its negative
effects on individuals and society. For example, during the Bitcoin peak in 2017, the People’s
Bank of China ordered its financial institutions to stop providing banking or funding to any
activity related to any cryptocurrency (Yu, 2018). Although the prior literature has
investigated the technological, economic and regulatory aspects of Bitcoin (Abramova and
Böhme, 2016; Böhme et al., 2015; Glaser et al., 2014), only a few studies have examined
Bitcoin usage behavior and its sequential consequences from a user perspective.
To bridge that research gap, we have investigated the cognitive processes of Bitcoin
users as they made decisions that led to speculative investment behavior, along with the
sequential consequences of their choices. We employed a dual-systems perspective to better
understand speculative investment behavior in the Bitcoin context. This study investigates
three main issues:
RQ1. What decision-making mechanism leads Bitcoin users into speculative investment
behavior?
RQ2. Does the decision-making mechanism of Bitcoin users differ depending on the
subjective norm and Bitcoin knowledge?
RQ3. Is speculative investment behavior associated with negative consequences in the
Bitcoin context?
To answer those research questions, we collected empirical data from 334 Bitcoin users with
experience Bitcoin speculation in the year 2017 in South Korea. South Korea was an
appropriate place to observe the Bitcoin speculative phenomenon and collect our data
because it was third worldwide in terms of Bitcoin trading volume in 2017. The collected
data were used to investigate the effects of impulse and self-control on speculative
investment behavior in the Bitcoin context from a dual-system perspective. We then
determined how impulse and self-control factors differ depending on the subjective norm
and Bitcoin knowledge. Finally, the relationship between speculative investment behavior
and negative consequences was examined.

2. Theoretical background
2.1 Bitcoin
Bitcoin has emerged as a fascinating phenomenon in the finance markets. As the first
decentralized cryptocurrency, Bitcoin was designed as an electronic peer-to-peer payment
system depending neither on central authorities nor on intermediaries to offer diverse
benefits to customers over traditional fiat money (Nakamoto, 2008). With Bitcoin, Bitcoin
transaction parties exchange digital money instantly and directly, independent of their speculative
geographical location, with low transaction fees and a certain degree of anonymity investment
(Abramova and Böhme, 2016; Böhme et al., 2015; Van Alstyne, 2014). In addition, “mining”
of Bitcoin, that is, the creation and transaction of Bitcoins, is entirely dependent on its users behavior
(Nakamoto, 2008). However, Bitcoin raises concerns, such as being used by criminals to
launder money or trade illicit goods (e.g. illegal drugs) because of its anonymity (Böhme 1433
et al., 2015). The innovative features of Bitcoin open new possibilities in the financial and
information technology (IT) fields, but it faces shortcomings and potential risks that
question its solidity and widespread use. Despite of its popularity and the public attention it
has received, the role of Bitcoin as an alternative currency or speculative asset is being still
debated (Baur et al., 2018; European Central Bank, 2012; Glaser et al., 2014; Hur et al., 2015;
Yermack, 2015). The previous literature has focused on those two conflicting yet coexisting
roles of Bitcoin: alternative currency and speculative asset.
The first approach regards Bitcoin as an alternative currency. This approach focuses
on Bitcoin as a decentralized peer-to-peer payment system and conceives of it as an
alternative to government-backed currencies (Baur et al., 2018; Nakamoto, 2008).
Advocates of Bitcoin as an alternative currency were attracted by its global and
government-free design, which eliminated chargeback risks, reduced transaction costs
and cross-border transactions fees, and increased security, as well as offering full support
for mobile devices and the possibility of purchasing special goods (Baur et al., 2015; Baur
et al., 2018). Although some researchers have argued that Bitcoin will be an attractive
alternative to traditional currencies, their general attitude toward Bitcoin as an alternative
currency is not entirely positive. A few researchers (Antonopoulos, 2014; Bohr and Bashir,
2014) limited the potential of Bitcoin as an alternative currency to specific. Grinberg (2012)
and Kristoufek (2015) noted that Bitcoin seems to be the new, revolutionary payment
method for micropayments because credit card fees are hardly profitable for small
transactions. However, Hur et al. (2015) reported the incompetence of Bitcoin as an
alternative currency because its level of network effects is low. Orcutt (2015) pointed out
that Bitcoin has seldom been used in retail transactions. Baur et al. (2018) indicated that
societies and businesses are still far from adopting and using Bitcoin as an alternative
currency in daily transactions. Brezo and Bringas (2012) analyzed the risks of
cryptocurrencies, taking Bitcoin as an example. They highlighted that Bitcoin is
vulnerable to speculation and misinformation and that no regulatory body oversees the
market. Grant and Hogan (2015) explained six risks taken on by companies or individuals
who use Bitcoin: price volatility, exchange rate risk, future legislation, theft or loss,
third-party reliability and e-commerce vulnerabilities. Bohr and Bashir (2014) claimed that
appropriate legal policies and regulations are essential to overcoming the existing risks
and threats slowing down the adoption and use of Bitcoin as a medium of exchange.
The second approach considers Bitcoin as a speculative asset. Given the low adoption
and use of Bitcoin as a payment system, this approach asserts that Bitcoin is less of a
currency and more of a speculative investment. In particular, the unusual rise and fall in the
price of Bitcoin leads to price volatility, which raises the question of whether Bitcoin
functions as an alternative currency at all. These researchers argue that Bitcoin has lost its
function as a currency against conventional payment tools and is only understood as a
speculative asset in the Bitcoin market. Surowiecki (2011) and Spenkelink (2014) claimed
that Bitcoin is hoarded by speculators rather than used as medium of exchange. Hur et al.
(2015) discovered the degree to which the participation of Bitcoin users depends on
speculative opportunities in the Bitcoin market. Baur et al. (2018) found that users only
receive Bitcoin and never send it on to others. Those studies thus reveal that
Bitcoin is primarily used as a speculative investment vehicle, due to its high volatility and
IMDS large returns. However, Ron and Shamir (2013), Blau (2017) and Gandal et al. (2018) found
119,7 that Bitcoin price dynamics result from price manipulation not speculation. Gandal et al.
(2018) found suspicious manipulations in a Bitcoin trade in late 2013 that exerted important
real effects on the price leap. Hur et al. (2015) claimed that the nature of Bitcoin is
speculative, but they went on to suggest that its speculative nature might not dominate user
behaviors entirely. Urquhart (2016) asserted that Bitcoin is a relatively new investment
1434 asset, making it similar to a new alternative investment in an emerging market. He
concluded that strong inefficiency is expected during the infancy of Bitcoin but predicted
that it will become efficient over time. Blau (2017) did not find any direct association
between Bitcoin and speculation and found no evidence that speculative trading contributed
to the unprecedented rise and subsequent crash in Bitcoin’s value. The relationship between
speculative investment behavior and price volatility in the Bitcoin context is still debated in
the literature.

2.2 Dual-systems perspective on impulse and self-control

Dual-systems theory has been applied in numerous contexts, including consumer behavior
research, such as the service evaluation of healthcare users (Rolland et al., 2013), risk
attitudes of private investors (Fehr and Hari, 2014), excessive mobile phone use (Soror et al.,
2015) or social media use (Turel and Qahri-Saremi, 2016) and employee acceptance of IS
(Bhattacherjee and Sanford, 2006). Watts (2015) suggested dual-systems theory as a
promising approach for studying IS use phenomena.
Dual-systems theory posits that a human’s thought process is guided by two structurally
different systems: reflexive and reflective (Lieberman, 2007; Soror et al., 2015; Strack and
Deutsch, 2004; Turel and Qahri-Saremi, 2016). The reflexive system processes information
using cognitive and affective associations that trigger a rapid behavioral response. By
contrast, the reflective system processes information based on rules that exert control over
actions (Soror et al., 2015). The reflexive system is characterized as fast, impulsive,
automatic and non-conscious, whereas the reflective system is slow, controlled, conscious
and analytical.
The dual-systems perspective is particularly concerned with how the two systems
compete to determine behavior (Strack and Deutsch, 2004), suggesting that decisions can be
made reflexively through spontaneous and effortless processing (reflexive system) or
reflectively through intentional and systematic processing (reflective system). The two
systems do not occur in isolation but interact in decision making. Most of the time, the two
systems work in harmony. However, when the two systems come into conflict, a tug-of-war
between them influences behavior (Hofmann et al., 2009; Soror et al., 2015). That is, while the
reflexive system generates an impulse to engage in (or avoid) a behavior, the reflective
system reflects on that impulse and behavior to determine whether they are aligned with a
person’s long-term goals and decide whether the behavior should be enacted or inhibited
(Strack and Deutsch, 2004).
Examples of the two systems in the literature are automatic affective reactions vs self-
control (Friese and Hofmann, 2009; Hofmann et al., 2009; Turel and Qahri-Saremi, 2016),
habit vs self-regulation (Soror et al., 2015), heuristic processing vs systematic processing
(Chaiken, 1980; Zhang et al., 2014) and automatic stereotyping vs suppression (Devine,
1989). To investigate the decision-making process pertaining to speculative investment
behavior in the Bitcoin context, we use impulse to represent the reflexive system and
self-control to represent the reflective system.

2.3 Speculative investment behavior in the Bitcoin context

Bitcoin prices experienced a dramatic increase, from $5.28 at the beginning to its peak price
of $18,900 in December 2017, and then fell to below $6,500 in February 2018. Despite the
potential risks and widespread misinformation about Bitcoin, many people were obsessed Bitcoin
with Bitcoin investment. People thought that buying Bitcoin to sell it later is actually worth speculative
more than they think and thus they spent a substantial amount of money and time with the investment
expectation of great profits. Bitcoin users searched for speculative opportunities in the
Bitcoin market. The prior literature pointed out that speculators distort market prices behavior
through their activities, thereby harming people who depend on the commodities being
traded (Angel and McCabe, 2009). Speculators do not produce anything useful and are 1435
parasites upon society. Thus, speculative behavior results in negative consequences for
both individuals and society. Therefore, we define speculative investment behavior in the
Bitcoin context as a behavior of buying or selling Bitcoin by spending time and money in
the expectation of profiting from market fluctuations.
Dual-systems theory has been particularly useful in explaining the etiology of unplanned
or problematic consumer behaviors. Prior researchers (Soror et al., 2015; Spada, 2014; Turel
et al., 2014; Turel and Qahri-Saremi, 2016; Turel and Serenko, 2010) have investigated
unplanned user behavior and problematic IT use behavior, such as internet overuse, mobile
e-mail and Facebook addiction, and problematic mobile phone and SNS use. The dual-
systems perspective indicates that an imbalance between the reflexive and reflective
systems can result in a strong impulse to engage in unplanned behavior with ineffective
self-control. Speculative investment behavior in the Bitcoin context occurred when users
were exposed to an external cue (e.g. high volatility and large return) that generated a strong
urge to impulsively invest in Bitcoin, which weakened self-control. Strong impulses
reinforce people’s drive to engage in speculative investment behavior, making it difficult to
resist temptation and ultimately leading to speculative investment behavior (Turel and
Qahri-Saremi, 2016). Therefore, speculative investment behavior in the Bitcoin context can
be explained as an unplanned behavior of Bitcoin usage driven by an imbalance between the
reflexive and reflective systems.
Although many researchers and regulators have expressed concern about Bitcoin
speculation, the sequential consequences of speculative investment behavior in the Bitcoin
context remain underexplored. The nature of speculation might lead Bitcoin investors to
experience distraction and interference in their work and social activities, negative emotions
and financial ruin. Speculative investors are also exposed to significant potential risks
through Bitcoin’s technical design, such as threats from malicious software and criminal
usage, hacking or loss through its anonymity and legal and regulatory risk (Abramova and
Böhme, 2016; Cheah and Fry, 2015; Grant and Hogan, 2015). Hence, Bitcoin as a speculative
asset motivates this research to enrich the existing literature on unplanned behavior with
Bitcoin, namely, Bitcoin speculation.

3. Research model and hypotheses

Based on dual-systems theory, this study regards Bitcoin speculative investment behavior
influenced by the outcome of the tug-of-war between the two systems. In this study, the
reflexive system is represented as an impulse, and the reflective system is represented as
self-control. The impulse can be either good or bad depending on the situation (Soror et al.,
2015; Turel and Serenko, 2010), but when the impulse conflicts with already established
goals, self-control needs to be engaged (Hofmann et al., 2009). Therefore, this study proposes
a research model in which Bitcoin speculative investment behavior is influenced by the
impulse and self-control simultaneously, which leads to negative consequences. The
research model then hypothesizes that the effects of impulse and self-control differ
depending on the subjective norm and Bitcoin knowledge (i.e. objective and subjective
knowledge (SK)). Figure 1 shows the proposed model.
An impulse arises when a latent motivation meets an activating stimulus that is suitable
for satisfying this (Friese and Hofmann, 2009). Impulses occur automatically and without
IMDS Subjective
119,7 Norm

Impulse H3a
H1
to invest Bitcoin H3b

1436
Bitcoin
Negative
Speculative
H5 Consequences
Investment

H4a–H4c
Self-control H2
H4b–H4d

Bitcoin Knowledge
Figure 1. - Objective Knowledge
Research model - Subjective Knowledge

effort from an individual. Loewenstein (1996) indicated that an impulse typically possesses a
strong incentive value consisting of a primitive hedonic reaction to a tempting stimulus.
Hofmann et al. (2009) referred to an impulse as people’s tendency to perform an unplanned
behavior, often an urge to approach or act on a temptation at hand. An unconstrained
impulse makes people do something unconsciously until it has naturally come to a bad end
(e.g. reaching the bottom of a potato-chip bag while dieting) (Hofmann et al., 2009). Most
unconstrained impulsive behaviors interfere with long-term goals or generate interpersonal
conflict at a certain point (Bogg and Roberts, 2004; Tangney et al., 2004). According to
Harden and Tucker-Drob (2011), impulse refers to the reflexive system’s output, defined in
the present study as people’s tendency to act on behavioral impulses without planning or
considering potential consequences.
Self-control describes an individual’s ability to control or regulate their emotions and
behaviors (Hofmann et al., 2009). An individual’s self-control reflects on the consequences
of a behavioral choice and attempts to prevent it if those consequences are deemed
damaging or socially unacceptable (Hofmann et al., 2009). Self-control is particularly
manifested in people’s restraint standards, that is, long-term standards about how
behavior should be regulated in a given domain of life (Hofmann et al., 2009).
Self-controlled individuals are more adept than their impulsive counterparts at regulating
their behavioral, emotional and attentional impulses to achieve long-term goals. In this
study, self-control, which is from the reflective system, is defined as people’s tendency to
override or inhibit undesired behavioral tendencies (such as impulses) and refrain from
acting on them, based on Tangney et al. (2004).
Social cognitive theory (Bandura, 1997) suggests that a person’s self-regulatory
mechanism influences their level of self-control and a lack of self-control leads to various
types of problematic behaviors and social maladaptation. Collins and Lapp (1992) pointed
out that individuals who possess the ability to resist a temptation develop strong
self-control, which mitigates unplanned behavior. People with high self-control are good at
controlling and overriding their impulses, leading to a low level of unplanned
behaviors, whereas those with low self-control more often act on their impulses, leading to
a high level of unplanned behavior (Friese and Hofmann, 2009). Therefore, impulse
makes it more likely that investors will engage in Bitcoin speculative investment behavior.
Simultaneously, self-control makes investors consciously aware of their behavior,
preventing or inhibiting Bitcoin speculative investment behavior. Consequently, the Bitcoin
following hypotheses are developed: speculative
H1. Impulse is positively associated with Bitcoin speculative investment behavior. investment
H2. Self-control is negatively associated with Bitcoin speculative investment behavior.
behavior
When people adopt or use technological innovations, the social context of the decision
makers should not be ignored. If the social context is in favor of adopting and using a 1437
technology, then it plays an important role in the decision process (Webster and Trevino,
1995). Bass (1969) emphasized that innovation adoption and use are significantly
influenced by the pressures of the social system. The importance of the subjective norm as
an attitude toward technology use has previously been established in the IS context. TPB
(Ajzen, 1991) suggests that the subjective norm is a significant determinant influencing
planned behaviors such as technology adoption and usage. In TPB, the subjective norm
refers to beliefs about the expectations of important referent others and users’ motivations
in complying with those expectations (Chiou, 1998). Based on Fishbein and Ajzen (1975),
we account for the social context by including the subjective norm, defined as user
perceptions about whether most people who are important to them think that they should
invest in Bitcoin.
Users exposed to greater social pressure will more impulsively invest in Bitcoin as a
speculative asset because they want to belong among their referent people and promote
others’ goals. Baur et al. (2015) found that some people unfamiliar with the Bitcoin process
attempted to use or invest in Bitcoin merely because of a friend’s recommendation. Glaser
et al. (2014) also found that Bitcoin users’ participation in speculation was positively
influenced by positive news but not influenced at all by serious negative news, such as
thefts and hacking. Liu and Tsyvinski (2018) argued that consumer activities on search
forums, such as Google and social medial sites (e.g. Twitter), are related to the price of
Bitcoin. They reported that an increase in keyword searches for Bitcoin led to an increase in
the Bitcoin price. Based on those prior studies, the subjective norm might be associated with
speculative investment behavior in the Bitcoin context.
However, given that such speculative investment is regarded as an unplanned behavior,
a new role for the subjective norm is required to explain our proposed model. Rook and
Fisher (1995) found that the social norms related to products or services moderated the
effects of consumers’ impulsive tendencies. They indicated that consumers can experience
social encouragement or discouragement when the urge to engage in a behavior arises.
Moreover, Sawang et al. (2014) pointed out that the effect of perceived behavioral control on
intention to use technology differs depending on the subjective norm. The subjective norm
motivates users to find a way to fit in with relevant others by changing their self-control
(Markus and Kitayama, 1991). Therefore, we propose that the subjective norm promotes an
impulse and simultaneously mitigates self-control to allow individuals to comply with
others’ Bitcoin speculative investment behavior. The strengthened impulse and weakened
self-control of Bitcoin investors trigger them to actively engage in highly Bitcoin speculative
investment behavior. As such, we propose the following hypotheses:
H3a. The subjective norm strengthens the effect of impulse on Bitcoin speculative
investment behavior.
H3b. The subjective norm weakens the effect of self-control on Bitcoin speculative
investment behavior.
The level of consumers’ product knowledge influences their decision-making behavior
(Brucks, 1985; Park et al., 1994). Consumers’ product knowledge is divided into objective
knowledge (OK) and SK. OK refers to accurate, product-related information stored in
IMDS memory, whereas SK refers to people’s perceptions of what or how much they know about a
119,7 product. Although SK is strongly related to OK, the two do not always coincide (Brucks,
1985; Park et al., 1994). OK is usually measured via testing procedures under the supervision
of an impartial third party, whereas SK can be measured using self-assessments. In other
words, OK describes what an individual actually knows, whereas SK reflects an individual’s
degree of confidence in their knowledge (Brucks, 1985; Chiou, 1998).
1438 The importance of OK in wise financial decision making is unquestionable (Hadar et al.,
2013). Financial education programs typically focus on increasing consumers’ OK. In the
Bitcoin context, Henry et al. (2018) highlighted that many people had heard of Bitcoin, but only
a few people had adopted and used it because of a lack of Bitcoin knowledge. Hadar et al.
(2013) found that consumers with a high level of OK about financial instruments can
deter their willingness to pursue a risky investment. OK about financial decisions usually
diminishes the urge to engage in speculation even when user confidence and willingness to
take risks are considered. That is, OK restricts impulse and facilitates self-control, leading
people to manage their risky investments. Therefore, we hypothesize that:
H4a. The effect of impulse on Bitcoin speculative investment behavior is weaker in the
high OK group than in the low OK group.
H4b. The effect of self-control on Bitcoin speculative investment behavior is stronger in
the high OK group than in the low OK group.
People’s behaviors are strongly influenced by their confidence in their ability to perform them
(Chiou, 1998). In uncertain and complex choice settings, SK has been shown to be a stronger
motivation of behavior than OK (Hadar et al., 2013; Mishra and Kumar, 2011). Hadar et al.
(2013) pointed out that willingness to pursue a risky investment increases when SK is high.
Bitcoin users with a high level of SK are more likely to use stereotypical information or
heuristics that provide simplistic cues in complex decision-making settings (Lee and Lee,
2009). This attitude toward Bitcoin investment can promote the effect of impulse and
overshadow the effect of self-control, leading to Bitcoin speculation. Thus, the effect of impulse
on Bitcoin speculative investment behavior will be stronger, and the effect of self-control will
be weaker when Bitcoin users have high SK. Thus, we hypothesize the following:
H4c. The effect of impulse on Bitcoin speculative investment behavior is stronger in the
high SK group than in the low SK group.
H4d. The effect of self-control on Bitcoin speculative investment behavior is weaker in
the low SK group than in the high SK group.
Prior studies have shown that impulse triggers strong urges to engage in a behavior that
will be problematic if performed (Hofmann et al., 2009; Strack and Deutsch, 2004). In this
study, Bitcoin speculative investment behavior is regarded as the unplanned behavior of
Bitcoin use driven by an imbalance between the reflexive and reflective systems.
Considering the nature of speculation and the potential risks of Bitcoin, Bitcoin speculative
investment behavior could lead to negative outcomes such as financial loss, distraction and
interference in work and social activities, negative emotions, and low work performance
(Fox and Moreland, 2015; Moqbel and Kock, 2018). Policy makers and regulators are
concerned about the individual negative outcomes that derive from a high level of Bitcoin
speculative investment and deem them socially unacceptable. Thus, several countries, such
as China and Korea, established strict regulations to hinder that behavior. However, only a
few studies have clearly investigated whether Bitcoin speculative investment behavior
actually leads to negative consequences for individual users.
In this study, we define negative consequences as the extent to which an individual
experiences personal, social and professional problems as a result of Bitcoin speculative
investment behavior (Haagsma et al., 2013; Soror et al., 2015). A high level of speculative Bitcoin
investment behavior is expected to correlate positively with inappropriate investing and speculative
thus with an increase in negative consequences such as financial loss, negative emotions, investment
distraction from one’s work, and trouble in social activities and relationships. As a result,
the following hypothesis is developed: behavior
H5. Bitcoin speculative investment behavior is positively associated with negative
consequences. 1439

4. Research methodology
4.1 Measurement
To achieve content validity, we developed our measurement items based on an intensive
literature review. Based on the extant innovation and IS literature, we developed
comprehensive multiple-item measures of impulse, self-control, subjective norm, SK, Bitcoin
speculative investment behavior and negative consequences. To test users’ OK about
Bitcoin, 14 questions with three multiple-choice answers (True/False/Don’t know) were
developed. Measures of Bitcoin speculative investment behavior were also developed to
assess the potential for Bitcoin speculation; it contains four general questions about
overspending of time and money. Negative outcome (Caplan, 2010; Haagsma et al., 2013)
was assessed by four questions about financial loss, psychological trouble, trouble in work
and social activities and degree of difficulty in managing life overall. Because respondents
often underreport negative behaviors and consequences and over-report positive behaviors,
we statistically controlled for the presence of social desirability bias by using the short form
of the Marlowe–Crowne social desirability scale (Reynolds, 1982). Age, gender, education
and annual income, which have all been found to potentially influence Bitcoin speculative
investment behaviors, were used as control variables in this study.
Before conducting the main survey, a pre-test was performed to examine the reliability
and validity of the instruments. The pre-test involved 30 respondents with experience
investing in Bitcoin in 2017. The pre-test results led to a significant refinement and
restructuring of the questionnaire. The initial face and internal validity of the measures
were also established. The measures were evaluated using a seven-point Likert-type scale,
ranging from “extremely low” (1) to “extremely high” (7). The structure of the measurements
used, and the relevant studies are shown in Table AI (Appendix).

4.2 Sample and data collection

In this study, we targeted Bitcoin users with experience in Bitcoin speculation in the year
2017 in South Korea because from 2017 to early 2018 is the heaviest period of Bitcoin
speculation in which the Bitcoin price rose sharply from $894 to $18,900 and then promptly
fell below $6,500 and Bitcoin trading volume in the year 2017 in South Korea was active
enough to rank third worldwide. Survey questionnaires were distributed and collected for
three weeks in April 2018, when the price of Bitcoin had just fallen after its peak in 2017.
E-mail invitations were first sent to the targeted participants. We continued sending
invitations until 1,000 respondents accepted the offer to participate. Respondents, who all
had Bitcoin investment experience over the previous three months, were asked to respond to
the entire questionnaire. To ensure that the respondents fully understood the survey
context, an initial screening question was asked to determine whether they had experience
using and investing in Bitcoin. After the main survey, a total of 359 responses were received,
of which 25 were classified as outliers and eliminated. Thus, 334 responses were used in our
analysis, indicating a usable response rate of 33.4 percent. In our sample, most respondents
(95.2 percent) had Bitcoin investment experience during 2017 and early 2018, the heaviest
period of Bitcoin speculation in the cryptocurrency trading market, as shown in Table I.
IMDS Division Freq. Percent (%) Division Freq. Percent (%)
119,7
Gender Annual income
Men 188 56.30 Below $20 51 15.30
Women 146 43.70 $20–39.9 113 33.80
Total 334 100 $40–59.9 102 30.50
User characteristics $60–79.9 41 12.30
1440 Early adopters 253 75.70 $80–100 12 3.60
Late adopters 81 24.30 Above $100 15 4.50
Total 334 100 Total 334 100
Age Period of investment
20–29 63 18.90 From 2018 28 8.40
30–39 92 27.50 From 4/4 in 2017 107 32.00
40–49 72 21.60 From 3/4 in 2017 89 26.60
50–59 61 18.30 From 2/4 in 2017 43 12.90
60+ 46 13.80 From 1/4 in 2017 51 15.30
Total 334 100 From 2016 12 3.60
From 2015 or before 4 1.20
Total 334 100
Education Frequency of investment
Under high school 8 2.40 Daily 56 16.80
High school 35 10.50 Weekly 144 43.10
College/associate 33 9.90 Monthly 87 26.00
Bachelor 217 65.00 Every 3 mon. 26 7.80
Master 36 10.80 Every 6 mon. 9 2.70
PhD 5 1.50 Once a year 4 1.20
Total 334 100 Almost not use 8 2.40
Total 334 100
Bitcoin use for other purpose
As a means of payment for goods services 76 22.8
As a means of cross-border money transfers 13 3.8
As a store of value 108 31.4
As a means of earning mining revenue 55 16
Table I. Do not use 92 26.7
Sample characteristics Total 334 100

Therefore, our sample was adequate for investigating speculative investment behavior
among Bitcoin users.
Table I summarizes the characteristics of the respondents. Our sample predominantly
consisted of men (56.3 percent) age 30–39 (27.5 percent) who regarded themselves as early
adopters and had bachelor’s degrees (65.0 percent) and $20,000–$39,900 in annual income.
In terms of frequency, respondents invested in Bitcoin weekly (43.1 percent), monthly
(26.0 percent), or daily (16.8 percent). In terms of Bitcoin use for other purpose than
speculative means, of the respondents, 31 percent used Bitcoin as a store of value, followed
by those who did not use Bitcoin for other purposes (26 percent) and those who used it as a
means of payment for goods or services (22.7 percent), as a means of earning mining
revenue (16.0 percent), and as a means of cross-border money transfer (3.8 percent).

5. Analysis and results

The partial least squares (PLS) method was adopted to examine the proposed model and its
hypotheses. The PLS is recommended for predictive research models with an emphasis on
theory development (Fornell and Bookstein, 1982). Given that the current study is an initial
attempt to advance a theoretical model that investigates the cognitive decision-making Bitcoin
process involved in Bitcoin speculative investment behavior, we chose PLS because of its speculative
appropriateness for exploratory science (Chin, 1998). Smart PLS version 3.00 was used to investment
analyze the measurement and structural models.
behavior
5.1 Measurement model
To validate our measurement model, we began by assessing the content, convergent and 1441
discriminant validity of the constructs (Hair et al., 1998). First, to test content validity, we
examined the process by which scale items were generated (Straub, 1989). That process
confirmed the content validity of our scales. Second, we evaluated convergent validity
examining the Cronbach’s α, composite reliability (CR) and average variance extracted
(AVE) for each construct (Barclay et al., 1995).
As shown in the Panel A of Table II, the Cronbach’s α and CR values were higher than
the recommended level (0.7) (Chin et al., 2003). The AVE values of our measures were higher
than 0.5 (Fornell and Larcker, 1981), thereby supporting convergent validity. In addition, the
discriminant validity of our instrument was assessed using the square root of the AVE
(Fornell and Larcker, 1981). As shown in the Panel A of Table II, the square root of the AVE
of every construct in the measurement model exceeds all off-diagonal correlations between
the focal constructs and all other constructs in the model. Moreover, each of the
measurement items loaded higher than 0.70 (Hair et al., 1998). Overall, the measurement
model was strongly supported by the gathered data and was suitable for further analysis.
The multicollinearity of all variables was assessed using variance inflation factors
(VIFs). The VIF values for the variables were acceptable, ranging from 1.033 to 1.361. Those
results and the correlation indicated that multicollinearity was not a serious issue with this
research model (Neter et al., 1996). Finally, we checked for possible common method
variance (CMV ) using Harman’s single-factor test (Podsakoff et al., 2003). The results
validated that the seven factors were extracted with an eigenvalue that was greater than one
and that no general factor was apparent in the unrotated factor structure, indicating no
excessive CMV.

5.2 Structural model

5.2.1 Hypothesis tests. Based on an adequate measurement model and a suitably low level of
multicollinearity, all proposed hypotheses were tested using PLS. The path coefficients, t-
values, and R2 were used to test the structural model (Figure 2). As illustrated in Figure 2, all
paths within the research model were supported at the 0.01 level. The proposed model
accounted for 50.0 percent of the variance in Bitcoin speculative investment behavior and
54.8 percent of the variance in negative consequences.
Figure 2 demonstrates that the impulse to invest in Bitcoin exerted a significant
positive effect on speculative investment behavior (β ¼ 0.307, p o 0.01). Thus, H1 was
supported. Self-control was negatively related to speculative investment behavior
(β ¼ −0.230, p o 0.01), providing support for H2. The results also reveal that the effect of
impulse was greater than that of self-control, indicating that Bitcoin investors had
impulsively engaged in speculative investment behavior. The results show that the
moderating effect between impulse and the subjective norm on Bitcoin speculative
investment behavior (β ¼ 0.199, p o 0.01) was smaller than the single effect of impulse
(β ¼ 0.307, p o 0.01). The subjective norm did not strengthen but rather weakened the
effect of impulse on speculative investment, not supporting H3a. Moreover, the subjective
norm positively influenced self-control regarding Bitcoin speculative investment behavior
(β ¼ 0.174, p o 0.01), whereas self-control alone exerted a negative influence (β ¼ −0.230,
p o 0.01). The moderating effect between the subjective norm and self-control was
 119,7

analysis
IMDS

1442

Table II.
Results of the
measurement model
(a) Assessment of reliability and validity
Construct Item Cronbach’s α CR AVE SAVE Loading t-Statistic
Impulse IP1 0.921 0.944 0.808 0.899 0.905** 67.395
IP2 0.903** 66.190
IP3 0.871** 48.688
IP4 0.915** 80.005
Self-control SC1 0.785 0.875 0.700 0.837 0.828** 23.783
SC2 0.885** 54.190
SC3 0.794** 20.786
Subjective norm SN1 0.896 0.927 0.762 0.873 0.864** 40.784
SN2 0.884** 51.387
SN3 0.864** 42.061
SN4 0.880** 55.313
Speculative investment behavior SI1 0.866 0.909 0.713 0.845 0.797** 33.324
SI2 0.861** 54.566
SI3 0.847** 46.909
SI4 0.872** 59.685
Negative consequence NC1 0.957 0.967 0.853 0.924 0.901** 67.094
NC2 0.907** 76.713
NC3 0.934** 122.534
NC4 0.935** 119.297
NC5 0.940** 119.191
(b) Correlations of variables
Construct Means (SD) 1 2 3 4 5 6 7 8 9 10 11
1. Gender 1.437 (0.497)
2. Age 2.805 (1.316) 0.039
3. Education 3.757 (0.919) −0.109* −0.066
4. Income 2.686 (1.235) −0.191** 0.177* 0.184**
5. Social desirability 4.167 (0.592) −0.030 −0.081 0.050 0.140*
6. Impulse 3.574 (1.319) 0.087 0.038 −0.051 0.085 0.272**
7. Self-control 5.036 (1.020) 0.091 0.081 0.051 −0.103 0.135* −0.259**
8. Subjective norm 3.812 (1.248) 0.123* 0.075 −0.031 0.120* 0.310** 0.400** −0.046
9. Objective knowledge 0.396 (0.236) −0.309** −0.176** 0.258** 0.075 −0.092 −0.268** 0.035 −0.336**
10. Subjective knowledge 4.516 (1.020) −0.218** −0.167** 0.129* 273** 0.203** 0.056 0.001 0.266** 0.400**
11. Speculative investment 3.650 (1.335) −0.036 −0.069 −0.001 0.214** 0.246** 0.537** −0.352** 0.455** −0.094 0.296**
12. Negative consequence 2.881 (1.501) 0.096 0.076 −0.069 0.121* 0.344** 0.541** −0.263** 0.407** −0.341** 0.021 0.644**
Notes: CR: composite reliability; AVE: average variance extracted; SAVE: square root of AVE. *p o0.05; **p o0.01
 Annual Social Bitcoin
Gender Age Education
Income Desirability
speculative
Impulse 0.307*** –0.049 –0.094** –0.006 –0.041 0.342***
investment
0.081*
to invest Bitcoin (5.514) (t = 1.218) (t = 2.358) (t = 0.144) (t = 1.889)
0.143***
–0.031 (t = 0.963) (t = 7.944)
(t = 0.892) behavior
(t = 3.574)

0.037 0.064
(t = 0.736)

Speculative
(t = 1.163)

Negative
1443
0.199***
(4.082) Investment Consequences
Behavior 0.554***
(12.015)
R 2 = 50.0 R 2 = 54.8

Self-control –0.230***
(5.233) 0.174***
(4.383)

Subjective Figure 2.
Norm Results of
hypothesis tests for
the total group
Notes: *p< 0.10; **p<0.05; ***p< 0.01

significant, but the subjective norm weakened the effect of self-control on speculative
investment behavior, thus supporting H3b. Finally, we tested the relationship between
Bitcoin speculative investment behavior and negative consequences. We found that
Bitcoin speculative investment behavior had a significant positive effect on negative
consequences (β ¼ 0.554, p o 0.01). Therefore, H5 was supported.
Among the five control variables, education was related to neither speculative
investment behavior nor negative consequences (Figure 2). However, annual income had a
positive and significant effect on Bitcoin speculative investment behavior (β ¼ 0.081,
p o0.10), and social desirability was positively related to negative consequences (β ¼ 0.336,
p o0.01). Age had a significant negative effect on speculative investment behavior
(β ¼ −0.094, p o0.05) and a significant positive effect on negative consequences (β ¼ 0.143,
p o0.01). These results showed that young investors with a high annual income engaged
the most in Bitcoin speculative investment behavior, and old investors with high social
desirability experienced the highest level of negative consequences, within the total group.
5.2.2 Moderation test of Bitcoin knowledge. Based on the previous literature (Brucks,
1985; Chiou, 1998; Hadar et al., 2013; Park et al., 1994), Bitcoin knowledge was divided into
OK and SK. To test the hypothesized moderation effects of Bitcoin knowledge, we used a
median-split method to divide subjects into high- and low-knowledge groups (OK:
Median ¼ 7.0, SD ¼ 3.3; SK: Median ¼ 4.1, SD ¼ 1.0). As shown in Table III, our user
classification produced four groups: the high OK group (n ¼ 113), the low OK group
(n ¼ 221), the high SK group (n ¼ 207) and the low SK group (n ¼ 127).

Objective knowledge
Low (0) High (1) Total

Subjective knowledge Low (0) 93 34 127 Table III.

High (1) 128 79 207 User classification by
Total 221 113 334 Bitcoin subjective and
Note: n ¼ 334 objective knowledge
IMDS As indicated in Table IV, the effects of impulse and self-control on Bitcoin speculation varied
119,7 depending on the two types of Bitcoin knowledge. Specifically, the high OK group presented a
weaker coefficient path for impulse (β ¼ 0.189, po0.05) than the low OK group (β ¼ 0.385,
po0.01), providing support for H4a. The effect of self-control on the speculative investment
behavior in the high OK group (β ¼ −0.253, po0.05) was greater than that in the low OK group
(β ¼ −0.194, po0.01). Thus, H4b was supported. In other words, the high OK group had higher
1444 self-control than the low OK group. Within the high OK group, the subjective norm strengthened
the effect of impulse (β ¼ 0.223, po0.05) and weakened the effect of self-control (β ¼ 0.209,
po0.10) on Bitcoin speculative investment behavior. In the low OK group, the moderating effect
of the subjective norm on impulse was not significant on speculative investment behavior,
whereas that on self-control was positively significant (β ¼ 0.203, po0.01). Bitcoin speculative
investment behavior had a significant positive effect on negative consequences in both the high
and low OK groups (β ¼ 0.566, po0.01; β ¼ 0.609, po0.01, respectively), but the low OK group
experienced greater negative consequences than the high OK group.
The impulse to invest in Bitcoin was stronger in the high SK group (β ¼ 0.328, po 0.01)
than in the low SK group (β ¼ 0.215, p o0.05), which supports H4c. Moreover, the high SK
group had weaker self-control (β ¼ −0.252, p o0.01) than the low SK group (β ¼ −0.294,
p o0.01). Therefore, H4d was supported. In the high SK group, the effects of the subjective
norm on both impulse and self-control were positive and significant on speculative
investment behavior (β ¼ 0.190, p o0.05; β ¼ 0.177, p o0.01, respectively). In the low SK
group, the subjective norm did not strengthen the effect of impulse, but it did weaken the
effect of self-control on speculative investment behavior (β ¼ 0.182, p o0.10). The negative
consequences from speculative investment were greater in the high SK group (β ¼ 0.543,
p o0.01) than in the low SK group (β ¼ 0.529, p o0.01).
To further validate the proposed model and hypotheses, we conducted an ad hoc test by
Bitcoin user types because user levels of both OK and SK varied. As shown in Table V,

Objective knowledge Subjective knowledge

High (n ¼ 113) Low (n ¼ 221) High (n ¼ 207) Low (n ¼ 127)

IP → SI 0.189** (2.014) 0.385*** (5.963) 0.328*** (5.000) 0.215** (2.068)

SC → SI −0.253** (2.464) −0.194*** (3.920) −0.252*** (5.110) −0.294*** (2.743)
IP × SN → SI 0.223** (1.966) 0.075 (1.082) 0.190** (2.480) 0.217 (1.225)
SC × SN → SI 0.209* (1.884) 0.203*** (3.701) 0.177*** (3.854) 0.182* (1.647)
SI → NC 0.566*** (6.234) 0.609*** (10.754) 0.543*** (10.452) 0.529*** (6.734)
Gender → SI −0.114 (1.168) −0.001 (0.016) −0.051 (1.116) 0.024 (0.299)
Age → SI −0.142 (0.623) −0.048 (1.068) −0.085 (1.478) −0.041 (0.585)
Education → SI −0.055 (0.700) 0.018 (0.494) −0.009 (0.190) −0.009 (0.127)
Annual income → SI 0.112 (1.126) 0.025 (0.497) 0.030 (0.487) 0.140 (1.853)
Social desirability → SI 0.010 (0.043) 0.102* (1.716) 0.048 (0.799) −0.111 (0.556)
Gender → NC 0.042 (0.545) −0.007 (0.154) 0.089 (1.957) 0.026 (0.585)
Age → NC 0.176** (2.269) 0.088* (1.824) 0.077 (1.575) 0.207*** (3.004)
Education → NC −0.152* (1.816) 0.032 (0.879) −0.073 (1.591) 0.020 (0.296)
Annual income → NC 0.011 (0.130) −0.058 (1.080) 0.025 (0.437) −0.102 (1.250)
Social desirability → NC 0.340* (1.709) 0.284 (5.064) 0.374*** (7.051) −0.200 (0.826)
R2
Table IV. SI 0.338 0.626 0.559 0.378
PLS analysis results NC 0.473 0.580 0.612 0.395
of moderation tests for Notes: IP: impulse; SC: self-control; SN: subjective norm; SI: speculative investment behavior; NC: negative
Bitcoin knowledge consequence. *p o0.10; **p o0.5; ***p o 0.01
 Expertise Well-Informed Self-confident Stranger
Bitcoin
(n ¼ 79) (n ¼ 34) (n ¼ 128) (n ¼ 93) speculative
High in OK/ High in OK/ Low in OK/ Low in OK/
investment
High in SK Low in SK High in SK Low in SK behavior
IP → SI 0.166 (1.357) 0.042 (0.366) 0.483*** (5.450) 0.224** (2.059)
SC → SI −0.332*** (3.056) −0.101 (0.712) −0.168** (2.311) −0.343*** (3.282) 1445
IP × SN → SI 0.296* (1.787) −0.340 (1.044) −0.026 (0.211) 0.179 (0.833)
SC × SN → SI 0.215* (1.859) 0.243* (1.657) 0.226*** (2.963) 0.075 (0.483)
SI → NC 0.581*** (5.257) 0.479*** (4.041) 0.591*** (7.883) 0.588*** (6.734)
Gender → SI −0.117 (1.325) −0.070 (0.611) −0.019 (0.331) 0.081 (1.054)
Age → SI −0.058 (0.504) −0.224** (2.562) −0.080 (1.171) 0.068 (0.981)
Education → SI −0.051 (0.545) −0.254** (2.423) 0.022 (0.460) −0.020 (0.308)
Annual income → SI 0.005 (0.042) 0.385*** (3.160) −0.017 (0.115) 0.077 (0.874)
Social desirability → SI −0.049 (0.209) −0.342 (1.082) 0.101 (1.245) −0.159 (0.749)
Gender → NC −0.068 (0.846) 0.400*** (4.629) 0.075 (1.315) −0.162 (2.260)
Age → NC 0.263*** (3.031) 0.017 (0.228) −0.007 (0.111) 0.167** (2.022)
Education → NC −0.181* (1.874) −0.059 (0.528) −0.017 (0.307) 0.073 (1.178)
Annual income → NC −0.020 (0.191) 0.158 (1.593) 0.045 (0.623) −0.140* (1.757)
Social desirability → NC 0.390 (1.532) −0.059 (0.317) 0.310*** (4.157) −0.260 (0.797)
R2
Table V.
SI 0.476 0.681 0.654 0.554
PLS analysis results
NC 0.526 0.441 0.638 0.479 of ad hoc tests
Notes: IP: impulse; SC: self-control; SN: subjective norm; SI: speculative investment behavior; NC, negative among four types
consequence.*p o0.10; **po 0.5; ***p o0.01 of Bitcoin users

simultaneously considering Bitcoin OK and SK produced four distinctive user groups: the
expertise (high OK and high SK) group (n ¼ 79), the well-informed (high OK and low SK)
group (n ¼ 34), 3) the self-confident (low OK and high SK) group (n ¼ 128) and the stranger
(low OK and low SK) group (n ¼ 93). We conducted four separate PLS analyses to test each
group individually (Table V ).
The results showed that the effect of impulse on Bitcoin speculative investment in the
self-confident group was the highest (β ¼ 0.483, po0.01). The effect of self-control was
the greatest on Bitcoin speculation in the stranger user group (β ¼ −0.343, po0.01) and then
the expertise group (β ¼ −0.332, po0.01), with that in the well-informed group being none.
The self-confident group showed the highest negative consequences from Bitcoin speculative
investment (β ¼ 0.591, po0.01) among the four user groups, followed by the stranger group
(β ¼ 0.588, po0.01). The subjective norm had a significant effect on impulse in Bitcoin
speculation only in the expertise group (β ¼ 0.296, po0.10), and the well-informed user group
had the weakest effect on self-control with the subjective norm (β ¼ 0.243, po0.10).
The self-confident group engaged in the most risky and dangerous Bitcoin investments,
showing a combination of high impulse, low self-control and high negative consequence. The
expertise group was the safest with high self-control and no impulse, making it the most
controlled and reasonable group. However, with the subjective norm, the impulse
strengthened (β ¼ 0.296, po0.10) and weakened self-control (β ¼ 0.215, po0.10), leading
to Bitcoin speculation in this group. The stranger group had high impulse but the strongest
self-control, giving it a balanced tendency. The subjective norm had no moderating
effect on impulse or self-control in the stranger group. Furthermore, the results showed
that young users with a high annual income tended to speculate on Bitcoin and that
men in the well-informed group generally experienced negative consequences from their
IMDS speculative investments. In both the expertise and stranger groups, old users tended to
119,7 experience negative consequences from their speculations. The self-confident group seemed to
cut down its negative consequences, and its real losses could be big.

6. Discussions and implications

6.1 Discussions of findings
1446 Despite the high expectation of potential in Bitcoin, many researchers have questioned its
viability. In practice, Bitcoin is mainly used as a speculative asset, not an alternative currency.
Nakamoto (2008), who first proposed Bitcoin to the world, did not predict this situation, which
led us to ask: what is the cognitive process people use when they treat Bitcoin as a speculative
asset which was not its original purpose? Does Bitcoin speculative investment behavior truly
result in negative consequences? This study attempts to answer those research questions here
by investigating the decision-making mechanism of people who engaged in Bitcoin speculation.
First, the results of this study showed that Bitcoin speculative investment behavior is
largely unplanned. The price volatility and temptation of a large return caused Bitcoin users
or investors to lose their self-control and be easily swayed by impulse, leading them to
engage in unplanned behavior, Bitcoin speculation. To explain Bitcoin speculative
investment behavior, we adopted the dual-systems perspective, which posits that two
different systems (i.e. reflexive and reflective) compete to determine people’s behavior. The
results of this study empirically reveal that an imbalance between the two cognitive systems
leads to speculative investment behavior.
Second, this study found that speculative investment behavior in the Bitcoin context is
driven by the combination of a strong impulse and weak self-control. Consistent with previous
studies, we found that Bitcoin speculative investment behavior was strongly associated with
an imbalance between impulse and self-control. For example, Bitcoin investors with low
impulse and high self-control controlled and mitigated their speculative investment behavior,
leading to a low level of negative consequences. By contrast, investors with strong impulse and
low self-control reacted on impulse, strengthening their speculative investment behavior and
leading to a high level of negative consequences. To prevent Bitcoin speculative investment
behavior and its negative consequences, the function of self-control is essential. If self-control is
weak, it cannot regulate behavioral and emotional impulses, resulting in Bitcoin speculation.
Third, Bitcoin speculation investment behavior usually has negative consequences, but
the extent of those negative consequences was not as great as expected. It should be noted
that speculation is not an investment. Speculation (spending much time and money in the
expectation of profits) is typically deemed risky and dangerous for investors and can lead to
negative consequences. The strong temptation of a large return overshadows the high risks
of speculation and drives people to engage in Bitcoin speculative investment behavior with
low self-control, which can result in negative outcomes. However, we found that the extent
of actual negative consequences from Bitcoin speculation was not great because the mean
for this construct was low (Panel B of Table II). Although Bitcoin speculation was
significantly related to negative consequences, the findings showed that those negative
consequences might be less serious than anticipated.
In this study, we discovered an interesting notion about the role of the subjective norm as a
moderator between the reflexive and reflective systems. The subjective norm weakened self-
control, thereby promoting Bitcoin speculation; however, a subjective norm favorable toward
Bitcoin did not strengthen impulse but rather weakened it. One possible explanation is that
Bitcoin investment directly involves money, unlike technology adoption, so the subjective
norm cannot drive investors to immediately comply with their referents. For example, when
friends and referents recommend Bitcoin investment, investors might think about their
budget or the possibility of financial loss, which would slightly weaken impulse.
However, positive news stories or a further recommendation could drive people to weak Bitcoin
self-control, leading to engage in Bitcoin speculative investment behavior. With the subjective speculative
norm, the extent of imbalance between the two systems was greater than without the investment
subjective norm, which increased Bitcoin speculative investment behavior.
Noteworthy differences in the impulse and self-control effects were found depending on the behavior
two types of Bitcoin knowledge (i.e. OK and SK). As shown in Figure 3, the high OK group and
the low SK group resisted speculating in Bitcoin, whereas the low OK group and the high SK 1447
group were strongly obsessed with Bitcoin speculation. The high OK group and the low SK
group present strong self-control with weak impulse. Our results indicate that people in the high
OK group managed to prevent damage from Bitcoin speculation and tried to maintain their
standard of investment while investing in Bitcoin. The users with low SK showed the highest
self-control against Bitcoin speculation because they are the most conservative and unconfident
given the high level of uncertainty associated with Bitcoin investment outcomes. The subjective
norm had more influence on the high OK group than the low SK group. Although the high OK
group was more controlled and conscious than the other three groups, with the addition of the
subjective norm, this group became more impulsive and non-conscious than the other three
groups, leading to engage in Bitcoin speculation.
A possible explanation for this intriguing result is that the subjective norm stimulates the
leading characteristics of the high OK group and thus triggers them to hurry to be ahead of
technical trends and be opinion leaders. Bitcoin speculative investment behavior in the high
SK group and low OK group was driven from a strong impulse with little self-control.
Specifically, the low OK group had the highest impulse and weakest self-control to invest in
Bitcoin, resulting in widespread participation in Bitcoin speculation and the most negative
consequences of any group. Those results were consistent with a prior study (Glaser et al.,
2014) reporting that new users with little or no actual Bitcoin knowledge tended to want to

0.3
0.5 0.209*
0.2 0.203***
0.385***
0.4
0.1
0.3
0.223** 0
0.189** SC SC with SN
0.2 –0.1
0.075 –0.194***
0.1 –0.2

0 –0.3 –0.253**
IP IP with SN
High OK group Low OK group High OK group Low OK group

(a) Moderation effects of the subjective norm on speculative behavior in the OK group

0.3
0.4 0.182*
0.328*** 0.2
0.3 0.1 0.177***
0.215** 0.217 0
SC SC with SN
0.2
0.190** –0.1
–0.2 –0.252***
0.1
–0.3
–0.294***
0 –0.4
IP IP with SN
High SK group Low SK group High SK group Low SK group
Figure 3.
Moderation effects of
(b) Moderation effects of the subjective norm on speculative behavior in the SK group the subjective norm
depending on Bitcoin
Notes: IP: impulse; SC: self-control; SN: subjective norm; OK: objective knowledge; knowledge
SK: subjective knowledge. *p<0.10, **p<0.5, ***p<0.01
IMDS participate in speculative investment. The high SK group, which was willing to pursue a risky
119,7 investment, easily engaged in Bitcoin speculation, with predictable negative consequences.
Therefore, Bitcoin speculative investment behavior in the low OK group and high SK group
should be carefully supervised and managed by regulators.

6.2 Theoretical and practical implications

1448 The results of this study carry several theoretical implications for researchers. First, to our
best knowledge, this study is the first attempt to empirically examine the decision-making
mechanism in Bitcoin speculation from the user perspective. In this study, we proposed the
two different cognitive systems, impulse and self-control, based on the dual-systems
perspective to investigate investors’ decision-making mechanisms for Bitcoin speculation.
The theoretical and empirical validation of the effect of impulse and self-control on Bitcoin
speculative investment behavior is a significant research contribution of this study.
A detailed understanding of the decision-making mechanism behind Bitcoin speculative
investment behavior would allow significant progress in the IS field.
Second, this study expands the current research on Bitcoin speculative investment
behavior, which received significant public attention but has been mostly overlooked by
academic researchers. We found that Bitcoin speculative investment behavior is an
unplanned behavior driven by an imbalanced interaction between high impulse and low
self-control. Our findings show that the decision-making process involved in Bitcoin
speculation is similar to that of other unplanned or problematic behaviors in IT use (e.g.
internet overuse, mobile e-mail and Facebook addiction, and problematic SNS use and
mobile phone use). This study thus facilitates a balanced view and encourages the use of the
dual-systems perspective in IS research.
Third, this study highlights the role of the subjective norm in the cognitive process
behind Bitcoin speculative investment behavior. Although the subjective norm was
originally proposed as a determinant of planned behavior (Ajzen, 1991), the effect of the
subjective norm on unplanned behavior had not previously been investigated in the IS
literature. This study empirically validated how Bitcoin investment decisions reacted with
the subjective norm. Interestingly, the subjective norm strongly weakens self-control but is
almost ineffective on impulse because impulse represents a preoccupation with persistent
thoughts about a behavior. This study provides a detailed understanding of the effect of the
subjective norm on Bitcoin speculative investment behavior.
Fourth, this study reveals that the effects of impulse and self-control on Bitcoin
speculation differ depending on two types of user knowledge about Bitcoin. We focused on
the different influences of impulse and self-control in high- and low- Bitcoin knowledge
groups. The high OK group and the low SK group were reserved in Bitcoin speculation,
whereas the low OK group and the high SK group were willing to speculate in Bitcoin. This
study thus sheds light on how Bitcoin investors differ in their decision-making process
concerning speculative investment behavior depending on their Bitcoin knowledge.
This study also provides several practical implications for practitioners and policy makers.
First, it provides policy makers and regulators with valuable insights to develop effective and
practical policies for supervising and preventing Bitcoin speculation. The results of this study
show that Bitcoin speculative investment behavior is driven by an imbalance between strong
impulse and weak self-control, which can be corrected by strengthening self-control or
weakening impulse. A strict governmental approach forbidding impulsive Bitcoin investment
could effectively restrain Bitcoin speculation. However, such a strict governmental approach
could stimulate users’ desire for profit, accelerating illegal Bitcoin speculation, or it could depress
all Bitcoin transactions and the overall market. Therefore, policy makers and regulators should
carefully consider how they help investors balance their two cognitive systems.
 Second, the subjective norm plays an important role in promoting or preventing Bitcoin
cryptocurrency speculation. Our findings show that the extent of the imbalance between the speculative
two cognitive systems is greater with the subjective norm than without it. Thus, the investment
subjective norm increases Bitcoin speculative investment behavior. Especially, preventing
self-control from weakening by social pressure is more important than restraining impulse. behavior
Our results suggest that adequate information or OK about the actual benefits and risks of
Bitcoin could prevent social influence and allow investors to make rational decisions about 1449
their speculative investments. Regulating misinformation about Bitcoin investment is also
necessary to prevent speculative investment.
Third, this study offers policy makers and regulators a deep understanding of Bitcoin
knowledge. In the Bitcoin context, the high OK group and the low SK group show reserve
when given the chance to speculate in Bitcoin, whereas the low OK group and the high SK
group seek financial gains through Bitcoin speculation. In particular, the low OK group tends
to be the most exposed to danger through Bitcoin speculation. That understanding of Bitcoin
users’ characteristics can enable effective regulations, thereby preventing speculation and
simultaneously facilitating healthy investment in Bitcoin. Building a risk-free transaction
environment in the Bitcoin context is important and could inspire higher levels of confidence
about Bitcoin transactions in potential users. Moreover, providing OK about Bitcoin could
inhibit the effect of the impulse and improve self-control. These findings thus provide valuable
information for policy makers and regulators who seek practical guidance in establishing
Bitcoin regulations and setting the direction of governance requirements.

6.3 Limitations and future research directions

Despite its contributions, this study has several limitations. First, this study relied on
respondent self-reported data. Because respondents might have inflated their positive
behavior responses (i.e. self-control) and downplayed their negative behavior responses (i.e.
speculative investment behavior and negative consequences), the actual levels of our
constructs might be overestimated or underestimated. We thus employed social desirability as
a control variable to prevent this research bias, but the possibility of bias still exists. Future
studies should objectively assess these constructs to reduce that gap and provide stronger
results. Second, out of the many possible manifestations of the reflexive and reflective
systems, this study measured one general impulse and self-control. Considering that impulse
and self-control are multifaceted, multi-dimensional constructs, their conceptualization and
measurement might have captured only a subset of the processes underlying impulse and self-
control. Therefore, their effects might be biased, so future studies should extend impulse and
self-control into multi-dimensional constructs to capture a whole set of processes.
Third, because Bitcoin research is still in its early stage, the measures of Bitcoin
speculative investment behavior were not tested comprehensively. Although they were
drawn from the theoretical literature and validated by the procedure proposed by Moore
and Benbasat (1991), further studies are necessary to ensure their validity and develop more
valid measurements. Fourth, the research scope of this study was limited to Bitcoin, which is
the first decentralized and most popular cryptocurrency. Therefore, our findings might not
be applicable to other cryptocurrencies (e.g. Ethereum, Ripple, Bitcoin cash, Dash, Eos and
Light coin). Future research needs to extend the research scope from Bitcoin to other
cryptocurrencies to provide comprehensive implications. Finally, our findings might not be
completely generalized because most of our sample was restricted to a specific time period
and nation, from 2017 to early 2018 in South Korea. Therefore, the results of this study must
be interpreted cautiously. Considering the importance of time and national characteristics in
Bitcoin, future studies need to apply a specific method, such as longitudinal analysis, and
collect more data to closely observe and investigate worldwide Bitcoin speculation
phenomenon across different time periods and nations.
IMDS References
119,7 Abramova, S. and Böhme, R. (2016), “Perceived benefit and risk as multidimensional determinants of
Bitcoin use: a quantitative exploratory study”, Proceedings of the 37th International Conference
on Information Systems, Dublin, pp. 1-20.
Ajzen, I. (1991), “The theory of planned behavior”, Organizational Behavior and Human Decision
Processes, Vol. 50 No. 2, pp. 179-211.
1450 Ajzen, I. and Fishbein, M. (1977), “Attitude-behavior relations: a theoretical analysis and review of
empirical research”, Psychological Bulletin, Vol. 84 No. 5, pp. 888-918.
Angel, J.J. and McCabe, D.M. (2009), “The ethics of speculation”, Journal of Business Ethics, Vol. 90
No. 3, pp. 277-286.
Antonopoulos, A.M. (2014), Mastering Bitcoin: Unlocking Digital Cryptocurrencies, 1st ed., O’Reilly
Media, Inc., CA.
Bandura, A. (1997), Self-Efficacy: The Exercise of Control, 1st ed., Freeman, NY.
Barclay, D., Higgins, C. and Thompson, R. (1995), “The partial least squares (PLS) approach to causal
modeling: personal computer adoption and use as an illustration”, Technology Studies, Vol. 2
No. 2, pp. 285-309.
Bass, F.M. (1969), “A new product growth model for consumer durables”, Management Science, Vol. 15
No. 5, pp. 215-227.
Baur, A.W., Bühler, J., Bick, M. and Bonorden, C.S. (2015), “Cryptocurrencies as a disruption? Empirical
findings on user adoption and future potential of Bitcoin and co”, Proceedings of the Conference
on e-Business, e-Services and e-Society, Delft, pp. 63-80.
Baur, D.G., Hong, K. and Lee, A.D. (2018), “Bitcoin: medium of exchange or speculative assets?”, Journal
of International Financial Markets, Institutions and Money, Vol. 43, May, pp. 177-189.
Bhattacherjee, A. and Sanford, C. (2006), “Influence processes for information technology acceptance:
an elaboration likelihood model”, MIS Quarterly, Vol. 30 No. 4, pp. 805-825.
Billieux, J., der Linden, M.V. and Rochat, L. (2008), “The role of impulsitivy in actual and problematic
use of the mobile phone”, Applied Cognitive Psychology, Vol. 22 No. 9, pp. 1195-1210.
Blau, B.M. (2017), “Price dynamics and speculative trading in Bitcoin”, Research in International
Business and Finance, Vol. 41, October, pp. 493-499.
Bogg, T. and Roberts, B.W. (2004), “Conscientiousness and health-related behaviors: a meta-analysis
of the leading behavioral contributors to mortality”, Psychological Bulletin, Vol. 130 No. 6,
pp. 887-919.
Böhme, R., Christin, N., Edelman, B. and Moore, T. (2015), “Bitcoin: economics, technology, and
governance”, Journal of Economic Perspectives, Vol. 29 No. 2, pp. 213-238.
Bohr, J. and Bashir, M. (2014), “Who uses Bitcoin? An exploration of the Bitcoin community”,
Proceedings of the 12th Annual Conference on Privacy, Security and Trust, Toronto, pp. 94-101.
Brezo, F. and Bringas, P.G. (2012), “Issues and risks associated with cryptocurrencies such as Bitcoin”,
Proceedings of the 2nd International Conference on Social Eco-Informatics, Venice, pp. 20-26.
Brucks, M. (1985), “The effects of product class knowledge on information search behavior”, Journal of
Consumer Research, Vol. 12 No. 1, pp. 1-16.
Caplan, S.E. (2010), “Theory and measurement of generalized problematic internet use: a two-step
approach”, Computers in Human Behavior, Vol. 26 No. 5, pp. 1089-1097.
Chaiken, S. (1980), “Heuristic versus systematic information processing and the use of source versus
message cues in persuasion”, Journal of Personality and Social Psychology, Vol. 39 No. 5,
pp. 752-766.
Cheah, E.-T. and Fry, J. (2015), “Speculative bubbles in Bitcoin markets? An empirical investigation
into the fundamental value of Bitcoin”, Economics Letters, Vol. 130, May, pp. 32-36.
Chin, W.W. (1998), “Commentary: issues and opinion on structural equation modeling”, MIS Quarterly,
Vol. 22 No. 1, pp. vii-xvi.
Chin, W.W., Marcolin, B.L. and Newsted, P.R. (2003), “A partial least squares latent variable modeling Bitcoin
approach for measuring interaction effects: results from a Monte Carlo simulation study and an speculative
electronic-mail emotion/adoption study”, Information Systems Research, Vol. 14 No. 2, pp. 189-217.
Chiou, J.S. (1998), “The effects of attitude, subjective norm, and perceived behavioral control on
investment
consumers’ purchase intentions: the moderating effects of product knowledge and attention to behavior
social comparison information”, Proceedings of the National Science Council, Republic of China,
Vol. 9 No. 2, pp. 298-308.
Collins, R.L. and Lapp, W.M. (1992), “The temptation and restraint inventory for measuring drinking 1451
restraint”, British Journal of Addiction, Vol. 87 No. 4, pp. 625-633.
Devine, P.G. (1989), “Stereotypes and prejudice: their automatic and controlled components”, Journal of
Personality and Social Psychology, Vol. 56 No. 1, pp. 5-18.
Eysenck, B.G. and Eysenck, H.J. (1978), “Impulsiveness and venturesomeness: their position in a
dimensional system of personality description”, Psychological Report, Vol. 43 No. 3, pp. 1247-1255.
European Central Bank (2012), Virtual Currency Schemes, European Central Bank, Frankfurt.
Fehr, R.R. and Hari, J.J. (2014), “Assessing the risk attitudes of private investors using the implicit
association test”, Journal of Financial Service Professionals, Vol. 68 No. 6, pp. 50-62.
Fishbein, M. and Ajzen, I. (1975), Belief, Attitude, Intention and Behavior: An Introduction to Theory
and Research, 1st ed., Addison-Wesley, MA.
Fornell, C. and Bookstein, F.L. (1982), “Two structural equation models: LISREL and PLS applied to
consumer exit-voice theory”, Journal of Marketing Research, Vol. 19 No. 4, pp. 440-452.
Fornell, C. and Larcker, D.F. (1981), “Evaluating structural equation models with unobservable
variables and measurement error”, Journal of Marketing Research, Vol. 18 No. 1, pp. 39-50.
Fox, J. and Moreland, J.J. (2015), “The dark side of social networking sites: an exploration of the
relational and psychological stressors associated with Facebook use and affordances”,
Computers in Human Behavior, Vol. 45, April, pp. 168-176.
Friese, M. and Hofmann, W. (2009), “Control me or I will control you: impulses, trait self-control, and the
guidance of behavior”, Journal of Research in Personality, Vol. 43 No. 5, pp. 795-805.
Gandal, N., Hamrick, J., Moore, T. and Oberman, T. (2018), “Price manipulation in the Bitcoin
ecosystem”, Journal of Monetary Economics, Vol. 95, May, pp. 86-96.
Glaser, F., Zimmermann, K., Haferkorn, M., Weber, M. and Siering, M. (2014), “Bitcoin-asset or
currency? Revealing users’ hidden intentions”, Proceedings of the 22nd European Conference on
Information Systems, Tel Aviv, pp. 1-14.
Grant, G. and Hogan, R. (2015), “Bitcoin: risks and controls”, Journal of Corporate Accounting & Finance,
Vol. 26 No. 5, pp. 29-35.
Grinberg, R. (2012), “Bitcoin: an innovative alternative digital currency”, Hastings Science &
Technology Law Journal, Vol. 4 No. 1, pp. 159-181.
Haagsma, M.C., Caplan, S.E., Peters, O. and Pieterse, M.E. (2013), “A cognitive-behavioral model of
problematic online gaming in adolescents aged 12–22 years”, Computers in Human Behavior,
Vol. 29 No. 1, pp. 202-209.
Hadar, L., Sood, S. and Fox, C.R. (2013), “Subjective knowledge in consumer financial decisions”,
Journal of Marketing Research, Vol. 50 No. 3, pp. 303-316.
Hair, J.F., Anderson, R.E., Tatham, R.L. and William, C. (1998), Multivariate Data Analysis, 1st ed.,
Prentice Hall, NJ.
Harden, K.P. and Tucker-Drob, E.M. (2011), “Individual differences in the development of sensation
seeking and impulsivity during adolescence: further evidence for a dual systems model”,
Developmental Psychology, Vol. 47 No. 3, pp. 739-746.
Henry, C.S., Huynh, K.P. and Nicholls, G. (2018), “Bitcoin awareness and usage in Canada”, Journal of
Digital Banking, Vol. 2 No. 4, pp. 311-337.
Hofmann, W., Friese, M. and Strack, F. (2009), “Impulse and self-control from a dual-systems
perspective”, Perspectives on Psychological Science, Vol. 4 No. 2, pp. 162-176.
IMDS Hur, Y., Jeon, S. and Yoo, B. (2015), “Is Bitcoin a viable e-business? Empirical analysis of the digital
119,7 currency’s speculative nature”, Proceedings of the 36th International Conference on Information
Systems, Fort Worth, TX, pp. 1-10.
Kristoufek, L. (2015), “What are the main drivers of the Bitcoin price? Evidence from wavelet coherence
analysis”, PlosOne, Vol. 10 No. 4, pp. 35-48.
Lee, J.K. and Lee, W.-N. (2009), “Country-of-origin effects on consumer product evaluation and purchase
intention: the role of objective versus subjective knowledge”, Journal of International Consumer
1452 Marketing, Vol. 21 No. 2, pp. 137-151.
Lieberman, M.D. (2007), “Social cognitive neuroscience: a review of Core processes”, Annual Review of
Psychology, Vol. 58 No. 2007, pp. 259-289.
Liu, Y. and Tsyvinski, A. (2018), “Risks and returns of cryptocurrency”, National Bureau of Economic
Research, Vol. 2018 No. 24877, pp. 1-25.
Loewenstein, G. (1996), “Out of control: visceral influences on behavior”, Organizational Behavior and
Human Decision Processes, Vol. 65 No. 3, pp. 272-292.
Markus, H.R. and Kitayama, S. (1991), “Culture and the self: implications for cognition, emotion, and
motivation”, Psychological Review, Vol. 98 No. 2, pp. 224-253.
Mishra, S.K. and Kumar, M. (2011), “How mutual fund investors’ objective and subjective knowledge
impacts their information search and processing behaviour”, Journal of Financial Services
Marketing, Vol. 16 No. 1, pp. 27-41.
Moore, G.C. and Benbasat, I. (1991), “Development of an instrument to measure the perceptions of adopting
an information technology innovation”, Information Systems Research, Vol. 2 No. 3, pp. 192-222.
Moore, T. and Christin, N. (2013), “Beware the middleman: empirical analysis of Bitcoin-exchange risk”,
Proceedings of the International Conference on Financial Cryptography and Data Security,
Okinawa, pp. 25-33.
Moqbel, M. and Kock, N. (2018), “Unveiling the dark side of social networking sites: personal and work-
related consequences of social networking site addiction”, Information & Management, Vol. 55
No. 1, pp. 109-119.
Nakamoto, S. (2008), “Bitcoin: a peer-to-peer electronic cash system”, available at: http://bitcoin.org/
bitcoin.pdf
Neter, J., Kutner, M.H., Nachtsheim, C.J. and Wasserman, W. (1996), Applied Linear Statistical Models,
4th ed., Irwin Chicago, IL.
Orcutt, M. (2015), “Is Bitcoin stalling?”, MIT Technology Review, available at: www.technologyreview.
com/s/535221/is-bitcoin-stalling/
Park, C.W., Mothersbaugh, D.L. and Feick, L. (1994), “Consumer knowledge assessment”, Journal of
Consumer Research, Vol. 21 No. 1, pp. 71-82.
Podsakoff, P.M., MacKenzie, S.B., Lee, J.-Y. and Podsakoff, N.P. (2003), “Common method biases in
behavioral research: a critical review of the literature and recommended remedies”, Journal of
Applied Psychology, Vol. 88 No. 5, pp. 879-903.
Reynolds, W.M. (1982), “Development of reliable and valid short forms of the Marlowe-Crowne social
desirability scale”, Journal of Clinical Psychology, Vol. 38 No. 1, pp. 119-125.
Rolland, E., Patterson, R.A., Messinger, P.R., Ward, K.F. and Finn, A. (2013), “Dual rules for service
evaluation”, Service Science, Vol. 5 No. 4, pp. 279-295.
Ron, D. and Shamir, A. (2013), “Quantitative analysis of the full Bitcoin transaction graph”, Proceedings
of the International Conference on Financial Cryptography and Data Security, Okinawa, pp. 6-24.
Rook, D.W. and Fisher, R.J. (1995), “Normative influences on impulsive buying behavior”, Journal of
Consumer Research, Vol. 22 No. 3, pp. 305-313.
Sawang, S., Sun, Y. and Salim, S.A. (2014), “It’s not only what I think but what they think! The
moderating effect of social norms”, Computers & Education, Vol. 76, July, pp. 182-189.
Shiller, R.J. (1981), “The use of volatility measures in assessing market efficiency”, Journal of Finance,
Vol. 36 No. 2, pp. 291-304.
Soror, A.A., Hammer, B.I., Steelman, Z.R., Davis, F.D. and Limayem, M.M. (2015), “Good habits gone Bitcoin
bad: explaining negative consequences associated with the use of mobile phones from a speculative
dual-systems perspective”, Information Systems Journal, Vol. 25 No. 4, pp. 403-427.
investment
Spada, M.M. (2014), “An overview of problematic internet use”, Addictive Behaviors, Vol. 39 No. 1,
pp. 3-6. behavior
Spenkelink, H. (2014), “The adoption process of cryptocurrencies–identifying factors that
influence the adoption of cryptocurrencies from a multiple stakeholder perspective”, 1453
University of Twente, available at: https://essay.utwente.nl/65677/
Stein, J.C. (1987), “Informational externalities and welfare-reducing speculation”, Journal of Political
Economy, Vol. 95 No. 6, pp. 1123-1145.
Strack, F. and Deutsch, R. (2004), “Reflective and impulsive determinants of social behavior”,
Personality and Social Psychology Review, Vol. 8 No. 3, pp. 220-247.
Straub, D.W. (1989), “Validating instruments in MIS research”, MIS Quarterly, Vol. 13 No. 2, pp. 147-169.
Taylor, S. and Todd, P.A. (1995), “Understadning information technology usage: a test of competing
models”, Information Systems Research, Vol. 6 No. 2, pp. 144-176.
Surowiecki, J. (2011), “Cryptocurrency”, MIT Technology Review, available at: www.technologyreview.
com/s/425142/cryptocurrency/.
Tangney, J.P., Boone, A.L. and Baumeister, R.F. (2004), “High self-control predicts good adjustment,
less pathology, better grades, and interpersonal success”, Journal of Personality, Vol. 72 No. 2,
pp. 271-324.
Turel, O. and Qahri-Saremi, H. (2016), “Problematic use of social networking sites: antecedents and
consequence from a dual-system theory perspective”, Journal of Management Information
Systems, Vol. 33 No. 4, pp. 1087-1116.
Turel, O. and Serenko, A. (2010), “Is mobile email addiction overlooked?”, Communications of the ACM,
Vol. 53 No. 5, pp. 41-43.
Turel, O., He, Q., Xue, G., Xiao, L. and Bechara, A. (2014), “Examination of neural systems sub-serving
Facebook ‘addiction’”, Psychological Reports, Vol. 115 No. 3, pp. 675-695.
Urquhart, A. (2016), “The inefficiency of Bitcoin”, Economics Letters, Vol. 148, November, pp. 80-82.
Van Alstyne, M. (2014), “Why Bitcoin has value?”, Communications of the ACM, Vol. 57 No. 5,
pp. 30-32.
Venkatesh, V., Brown, S.A., Maruping, L.M. and Bala, H. (2008), “Predicting different
conceptualizations of system use: the competing roles of behavioral intention, facilitating
conditions, and behavioral expectation”, MIS Quarterly, Vol. 32 No. 3, pp. 483-502.
Watts, S. (2015), “Application of dual-process theory to information systems: current and future
research directions”, Foundations and Trends® in Information Systems, Vol. 1 No. 2, pp. 69-162.
Webster, J. and Trevino, L.K. (1995), “Rational and social theories as complementary explanations of
communication media choices: two policy-capturing studies”, Academy of Management Journal,
Vol. 38 No. 6, pp. 1544-1572.
Yermack, D. (2015), “Is Bitcoin a real currency? An economic appraisal”, Handbook of Digital Currency,
Vol. 2015 No. 2, pp. 31-43.
Yu, X. (2018), “China orders banks to stop financing cryptocurrencies as noose tightens around
disrupter”, South China Morning Post, available at: www.scmp.com/business/banking-finance/
article/2129645/pboc-orders-banks-halt-banking-services-cryptocurrency
Zhang, K.Z., Zhao, S.J., Cheung, C.M. and Lee, M.K. (2014), “Examining the influence of online reviews
on consumers’ decision-making: a heuristic–systematic model”, Decision Support Systems,
Vol. 67, November, pp. 78-89.
IMDS Appendix
119,7

Constructs Questionnaire References

Impulse (IP) IP1. I often invest in Bitcoin on impulse Eysenck and

1454 IP2. I often invest in Bitcoin on the spur of the moment Eysenck (1978)
IP3. I often invest in Bitcoin without any deliberation and Harden and
IP4. I generally invest in Bitcoin without thinking many times Tucker-Drob
(2011)
Self-control (SC) SC1. I exert effort to maintain my investment budget when I invest Eysenck and
in Bitcoin Eysenck (1978)
SC2. I exert effort not to over-budget when I invest in Bitcoin and Tangney
SC3. I exert effort not to spend considerable time and money when I et al. (2004)
invest in Bitcoin
Subjective SN1. My friends or my co-workers think that I should invest in Bitcoin Taylor and
norm (SN) SN2. My friends or my co-workers think that investing in Bitcoin is a Todd (1995)
general trend
SN3. People who are important to me invest in Bitcoin
SN4. Many people around me invest in Bitcoin
Subjective Compared to average people Brucks (1985)
knowledge (SK) SK1. What is my knowledge level about cryptocurrency? and Park et al.
SK2. What is my knowledge level about Bitcoin? (1994)
SK3. What is my knowledge level about other cryptocurrencies (i.e.
Ethereum, Bitcoin Cash, LiteCoin, EOS, Ripple, Dash and so on)?
SK4. What is my knowledge level about the cryptocurrency market
and its outlook?
Objective Please answer the following questions. If the statement is true, please Brucks (1985),
knowledge (OK) choose “True.” If the statement is wrong, please choose “False.” If you Park et al. (1994)
don’t know, please choose “Don’t know” and Chiou (1998)
OK1. Proofs of statement or PoS, introduced as part of Bitcoin,
mean that Bitcoin uses a hash algorithm to generate
new blocks
OK2. Bitcoins are sent and received through software and websites
called wallets
OK3. Blockchain is an open, distributed ledger that can efficiently
record transactions between two parties, leading to
safekeeping of ledgers
OK4. Bitcoins are offered as an incentive to people who perform the
computation work required to generate blocks, and no more
than 31 million Bitcoins will ever exist
OK5. The Bitcoin blockchain can create financial transactions as well
as generate financial applications directly related to contracts
OK6. A private blockchain refers to a blockchain platform for a special
purpose operated by special people, of which Bitcoin and
Ethereum are good examples
OK7. Ethereum is the first of public blockchain 2.0
OK8. A hard fork refers to a rule change such that the new rules
consider the blocks produced by old software to be invalid
OK9. Satoshi Nakamoto proposed a peer-to-peer network system to
prevent the double-spending problem without requiring third
parties because double-spending is a potential flaw in a digital
cash scheme allowing a single digital token to be spent more
than once
Table AI.
Structure of the
survey instrument (continued )
Constructs Questionnaire References
Bitcoin
speculative
OK10. ICO (Initial Coin Offering) is the launch of a new coin and its investment
sale to the public, which has become a popular method of
crowdfunding behavior
OK11. The Korean government (i.e. the Financial Service Commission)
permits ICOs in Korea
OK12. Electronic money, cyber money, and cryptocurrency are legal 1455
and protected by the Korean government
OK13. Segregated Witness or SegWit refers to a Bitcoin Improvement
Proposal to overcome two shortcomings of Bitcoin, scalability
and malleability, and was thus employed by Bitcoin to improve
the current Bitcoin blockchain design
OK14. The Cryptocurrency Exchange deals with trades in which
actual cryptocurrencies are sent and received
Speculative SI1. I often invest in Bitcoin during my work time Billieux et al.
investment SI2. I invest in Bitcoin using more money than my income (2008) and
behavior (SI) SI3. I invest in Bitcoin using more money than my original budget Venkatesh et al.
SI4. Overall, I invest in Bitcoin by spending considerable time and (2008)
money
Negative NC1. I have received a financial loss from investing in Bitcoin Caplan (2010),
consequence (NC) NC2. I am in psychological trouble from investing in Bitcoin Soror et al. (2015)
NC3. I am in trouble at work because of investing in Bitcoin and Haagsma
NC4. I am in trouble with social engagements or activities because of et al. (2013)
investing in Bitcoin
NC5. Overall, investing in Bitcoin has made it difficult for me to
manage my life
Social desirability SD1. It is sometimes hard for me to go on with my work if I am Reynolds (1982)
(SD) not encouraged
SD2. I am always willing to admit when I make a mistake
SD3. On a few occasions, I have given up doing something because I
thought too little of my ability
SD4. There have been times when I felt like rebelling against people
in authority even though I knew they were right
SD5. I sometimes feel resentful when I do not get my way
SD6. There have been occasions when I took advantage of someone
SD7. No matter who I am talking to, I am always a good listener
SD8. I sometimes try to get even rather than forgiving and forgetting
SD9. I have never resented being asked to return a favor
SD10. There have been times when I was quite jealous of the good
fortune of others
SD11. I am sometimes irritated by people who ask favors of me
SD12. I am always courteous, even to people who are disagreeable
SD13. I have never deliberately said something that hurt
someone’s feelings Table AI.
IMDS About the authors
119,7 Hyun-Sun Ryu is Visiting Professor of Software Department in the Software College of Sungkyunkwan
University in Seoul, Korea. She received the MS Degree from the Graduate School of Management,
Korea Advanced Institute of Science and Technology (KAIST) and PhD Degree from the Korea
University Business School, both in Management Information System, in 2005 and 2014, respectively.
Her work experience includes serving as Software Engineer and Senior Researcher for the
Samsung Electronics Co., Ltd, and the Korea Government Institute for Economic and Trade. Her
1456 research interests include disruptive innovations (e.g. FinTech) between service and IT and their
effects on firms, and of technology or IT as true innovator role in disruptive innovations. Her published
research articles appear in Information & Management, Industrial Management & Data Systems, IEEE
Transactions on Engineering Management, Service Industries Journal and others.
Kwang Sun Ko is Advisor in Financial Information Technology (IT) department at Kim & Chang
Law Firm. He received the MS Degree and PhD Degree from the Graduate School of Sungkyunkwan
University, both in Computer Engineering, in 2002 and 2007, respectively. After receiving the PhD
Degree, he worked for the Graduate School of Sungkyunkwan University as Research Professor, and
for the Korea Financial Security Institute as Senior Researcher. He is an expert in providing project-
related advice. Since he joined Kim & Chang Intellectual Property in 2012 as Senior Consultant, he has
provided advisory services related to information leakage to major Korean financial institutions and
FinTech companies. His research interests include information leakage in financial companies,
FinTech and IT compliance in in finance and external cooperation business. Kwang Sun Ko is the
corresponding author and can be contacted at: kwangsun.ko@gmail.com

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
 The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm

The influence
Investigating the influence of of
organizational factors on organizational
factors
blockchain adoption
An innovation theory perspective 1457

Trevor Clohessy Received 28 August 2018

Revised 8 April 2019
Department of Enterprise and Technology, Accepted 23 May 2019
Galway-Mayo Institute of Technology, Galway, Ireland, and
Thomas Acton
School of Business and Economics,
National University of Ireland Galway, Galway, Ireland

Abstract
Purpose – Blockchain possesses the potential to disrupt and reshape a plethora of industries in the next
decade. However, blockchain adoption rates in technology developed countries, such as Ireland, are
relatively low. Motivated by blockchain’s potential to transform sociotechnical systems, the lack of
systematic inquiry pertaining to blockchain studies from an information system perspective, the authors
propose the following research question: “How do organizational factors influence blockchain adoption in
organizations based in a developed country?” Specifically, the purpose of this paper is to elucidate the
impact of organizational factors on the adoption of blockchain and the adoption of blockchain in companies
based in Ireland.
Design/methodology/approach – A comprehensive literature review was conducted, and the methods of
qualitative content analysis were used to identify the most important technology–organization–environment
(TOE) blockchain adoption factors. Organizational factors are often viewed as the most significant
determinants of IT innovation adoption in organizations. Consequently, using a multiple-case study of 20
companies based in Ireland, the authors investigate how the top three organizational factors identified from
the blockchain literature affected these companies decision to adopt or not adopt blockchain.
Findings – The literature review on blockchain adoption identified specific technological, organizational and
environmental factors. Furthermore, the case study findings identified three patterns: top management
support and organizational readiness are enablers for blockchain adoption, and large companies are more
likely to adopt blockchain than small to medium-sized enterprises (SMEs). The authors explain these patterns
by examining the nature of blockchain and the characteristics of Ireland as a developed country. Practical and
scientific contributions are also presented.
Research limitations/implications – This study makes several important scientific contributions. First,
the findings revealed that top management support and organizational readiness are significant enablers of
blockchain adoption. Ireland is recognized as a technology developed country; however, the findings in
relation to top management support contradict existing IT adoption literature pertaining to developed
countries. Second, previous IT innovation adoption literature suggests that organizations size has a
positive influence on a company’s IT innovation adoption process. This study demonstrates that large
organizations are more likely to not only adopt blockchain but are also more likely to conduct increased
levels of blockchain research and development activities. Finally, and most significantly, the authors
identified several patterns, which relate specifically to Ireland as a developed country that influenced the
findings. These findings could hold particular relevance to governments and organizations of other
developed countries in terms of accelerating blockchain adoption.
Practical implications – The findings about the low level of blockchain awareness and the lack of
information pertaining to viable business use cases indicate that the Irish government could play a more
significant role in promoting the benefits of blockchain technologies. Further, the findings could also
encourage IT providers to formulate enhanced strategies aimed at disseminating information pertaining to
blockchain technologies. Second, the positive influence of top management support and organizational Industrial Management & Data
Systems
readiness, particularly about core competencies, on blockchain adoption suggests that equipping managers Vol. 119 No. 7, 2019
with the requisite knowledge and skills will be crucial in adopting these IT innovations. Finally, organizations pp. 1457-1491
© Emerald Publishing Limited
who adopted blockchain used cloud-based blockchain platforms and tools to overcome the constraints of their 0263-5577
initial low levels of organizational readiness. DOI 10.1108/IMDS-08-2018-0365
IMDS Originality/value – This is one of the first studies to identify specific TOE blockchain adoption factors.
Further, the authors examine how the three most identified organizational adoption factors impact organizations
119,7 decisions to adopt blockchain. Finally, the authors discuss how the resulting three patterns identified by
examining the nature of blockchain and the characteristics of Ireland as a technology developed country.
Keywords IT adoption, Blockchain, IT Innovation,
Technological–Environmental–Organizational (TOE) Framework
Paper type Research paper
1458
1. Introduction

Blockchain enables you to do something that you have not done before. Therefore, the fundamental
question for your business prior to adoption should be: what problem are you trying to solve which
can only be solved by blockchain? (A16)
The emergence of blockchain as a trend in the information technology (IT) sector has attracted
considerable attention from practitioners, academics, researchers and national development
authorities. Blockchain, as it is used today, “is a tamper-resistant database of transactions
consistent across a large number of nodes and is cryptographically secured against
retrospective manipulations, and it uses a consensus mechanism to keep the database
consistent whenever new transactions need to be validated” (Beck, 2018). Blockchain was
introduced in October 2008 as part of a proposal for bitcoin by Satoshi Nakamato, a virtual
currency system, that “eschewed a central authority for issuing currency, transferring
ownership, and confirming transactions” (Lansiti and Lakhani, 2017). Bitcoin is viewed as the
first application of blockchain technology. Blockchain is not a disruptive technology; it is a
foundational technology, which possesses “the capacity to create new foundations for our
economic and social systems” (Lansiti and Lakhani, 2017). Modern applications of blockchain
range from low novelty and complexity initiatives (e.g. bitcoin payments) to high novelty and
complex initiatives (e.g. self-executing contracts). There are also nuanced mature blockchain
supply chain tracking initiatives. For instance, in the diamond industry off-the-shelf
blockchain technologies can be acquired to trace gems along the supply chain from origin to
the customer. Furthermore, blockchain is being used in the music industry where databases
containing information pertaining to music rights ownership are being stored in public
ledgers. Future transformational applications will encompass large-scale public identification
systems (e.g. passport control) and machine learning-based decision making (e.g. money
laundering). Large global financial institutions such as NasDaq, Bank of America, JP Morgan,
the New York Stock Exchange, Fidelity Investments are currently conducting private
blockchain research and development initiatives. These initiatives encompass the trialing of
digital currencies (e.g. interbank transfers), and the replacement of manual and paper-based
transactions (e.g. foreign exchanges). Blockchain technology can provide its adopters with
strategic and operational advantages which include enhanced security, cost savings,
immutability, faster transactions, transparency and pseudonymity (Lansiti and Lakhani,
2017; Tapscott and Tapscott, 2016).
Several reports have predicted that these advantages may be especially promising for
developed countries (Espinel, 2015; Cuomo et al., 2016; Pisa and Juden, 2017). For the
purposes of this study, a developed country represents a sovereign state categorized as
having a highly developed economy and advanced IT infrastructure relative to other less
industrialized countries. In 2018 the global blockchain technology market is predicted to
reach $548m in size and is forecast to grow to $2.3bn dollars by 2021 (Mehta and Striapunia,
2017). However, although the global blockchain adoption rate is increasing gradually, as
reported by IT analysts such as McKinsey (2017) and Accenture (Treat et al., 2017) and
multinational technology company IBM (Bear et al., 2016), in developed countries such as
Ireland, the UK and the USA, the adoption rates appear to be rather low. Reasons which are The influence
used to explain the low rate of adoption of blockchain in Ireland are, among others, of
organizational and technology readiness (Beck, Becker, Lindman and Rossi, 2017) and lack organizational
of blockchain awareness (Fitzpatrick et al., 2017). For example, a study conducted by PWC
of 1,300 Irish business leaders reports that 14 percent of Irish survey respondents claimed factors
that they are either very or extremely familiar with blockchain, compared to 24 percent
globally. Further, the survey highlights how 45 percent of Irish business leaders say 1459
blockchain is not part of their strategic plans and 23 percent of respondents are either in the
early stages of evaluating the technology or experimenting with it compared to 30 percent
globally (Fitzpatrick et al., 2017).
To investigate the low rate of adoption of blockchain in Ireland, we operationalized
innovation theory, which has been extensively used to examine technology innovation
adoption in organizational studies (Rogers, 1995; Yu and Hang, 2010; Van de Weerd et al.,
2016). IT innovation can be defined as the application of a new IT by an organization,
individual or unit (Swanson, 2004). According to Wang (2009), one question that is central to
research on IT innovation is “Why do some information technologies come to be applied
widely among organizations, while others do not?” Information systems (IS) researchers
have studied IT as organizational innovations and have identified various, organizational,
technological and environmental factors which contribute to an organization’s decision to
adopt or not adopt an IT innovation. A framework that is often used to investigate the
intention of an organization to adopt an IT innovation is the technology–organization–
environment (TOE) framework (Tornatzky and Fleischer, 1990).
The adoption of a IT innovation can result in significant transformation to an
organization’s internal and external operations (Wang, 2009; Kaganer et al., 2010).
Consequently, organizations should thread carefully when deciding to adopt such
innovations. As a result, a large body of research has focused on specific organizational
factors that impact company’s decisions concerning IT innovation adoption (Wang, 2009).
Organizational factors that are considered significant determinants of IT innovation
adoption include organizational readiness, top management support, innovativeness,
organizational size, culture, prior IT experience and business model readiness (Damanpour,
1991; Law and Ngai, 2007; Jang, 2010; Yang et al., 2015).
Blockchain is frequently referred to as one of the primary IT innovations that possesses
the potential to disrupt and reshape a plethora of industries (e.g. insurance, financial, legal,
sharing economy etc.) in the next decade (Tapscott and Tapscott, 2016; Puschmann and Alt,
2016). Motivated by blockchain’s potential to transform sociotechnical systems, the lack of
systematic inquiry pertaining to blockchain studies from an IS perspective and the theories
presented earlier, we propose the following research question:
RQ1. How do organizational factors influence blockchain adoption in organizations
based in a developed country?
Specifically, this study elucidates the impact of organizational factors on the adoption of
blockchain and the adoption of blockchain in companies based in Ireland.
The paper proceeds as follows. The next section discusses the determinants of
blockchain adoption. The research method is presented next, followed by the data analysis
and results. Next, we discuss our research findings. The study concludes with discussion of
contributions of the study from the research and practice perspectives.

2. Blockchain adoption
2.1 Blockchain definition and characteristics
IT innovations are now part of the popular business lexicon. Given the significant impact of
IT innovations on organizations, IT innovation adoption has regularly been put under the
IMDS spotlight over the past decades. There is a wealth of research demonstrating how IT
119,7 innovations can influence every facet of a company and can lead to enhanced innovation,
growth, performance, profitability efficiency and productivity (Barrett et al., 2015; Plewa
et al., 2012; Christensen et al., 2015).
For this study, we define blockchain as an open-source data set, distributed across millions
of computers, utilizing avant-garde cryptography (Tapscott and Tapscott 2016). Each block in
1460 the chain is an acknowledgment by network participants that the transaction took place and
was not fraudulent. Each block contains information from the previous block, thus ordering
chronologically, creating a chain of blocks (Nakamoto, 2008). Blockchain is anticipated to
disrupt a multitude of industries in the next decade (Ito et al., 2017; Li et al., 2018). Blockchain
provides adopters with advantages such as anonymity (Zyskind and Nathan, 2015);
immutability (Pilkington, 2016); transparency (Kosba et al., 2016); security (Mendling et al.,
2017) and fast transactions (Kiayias and Panagiotakos, 2016).
Case study research investigating blockchain technology should also consider and reflect
on the unique characteristics of blockchain (Beck, Avital, Rossi, Thatcher, 2017). Table I
provides an overview of the six unique characteristics that encapsulate blockchain
technologies. These characteristics may not apply equally to all categories of blockchain
applications. For instance, according to Treiblmaier (2019) private permissioned blockchains
run by a private consortia of organizations encapsulates a type of closed ecosystem
encompassing pre-defined membership with clearly defined governance structures that in
some instances could be described as being centralized. In contrast, public permissioned
blockchains such as Bitcoin and Ethereum encapsulate open-ecosystems that can be accessed

Characteristic Definition Positive Negative

Access privileges: Both instances describe the Public: accessibility, and Public: transaction
permissioned and level of public access to data. decentralized cooperation. performance, governance
permissionless In public permissioned Private: transaction issues, data privacy,
blockchains, there are no performance, defined security and anonymity.
restrictions on reading data. governance structures, Private: cost, censorship,
Private permissionless innovation speed, data regulation and trust
blockchains restrict access privacy, security and
to pre-defined users anonymity
Immutability Transactions cannot be Traceability and Inflexibility pertaining to
altered/deleted once added business value the deletion/altering of data
to the blockchain
Transparency Blockchain facilitates read- Efficient and accurate Data privacy
only access to transactions record keeping
and the inspection of smart
contracts contents
Programmability Programmable blockchains The deterministic execution Non-programmable
such as Bitcoin and of smart contracts blockchains and the
Ethereum use scripting complexity of coding real
languages to write digital world contracts into
smart contracts blockchain smart contracts
Decentralized The elimination of a central Disintermediation and the Energy consumption,
consensus authority/broker with creation of new power governance issues and
innovative consensus structures security vulnerabilities
protocols
Distributed trust Blockchain does not Trust-free systems Elimination of personal
Table I. necessitate high confidence relationships
Blockchain levels in single authorities
characteristics Sources: Treiblmaier (2019) and Clohessy et al. (2018)
by anybody. These aforementioned access privileges have consequences pertaining how the The influence
characteristics outlined in Table I manifest for blockchain applications. In addition, some of of
these characteristics are still being contested in the academic and practitioner literature (Ito organizational
et al., 2017). There are also characteristics that we have not included such as the chronological
time stamping of data and cryptography mechanics “since those are usually a means to an factors
end” Treiblmaier (2019).
1461
2.2 Determinants of blockchain adoption
According to Rogers (1995), an innovation is “an idea, practice or object that is perceived as
new by an individual or other unit of adoption” (p. 11). Whereas innovation can allude to
something abstract, like an idea, it can also manifest through new technology. An
organization’s decision to adopt a IT innovation can be conceptualized as “a decision to
make full use of an innovative IT as the best course of action available” (Rogers, 1995, p. 21).
Many theories in the IS field have been used to identify specific factors that significantly
or insignificantly influence the adoption of IT innovations in enterprises. Examples include
the TOE framework (Tornatzky and Fleischer, 1990), the perceived e-readiness model (Molla
and Licker, 2005), the technology acceptance model (Venkatesh and Davis, 2000),
assimilation theory (Armstrong and Sambamurthy, 1999) and theory of reasoned action
(Karahanna et al., 1999).
The main objective of the TOE framework (Tornatzky and Fleischer, 1990) is to identify
technological, organizational and environmental views that influence the adoption of IT
innovations in organizations. These views can provide barriers and incentives to IT
adoption. The technological view encompasses technological factors such as complexity,
relative advantage, privacy, security and compatibility that can affect existing IT systems
in use or the new IT being considered for adoption (Rogers, 1995). The organizational view
refers to the internal factors within an organization such as prior IT experience,
innovativeness, top management support, organizational size, information intensity and
organizational readiness (Weiner, 2009; Wang et al., 2010). The environmental view
encompasses factors which impact an organization’s day-to-day business operations such
as competitive and industry dynamics, government interactions, and regulation (Lippert
and Govindarajulu,2006).
Table II delineates blockchain studies, which outline significant technological,
organizational and environmental factors that influence blockchain adoption. Table II
was created based on a comprehensive literature review (Kitchenham and Brereton, 2013).
An effective literature review not only makes a significant contribution to cumulative
culture but also “creates a firm foundation for advancing knowledge. It closes areas where a
plethora of research exists and uncovers areas where research is needed” (Webster and
Watson, 2002). Our motivation was to produce a well-rounded understanding of blockchain
adoption, which is currently lacking in the IS field by carefully describing and then
contrasting and comparing an array of sources on the topic (Heyvaert et al., 2013). The first
step in the analysis of the literature encompassed the sourcing of relevant research
resources via scholarly databases and manual searches. To ensure the consistency and
reliability of the search and data collection process, a three-stage literature mapping
protocol was used as prescribed by Kitchenham and Brereton (2013) to search, select,
appraise and validate the literature. This mapping protocol ensured that no relevant
literature was overlooked which may have been categorized under different headings. This
protocol also helped the researchers to define the boundaries in which the review was
conducted (e.g. inclusion and exclusion criteria). For the initial Stage 1, a rigorous search of
seven prominent databases was conducted to produce a research resource set which was
representative of the status of blockchain adoption research: EBSCOhost, JSTOR, ProQuest,
Google Scholar, PubMed, Scopus and Web of Knowledge. We selected these specific
IMDS No. Author Technological factors
119,7
1 Wang et al. (2016) Perceived benefits*, data
security*, data integrity,
complexity*,
compatibility*, technology
maturity*, uncertainty
1462 2 Lansiti and Relative advantage*, cost
Lakhani (2017) savings, complexity*,
accessibility, trialability,
compatibility*
3 Guo and Liang Cost, data security*,
(2016) privacy, relative
advantage*, business
concerns*, compatibility*, information intensity
complexity*,
disintermediation*
4 Crosby et al. (2016) Perceived benefits*,
complexity*, relative
advantage*, privacy, data organizational readiness*, competitive pressure*,
security
5 Swan (2015) Complexity*, relative
advantage*, data
security*, privacy,
disintermediation*
6 Shrier et al. (2016) Complexity*, relative
advantage*, perceived
benefits*, legacy
infrastructure,
compatibility*
7 O’Dair et al. (2016) Relative advantage*,
perceived benefits*,
complexity*,
compatibility*, data
governance,
disintermediation*
8 Folkinshteyn and Data security*, privacy,
Lennon (2016) perceived benefits*,
disintermediation*, cost
savings, continuity of
service
9 Tapscott and Perceived benefits*, data Organizational readiness*, Government support*,
Tapscott (2016) security*, privacy,
technology maturity*
10 Mendling et al. Data security*, latency,
(2017) throughput, usability, hard organizational size*,
forks, wasted resources
11 Pilkington (2016) Perceived benefits*,
complexity*, technology
maturity*, compatibility*, participation incentives*,
Table II.
Significant blockchain
TOE adoption factors
Organizational factors Environmental factors
Organizational size*, top Regulatory environment*,
management support*, industry pressure*,
organizational readiness*, market dynamics*
responding capability
Technology readiness*, Competitive pressure*,
organizational size*, top relationship with
management support*, partners, government
value chain readiness policy, business use
cases*
Organizational readiness*, Market dynamics*,
top management support*, government support*,
blockchain knowledge, regulatory environment*,
industry standards*
Customer relationship, top Government support*,
management support*, regulatory environment*,
organizational size* trading partner pressure*
Technology readiness*, Regulatory environment*,
organizational readiness*, public perception of the
business model readiness*, industry standards*,
relative advantage market dynamics,
government support*
Organizational readiness*, Regulatory environment*,
organizational size*, top governmental support*
management support*,
employee disruption
Blockchain knowledge, Emergence of use case
organizational size*, examples, government
organizational readiness*, regulation*, market
business model readiness dynamics, critical user
mass*
Organizational readiness*, Market dynamics*,
customer relationship, size, trading partner support*,
top management support* regulatory environment*
organizational size*, market standards,
business model readiness*, regulatory environment*
blockchain knowledge*
Organizational readiness*, Regulatory environment*,
market dynamics,
governance, business competitive pressure*
models, top management
support*
Organizational size*, top Competitive pressure*
management support,
(continued )
No. Author Technological factors Organizational factors Environmental factors
The influence
of
permissions (public vs innovativeness*, organizational
private blockchains)* technological readiness*
12 Morabito (2017) Complexity, perceived Technological readiness, Regulatory environment*, factors
benefits, compatibility*, innovativeness*, value government support,
maturity*, cost chain readiness*, top business use cases*,
management support and trading partner support* 1463
involvement*, size
14 Seebacher and Perceived benefits*, smart Technology responding Industry pressure*,
Schüritz (2017) contract coding*, capability, information business use cases*
complexity intensity, organizational
readiness*, value chain
readiness*
15 Lindman et al., Complexity*, perceived Technology readiness*, Regulatory environment*,
(2017) benefits*, technology Value chain readiness, market dynamics*
maturity*, compatibility, business models,
technology architecture* organizational readiness*
16 Chen et al., (2018) Perceived benefits*, Top management Market dynamics,
complexity*, smart support*, organizational governmental projects,
contract coding*, energy readiness* Industry pressure*
consumption
17 Beck, Becker, Smart contract coding*, Incentive structures, Market dynamics*,
Lindman and permissions*, security, governance mechanisms, regulatory environment*
Rossi (2017) architecture (centralized vs accountability decision
decentralized)*, privacy rights (management vs
control), business models*
18 Zamani and Security, perceived Value chain readiness, Business use cases*,
Giaglis (2018) benefits*, smart contract business models*, market standards,
coding*, complexity* organizational readiness*, regulatory environment*
organizational size*, top
management support*,
innovativeness
19 Woodside et al., Perceived benefits*, Business models*, Regulatory environment*,
(2017) security, complexity*, innovativeness* market dynamics*
privacy, compatibility, cost
20 Kokina et al. (2017) Data security, smart Incentive structures*, Regulatory environment*,
contract coding*, business models, quality market standards
perceived benefits*, assurance
scalability, permissions*
Note: *Factors found to be significant Table II.

databases because of the multidisciplinary nature of blockchain research. Furthermore,

these databases have been used by IS researchers as sources for other systematic literature
reviews (Vom Brocke et al., 2015). We used the search strings “blockchain” “adoption”
“TOE” and “bitcoin” “adoption” “TOE.” We included both theoretical and empirical studies
and extracted significant factors that influenced blockchain adoption. Given the existing
dearth of academic research pertaining to blockchain adoption, gray literature research
resources (e.g. conference proceedings, research reports, issue papers, white papers) were
also included. All research resources were imported directly into an EndNote database.
Using EndNote’s “find duplication” feature 70 duplicates were removed. The remaining
research sources were further filtered. The selection processes encompassed a decision-
making process to include or exclude relevant research papers from the data extraction
process. For instance, in terms of exclusion criteria, Stages 2 and 3 resulted in the removal of
research articles that were extraneous to the research question, further duplicates not
IMDS initially picked up by EndNote (e.g. surnames and first names misplaced), materials no
119,7 longer accessible, questionable sources (e.g. credibility of resource could not be verified) and
research sources where blockchain was only briefly mentioned and was not the main theme
of the content. The final selection decision took place when the research sources were read in
parallel with data extraction and quality assessment. Stage 3 search and selection took place
in parallel with data and quality extraction from the research sources identified in Stages 1
1464 and 2 and comprised three main tasks: search process validation, backwards snowballing
and researcher consultation (Kitchenham and Brereton, 2013).
Next, the final 20 research articles were systematically full text reviewed and coded
(Strauss and Corbin, 1998) to create Table III, which present an overview of salient blockchain
adoption considerations. Following the recommendations of Ritchie et al. (2003), a multistage
hierarchical data analysis approach was used comprising four analytical cycles that
incorporated open and axial coding techniques based on the recommendations of Strauss and
Corbin (1998). The hierarchical data analysis procedure used was an iterative process
whereby as “categories are refined, dimensions clarified, and explanations are developed,
there is a constant need to revisit the original or synthesized data to search for new clues, to
check assumptions or to identify underlying factors” (Ritchie et al., 2003, p. 213). The primary
analytical cycle comprised a process of open coding which was used to identify codes from the
research resource title, keywords, abstract and content. Open coding “involves analyzing the
text (e.g. a sentence or paragraph) and summarizing this text by the use of a succinct code”
(Myers, 2013, p. 107) and requires scrupulous familiarization and interrogation of the data.
Due to the tentative nature of the codes and concepts that emerged during the initial stages of
the open coding process, the researchers constantly compared the qualitative data for
similarities and variations (Myers, 2013). As the analysis advanced, the codes and concepts
became more conclusive and definitive. In the secondary analytical cycle, axial coding was
used to reassemble the data that were fractured during the open coding phase by identifying
causal conditions and relationships between the concepts and categories (Strauss and Corbin
1998). Axial categories and subcategories were developed through a coding paradigm of
causal conditions, strategies, context or intervening conditions. This stage of the analysis also
focused on intended and unintended consequences. The coding process continued until the
categories were theoretically saturated (Strauss and Corbin, 1998). The tertiary analytical
cycle encompassed a process of triangulation and peer debriefing. To confirm
representativeness, once the coding was completed, the resulting themes (e.g. business
model readiness, organizational readiness, technological readiness) were juxtaposed and
triangulated in order to elucidate similarities and differences. Peer debriefing enabled us to use
external groups as a soundboard for further validating the final set of themes which emerged

Technological factors Organizational factors Environmental factors

a
Perceived benefits 13 Organizational readiness 13 Regulatory environmentb 15
Complexity 12 Top management support 9 Market dynamicsc 11
Compatibility 8 Organizational size 9 Industry pressured 5
Data security 6 Business model readiness 7 Government support 5
Smart contract coding 6 Technology readiness 3 Business use cases 4
Maturity 5 Innovativeness 3 Trading partner support 3
Relative advantage 4 Participation incentives 3 Critical user mass 1
Disintermediation 4 Blockchain knowledge 1
Table III. Permissions (public vs private) 3
Summary of Architecture 2
significant blockchain Notes: aIncludes value chain readiness; bincludes government regulation; c
includes competitive pressure;
d
adoption factors includes industry standards
from the analysis (Schwandt et al., 2007). For instance, one of the outcomes of the peer The influence
debriefing was to place similar considerations under a common themed consideration of
(see Table III). Table II enabled us to extract specific variables that were found to be organizational
significant in at least one of the studies, denoted by*. This process enabled us to then create
Table III, which provides a summary of the variables according to the number of times that factors
were found to be significant.
As can be seen in Table III, three organizational factors emerged as being significant
from the blockchain literature: organizational readiness (N ¼ 13), top management support
1465
(N ¼ 9), and organizational size (N ¼ 9). Our findings support extant IT innovation adoption
literature, which have identified how organizational factors such as top management
support, firm size, prior IT experience, innovativeness and organizational readiness are
often viewed as the most significant determinants of IT innovation adoption in enterprises
(Kimberly and Evanisko, 1981; Tornatzky and Fleischer, 1990; Damanpour, 1991). These
organizational factors have also been widely examined to ascertain the degree to which they
constrain or act as a catalyst for the adoption of IT (Grandon and Pearson, 2004; Van de
Weerd et al., 2016). Given the dearth of empirical research into the organizational factors in a
blockchain IT innovation adoption context, we have intentionally narrowed the scope of the
study whereby we have focused only on the top three organizational factors, which emerged
from the literature (Table III). We now provide further information in the following section
on these three specific factors. For each factor, we discuss the existing literature from
general innovation adoption and blockchain perspectives and provide a definition that we
used in our research.
2.2.1 Top management support. Top management support has been identified as a key
recurrent factor critical to the adoption of IT innovations (Sabherwal et al., 2006; Bajaj, 2000;
Dong et al., 2009; Kulkarni et al., 2017). According to Jarvenpaa and Ives (1990), “few nostrums
have been prescribed so religiously and ignored as regularly as top management support in
the development and implementation of IT.” For this study, we define top management
support as “managerial beliefs about technological initiatives, participation in those initiatives,
and the extent to which top management advocates technological advancement” (Kulkarni
et al., 2017). High levels of top management support for a specific IT innovation ensure the
long-term vision, commitment and optimal management of resources, creation of a favorable
organizational climate, support in overcoming barriers and resistance to change (Wang et al.,
2010; Gangwar et al., 2015). In the context of blockchain adoption, top management support
plays an important role because blockchain adoption may involve new regulatory
requirements, a high degree of complexity, the acquisition of new resources, the integration of
resources, the re-engineering of business-to-consumer and business-to-business transactions
and information exchanges and the development of new skills and competencies (Swan, 2015;
Pilkington, 2016; Lansiti and Lakhani, 2017).
2.2.2 Organizational readiness. Organizational readiness is conceptualized as the
availability of specific organizational resources to adopt new IT innovations (Lacovou et al.,
1995; Weiner, 2009; Wang et al., 2010). This conceptualization is frequently categorized
under several headings, including human resources, financial and infrastructure facets.
Human resources facets refer to the presence of employees with the requisite knowledge,
skill and experience to adopt new IT innovations (Wang et al., 2010). Next, financial facets
refer to the allocated financial resources an organization commits to new IT innovations
(Weiner, 2009). While certain research has focused on the financial resources from the
perspective of a specific IT innovation (e.g. Lacovou et al., 1995), in general, many studies
have focused on financial resources from the perspective of any new IT innovation. Finally,
infrastructure facets refer to existing IT platforms on which new IT innovations can be
developed (Lacovou et al., 1995). When organizational readiness for a new IT innovation is
high, an organization’s management and staff are more likely to initiate change, exhibit
IMDS greater effort and persistence, and engage in enhanced cooperative behavior (Weiner, 2009;
119,7 Wang et al., 2010). Consequently, this results in a more effective adoption of the new IT
innovation. The exact influence of organizational readiness on the adoption of blockchain is
currently unclear. While existing theoretical research suggests that organizational readiness
has a significant influence on the adoption of blockchain (Swan, 2015; Wang et al., 2016),
there is currently a dearth of empirical studies that have confirmed that this is the case.
1466 In line with Weiner (2009) and Wang et al. (2010), we define organizational readiness as
the availability of employees with the requisite IT knowledge and skills; financial resources
for adopting IT innovations (e.g. IT budget); and infrastructure on which blockchain
applications can be built.
2.2.3 Organizational size. To categorize the size of each case company in our study, we
enlisted the use of a quantitative measurement approach in line with the organizational size
definition provided by the World Bank (Kushnir et al., 2011). The number of employees in
the organization determines this approach. For instance, micro enterprises are classified has
being organizations with 1–9 employees; small enterprises, with 10–49 employees; medium
enterprises; 50–249 employees and large organizations with ≥250 employees. For this study,
we simplified the classifications as medium-sized enterprises (SMEs), with 1–249 employees
(micro, small and medium enterprises) and large enterprises, with ≥ 250 employees. Many
past studies suggest that an enterprises’ willingness to adopt a new innovative IT is
positively influenced by organizational size (Damanpour, 1992). The reasoning behind this
is that large organizations possess more complex and diverse facilities which positively
contribute to adoption (Lee and Xia, 2006). Micro enterprises and SMEs, on the other hand,
are susceptible to many barriers which constrain their ability to adopt IT innovations such
as resource poverty (e.g. lack of IS personnel and expertise) and small IT budgets (Thong
and Yap, 1995). However, our research indicates that in the case of specific IT innovations,
because of the characteristics of the technology and the flexibility and adaptability of micro
enterprises and SMEs, the opposite has been found. For example, empirical studies have
shown that SMEs were more suitable and more inclined to adopt cloud computing
technologies (Carcary et al., 2014; Van de Weerd et al., 2016). While organizational size is
considered an important predictor of blockchain adoption (Tapscott and Tapscott, 2016;
Mendling et al., 2017) further empirical research is necessary to establish a consistent
relationship between organizational size and blockchain adoption.

2.3 Blockchain adoption in Ireland

The Bloomberg innovation index for 2018 ranked Ireland in 13th place. The index scores
countries using seven criteria, including research and development spending, concentration
of high-tech public companies and patent activity ( Jamrisko and Liu, 2018). A report by the
IDA (2017) has identified Blockchain as a strategic priority and commenced a number of
initiatives to establish Ireland as a European center for financial technology (FinTech) and
Blockchain development. Currently, much of the blockchain adoption is taking place in the
FinTech services sector. For instance, Deloitte selected Dublin as a base for their EMEA
blockchain lab. Additionally, the Blockchain Association of Ireland has pioneered a “Crypto
Coast” program, which has established a network of blockchain and cryptocurrency
companies along the east coast of Ireland. Furthermore, a report by the blockchain
association of Ireland (Moeller and Schwerin, 2017) highlighted how Irish organizations are
investigating the potential use of blockchain technology as a means of compliance with the
general data protection regulation (GDPR). In 2017, there was a 41 percent increase in
blockchain-related roles when compared to 2016 (Moran, 2018). These roles encompassed
various employment opportunities including accounting, software development, data
science, management consulting, business analysis and project management.
 However, the majority of blockchain developments are taking place within a small The influence
network of organizations (IDA, 2017). Moreover, we have also highlighted the low levels of of
blockchain awareness in Ireland and how blockchain adoption in Ireland is stagnant organizational
(Fitzpatrick et al., 2017).
factors
3. Method
In this study, we followed a multiple-case study approach to investigate how organizational 1467
factors influence the adoption of blockchain in Irish organizations. Our research adopted an
organizational perspective; that is, we focused our research on three organizational
perspectives that were defined a priori (Eisenhardt, 1989), based on extant studies: top
management support, organizational size and organizational readiness. As part of our case
study protocol, we consistently compared our findings with existing theory. We then used
this analysis to enhance our understanding of the adoption process and to elucidate why
Irish companies adopt blockchain or not. Furthermore, this multiple-case design enabled
direct replication across cases with contrasting situations that allowed us to draw more
powerful analytical conclusions (Yin, 2014).
With our research method, our intention was to understand why the adoption of
blockchain in Ireland is relatively low and to ascertain how the several organizational
factors influenced company’s adoption decision. Therefore, we used an exploratory case
study to determine how or why a certain condition (the adoption or non-adoption of
blockchain) came to be.
In the remainder of this section, we discuss our case selection process. We then present
how we collected, processed and analyzed the data.

3.1 Case selection

All the cases in our research were located in Ireland. The country is of particular interest for
the following reasons. First, Ireland has the typical characteristics of a developed country
and particularly a developed country in the European Union region ( Jamrisko and Liu,
2018). Consequently, the findings of this study may be relevant for similar developed
countries.
We used our own networks to identify case companies. Cases were selected using literal
and theoretical replication (Benbasat et al., 1987; Yin, 2014). We used literal replication to
corroborate cases with similar characteristics, thus enhancing the reliability and strength of
our study (Yin, 2014). Theoretical replication predicts contrasting results for predictable
reasons. For example, we were interested in organizations with sufficient organizational
readiness and those with insufficient readiness or how SMEs and large organizations differ
in their adoption of blockchain.
We used three variables organizational size, organizational readiness and top
management support as well as an outcome variable (adopt or not adopt) as conditions
to determine the types covered by our theoretical replication. We identified the types based
on the size condition (SME and large organizations) at the start of the study. The remaining
types (organizations which adopted blockchain vs those that did not, organizations with
sufficient top management support vs insufficient management support and, organizations
with sufficient vs insufficient organizational readiness) were delineated after the interview
process and the initial data analysis stage. Cases were added to the study following the
initial analyses in stances where there were not at least two individual cases in each
subgroup that complemented the theoretical contributions across the groups with literal
replication within each group.
We sent a letter of permission to a total of 50 companies, of which 20 agreed to be
participants in our study. Four interviewees were interviewed twice, namely A2, A4, A15
and A19. Of these 20 companies, 8 had adopted blockchain and 12 had not or did not intend
IMDS to adopt blockchain in the next two years. Information relating to our case study companies
119,7 is presented in Table IV. The cases are divided over different sectors (e.g. financial, IT,
education, fishing, gaming, legal, marketing and mobile app development).

3.2 Data collection and interview process

Data collection took place between May and November 2017. Most of the interviews
1468 (n ¼ 15) were conducted in the case organization’s offices and lasted between 60 and
120 min. The five remaining interviews were conducted via Skype. We selected
interviewees who had a main role to play in IT adoption decisions within each
organization. As can be seen from Table IV, the interviewees came from different
management functions and included IT managers, company owners, researchers and
directors. Thus, our study reflects a broad range of expertise and knowledge in
blockchain. Prior to the interviews, each interviewee was emailed a research information
sheet that provided a background to the study.

Existing
blockchain IT Blockchain
No. Industry applications Employees staff Locations Total assets Position awareness

A1 IT None 10 8 1 o €100,000 Owner Basic

A2 Education Identity 16,000 30 1 €100,000o$3m IT Medium
authentication manager
A3 IT Supply chain 6,000* 276 348 o €20m IT High
monitoring manager
A4 IT Supply chain 244 26 2 €100,000o$3m IT Medium
Manager
A5 Education None 2,439 20 1 €100,000o$3m IT Basic
manager
A6 IT Supply chain 2,600* 1,240 425 o €20m IT High
manager
A7 Gaming None 750 10 165 o €20m IT High
manager
A8 Mobile app Crypto 20 16 1 €100,000o$3m CEO High
development payment
applications
A9 IT Identity 8 3 1 o €100,000 Owner High
management
A10 Legal None 600 17 4 o €20m FinTech Medium
researcher
A11 Financial None 2,500* 55 7 o €20m IT Medium
director
A12 Financial None 210 8 3 €100,000o$3m IT Basic
manager
A13 IT None 12 7 1 €100,000o$3m CEO Basic
A14 Marketing None 15 2 1 o €100,000 Owner Basic
A15 IT Multiple 3,000* 2,130 3 o €20m IT High
manager
A16 Fishing None 7 2 1 €100,000o$3m CEO Medium
A17 IT None 6 4 1 o €100,000 Owner Medium
A18 IT None 344 23 3 o €20m CIO High
Table IV. A19 Financial Multiple 9,546 367 220 o €20m CIO High
Summary of case A20 Financial None 1,423 210 1 o €20m FinTech High
organizations researcher
To ensure that we obtained all the required information from our interviewees we used semi- The influence
structured interviews as our primary data collection method (Pare, 2004; Yin, 2014). We used of
an interview guide, which evolved during the interviewing process. The initial research organizational
questions contained within the interview guide were refined following a number of pilot
tests involving a similar cohort to the interviewees selected for the main study. We also used factors
senior researchers and academics as a soundboard to identify flaws, limitations and other
weaknesses contained within our interview guide. 1469
We complemented our interviews with field notes and documentation (e.g. annual reports,
corporate websites, white papers etc.) which we obtained online. Prior to commencing the
interviewing sessions, we introduced ourselves and provided an additional backdrop to the
study explaining the purpose of our research. Questions were then administered via the study
interview guide. The interview topics included the case organizations background, the
interviewee’s awareness of blockchain and the impact of the three organizational factors on
blockchain. For example, the following examples illustrate a sample set of questions that were
asked concerning top management support:
What are your and senior management’s perceptions of blockchain?
Have you or do you plan to adopt blockchain. If yes, can you discuss why and how you adopted the
technology?
Have you encountered any blockchain negativity from senior management within your organization?
Can you provide examples of various resources that you or senior management have made
available with respect to blockchain adoption?
Additionally, in order to clarify the interviewee’s awareness of blockchain, the interviewer
asked if the interviewee had ever heard of blockchain and, if so, could they explain what
blockchain was. The interviewer then provided our definition of blockchain (see Section 2)
and provided several examples of blockchain in business and personal settings. Once there
was an agreed understanding of blockchain for the interview, the interviewer proceeded
with the remaining questions.
In line with the study’s confidentiality policy, all interviewees were ensured that they and
their case organizations would be referred to pseudonymously in the study. Consequently,
the identifiers A1–A20 represent the interviewees. All interviews were audio recorded
with the express consent of the interviewees. The interviews were then transcribed and sent
to the interviewees for feedback and approval.

3.3 Data processing and analysis

To assist with the data analysis, we developed six codes to organize the data: blockchain
awareness, organizational readiness, organizational size, top management support,
developed country and blockchain adoption and use. We used NVivo 11 to code our
interviews across these six themes. Table V highlights our coding scheme and provides a
detailed description and an example of each code.
We analyzed the data in four iterations. To make sense of and structure, the initial data
we commenced with a within-case analysis of each of the individual cases. Each case was
analyzed separately in terms of the three organizational variables and outcome variable by
using the interview transcripts, field notes and online documentation. Additionally, we
analyzed the data for evidence related to blockchain awareness level, blockchain adoption
and use and matters that are typical for developed countries.
Using an informal qualitative comparative analysis (QCA) technique, the results of the
within case analysis were processed, and cross-case comparisons made. QCA, a method
which originated in management research, was primarily developed “to solve a fundamental
IMDS Code Description Example
119,7
Blockchain Contains a definition and collection of “Yes, I have been aware of blockchain for the
awareness interviewee responses regarding their past four years […] It is a database which
level awareness level allows different identities to share
information in a secure manner.” (A8)
Top Contains a collection of interviewee “Unfortunately, management are unable to
1470 management responses pertaining to how top see the value that blockchain can bring to the
support management make decisions on the adoption company and as a result we do not have
or rejection of blockchain plans to adopt any specific implementations
soon. The core issue is the lack of business
use cases.” (A10)
Organizational Contains a collection of interviewee “We are currently implementing instances of
readiness responses relating to the impact of the blockchain along our supply chain […]
availability of the required organizational however we are finding it difficult to get our
resources ( financial, IT expertise, and IT SME supply chain partners to implement our
infrastructure) in triggering the adoption or flagship blockchain authentication service
rejection of blockchain […] they are struggling to see how
blockchain sits with within their IT
innovation strategies.” (A3)
Organizational Contains a collection of interviewee “Our IT department is quite large. We have
size responses pertaining to the size of the created a new department which is focused
organization or its IT unit and how this entirely on blockchain research and
influences its adoption decision development.” (A6)
Blockchain Contains a collection of responses relating to “We created a blockchain product on IBM’s
adoption and the (non-) adoption or use of blockchain hyperledger platform.” (A9)
use applications
Developed Contains a collection of interviewee “The Irish government have yet to offer their
country responses regarding matters that are typical voice on blockchain technologies. Until they
Table V. for developed countries (in relation to do, smaller companies like ourselves will be
Coding scheme blockchain adoption) reluctant to enter the playing field.” (A1)

problem presented by cross-case analyses: preserving the integrity of cases as complex

permutations of causal factors while concurrently allowing for the systematic examination
of similarities and differences in causal factors across many cases” (Greckhamer et al., 2008).
These permutations are a specific set of factors (in this study, organizational variables) that
produce a given outcome of interest (in this study, the adoption of blockchain). In the field of
IS, QCA has been used for exploring permutations of factors that explain service
innovations (Ordanini et al., 2014) and has, for example, been used to explain IT innovation
outcomes (Fichman, 2004). Next, we carried out a cross-case analysis using the QCA results
as a basis for identifying varying permutations that led to either the adoption of non-
adoption of blockchain. The goal of this phase of data analysis was to find similarities
across our 20 cases, which would allow us to draw conclusions based on the impact of our
three organizational variables (Yin, 2014).

3.4 Validity
The quality of empirical case study research designs can be judged according to four logical
tests: construct validity, external validity, internal validity and reliability. To ensure
construct validity we, first, used multiple sources of evidence (online documentation, field
notes and interview transcripts, second, summarized the interview reports for each case and
elicited feedback and approval from the interviewees and, third, established a chain of
evidence. External validity establishes the domain to which a study’s findings can be
generalized. The steps we took to ensure this category of validity were that we used a
multiple-case study design in which we used a replication logic to ensure the generalizability
of our findings. An explanation-building analytic technique was used to ensure internal The influence
validity. Internal validity establishes a causal relationship. Finally, we ensured the of
reliability of our research by using a case study protocol and a case study database which organizational
ensures that the main operations of the study such as the data collection procedures can be
repeated (Miles et al., 2013; Yin, 2014). factors

4. Results 1471
We present our findings in three sections. First, we elucidate on the results of our within-
case analysis. Next, we elaborate on our QCA to demonstrate how the various cases scored
on the three organizational variables in relation to the outcome variable, blockchain
adoption. Finally, we present our across-case analysis, where we discuss the patterns we
identified. These are patterns are supported by interview quotes to reveal additional
information about the respondents’ perceptions regarding these variables.

4.1 Within-case analysis

The within-case analysis necessitated a thorough analysis of each individual case based on
the outcome variable (adoption or non-adoption of blockchain) and the three organizational
variables. We also analyzed other case details, such as awareness of blockchain and any
other characteristics which manifested. The analysis of the three organizational variables
consisted of two segments: first, we investigated various methods that allowed us to allocate
a value to each of the variables. For organizational size, this was straightforward: This
specific variable was sourced from the respondents’ interviews or from the case
organizations websites or annual reports. Similarly, the assessment of the outcome variable
(adoption or non-adoption of blockchain) was also determined based on the data elicited
from the respondents. However, organizational readiness and top management support
were more challenging to determine. Concerning top management support, this variable was
determined based on how each of the interviewees viewed the perception of this variable
within their organization. For example, the interviewee in A6 described how the
development of brand new research and development facility dedicated solely for the
blockchain projects was clear evidence of the company’s top-level management support. For
organizational readiness, we investigated three specific concepts: human resources,
financial resources and IT infrastructure. Our analysis of the interviews and the
supplementary documentation enabled us to determine if these conditions were present or
not. For instance, the interviewee in A1 elaborated how they had insufficient financial
resources, no employees with the necessary skills and experience and insufficient IT
infrastructure to develop blockchain technologies:
Being a start-up organization, my budget is quite restrictive in term of IT expenditure. I have one
senior IT manager and three interns who work on a part-time basis who do not possess the
requisite blockchain competencies. We have carried out an inventory of the infrastructure that we
would need for blockchain development and unfortunately I cannot justify a budget for researching
the potential of blockchain for our company. (A1)

4.2 Qualitative comparative analysis

Using a QCA approach, which in line with Rihoux and Ragin (2009), we present Table VI and
Table VII that provide overviews of how the case organizations were categorized according to
top management support, organizational readiness, organizational size and the outcome
variable, blockchain adoption. First, we categorized each variable with either a 1 which
indicated that a given condition was present or a 0 indicating that a given condition was
absent. We coded large organizations with a 1 and SMEs with a 0 given large organizations
are more likely to adopt an IT innovation ( Joo and Kim, 2004; Anderson et al., 2014).
IMDS Case Organizational size Organizational readiness Top management support Blockchain adoption
119,7
A1 0 0 0 0
A2 1 1 1 1
A3 1 1 1 1
A4 0 1 1 1
A5 1 0 0 0
1472 A6 1 1 1 1
A7 1 0 0 0
A8 0 1 1 1
A9 0 1 1 1
A10 1 0 0 0
A11 1 0 0 0
A12 0 0 0 0
A13 0 0 0 0
A14 0 0 0 0
A15 1 1 1 1
A16 0 0 0 0
A17 0 0 0 0
Table VI. A18 1 0 0 0
Case organization A19 1 1 1 1
data set A20 1 0 0 0

Organizational Organizational Top management Blockchain Blockchain not

Permutation size readiness support adopted adopted

A: 000 0 0 0 6
B: 001 0 0 1
C: 010 0 1 0
D: 011 0 1 1 3
E: 100 1 0 0 6
Table VII. F: 101 1 0 1
Case permutation data G: 110 1 1 0
set summary H: 111 1 1 1 5

We adopted the same procedure for the other two variables top management support and
organizational readiness. The data from the within-case analysis were used to assign values to
the three variables, as illustrated in Table VI.
Table VI provides a summary of our case organization data set. This data set enabled us
to create Table VII which illustrates eight possible permutations of the three organizational
variables which influenced an organization’s decision adoption of blockchain. Four of these
permutations were found in our data set, specifically A, D, E and H. Interestingly, the
permutations that lead to blockchain adoption was permutation D, containing SMEs with
sufficient organizational readiness and top management support, and permutation H,
comprising large organizations with sufficient organizational readiness and top
management support. Permutations A and E did not lead to blockchain adoption. Four
permutations were not identified in our study, namely permutation B (SMEs with sufficient
management support that have adopted blockchain), permutation C (SMEs with sufficient
organizational readiness that have adopted blockchain), F (large organizations with
sufficient top management support that have adopted blockchain) and permutation G (large
organizations with sufficient organization readiness but lacking top management support
that have adopted blockchain).
 In addition to Table VII, we created Figure 1 that illustrates a set-theoretical The influence
representation of our case organization data set. Case organizations that adopted blockchain of
are marked in gray. SMEs who did not have sufficient organizational readiness, top organizational
management support and who did not adopt blockchain are situated outside of the three
circles. As can be seen from this representation, five large organizations and three SMEs factors
with sufficient top management support and organizational readiness adopted blockchain.
Consequently, our data suggest that sufficient top management support and organizational 1473
readiness represent the strongest predictors for the adoption of blockchain. Additionally,
most large organizations except for A2, A3, A6, A15 and A19, did not adopt blockchain.

4.3 Across-case analysis

Table VIII provides a summary of our findings with regards to how the selected three
organizational factors cumulatively determined the rationale for both large organization’s
and SME’s decisions to adopt blockchain or not to adopt blockchain. This table also outlines
the type of blockchain applications deployed. Consequently, in this section, we present our
findings in relation to the interviews and discuss the common indicators that help us explain
the impact of organizational variables on the adoption of blockchain. First, we present an
overview of the blockchain awareness of the study’s interviewees. Next, we delineate our
findings in relation to the three variables: organizational readiness, organizational size and
top management support. We conclude with an overview of our findings in relation to
Ireland as a developed country.
4.3.1 Blockchain awareness level. The main themes emanating from the blockchain
definitions provided by the interviewees revolved around immutable and distributed
ledgers, trust, reduced costs, increased speed, reduced fraud and risk, and increased security
and traceability. We created the following three categories of groups to classify the
blockchain awareness of our interviewees: first, basic level – the interviewee has heard
about blockchain technology but is unable to adequately describe the concept; second,
medium level – the interviewee has heard about blockchain technology and is able to give an
accurate description of blockchain; and third, high level – the interviewee is able to give a
correct description of blockchain technology and can provide real world examples of
blockchain applications.
We identified that 5 out of the 20 interviewees had only a basic level of blockchain
awareness. Six of the interviewees had a medium level awareness of blockchain. Only nine

Organizational
Top
Size
Management
A5 A7
Support
A10 A11

A18 A20

A2 A3

A6 A15 A4
Bolckchain adopted
A8
A19
Bolckchain not adopted
A9

A1 A12
Figure 1.
Organizational Set-theoretical
A13 A14 Readiness representation of case
organizations
A16 A17
IMDS Organization Adopted – deployment and rationale Non-adopting rationale
119,7
Large Multiple instances of fully deployed and Lack of internal IT adoption coordination
functional blockchain applications Blockchain technological complexity
Private permissioned blockchains Lack of specific industry business cases and
Initial blockchain prototyping to create standards
business use cases Lack of government incentives
1474 Availability of cloud-based blockchain Lack of blockchain top management awareness
development tools Lack of internal staff with requisite blockchain
Supply chain transaction innovation skills and competencies
cost reduction Lack of supply chain organizational buy in
enhanced security
enhanced transparency
enhanced efficiency
SMEs Single instance of a fully deployed and Lack of blockchain awareness
functional blockchain application Lack of specific industry business cases
Table VIII. Public permissioned blockchains Challenges sourcing employees with requisite
Summary of main Provision of new innovative services blockchain skills and competencies
blockchain Availability of cloud-based blockchain Challenges sourcing blockchain educational
organizational development tools resources
adoption Availability of publicly available business
considerations use cases

interviewees were able to demonstrate that they had a high level of blockchain awareness.
In other words, most of our interviewees had heard of blockchain technology but were
unable to provide a correct description and provide actual examples of real world
blockchain applications. Most interviewees who were classified as having a basic to medium
level of blockchain awareness (n ¼ 11) described blockchain as being applicable only to the
FinTech industry which is not entirely in line with the definition of blockchain used in this
study. Below is a sample of the interviewee’s responses:
The word that comes to mind with blockchain is security which is associated with identity and the
management of financial activities along the supply chain. (A1)
Blockchain are distributed digital transaction ledgers where you can record and verify transactions
in a secure peer to peer environment. (A3)
Blockchain enables financial organizations to comply with financial regulatory requirements in real
time. (A14)

4.3.2 Top Management support. For the purposes of this study, top management support
refers to a person or a group of people who make decisions or play a key role in influencing
decisions, which result in an organization being able to adopt or not adopt blockchain.
Take for example a CEO of a company who has made the decision that blockchain will
underpin their new payments loyalty system. This CEO then makes the necessary
organizational resources available to support the blockchain adoption process. Eight of
the case organizations demonstrated satisfactory top management support for blockchain.
Furthermore, these eight cases had adopted blockchain. It was evident that top
management were able to recognize the benefits of blockchain technologies, as is
evidenced in the following:
We have been operationalising a number of blockchain strategic initiatives for the past two years.
These initiatives were created by our CEO and board of directors who envision that blockchain is
going to be vital for securing our enterprise cloud and supply chain services. All of future services
will be underpinned by blockchain technology. (A3)
In other cases, top management support for blockchain grew gradually and was influenced by The influence
employees who were able to demonstrate real world value as is illustrated in the following: of
Management were initially reluctant to adopt blockchain despite the obvious benefits. This was organizational
until our senior engineers created an innovative blockchain prototype that could fundamentally factors
restructure our supply chain. Consequently, management created a new department which was
managed by a newly recruited CIO to spearhead the internal development of our private blockchain
projects. (A4) 1475
Interest in blockchain technologies was first initiated by our software engineers who had an
interest in cryptocurrencies. They developed multiple innovative proof of concept blockchain
prototypes which were showcased to lower, middle and then upper management […] in that order
[…] gaining top management support was crucial but it was an evolutionary process […] these
prototypes got them excited. (A6)
It was interesting to note that in the case of four organizations while there was interest
among decision maker influencers for their organizations to pursue blockchain strategies,
their enthusiasm was not reciprocated by top management:
It is so important to get management awareness and buy-in even at a blockchain proof of concept
stage. However, there is zero uptake of anything blockchain here now. Our senior managers and IT
staff have little or no understanding of what blockchain is. It is such a pity as many of our American
competitor are rolling out a number of loyalty programmes and grey/black market supply chain
monitoring and authentication applications which are underpinned by blockchain technologies. (A7)
If I’m to be blunt, most of our top managers here would have a very basic awareness of what
blockchain is. It is still very early days for a lot of people in terms of understanding what it is and
how they are going to use it […] there needs to be more practical use cases before our managers
give the green light for adopting blockchain. (A11)
There is no point having the only telephone in the world if no one else has one. Most of the
blockchain innovation in Ireland is occurring in the Fintech industry that encompasses larger
companies. Where is all of this going in terms of SMEs? A lot will become clearer in 3–5 years’ time
once business and market use cases become available. (A18)
Cumulatively, our findings indicate that top management support had a positive influence on
the adoption of blockchain. It was interesting to note that several of the case interviewees were
researching and developing possible blockchain applications/use cases, ether in their allocated
innovation time slots or in their spare time, to influence senior management decisions.
4.3.3 Organizational size. In this section, we discuss how organizational size influences
an organization’s decision to adopt or not adopt blockchain. We categorized these
organizations as SMEs or large. As illustrated in Table IV, 11 large organizations and 9
SMEs took part in our study. In total, 5 out of the 11 large organizations and 3 out of the 9
SMEs adopted blockchain. If we delve into the primary reasons that the five large
organizations adopted blockchain, the intention to adopt was related to reduced costs,
enhanced security, efficiency and transparency of transactions:
We have been implementing instances of blockchain (e.g. bitcoin rewards, distributed ledgers)
along our supply chain for the past two years. We have reaped numerous logistic advantages in
terms of improved transaction flows, enhanced integral traceability and security. (A3)
Blockchain enables the company to do to financial regulatory compliance in real time. This has
manifested in significant cost savings for the company. We no longer need expensive consultants to
retrospectively review our transactions. (A19)
The data revealed that these organizations have earmarked blockchain technologies as one
of their top strategic IT budget priorities for the next three to five years. Interestingly, all the
adopting large companies were using privately permissioned blockchain applications.
IMDS In contrast, all of the SME’s were using public permissioned blockchains. The interviewees
119,7 identified that the primary advantage of these private permissioned blockchains was tighter
control and security mechanisms. For instance, interviewees A19 and A15 indicated that
access privileges to modify or read the blockchain state of their applications is restricted to
only a few authorized users:
Our fully private blockchain technologies underpin various database and auditing functions within
1476 the company. The write permissions to these blockchain applications are centralized within the
organization. (A15)
There are public and private permission blockchains. We currently have blockchain applications to
optimise back and middle office systems that can be categorised as the latter. Only a hand full of
people have authorised access to these blockchains including the financial regulator, which serves
to augment our security, reporting and regulatory, and compliance profile. Unfortunately, there is a
myriad of issues (e.g. trust, cost, privacy) with public permission blockchains. Consequently, I do
not envisage that we as a company will be using these public instances anytime soon. (A19)
All of the three SMEs indicated that the primary advantage of the public permissioned
blockchains was the ease of accessibility, setup and access to information resources. In
terms of two of the SMEs, the interviewees indicated that they were not aware of private
permissioned blockchains and were eager to investigate further. The interviewee from the
other SME had investigated the possibility of private permissioned blockchain but ruled it
out as an option due to the complexity, cost and lack of business use cases for their industry.
For the non-adopting large organizations, issues relating to lack of top management
support, lack of business use cases, lack of government incentives and blockchain’s
association with cryptocurrencies/initial coin offerings (ICOs) were cited by the interviewees
as reasons for why their companies had not pursued blockchain technologies. These reasons
will be visited in subsequent sections. One interviewee mentioned their inability to develop
sustainable business models and one of the reasons for not adopting blockchain:
Unfortunately, management are unable to see the value that blockchain can bring to the company and
as a result we do not have plans to adopt any specific implementations soon. The core issue is that we
have been unable to simulate sustainable revenue and business models for potential blockchain
products and services. Yes, they can see the bottom line benefit of smart contracts but until there is real
money to be made, we are not going to dedicate much time to investigating it any further. (A10)
Of the nine SMEs, our study identified that three adopted blockchain, but the other six did
not. In comparison to the larger organizations who had adopted multiple instances of
blockchain and who were actively trialing other applications, these SMEs had adopted
single instances of blockchain applications. While the interviewees in these SMEs identified
various reasons for adopting blockchain, there was a consensus that their organizations had
primarily made the decision to adopt for the new innovative functionality that blockchain
technology could provide, as evidenced by the following:
For me blockchain is all about identity management and protection. Our blockchain product
enables citizens to transform their physical identities into virtual ones that are wrapped around
smart contracts. Blockchain will enable citizens to forge self-sovereign identities. (A9)
[…] we also created a secure cryptocurrency payment wallet, underpinned by blockchain
technology, to enable our customers to pay us in bitcoin. We are only one of a handful of companies
here in Ireland to do so. (A8)
Specific technological and business use case issues were outlined by interviewees from two
non-adopting SMEs:
We are looking at blockchain as a means of adding a further layer of security for protecting
transactional data within our SaaS CRM. However, the only use cases for businesses our size are
 only to be found in large corporations. Our clients are still trying to get their heads around cloud The influence
computing. Blockchain is adding to the confusion and therefore we are resistant to make the move of
now. (A13)
organizational
We have created a cloud-based data market platform where fish farmers can manage their farms factors
and share data. We are interested in underpinning this platform with blockchain technologies to
digitize farming stock such as oysters. The farmers will then have a transparent method of stock
management. However, we will not adopt blockchain until (i) specific technological issues
(e.g. scaling, tokenization, securitization) are resolved and (ii) smart contracts which are specific to
1477
the fishing industry have reached maturity. (A16)
Based on the findings presented here, we can conclude that organizational size is generally
positively related to blockchain adoption. Considering the number of blockchain instances
adopted and the prevalence of ongoing blockchain research and development activities, our
data suggest that large enterprises are more likely to adopt blockchain than SMEs.
4.3.4 Organizational readiness. For the purposes of this study, organizational readiness
with regards to adopting new IT innovations was examined in terms of three categories of
organizational resources which encompassed the availability of employees with the
requisite IT knowledge and skills; financial resources for adopting IT innovations (e.g. IT
budget) and infrastructure on which blockchain applications can be built. According to
existing research (Lacovou et al., 1995; Wang et al., 2010), the absence of one or more of these
resources is likely to constrain an organization’s ability to adopt an IT innovation.
Examining the results of our QCA in conjunction to Figure 1, it can be observed that all
organizations which adopted blockchain had sufficient organizational readiness.
In terms of financial and IT infrastructure resources, we noted that the availability and
functionality of cloud-based blockchain development platforms were pivotal in triggering
an organization’s decision to adopt blockchain, as confirmed by the following interviewees,
whose companies had adopted blockchain:
As with any software development project, the cost depends on the use case and the inherent
complexity. Furthermore, blockchain development and cloud computing go hand in hand. We
currently use the IBM Bluemix platform for developing blockchain applications. Our motivation
here is to keep our development cost base as low as possible. (A15)
The emergence of cloud-based products such as Ethereum and Microsoft’s Blockchain-as-a-service
(BaaS) has meant that organizations can use a rapid and fail-fast platform for developing, testing
and deploying blockchain applications on a free or pay-as-you-go basis. (A2)
We are currently using IBM’s cloud-based development platform Hyperledger to create my
blockchain enabled identity management application. These cloud-based blockchain development
platforms and tools have enabled SMEs such as mine to leverage blockchain in a cost-effective and
innovative manner. (A9)
One interviewee mentioned the presence of hidden costs embedded within one specific cloud-
based blockchain development platform that were beginning to frustrate his supervisors:
From a business and development point of view blockchain is relatively inexpensive. However,
there are hidden costs associated with blockchain. For instance, if you want to run an application
against the blockchain there are processing costs. On the Ethereum platform you must pay a ‘gas’
rate to verify every transaction. (A4)
From our analysis of the organizations that adopted, it emerged that the core competencies
required for blockchain are broader than the core technology and encompassed skill sets
that fall under the following categories: first, foundational technology (e.g. cryptography,
public key architecture); second, distributed ledger technology (e.g. mining, consensus
algorithms); third, forensics and law enforcement (e.g. money laundering, darknet); fourth,
IMDS markets, economics and finance (e.g. game theory, business modeling); fifth, industrial
119,7 design (e.g. supply chain, IoT) and, sixth, regulations and standards (e.g. smart contracts
and frameworks). A2 and A15 were examples of two organizations that were familiar with
the skills and knowledge required for blockchain:
We have engineers who are experts in creating secure (e.g. cryptography, encryption) distributed
network infrastructures. They are currently creating instances of private blockchains on a peer-to-
1478 peer network. Some organizations do not have this expertise and this type of development would
prove onerous particularly for tricky aspects of private blockchain such as automatic peer
discovery. (A2)
From our experience, the upskilling involved for an expert software or database engineer is quite
minimal. For instance, an individual with strong coding skills (e.g. C, Java, Python) and a good
understanding of distributed storage (e.g. NoSQL, RDBMS) would need to acquaint themselves
with the intricacies of smart contract and blockchain frameworks. Next, they would have to
familiarize themselves with ledger and decentralised technologies. (A15)
However, in contrast to the responses above, we also found that non-adopting SMEs were
struggling in terms of employee blockchain competencies:
From my experience of seeking external competencies, there is a dearth of people here in Ireland
who have an in-depth-knowledge and actual hands-on experience developing blockchain
applications. Additionally, not everything is applicable for the blockchain. Therefore, it doesn’t
make sense for us from a commercial value standpoint to explore it further. (A14)
Finally, we observed that the presence of organizational readiness for the lead provision
company in a supply chain does not always guarantee that other partners along their
supply chain will also have equivalent levels of organizational readiness to adopt
blockchain as evidenced by the following:
We have all the requisite resources in place to develop and host our own proprietary blockchain
services […] however we are struggling to get our SME supply chain partners to implement our
flagship blockchain authentication service which can result in performance management, supply
chain traceability, counterfeit, cyber and customer engagement benefits […] they are struggling to
see how blockchain sits with within their IT innovation strategies (e.g. we are not a virtual currency
business). (A3)
We have been developing blockchain applications for the past 5 years. While we can see the value
add, our business partners are struggling to also see it for their business. We are encountering a lot
of resistance because of various misinformation regarding blockchain. It could be a few years more
until we see widespread adoption. (A15)
Based on the arguments presented in this section, we conclude that the presence of sufficient
organizational readiness in terms of the availability of financial and employee resources
and access to IT infrastructure have a positive influence on a company’s decision to
adopt blockchain.
4.3.5 Adopting blockchain in a developed country. In the previous subsections, we presented
our findings in relation to our a priori variables. However, during our analysis we found
additional determinants of blockchain adoption that related to Ireland as a developed country.
There was consensus among the interviewees that Ireland as a country is quite proactive
in terms of the current initiatives that are taking place nationwide which are aimed at
enhancing blockchain awareness. Examples of such initiatives include regular blockchain
meetups, the establishment of metagroups such as the Blockchain Association of Ireland
and the IDA Blockchain Expert Group that is tasked with positioning Ireland as a leading
European center for blockchain development. Technology providers such as Hewlett
Packard, IBM, and Dell in conjunction with multinational professional service providers
such as Deloitte, EY and Accenture have established research and development hubs that The influence
focus primarily on blockchain research activities. Furthermore, these companies have of
developed specific internal blockchain applications, which they are currently piloting. organizational
However, we point out that most of the interviewees noted that the Irish government and the
central bank must do more to promote the adoption of blockchain technologies. Indeed, the factors
Irish government could make a global statement by adopting and rolling out blockchain
technologies via specific governmental e-Services: 1479
In similar fashion to how the UK’s cloud first policy was pivotal in accelerating the adoption of
cloud technologies in the private and public sectors, the Irish government could signal their intent
on a global scale by implementing a new form of public procurement infrastructure which is
underpinned by blockchain. (A3)
In order to put Ireland on the global blockchain map, we need a high profile, government backed
national use case which is underpinned by blockchain technology in a similar vein to Dubai,
Delaware and Hong Kong. For instance, the Irish government could roll out a universal national
digital identity scheme using blockchain. However, most governmental officials are not even aware
of blockchain and that is disappointing. (A19)
Six of the interviewees pointed out that the newly enacted GDPR triggered their
organizations to adopt or consider blockchain technology to ensure compliance with the
new data protection laws:
In the past six years, we have undergone a significant digital transformation to provisioning
cloud-based technologies. Security is of paramount importance for our customers. With the emergence
of the general data protection regulations we have prioritized the underpinning of all of our cloud-based
services with blockchain technology. (A6)
Despite the repeated lack of support by management for blockchain technologies […] I was at a
meeting recently where they indicated that they were now considering adopting blockchain
because of the strict sanctions in place for non-compliance with the GDPR. Identity and data
protection is becoming more and more of a concern. (A5)
Further issues, which emerged from our analysis, related to a low blockchain awareness
level. While most of our interviewees had heard of blockchain, as discussed earlier most
them had heard of blockchain technology but were unable to provide a correct description
and provide actual examples of real world blockchain applications. There was also a
consensus among the interviewees that there was a low awareness level with the other
companies they were dealing with on a day-to-day basis. One interviewee mentioned that he
had been several blockchain seminars that he felt were more about the promotion of
cryptocurrency products (e.g. ICOs) rather than educating blockchain laymen. Another
interviewee indicated that blockchain was being misleadingly pitched as the panacea to a
plethora of organizational issues at open seminars:
There is still a lot of hype around blockchain. Providers have underpinned their marketing
campaigns with the notion that blockchain is the remedy for all company problems. That is not the
case. For instance, from our experience, blockchain is not suitable for internal process improvement
such as improving asset utilization. (A6)
Most of our interviewees pointed out that the number of blockchain business use cases
across various industries has yet to reach full maturity:
Since 2013, the company have committed substantial financial investment. For instance, we have
installed backend filing systems which are hashing compliant so that they are interoperable with a
multitude of blockchain applications. However, this functionality has yet to be used to its fullest
potential. The business use cases demonstrating the value that can be derived for adopting
blockchain have not yet matured here in Ireland. (A8)
IMDS We like most companies are driven by the bottom line. The existing business cases have yet to
119,7 convince our senior executives that a move is justified. They believe there is a herd mentality
concerning blockchain. Until we as a company can see the benefits in terms of cost, faster
transaction times and improved security and most significantly see our competitors using it, we
will continue to observe blockchain with interest from a distance. (A11)
However, our study revealed that all of the large organizations who had adopted blockchain
1480 had developed private permissioned blockchains. All of the interviewees from these
organizations acknowledged that their blockchain deployments were exciting and
innovative developments for their companies however due to strict NDAs their
organizations were prevented from providing their business use case details to a wider
external audience. It would be interesting to investigate if this is the case in other developed
countries and if these NDAs play a significant contribution to blockchain awareness and
adoption of blockchain technologies.
Finally, 16 of our interviewees pointed out that blockchain is still largely associated with
terms such as “cryptocurrencies” or “virtual currencies” which give the technology a negative
connotation (e.g. ponzi schemes, unregulated ecosystem, illicit transactions and fraud):
Once senior managers could decouple the concept of blockchain from its Bitcoin alter-ego, they
were able to see the benefits that the company could derive in terms enhancing supply chain
operations while also providing the same safety, higher speeds and lower costs. (C7)
There is currently a guilty by association phenomenon currently occurring with blockchain and
cryptocurrencies. Yes, bitcoin would not have been possible without blockchain, however,
cryptocurrencies are just one blockchain use case example. Recent scandals with regards to initial
coin offerings, the unregulated nature of virtual currencies and a lack of custodial frameworks have
unfairly tainted people’s perceptions of blockchain. (A11)

5. Discussion
This section will first discuss the influence of organizational factors on the adoption of
blockchain by organizations based in Ireland and then outline the limitations of the study.

5.1 Patterns of blockchain adoption in Ireland

Our across-case analysis revealed three patterns pertaining to the adoption of blockchain in
Ireland: top management support positively influences blockchain adoption; large organizations
are more likely to adopt blockchain than SMEs; and organizational readiness is an enabler got
blockchain adoption. We now discuss each of these patterns and elucidate on the context in
which the results were identified, namely, Ireland as a developed country.
5.1.1 Pattern 1: top management support is an enabler for blockchain adoption. In terms
of top management support, we identified that sufficient top management support was
present for all companies who adopted blockchain. Thus, we conclude that key decision
makers play a significant role in deciding whether an organization adopts or does not adopt
blockchain. This finding has been consistently found in the literature that has demonstrated
that the presence of sufficient management support acts as a compelling enabler for the
adoption of IT innovations (Sabherwal et al., 2006; Bajaj, 2000; Kulkarni et al., 2017). Top
management support has also proposed as a significant factor for blockchain adoption
(Swan, 2015; Tapscott and Tapscott, 2016).
It was also interesting to note, first, the positive influence of top management’s
commitment toward technological innovativeness, second, the connection that emerged
concerning top management support and the levels of blockchain awareness and, third, the
ability for top management to adjust their behaviors throughout a blockchain adoption
process. In terms of this first finding, we identified that the level of research and
development intensity occurring within an organization had a significant impact on the The influence
adoption of blockchain. Top management’s commitment toward research and development of
initiatives has also been found to be a strong enabler for the adoption of IT innovations organizational
( Jarvenpaa and Ives, 1991; Rogers, 1995).
All of the interviewees from the adopting case organizations were directly involved and/
factors
or responsible for decisions about the IT budget allocations for research and development
initiatives. These interviewees indicated that prior to the adoption of blockchain, there 1481
organizations had devoted substantial resources and IT budget allocations toward
blockchain research and development initiatives. Additionally, in four non-adopting cases,
our interviewees who had oversight on IT budget allocations had worked in the
organization a period of more than ten years. Additionally, these interviewees had more
than 15 years’ experience in the IT industry. These interviewees indicated that while they
were aware of blockchain technologies, they were unlikely to use their current or future IT
budgets to pursue blockchain research and development initiatives. This finding suggests
that top managers whose perceptions of new IT innovations may be biased by substantial
same-company/industry experience (Daellenbach et al., 1999). This in contrast to previous
research which identified that “longer careers in a particular company or industry should
enhance a manager’s knowledge of the trends in the industry and make him/her more open
to research and development investments in innovation” (Hayes and Abernathy, 1980).
Concerning the second finding, as can be seen from Table IV and Figure 1, most of the
adopting organizational interviewees had medium to high levels of blockchain awareness
while most of the non-adopting organizations had no or low levels of blockchain awareness.
In terms of the third finding, our analysis identified that top management supportive actions
were not static during the adoption process. Complex blockchain adoptions, especially
related to large-scale projects, are rarely predictable. Issues relating to training, resources,
trading partners and legislation frequently emerge. For instance, in four of the non-adopting
cases, we saw that decision makers were influenced by uncertainty with regards to
blockchain legislation. The interviewees reported that because existing regulatory
frameworks underpin traditional centralized and trusted models of security processing,
migrating to a new decentralized model would takes years of cross-jurisdictional
cooperation. However, top management in adopting organizations, most notably large
enterprises, had gained insights from regulators and were assured that substantial change
in the regulatory position on blockchain and distributed ledger technologies is unlikely any
time soon. This confirmation enabled top management to positively adjust their behavior
toward blockchain adoption. These findings are in line with existing research which
suggests that management existing knowledge base, attitude about a specific IT and
management’s ability to adjust their behaviors positively influences the adoption of an IT
innovation (Rogers, 1995; Thong and Yap, 1995; Dong et al., 2009).
5.1.2 Pattern 2: large companies are more likely to adopt blockchain than SMEs.
Concerning organizational size, we found that 5 out of the 11 large organizations and 3 out
of the 9 SMEs adopted blockchain. This finding supports previous innovation adoption
research that has shown how organizational size positively affects an organizations
willingness to adopt an IT innovation (Pothukuchi et al., 2002; Lee and Xia, 2006). The
interviewees expressed two significant reasons for adopting or rejecting blockchain which
can be sourced to the size of their organization. First, for the organizations that adopted
blockchain, their decisions were motivated by the need to reduce complexity and lower
supply chain investment costs associated with transactions (e.g. transfer of tangible and
intangible value) while also reaping the benefits of enhanced security, efficiency and
transparency which is associated with blockchain transactions. The large costs and time-
consuming nature associated with traditional supply chain transactions, which in some
IMDS cases involved thousands of actors, were some of the main considerations in their decisions
119,7 to make blockchain a strategic imperative. Conversely, the complexity of a large-scale
digital transformation that would be encompassed in migrating from traditional centralized
systems to decentralized ones also led other large organizations to reject blockchain. For the
SMEs that adopted blockchain, they were similarly motivated by the enhanced speed,
security, transparency and cost effectiveness afforded by blockchain transactions. These
1482 findings in relation to blockchain adoption for the purposes of supply chain efficiency have
also been reported in previous research (e.g. Crosby et al., 2016; Lansiti and Lakhani, 2017).
On the other hand, for the SMEs who did not adopt blockchain, decisions were motivated by
the fact they were operating in small business networks (e.g. supply chains) which would
not necessitate the use of blockchain. Furthermore, the interviewees from SMEs in larger
business networks acknowledged that unless the leading actor in the business network
decides to implement blockchain technologies, they will be reluctant to do so because of the
levels of complexity involved.
Second, most interviewees expressed the significance of blockchain business use cases.
According to Fichman (2004, p. 315), “organizations that are larger, more diverse, have
greater technical expertise, possess supportive senior management, operate in more
competitive contexts, and perceive the innovation as more beneficial and compatible,
are more likely to adopt a larger number of innovations, to adopt them earlier, and to
implement them more thoroughly.” The fact that large organizations possess more
resources in terms of finances, infrastructure and skilled employees implies that these
companies can engage in increased experimentation with the IT innovation. Ultimately,
larger organizations can create their own business cases more readily. For instance, all the
large companies who adopted blockchain were operationalizing “complex” private
permissioned instances of the technology (e.g. private blockchains). These interviewees
also indicated that given the inherent negatives of traditional blockchain transactions (e.g.
scaling of speed and costs) that their organizations were researching blockchain
transactions that can be conducted “off-chain.” It was also interesting to note that for both
SMEs and large organizations that adopted Blockchain, decisions were motivated by the
ability to experiment with blockchain technologies, prior to adoption, on “cost effective”
pay-as-you go cloud development platforms. This motivation may be reinforced by the
characteristics of Ireland as a developed country. As highlighted earlier, the Bloomberg
innovation index for 2018 ranked Ireland in 13th place. The index scores countries using
seven criteria, including research and development spending, concentration of high-tech
public companies and patent activity ( Jamrisko and Liu, 2018). Additionally, motivated by
the proliferation of business use cases, increased levels of trust and the need to lower the
investment costs in IT infrastructure cloud computing technologies are now being used on
a mainstream basis in Ireland from storage and IT development perspectives (Carcary
et al., 2014). Conversely, the perceived levels of complexity and the lack of business use
cases led to both large and SMEs rejecting blockchain. This finding with regards to
complexity serving as an adoption barrier confirms the arguments of existing research
(e.g. Morabito, 2017; Lindman et al., 2017). Based on our findings, we conclude that large
organizations are more inclined to adopt blockchain than SMEs.
5.1.3 Pattern 3: organizational readiness for IT innovation increases the likelihood of
blockchain adoption. In total, 8 of the 20 case study organizations demonstrated sufficient
organizational readiness for IT innovation in terms of possessing satisfactory levels of
financial resources, blockchain competent employees and IT infrastructure. These eight
cases adopted blockchain technologies. This finding is line with the existing literature that
argues that enterprises with sufficient organizational readiness are more likely to be
adopters of blockchain (Swan, 2015; Tapscott and Tapscott, 2016).
 With regards to blockchain expertise an IDA report (IDA, 2017) suggests that Ireland’s The influence
IT workforce, by in large, possess core foundational software development skills which can of
be leveraged by organizations to capitalize on the emergence of blockchain innovations. organizational
Based on our analysis, it emerged that the core competencies required for blockchain are
broader than the core technology. The nuanced and ambiguous nature of blockchain skills factors
and competencies was cited as a major barrier to blockchain adoption by both SME and
large non-adopting interviewees. Thus, it would seem logical that to bridge the current 1483
blockchain skills gap (e.g. the skills that exist and the skills that are being produced),
professional certifications and university add-on courses/modules are needed which are
specific to blockchain technologies.
In the case of another IT innovation, namely cloud computing, organizational adoption
rates began to dramatically increase following the emergence of use cases by IT service
providers and government agencies who demonstrated how cloud technologies could be
used effectively (Morgan and Conboy, 2013; Clohessy et al., 2017). Blockchain organizational
adoption rates seem to be following a similar trajectory with use case examples beginning
to slowly emerge. We envisage that, given the increasing significance of secure IT
(e.g. enterprise cloud, virtual currencies, cashless payments), elements of blockchain
technologies will underpin the majority of SME and large IT services in the future not only
here in Ireland but at a global level. Moving forward, there needs to be a national awareness
about what blockchain is and is not. This can be achieved through collaborative
engagement and dialogue with both SME, large and public sector organizations. This will
build a knowledge base that companies and the public can access.
It should be noted from Figure 1 that there are overlaps between large organizations
and organizational size and SMEs and top management support. Concerning the former,
in accordance with the existing literature (Mehrtens et al., 2001) we treated these variables
separately. In terms of the latter, this overlap is expected (Ramdani and Kawalek, 2007)
because our interviewees were key decision makers (e.g. company owner, CEO, IT
manager) and their support for the adoption of a new IT innovation is vital (Premkumar
and Roberts 1999).
Finally, our data analysis suggests that an organization’s ability to experiment with
blockchain technologies “on the cloud” prior to adoption positively affects their adoption
decision. It was interesting to hear our blockchain adopting SME interviewees reflect on
how the nature of these cloud development platforms (e.g. pay per use model, sandbox
environments, low cost up-front investments, open source) played a pivotal role in their
decision to adopt the IT innovation. This finding is in line with extant research that argues
that an enterprises’ ability to experiment with a new IT innovation increases their likelihood
of adoption (Ramdani and Kawalek, 2007). Collectively, these findings can have practical
implications for IT vendors in terms of developing marketing strategies that would target
potential adopters.

5.2 Limitations
It is worth highlighting some limitations of our study and areas that may represent fruitful
direction for additional research. First, our study focused on three specific organizational
factors that influence a company’s decision to adopt blockchain. As noted in Section 2.2, this
narrowing of the scope in relation to these factors was intentional as they are the most
commonly used organizational factors in IT innovation adoption studies. We acknowledge
that this narrowing of scope means that we did not explore other organizational factors that
were identified in our literature review. Furthermore, as highlighted by Table III, we also
identified environmental and technological factors that also merit further investigation.
We envisage that future research that adopts a broader scope might result in a more
IMDS comprehensive analysis of blockchain adoption in Ireland. Second, we identified that top
119,7 management support is crucial to adopting blockchain. It would be interesting to delve
deeper into how these managers make decisions concerning IT innovations (governance
structures, personal characteristics, etc.). Finally, our study was based on 20 Irish cases,
divided across eight industry sectors. Although we used prescribed research protocols to
ensure reliability and validity, the findings should be interpreted cautiously. For instance,
1484 certain sectors were not included in our study (e.g. government, health) which may have
influenced our findings. Additionally, while Ireland is categorized as a developed country,
certain aspects ( Jamrisko and Liu, 2018) mean that direct comparisons with other developed
countries cannot be made. While our qualitative approach resulted in interesting findings on
the blockchain adoption process, we do encourage future research to explore an increased
number of organizations using quantitative-based research approaches to further
investigate the influence of organizational factors.

6. Conclusion and implications

This study investigated how organizational factors influenced blockchain adoption in
organizations based in a developed country. Specifically, we explored blockchain awareness
in companies based in Ireland and investigated how the following several organizational
factors influenced the blockchain adoption process: organizational size, organizational
readiness and top management support.

6.1 Scientific contributions

This study makes several important scientific contributions. First, our study provides a
cumulative overview (Table VIII) of specific organizational considerations that provide
motivations as to why both large and SME companies based in a technology developed
country such as Ireland chose to adopt blockchain or not adopt blockchain. We also have
identified several blockchain adoption patterns (Section 5). We postulate that these
considerations and patterns may explain the low blockchain adoption rates in Ireland.
Future blockchain adoption research could explore how our findings in relation to these
considerations and patterns are comparable with other technology developed countries with
similar features to Ireland (e.g. Israel, Denmark, Finland and so on).
Further, there was consensus among the respondents, once we had provided them with
our standardized encapsulation of what blockchain was, that blockchain could provide a
fundamental underpinning pillar to the digitized Irish economy that has seen a major
increase in the production and consumption of data. However, the pace of the technology
was identified as a concern with respondents who are struggling keeping up with the
demands of the technology.
Second, the findings revealed that top management support and organizational
readiness are significant enablers of blockchain adoption. These findings are in line with the
existing literature that argue that these factors have positive influences on IT innovation
adoption (Rogers, 1995; Lacovou et al., 1995; Weiner, 2009). Previous IT adoption literature
(Zhu et al., 2006; Oliveira and Martins, 2010; Yuen et al., 2010) assume that organizations
operating in technologically developed countries exhibit a largely coordinated approach
with regards to new IT innovations. Contrary to this assumption, our study demonstrated
that top management support is reinforced by a cultural factor where new IT innovations
are often categorized as a high priority at the chief-suite level but not at board level. This is
compounded by the fact that the adoption of new IT innovations within the organizations is
often uncoordinated (e.g. bottom-up, top-down approaches). Future research could
investigate in more detail how the relationship between top management support and
new IT innovations is influenced by Irish culture, or in a broader context, the characteristics
of developed countries.
 Finally, previous IT innovation adoption literature suggests that organizations size has a The influence
positive influence on a company’s IT innovation adoption process (Damanpour, 1992; of
Thong, 1994). This study demonstrates that large organizations are more likely to not only organizational
adopt blockchain but are also more likely to conduct increased levels of blockchain research
and development activities. We believe that this finding is directly related to the complexity factors
of the IT innovation that makes it highly compatible for companies with a high resource
capacity. We envisage that as blockchain continues to evolve as a commercial entity and the 1485
availability of cloud-based blockchain development tools increase, the IT innovation will
become more amenable to organizations with fewer resources.

6.2 Practical contributions

Our study also has significance from a practical perspective. First, our findings concerning the
low level of blockchain awareness and the lack of information pertaining to viable business
use cases indicate that the Irish government could play a more significant role in promoting
the benefits of blockchain technologies. Further, our findings could also encourage IT
providers to formulate enhanced strategies aimed at disseminating information pertaining to
blockchain technologies. Second, the positive influence of top management support and
organizational readiness, particularly about core competencies, on blockchain adoption
suggests that equipping managers with the requisite knowledge and skills will be crucial in
adopting these IT innovations. In order to expedite this process our findings suggest that
there is a need for Irish Universities and/or other Irish course providers to formulate new
blockchain-based curriculum encompassing the core competencies delineated in Section 4.3.4.
Finally, it was encouraging to note that the several SMEs who adopted blockchain used cloud-
based blockchain platforms and tools to overcome the constraints of their initial low levels of
organizational readiness. Therefore, we encourage SMEs with low levels of organizational
readiness to explore cloud-based blockchain development platforms and tools when
contemplating whether to adopt blockchain.

References
Anderson, N., Potočnik, K. and Zhou, J. (2014), “Innovation and creativity in organizations: a state-of-
the-science review, prospective commentary, and guiding framework”, Journal of Management,
Vol. 40 No. 5, pp. 1297-1333.
Armstrong, C.P. and Sambamurthy, V. (1999), “Information technology assimilation in firms: the
influence of senior leadership and IT infrastructures”, Information Systems Research, Vol. 10
No. 4, pp. 304-327.
Bajaj, A. (2000), “A study of senior information systems managers’ decision models in adopting
new computing architectures”, Journal of the Association for Information Systems, Vol. 1 No. 1,
pp. 1-58.
Barrett, M., Davidson, E., Prabhu, J. and Vargo, S.L. (2015), “Service innovation in the digital age: key
contributions and future directions”, MIS Quarterly, Vol. 39 No. 1, pp. 135-154.
Bear, K., Drury, N., Korsten, P., Pureswaran, V., Wallis, J. and Wagle, L. (2016), “Blockchain rewires
financial markets: trailblazers take the lead”, IBM Institute for Business Value, New York,
available at: https://public.dhe.ibm.com/common/ssi/ecm/gb/en/gbp03469usen/GBP03469USEN.
PDF (accessed February 2, 2017).
Beck, R. (2018), “Beyond bitcoin: the rise of Blockchain world”, Computer, Vol. 51 No. 2, pp. 54-58.
Beck, R., Avital, M., Rossi, M. and Thatcher, J.B. (2017), “Blockchain technology in business and information
systems research”, Business & Information Systems Engineering, Vol. 59 No. 6, pp. 381-384.
Beck, R., Becker, C., Lindman, J. and Rossi, M. (2017), “Opportunities and risks of Blockchain
technologies (Dagstuhl Seminar 17132)”, cited in Dagstuhl Reports Vol. 7, No. 3, Schloss
Dagstuhl-Leibniz-Zentrum fuer Informatik.
IMDS Benbasat, I., Goldstein, D.K. and Mead, M. (1987), “The case research strategy in studies of information
119,7 systems”, MIS Quarterly, pp. 369-386.
Carcary, M., Doherty, E. and Conway, G. (2014), “The adoption of cloud computing by Irish SMEs–an
exploratory study”, Electronic Journal Information Systems Evaluation, Vol. 17 No. 1,
pp. 1-14.
Chen, G., Xu, B., Lu, M. and Chen, N.S. (2018), “Exploring blockchain technology and its potential
1486 applications for education”, Smart Learning Environments, Vol. 5 No. 1, pp. 1-10.
Christensen, C.M., Raynor, M.E. and McDonald, R. (2015), “What is disruptive innovation”, Harvard
Business Review, Vol. 93 No. 12, pp. 44-53.
Clohessy, T., Acton, T. and Morgan, L. (2017), “The impact of cloud-based digital transformation on IT
service providers: evidence from focus groups”, International Journal of Cloud Applications and
Computing (IJCAC), Vol. 7 No. 4, pp. 1-25.
Clohessy, T., Acton, T., Godfrey, R. and Houston, M. (2018), “The adoption of Blockchain in Ireland:
examining the influence of organisational factors”, A National University of Ireland Galway and
Blockchain Association of Ireland Industry Report, May 10.
Crosby, M., Pattanayak, P., Verma, S. and Kalyanaraman, V. (2016), “Blockchain technology: beyond
bitcoin”, Applied Innovation, Vol. 2 No. 2, pp. 6-10.
Cuomo, J., Ramamurthy, S. and Wallis, J. (2016), “Fast forward: rethinking enterprises, ecosystems and
economies with blockchains”, IBM Institute for Business Value, New York, available at: ibm.biz/
blockchainstudy (accessed February 8, 2017).
Daellenbach, U.S., McCarthy, A.M. and Schoenecker, T.S. (1999), “Commitment to innovation: the
impact of top management team characteristics”, R&D Management, Vol. 29 No. 3, pp. 199-208.
Damanpour, F. (1991), “Organizational innovation: a meta-analysis of effects of determinants and
moderators”, Academy of Management Journal, Vol. 34 No. 3, pp. 555-590.
Damanpour, F. (1992), “Organizational size and innovation”, Organization Studies, Vol. 13 No. 3,
pp. 375-402.
Dong, L., Neufeld, D. and Higgins, C. (2009), “Top management support of enterprise systems
implementations”, Journal of Information Technology, Vol. 24 No. 1, pp. 55-80.
Eisenhardt, K.M. (1989), “Building theories from case study research”, Academy of Management
Review, Vol. 14 No. 4, pp. 532-550.
Espinel, V. (2015), “Deep shift: technology tipping points and societal impact”, World Economic
Forum, Geneva, available at: www3.weforum.org/docs/WEF_GAC15_Technological_
Tipping_Points_report_2015.pdf (accessed March 6, 2017).
Fichman, R.G. (2004), “Going beyond the dominant paradigm for information technology innovation
research: emerging concepts and methods”, Journal of the Association for Information Systems,
Vol. 5 No. 8, pp. 314-355.
Fitzpatrick, R., Murphy, J. and O’Neil, D. (2017), “Redrawing the lines: FinTech’s growing influence on
financial services”, PWC, London, available at: www.pwc.com/jg/en/publications/pwc-global-
fintech-report-17.3.17-final.pdf (accessed April 21, 2018).
Folkinshteyn, D. and Lennon, M. (2016), “Braving bitcoin: a technology acceptance model (TAM)
analysis”, Journal of Information Technology Case and Application Research, Vol. 18 No. 4,
pp. 220-249.
Gangwar, H., Date, H. and Ramaswamy, R. (2015), “Understanding determinants of cloud computing
adoption using an integrated TAM-TOE model”, Journal of Enterprise Information
Management, Vol. 28 No. 1, pp. 107-130.
Grandon, E.E. and Pearson, J.M. (2004), “Electronic commerce adoption: an empirical study of small
and medium US businesses”, Information & Management, Vol. 42 No. 1, pp. 197-216.
Greckhamer, T., Misangyi, V.F., Elms, H. and Lacey, R. (2008), “Using qualitative comparative analysis
in strategic management research: an examination of combinations of industry, corporate, and
business-unit effects”, Organizational Research Methods, Vol. 11 No. 4, pp. 695-726.
Guo, Y. and Liang, C. (2016), “Blockchain application and outlook in the banking industry”, Financial The influence
Innovation, Vol. 2 No. 1, pp. 1-12. of
Hayes, R.H. and Abernathy, W.J. (1980), “Managing our way to economic decline”, Harvard Business organizational
Review, Vol. 58 No. 4, pp. 67-77.
factors
Heyvaert, M., Maes, B. and Onghena, P. (2013), “Mixed methods research synthesis: definition,
framework, and potential”, Qual. Quant., Vol. 47 No. 2, pp. 659-676.
IDA (2017), “Ireland the European location for FinTech and Blockchain”, available at: www.idaireland. 1487
com/newsroom/publications/irelandforfintech (accessed March 8, 2018).
Ito, J., Narula, N. and Ali, R. (2017), “The Blockchain will do to the financial system what the internet
did to media”, Harvard Business Review, March 9, available at: https://hbr.org/2017/03/the-
blockchain-will-do-to-banks-and-law-firms-what-the-internet-did-to-media
Jamrisko, M. and Liu, W. (2018), “The US drops out of the top 10 in innovation ranking”, available at:
www.bloomberg.com/news/articles/2018-01-22/south-korea-tops-global-innovation-ranking-
again-as-u-s-falls (accessed March 8, 2018).
Jang, H. (2010), “An empirical study on the factors influencing RFID adoption and implementation”,
International Journal Management Review, Vol. 5 No. 2, pp. 55-73.
Jarvenpaa, S.L. and Ives, B. (1990), “Information technology and corporate strategy: a view from the
top”, Information Systems Research, Vol. 1 No. 4, pp. 351-376.
Jarvenpaa, S.L. and Ives, B. (1991), “Executive involvement and participation in the management of
information technology”, MIS Quarterly, Vol. 15, pp. 204-224.
Joo, Y.B. and Kim, Y.G. (2004), “Determinants of corporate adoption of e-marketplace: an innovation
theory perspective”, Journal of Purchasing and Supply Management, Vol. 10 No. 2, pp. 89-101.
Kaganer, E., Pawlowski, S.D. and Wiley-Patton, S. (2010), “Building legitimacy for IT innovations: the
case of computerized physician order entry systems”, Journal of the Association for Information
Systems, Vol. 11 No. 1, pp. 1.
Karahanna, E., Straub, D.W. and Chervany, N.L. (1999), “Information technology adoption across time:
a cross-sectional comparison of pre-adoption and post-adoption beliefs”, MIS Quarterly, Vol. 23
No. 2, pp. 183-213.
Kiayias, A. and Panagiotakos, G. (2016), “On trees, chains and fast transactions in the Blockchain”,
IACR Cryptology ePrint Archive No. 545, Edinburgh.
Kimberly, J.R. and Evanisko, M.J. (1981), “Organizational innovation: the influence of individual,
organizational, and contextual factors on hospital adoption of technological and administrative
innovations”, Academy of Management Journal, Vol. 24 No. 4, pp. 689-713.
Kitchenham, B. and Brereton, P. (2013), “A systematic review of systematic review process research in
software engineering”, Information and Software Technology, Vol. 55 No. 12, pp. 2049-2075.
Kokina, J., Mancha, R. and Pachamanova, D. (2017), “Blockchain: emergent industry adoption and
implications for accounting”, Journal of Emerging Technologies in Accounting, Vol. 14 No. 2,
pp. 91-100.
Kosba, A., Miller, A., Shi, E., Wen, Z. and Papamanthou, C. (2016), “Hawk: the blockchain model of
cryptography and privacy-preserving smart contracts”, Security and Privacy (SP), 2016 IEEE
Symposium on, pp. 839-858.
Kulkarni, U.R., Robles-Flores, J.A. and Popovič, A. (2017), “Business intelligence capability: the effect of
top management and the mediating roles of user participation and analytical decision-making
orientation”, Journal of the Association for Information Systems, Vol. 18 No. 7, pp. 516-541.
Kushnir, D. and Sandén, B.A. (2011), “Multi-level energy analysis of emerging technologies: a case
study in new materials for lithium ion batteries”, Journal of Cleaner Production, Vol. 19 No. 13,
pp. 1405-1416.
Lacovou, C.L., Benbasat, I. and Dexter, A.S. (1995), “Electronic data interchange and small
organizations: Adoption and impact of technology”, MIS Quarterly, Vol. 19 No. 4, pp. 465-485.
IMDS Lansiti, M. and Lakhani, K.R. (2017), “The truth about blockchain”, Harvard Business Review, Vol. 95
119,7 No. 1, pp. 118-127.
Law, C.C. and Ngai, E.W. (2007), “ERP systems adoption: an exploratory study of the organizational
factors and impacts of ERP success”, Information & Management, Vol. 44 No. 4,
pp. 418-432.
Lee, G. and Xia, W. (2006), “Organizational size and IT innovation adoption: a meta-analysis”,
1488 Information & Management, Vol. 43 No. 8, pp. 975-985.
Li, Z., Wang, W.M., Liu, G., Liu, L., He, J. and Huang, G.Q. (2018), “Toward open manufacturing: a cross-
enterprises knowledge and services exchange framework based on blockchain and edge
computing”, Industrial Management & Data Systems, Vol. 118 No. 1, pp. 303-320.
Lindman, J., Tuunainen, V.K. and Rossi, M. (2017), “Opportunities and risks of Blockchain
Technologies–a research agenda”, Proceedings of the 50th Hawaii International Conference on
System Sciences, HI.
Lippert, S.K. and Govindarajulu, C. (2006), “Technological, organizational, and environmental
antecedents to web services adoption”, Communications of the IIMA, Vol. 6 No. 1, pp. 147-160.
McKinsey (2017), “Blockchain technology in the insurance industry”, available at: www.treasury.gov/
initiatives/fio/Documents/McKinsey_FACI_Blockchain_in_Insurance.pdf (accessed March 8, 2018).
Mehrtens, J., Cragg, P.B. and Mills, A.M. (2001), “A model of Internet adoption by SMEs”, Information &
Management, Vol. 39 No. 3, pp. 165-176.
Mehta, D. and Striapunia, K. (2017), “FinTech 2017, market size, business models, blockchain and
company profiles”, Statista Report, Hamburg, available at: www.statista.com/study/45600/
statista-report-fintech/ (accessed July 9, 2018).
Mendling, J., Weber, I., van der Aalst, W., Brocke, J.V., Cabanillas, C., Daniel, F. and Gal, A. (2017),
“Blockchains for business process management-challenges and opportunities”, ACM
Transactions on Management Information Systems (TMIS), Vol. 9 No. 1, p. 4.
Miles, M.B., Huberman, A.M. and Saldana, J. (2013), Qualitative Data Analysis, Sage, Arizona.
Moeller, T. and Schwerin, S. (2017), “Privacy-friendly blockchain technologies: an overview of
Blockchain technology and the ‘GDPR’ ”, available at goo.gl/zz4Ci2 (accessed February 28, 2018).
Molla, A. and Licker, P.S. (2005), “Perceived e-readiness factors in e-commerce adoption: an empirical
investigation in a developing country”, International Journal of Electronic Commerce, Vol. 10
No. 1, pp. 83-110.
Morabito, V. (Ed.) (2017), “Blockchain value system”, Business Innovation through Blockchain, Springer,
Cham, pp. 21-39.
Moran, O. (2018), “Bitcoin could crash and burn – but blockchain jobs are here to stay”, available at:
www.techcentral.ie/bitcoin-crash-burn-blockchain-jobs-stay/ (accessed February 28, 2018).
Morgan, L. and Conboy, K. (2013), “Key factors impacting cloud computing adoption”, Computer,
Vol. 46 No. 10, pp. 97-99.
Myers, M.D. (2013), Qualitative Research in Business and Management, Sage, London.
Nakamoto, S. (2008), “Bitcoin: a peer-to-peer electronic cash system”, available at: www.Bitcoin.org/
bitcoin.pdf (accessed February 28, 2018).
O’Dair, M., Beaven, Z., Neilson, D., Osborne, R. and Pacifico, P. (2016), “Music on the blockchain”,
Blockchain for Creative Industries Research Cluster Report, Middlesex University, London,
pp. 1-29.
Oliveira, T. and Martins, M.F. (2010), “Understanding e-business adoption across industries
in European countries”, Industrial Management & Data Systems, Vol. 110 No. 9, pp. 1337-1354.
Ordanini, A., Parasuraman, A. and Rubera, G. (2014), “When the recipe is more important than the
ingredients: a qualitative comparative analysis (QCA) of service innovation configurations”,
Journal of Service Research, Vol. 17 No. 2, pp. 134-149.
Pare, G. (2004), “Investigating information systems with positivist case research”, Communication The influence
Association Information Systems, Vol. 13 No. 1, pp. 233-264. of
Pilkington, M. (2016), “Blockchain technology: principles and applications”, in Olleros, F.X. and Zhegu., M. organizational
(Eds), Research Handbook on Digital Transformations, Edward Elgar, Northampton, pp. 225-254.
factors
Pisa, M. and Juden, M. (2017), “Blockchain and economic development”, Centre for Global Development,
London, available at: www.cgdev.org/sites/default/files/blockchain-and-economic-development-
hype-vs-reality_0.pdf (accessed April 3, 2018). 1489
Plewa, C., Troshani, I., Francis, A. and Rampersad, G. (2012), “Technology adoption and performance
impact in innovation domains”, Industrial Management & Data Systems, Vol. 112 No. 5,
pp. 748-765.
Pothukuchi, V., Damanpour, F., Choi, J., Chen, C.C. and Park, S.H. (2002), “National and organizational
culture differences and international joint venture performance”, Journal of International
Business Studies, Vol. 33 No. 2, pp. 243-265.
Premkumar, G. and Roberts, M. (1999), “Adoption of new information technologies in rural small
businesses”, Omega, Vol. 27 No. 4, pp. 467-484.
Puschmann, T. and Alt, R. (2016), “Sharing economy”, Business & Information Systems Engineering,
Vol. 58 No. 1, pp. 93-99.
Ramdani, B. and Kawalek, P. (2007), “SME adoption of enterprise systems in the Northwest of
England”, IFIP International Working Conference on Organizational Dynamics of Technology-
Based Innovation. Springer, Boston, MA, June, pp. 409-429.
Rihoux, B. and Ragin, C. (2009), Configurational Comparative Methods: Qualitative Comparative
Analysis (QCA) and Related Techniques, Sage Publications, Thousand Oaks, CA.
Ritchie, J., Spencer, L. and O’Connor, W. (2003), “Carrying out qualitative analysis”, in Ritchie, J. and
Lewis, J. (Eds), Qualitative Research Practice: A Guide for Social Science Students And
Researchers, Sage, London, pp. 219-262.
Rogers, E. (1995), Diffusion of Innovations, 4th ed., Free Press, New York, NY.
Sabherwal, R., Jeyaraj, A. and Chowa, C. (2006), “Information system success: individual and
organizational determinants”, Management Science, Vol. 52 No. 12, pp. 1849-1864.
Schwandt, T.A., Lincoln, Y.S. and Guba, E.G. (2007), “Judging interpretations: but is it rigorous?
Trustworthiness and authenticity in naturalistic evaluation”, New Directions for Evaluation, Vol.
2007 No. 114, pp. 11-25.
Seebacher, S. and Schüritz, R. (2017), “Blockchain technology as an enabler of service systems: a
structured literature review”, International Conference on Exploring Services Science, Springer,
Cham and Chicago, IL, pp. 12-23.
Shrier, D., Sharma, D. and Pentland, A. (2016), Blockchain & Financial Services: The Fifth Horizon of
Networked Innovation, MIT Science Connection, Cambridge, MA, pp. 1-10.
Strauss, A. and Corbin, J. (1998), Basics of Qualitative Research: Procedures and Techniques for
Developing Grounded Theory, Sage, Thousand Oaks, CA.
Swan, M. (2015), Blockchain: Blueprint for A New Economy, “O’Reilly Media”, CA.
Swanson, E.B. (2004), “How is an it innovation assimilated”, IFIP Working Conference on IT Innovation
for Adaptability and Competitiveness. Springer, Boston, MA, pp. 267-287.
Tapscott, D. and Tapscott, A. (2016), Blockchain Revolution, Portfolio/Penquin, New York, NY.
Thong, J.Y. and Yap, C.S. (1995), “CEO characteristics, organizational characteristics and information
technology adoption in small businesses”, Omega, Vol. 23 No. 4, pp. 429-442.
Tornatzky, L.G. and Fleischer, M. (1990), The Processes of Technological Innovation, Lexington Books,
Lexington, MA.
Treat, D., McGraw, L., Helbing, C. and Brodersen, C. (2017), “Blockchain technology: preparing for
change”, available at: www.accenture.com/t20160608T052656Z__w__/us-en/_acnmedia/PDF-5/
Accenture-2016-Top-10-Challenges-04-Blockchain-Technology.pdf (accessed March 2, 2018).
IMDS Treiblmaier, H. (2019), “Toward more rigorous blockchain research: recommendations for writing
119,7 blockchain case studies”, Front. Blockchain, Vol. 2 No. 3, doi: 10.3389/fbloc.2019.00003.
Van de Weerd, I., Mangula, I.S. and Brinkkemper, S. (2016), “Adoption of software as a service in
Indonesia: examining the influence of organizational factors”, Information & Management,
Vol. 53 No. 7, pp. 915-928.
Venkatesh, V. and Davis, F.D. (2000), “A theoretical extension of the technology acceptance model: four
1490 longitudinal field studies”, Management Science, Vol. 46 No. 2, pp. 186-204.
Vom Brocke, J., Simons, A., Riemer, K., Niehaves, B., Plattfaut, R. and Cleven, A. (2015), “Standing on
the shoulders of giants: challenges and recommendations of literature search in information
systems research”, CAIS, Vol. 37, No. 1, pp. 205-224.
Wang, H., Chen, K. and Xu, D. (2016), “A maturity model for blockchain adoption”, Financial
Innovation, Vol. 2 No. 1, pp. 1-5.
Wang, P. (2009), “Popular concepts beyond organizations: exploring new dimensions of information
technology innovations”, Journal of the Association for Information Systems, Vol. 10 No. 1,
pp. 1-30.
Wang, Y.M., Wang, Y.S. and Yang, Y.F. (2010), “Understanding the determinants of RFID adoption in
the manufacturing industry”, Technological Forecasting and Social Change, Vol. 77 No. 5,
pp. 803-815.
Webster, J. and Watson, T. (2002), “Analyzing the past to prepare for the future: writing a literature
review”, MIS Quarterly, Vol. 26 No. 2, pp. xiii-xxiii.
Weiner, B.J. (2009), “A theory of organizational readiness for change”, Implementation Science, Vol. 4
No. 1, pp. 1-9.
Woodside, J.M., Augustine, F.K. Jr and Giberson, W. (2017), “Blockchain Technology adoption status
and strategies”, Journal of International Technology and Information Management, Vol. 26 No. 2,
pp. 65-93.
Yang, Z., Sun, J., Zhang, Y. and Wang, Y. (2015), “Understanding SaaS adoption from the perspective of
organizational users: a tripod readiness model”, Computers in Human Behavior, Vol. 45,
pp. 254-264.
Yin, R.K. (2014), Case Study Research: Design and Methods, 5th ed., Sage Publications, Thousand Oaks, CA.
Yu, D. and Hang, C.C. (2010), “A reflective review of disruptive innovation theory”, International
Journal Of Management Reviews, Vol. 12 No. 4, pp. 435-452.
Yuen, Y.Y., Yeow, P.H., Lim, N. and Saylani, N. (2010), “Internet banking adoption: comparing
developed and developing countries”, Journal of Computer Information Systems, Vol. 51 No. 1,
pp. 52-61.
Zamani, E.D. and Giaglis, G.M. (2018), “With a little help from the miners: distributed ledger technology
and market disintermediation”, Industrial Management & Data Systems, Vol. 118 No. 3,
pp. 637-652.
Zhu, K., Kraemer, K.L. and Xu, S. (2006), “The process of innovation assimilation by firms in different
countries: a technology diffusion perspective on e-business”, Management Science, Vol. 52
No. 10, pp. 1557-1576.
Zyskind, G. and Nathan, O. (2015), “Decentralizing privacy: using blockchain to protect personal data”,
Security and Privacy Workshops (SPW), IEEE, pp. 80-184.

Further reading
Ramdani, B., Kawalek, P. and Lorenzo, O. (2009), “Predicting SMEs’ adoption of enterprise systems”,
Journal of Enterprise Information Management, Vol. 22 Nos 1/2, pp. 10-24.
Swanson, E.B. (1994), “Information systems innovation among organizations”, Management Science,
Vol. 40 No. 9, pp. 1069-1092.
Treiblmaier, H. (2018), “Case studies for blockchain use cases: a bottom-up research agenda”, The influence
Communication of the Association for Information Systems (CAIS), Vol. 34 No. 3, of
pp. 555-577.
Walsh, C., O’Reilly, P., Gleasure, R., Feller, J., Li, S. and Cristoforo, J. (2016), “New kid on the block: a organizational
strategic archetypes approach to understanding the Blockchain”, International Conference on factors
Information Systems, Dublin, December.

1491
Corresponding author
Trevor Clohessy can be contacted at: trevor.clohessy@gmit.ie

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

IMDS
119,7
1492
Received 30 August 2018
Revised 10 January 2019
Accepted 4 April 2019
Industrial Management & Data
Systems
Vol. 119 No. 7, 2019
pp. 1492-1514
© Emerald Publishing Limited
0263-5577
DOI 10.1108/IMDS-08-2018-0370
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm
Examining the role of narratives
in civic crowdfunding: linguistic
style and message substance
Chang Heon Lee
Department of Business Administration, College of Business and Economics,
United Arab Emirates University, Al Ain, United Arab Emirates
Yiyang Bian
Department of Information Systems, College of Business,
City University of Hong Kong, Hong Kong, and
Rajaa Karaouzene and Nasreen Suleiman
College of Business and Economics,
United Arab Emirates University, Al Ain, United Arab Emirates
Abstract
Purpose – The purpose of this paper is to explore how linguistic style and message substance influence
persuasion in civic crowdfunding marketplaces in which written narrative pitch become a vital
communication to attract private contributions to public goods and services. Drawing on the elaboration
likelihood model (ELM), the authors operationalize the linguistic style of the narrative pitch as language
power and message substance as issue-relevant argument quality. In this paper, the authors examine how
characteristics of both style and message are related to the outcome of civic crowdfunded projects.
Design/methodology/approach – The data on civic crowdfunding projects were retrieved from Spacehive,
the platform that dedicated mainly to civic projects ranging from community programs, social-oriented
enterprises, to infrastructure or facility development. Each of the narrative samples is analyzed using a
computerized text analysis package called the Linguistic Inquiry and Word Count to extract the features of
the linguistic style and message substance in the narratives. The logistic regression models are estimated to
assess the impact of both linguistic style and message substance on crowdfunding decisions.
Findings – The results show that funding outcomes can be improved with psychological language
dimensions (i.e. positive affective and perceptual language). However, extensive use of social language does
not help project creators to increase their chance of funding performance; but instead, such language
reduces the likelihood of project success. Additionally, message substance or issue-relevant information such
as money and risk language influences funding outcome.
Originality/value – Very few empirical studies investigated the differential effects of language style and
message substance on funding performance of crowdfunding campaigns. The authors draw upon the dual
process of persuasion as a theoretical base to identify a comprehensive set of linguistic style and message
substance and to examine the role of such features in an emerging civic crowdfunding market. This study
advances the application of the dual process in ELM by identifying and examining distinct persuasive cues
originating from linguistics styles and message contents.
Keywords Persuasion, Elaboration likelihood model, Linguistic style, LIWC, Civic crowdfunding
Paper type Research paper
1. Introduction
Civic crowdfunding is becoming an increasingly important source of funds for public sector
projects, such as community educational events, green space improvements and even
construction of small airports. In particular, social entrepreneurs find the civic crowdfunding to
be a vital method of funding for their social enterprises since crowdfunding, defined as an open
call through the internet for the provision of financial resources, enabled them to garner funds.
This project is supported by UAE University Research Grant Code G00002617 (Funding No. 31B088),
Shenzhen Special Fund for Strategic Emerging Industries Development (Grant No. JCYJ20170818100156260),
and GRF Grant-City University of Hong Kong (Grant Code: 11508517 and 11507717).
 Given that civic crowdfunding takes place via open online settings, narratives play an Role of
elevated role in the process of communication between project creators and potential backers narratives in
because the backers are often left to make judgments or form assessments solely relying on the civic
narrative pitches (Anglin et al., 2018). The entrepreneurship literature shows that how project
creators deliver the message matters since the narratives assist in acquiring necessary resources crowdfunding
through conveying value and setting expected outcomes (Chen et al., 2009; Martens et al., 2007).
As a result, we have witnessed a more direct interest in the role of language in entrepreneurship 1493
and crowdfunding literature (Larrimore et al., 2011; Parhankangas and Renko, 2017).
A emerging literature has explored the relationship between language and campaign
outcomes, particularly in the context of peer-to-peer micro-lending (Allison et al., 2013;
Herzenstein et al., 2011; Majumdar and Bose, 2018) and reward-based crowdfunding (Anglin
et al., 2018; Parhankangas and Renko, 2017), where financial benefits and delivery of tangible
products are specified as forms of rewards or incentives. In reward-based crowdfunding,
entrepreneurial pitches often center on the development of a new product. Research
examining language features of such pitches has predominately focused on the objective
evidence of product quality (Parhankangas and Renko, 2017) and economic viability (Allison
et al., 2017). Similarly, in examining the effect of the language on micro-lending, research has
mainly focused on evidence-based elements of the pitches (Allison et al., 2013) since central to
crowdfunding microloan solicitation is the entrepreneurial narratives describing borrower’s
personal details and loan characteristics. Prior research has examined linguistic style, defined
as the use of style words, rather the message substance of microloan pitches and found
evidence that linguistic style enables lenders to form opinions about prospective borrowers
and thus affect their decisions (Herzenstein et al., 2011; Larrimore et al., 2011). Recent studies
have also demonstrated that positive psychological capital rhetoric (Anglin et al., 2018) and
negative emotions (Majumdar and Bose, 2018) affect the probability of meeting funding goals.
These studies have advanced our understanding of the role played by entrepreneurial
narratives in both reward-based crowdfunding and micro-lending.
However, extant research is dominated by focusing one aspect of language, more
specifically either on linguistic style or content cues that are presented in narrative
pitches. To date, scholars have examined these two factors in isolation. Moreover, prior
research explored limited aspects of language; there are still rich features (psychologically
derived linguistic style such as perceptual, social and cognitive languages) have not been
concerned. Nevertheless, these psychological linguistic styles have shown importance on
persuasion in other contexts, such as advertising (Gibbons et al., 1991), and online product
review (Ludwig et al., 2013; Peng et al., 2004). Given that researchers have investigated
various factors with different focuses, there is a need for a synthesis of divergent
perspectives and the development of a theoretical model that systematically examines
both linguistic style and message substance. Yet relatively few research works have been
dedicated to the synthesis of existing knowledge; the question of how message substance
(what one communicates) and linguistic style (how one communicates) affect the
persuasive process has not been fully answered.
We conduct this study in the context of the civic crowdfunding marketplace where
rhetorical or linguistic positioning is critical to securing resources from diverse
stakeholders. Civic crowdfunding campaigns present a unique context for investigating
narratives pitches for two reasons. First, the civic projects would be expected to produce
shared goods or services beneficial to the public, implying that they typically address rather
diverse stakeholders or crowds, which ask for multivocal narrative pitches. The
presentation of information is especially important to appeal to a variety of potential
project backers, including citizens, local community groups, city councils and local or
federal governments since civic crowdfunding projects involve a mechanism of private
provision of public goods. Second, creating and communicating both economic and social
IMDS value to society is a vital element of narrative pitches in civic crowdfunding. Given that
119,7 these projects have multiple pursuits, project creators must find a way of delivering or
communicating their goals, execution plans and values in a persuasive manner. These
creators can choose between highlighting the substance of the message and linguistically
emphasizing the presentation style or make a balanced approach.
To identify a richer set of both linguistic style and content features and to examine the
1494 role of such features in persuasion in civic crowdfunding, we draw upon elaboration likelihood
model (ELM) of persuasion (Petty and Cacioppo, 1986) as a theoretical base. As a dual
process model, the ELM has been employed in a wide range of domains, including product
adverting (Petty et al., 1983), healthcare (Angst and Agarwal, 2009), information technology
(Bhattacherjee and Sanford, 2006) and electronic commerce (Tam and Ho, 2005). It provides a
comprehensive framework about the degree and type of information processes individuals
employ in considering a persuasive message. The ELM asserts that attitude can be formed as
a result of either high or low degree of thinking or thoughtful information processing, and the
type of thinking operate along a continuum ranging from low to high elaboration. This
continuum is anchored by two distinct routes, namely central and peripheral processing, to
persuasion. The central route is characterized by extensive and thoughtful cognitive
processing of the message, whereas the peripheral route is characterized by minimal or
low-effort information processing in which individuals use shortcuts or simple heuristics to
form a judgment or evaluation. Additional, prior studies employing the ELM framework have
considered message framing and argument strength, and linguistic style. For instance, Allison
et al. (2013) find that strength and quality of arguments that are presented in either
loss-framed or gain-framed message influence consumers’ attitude toward the brand. The
ELM framework is particularly relevant to civic crowdfunding settings because certain
features of linguistic styles and issue-relevant contents could potentially influence potential
backers’ central and peripheral processing.
In this study, we endeavor to examine the effects of linguistic styles and content cues on
the success of civic crowdfunding projects. We primarily interested in addressing the
following research questions:
RQ1. Whether does the linguistic style and message substance in a civic crowdfunding
narrative increase the likelihood of funding success?
RQ2. How do the linguistic style cues that are derived from psychological, social and
cognitive dimensions affect backers’ funding decision behaviors?
Given that, we employ the ELM framework to identify and examine how language
characteristics are related to funding outcomes of civic crowdfunding projects.
We generalize these characteristics into linguistic style and message substance in terms
of these two key persuasion routes (central and peripheral cues) of ELM. Linguistic style
and message substance are two key factors that will affect the persuasion process in many
domains such as law, marketing and politics (Sparks and Areni, 2008). Specifically,
linguistic style and message substance can be seen as language power and argument
quality in most cases which could be explained by ELM. Thus, this study considers the dual
roles of narratives in civic crowdfunded projects. We examine how a psychological or affect
dimension of language style (i.e. positive affective, social and perceptual) influence
impression formation or persuasion process via a peripheral route. Further, we investigate
how the use of message substance or content cues related to issues such as money, risk and
achievement affect persuasion via a central route.
This study extends the prior efforts that examine the factors of crowdfunding campaigns
in several ways. First, there has been a little research reported on the effectiveness of
text-based product pitches in civic crowdfunding. Civic crowdfunding mainly deals with
public campaigns range from a community program to bridge construction. Narratives for
public goods largely differ from ones from dominant reward-based crowdfunding and Role of
micro-lending. We fill this gap by focusing on the effectiveness of such narratives presented narratives in
in civic crowdfunding marketplaces. Second, our findings contribute to, and extend, the civic
emerging stream of research on online entrepreneurial financing in general and civic
crowdfunding campaign. Most of the prior literature focuses on entrepreneur characteristics crowdfunding
and relatively tangible product-related factors (Agrawal et al., 2015; Mollick, 2014) that
improve funding outcomes. In this study, we have sought to examine both language style 1495
and message substance and explore how two language features that have been studied in
isolation are related to funding outcome in the context of civic crowdfunding in which
individual investors are making private contributions to public goods or services. Different
from earlier studies, we investigated a comprehensive set of both language style and
message content and explore how the two language features are related to the funding
outcome. Third, we advance the application of the dual process in ELM by identifying and
examining distinct persuasive cues originating from linguistics styles and contents. We
argue that a psychological dimension of languages, such as affective positive, cognitive, and
perceptual language, influence backers’ funding decision. Specifically, affect-related
components in the narratives are effective to evoke a peripheral route to persuasion. Fourth,
this study extends the ELM of persuasion to the civic crowdfunding context. The ELM has
been validated in many decision behavior contexts, yet the issue of how the various
psychological dimension of language influence a peripheral route to persuasion remains
unexplored. Through the validation of our work, we enhance the theoretical understanding
of how language style and message substance influence funders’ investment decisions in the
civic crowdfunding marketplace.

2. Related literature
2.1 Civic crowdfunding
Crowdfunding, as a new model of finance, refers to the practice of funding a project or
venture by raising capital through requesting contributions from a large number of people
via the internet (Mollick, 2014). The phenomenon of collective finance is not new by itself as
projects or ideas were funded in the past by open calls. However, the internet has boosted up
both the scope and the potentials of the phenomenon (Agrawal et al., 2015; Ryu, 2018). Web
2.0 technology play an essential role in advancing crowdfunding by providing the platforms
that facilitate the interaction and communication between project initiators and the crowd to
achieve the desired outcome. The novelty and convenience of the online crowdfunding have
led to the dramatic increase in the number of crowdfunding platforms that respond to the
need of different markets. In a recent industry report published by Massolution (2015) the
number of crowdfunding platforms reached 1,200 platform covering different markets.
Initially, crowdfunding platforms first emerged to collect donations or fund for small
creative projects without monetary rewards (Kang et al., 2016). This application was further
expanded to facilitate loans between private persons in which capital-giving agents receive
interests for borrowing money in a model known as peer-to-peer micro-lending (Herzenstein
et al., 2011; Majumdar and Bose, 2018). Alongside with progressively increasing projects
and investments, private-to-business loans, and equity-based crowdfunding started arising
as a new form of crowdfunding platforms (Ahlers et al., 2015).
Civic crowdfunding is identified as “sub-type of crowdfunding through which citizens, in
collaboration with government, fund projects providing a community service” (Stiver et al.,
2015). Civic crowdfunding refers to the direct financing of civic projects. It is known to have
the potential to revolutionize community participation (Davies, 2015; Stiver et al., 2015). The
main potential of civic crowdfunding is that private–public collaborations or partnerships
may increase the interactivity between individual citizens and municipalities (Lee et al.,
2016). Further, a distinct characteristic of civic crowdfunding is that civic projects would be
IMDS expected to produce shared goods or services beneficial to the public (Hassna et al., 2018).
119,7 Recent years have witnessed that an increased number of civic crowdfunding platforms has
been developed to allow ordinary citizens to direct their money to a local public project.
Civic crowdfunded projects range from community facility development to transportation
improvement, to community educational events and environmental enhancement programs.
In particular, social entrepreneurs find the civic crowdfunding to be a vital method of
1496 funding for their social enterprise since online open platforms enabled them to garner funds.
Despite civic crowdfunding’s growing popularity as a vehicle for financing public
projects and social enterprises, a significant number of civic crowdfunding campaigns have
reportedly failed to meet their stated funding targets. To understand factors affecting the
success of civic crowdfunding campaigns, research has focused on diverse factors that
influence individuals’ funding behaviors, such as altruism and warm glow motivation
(Davies, 2015), source credibility (Hassna et al., 2018) and project legitimacy (Calic and
Mosakowski, 2016).

2.2 Role of narratives in the context of crowdfunding

A small body of literature has examined the content characteristics of reward-based
crowdfunding (see Anglin et al., 2018; Frydrych et al., 2016; Gorbati and Nelson, L. 2015;
Parhankangas and Renko, 2017). Anglin et al. (2018) examined the content of campaign
pitches from Kickstarter and their results show that positive narratives and passion have a
positive effect on crowdfunding performance. Parhankangas and Renko (2017) focus on the
linguistic style of project pitches, more specifically, concrete language, precise language and
interactive style in both social and commercial categories. Their findings indicate that
concrete language and interactive style have a positive effect on social crowdfunding projects,
whereas none of the linguistic style variables predicted the success of commercial
crowdfunding campaigns. The four linguistic styles – concrete, precise, interactive and low
psychological distance language – have a greater impact on the success of social campaigns
than on commercial campaigns. Gorbatai and Nelson (2015) examined four linguistic
characteristics, namely, vivid, inclusive, positive and business-related language in technology
and small business categories and their results show that inclusive and positive emotional
language has a positive effect on the crowdfunding performance, while business-related
language has an adverse effect.
Furthermore, emerging literature has explored the relationship between message substance
and funding outcomes, in the context of peer-to-peer micro-lending (Allison et al., 2013;
Herzenstein et al., 2011; Majumdar and Bose, 2018) where financial benefits are specified as a
form of incentives. A study by Herzenstein et al. (2011) shows that identity contents are effective
to attract funding in micro-lending and their findings, in particular, demonstrate trustworthy
identities in the narrative have a positive effect on loan funding. Allison et al. (2015) find that the
presence of both intrinsic and extrinsic cues embedded within micro-lending entrepreneurial
narratives affect funding outcomes. A recent study shows that the female reference and
authenticity in the narrative pitch positively affect the trustworthiness of the narration, leading
to increases in odds of getting a successful donation (Majumdar and Bose, 2018). These results
demonstrate that the role of narratives is critical as project creators seek to persuade potential
funders. Prior literature dealing with individuals purchase and consumption tends to focus on
informational factors, namely, the notion of information asymmetry; investors desire to reduce
risks or uncertainty associated with quality of product from reward-based crowdfunding and
with personal loans from micro-lending.

3. Theoretical background and hypothesis development

ELM is a general dual process theory of attitude change. The theory provides a general
framework for understanding the fundamental processes underlying stimuli processing
and, consequently, the effectiveness of persuasive communications (Petty and Cacioppo, Role of
1986). The model suggests two distinct routes for stimuli processing in the process of narratives in
attitude formation: a central route and a peripheral route. ELM holds that there are civic
numerous specific processes of attitude formation on the “elaboration continuum” ranging
from low to high. When the processes are at the low end of the elaboration continuum, crowdfunding
persuasion follows the peripheral route, while when the processes are at the high end of the
elaboration continuum; persuasion follows the central route (Kruglanski et al., 2012). 1497
Under the central route, persuasion and attitude formation result from the careful
consideration of the real merits of the information presented in the message about the
argument (Cacioppo and Petty, 1984). On the other hand, under the peripheral route,
persuasion and attitude formation result from the person’s association with negative or
positive cues or from making an inference about the argument.
Extant literature on language and persuasion has demonstrated that when applying the
ELM to products, central cues constitute issue-relevant information that includes specific
attributes or verifiable evidence directly relating to products or product quality, while
peripheral cues typically consist of affective components in the message such as message
tone and perceived status of source delivering the message (Petty and Cacioppo, 1986).
Evidence-based arguments that are typically presented in the objective language with
specific issues such as financial benefits and risks are related to argument quality, leading
to evoke a central route to persuasion. On the contrary, the affect-based element of the
messages such as tone or mood influence message comprehensions, and source credibility,
leading to a peripheral route to persuasion. Cues used under the peripheral route are not
related to the logical quality or merit of the stimuli and involve factors like credibility,
legitimacy, popularity or attractiveness of the message sender. The likelihood of elaboration
is determined by the person’s ability and motivation to comprehend logically and evaluate
the argument from one side, and on the quality of the information presenting the argument
from another side (Petty et al., 1983).
In civic crowdfunding, creators publicize product description to demonstrate the ideas of
their proposed projects and to persuade potential funders to provide the required fund.
From the ELM perspective, following the central route requires backers to think critically
using the information disclosed in the project description and analyze the relative merits of
the project before making their funding decisions. Different studies suggested that potential
backers receive very limited information which is self-reported by project creators
(Mollick, 2014). The limitation of available information makes the high elaboration process
more difficult and less effective when making funding decision (Allison et al., 2015).
Evaluating the viability of civic projects under the conditions of limited information can be
full of uncertainty and risk. The other option for the backers, under these conditions, is the
peripheral low elaboration route which involves less cognitive effort but more association
with cues and signals when making their inference. When following this route, the backers
collect, and process different cues related to the project legitimacy and credibility.
We argue that the linguistic dimensions associated with psychological dynamics are
created to tap affective, sensory and social process, leading to the peripheral route to
persuasion. Individuals’ assessment can be shaped by a less-effortful heuristic process
through which they rely upon peripheral cues (Cacioppo and Petty, 1984; Petty et al., 1983).
In the civic crowdfunding context, in addition to looking to verifiable evidence of quality
regarding the projects prospective backers can search for other elements in the linguistic
style that create the message setting, rather than in the message itself (Allison et al., 2013,
2017). Such linguistic style cues are relatively subjective, affect-based elements or the
underlying tone (Ludwig et al., 2013; Schwarz et al., 1991).
Although project description that psychological process (e.g. words depicting affective,
cognitive, social and perceptual words) are important for both framing the significance of
IMDS message settings or strengthening the effect of the affect-related peripheral route, it is also
119,7 important to consider the approach through which project creators are trying to persuade
backers of thoughtful considerations of issue-relevant arguments. In general, issues-relevant
information refers to key content directly relating to the merits or risks of the focal topic, such
as objective evidence of project quality. According to the dual process of ELM, the central
route focuses on the issue-relevant message cues to persuade (Petty and Cacioppo, 1986). Petty
1498 and Cacioppo defined the persuasive impact of a message in relation to the argument quality
and its ability to elaborate thoughts in favor of that argument’s position. Message recipients’
involvement in examining and making sense of a message argument has long been specified
as a vital factor in determining the persuasive effect of a message (Park et al., 2007). That is,
issue-relevant elaborations activate increased message involvement, leading to facilitating
information processing in civic crowdfunding platforms. This enhanced processing is more
likely lead to effective or enhanced persuasion when a message presents quality arguments.

3.1 Peripheral route: psychological language style

Positive affective language. Linguistic cues have long been served as a common medium for
persuasion (Hosman, 2002). The literature on language and persuasion has demonstrated
that how language is framed can exert influence on persuasion and judgment (Miller et al.,
2007). In particular, affective language cues drive individuals’ decision-making process
(Gibbons et al., 1991; Petty et al., 1993). Although people are likely to use both positive and
negative language tones when sharing their opinions or experiences with others, these
expressions serve different functions. The positive (negative) affective language would lead
to more positive (negative) evaluations toward firms or products (Tetlock et al., 2008), which
then influence persuasive effects. For example, the use of positive affective language elicits
positive feedback from others and facilitates positive assessments. On the other hand, the
use of negative tone induces pessimism. This phenomenon suggests that affective language
as a cue to online readers may play an elevated role in online settings.
Positive language influences the positive perceptions of a firm or product. Prior studies
on entrepreneurship and marketing literature also have shown that linguistic style
cultivates impressions concerning the value of a firm (Davis et al., 2006; Henry, 2008) or a
product (Ludwig et al., 2013). For example, the affective tone in earnings press releases is
predictive of increasing security prices, suggesting that optimistic language usage in
earning announcements affect investors’ valuation of the firm. Ludwig et al. (2013) find that
positive affective content in online product review has a notable influence on product sales
in electronic retail marketplaces. Further, recent research examining online written loan
requests in peer-to-peer lending also find that the positive tone of loan description improves
funding success (Allison et al., 2013).
In civic crowdfunding marketplaces, a narrative of the project pitch is the primary means
by which creators communicate their ideas and expected outcomes to prospective backers.
The positive affective language in the narrative may act as a positive signal through the
portrayal of optimism and confidence, leading to more positive judgment by those backers.
More recently, it also has been suggested that affective components in the message
influence persuasion (Allison et al., 2017; Chen et al., 2009). For example, Allison et al. (2017)
find that the presence of the positive tone in the form of portraying the venture as a personal
dream positively affect funding performance. This body of work suggests that positive
narrative tone may selectively facilitate retrieval of positive impression from the project
pitch and inhibit retrieval of negative tones, leading more favorably evaluations toward a
proposed civic project. We, therefore, propose the following hypothesis:
H1. The positive affective language positively affects funding success on civic crowdfunding
platforms.
Perceptual language. In addition of using the positive affective language in an attempt to Role of
appeal to potential backers emotionally, project creators choose words carefully to pitch the narratives in
campaign by explaining what the project is about, why it is essential for the public or
community and what the expected outcomes are. Extant literature has recognized that
civic
linguistic cues play an elevated role in improving message comprehension or recall, leading crowdfunding
to effective persuasion process (Fussell and Krauss, 1989; Pennebaker and King, 1999).
Prior studies show that language variation closely related to message comprehension is the 1499
use of perceptual language (Federmeier, 2007; Pennebaker and King, 1999). The perceptual
words that activate the sensory systems may lead to tap depth of information process
quickly ( Jordan and Thomas, 2002). For instance, Federmeier (2007) finds that perceptual
information becomes more critical for judgment. The findings imply that the inclusion of
perpetual language, including words related to hearing, seeing and other sensory words
may facilitate message comprehension and then improve effective persuasion. Further
perceptual words are associated with a vivid and detailed representation of objects
(Weiskopf, 2010), which allow faster information processing through sensory systems. For
instance, a recent study by Kim et al. (2011) finds that perceptual previous studies on
language and persuasion posit that perceptual language facilitates the ease of processing
and make the message seems more appealing and open (DeYoung et al., 2005). Words that
can be processed quickly and fluently is usually viewed as more accurate and persuasive
(Alter and Oppenheimer, 2009). That is, language processing easiness or fluency positively
affects judgments or evaluations. Further, openness is most associated with a higher
prevalence of perceptual processes, including words related to hearing and seeing.
Openness has previously been linked to higher levels of creative functioning and esthetic
sensitivity (DeYoung et al., 2005). The current results suggest that this esthetic sensitivity
also translates into the writing of highly open individuals, with a greater emphasis on
perceptual processes.
Previous studies on language and persuasion posit that perceptual language facilitates
the ease of processing and make the message seems more appealing and open (DeYoung
et al., 2005). Words that can be processed quickly and fluently is viewed as more accurate
and persuasive (Alter and Oppenheimer, 2009). That is, language processing easiness or
fluency positively affects judgments or evaluations. Further, openness is most associated
with a higher prevalence of perceptual processes, including words related to hearing
and seeing. Openness has previously been linked to higher levels of creative functioning and
esthetic sensitivity (DeYoung et al., 2005). The current results suggest that this esthetic
sensitivity also translates into the writing of highly open individuals, with a greater
emphasis on perceptual processes.
In civic crowdfunding platforms, the use of perceptual language in the project pitch
may influence or convince potential backers to contribute to the campaign. Backers
in crowdfunding frequently do not have detailed domain knowledge of proposed ideas and
execution plan (Kim and Viswanathan, 2018). Accordingly, those prospective patrons are
often left to comprehend the details by relying on language and content cues of the project
narrative. When narratives are written for readers to comprehend the message easily
or perhaps for a wider audience, the general public, such narrative pitch affect message
comprehension, which may reduce backers’ perception of ambiguity or uncertainty.
Consequently, the use of perceptual language may be effective in communicating
core values and reinforcing the importance of public benefits in the mind of the
audience, leading to better funding performance. This line of reason leads to the
following hypothesis:
H2. The use of perceptual language positively affects funding success on civic crowdfunding
platforms.
IMDS Social language. A great deal of research has highlighted that language is primarily a mean
119,7 of communicating among members of society (Miller, 1951). Word usage is generally
considered an essential marker of cognitive and social processes, and the way people
express themselves can be more persuasive than the message substance (Pennebaker and
King, 1999). Word provides information about the social process such as group identity and
social norm (Miller, 2000). Prior studies show that social language − indicating the use of
1500 social concerns (e.g. poverty or employment), as well as social intercourse-related (e.g. we,
they, neighbor, or children) − may promote better group performance. For instance,
pronouns reveal how individuals refer to others when interacting and increased use of the
first-person plural is found to be related to group cohesion and group performance
(Tausczik and Pennebaker, 2010).
Furthermore, recent psychological research suggests that people are more likely to
support other members within a community if their request is presented in a way that
evokes social concerns (Davis, 2018). The linguistic style that focuses on social processes
may help builds social empathy or social rapport with the target audience. Examples of
social language that build social rapport and draw attention to social concerns exhibit
high levels of social empathy and avoid psychological distancing from the target audience
(Toma and D’Angelo, 2015). Social empathy refers to the generation of internally
emotional responses to social concerns or events by identifying the feeling of another
person by taking that person’s viewpoint (Eisenberg et al., 1998). For instance, prior
studies suggest that greater use of social words is likely to exert individuals’ social
empathy. Additionally, sharing personal experiences with greater use of the first-person
plural is one of the effective ways of avoiding psychological distancing from the target
audience. For instance, previous studies find that first-person plural is related to higher
relationship quality while the use of the second person is negatively related to relationship
quality (Simmons et al., 2008). The entrepreneurship literature also suggests that use of
social language may function as a peripheral cue that could affect individuals’ assessment
and judgments since those social languages in the narratives becomes a vital source of
background information, possibly revealing their social position, social concerns and
social status (Tan et al., 2007).
In civic crowdfunding platforms, individual citizens and municipalities are particularly
interested in what good their money does since the civic projects would be expected to
produce shared goods or services beneficial to the public (Hassna et al., 2018; Lee et al., 2016).
There are citizens who are willing to contribute their money in order to improve their
neighborhood or support for social enterprise. As a result, they are likely to be influenced by
the social linguistic style that highlights an instrumental role in improving a community
and public infrastructure. For instance, recent research has provided evidence that
microenterprise that emphasizes social value in their linguistic narratives assemble financial
resources quickly (Moss et al., 2018). The findings imply that an increasing emphasis on the
social language can improve clarity about the motives of the civic projects, leading to attract
more investment.
Furthermore, civic project creators can appeal to prospective funders by utilizing the
peripheral cues that consist of social language since social or community identification
improve the effectiveness of narratives (Martens et al., 2007). Previous research has shown
that social language emphasizes community identity and group differentiation and an
increase in the importance of a social norm can increase persuasiveness. When potential
backers feel close to project creators, support or cooperation is more likely to occur, and even
the unconscious social identity can increase social empathy. Social bonds, which form the
basis of the relationship, are easily formed. For example, individuals may feel close to others
based on the variable dimensions of social similarities, such as society or social status.
Taken together, we expect that, when the use of social language is high, the project is more Role of
likely to obtain outside resources, and further achieve better performance: narratives in
H3. The use of social language positively affects funding success on civic crowdfunding civic
platforms. crowdfunding

3.2 Central cues: issue-relevant content 1501

Quantitative information. Objective language is the language that is not influenced by
emotions, opinions or personal feelings, and uses factual information that is quantifiable
and measurable. Objective language can be compromised of quantitative information,
such as financial and monetary data, which proves credibility, reliability and accuracy
and increases the backers’ confidence of the campaign to carry out its goals as promised
(Larrimore et al., 2011). Research in marketing and adverting has viewed as objective
evidence of product quality as issue-relevant information and demonstrated that quality
of the arguments presented in the ad influences consumers’ product attitudes (Petty et al.,
1983). For instance, a recent study shows that message recipients’ product evaluations are
derived in part from perceived benefits from money-related message. The findings imply
that the inclusion of objective language, including quantitative information, may facilitate
message comprehension and then improve effective persuasion. Extant literature on
entrepreneurship and crowdfunding has also demonstrated that financial details in
narratives influence investors’ decision-making process (Allison et al., 2017). For instance,
Allison et al. (2017) find that in micro-lending settings, investors be persuaded by a
centrally processed argument which includes additional financial details in the loan
request description.
In civic crowdfunding platforms, many campaigns actively seek contributions not only
from individual citizens but also from established organizations of different kinds, including
businesses, local governments and non-governmental entities. Those resource providers are
likely to be attracted to projects emphasizing economic value as prior literature shows
financial viability to be essential for affecting funding decisions (Moss et al., 2018). In order to
realize economic value, civic project creators send quality signals to various stakeholders, by
disclosing factual data or quantifiable commitments on plans. Despite social orientation in
civic crowdfunding contexts, diverse audiences analyze the economic performance in making
funding decisions (Stiver et al., 2015). Civic crowdfunding campaigns that attempt to
communicate information accurately by revealing the details of their idea without interjecting
their personal opinions demonstrate expertise, preparedness and competence, which help
ease the uncertainty that backers may face about investing. Moreover, if the creators fail to
disclose quantitative data for the financial benefit of public goods or services, backers hardly
assess the projects concerning economic value and eventually hesitant to make an investment
decision. That is, if the creators fail to conform to economic expectations, prospective
funders are less likely to view the proposed projects as economically viable ones, resulting
in less funding. Therefore, we posit that the use of quantitative language shows preparedness
and increases the chances of fundraising success. For these reasons, we propose the
following hypothesis:
H4. The use of quantitative language positively affects funding success on civic
crowdfunding platforms.
Risk-related content. Disclosure of risk-related information is of increasing concern in a wide
range of practices in the industry (Renn and Levine, 1991). Notably, in a business setting,
the increasing complexity of business strategies makes it difficult for investors to appreciate
the financial information so to effectively fulfill information demands of such investors
IMDS firms often attempt to disclose risk-related information (Abraham and Cox, 2007; Linsley
119,7 and Shrives, 2006). Prior studies have demonstrated that increased risk disclosure could
impact the perceived level of company risk (Fahnestock, 2011). Although disclosure of risk
information would be useful for shareholders to evaluate companies’ activity, increase
risk disclosure influence the perceived level of firm risks (Linsley and Shrives, 2006).
This phenomenon implies that a higher degree of use of risk language in communication
1502 styles may lead message recipients or prospective investors to perceive a higher risk than is
actually present.
The entrepreneurship literature highlights risk and reward as two essential components of
an investment proposal. These define how much risk an entrepreneur anticipates and how much
profit is expected (MacMillan et al., 1985). For instance, Allison et al. (2015) find that in micro-
lending higher degrees of risk-taking language are related to an increase in the fundraising time
needed to fund a microloan. The finding implied that a micro-lending presentation framed in a
way that focuses on the risk are more likely to be less appealing to investors and hence
decreases the attractiveness of a microloan to investors. Additionally, the earlier study in
psycholinguistics suggests that extensive use of risk language raise cautionary reactions and
doubt of deception, leading to influence credibility (Covello and Sandman, 2001; Renn and
Levine, 1991). A linguistic style measuring risk cues is based on whether a communicator uses
various terms to highlight risk factors. For instance, risk-related word, including fail, caution,
doubtful, penalty, liability, uncertain or error, are used to describe a product.
In civic crowdfunding platforms, we expect that using a project pitch that includes more
risk language may influence potential backers by informing risky elements of the proposed
project. Such disclosure of risk factors activates the central route, examining risk factors in a
more objective manner. When forming assessments on projects, backers are less likely to
fund a campaign that uses more risk words.
Furthermore, using risk-related language provides signals about the contrast between
their campaign and alternatives, and can also justify why their ideas are necessary or not
compared with alternative solutions. This information is essential for backers looking to
determine the credibility of the campaign and in deciding whether to support it. Risk-related
language helps to emphasize the differences to backers more easily and conveys the
potential benefits in a way that assists in the evaluation of whether their investment will
produce the benefits claimed. On the contrary, it generates signals that can communicate
negative features about the idea. Hence, we argue more use of risk-related language
adversely affect the credibility and decrease the chances of fundraising success. For these
reasons, we propose the following hypothesis:
H5. The use of risk language negatively affects funding success on civic crowdfunding
platforms.
Achievement content. The achievement language refers to the overarching dimension that
captures the language refers to achievement striving, words indicative of moving toward a
successful outcome – including goal, effort, success (Tausczik and Pennebaker, 2010). In
civic crowdfunding, achievement language in the narratives could improve backers’
knowledge of project confidential information and expertise of initiator in funding projects.
The issue-related achievement facts reveal the confidential information of “success” of
related projects. As a result, backers are able to evaluate the perceived costs of investment,
which makes their perception of funding decision more reliable. The achievement language
with confidential information sharing also provides a signal of “good faith” to the potential
backers. In this way, using achievement language could build positive an initiator-backer
relationship by providing tangible facts that backers can perceive the motives of the civic
funding projects.
 On the other hand, achievement language expresses the domain expertise of Role of
crowdfunding project initiator. Both narrations of affirmative and failed achievement can narratives in
demonstrate past project experience of the starter. More achievement language reveals more civic
domain expertise of starter. More achievement language may lead investors to believe the
initiator is more professional in terms of the target crowdfunding project and have the crowdfunding
state-of-the-art knowledge about expertise in related domains (Ericsson and Smith, 1991).
Prior studies indicated the fact that trust is critical in online financial activities because of the 1503
absence of proven guarantees (Kollock, 1999). When perceived related information confidence
and knowledge level of the service provider are high, online customers would be more likely to
trust in the products, services, providers and their organizations (Doney and Cannon, 1997).
According to the integrated model proposed by Gefen et al. (2003), these accumulate
knowledge-based and institutional-based trust leads to a higher level of investment intention of
potential decision makers. Thus, the following hypotheses are proposed:
H6. The use of achievement language positively affects funding success on civic
crowdfunding platforms.

4. Research model and methodology

4.1 Research context and data
The data for this research were collected from an UK-based online civic crowdfunding
platform, namely Spacehive. Established in 2011, Spacehive is one of the most popular
crowdfunding platforms dedicated mainly for civic projects that are focused on the
community program, social-oriented events, public space and infrastructure or facility
development. Spacehive offers a fixed funding mechanism that is already implemented by
reward-based crowdfunding platforms. In a fixed funding campaign, the project succeeds
only if it is pledged to surpass the stated funding goal within the specified fundraising
period. At Spacehive, project creators are required to specify campaign information,
including project type, description, target capital, campaign duration and cost breakdown
structure. A web crawler was developed and used to collect all projects that were in the idea,
fundraising, failed or successful status. A total of 905 civic crowdfunding projects were
gathered for December 2011–November 2015. Out of the collected 905 civic projects, 597
projects are either under the idea development or fundraising stage and thus excluded from
our sample. We finally obtained a sample of 308 civic projects.

4.2 Analytic strategy

Each of the narrative samples is analyzed using a computerized text analysis package called the
Linguistic Inquiry and Word Count (LIWC) to extract the features of the linguistic style and
message substance in the narratives. LIWC uses a word count strategy where it searches for
words in over 72 language dimensions or categories (McHaney et al., 2018; Tausczik and
Pennebaker, 2010). LIWC is designed to measure tone by measuring the extent words associated
with a set of psychological and linguistic categories. Originally developed to facilitate the
technique of emotional language analysis, the LIWC program is now widely used in
management research, in areas such as the study of advertising, business communication,
persuasion, leadership and financial reporting. We believe that LIWC is well-suited to the task of
analyzing the narratives pitched by project creators, as it contains dictionaries that focus not
only on relatively more state-dependent content words but also on function words that reflect the
linguistic style. Whereas some of the LIWC categories were initially derived from psychological
theories, more than 72 categories offered by the LIWC software program represent standard
linguistics categories such as person pronouns and articles, categories relating to affective
processes (e.g. anger and sadness), cognitive processes (e.g. insight), perceptual processes (e.g.
IMDS hear and listen) and categories that measure references to space, time and motion. Each category
119,7 corresponds to a predetermined dictionary of related words, so each message is represented as
approximately a 72-dimensional vector. The value in each dimension is computed as the number
of defined words belonging to that category divided by the total number of words in the
message. The output is expressed as a percentage of the total words within the text material.
The 2007 version of LIWC software has over 72 judge-defined variables that can be
1504 categorized mainly into four classes: linguistics features (e.g. pronouns, articles and
prepositions), relativity (e.g. past, present and future tenses), psychological (e.g. words
depicting affect such as cognitive, social, and perceptual words) and traditional content
dimension (e.g. personal concerns words relating to money, achievement and job). In this
study, we first focus on the psychological affect dimension of language as a peripheral route
to persuasion and then turn our attention to the issue-relevant content cues as a central
route to persuasion. Both the affect linguistic style and the issue-relevant content cues of
civic crowdfunding narratives are analyzed using the 2015 version of LIWC software.

4.3 Empirical model

As the dependent variable is dichotomous, we use logistic regression (logit) model with robust
standard errors. To accurately assess the effects of language style and message substance, we
run the base model with the control variables and then run two separate models. Model 1 is
the base model that includes only nine controls. Our choice of control variables closely follows
the extant literature on crowdfunding and persuasive language. Model 2 considers an affect-
related or psychological dimension of language as a peripheral cue to persuasion. We add
three independent variables, including positive affective, perceptual, and social language.
Model 3 includes three issue-relevant content cues, such as money, risk and achievement
contents, to analyze the impact of message substance on the success of crowdfunding.
Model 1 (Base model):
ProbabilityðSuccess ¼ 1Þ ¼ b0 þb1 FBFriendsþb2 Commentsþb3 Backers

þb4 PitchLengthþb5 Concrete þb6 Certainty

þb7 M aleRef erence þb8 Disrepancy þb9 LanguageI ntensityþe:

Model 2 (Affect-related language):

ProbabilityðSuccess ¼ 1Þ ¼ b0 þb1 FBFriendsþb2 Comments

þb3 Backersþb4 PitchLength þb5 Concrete þb6 Certainty

þb7 M aleRef erence þb8 Disrepancy þb9 LanguageI ntensity
þb10 Af f ective þb11 Perceptual þb12 Social þe:
Model 3 (Issue-relevant language):

ProbabilityðSuccess ¼ 1Þ ¼ b0 þ b1 FBFriends þb2 Comments þb3 Backersþ b4 PitchLength

þ b5 Concreteþ b6 Certaintyþ b7 M aleRef erenceþ b8 Disrepancy

þ b9 LanguageI ntensity þ b10 Af f ective þ b11 Perceptual þ b12 Social

þ b13 Quantatitive þb14 Risk þb15 Achievement þe:

4.4 Measures Role of
We operationalize crowdfunding success as a dichotomous variable indicating whether a narratives in
project is successfully funded or not. We generated a dummy variable, Success, which is civic
equal to 1 if the pledge amount is greater than or equal to the stated goal amount.
In order to evaluate hypotheses, we operationalize the psychological dimension of crowdfunding
languages such as positive affective, social and perceptual language and use them as
independent variables. LIWC scores the relative frequencies of words in each language 1505
category as a ratio to total words in that category in a campaign narrative to the total length
of the campaign narrative (Pennebaker and King, 1999).
Positive affective language is operationalized following (Anglin et al., 2018; Pennebaker
et al., 2001) as a ratio of words falling into positive affective language category within a
written narrative pitch, which is calculated by the number of positive affective words by the
total number of word in the narration. The positivity affective word list includes words such
as encourage, success, accomplish, strong, leading, high, great and expansion.
Perceptual language is operationalized as the ratio of perceptual words within the
narrative. A message that scored high in terms of the use of perceptual words would contain
terms that denote sensory experiences such as hearing, seeing, or having physical sensations.
Social language is operationalized as the ratio of the use of social intercourse-related
words such as neighbor, family, employee, advice and personal pronouns other than the
first-person singular within the narrative pitch.
In order to evaluate H4 through H6, we seek to operationalize issue-relevant content cues of
quantitative, risk and achievement language and use them as independent variables for Model 3.
The three content languages are operationalized using three distinct LIWC dictionary categories.
Quantitative language is operationalized as the ratio of number or money-related words
within the narrative. There is a consensus in the prior management literature that the use of
money-related evidence in the venture application signals entrepreneur effort and
preparedness behind the campaign (Chen et al., 2009). A text that scored high in terms of
quantitative information that is related to number, money, cost, sales, worth and profit.
Risk language is operationalized as the ratio of risk-related words within the narrative.
The risk-related language category includes obstacle, barrier, penalty, uncertain, error,
defect, cancel, discontinue, reluctant, prohibit, hinder, caution, doubtful and penalty.
Achievement language is operationalized as the ratio of achievement-related words in the
narrative. The achievement-related words refer to achievement striving, words indicative of
working toward a successful outcome – including the goal, effort and success (Tausczik and
Pennebaker, 2010).
Following prior studies in persuasion, we control for language features to be associated
with crowdfunding performance, such as concrete and certainty language (Parhankangas
and Renko, 2017). We also include the length of the narrative pitch as a control variable
since it could serve as a signal for preparedness. Further, following prior studies in
crowdfunding (Agrawal et al., 2015; Mollick, 2014), we control for several project quality
factors to be associated with crowdfunding success, such as project comment and social
capital (e.g. Facebook friends).

5. Results
5.1 Empirical results
The summary statistics, reported in Table I, show that the 308 civic crowdfunding projects
in our sample attracted over $8.16m. Also, the average dollar amount pledged to each
project is slightly over $26,500. Our final dataset from the Spacehive platform comprises 158
successful and 150 failed projects.
Table II presents the results of our regression analysis. In Model 1 of Table II, we first
begin with a baseline model that consists of only control variables in the analysis. In Model 2,
 119,7
IMDS

1506

Table I.

deviations
and standard
Correlations, means
Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1. FB Friends 119.208 166.980 1.000

2. Comments 11.153 32.491 0.153 1.000
3. Backers 35.016 86.927 0.051 0.415 1.000
4. Length of 1.000
Pitch 448.640 164.562 0.115 0.114 0.019
5. Concrete 22.818 2.669 −0.006 0.117 −0.016 −0.069 1.000
6. Certainty 0.909 0.616 −0.047 0.064 0.019 −0.143 −0.015 1.000
7. Male 1.000
Reference 0.093 0.291 −0.100 0.134 −0.020 0.008 0.063 −0.054
8. Discrepancy 0.906 0.720 −0.055 0.096 0.106 −0.147 −0.021 0.148 −0.062 1.000
9. Language 1.000
Intensity 0.257 0.486 −0.005 0.088 0.156 −0.080 −0.156 0.245 −0.015 0.195
10. Affective
Language 4.932 1.669 −0.053 −0.004 0.037 0.031 −0.175 0.140 0.055 −0.040 0.216 1.000
11. Perceptual 1.000
Language 2.048 1.671 −0.061 −0.046 −0.009 −0.264 −0.006 0.054 0.013 −0.116 0.078 0.006
12. Social
Language 10.229 3.235 −0.143 0.014 0.123 −0.083 −0.420 0.036 0.068 0.109 0.263 0.366 0.055 1.000
13. Quantitative 1.702 1.085 −0.041 −0.010 −0.038 −0.118 −0.030 −0.052 0.017 −0.077 −0.020 −0.189 0.079 −0.080 1.000
14. Risk 0.329 0.430 −0.022 −0.023 −0.083 0.009 0.018 −0.075 0.115 0.097 −0.074 −0.003 −0.151 0.016 −0.003 1.000
15. Achievement 1.436 0.825 −0.065 −0.055 −0.017 −0.025 −0.143 0.021 0.007 0.059 −0.031 0.077 0.032 0.106 −0.156 0.188 1.000
16. Success 0.512 0.501 0.292 0.218 0.339 0.060 0.012 −0.021 −0.116 0.009 0.149 0.066 0.113 −0.061 0.054 −0.177 −0.047 1.000
Note: n ¼ 308
 Base Model (1) Model (2) Model (3)
Role of
D.V. Success Success Success narratives in
civic
1. FB friends (logged) 0.286*** (0.061) 0.293*** (0.062) 0.311*** (0.064)
2. Comments (logged) 0.102 (0.102) 0.125 (0.108) 0.126 (0.105) crowdfunding
3. Backers (logged) 0.596*** (0.123) 0.635*** (0.127) 0.653*** (0.128)
4. Length of pitch 0.001 (0.001) 0.001 (0.001) 0.001 (0.001)
5. Concrete 0.094 (0.076) 0.059 (0.083) 0.057 (0.084) 1507
6. Certainty −0.183 (0.235) −0.283 (0.242) −0.335 (0.235)
7. Male reference −1.225* (0.676) −1.230* (0.679) −1.253* (0.678)
8. Discrepancy −0.111 (0.175) 0.033 (0.175) 0.168 (0.189)
9. Language intensity 0.670 (0.318) 0.614 (0.337) 0.574 (0.342)
10. Affective 0.175** (0.085) 0.221** (0.088)
11. Perceptual 0.230** (0.088) 0.188** (0.086)
12. Social −0.090* (0.049) −0.105** (0.049)
13. Quantitative 0.285** (0.138)
14. Risk −1.185** (0.435)
15. achievement 0.032 (0.179) Table II.
Constant −2.992*** (0.868) −3.474** (1.282) −4.524** (1.447) Logistic regressions
Pseudo R2 0.185 0.215 0.239 analysis on
Notes: n ¼ 308. Standard errors in parentheses. *p o0.1; **po 0.05; ***po 0.01 crowdfunding success

the three independent variables, related to psychological language or affect-related language,

are added to the base model. Model 3 adds the measures of message substance or content cues.
The outcome measure for Model 1 through 3 is the success of crowdfunding projects, which is
equal to 1 if a focal civic crowdfunding project acquires the desired target goal amount, and
zero otherwise.
H1 predicted a positive relationship between positive affective language and funding
success. The coefficient of affective language in Column (10) of Model 3 is positive and
significant (β ¼ 0.221, p o0.05), suggesting that the affective language has a positive impact
on funding outcome. Thus, H1 is supported.
Further, H2 stated that perceptual language would be positively related to the success of
crowdfunding. We find support for this hypothesis (β ¼ 0.188, po0.05) revealing that the
use of perceptual language improves the likelihood of funding success.
H3 predicted that social language is positively related to the success of the funding
outcome. As shown in Column (12) of Model 3, the result indicates that the relationship
between social language and the likelihood of project success is negative and statistically
significant (β ¼ −0.105, p o 0.05). Thus, H3 is not supported. On the contrary to our
hypothesis, we found that social language had a negative effect on the odds of success of
civic crowdfunded projects. Our estimates indicate that in civic crowdfunding
marketplaces extensive usage of social language does not help project creators to
increase their chance of the funding performance, but rather such language reduces the
likelihood of funding performance.
H4 stated that quantitative language of product pitch is positively related to the
success of civic crowdfunding. We thus expect that coefficient of the quantitative message
to be positive and statistically significant. In Column (13) of Model 3, the parameter
estimates show that the effect of money language, along with numerical information,
is significant and positive (β ¼ 0.285, p o 0.05). Thus, H4 is supported. Our finding
supports evidence that inclusion of money-related content in the project pitch positively
influence the financing of civic crowdfunding projects.
H5 predicted a negative relationship between risk language and funding success. We
find support for this hypothesis (β ¼ −1.185, p o0.01) as shown in Column (14) of Model 3.
IMDS As expected, we found that disclosure of risk factor explicitly in the project pitch could
119,7 decrease the odds of funding success of crowdfunding projects.
Finally, H6 predicted that achievement language improves funding performance. In
Model 3, the parameter estimates show that the relationship between achievement language
and the likelihood of project success is not significant, and thus H6 is not supported.

1508 5.2 Robustness checks

To further verify the robustness of our results, we test our hypotheses using the probit
regression and an alternate measure of funding performance. The results are provided in
Tables III and IV.
Alternative regression method. To test for the robustness of our choice to use the logit
model, we performed an additional regression using the probit model as well, in order to
ascertain if our results are influenced by the functional form of the logit regression model or
not. The additional regression was run with funding outcome as the dependent variable and
the results showed similar patterns compared to our logit model.
Alternative measure of funding performance. To enhance confidence in our main
findings, we tested our results with an alternate measure of funding performance, the actual
percent funded ( funding ratio) as dependent variables. In line with the existing literature
products (Burtch et al., 2013), we operationalize funding performance with the funding ratio,
which is computed by dividing funds raised by fund requested. The results show similar
patterns compared to the model when funding success is used as the dependent variable in
our main analysis.

6. Discussion and implications

6.1 Practical contributions
Our study offers distinct recommendations for project campaigners who intend to use online
crowdfunding platforms to raise funds for a civic or public project in which it mainly
concerns with private contributions toward public goods. This research can contribute to
practitioners with some implications.

Base Model (1) Model (2) Model (3)

D.V. Success Success Success

1. FB Friends (logged) 0.171*** (0.036) 0.175*** (0.036) 0.187*** (0.037)

2. Comments (logged) 0.0615 (0.061) 0.0705 (0.063) 0.063 (0.063)
3. Backers (logged) 0.355*** (0.070) 0.377*** (0.072) 0.387*** (0.073)
4. Length of pitch 0.001 (0.001) 0.001 (0.001) 0.001 (0.001)
5. Concrete 0.057 (0.044) 0.036 (0.048) 0.062 (0.050)
6. Certainty −0.117 (0.135) −0.181 (0.138) −0.213 (0.137)
7. Male reference −0.711* (0.379) −0.712* (0.383) −0.735* (0.391)
8. Discrepancy −0.070 (0.108) 0.016 (0.108) 0.093 (0.113)
9. Language intensity 0.392* (0.187) 0.352* (0.197) 0.323* (0.200)
10. Affective 0.108** (0.050) 0.138** (0.052)
11. Perceptual 0.137** (0.052) 0.115** (0.051)
12. Social −0.054** (0.029) −0.063** (0.029)
13. Quantitative 0.162* (0.081)
14. Risk −0.669*** (0.248)
Table III.
Robustness check: 15. Achievement 0.030 (0.103)
probit regressions Constant −1.777*** (0.512) −2.065** (0.743) −2.650** (0.832)
analysis on Pseudo R2 0.184 0.214 0.238
crowdfunding success Notes: n ¼ 308. Standard errors in parentheses. *p o 0.1; **p o0.05; ***p o0.01
 Base Model (1) Model (2) Model (3)
Role of
D.V. Funding ratio Funding ratio Funding ratio narratives in
civic
1. FB friends (logged) 0.043*** (0.010) 0.041*** (0.010) 0.042*** (0.010)
2. Comments (logged) 0.018 (0.015) 0.019 (0.015) 0.018 (0.015) crowdfunding
3. Backers (logged) 0.096*** (0.016) 0.099*** (0.016) 0.097*** (0.016)
4. Length of pitch 0.001 (0.001) 0.001 (0.001) 0.001 (0.001)
5. Concrete 0.002 (0.008) 0.003 (0.008) 0.0001 (0.008) 1509
6. Certainty −0.117 (0.135) −0.068 (0.039) −0.072* (0.038)
7. Male reference −0.161* (0.064) −0.157* (0.058) −0.137* (0.066)
8. Discrepancy −0.023 (0.030) 0.005 (0.028) 0.018 (0.043)
9. Language intensity 0.114** (0.041) 0.112* (0.042) 0.103** (0.043)
10. Affective 0.032** (0.014) 0.037*** (0.014)
11. Perceptual 0.041*** (0.014) 0.035*** (0.013)
12. Social −0.021*** (0.007) −0.020*** (0.007)
13. Quantitative 0.041** (0.020)
14. Risk −0.135*** (0.0.44) Table IV.
15. Achievement 0.021 (0.025) Robustness check:
Adjusted R2 0.215 0.259 0.284 regressions analysis
Notes: n ¼ 308. Standard errors in parentheses. *p o0.1; **po 0.05; ***po 0.01 on crowdfunding ratio

First, platform managers and project campaigners may leverage the impacts of the
narrative pitch through the lens of our research model. The findings encourage practitioners
to focus on the role of affect-related language cues in civic crowdfunding. Specifically, from
the peripheral perspective of persuasion, tone or mood of the narratives is helpful to backers’
purchase decision-making process since verifiable evidence of civic projects is often not
readily available to the backers. Therefore, when presenting project pitches to potential
backers, the project proponents are suggested to include both positive affective and
perceptual language, leading to a greater likelihood of funding success.
Second, based on the direct effects of issue-relevant content cues, project campaigners
are suggested to strengthen the argument quality of the narration with more detailed issues
and more quantitative numbers. Concerning the effect of the central route to persuasion, our
results show that project proponents can facilitate consumers’ decision making by framing
their narratives with issue-relevant information and using of money-related language.
Third, for campaign initiators, this study encourages project campaigners to understand
both linguistic style and message substance of their written narrative pitch and to balance
such dual process factors that are essential to drive backers’ funding decisions.

6.2 Theoretical contributions

Our research makes significant contributions to the crowdfunding literature and the ELM of
persuasion in three ways.
First, our findings contribute to, and extend, the emerging stream of research on online
entrepreneurial financing in general and civic crowdfunding campaign. Different from prior
literature, this study focuses on exploring the message cues instead of entrepreneur
characteristics and product-related factors (Agrawal et al., 2015; Mollick, 2014) in the
crowdfunding platform. Moreover, this research examines the effects of both linguistic style
and message substance simultaneous. The research results indicated that two types of
linguistic features are related to funding success in the context of civic crowdfunding.
Second, we advance the application of the dual process in ELM by identifying and
examining distinct persuasive cues originating from linguistics styles and message
substance. The psychological dimension of languages, such as affective positive, cognitive
and perceptual language, influences backers’ funding decision. Specifically, affect-related
IMDS components in the narratives are effective to evoke a peripheral route to persuasion.
119,7 Narrative pitches that are framed with the use of affective linguistic style can serve as a
valuable signal for backers as objective information conveyed explicitly through the
message contents.
Third, this study extends the ELM of persuasion to the civic crowdfunding context.
Through the validation of our work, we provide a theoretical understanding of how
1510 psychological dimension of language influence funders’ investment decisions via the
central and peripheral route in the civic crowdfunding marketplace. The results of this
study shed light on how linguistic style and message substance work to affect persuasion.
This study demonstrates that the role of narrative could be considered a particularly
formidable variable within ELM and its effects exit in the persuasive process of
cro6wdfunding project investment.

6.3 Limitation and future research

There are several limitations to this study. First, our sample was all collected from
Spacehive in the UK; however, the role of the narrative may varies depending on different
cultures and platforms. Another study can be conducted to verify the findings of linguistic
style and message substance based on cross-platform data. The comparison can further
examine whether culture difference shapes the effects of narrative in civic crowdfunding.
Second, this research chooses textural message on the platform for analyzing the
linguistic style and message substance of each project. Some of the information cues are
delivered to backers via pictures and videos in the project campaign webpage.
We encourage future research to explore the effects of the narrative by combining multi-
media message with textural message together and advance our knowledge regarding
civic crowdfunding.

References
Abraham, S. and Cox, P. (2007), “Analysing the determinants of narrative risk information in UK FTSE
100 annual reports”, The British Accounting Review, Vol. 39 No. 3, pp. 227-248.
Agrawal, A., Catalini, C. and Goldfarb, A. (2015), “Crowdfunding: geography, social networks, and the
timing of investment decisions”, Journal of Economics & Management Strategy, Vol. 24 No. 2,
pp. 253-274.
Ahlers, G.K., Cumming, D., Günther, C. and Schweizer, D. (2015), “Signaling in equity crowdfunding”,
Entrepreneurship Theory and Practice, Vol. 39 No. 4, pp. 955-980.
Allison, T.H., McKenny, A.F. and Short, J.C. (2013), “The effect of entrepreneurial rhetoric on
microlending investment: an examination of the warm-glow effect”, Journal of Business
Venturing, Vol. 28 No. 6, pp. 690-707.
Allison, T.H., Davis, B.C., Short, J.C. and Webb, J.W. (2015), “Crowdfunding in a prosocial microlending
environment: examining the role of intrinsic versus extrinsic cues”, Entrepreneurship Theory
and Practice, Vol. 39 No. 1, pp. 53-73.
Allison, T.H., Davis, B.C., Webb, J.W. and Short, J.C. (2017), “Persuasion in crowdfunding: an
elaboration likelihood model of crowdfunding performance”, Journal of Business Venturing,
Vol. 32 No. 6, pp. 707-725.
Alter, A.L. and Oppenheimer, D.M. (2009), “Uniting the tribes of fluency to form a metacognitive
nation”, Personality and Social Psychology Review, Vol. 13 No. 3, pp. 219-235.
Anglin, A.H., Short, J.C., Drover, W., Stevenson, R.M., McKenny, A.F. and Allison, T.H. (2018), “The
power of positivity? The influence of positive psychological capital language on crowdfunding
performance”, Journal of Business Venturing, Vol. 33 No. 4, pp. 470-492.
Angst, C.M. and Agarwal, R. (2009), “Adoption of electronic health records in the presence of privacy Role of
concerns: the elaboration likelihood model and individual persuasion”, MIS Quarterly, Vol. 33 narratives in
No. 2, pp. 339-370.
Bhattacherjee, A. and Sanford, C. (2006), “Influence processes for information technology acceptance:
civic
an elaboration likelihood model”, MIS Quarterly, Vol. 30 No. 4, pp. 805-825, available at: crowdfunding
www.jstor.org/stable/25148755
Burtch, G., Ghose, A. and Wattal, S. (2013), “An empirical examination of the antecedents and 1511
consequences of contribution patterns in crowd-funded markets”, Information Systems
Research, Vol. 24 No. 3, pp. 499-519.
Cacioppo, J.T. and Petty, R.E. (1984), “The elaboration likelihood model of persuasion”, Advances in
Consumer Research, Vol. 11, pp. 673-675.
Calic, G. and Mosakowski, E. (2016), “Kicking off social entrepreneurship: how a sustainability
orientation influences crowdfunding success”, Journal of Management Studies, Vol. 53 No. 5,
pp. 738-767.
Chen, X.-P., Yao, X. and Kotha, S. (2009), “Entrepreneur passion and preparedness in business plan
presentations: a persuasion analysis of venture capitalists’ funding decisions”, Academy of
Management Journal, Vol. 52 No. 1, pp. 199-214.
Covello, V.T. and Sandman, P.M. (2001), “Risk communication: evolution and revolution”, in Wolbarst, A.
(Ed.), Solutions to an Environment in Peril, John Hopkins University Press, Baltimore, MD,
pp. 164-178.
Davies, R. (2015), “Three provocations for civic crowdfunding”, Information, Communication &
Society, Vol. 18 No. 3, pp. 342-355.
Davis, A.K., Piger, J.M. and Sedor, L.M. (2006), “Beyond the numbers: an analysis of optimistic and
pessimistic language in earnings press releases”, Federal Reserve Bank of St Louis, St Louis, MO.
Davis, M.H. (2018), Empathy: A Social Psychological Approach, Routledge, New York, NY.
DeYoung, C.G., Peterson, J.B. and Higgins, D.M. (2005), “Sources of openness/intellect: cognitive and
neuropsychological correlates of the fifth factor of personality”, Journal of Personality, Vol. 73
No. 4, pp. 825-858.
Doney, P.M. and Cannon, J.P. (1997), “An examination of the nature of trust in buyer-seller
relationships”, The Journal of Marketing, Vol. 61 No. 2, pp. 35-51.
Eisenberg, N., Cumberland, A. and Spinrad, T.L. (1998), “Parental socialization of emotion”,
Psychological Inquiry, Vol. 9 No. 4, pp. 241-273.
Ericsson, K.A. and Smith, J. (1991), “Prospects and limits of the empirical study of expertise: an
introduction”, in Ericsson, K.A. and Smith, J. (Eds), Toward a General Theory of Expertise,
Cambridge University Press, New York, NY, pp. 1-38.
Fahnestock, J. (2011), Rhetorical Style: The Uses of Language in Persuasion, Oxford University Press,
OXford.
Federmeier, K.D. (2007), “Thinking ahead: the role and roots of prediction in language comprehension”,
Psychophysiology, Vol. 44 No. 4, pp. 491-505.
Frydrych, D., Bock, A.J. and Kinder, T. (2016), “Creating project legitimacy: the role of entrepreneurial
narrative in reward-based crowdfunding”, in Méric, J., Maque, I. and Brabet, J. (Eds),
International Perspectives on Crowdfunding: Positive, Normative and Critical Theory, Emerald
Group Publishing Limited, Bingley, pp. 99-128.
Fussell, S.R. and Krauss, R.M. (1989), “Understanding friends and strangers: the effects of audience design
on message comprehension”, European Journal of Social Psychology, Vol. 19 No. 6, pp. 509-525.
Gefen, D., Karahanna, E. and Straub, D.W. (2003), “Trust and TAM in online shopping: an integrated
model”, MIS Quarterly, Vol. 27 No. 1, pp. 51-90.
Gibbons, P., Busch, J. and Bradac, J.J. (1991), “Powerful versus powerless language: consequences for
persuasion, impression formation, and cognitive response”, Journal of Language and Social
Psychology, Vol. 10 No. 2, pp. 115-133.
IMDS Gorbati, A. and Nelson, L. (2015), “Gender and the language of crowdfunding”, Academy of
119,7 Management Proceedings, Vol. 2015 No. 1, p. 15785.
Hassna, G., Burtch, G., Lee, C.H. and Zhao, J.L. (2018), “Understanding the role of lead donor types in
civic crowdfunding”, available at: https://ssrn.com/abstract=3175916; http://dx.doi.org/10.2139/
ssrn.3175916
Henry, E. (2008), “Are investors influenced by how earnings press releases are written?”, The Journal of
1512 Business Communication, Vol. 45 No. 4, pp. 363-407.
Herzenstein, M., Sonenshein, S. and Dholakia, U.M. (2011), “Tell me a good story and I may lend you
money: the role of narratives in peer-to-peer lending decisions”, Journal of Marketing Research,
Vol. 48, pp. S138-S149, available at: https://doi.org/10.1509/jmkr.48.SPL.S13
Hosman, L.A. (2002), “Language and persuasion”, in Dillard, J.P. and Pfau, M. (Eds), The Persuasion
Handbook: Developments in Theory and Practice, Sage, New York, NY, pp. 371-390.
Jordan, T.R. and Thomas, S.M. (2002), “In search of perceptual influences of sentence context on word
recognition”, Journal of Experimental Psychology: Learning, Memory, and Cognition, Vol. 28
No. 1, pp. 34-45.
Kang, M., Gao, Y., Wang, T. and Zheng, H. (2016), “Understanding the determinants of funders’
investment intentions on crowdfunding platforms: a trust-based perspective”, Industrial
Management & Data Systems, Vol. 116 No. 8, pp. 1800-1819.
Kim, K. and Viswanathan, S. (2018), “The ‘experts’ in the crowd: the role of experienced investors in a
crowdfunding market”, Mis Quarterly, Vol. 43 No. 2.
Kim, K., Bae, J., Nho, M.-W. and Lee, C.H. (2011), “How do experts and novices differ? Relation versus
attribute and thinking versus feeling in language use”, Psychology of Aesthetics, Creativity, and
the Arts, Vol. 5 No. 4, pp. 379-388.
Kollock, P. (1999), “The production of trust in online markets”, Advances in Group Processes, Vol. 16
No. 1, pp. 99-123.
Kruglanski, A.W., Belanger, J.J., Chen, X., Kopetz, C., Pierro, A. and Mannetti, L. (2012), “The energetics
of motivated cognition: a force-field analysis”, Psychological Review, Vol. 119 No. 1, pp. 1-20.
Larrimore, L., Jiang, L., Larrimore, J., Markowitz, D. and Gorski, S. (2011), “Peer to peer lending: the
relationship between language features, trustworthiness, and persuasion success”, Journal of
Applied Communication Research, Vol. 39 No. 1, pp. 19-37.
Lee, C.H., Zhao, J.L. and Hassna, G. (2016), “Government-incentivized crowdfunding for one-belt, one-
road enterprises: design and research issues”, Financial Innovation, Vol. 2 No. 1, p. 1-14.
Linsley, P.M. and Shrives, P.J. (2006), “Risk reporting: a study of risk disclosures in the annual reports
of UK companies”, The British Accounting Review, Vol. 38 No. 4, pp. 387-404.
Ludwig, S., De Ruyter, K., Friedman, M., Brüggen, E.C., Wetzels, M. and Pfann, G. (2013), “More than
words: the influence of affective content and linguistic style matches in online reviews on
conversion rates”, Journal of Marketing, Vol. 77 No. 1, pp. 87-103.
McHaney, R., Tako, A. and Robinson, S. (2018), “Using LIWC to choose simulation approaches: a feasibility
study”, Decision Support Systems, Vol. 111, pp. 1-12, available at: https://linkinghub.elsevier.com/
retrieve/pii/S0167923618300691
MacMillan, I.C., Siegel, R. and Narasimha, P.S. (1985), “Criteria used by venture capitalists to evaluate
new venture proposals”, Journal of Business Venturing, Vol. 1 No. 1, pp. 119-128.
Majumdar, A. and Bose, I. (2018), “My words for your pizza: an analysis of persuasive narratives in
online crowdfunding”, Information & Management, Vol. 55 No. 6, pp. 781-794.
Martens, M.L., Jennings, J.E. and Jennings, P.D. (2007), “Do the stories they tell get them the money they
need? The role of entrepreneurial narratives in resource acquisition”, Academy of Management
Journal, Vol. 50 No. 5, pp. 1107-1132.
Massolution, C.L. (2015), “Crowdfunding industry report”.
Miller, C.H., Lane, L.T., Deatrick, L.M., Young, A.M. and Potts, K.A. (2007), “Psychological reactance Role of
and promotional health messages: the effects of controlling language, lexical concreteness, and narratives in
the restoration of freedom”, Human Communication Research, Vol. 33 No. 2, pp. 219-240.
civic
Miller, G.A. (1951), Language and Communication, McGraw-Hill, New York, NY.
crowdfunding
Miller, J.M. (2000), “Language use, identity, and social interaction: migrant students in Australia”,
Research on Language and Social Interaction, Vol. 33 No. 1, pp. 69-100.
Mollick, E. (2014), “The dynamics of crowdfunding: an exploratory study”, Journal of Business 1513
Venturing, Vol. 29 No. 1, pp. 1-16.
Moss, T.W., Renko, M., Block, E. and Meyskens, M. (2018), “Funding the story of hybrid ventures:
crowdfunder lending preferences and linguistic hybridity”, Journal of Business Venturing,
Vol. 33 No. 5, pp. 643-659.
Parhankangas, A. and Renko, M. (2017), “Linguistic style and crowdfunding success among social and
commercial entrepreneurs”, Journal of Business Venturing, Vol. 32 No. 2, pp. 215-236.
Park, H.S., Levine, T.R., Kingsley Westerman, C.Y., Orfgen, T. and Foregger, S. (2007), “The effects of
argument quality and involvement type on attitude formation and attitude change: a test of
dual-process and social judgment predictions”, Human Communication Research, Vol. 33 No. 1,
pp. 81-102.
Peng, K.-F., Fan, Y.-W. and Hsu, T.-A. (2004), “Proposing the content perception theory for the online
content industry–a structural equation modeling”, Industrial Management & Data Systems,
Vol. 104 No. 6, pp. 469-489.
Pennebaker, J.W. and King, L.A. (1999), “Linguistic styles: language use as an individual difference”,
Journal of Personality and Social Psychology, Vol. 77 No. 6, pp. 1296-1312.
Pennebaker, J.W., Francis, M.E. and Booth, R.J. (2001), “Linguistic Inquiry and Word Count”, Lawrence
Erlbaum Associates, Mahway, NJ.
Petty, R.E. and Cacioppo, J.T. (1986), “The elaboration likelihood model of persuasion”, Communication
and Persuasion, Springer, New York, NY, pp. 1-24, available at: https://doi.org/10.1007/978-1-
4612-4964-1_1
Petty, R.E., Cacioppo, J.T. and Schumann, D. (1983), “Central and peripheral routes to advertising
effectiveness: the moderating role of involvement”, Journal of Consumer Research, Vol. 10 No. 2,
pp. 135-146.
Petty, R.E., Schumann, D.W., Richman, S.A. and Strathman, A.J. (1993), “Positive mood and persuasion:
different roles for affect under high-and low-elaboration conditions”, Journal of Personality and
Social Psychology, Vol. 64 No. 1, pp. 5-20.
Renn, O. and Levine, D. (1991), “Credibility and trust in risk communication”, in Kasperson, R.E. and
Stallen, P.J.M. (Eds), Communicating Risks to the Public, Kluwer Academic, Dordrecht, pp. 175-218.
Ryu, H.-S. (2018), “What makes users willing or hesitant to use Fintech? The moderating effect of user
type”, Industrial Management & Data Systems, Vol. 118 No. 3, pp. 541-569.
Schwarz, N., Bless, H. and Bohner, G. (1991), “Mood and persuasion: affective states influence the
processing of persuasive communications”, in Zanna, M.P. (Ed.), Advances in Experimental
Social Psychology, Vol. 24, Academic Press, Orlando, FL, pp. 161-199.
Simmons, R.A., Chambless, D.L. and Gordon, P.C. (2008), “How do hostile and emotionally overinvolved
relatives view relationships? What relatives’ pronoun use tells us”, Family Process, Vol. 47 No. 3,
pp. 405-419.
Sparks, J.R. and Areni, C.S. (2008), “Style versus substance: multiple roles of language power in
persuasion”, Journal of Applied Social Psychology, Vol. 38 No. 1, pp. 37-60.
Stiver, A., Barroca, L., Minocha, S., Richards, M. and Roberts, D. (2015), “Civic crowdfunding
research: challenges, opportunities, and future agenda”, New Media & Society, Vol. 17 No. 2,
pp. 249-271.
Tam, K.Y. and Ho, S.Y. (2005), “Web personalization as a persuasion strategy: an elaboration likelihood
model perspective”, Information Systems Research, Vol. 16 No. 3, pp. 271-291.
IMDS Tan, K.W., Swee, D., Lim, C., Detenber, B.H. and Alsagoff, L. (2007), “The impact of language variety
119,7 and expertise on perceptions of online political discussions”, Journal of Computer-Mediated
Communication, Vol. 13 No. 1, pp. 76-99.
Tausczik, Y.R. and Pennebaker, J.W. (2010), “The psychological meaning of words: LIWC and
computerized text analysis methods”, Journal of Language and Social Psychology, Vol. 29 No. 1,
pp. 24-54.
Tetlock, P.C., Saar-Tsechansky, M. and Macskassy, S. (2008), “More than words: quantifying language
1514 to measure firms’ fundamentals”, The Journal of Finance, Vol. 63 No. 3, pp. 1437-1467.
Toma, C.L. and D’Angelo, J.D. (2015), “Tell-tale words: linguistic cues used to infer the expertise of
online medical advice”, Journal of Language and Social Psychology, Vol. 34 No. 1, pp. 25-45.
Weiskopf, D.A. (2010), “Embodied cognition and linguistic comprehension”, Studies in History and
Philosophy of Science, Vol. 41 No. 3, pp. 294-304.

Corresponding author
Chang Heon Lee can be contacted at: changlee@uaeu.ac.ae

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
 The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm

Exploring
Exploring individuals’ behavioral individuals’
intentions toward donation behavioral
intentions
crowdfunding: evidence from China
Tao Wang and Yalan Li 1515
School of Economic Information Engineering,
Received 15 October 2018
Southwestern University of Finance and Economics, Chengdu, China Revised 19 May 2019
Minghui Kang 19 July 2019
Accepted 25 July 2019
School of Economics, Sichuan Agricultural University, Chengdu, China, and
Haichao Zheng
School of Economic Information Engineering,
Southwestern University of Finance and Economics, Chengdu, China

Abstract
Purpose – The purpose of this paper is to apply the self-determination theory (SDT) to propose a research model
that incorporates the SDT framework and contextual variables as determinants and self-identity and social
identity as mediating constructs to predict individuals’ intentions toward donation crowdfunding in China.
Design/methodology/approach – Structural equation modeling is used to analyze the data collected
from China.
Findings – The results indicate that the self-identity and social identity collectively or separately mediate the
effect exerted by the sense of self-worth, face concern, moral obligation, perceived donor effectiveness, social
interaction and referent network size on donation intentions. However, there is no evidence supporting the
hypothesis connecting moral obligation with self-identity.
Practical implications – The study provides suggestions for service providers on how to improve and
perfect the functions, and it also provides insights for donation crowdfunding fundraisers on how to increase
the success rate.
Originality/value – The conclusions of this study provide academics with a more thorough understanding of
the driving forces of individual behavior intention toward donation crowdfunding in China. This study further
expands the SDT and identity theory in the context of donation crowdfunding, which improves their robustness in
explaining behavioral intention. These theories may be an important part of future information system research.
Keywords Self-identity, Social identity, Self-determination theory, Donation crowdfunding,
Individuals’ behavior, Self-determination theory (SDT)
Paper type Research paper

1. Introduction
Although donation crowdfunding was only introduced to China less than five years ago, it is
already considered the most popular crowdfunding approach today that can effectively
organize the funding of fixed project development costs (Boudreau et al., 2015). From
rescuing homeless animals to overcoming medical crises, eliminating community problems
to reallocating educational resources, donation crowdfunding covers an incredibly broad
spectrum of topics. Furthermore, donation crowdfunding has attracted a surge of interest
from researchers, industry practitioners and even the general public (Boudreau et al., 2015).
According to the World Bank, China is a rising star in the global crowdfunding market, and
by the end of 2025, it will reach a size of $50bn, among which donation crowdfunding is
expected to account for one-fifth of that figure. Moreover, based on the report released
Industrial Management & Data
Systems
Vol. 119 No. 7, 2019
This work has been supported by the National Natural Science Foundation of China (71502145, pp. 1515-1534
71702155), the Fundamental Research Funds for the Central Universities (No. JBK1902032), and the © Emerald Publishing Limited
0263-5577
Key Laboratory of Financial Intelligence and Financial Engineering of Sichuan Province. DOI 10.1108/IMDS-10-2018-0451
IMDS by Tencent Charity, 20 Internet fundraising information platforms designated by the
119,7 Charity Law issued 21,000 fundraising documents for more than 1,400 public charitable
organizations nationwide in 2018. Meanwhile, over 8.46bn people are involved, raising a
total of more than 3.17bn yuan, an increase of 26.8 percent compared with 2017.
Compared with traditional charity, donation crowdfunding is free from the allocation
cycle of traditional capital, and it has broken the restrictions of time and space. Through
1516 donation crowdfunding, smaller entrepreneurs or individuals who traditionally had great
difficulty in obtaining capital now have access to anyone in the world with a computer and
Internet access, and gather up small amounts of funds from crowds without the limitations
of time and space. With the wide reach of social networking and the popularity of mobile
payments, donation crowdfunding has more of the general public involved in it. It has
become a new belief system and has changed the nature of charity, which encourages
everyone to do something good.
Although donation crowdfunding provides exciting opportunities for both fundraisers and
funders, it also has many challenges to overcome. Donation crowdfunding remains a topic of
hot debate, with questions raised about whether the wider public should engage in donation
crowdfunding, characterized by many as highly risky and lacking donor protections (Boudreau
et al., 2015). As an emerging topic, the relevant literature has paid considerable attention to loan-
based, equity-based and reward-based crowdfunding, with the exception of donation
crowdfunding, not to mention what factors influence donation intention. Insufficient research on
the donation intention of the public raises critical questions, particularly when evidence
indicates that donation crowdfunding has experienced an extremely low success rate. Moreover,
despite the fact that the proportion of the population engaged in donation crowdfunding is
much higher than other types, when considering the total population of China, it is evident that
too few people are willing to participate in it. Given that the success of a crowdfunding project
depends entirely on the involvement of potential investors, understanding their motivations and
behaviors is a more compelling purpose of this research area.
Drawing on the existing literature, we have identified potential factors underlying
donation intentions that adopted the self-determination theory (SDT) and identity theory as
the theoretical foundation. Based on SDT, most individuals’ behavior is driven by the
combination of extrinsic and intrinsic motivations. In addition to these, previous studies
have also revealed the importance of contextual support in influencing individuals’ behavior
(Ryan and Deci, 2000b); therefore, we have also incorporated contextual factors into our
model to provide a better understanding for individuals’ intentions. To shed light on this, we
have looked over extensive literature on donation crowdfunding and identity research.
However, owing to the fact that donation crowdfunding is still at a preliminary stage,
empirically validated research that is relevant to individuals’ donation intentions is virtually
absent. For this reason, we have reviewed an expanded range of related studies, including
three other kinds of crowdfunding, as well as charitable giving. Eventually, SDT was
adopted as the theoretical foundation and incorporated self-identity and social identity as
mediating constructs to analyze the antecedents of individuals’ donation intentions.
The model combined contextual factors with the SDT framework including extrinsic
motivations and intrinsic motivations. To test the model, structural equation modeling was
introduced to analyze data collected from 588 respondents in China.
The study will be of great significance to both industry and academia. Specifically, this
paper provides service providers suggestions on how to improve and perfect the functions,
and it also provides insights for donation crowdfunding fundraisers into how to increase the
success rate from a practical perspective. From a theoretical perspective, the findings of this
study provide a better understanding of the driving forces of individuals’ intentions toward
donation crowdfunding for academics. This study also further expands the SDT and
identity theory into the context of donation crowdfunding, which improves their robustness
in explaining behavior intention. These theories may be important in future information Exploring
system research. individuals’
The rest of this paper is organized as follows: Section 2 presents a review of the existing behavioral
literature on donation crowdfunding, SDT and identity theory. Section 3 presents the research
hypotheses, specifying the factors determining individuals’ donation intentions. Section 4 intentions
outlines the research methodology. Section 5 provides the results of empirical tests, followed
by a summary of the findings and a discussion of the implications of the research. Finally, 1517
limitations and suggestions for future research are identified in the last section.

2. Conceptual foundation
2.1 Donation crowdfunding
The previous literature identifies four main types of crowdfunding based on what funders
receive in exchange for their donation (Ewe et al., 2015): rewards-based crowdfunding involves
individuals contributing to a business in exchange for a reward, typically a form of the
product or service that the company offers. Equity-based crowdfunding allows contributors to
become part-owners of a company by trading capital for equity shares. As equity owners,
contributors receive a financial return on their investment and ultimately receive a share of the
profits in the form of a dividend or distribution. Lending-based crowdfunding allows
fundraisers to raise funds in the form of loans that they will pay back to the lenders over a
pre-determined timeline with a set interest rate. Donation-based crowdfunding is the simplest
way to crowdfund a cause or a project. Donations-based crowdfunding runs on a charity
basis. The motivation to make a contribution is not financial profit. Instead, donators see their
return and reward in the form of happiness, seeing that their funds are used for an honorable
cause such as a social project (Morningpost in BeiJing, 2014). Since each type of crowdfunding
features a distinctive operational mode, these four types of crowdfunding are usually analyzed
separately. In this paper, we focus on donation crowdfunding.
The first officially recorded use of donation crowdfunding in China dates back to 2014, and
donation crowdfunding is defined as voluntary contributions to funding a public good, for
which individuals usually receive intangible rewards in return, similar to charitable donations
(Burtch et al., 2011). As a distinctive problem-solving strategy and a widespread social
practice, although it has only emerged over the past few years, donation crowdfunding has
been recognized as an alternative source of capital for a variety of personal and public
objectives. An increasing number of organizations or individuals who initially have
insufficient capital to support a cause now have more opportunities to obtain funding.
Given the natural difference between donation crowdfunding and the remaining
crowdfunding types, examining the principles that apply to investors in financial markets
such as equity crowdfunding may not be suitable in the context of donation crowdfunding.
For this reason, we have looked further over the literature about charitable giving in
different cultural settings, which indicates that individuals make donations to charities
mainly because they derive satisfaction from the feeling of making a difference or from the
act of giving, to receive acclaim or prestige in society or because they are constrained by
moral principles (Andreoni, 1989). The natural similarity between donation crowdfunding
and charitable giving provides us with a theoretical foundation to explore individual
behavior intention toward donation crowdfunding from the perspective mentioned above.
Considering the contradiction between the fact that donation crowdfunding attracted an
extremely large population in 2018 and that the empirical research on the pre-factors of
personal involvement in donation crowdfunding is surprisingly rare, it is extremely
essential to probe into how the public views the issue. Moreover, the natural distinction
between donation crowdfunding and the other three reveals the necessity to explore
individual behavioral intention from an original perspective, which explains the reason why
we performed the present empirical study.
IMDS 2.2 Self-determination theory
119,7 SDT was developed from the cognitive evaluation theory (Ryan and Deci, 2000a) and was
used to explore people’s volitional motivation, particularly the degree to which
people experience their actions as autonomous (e.g. participation in donation
crowdfunding campaigns). Over the past few decades, SDT has gained wide popularity
in different contexts (Han et al., 2017; Evans, 2015). However, while studies linking SDT to
1518 donation crowdfunding are surprisingly poor, we therefore apply SDT for a more suitable
conceptualization of motivation in the donation crowdfunding context for two reasons.
First, previous studies have found that individuals’ motivation to participate in donation
crowdfunding is motivated by internal factors, which indicates the necessity of measuring
the level of individuals’ self-determination. However, studies have noted the major
limitation of technology acceptance models such as TAM, TPB, UTAUT, etc., namely, a
lack of attention to the self-determined motivations (Venkatesh et al., 2003). SDT is one of
the most detailed frameworks explaining the links between self-determined motivations
and positive outcomes (Ryan and Deci, 2000b). Second, intrinsic and extrinsic motivating
factors have typically been investigated in isolation from each other in previous studies.
We argue that different motivational factors are likely to harmonize by having a mutually
reinforcing effect on individuals’ donation intentions (Ryan and Deci, 2000a). We believe
that examining the consequences of diverse motivations in a unique framework will offer
more insights into the utilization of donation crowdfunding.
In line with SDT, there are different types of motivation, including intrinsic and extrinsic
motivation, that differ in their degree of self-determination. Intrinsic motivation is derived
from the inherent satisfaction that an individual experiences from a behavior, that is, one is
enjoying an activity for its own sake (Marshall et al., 2017). Furthermore, with respect to
extrinsic motivation, SDT has identified four different types of extrinsic motivation that
vary the extent to which behavioral regulations are internalized (Ryan and Deci, 2000b; Liu
et al., 2019), namely, external, introjected, identified and integrated regulation.
External regulation is the least autonomous form of motivation; it underlies behaviors
that are performed in order to obtain a reward or avoid a punishment; introjected regulation
inspires an individual to enact a behavior because he or she fears a negative effect, worry or
shame; identified regulation regulates behaviors based on the value ascribed to the outcome,
not the activity itself; and integrated regulation is a form of motivation that arises when a
person feels that a behavior is integrated within the self and is in harmony and coherence
with other aspects of the self. Among these, donors of crowdfunding platforms are
unlikely to receive a reward or a punishment (i.e. external regulation). Only intrinsic,
introjected, identified and integrated regulation are relevant to donation behaviors on equity
crowdfunding platforms. The representative construct for each of the dimensions is selected
as suggested in the literature, and they were adjusted to adapt the crowdfunding context
involving the sense of self-worth (i.e. intrinsic regulation), face concern (i.e. introjected
regulation), donor effectiveness (i.e. identified regulation) and moral obligation
(i.e. integrated regulation).
Scholars have demonstrated that, apart from intrinsic and extrinsic motivations,
contextual support can also promote well-being and optimal motivation by providing
chances for individuals to fulfill their basic needs. On the contrary, a lack of contextual
support undermines individuals’ motivation and well-being (Ryan and Deci, 2000b).
There are mainly two contextual environments for a participant in donation crowdfunding,
namely, the community in which individuals interact with others and the donation
crowdfunding campaign where individuals make a donation. Accordingly, we identified
two types of contextual factors that reflect the influences of social interaction
(e.g. communication in crowdfunding communities) and referent network size (e.g. donor
size of a donation crowdfunding platform).
2.3 Self-identity and social identity Exploring
Identity-related studies uncover two main levels of analysis: self-identity (role identity) and individuals’
social identity (one’s identity as part of a social group). Self-theory and social identity are behavioral
two perspectives on the social basis of the self-concept and on the nature of normative
behavior. The two perspectives occupy parallel but separate universes, with virtually no intentions
cross-referencing (Whitmarsh and O’Neill, 2010). Both perspectives address the structure
and function of the socially constructed self as a dynamic construct that mediates the 1519
relationship between social structure or society and individual social behavior. Self-identity
is principally a micro sociological concept that sets out to explain individuals’ role-related
behaviors, while social identity is a social psychological concept that sets out to explain
group processes and intergroup relations (Huang, 2012). Hogg et al. (1995) suggested that
there is an important difference between self-identity and social identity, that is, self-identity
focuses on role identities, such as social roles (e.g. mother, father) and social types
(e.g. athlete, actor), whereas social identity focuses on social groups (e.g. nationality,
delegation) into which one falls, and to which one feels one belongs (Huang, 2012;
Whitmarsh and O’Neill, 2010).
Self-identity and social identity are considered subjective substitutes to social rules so
that a more relaxed relationship atmosphere can be created (Thorbjørnsen et al., 2007;
Dermody et al., 2017). They are important in shaping reliable and socially accepted
behavior within a situation where there is an absence of workable rules ( Jebarajakirthy
and Thaichon, 2016; Jiang et al., 2016). Both of the two perspectives broaden the notion of
social context by acknowledging the influence of societal roles (i.e. self-identity) and
group memberships (i.e. social identity) upon an individual’s social behavior. In a
donation crowdfunding context, self-identity and social identity can be used to impel
potential donors to comply with certain beliefs without having mandatory measures.
Therefore, we argue that self-identity and social identity are compatible with the SDT
framework and the adoption of self-identity and social identity is appropriate for the
present research. In the IS literature, self-identity and social identity were widely
used to examine factors affecting users’ behavioral intention toward social applications,
such as SNSs ( Jiang et al., 2016) and virtual communities (Cheng and Guo, 2015), etc.
Nevertheless, there is no empirically validated research on the effectiveness of
self-identity and social identity in the context of donation crowdfunding, and the
characteristics of individuals’ self-identity and social identity remain unclear.
Consequently, we introduce self-identity and social identity to examine their effects
against a donation crowdfunding background.

3. Research model and hypotheses

In this study, we applied the SDT to propose the research model that incorporates
contextual variables as determinants and self-identity and social identity as mediating
constructs to predict individuals’ intentions toward donation crowdfunding. On the basis of
SDT, individuals’ behaviors are driven by the combination of different types of motivation.
The proposed research model is presented in Figure 1.

3.1 Identity
Self-identity is defined as a cognitive construct of the self that answers the question “who
am I?” It focuses on individual characteristics that separate one from others (Sirgy, 1982).
Individuals are more willing to conduct behavior with a high level of role identity because
they always make expectations on the basis of role-appropriate behavior and then perform
suitable behavior to satisfy these expectations (Gerber et al., 2012), which can be used to
confirm their self-concept. Previous research has indicated that self-identity explains a
significant amount of variance in behavioral intention (Barbarossa et al., 2017). According to
IMDS SDT framework
119,7 Intrinsic regulation
Sense of self-worth
H3
Introjected regulation Self-identity
H4
1520 Face concern
H6 H1
Identified regulation
Perceived donor H7 Donation intention
effectiveness H5
H9
Integrated regulation H11 H2
Moral obligation
H8
Social identity

Contextual influence H10

H12
Social interaction
Figure 1.
Research model Referent network size

identity theory, potential donors usually categorize themselves in specific social roles, and
then these roles guide their intentions and behaviors, so that they spontaneously act in
accordance with their self-identities ( Jiang et al., 2016). Considering that donation
crowdfunding is charitable giving of a kind, those who consider themselves as concerned
with charitable causes will have greater intentions to donate. Therefore, we propose the
following hypothesis:
H1. Self-identity is positively related to donation intention.
Tajfel (1972) defines social identity as “one’s knowledge that one belongs to certain social
groups together with some emotional and value significance to one as this group
membership.” People in the same social group are more likely to behave in line with their in-
group members and to distinguish themselves from out-group members ( Jiang et al., 2016).
Social identity theory assumes that if a particular social identity is salient, then the
individual will tend to be in accordance with the values and beliefs of the group
(Chatzisarantis et al., 2009), that is, socially identifying with a group leads one to behave in
the way that one believes that members of that group should behave (Edwards, 2005). Based
on the above reason, potential donors who perceive that the engagement in donation
crowdfunding is recognized by the majority of social members will behave more consistent
with their in-group members and show greater intention ( Jiang et al., 2016). This leads to the
following hypothesis:
H2. Social identity is positively related to donation intention.

3.2 SDT framework

3.2.1 Intrinsic regulation. A previous study reported that the sense of self-worth is a typical
intrinsic motivation (Evans, 2015). It refers to the degree of positive cognition based on one’s
feeling of personal contribution to others based on donation behavior (Bock et al., 2005).
Applied to the donation crowdfunding context, the sense of one’s value is defined as the
degree to which an individual believes that his or her donation activities via crowdfunding
platforms will yield feelings of intrinsic happiness, accomplishment and personal value Exploring
(e.g. “I feel proud of myself in general”). According to the cognitive theory, the meaning of individuals’
roles is socially formed in the social processes such as experience and observation behavioral
(Tabernero and Hernández, 2011). Therefore, enjoyable experiences gained by realizing self-
value are critical parts of understanding ourselves and defining ourselves. Therefore, if an intentions
individual gains inner happiness or satisfaction when donating on crowdfunding platforms,
the self-concept will be confirmed during the process. Accordingly, the following hypothesis 1521
is addressed:
H3. Sense of self-worth is positively related to self-identity.
3.2.2 Introjected regulation. Face concern describes one’s concern about maintaining or
promoting one’s social status through playing particular social roles (Hall and Bucholtz,
2013). Previous studies have indicated that the Chinese are strongly motivated to enhance or
protect their face by showing off the good side while hiding their misconduct (Bradford,
2011). The action of donation is generally considered to be a good and ethical behavior
(an action with social approval) in our society, which is believed to be effective for those who
extremely care about their face in achieving the identity they desire. Thus, individuals
laying emphasis on their face are more likely to be concerned about charitable causes, for
which contribute them to maintaining a positive self-evaluation and forming a self-identity
(Whitmarsh and O’Neill, 2010). In addition, people have a natural tendency to define and
locate themselves in the society by classifying themselves into certain social groups
(Hall and Bucholtz, 2013). Hence, individuals with a high level of face concern are more likely
to foster the identification with charity groups and distinguish themselves from those
out-group members, in particular, those who do not participate in charitable causes
( Jiang et al., 2016). Therefore, this paper proposes the following hypotheses:
H4. Face concern is positively related to (a) self-identity/(b) social identity.
3.2.3 Identified regulation. In the present study, we adopt perceived donor effectiveness,
which was originally called perceived consumer effectiveness, to describe the extent to
which one believes that one’s efforts can make a significant difference and contribute
toward solving social or others’ problems (Meijboom and Brom, 2012). According to a
study on consumer effectiveness, individuals must perceive that their own actions will
have a real and observable outcome and will result in a difference-making solution to a
problem (Cojuharenco et al., 2016). However, charitable donations do not always lead to a
desired result, such as a fraudulent charity fundraising. In addition to frauds, some
donation crowdfunding platforms routinely spend many of donors’ funds on
administrative and fundraising costs rather than on programs that benefit people in
need. Hence, a donation activity without an observable effect will not only incur a sense of
financial loss, but will also discourage one from willingness to help others in future.
Instead, if an individual is convinced that the behavior will bring about the desired
outcome, it helps confirm one’s social roles as a philanthropist and will foster his or her
self-identity (Meijboom and Brom, 2012). For this reason, donor effectiveness has
been shown to be necessary to enable consumers to confirm their self-identities. Hence,
this work proposes the following hypothesis:
H5. Perceived donor effectiveness is positively related to self-identity.
3.2.4 Integrated regulation. Moral obligation is described as the extent to which an
individual feels a sense of responsibility to act (or not) morally (or immorally) when faced
with an ethical situation. Hogg et al. (1995) argued that people perceive their identities by
projecting themselves into a social context. That is, people with a high moral obligation may
IMDS have more readily available moral schemas regarding their daily behavior. Becoming
119,7 involved in donation crowdfunding may exactly meet their own moral principles, which will
in turn contribute to their perceptions of self-identity as a specific role (Wilska et al., 2002).
Meanwhile, a stronger sense of moral obligation can also serve as a catalyst to encourage
people to perform ethical behavior (e.g. be ready to help others, conserve water, reduce
pollution, etc.), which is a process of identifying as a specific social group member
1522 (e.g. philanthropist, environmentalist, etc.) and emotionally investing in one’s social group
memberships (Meijboom and Brom, 2012). People may feel connected to a group or a
community with a common purpose, topic or interest, and then enable them to foster their
social identity. Accordingly, the following hypotheses are proposed:
H6. Moral obligation is positively related to self/social identity.

3.3 Contextual influence

Social interaction refers to a process in which fundraisers and potential donors share ideas
with each other in donation crowdfunding communities or social network sites (Cojuharenco
et al., 2016). Donation crowdfunding participants usually form various communities based
on common interests and social interactions. Research has suggested that identity forms
during the course of social interactions that take place within a social structure (Dukerich
and Ashforth, 2000). Participants within communities exchange not only ideas but also
feelings, sympathy and encouragement (Ewe et al., 2015; Wong et al., 2018). Postmes et al.
(2005) argued that social interactions may play an important role in the emergence of a
shared cognition, such as a shared task and goal, in groups, and may then contribute to the
emergence of social identity. Thus, social interactions with others in crowdfunding
communities play a key role in the development of one’s sense of social identity, as these
social interactions allow one to perform social categorizations and to learn with which social
groups one wants to be associated with. Meanwhile, a few studies highlight that self-identity
can also be perceived through interaction and communication between group members
(Cojuharenco et al., 2016). Through frequent social interactions, members may perceive the
roles that they will play in groups. Therefore, social interactions serve as the indicator of a
role and contribute to members’ perception of their self-identity as a donor. Thus, the
following hypotheses are presented:
H7. Social interaction is positively related to (a) self-identity/(b) social identity.
Referent network size is defined as an individual’s perception of the current donor size of a
donation crowdfunding platform (Schlagwein and Choy, 2015). The size of a social group
has been viewed as a determinant of social identification (Giosan et al., 2001). An
increasing number of users who donate on crowdfunding platforms strengthen the
attraction for non-users to confirm their identity as a member of a certain social group.
In addition, a larger referent network size increases the diversity of people who one is
exposed to. Within the given situation, an individual will enact his or her identity based on
the contexts, especially when there are many valued relationships within the current
network (e.g. family members, friends or colleagues, etc.). Individuals may gain clues on
how to construct a role that is deemed to be acceptable, that is, if there are many people
focusing on charitable causes in the social environment, an individual may foster a
personal identity that conforms to these expectations and behavioral norms. Thus, the
larger the referent network size that an individual has based on to enact his or her
charitable acts, the more salient self-identities with the charitable causes will be fostered.
For these reasons, we proposed the following hypotheses:
H8. Referent network size is positively related to (a) self-identity/(b) social identity.
4. Research design and methodology Exploring
4.1 Measurement development individuals’
All the items included in Table I were compiled and adapted from validated instruments used behavioral
in the previous literature, with minor modifications in wording to fit the donation crowdfunding
context. We rated all the measurement items using a five-point Likert scale, ranging from 1
intentions
“strongly disagree” to 5 “strongly agree.” Since the questionnaire was first drafted in
English, we used the “back-translation method” to check the accuracy of the questionnaire. 1523

Construct Items References

Sense of self- Donating on crowdfunding platforms will give me a feeling Bock et al. (2005)
worth of happiness
Donating on crowdfunding platforms will give me a sense of
accomplishment
Donating on crowdfunding platforms will realize my personal value.
Face concern I do not want others to say I am stingy Monkhouse et al.
I pay considerable attention to how others see me (2012)
I do not want the people around me to feel I am indifferent
I am concerned with not bringing shame to myself
Perceived donor When I donate money, I try to consider how my donation behavior Cojuharenco et al.
effectiveness will affect the fundraiser’s crowdfunding (2016)
It is worthless for the individual donor to do anything for the
fundraiser.(R)
Since one person cannot have any effect upon the fundraiser’s
crowdfunding, it does not make any difference what I do.(R)
Each donor’s behavior can have a positive effect on the fundraiser’s
life by donating to his/her crowdfunding
Moral obligation I would feel guilty if I do not help others Beck and Ajzen
Not helping others goes again my principles (1991)
It would be morally wrong for me not to help others
Social interaction I would like to exchange and share opinions with the fundraisers Cojuharenco et al.
or other people (2016)
I would like to meet other people who regularly visit the
communities
I do quite a bit of socializing in the communities
Referent network Many people are engaging in this donation crowdfunding Lin and Lu (2011)
size Many famous people who I know are engaging in this
donation crowdfunding
Many relatives or friends around me are engaging in this
donation crowdfunding
Self-identity I think of myself as a donation crowdfunding donor Whitmarsh and
I think of myself as someone who is concerned about O’Neill (2010)
charitable causes
I think of myself as someone who is concerned about others
Social identity I see myself as a member of a social group that is concerned about Huang (2012)
charitable causes
I often feel like I am a worthy member of a social group that is
concerned about charitable causes
I have a lot in common with the other members of the social group
that is concerned about charitable causes
The social group I belong to is an important part of who I am
Donation intention Assuming I have access to the donation crowdfunding platform, I Ewe et al. (2015)
intend to participate in it
I intend to participate in donation crowdfunding in the future
I would use the donation crowdfunding platform to help others Table I.
Participating in donation crowdfunding is something I would do Measurement items
IMDS Then, a pre-test and a pilot test were performed, respectively, before the formal survey to
119,7 ensure the reliability and validity of the scales. The pre-test was conducted using fourteen
individuals who have experience in donation crowdfunding. The participants were asked to
assess the logical consistency, contextual relevance and terminology, and question clarity.
Based on their feedback, the questionnaire items were thoroughly refined. Then a pilot test was
carried out to examine test–retest reliability and questionnaire reliability with 69 users who had
1524 some knowledge of donation crowdfunding. To measure test–retest reliability, email invitations
containing the same version of the questionnaires were sent to these participants over a
one-week interval. We finally collected 58 valid responses, yielding a valid response rate of
84.1 percent. The pilot data analysis showed satisfactory test–retest reliability and construct
reliability for the questionnaire items.

4.2 Survey procedure

This research used China as the site of the empirical investigation for the reason that the
supporting context required for the growth of donation crowdfunding had already been
set in place. As one of the most widely accepted financial innovations in the last few years,
crowdfunding in China has gained remarkable achievements. Moreover, China is now
recognized as a “rising star” in the global crowdfunding market and has attracted
worldwide attention (Ryu, 2018). Meanwhile, the World Bank predicted that crowdfunding
industry will reach a size of $50bn in 2025, in which, donation crowdfunding, with a
brightest future, will account for one-fifth. Furthermore, according to the Tencent Charity
report, the population engaged in donation crowdfunding in China has reached up to
8.46bn, funding more than 3.17bn yuan, enjoying an increase of 26.8 percent compared
with 2017. The favorable environment provides a solid foundation for the investigation of
the dynamics of donation crowdfunding in China.
A web-based survey questionnaire was first posted on “wjx.cn,” a popular online
survey website in China. Then, online forums (including WeChat, Baidu tieba, QQ
groups, virtual communities, etc.) devoted to donation crowdfunding were randomly
identified via search engines (Baidu and Bing), and a hyperlink connecting to the survey
page was placed on these selected forums. This approach is effective because donation
crowdfunding donors usually discuss campaign-related issues online. Before the
participants answered the questionnaire, they were given a description of donation
crowdfunding services. Each user (including browsers and donors) was asked to provide
the name of a donation crowdfunding platform that he or she used recently as the
representative site that he or she selected for answering the questionnaire. To ensure the
quality of the responses, a preventive measure was also used to avoid multiple responses
from the same person by rejecting any two submissions with the same IP address. To
increase the response rate, a total of 300 electronic red-envelopes (CNY 10 per envelope)
were given as incentives to randomly selected respondents. The survey was conducted
from June 23, 2019 to July 12, 2019. A total of 632 completed online questionnaires were
returned. Questionnaires completed in less than two minutes were discarded because
pilot test results showed that at least two minutes were needed to complete this
questionnaire. After eliminating questionnaires that were considered invalid, the final
sample consisted of 588 respondents. Independent t-tests were conducted to examine the
difference between potential donors and actual donors on the nine constructs.
No significant differences were found among the data from different groups. The test
result indicates that there is no need to perform a separate analysis on potential donors
and actual donors (Table II).
Detailed descriptive demographics of the respondents are shown in Table III. The valid
sample included 57.1 percent male and 42.9 percent female respondents. More than
80 percent of the participants had previously made a donation through donation
 Potential donors Actual donors Independent t-test
Exploring
Construct Mean SD Mean SD t-value Significance individuals’
behavioral
SS 3.05 0.859 3.07 0.833 0.844 0.082
FC 3.16 0.918 3.19 0.935 0.232 0.623 intentions
PE 3.07 0.868 3.11 0.829 0.355 0.829
MO 3.12 0.827 3.14 0.839 0.108 0.110
SI 3.01 0.901 3.03 0.747 −0.496 0.465 1525
NS 3.21 0.794 3.18 0.812 −1.126 0.205
SEI 3.10 0.812 3.12 0.780 −2.204 0.520
SOI 3.11 0.858 3.13 0.836 0.442 0.606 Table II.
DI 3.49 0.766 3.47 0.785 0.036 0.415 Descriptive statistics

Potential donors Actual donors Potential donors Actual donors

Category (n ¼ 106) (n ¼ 482) Category (n ¼ 106) (n ¼ 482)

Age Occupation
18 or below 3 (2.8%) 15 (3.1%) Student 27 (25.5%) 108 (22.4%)
18-25 26 (24.5%) 132 (27.4%) Employed 49 (46.2%) 249 (51.7%)
26-30 34 (32.1%) 168 (34.9%) Unemployed 24 (22.6%) 95 (19.7%)
31-40 23 (21.7%) 97 (20.1%) Other 6 (5.7%) 30 (6.2%)
41-50 15 (14.2%) 62 (12.9%) Length of Internet experience
50 or older 5 (4.7%) 8 (1.6%) Less than 1 year 1 (0.9%) 7 (1.5%)
Gender 1–2 years 2 (1.9%) 9 (1.9%)
Male 63 (59.4%) 273 (56.6%) 3–5 years 26 (24.5%) 97 (20.1%)
Female 43 (40.6%) 209 (43.4%) 6–10 years 27 (25.5%) 134 (27.8%)
Education level More than 10 50 (47.2%) 235(48.7%)
years
High school or 21 (19.8%) 89 (18.5%) Samples have used donation crowdfunding platforms
below (multiple responses)
College/University 74 (69.8%) 315 (65.3%)
Graduate school 11 (10.4%) 78 (16.2%) Tencent charity 21 (22.1%) 228 (21.5%)
Monthly income (CN Yuan) Qingsongchou 17 (17.9%) 186 (17.5%)
Less than 3,000 29 (27.4%) 125 (25.9%) Shuidichou 13 (13.7%) 161 (15.2%)
3,001–6,000 29 (27.4%) 143 (29.7%) Zhongchou 9 (9.5%) 125 (11.8%)
6,001–10,000 28 (26.4%) 124 (25.7%) Aixinchou 11 (11.6%) 109 (10.3%)
10,001–30,000 16(15.1%) 65 (13.5%) Wuyouchou 12 (12.6%) 101 (9.5%)
More than 30,001 4 (3.7%) 25 (5.2%) Other platforms 12 (12.6%) 151 (14.2%) Table III.
Note: n ¼ 588 Sample demographics

crowdfunding platforms; this suggested that the information that we collected represented
the views of “donors” rather than “browsers.” The most frequently used donation
crowdfunding platforms are Tencent charity, Qingsongchou, Shuidichou and Zhongchou,
respectively. Respondents were well educated overall (81.3 percent of respondents had a
college/university degree or above) and rather young (81.6 percent of the respondents were
18–40 years of age). Although the data collected in the present study have potential selection
bias, it is impossible to select a random sample of users because a complete directory of
crowdfunding platforms that allow online donation does not exist. Even though we mainly
surveyed well-educated young adults, they without a doubt are the main donors on donation
crowdfunding platforms. According to the China Crowdfunding White Paper, more than
IMDS 60 percent of crowdfunding users in China are between 18 and 30 years of age, and more
119,7 than 60 percent of crowdfunding users have undergraduate education or above. Hence, we
believe that the responses seem to be appropriate to the present study.

5. Data analysis and results

5.1 Measurement model validation
1526 Confirmatory factor analysis (CFA) was conducted to evaluate the measurement model,
construct validity and reliability. The overall model fit index for the initial measurement
model indicated that the fit was acceptable, with χ2/df ¼ 1.927, RMSEA ¼ 0.045,
GFI ¼ 0.925, AGFI ¼ 0.912, NFI ¼ 0.911, IFI ¼ 0.952 and CFI ¼ 0.948 all meeting the
required standards. However, in some cases, overall good-of-fit indices indicated an
acceptable fit despite the fact that some relationships among the indicators in the data have
not been sufficiently reproduced (Kline, 2016). Therefore, we adopted modification indices to
identify focal areas of misfits in a CFA solution. Modification indices indicate a decrease in
χ2 when a parameter that had been constrained is relaxed. After examining the modification
indices, the initial measurement model was respecified with a full consideration of the
theoretical rationale (Kline, 2016). Finally, one item of “social interaction” (SI2 with high
modification indices ranging from 8.107 to 13.852) and one item of “self-identity” (SEI3 with
high modification indices ranging from 7.082 to 14.387) that shares a high degree of error
covariances with other items (SI2↔NS1, SEI3↔PE2 and SEI3↔MO3) and variables
(SI2↔Social identity and SEI3↔Face concern) were removed. The fit indices for the revised
measurement model indicated a better model fit than those obtained from the initial one
( χ2/df ¼ 1.867, RMSEA ¼ 0.036, GFI ¼ 0.937, AGFI ¼ 0.921, NFI ¼ 0.931, IFI ¼ 0.971 and
CFI ¼ 0.963). Composite reliability (CR), the average variance extracted (AVE) and
Cronbach’s α were used to measure internal consistency (Fornell and Larcker, 1981). For a
construct that has good validity and reliability, the recommended value for CR should be at
least 0.6, the AVE should exceed 0.5 and Cronbach’s α should be greater than 0.7 (Hair et al.,
1998). From Table IV, we can see that all values surpass the generally accepted values,
indicating good reliability.
We employed construct validity and content validity as the validity measurement. The
variables in the research were derived from the existing literature, thus showing high
content validity. Convergent validity and discriminant validity were used to examine
construct validity. We adopted principal components analysis to verify the convergent
validity of each variable. A measurement item has high validity if its cross-loading
coefficient is below 0.4 and its loading coefficient is above 0.6 (Fornell and Larcker, 1981). On
the basis of the criteria, factor loadings for all the items surpass the recommended value of
0.6 and are significant at p o0.001. Meanwhile, no items have cross-loadings above 0.4.
Therefore, the whole model has adequate convergent validity.
Discriminant validity was verified using the index recommended by Fornell and Larcker
(Fornell and Larcker, 1981): the square root of AVE extracted from each variable should be
larger than the correlations between the variable and the other variables. As shown in Table V,
each variable has a higher loading than its cross-loadings on other variables, thus providing
evidence of discriminant validity. Therefore, this research model illustrates adequate reliability,
convergent validity and discriminant validity.

5.2 Test of structural model for all respondents

To evaluate the representation of the model, we used AMOS 6.0 to evaluate the “goodness of
fit” indices. As indicated in Table VI, χ2/df ¼ 1.914, GFI ¼ 0.928, AGFI ¼ 0.913, NFI ¼ 0.919,
IFI ¼ 0.960, CFI ¼ 0.959 and RMSEA ¼ 0.039 are all within the widely accepted thresholds
proposed in the literature (Fornell and Larcker, 1981; Hair et al., 1998). The fit indices
demonstrate that the model provides an adequately good fit to the data.
Constructs Items Mean SD Factor loading CR AVE Cronbach’s α
Exploring
individuals’
Sense of self-worth SS1
SS2
3.09
3.08
0.873
0.871
0.899
0.875
0.890 0.731 0.831 behavioral
SS3 3.06 0.815 0.786 intentions
Face concern FC1 3.22 0.929 0.843 0.872 0.630 0.806
FC2 3.18 0.903 0.793
FC3 3.17 0.923 0.784 1527
FC4 3.14 0.942 0.751
Perceived donor effectiveness PE1 3.15 0.781 0.889 0.911 0.719 0.844
PE2 3.15 0.925 0.878
PE3 3.05 0.840 0.834
PE4 3.04 0.840 0.787
Social interaction SI1 3.07 0.818 0.868 0.886 0.721 0.831
SI4 3.00 0.902 0.859
SI3 3.03 0.957 0.820
Moral obligation MO1 3.20 0.830 0.889 0.892 0.735 0.840
MO2 3.09 0.793 0.876
MO3 3.06 0.890 0.804
Network size NS1 3.16 0.890 0.792 0.810 0.587 0.796
NS2 3.19 0.802 0.778
NS3 3.22 0.730 0.727
Self-identity SEI1 3.08 0.693 0.836 0.843 0.642 0.796
SEI2 3.14 0.762 0.791
SEI4 3.15 0.887 0.776
Social identity SOI1 3.06 0.811 0.784 0.835 0.559 0.797
SOI4 3.10 0.847 0.755
SOI3 3.17 0.887 0.743
SOI2 3.18 0.835 0.706
Donation intention DI1 3.47 0.761 0.843 0.893 0.676 0.859 Table IV.
DI2 3.42 0.775 0.840 Construct reliability
DI3 3.52 0.788 0.835 and convergent
DI4 3.54 0.784 0.768 validity

SI FC PE MO NS SS SEI SOI DI

SI 0.865
FC 0.129** 0.793
PE 0.138** 0.092* 0.848
MO 0.059 0.114** 0.148** 0.857
NS 0.266** 0.341** 0.126** 0.266** 0.766
SS 0.086** 0.131** 0.089* 0.086* 0.045* 0.855 Table V.
SEI 0.156** 0.392** 0.214** 0.156** 0.392** 0.173** 0.801 Discriminant validity:
SOI 0.267** 0.451** 0.167** 0.267** 0.451** 0.217** 0.264** 0.748 the square roots of
DI 0.107** 0.305** 0.140** 0.107** 0.142** 0.124** 0.393** 0.338** 0.822 the AVEs and factor
Notes: *p o0.05; **p o 0.01 correlation coefficients

Fit index Observed value Recommended value References

χ2/df 1.914 Good fit (should be less than 3) Fornell and Larcker (1981)
GFI 0.928 Good fit (should be greater than 0.90) Hair et al. (1998)
AGFI 0.913 Good fit (should be greater than 0.90) Hair et al. (1998)
NFI 0.919 Good fit (should be greater than 0.90) Fornell and Larcker (1981)
IFI 0.960 Good fit (should be greater than 0.90) Hair et al. (1998)
CFI 0.959 Good fit (should be greater than 0.90) Fornell and Larcker (1981) Table VI.
RMSEA 0.039 Good fit (should be less than 0.06) Hair et al. (1998) Model fit indices
IMDS We represent the standardized path coefficients of the empirical model in Table VII. Most of
119,7 the paths are significant in the expected directions. Exception is a path connecting moral
obligation with self-identity, providing no support for H6. The path coefficients of H4 are
significant at the level of p o0.05, thus providing support for H4. The path coefficients of
H4 and H8 are significant at the level of p o0.01, thus providing support for these two
hypotheses. The path coefficients of H1, H2, H3, H5, H6, H7 and H8 are significant at
1528 the level of p o0.001, providing support for these hypotheses. Figure 2 shows a summary
of the results for each path in the research model.

6. Discussion and implications

The aim of the present study is to explore the factors affecting individuals’ behavioral
intentions toward donation crowdfunding based on the SDT and identity theory. Hypothesis

Hypothesized path Coefficient SE t-value p-value

Self-identity→Donation intentions 0.429 0.058 7.393 0.000***

Social identity→Donation intentions 0.403 0.070 5.788 0.000***
Sense of self-worth→Self-identity 0.206 0.053 3.863 0.000***
Face concern→Self-identity 0.211 0.040 5.305 0.04*
Face concern→Social identity 0.198 0.028 7.062 0.002**
Perceived donor effectiveness→Self-identity 0.132 0.035 3.739 0.000***
Moral obligation→Self-identity 0.031 0.036 0.854 0.393
Moral obligation→Social identity 0.095 0.023 4.171 0.000***
Social interaction→Self-identity 0.174 0.039 6.187 0.000***
Social interaction→Social identity 0.088 0.024 3.651 0.000***
Table VII. Referent network size→Self-identity 0.396 0.054 7.312 0.000***
Hypothesis testing Referent network size→Social identity 0.256 0.038 6.687 0.003**
for all respondents Notes: n ¼ 588. *p o0.05; **p o0.01; ***p o0.001

SDT framework Significant path ( p < 0.05)

Intrinsic regulation Non-significant path ( p.0.05)

Sense of self-worth
0.206***
Introjected regulation
Self-identity
0.211*
Face concern
0.132*** 0.429***
Identified regulation
Perceived donor 0.031 Donation intention
effectiveness 0.198**
0.174***
Integrated regulation 0.403***
0.396***
Moral obligation
0.095***
Social identity

Contextual influence 0.088***

0.256**
Social interaction
Figure 2.
Test of a structural Referent network size
model for all
respondents
Notes: n = 588. *p < 0.05; **p < 0.01; ***p < 0.001
testing result shows that 11 of the 12 hypotheses are statistically supported. As expected, self- Exploring
identity (βself-donation ¼ 0.429, t ¼ 7.393, po0.001) and social identity (βsocial-donation ¼ 0.403, individuals’
t ¼ 5.788, po0.001) are found to significantly influence donation intentions, which are in line behavioral
with the results of studies conducted by Barbarossa et al. (2017) and Jiang et al. (2016).
These findings indicate that, if users perceive that performing donation behavior meets their intentions
self-identity and social identity, it could boost their intentions to donate.
Sense of self-worth (βworth-self ¼ 0.206, t ¼ 3.863, p o0.001) is positively associated with 1529
social identity. This result is consistent with those of previous studies in the IS literature
that tested the relationship between sense of self-worth and role identity (Majchrzak et al.,
2005). This finding indicates that sense of self-value is critical in knowing and defining the
social roles of ourselves.
Consistent with the findings obtained from previous research (Whitmarsh and O’Neill,
2010; Huang, 2012), the present study validated the prediction that the face concern exerts
significant influences on self-identity (βface-self ¼ 0.211, t ¼ 5.305, p o0.05) and social
identity (βface-social ¼ 0.198, t ¼ 7.062, p o0.01). Thus, for those who pay close attention to
their face, charitable causes, such as donation crowdfunding, can exactly provide them with
an opportunity to enhance their positive evaluation right in front of others. Moreover,
donation behavior also facilitates them to identify those groups with higher social status.
Perceived donor effectiveness (βdonor-self ¼ 0.132, t ¼ 3.739, p o0.001) has positive effects
on self-identity, indicating that it is a critical factor in the formation of self-identity. This
result is in line with the finding of Dermody et al. (2017)’s research on customer
effectiveness, in which it was argued that if an individual believes that his/her behavior will
have an effect on the desired outcome, it can obviously foster his/her self-identity.
Moral obligation is found to have a positive effect on social identity (βmoral-social ¼ 0.095,
t ¼ 4.171, p o0.001). This finding is in agreement with the results of research conducted by
Dwyer et al., 2011, who argued that a high sense of moral obligation may be considered a
catalyst to stimulate people to perform ethical behavior, which in turn fosters their social-
identities with the social groups. In contrast to previous studies (Hogg et al., 1995), the
proposed positive relationship between moral obligation and social identity is not confirmed
in this study (βmoral-self ¼ 0.031, t ¼ 0.854, p ¼ 0.393). A possible explanation may be that
although helping behavior tendencies among members of the Chinese society are common,
nevertheless, “Peng-Yu” case (a well-known social event that occurred in China in 2006)
surprisingly discourages the tendency in China and turns it around. In the case, Peng Yu
has rekindled public debate on morality crisis, and the final judgment evoked strong
reactions among the public which still has a profound impact today. While some Chinese
think themselves as kind, they will refuse to live up to their moral obligations owing to the
possibility of taking unnecessary risks.
Consistent with previous research showing a link between social interaction and social
identity (Postmes et al., 2005), social interaction is found to positively influence self-identity
(βinteract-self ¼ 0.174, t ¼ 6.187, po0.001) and social identity (βinteract-social ¼ 0.088, t ¼ 3.651,
po0.001). Considering that social interaction is popular among donation crowdfunding
participants, this finding provides support for the notion that frequent interaction may boost
the formation of social identity and self-identity as a donor.
Finally, referent network size was verified to have positive influences on self-identity
(βnet-self ¼ 0.396, t ¼ 7.312, p o0.001) and social identity (βnet-social ¼ 0.256, t ¼ 6.687,
p o0.001). These results are consistent with that of the study conducted by Huang (2012).
This finding is important because previous studies have always adopted network size as an
antecedent of attitude, neglecting its influence on individuals’ societal roles (self-identity)
and group memberships (social identity). The results demonstrate that large network size
can strengthen the attraction for individuals to confirm their social-identities and form their
self-identities with the charitable causes.
IMDS 6.1 Theoretical implications
119,7 This study makes several contributions to the research literature. First, although behavioral
intention is an important research issue and has been extensively studied in the IS literature
(Venkatesh et al., 2003), this is a topic that has not been studied in the context of donation
crowdfunding in the past. The present study fills this knowledge gap. To our knowledge,
this is one of the first studies to contribute to an understanding of individuals’ donation
1530 intention in the Chinese donation crowdfunding market. Our research model accounts for
74 percent of the variance in behavioral intentions, a substantial improvement over
traditional technology acceptance models (e.g. TAM, TPB, etc.). By providing deep insights
into the antecedents of individuals’ donation intentions, this research greatly contributes to
the emerging literature on donation crowdfunding, and may also serve as a theoretical
foundation to investigate the acceptance of other innovations.
Second, apart from personal motivation, Ryan and Deci (2000a, b) have asserted that
contextual support for the basic psychological needs enhances motivation. Accordingly,
another significant implication of the present study is that it deepens and broadens extant
knowledge on SDT by providing a comprehensive explanation of contextual influence,
providing additional insights into motivations as conceptualized by SDT. This study
further emphasizes the irreplaceable role of contextual factors in terms of social interaction
(βsocial-donation ¼ 0.110) and referent network size (βnet-donation ¼ 0.273) during the process of
individuals’ behavioral decision.
Third, the findings of the present study highlight the vital role played by the
self-identity (βself-donation ¼ 0.429) and social identify (βsocial-donation ¼ 0.403) in donation
decision-making, thus deepening the research on identity in the context of donation
crowdfunding. This finding is important because previous studies have always discussed
self-identity and social identity separately in many other areas; however, to our
knowledge, their combination has not been investigated in the context of donation
crowdfunding. The present study is one of the first to address the combined function of
self-identity and social identity and the conclusions show that they fit well against the
donation crowdfunding background.
Fourth, present study investigated the formation of self-identity and social identity in the
donation crowdfunding context, contributing to both the crowdfunding literature and identity
formation research. Plenty of factors, such as the sense of self-worth (βworth-self ¼ 0.206),
face concern (βface-self ¼ 0.211, βface-social ¼ 0.198), moral obligation (βmoral-social ¼ 0.095),
perceived donor effectiveness (βdonor-self ¼ 0.132), social interaction (βinteract-self ¼ 0.174,
βinteract-social ¼ 0.088) and referent network size (βnet-self ¼ 0.396, βnet-social ¼ 0.256) can play
prominent roles in the formation of self-identity and social identity. However, no study has ever
integrated these factors to establish a complete and clear formation path of self-identity and
social identity, especially in donation crowdfunding. This study filled this gap, providing
important theoretical implications for future research on self-identity and social identity.

6.2 Practical implications

From a practical perspective, the conclusions of the study help us develop the concept of
donation crowdfunding and contribute to both the fundraisers and service providers when
we know the factors that govern individuals’ donation intentions.
Specifically, this research confirmed that a high face concern can contribute to fostering
the self-identity (βface-self ¼ 0.211) and social identity (βface-social ¼ 0.198), thus further stepping
up people’s donation intentions (βself-donation ¼ 0.429, βsocial-donation ¼ 0.403). Thus, fundraisers
should acknowledge the importance of face. Previous studies have argued that Chinese on one
hand are typically sensitive to face and are concerned with saving face for himself (or herself ),
while on the other hand tend to maintain the face of others, especially those they are familiar
with, such as family members or close friends. (Monkhouse et al., 2012). Therefore, fundraisers
should recognize the importance of acquaintances such as family members and friends when Exploring
promoting donation crowdfunding. It is advised to share the fundraising campaign with their individuals’
friends and family and ask them for help. It is due to individuals’ face concern may be aroused behavioral
under such conditions to maintain their own face and others’, which will eventually promote
their intention to engage in the donation crowdfunding. intentions
In addition, this paper confirmed that when individuals perceive a larger referent
network size (βnet-self ¼ 0.396, βnet-social ¼ 0.256), their further willingness to donate 1531
eventually increases. This leads to the conclusion that achieving a large network size
through a sound strategy may help increase more participation. Fundraisers can ask
supporters to share campaigns with their own networks as the higher the exposure to
crowdfunding, the more donations they may bring. It is also advisable to launch a
donation crowdfunding campaign with other strategic initiatives, especially at an early
stage, for example, by adopting effective communication strategies (e.g. blog, forums,
social networks, etc.) or sharing the campaign with influencers (e.g. online forum
administrators, media reporters, etc.). Service providers can also place information in a
conspicuous place on websites to inform potential donors that many individuals
(e.g. potential donors’ acquaintances) have already been funded the campaign.
Our results indicate that moral obligation contributes to the fostering of social identity
(βmoral-social ¼ 0.095), and it therefore contributes to the donation intention. In light of this,
the fundraiser can identify individuals who have previously contributed, as well as
individuals who have participated in other charitable projects, because these individuals
may have a higher sense of moral obligation to help others. In fact, most individuals donate
money for charitable activities because they perceive the act of donation as a moral
responsibility, only hoping to get inner satisfaction, instead of anything else in return.
Fundraisers can send certain intangible, free “rewards,” such as thank-you letters or
postcards, just to let the donors know that they are grateful for their help.
Sense of self-worth (βworth-self ¼ 0.206) has a significant effect on self-identity, indicating
that it is a critical factor in the formation of self-identity. According to this finding, it is
important for service providers and fundraisers to boost potential donors’ sense of self-
worth. To achieve this goal, they can focus on publicizing the importance and significance of
charitable donations. For example, “instead of putting money toward a gift someone may
look at once and never use again, you can donate that money to a charity in need on your
loved one’s behalf”; this leads to a feeling of self-worth, knowing that one is offering much-
needed resources to a great cause for those in need.
Perceived donor effectiveness (βdonor-self ¼ 0.132) has been confirmed to affect the
development of self-identity, which in turn exerts a significant influence on donation intention.
Thus, in order to increase opportunities and allow potential donors to fund a cause,
fundraisers need to use the campaign page to narrate their stories and let people know the
purpose of raising funds. Based on the type of donation crowdfunding they are doing,
fundraisers can talk about the hardships that they have faced or are facing and how they
want to overcome them. If potential donors know exactly how their donations will affect the
life of fundraisers, they are much more likely to make donations. In addition, crowdfunding
activities are not static fundraising methods. Instead, these require thoughtful updates and
progress reports. These updates not only allow existing supporters know that their donations
have made a difference, but also encourage potential donors to contribute to the campaign.
Therefore, it is recommended that fundraisers provide daily updates to their donors and let
the donors know how their donations help them or those in need.
Finally, social interaction is positively correlated to self-identity (βinteract-social ¼ 0.174) and
social identify (βinteract-social ¼ 0.088). Thus, communicating with others within virtual
communities can provide individuals with the opportunity of developing a shared
understanding with each other, thus arousing the group identity. Accordingly, fundraisers
IMDS should emphasize the importance of frequent interactions with potential donors. To do this,
119,7 our research recommends that communication tools be fully integrated into the management
process of donation crowdfunding; restricting social interactions may hinder donations. Based
on this finding, more efforts should be made to develop personal social interactions within the
crowdfunding community. We also recommend that fundraisers choose the crowdfunding
platform for fundraising in which individuals can interact with each other.
1532
7. Limitations and suggestions for future research
This study is not free from limitations. Our study was carried out on the basis of survey
data, during which, common problems related to the cross-sectional survey method were
inevitably encountered, so a longitudinal research is recommended in future research.
Moreover, it would be more persuasive to use the cost-benefit approach to examine the
motivation of participating in the donation crowdfunding. The cost–benefit approach is a
systematic approach to estimating the strengths and weaknesses of alternatives, which can
provide a brand new way of understanding individuals’ donation behavior. In addition, this
empirical study is restricted to respondents in China. It may be exhaustive to examine
whether culture differences are an attentional antecedent of donation intention. Thus, an
appealing expansion to the research would compare the donation intentions in different
cultural settings to determine whether the results are different.

References
Andreoni, J. (1989), “Giving with impure altruism: applications to charity and Ricardian equivalence”,
Journal of Political Economy, Vol. 97 No. 6 No. C, pp. 1447-1458.
Barbarossa, C., Pelsmacker, P.D. and Moons, I. (2017), “Personal values, green self-identity and electric
car adoption”, Ecological Economics, Vol. 140 No. C, pp. 190-200.
Beck, L. and Ajzen, I. (1991), “Predicting dishonest actions using the theory of planned behavior”,
Journal of Research in Personality, Vol. 25 No. 3, pp. 285-301.
Bock, G.W., Zmud, R.W., Kim, Y.G. and Lee, J.N. (2005), “Behavioral intention formation in knowledge
sharing: examining the roles of extrinsic motivators, social-psychological forces, and
organizational climate”, MIS Quarterly, Vol. 29 No. 1, pp. 87-111.
Boudreau, K., Jeppesen, L.B., Reichstein, T. and Rullani, F. (2015), Crowdfunding as “Donations”:
Theory & Evidence, Social Science Electronic Publishing, New York, NY.
Bradford, C.S. (2011), “Crowdfunding and the federal securities laws”, Social Science Electronic
Publishing, Vol. 21 No. 6, pp. 1081-1154.
Burtch, G., Ghose, A. and Wattal, S. (2011), “An empirical examination of the antecedents and
consequences of contribution patterns in crowd-funded markets”, Social Science Electronic
Publishing, Vol. 24 No. 3, pp. 499-519.
Chatzisarantis, N.L.D., Hagger, M.S., Wang, C.K.J. and Thøgersen-Ntoumani, C. (2009), “The effects of
social identity and perceived autonomy support on health behavior within the theory of planned
behavior”, Current Psychology, Vol. 28 No. 1, pp. 55-68.
Cheng, Z. and Guo, T. (2015), “The formation of social identity and self-identity based on knowledge
contribution in virtual communities: an inductive route model”, Computers in Human Behavior,
Vol. 43 No. 1, pp. 229-241.
Cojuharenco, I., Cornelissen, G. and Karelaia, N. (2016), “Yes, I can: feeling connected to others increases
perceived effectiveness and socially responsible behavior”, Journal of Environmental Psychology,
Vol. 48 No. 12, pp. 75-86.
Dermody, J., Koenig-Lewis, N., Zhao, A.L. and Hanmer-Lloyd, S. (2017), “Appraising the influence of
pro-environmental self-identity on sustainable consumption buying and curtailment in
emerging markets: evidence from China and Poland”, Journal of Business Research, Vol. 86
No. 10, pp. 333-343.
Dukerich, J.M. and Ashforth, B.E. (2000), “Role transitions in organizational life: an identity-based Exploring
perspective”, Academy of Management Review, Vol. 26 No. 4, pp. 670-683. individuals’
Dwyer, D.M., Haselgrove, M. and Jones, P.M. (2011), “Cue interactions in flavor preference learning: a behavioral
configural analysis”, Journal of Experimental Psychology: Animal Behavior Processes, Vol. 37
No. 1, pp. 41-57. intentions
Edwards, M.R. (2005), “Organizational identification: a conceptual and operational review”,
International Journal of Management Reviews, Vol. 7 No. 4, pp. 207-230. 1533
Evans, P. (2015), “Self-determination theory: an approach to motivation in music education”, Musicae
Scientiae, Vol. 19 No. 1, pp. 65-83.
Ewe, S.Y., Yap, S.F. and Lee, C.K.C. (2015), “Network externalities and the perception of innovation
characteristics: mobile banking”, Marketing Intelligence & Planning, Vol. 33 No. 4, pp. 592-611.
Fornell, C. and Larcker, D.F. (1981), “Evaluating structural equation models with unobservable
variables and measurement error”, Journal of Marketing Research, Vol. 18 No. 1, pp. 39-50.
Gerber, E., Hui, J. and Kuo, P.Y. (2012), “Crowdfunding: why people are motivated to post and fund
projects on crowdfunding platforms”, Computer Supported Cooperative Work, Workshop on Design
Influence and Social Technologies: Techniques, Impacts and Ethics, Seattle, February 11-15.
Giosan, C., Glovsky, V. and Haslam, N. (2001), “The lay concept of ‘mental disorder’: a cross-cultural
study”, Transcultural Psychiatry, Vol. 38 No. 3, pp. 317-332.
Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E. and Tatham, R.L. (1998), “Multivariate data analysis”,
Technometrics, Vol. 30 No. 1, pp. 130-131.
Hall, K. and Bucholtz, M. (2013), “Epilogue: facing identity”, Journal of Politeness Research, Vol. 9 No. 1,
pp. 123-132.
Han, S.Y., Yoo, J., Zo, H. and Ciganek, A.P. (2017), “Understanding makerspace continuance: a self-
determination perspective”, Telematics & Informatics, Vol. 34 No. 4, pp. 184-195.
Hogg, M.A., Terry, D.J. and White, K.M. (1995), “A tale of two theories: a critical comparison of identity
theory with social identity theory”, Social Psychology Quarterly, Vol. 58 No. 4, pp. 255-269.
Huang, E. (2012), “Online experiences and virtual goods purchase intention”, Internet Research, Vol. 22
No. 3, pp. 252-274.
Jebarajakirthy, C. and Thaichon, P. (2016), “Marketing microcredit to bottom of the pyramid market”,
International Journal of Bank Marketing, Vol. 34 No. 2, pp. 191-215.
Jiang, C., Zhao, W., Sun, X., Zhang, K., Zheng, R. and Qu, W. (2016), “The effect of self and social
identity on the intention to microblog: an extension of the theory of planned behavior”,
Computers in Human Behavior, Vol. 64 No. 11, pp. 754-759.
Kline, R.B. (2016), Principles and Practice of Structural Equation Modeling, 4th ed., The Guilford Press,
New York, NY.
Lin, K.Y. and Lu, H.P. (2011), “Why people use social networking sites: an empirical study integrating
network externalities and motivation theory”, Computers in Human Behavior, Vol. 27 No. 3,
pp. 1152-1161.
Liu, Q., Shao, Z., Tang, J. and Fan, W.G. (2019), “Examining the influential factors for continued social
media use”, Industrial Management & Data Systems, Vol. 119 No. 5, pp. 1104-1127.
Majchrzak, A., Malhotra, A. and John, R. (2005), “Perceived individual collaboration know-how
development through information technology-enabled contextualization: evidence from
distributed teams”, Information Systems Research, Vol. 16 No. 1, pp. 9-27.
Marshall, G.R., Hine, D.W. and East, M.J. (2017), “Can community-based governance strengthen
citizenship in support of climate change adaptation? Testing insights from self-determination
theory”, Environmental Science & Policy, Vol. 72 No. C, pp. 1-9.
Meijboom, F.L.B. and Brom, F.W.A. (2012), “Ethics and sustainability: guest or guide? On sustainability
as a moral ideal”, Journal of Agricultural & Environmental Ethics, Vol. 25 No. 2, pp. 117-121.
IMDS Monkhouse, L.L., Barnes, B.R. and Stephan, U. (2012), “The influence of face and group orientation on
119,7 the perception of luxury goods: a four market study of East Asian consumers”, International
Marketing Review, Vol. 29 No. 6, pp. 647-672.
Postmes, T., Haslam, S.A. and Swaab, R. (2005), “Social influence in small groups: an interactive model
of social identity formation”, European Review of Social Psychology, Vol. 16 No. 1, pp. 1-42.
Ryan, R.M. and Deci, E.L. (2000a), “Intrinsic and extrinsic motivations: classic definitions and new
directions”, Contemporary Educational Psychology, Vol. 25 No. 1, pp. 54-67.
1534
Ryan, R.M. and Deci, E.L. (2000b), “Self-determination theory and the facilitation of intrinsic
motivation, social development, and well-being”, American Psychologist, Vol. 55 No. 1, pp. 68-78.
Ryu, H.S. (2018), “What makes users willing or hesitant to use Fintech? The moderating effect of user
type”, Industrial Management & Data Systems, Vol. 118 No. 3, pp. 541-569.
Schlagwein, D. and Choy, K. (2015), “IT affordances and donor motivations in charitable crowdfunding:
the ‘Earthship Kapita’”, Case: European Conference on Information Systems, Münster, May 26-29.
Sirgy, M.J. (1982), “Self-concept in consumer behaviour: a critical review”, Journal of Consumer
Research, Vol. 9 No. 3, pp. 287-300.
Tabernero, C. and Hernández, B. (2011), “Self-efficacy and intrinsic motivation guiding environmental
behavior”, Environment & Behavior, Vol. 43 No. 5, pp. 658-675.
Tajfel, H. (1972), “Social categorization”, in Moscovici, S. (Ed.), Introduction à la psychologie sociale,
Larousse, Paris, pp. 30-37.
Thorbjørnsen, H., Pedersen, P.E. and Nysveen, H. (2007), “ ‘This is who I am’: identity expressiveness
and the theory of planned behavior”, Psychology & Marketing, Vol. 24 No. 9, pp. 763-785.
Venkatesh, V., Morris, M.G., Davis, G.B. and Davis, F.D. (2003), “User acceptance of information
technology: toward a unified view”, MIS Quarterly, Vol. 27 No. 3, pp. 425-478.
Whitmarsh, L. and O’Neill, S. (2010), “Green identity, green living? The role of pro-environmental self-
identity in determining consistency across diverse pro-environmental behaviors”, Journal of
Environmental Psychology, Vol. 30 No. 3, pp. 305-314.
Wilska, K.P., Paajanen, R., Terho, M. and Hamalainen, J. (2002), Device for Personal Communications, Data
Collection and Data Processing, and a Circuit Card, University of Idaho Library, Moscow, ID.
Wong, T.C., Haddoud, M.Y., Kwok, Y.K. and He, H.W. (2018), “Examining the key determinants
towards online pro-brand and anti-brand community citizenship behaviours”, Industrial
Management & Data Systems, Vol. 118 No. 4, pp. 850-872.

Further reading
Karl, E. (1995), Sensemaking in Organizations, Sage, Thousand Oaks, CA.
Paul, A.S. and Kwon, S.W. (2002), “Social capital: prospects for a new concept”, The Academy of
Management Review, Vol. 27 No. 1, pp. 17-40.
Revilla, M.a., Saris, W.E. and Krosnick, J. (2013), “Choosing the number of categories in agree-disagree
scales”, Sociological Methods & Research, Vol. 43 No. 1, pp. 73-97.
Weijters, B., Cabooter, E. and Schillewaert, N. (2010), “The effect of rating scale format on response
styles: the number of response categories and response category labels”, International Journal of
Research in Marketing, Vol. 27 No. 3, pp. 236-247.

Corresponding author
Yalan Li can be contacted at: m17729827763@163.com

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
 The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm

Analysis of
An empirical analysis of rural rural farmers’
farmers’ financing intention of financing
intention
inclusive finance in China
The moderating role of digital finance and 1535
social enterprise embeddedness Received 31 August 2018
Revised 29 October 2018
Gulizhaer Aisaiti Accepted 11 December 2018

School of Business Administration,

Xinjiang University of Finance and Economics, Urumqi, China
Luhao Liu
School of International Trade and Economics,
Shandong University of Finance and Economics,
Jinan, China
Jiaping Xie
College of Business,
Shanghai University of Finance and Economics, Shanghai, China and
School of Business Administration,
Xinjiang University of Finance and Economics, Urumqi, China, and
Jun Yang
School of Economics and Management,
Zhongyuan University of Technology, Zhengzhou, China

Abstract
Purpose – The purpose of this paper is to investigate and understand China’s rural farmers’ financing
intention of inclusive finance, and it examines related drivers like knowledge of inclusive finance, perceived
benefits and perceived risks of ordering finance. Besides, the social enterprise embeddedness and digital
finance are integrated into the conceptual model to further investigate their moderating impact.
Design/methodology/approach – The authors designed an inclusive finance intention model to examine
the relations between dependent variable knowledge of inclusive finance, intermediary variables perceived
benefits and perceived risks of ordering finance and the independent variable financing intention of inclusive
finance. The embeddedness of social enterprise and digital finance were identified as modifying factors. Both
exploratory and conclusive research strategies were applied. A structured questionnaire was developed to
collect empirical data from the rural areas of China.
Findings – It suggests that knowledge of inclusive finance can strengthen both perceived benefits and
perceived risk of ordering finance. Interestingly, the embeddness of social enterprise can significantly reduce
risk perceptions and improve perceived benefits of ordering finance. Furthermore, perceived benefits of
ordering finance can positively enhance rural farmers’ financing intention of inclusive finance, whereas
perceived risks can negatively influence the financing intention. Moreover, digital finance as a modifying
factor can significantly strengthen the positive correlation between perceived benefits of ordering finance and
financing intention of inclusive finance.
Practical implications – The research indicates that a systematic inclusive finance educational project is
needed to enhance rural farmers’ understanding of inclusive finance and its components. Moreover, the study
reveals that it is crucial to promote social enterprise participation and digital finance to develop inclusive
finance in rural China, as the service attributes of social enterprise and efficiency of digital finance can greatly
reduce the existing transaction cost of farmers. Industrial Management & Data
Systems
Vol. 119 No. 7, 2019
The work was supported by National Social Science Foundation of China: Grant No. 15ZDB161, “Research pp. 1535-1563
on the evolution of industry and enterprises in green whole industry chain”. The authors thank the editor © Emerald Publishing Limited
0263-5577
and the anonymous referees for comments and suggestions. The authors contributed equally to this work. DOI 10.1108/IMDS-08-2018-0374
IMDS Originality/value – The conceptual model would potentially contribute to researchers interested in
investigating the financing intention of inclusive financial services relating to rural population. The integration
119,7 of social enterprise embeddedness and digital finance is the uniqueness of this research conceptual model.
Keywords Inclusive finance, Digital finance, Perceived benefits of ordering finance,
Perceived risks of ordering finance, Social enterprise embeddedness
Paper type Research paper

1536
1. Introduction
Representing more than 20 percent of the world population, China is one of the fast growing
countries in terms of economy and political impact. However, still a majority of its rural
farmers have been suffering from poverty (WMFG, 2010). For decades, one of the most
popular ideas in the bottom of pyramid (BOP) literature is that poor people inevitably have
limited access to financial products and services (Sharif, 2000). The BOP market refers to a
market segmentation in which customers earn less than 1,500 dollar annual income
(Prahalad, 2010). It should be noted that BOP markets include many subdivisions, for
instance, poor rural communities, marginal farmers, landless laborers (Anderson et al.,
2010). In this sense, poor rural farmers can be categorized as a segment of the BOP market in
China. Such an assumption can be justified, as the majority of rural farmers suffering from
poverty and poor infrastructure and limited resources are quite common (Li et al., 2011).
Increasing numbers of organizations have become interested in the potential business
opportunities in the BOP market by designing products and services that exclusively serve
the poor customers ( Jebarajakirthy et al., 2015). One example is that the provision of
inclusive finance conveniently fulfills the financial needs of customers in the BOP markets
(Pitta et al., 2008). Inclusive finance refers to a type of finance model in which financial
institutions provide a range of quality financial services that can reach everyone, without
excluding the poor, disabled, rural population (Ashraf and Noor, 2010). Comparing with
traditional financial institutions, the major aim of inclusive financial institutions is to
improve the accessibility of affordable financial services to underserved consumers
(Varghese, 2001). Thus, inclusive finance is important for improving the living conditions
and eliminating the poverty of poor farmers, rural non-farm enterprises and other
vulnerable groups. Especially, rural farmers need financial access to improve agricultural
productions and further develop entrepreneurial activities in the rural areas.
However, the development of inclusive finance in rural China is full of various challenges
along with opportunities. Whether the new coming social enterprises, especially the
agricultural social enterprises, can play a significant role in terms of inclusive finance
development? Whether the application of digital finance can accelerate the process of
inclusive finance implementation? These underlined questions need to be answered by
researchers’ in-depth investigation, and further effective measures need to be taken to
enhance the performance of inclusive finance poverty alleviation projects in China.
To elaborate the significant role of inclusive finance in China, it is crucial to explain the
rural financial system. In China, inclusive finance providers include rural credit cooperatives
(RCC), rural commercial banks, rural cooperative banks, village banks (VB), microcredit
companies, urban commercial banks, postal saving banks and agricultural social enterprises
(WMFG, 2010). Social enterprises like microcredit companies are the new players in rural
financial system (Tang, 2009). By contrast, informal lending is a very common way of
financing in rural China, which can further indicate the potential size of the market (Turvey
and Kong, 2010). However, the regional wealth disparity is still one of the significant problems
considering the sustainability of national economic growth (Li et al., 2011). To shorten the
increasing wealth disparities, the central government has been providing the poor and rural
population financial access through RCC (Wang, 2004). Social enterprises like microcredit
companies also play a significant role in rural financial system. However, social enterprises
have not met the needs of rural consumers due to the notion that a social enterprise is Analysis of
relatively new, leading to a source of confusion (Wang et al., 2015).Therefore, social enterprises rural farmers’
need ongoing promotions in the rural areas to meet the needs of underserved population. financing
Furthermore, the progress of developing inclusive finance in rural China needs long-term
efforts from various stakeholders, including government, inclusive financial institutions, social intention
enterprises to accelerate the systematic adaptations of financial technology and to improve the
service quality and efficiency (Chan, 2010). Additionally, a technology-oriented approach of 1537
inclusive financial institutions is essential to put in place to ensure the sustainability of
inclusive finance system and to improve the intention of inclusive finance (Koh et al., 2018).
RCC and microcredit companies provide a small amount of loan to people who cannot get
access from traditional financial institutions with low interest rate asking for collaterals
(Tang, 2009). Therefore, inclusive finance can empower the rural farmers, local vendors and
SMEs to accelerate the economic growth of countryside (Li et al., 2011). From this
perspective, it is essential to enhance the efficiency of inclusive finance in rural China to
re-balance the Chinese economy to a domestic demand-led model (Li et al., 2011). Although
social enterprises have reached a small proportion of the BOP market, thousands of poor
farmers and vendors across the nation have benefited from loans over the course of time
(Albert and Ren, 2001). Therefore, rural China represents potential opportunity for
promoting the inclusive finance. Thus, the main aim of this research study is to investigate
rural farmers’ benefit and risk perceptions on obtaining ordering finance and financing
intention of inclusive finance. It further investigates related drivers, the embeddedness of
social enterprise and digital finance integrated as modifying factors to examine the
moderating impact on the model. A unique conceptual model has been developed for
research purpose, which is empirically tested. The proposed model and its findings can
possibly be applied to other customer segments of the BOP markets, hence providing
managerial implementations. A total of 160 valid empirical data were collected from
respondents aged under 60 in Xinjiang province of China in 2018. The collected data
represented the rural population of the BOP market in China. Moreover, agricultural
activities like farming, stockbreeding and entrepreneurial activities were the main economic
activities in the rural areas of China (Tang, 2009).

2. Literature review
2.1 Consumers’ attitudes toward inclusive finance
Considering the feature of the customers of BOP market, it is reasonable to make an
assumption that there are significant differences between the customers from BOP market
and the other middle and top market segments in respect of their attitude, perceptions and
actual behavior toward products and services (Barki and Parente, 2010). One of the major
topics to be investigated in this field is the relation between attitudes and intention. Most
researchers suggested that cognitive attitudes have a significant impact on purchasing
intentions (Ajzen, 1996; Turvey and Kong, 2010). Attitude can examine a person’s positive or
negative assessment regarding a specific behavior and possibly predict the desirable
outcomes based on the formed perceptions (Lee, 2009). Smith et al. (2008) suggested that
affective and cognitive attitudes can predict purchasing intention. Therefore, a positive or
negative attitude can directly influence the strength of consumers’ behavior (Ajzen, 1996).
First, attitude refers to individual’s psychological evaluation of an object (Ajzen, 1996).
Consumers’ attitude has an impact on the purchase intention of a product or service (Barki
and Parente, 2010). Second, the relation between attitude and consumer behavior has been
studied extensively in the context of social psychology. Ajzen (1991) suggested that the
relationship between attitude and behavior in financial decision might be more complicated.
As such, it is possible that consumers with favorable attitudes might result in not performing
the behavior (Michael, 2009). Thus, it is crucial to identify predominant attitudes that hinder or
IMDS encourage consumers from making inclusive financial decisions. Third, previous studies
119,7 suggest that consumers’ attitudes are a decisive factor of intentions relating to making
financial decisions like obtaining credit (Wickramasinghe and Gurugamage, 2012).
Scholars suggested three types of attitudes, namely, perceived affect, perceived benefits and
perceived deterrents (Albert and Ren, 2001). Positive affect refers to an individual’s psychological
state that encourages him or her to make an impulsive decision (Russell and Barrett, 1999).
1538 Considering the maturity of inclusive finance in the BOP market of China, the positive affect is
not applied in this study as a type of attitude. Moreover, perceived risks have been applied as a
type of unfavorable attitude in inclusive finance literature (Gärling et al., 2009; Turvey and Kong,
2010).Therefore, perceived risks are applied as an attitude in this study.
However, the relation between favorable attitude and strong intention is not the only
possible approach in the terms of inclusive finance. Making a financial decision might be more
complicated for consumers in the BOP market. As such, consumers might have positive
attitudes toward inclusive finance, but the financial demand cannot be fulfilled due to
the intentional exclusion from inclusive financial institutions on the basis of the factors like low
service quality and service accessibility of local inclusive financial institutions (Li et al., 2011).
In this sense, attitudes might be related to socio-demographic factors and service quality of
financial institutions (Wickramasinghe and Gurugamage, 2012). Some studies in social
psychology have found that attitude and behavior are not always interrelated (Ajzen, 1991).
However, few existing literature have discussed rural consumers’ attitudes toward inclusive
finance and further analyzed the relations between formed attitudes and intentions. As the
inclusive financial products are intangible and abstract, the formation of attitudes might be
different from tangible products in the rural areas, considering the consumers’ features and
financial market maturity. Therefore, it is crucial to understand China’s rural farmers’ attitudes
in the context of a specific social and demographic area to explore the relations between
attitude and intention. It should be noted that ordering finance is one of the financial services
that are largely applied among farmers in rural China; most of the related literature have not
discussed attitudes of ordering finance. To be more specific, the ordering finance refers to a
type of financial service that is provided by inclusive financial institutions, which enables
farmers to apply for loan on the basis of the future income pledge from the real and effective
orders of transactions (Xie et al., 2017). Therefore, it is reasonable to examine farmers’ attitude
toward ordering finance as ordering finance is most representative inclusive financial service
in rural China. Based on the context of the BOP market of rural China, two types of attitudes
are formulated: perceived benefits of ordering finance and perceived risks of ordering finance.

2.2 Perceived benefits of ordering finance

Perceived benefits refer to the consumers’ positive perception of desired outcomes
associated with their decisions (Chandon et al., 2000). Ordering finance is a major type of
inclusive financial service in rural areas of China (Xie et al., 2017). The perceived benefits of
ordering finance can formulate positive perceptions and further influence farmers’
purchasing intentions ( Jose et al., 2012). Consumers expect to increase their household
income and enhance their family’s living standard, and benefit their communities by
extension as a result of their financing decisions ( Jebarajakirthy et al., 2015). Lee (2009)
suggested two main types of perceived benefits, which can be categorized as direct benefits
and indirect benefits. Direct benefits refer to immediate and tangible benefits that result
from financial decisions (Koh et al., 2018). In this sense, increasing family monthly income is
the direct benefit of ordering finance. Indirect benefits refer to the benefits that are less
tangible and difficult to measure (Koh et al., 2018). Thus, the improvement of living standard
is an indirect benefit of ordering finance, as it needs relatively long period of time to achieve.
Research works on mental accounting suggested that it is necessary to understand
consumers’ mental trade-offs of a financial decision (Szajna, 2010).
2.3 Perceived risks of ordering finance Analysis of
Perceived risk is individual’s subjective expected loss that impacts decisions (Lee, 2009). rural farmers’
Perceived risk is a cognitive assessment and important component of the financial financing
decision-making process (Weber, 2004). Perceived risk is an inevitable cognitive evaluation
in the terms of financial decision-making process (Weber, 2004). Researchers suggested that intention
consumers are influenced more by perceived risks than objective risks (Diacon and Ennew,
2001). Furthermore, a number of scholars suggested that perceived risk has different 1539
dimensions. Jacoby and Kaplan (1972) proposed six components of perceived risks, namely,
financial risks, safety risks, psychological risks, performance risks, time risks and social
risks. Based on the perceived risk theory of ( Jacoby and Kaplan, 1972), six types of
perceived risks are defined as follows. Financial risks refer to the amount of financial cost
and possible financial loss of obtaining ordering finance. Psychological risk refers to the
negative emotion of obtaining ordering finance. Social risk refers to the feelings that
obtaining ordering finance may result in disapproval of farmer’s friends, family members
and communities. Safety risks refer to the possible safety threats that might be caused by
failing in repayment; performance risks refer to consumers’ negative perceptions related to
the service attributes of ordering finance. Time risk refers to time cost of getting financial
access. Considering the vulnerability of the consumers from the BOP market, it is
reasonable to make assumption that perceived risks of ordering finance decrease the
financing intentions. Furthermore, the level of risk perception is mainly based on
consumers’ knowledge of product features, which relates to consequences of purchasing
behavior (Turvey and Kong, 2010). From this perspective, knowledge of inclusive finance
and perceived risks of ordering finance might have an interrelation.

2.4 Knowledge of inclusive finance

Knowledge of inclusive finance can formulate and restructure the cognitive component of
consumers’ evaluations of financial products or services (Kidwell and Turrisi, 2004).
Consumers with sufficient knowledge of inclusive finance have better understanding of
product features (Kinsey and McAlister, 1981). Zhu and Meeks (1994) carried out an
empirical research among low-income families to examine the relationship between
knowledge, attitude, and credit use. According to their research results, the well-educated
consumers showed greater desire and intentions for obtaining credit, whereas low-income
family have a low level of financial knowledge and intentions. The educational level of rural
population is relatively low in China (Li et al., 2011). Thus, the knowledge of inclusive
finance should be investigated from service attributes rather than consumers’ educational
level (Turvey and Kong, 2010). For instance, applying financial services include a great
number of information: contract terms, interest rate, payment method, repayment duration
and the risk of indebtedness (Albert and Ren, 2001). Consumers can make more
economically sound decisions based on their cognitive understanding of inclusive finance
(Gärling et al., 2009). In contrast, consumers with insufficient knowledge tend to make poor
financial decisions (Montano and Kasprzyk, 2008). Jebarajakirthy et al. (2015) suggested that
knowledge of microcredit enhanced positive attitudes toward microcredit and decreased
risk perception of the consumers from the BOP market. Therefore, the level of knowledge
and awareness of inclusive finance are related to both attitudes and intentions.

2.5 Embeddedness of social enterprise

There is an array of definitions of social enterprise, based on the perspectives from different
school of thoughts. Social enterprise refers to the use of commercial activities by non-profit
organizations in terms of supporting social missions (Dees, 1998). It represents a new
business model that enables customers to become members and to have share in the profits
of the business (Byerly, 2014). The notion of the social enterprise has gained popularity in
IMDS academic field in recent years. However, the concept of a social enterprise is relatively new
119,7 and has not widely known for rural consumers. In China, the term first appeared in an
academic journal in the year 2004 (Zhang et al., 2010). Social enterprise continuously
promotes innovative ideas that drive the BOP market development to strike a balance
between running a sustainable business and achieving social objectives (Irwin, 2014).
Moreover, social enterprise provides a business framework for tackling social issues by
1540 combining entrepreneurial activities with the aim of improving life quality (Byerly, 2014).
Effectiveness of social enterprise business model has been investigated. For instance, social
enterprises like microcredit companies offer inclusive finance business model that mainly
focuses on eradicating extreme poverty with incredible growth potential in the BOP market
(WMFG, 2010). Thus, social enterprises contribute to find new business opportunities to
promote, simultaneously, entrepreneurial spirit and the pursuit of the public good (Byerly,
2014). Nevertheless, social enterprise is still at the initial stage in rural China; some examples
have shown that it can help businesses to develop solutions to eliminate poverty by
cultivating market-oriented approach (Wang et al., 2015). The social enterprise
embeddedness not only motivates farmers to remain actively participated in agricultural
production, but also provides technical and financial support to improve farmers’
performance (Xie et al., 2017 ). An empirical analysis suggested that the high level of
relational and structural embeddedness of social enterprise might greatly promote the
sustainable agricultural development in rural China (Xie et al., 2017). In rural areas, social
enterprise embeddedness is integrated in whole agricultural value chain. For example,
agricultural social enterprises provide a set of comprehensive services that meet farmers’
need before, during and after the agricultural production (Xie et al., 2017). It should be noted
that the social enterprise embeddedness has not been investigated as a type of moderating
factor in relevant literature. Moreover, the previous literature of social enterprise have not
covered the moderating role of social enterprise embeddedness in the context of inclusive
finance. The social enterprise embeddedness might provide whole new perspective for
scholars to explain the relations between the formation of inclusive financial knowledge and
attitudes of ordering finance at institution participation level.
Therefore, in this research paper, the authors assumed social enterprise as embedded
factors, since it plays a significant role throughout the whole process of agricultural
production. Thus, in this research, social enterprise embeddedness is proposed as a
moderating factor. Considering the significant role of social enterprise in rural China, it is
necessary to investigate the participation of social enterprise to attain consumers’ insights.

2.6 Digital finance

According to the definition of the Financial Stability Board, digital finance refers to the
emerging business models that are driven by big data, block chains, cloud computing,
artificial intelligence and other up-to-date technologies in financial market (Koh et al., 2018).
The main mission of inclusive finance is to increase the customer scope, availability and
satisfaction of financial services to meet the increasing financial needs of the people from the
BOP market, including farmer SME, poor and the physically challenged people, and elders to
obtain suitable financial service at affordable interest rate in a convenient, efficient and safe
manner (Owens, 2013). Internet, big data, cloud computation, artificial intellectual, block
chains have ever increasingly become the interest of academic topics and emerged a new
concept called Fitech. Fitech is not the simple combination of finance and technology and it
refers to the application of various technological approaches to improve the service efficiency
and reduce the operational costs by extension (Radcliffe and Voorhies, 2012). Digital finance
decreases the transaction cost by saving the time cost and human resources cost (Radcliffe
and Voorhies, 2012). Thus, it is crucial to examine the significance of innovation and digital
finance that strives to promote growth and development of inclusive finance in a digital age.
Transaction cost is a decisive factor in consumers’ decision-making process (Antonides et al., Analysis of
2011). The definition of transaction cost refers to the overall cost of consumers’ market rural farmers’
participation, including the cost from accessing product information, collecting information financing
from various source, etc. (Liu et al., 2016). In the past, rural consumers excluded from inclusive
financial institutions due to the restriction of physical distance and the dual constraints of intention
costs and benefits (Li et al., 2011). The transaction cost of extra transportation and time has
caused rural farmers to leave inclusive finance. Nevertheless, development of information 1541
technology has enabled financial institutions to consistently revolutionized their service
quality by reducing the additional transaction costs. For instance, online banking, mobile
banking, mobile transaction are based on the attributes of financial technology.
Since 2004, many inclusive financial institutions have adopted mobile technology to
provide financial services to the underbanked and unbanked (Owens, 2013). By adopting
digital finance, consumers can get reasonable, convenient and safe financial services in time.
However, the reduction of transaction cost enables inclusive financial institutions to provide
consumers with more accurate services, thereby reducing reluctant financial procedures
(Byerly, 2014). However, the moderating role of digital finance in the relation between
consumers’ attitudes and financing intention has not gained much attention in inclusive
finance-related studies. Therefore, there is a necessity for understanding the relation between
financing attitude and financing intention in the context of digital finance adoption in rural
China to fill the existing research gap. Thus, the digital finance is proposed as modifying
construct in this research. To be concise, the application of digital finance in the field of
inclusive finance is an increasing trend. Therefore, the role of digital finance should not be
neglected in the BOP market. Whether digital finance can moderate consumers’ perception
of ordering finance and further impact the financing intention of inclusive finance is an
interesting perspective to investigate. It should be noted that the digital finance has not been
proposed as a type of moderating factor in relevant literature. Considering the nature of rural
farmers, it is essential to investigate the impact of digital finance in the BOP sector of China.
The previous literature of inclusive finance and consumer behavior have examined the
relations between knowledge, attitudes and intentions. However, most of the literature has not
investigated farmers’ financing intention of inclusive finance and its driving factors in the
context of rural China. To fill the existing gap in literature, an empirical inclusive finance
model is developed to examine the rural farmers’ financing intention of inclusive finance. This
model would potentially contribute to researchers interested in investigating the financing
intention of rural population and the other consumers from the BOP market. Additionally, the
integration of social enterprise embeddedness and digital finance as modifying factors is the
uniqueness of this research. It provides a revolutionized prospective for researchers to get new
consumer insights and to further apply the conceptual model to other financial services to get
in-depth understanding of rural farmers’ decision-making process.

3. Development of conceptual model

3.1 Knowledge of inclusive finance’s impact on attitudes
Knowledge is a component of human cognitive structure that has a significant impact on
individuals’ perception toward a particular object or phenomenon ( Jebarajakirthy and Lobo,
2014). Knowledge of inclusive finance refers to consumers’ level of awareness of inclusive
finance-related concepts such as interest rate, transaction cost, method of payment, etc.
(Cheng, 2007). Consumers with a high level of knowledge of inclusive finance have a rational
understanding of benefits and risks that are associated with a financial decision (Chien and
Devaney, 2001). Consumers’ evaluating criterion of financial decision is based on the level of
information they possess (Michael, 2009). It suggests that consumers with sufficient
knowledge of financial product have a larger probability to make rational and risk-averse
purchasing decisions (Smith et al., 2008). In contrast, lack of inclusive financial knowledge
IMDS leads to impulsive and irrational decision, which is to say, consumers might overrate the
119,7 desirable outcome and undesirable outcome of their financial decisions based on their
perceptional attitudes ( Jebarajakirthy et al., 2015). Since inclusive finance is complex to
understand by mass majority of rural population in China, the low-literate consumers have a
big disadvantage in inclusive finance market (Cheng, 2007). The low-literate rural
population copes with the barriers of illiteracy by avoiding situations wherein they have to
1542 expose their low level of knowledge in terms of inclusive finance. Hence, the first hypothesis
is formulated as:
H1. Knowledge of inclusive finance has a positive impact on perceived Benefits and
perceived risk.

3.2 Attitudes’ impact on intention of inclusive finance

There is a complex relationship between attitudes and behavior. This paper will consider
two types of attitudes and examine the relations between attitude and intention of inclusive
finance. Pitta et al. (2008) suggested three types of attitudes toward financial decisions,
namely, perceived benefits, perceived risks and perceived deterrents. As perceived
deterrents are components of financial exclusion, perceived deterrents have not been applied
in this research paper. Further, considering the market maturity of China’s inclusive finance
system, perceived risks and perceived benefits are integrated as two types of major
attitudes toward inclusive finance. Thus, two types of attitudes are formulated: perceived
benefits and perceived risks.
3.2.1 Perceived benefits of ordering finance. Jebarajakirthy et al.’s (2015) survey results
indicated that perceived benefits can significantly enhance consumers’ intentions.
Therefore, consumers’ desirable and positive interpretations of inclusive finance can
strengthen their financing intentions. Direct benefits and indirect benefits are the two types
of perceived benefits (Gärling et al., 2009). Direct benefit is the consumers’ perceptions of
instant and short-term beneficiaries of inclusive finance (Koh et al., 2018). The expectation of
increasing household income by utilizing financial services is a type of direct benefit.
Indirect benefit is consumers’ perception of profound and long-term beneficiaries that result
from applying inclusive finance products (Koh et al., 2018), for instance, improving
household’s life standards and developing communities well-being (Li et al., 2011). Thus, the
second hypothesis is formulated as:
H2. Perceived benefits of ordering finance have a positive impact on the intention of
inclusive finance.
3.2.2 Perceived risks of ordering finance. Making financial decisions includes extensive
mental accounting and entails a certain kind of perceived risk (Michael, 2009). In this
research paper, perceived risk refers to consumers’ conception that inclusive finance has
potential negative consequences. From the perspective of accessibility and availability of
inclusive finance, risk-related perception exists among the people from the bottom of the
pyramid (Turvey and Kong, 2010). Inclusive finance is a comparatively new concept in rural
China, as the business scope of inclusive financial institution has not reached its potential
due to certain constrictions (Li et al., 2011). However, considering the low-literate rate of
rural population, it is reasonable that complicated financial terms and operational
procedures might be perceived as potential risks by rural consumers (Li et al., 2011).
Jebarajakirthy et al. (2015) suggested that perceived risks reduce consumers’ intention of
obtaining financial services and decrease the intention of inclusive finance by extension. In
contrast, perceived risk is not a decisive factor for consumers who have less income as
they have less to sacrifice from an irrational, impulsive financial decisions (Michael, 2009).
For instance, highly confident consumers have less to worry about social risks that are Analysis of
related to purchasing financial services that inclusively serve underbanked population, rural farmers’
whereas financially and psychologically vulnerable consumers might experience financing
undesirable social risks (Godwin, 1997). In total, five dimensions of risks are underlined
in previous research studies, namely, monetary risk, functional risks, physical risks, social intention
risks and psychological risks ( Jacoby and Kaplan, 1972). People tend to remain concerned
about the risks associated with the decisions, as applying for the credit might affect their 1543
financial safety ( Jacoby and Kaplan, 1972). Scholars increasingly mentioned that
consumers’ perceived risks of inclusive finance reduce their intentions (Cheng, 2007;
Li et al., 2011). Physical risk refers to individuals’ interpretation of the consequences of
behavior related to physical safety; psychological refers to individuals’ perception of the
consequences of the decision related to self-esteem; financial risk refers to individuals’
perception of over-indebtedness; social risks refer to the individuals’ anticipated
pressures from social interaction as the consequence of the financial decision; and time
risks refer to the cost of time, which are related to inclusive finance services ( Jacoby and
Kaplan, 1972). Jebarajakirthy et al. (2015) suggested that perceived risks reduce consumers’
intentions and decrease the consumers’ intention of inclusive finance. Thus, the third
hypothesis is formulated:
H3. Perceived risks of ordering finance have a negative impact on consumers’ intention
of inclusive finance.

3.3 Moderating effects of social enterprise embeddedness

Social enterprise is a form of enterprise that makes profits and maintains business
sustainability through maximizing the social value (Dees, 1998). Social enterprise integrates
both commercial sustainability and social welfare to enhance the overall value of a society
(Doherty et al., 2014, Zhang et al., 2010). The priority of social enterprise is vulnerable social
group like farmers, and they will not only provide agricultural production services but also
include inclusive financial services to best serve the interest of rural households (Wang et al.,
2015). According to the empirical research study conducted in China, the participation of
agricultural social enterprise can largely stimulate the agricultural production performance
and promote sustainable agricultural development in China (Xie et al., 2017). Most social
enterprises like Microcredit companies mainly targets poor families in rural regions, with
small and medium loan characterized by a relatively small size, which is under 30,000 Yuan
per loan within 1‒2-year loan duration (Zhang et al., 2010). Social enterprises not only
provide inclusive financial services like microcredit loan but also provide non-financing
services, such as training of agriculture technique, innovation of make-to-order agriculture
business management model and financial education (Xie et al., 2017). In this respect,
consumers’ trust of institution is related to an individual’s willingness to accept services
from a social enterprise. It has been suggested that trust can reduce uncertainty by making
risks manageable (Gärling et al., 2009). Since obtaining financial services is associated with
vulnerability, risk and interdependence, it is important to understand the moderating
impact of social enterprise embeddedness. Furthermore, two dimensions of trust have been
suggested by scholars to investigate consumers’ trust in social enterprises (Weisberg et al.,
2011). The first dimension is institutional trust, which refers to the extent to which a social
enterprise is perceived as being worthy of trust (Gärling et al., 2009). Institutional trust can
be influenced by the image and reputation of the institution, and also can be built by both
sustainable policy, practice and effective external communications (Weisberg et al., 2011).
The second dimension of trust is interpersonal trust, which refers to the consumers’ level of
confidence and positive expectations toward functional attributes of the social enterprise
IMDS (Epstein and Yuthas, 2011). Social enterprise embeddedness can be reflected by the level of
119,7 trust of consumers and willingness of applying services. The level of consumers’ trust on
social enterprise might impact rural farmers’ financing intention (Liu et al., 2013).
The embeddedness of social enterprise can fill the gap between inclusive financial service
demand and supply. More specifically, most of the rural farmers have difficulty in
understanding the financial terms and related knowledge in application to repayment
1544 process due to restrictions of their low educational level (Zhang et al., 2010). They might be
aware of their financial needs, but they have less constructive idea when it comes to utilizing
the loan to improve their agricultural production and enhance family income. Therefore, the
embeddedness of social enterprise will provide instructions to rural farmers with inadequate
financial awareness and investment capacity. The social enterprise embeddedness enhances
rural farmers understanding of inclusive finance and related terms with financial
educational projects and services. Furthermore, social enterprise can build up a platform to
link rural farmers with distributors and improve their expected investment return, which
could increase rural farmers’ perceived benefits (Xie et al., 2017). By providing helping
mechanism, social enterprise embeddedness can track rural farmers’ credit and prevent
them from loan default, which could decrease their perceived risks. Thus, we assumed that
the level of social enterprise embeddedness has a moderating impact on the risk tolerance
and the level of positive expectations of inclusive finance (Mayer et al., 1995). Hence, the
fourth hypothesis is formulated:
H4. Social enterprise embeddedness has a positive impact on the association between
knowledge of inclusive finance and perceived benefits of ordering finance, and a
negative impact on the association between knowledge of inclusive finance and
perceived risk of ordering finance.

3.4 Moderating effects of digital finance

The main mission of inclusive finance is to increase the customer scope, availability
and satisfaction of financial services to meet the increasing financial needs of the people
from BOP with convenient, efficient and safe approach (Koh et al., 2018). Digital finance
plays a significant role in terms of improving the efficiency of service and saves a
considerable amount of transaction cost caused by applying financial services from
traditional methods (Bruett, 2007). Moreover, by providing convenient and flexible
financial services, digital finance has largely decreased the transaction cost and enhanced
the accessibility of inclusive finance (Owens, 2013). The connectivity of supplier and
consumers, and consumers and consumers is the major feature of digital financial service.
Telecommunication network enables customers to connect and communicate with the
provider’s transaction authorization system through mobiles (Koh et al., 2018). The third
party payment applications like WeChat payment, Zhifubao payment provide
identification capturing points and virtual account for customers’ payment transactions
(Koh et al., 2018). Digital finance like P2P makes it easy for agricultural firms to get
financial access for agricultural innovation (Radcliffe and Voorhies, 2012). Furthermore,
from the prospective of economic growth, the vulnerable farming household has a greater
reliance on digital finance, and it experiences larger marginal revenue by applying
inclusive finance (Owens, 2013). However, the limitation of financial technology and the
low level of educational literacy might hinder rural household to evaluate pros and cons of
inclusive finance. Digital finance might enhance the association between perceived
benefits and financing intention of inclusive finance with decreasing the transaction cost
and improving time. As compared to traditional face-to-face transaction method, digital
finance makes it easy for rural farmers to transfer money through their digital account
without risking loss of paper currency. Furthermore, the efficiency and the virtual
transparency of digital finance might decrease rural farmers’ risk perceptions. Hence, the Analysis of
hypothesis is formulated as follows: rural farmers’
H5. Digital finance can strengthen the association between perceived benefits and financing
intention of inclusive finance, and decrease the negative association between intention
perceived risk of ordering finance and intention of inclusive finance.

1545
3.5 Underpinning theories
3.5.1 The theory of planned behavior (TPB). The TPB has been proven successfully in
predicting and explaining consumers’ behavior in an array of fields (Ajzen, 1991). The
TPB suggests that individual’s intentions to perform certain behaviors can be predicted
by attitudes toward behavior, subjective norms and perceived behavioral control. TPB is
the extension of the theory of reasoned action (TRA), which explains the relation between
intentions and behaviors (Ajzen, 1991). TRA proposes that attitude toward an actual
behavior is the result of the positive and negative perceptions of performing that behavior
(Ajzen, 1996). Moreover, the TPB suggests three conceptually independent determinants
in predicting intentions, namely, attitudes, subjective norms and perceived behavioral
control. However, the level of importance of attitude, subjective norm, and perceived
behavioral control in the prediction of intention is different, based on behaviors and
situations (Ajzen, 1991). This finding indicates that, considering the consumers’ intention,
personal attitudes tend to overshadow the other factors. At the fundamental level, the
TPB theory postulates that actual behavior is a function of salient beliefs, which is
relevant to the intentions (Ajzen, 1991). Personal attitudes are considered to be prevailing
determinants of an individual’s intentions and actions (Ajzen, 1991). Majority of social
psychologists apply information-processing approach to attitude formation (Ajzen, 1991).
This approach suggests that people formulate their attitudes based on the information
and knowledge related to certain attributes of an object. In the respect of attitudes, each
type of attitude links the behavior to the benefit or cost that happened by performing the
behavior (Lee, 2009). Since the inclusive finance can be valued positively and negatively, it
is possible that favorable attitudes can lead to desirable outcome and unfavorable
attitudes lead to undesirable consequences (Turvey and Kong, 2010). From a general view,
the application of the TPB in inclusive finance can help to formulate the determinants that
might influence rural populations’ actual participation and intention of inclusive finance.
It provides a useful conceptual model for understanding the complexities of consumers’
behavior in inclusive finance.
3.5.2 The theory of bottom of pyramid (BOP). Prahalad (2010) referred people from BOP
as profitable consumers. Karnani (2006) also suggested that the most effective way to
eliminate the poverty is to identify poor as producers rather as consumers. The BOP
perspective provides a market-based approach to poverty alleviation. The theory of BOP
provides a holistic and efficient instruction for analyzing the nature of rural population in
China. Successful BOP approaches enable to accelerate the sustainable development of
inclusive finance (Anderson et al., 2010). Adopting digital finance among BOP market is the
foremost significant obstacle that hinders the integration of sustainable development of
the poor (Koh et al., 2018). However, there is a profound traditional belief that BOP
consumers do not want to adopt financial technology easily. However, Prahalad (2012)
claimed against this traditional belief, pointing that there is a huge demand to adopt
financial technology in the BOP market. For example, BOP consumers have access to mobile
phone and mobile banking, which enables them to increase the participation in digital
finance like mobile transaction. Furthermore，Moseley and Rogers (2004) suggested that
the complexity of the innovation has a significant impact on the adoption of innovation in
IMDS the BOP market. Thus, it is crucial to understand the consumers’ behavior in terms of
119,7 China’s BOP inclusive financial market. Economic researchers have shown that economic
and finance are both influenced by constitutive of social factors (Michael, 2009). The
integration of BOP theory strengthens the major focus of this research. Therefore, the
conceptual model for this study is illustrated in Figure 1, which shows the knowledge of
inclusive finance, types of attitudes and their influence on China rural population’s
1546 financing intention of inclusive finance. This model conceptualized the major drivers of
intention of inclusive finance and examined the interrelations between knowledge of
inclusive finance, perceived benefits of ordering finance, perceived risks of ordering finance
and intention of inclusive finance. In addition, social enterprise embeddedness and digital
finance added as modifying constructs to investigate their impacts on consumers’
perceptions and the financing intention.

4. Research design and methodology

4.1 Research design
A good research design should collect all applicable information to ensure the efficiency and
rigorousness of research conduction (Malhotra, 2009). Thus, based on the research conceptual
model and formulated hypotheses, both exploratory and conclusive research strategies are
applied in this study. The primary research strategy includes conclusive research method.
The major primary research is defined as quantitative research, and a questionnaire is
developed to gather data to answer the research problems. The data collection and analysis
procedures are included.
4.1.1 Secondary research strategy. The main aim of secondary research is to develop a
conceptual model and identify hypotheses. Secondary data have been collected from the
various sources, which include existing literature about consumers’ attitude and intention of
inclusive finance, industrial reports and socio-demographic reports, such as current
situation of the inclusive finance in China, an analysis of rural financial system. However, it
should be noted that the validity of secondary data is not certain (Malhotra, 2004). In order
to summarize the findings and explore the gaps in the existing research, a thorough
secondary research is already conducted in the context of inclusive finance. The secondary
sources are adopted to identify six constructs and conceptualize the variables. Secondary
data are collected by reliable academic journals like Journal of Retailing and Consumer
Services, International Journals of Consumer Studies, etc. Insights from secondary research
will play an important role in formulating questions for the survey (Malhotra, 2009).
4.1.2 Primary research strategy. Conclusive research is applied, and it aims at testing the
specific hypotheses of a study and examining the relations between various constructs
(Malhotra, 2009). In conclusive research, findings relatively conclusive can be used as
contribution in managerial decision-making process (Malhotra, 2009). Thus, conclusive

Social Enterprise
Embeddedness

H4a H4b Perceived Benefits of

Ordering Finance H2
Knowledge of H1a
Intention of
Inclusive H3 Inclusive Finance
Finance H1b Perceived Risks
of Ordering Finance
Figure 1. H5a H5b
Conceptual model
for this study Digital Finance
research design has been applied for this study to test the hypotheses. Conclusive research Analysis of
is selected to gather primary data within a limited period of time, as it is a cost-efficient rural farmers’
method (Churchill and Iacobucci, 2002). This study applies a quantitative survey to gather financing
data, and respondents are asked 28 questions on personal information, knowledge of
inclusive finance, attitudes toward ordering finance and financing intention of inclusive intention
finance. Survey method is applied, as it is a cost-efficient approach to gather large amounts
of data (Malhotra, 2004). 1547
4.2 Sampling and instrumental development
Identifying the sampling frame enables researcher to identify the target population of the
study (Malhotra, 2004). From the perspective of age, the sample should be above 18 years
and under 60 years, the age range that is legally applicable for inclusive finance clients
(Chan, 2010). From the characteristics of respondents, the people who lived in countryside,
county town and subcity areas were selected. Therefore, the sample for this study
comprised 160 rural population, aged 18‒60 years, who were conveniently selected from
Xinjiang province of China.
4.2.1 Questionnaire design. To analyze the impact of independent variables on the
dependent variables, a questionnaire has been designed and tested (Malhotra, 2009). In
this study, dependent variable is intention of inclusive finance, independent variable is
knowledge of inclusive finance, the intermediary variables are perceived benefits of
ordering finance and perceived risks of ordering finance. The modifying variables are the
social enterprise embeddedness and digital finance. A survey was designed from the
existing validated scales and the combination of some self-developed items. The existing
validated scales were adjusted to apply in the context of inclusive finance, where
applicable. The scale of knowledge of inclusive finance included six items adopted from
Jebarajakirthy and Lobo (2014). Perceived benefits of ordering finance were measured by
using five items adopted from Jebarajakirthy and Lobo (2014); five items of perceived
risks of ordering finance were adopted from Jacoby and Kaplan (1972), Anita et al. (2008);
and three items measuring financing intentions of inclusive finance were adapted from
Weisberg et al. (2011) and Luarn and Lin (2005). The modifying factors of social enterprise
and digital finance were measured through self-adopted items based on literature review.
Demographic data were collected through fixed-alternative questions. The majority
of the scales mentioned above were adopted from literature related to inclusive finance.
A seven-point Likert-type scale was applied in this study, in which the score of 1 stood for
“strongly disagree” and the score of 7 stood for “strongly agree.” The questionnaire was
divided into four sections, with a total of 27 questions; personal information, knowledge of
inclusive finance, perceived benefits of ordering finance, perceived risks of ordering
finance, consumers’ intention of inclusive finance were comprised in four sections.
4.2.2 Survey implementation procedure. The field survey was conducted for both
geographic and demographic purposes, within a convenience and short-time scale experiment
(Malhotra, 2009). Research data were collected over a 3-week period, from 1 June to 22th June
2018. The questionnaire was conducted by researchers (see Appendix). The detailed
requirements were explained thoroughly by face-to-face approach. The demographic profile of
the sample included five major aspects: gender, age, monthly income, living area and
educational factor.

5. Data analysis and discussion

In order to analyze the data collected by questionnaire, three techniques were applied. First
of all, the reliability test was applied to examine the reliability and validity of items from
each construct. Second, one simple t-test and correlation analysis were applied to test means
IMDS and standard deviations of each item. Third, multiple regression analysis model and
119,7 hierarchy regression analysis were applied to test the first five hypotheses except modifying
factors. Also, BOOTSTRAP method was applied, which is considered to be an effective
method to test the mediating effect of the construct in complicated situations (Preacher and
Hayes, 2004; Wen and Ye, 2014). BOOTSTRAP method includes the examination of multiple
mediating variables, multistage of multiple variables, moderated intermediary variables and
1548 mediating variables of intermediary variables. This method was mainly used to test the
moderating effect of digital finance on perceived benefits and perceived risks.

5.1 Construct validity and reliability

In this research paper, SPSS19.0 software was applied to process collected data. A reliability
test was conducted to check the internal consistency of the scales. Cronbach’s α is the
measurement of the internal consistency of the various constructs (Malhotra, 2009).
The reliability test results indicated that the Cronbach’s α value of research constructs was
above 0.70, which indicated high internal consistency of all variables, and it was doable to
measure the essential features of the concepts steadily and reliably. In this paper, the scale
validity was assessed from the content and the structure. The main variables and their
measurement items were adopted from reliable literature source and further adjusted and
revised based on the combination of empirical rural field research and suggestions given by
the experts from the research field. Thus, the questionnaire has a high content validity.
The results of confirmatory factor analysis showed that the factor loads of each item in
the main variables were higher than 0.6, indicating that the scale had good structural
validity. The results of statistical description, reliability and validity analysis of variables
are presented in Table I.

5.2 Respondent profile analysis

Table II presents the key demographic data of the respondents. The demographic
profiles of the sample were representative, since a total of 160 questionnaires were
collected from target respondents. The gender ratio was balanced. From the age
perspective, 48.13 percent of participants were 18‒30-years old. Moreover, the sampling
covered target consumers of inclusive finance in every age group. As for family income,
30.63 percent of respondents’ income level was lower than 1,000 Yuan, and 33.13 percent
of respondents were earning a monthly income between 1,000 to 3,000 Yuan. In terms of
living area, 46.88 percent of the respondents were living in the countryside, and
20.63 percent of the respondents were from the county, 32.5 percent of the respondents
were living in subcity. In terms of educational level, 44.38 percent of the respondents
had a bachelor degree or above. Other respondents’ educational levels were below than
bachelor degree.
It reveals that there is a significant negative correlation between gender and perceived
risks of ordering finance; it indicates that male farmers have greater risk perceptions when
it comes to making inclusive financial decisions, whereas female farmers experience less
risk perceptions. In rural areas, family’s financial services are mainly applied by the head of
the household, who usually happens to be the male member of the family. Therefore, it is
possible that male as the major financial decision maker might experience greater risk
perceptions than female. Moreover, there is a significant negative correlation between living
area and intentions of inclusive finance. The farmers from countryside have stronger
intentions than other counterparts. It is possible that farmers from countryside have a
greater demand‒supply gap in terms of financial access; therefore, they have a greater
desire for applying inclusive finance, as it can provide affordable financial services for
farmers to promote the development of agriculture.
 Load of Cronbach’s
Analysis of
Construct Means SD factors α rural farmers’
financing
Knowledge of inclusive finance
I cannot afford the interest of inclusive financial loan 3.69 1.991 0.627 0.805 intention
I need to pay the service fee to apply for inclusive financial loan 3.68 2.048 0.654
I need to wait for a long time to get inclusive financial loan 3.81 1.774 0.873
I know the concept of inclusive finance 5.02 2.339 0.951 1549
I know where to apply for inclusive financial loans 4.86 2.44 0.907
I know the application terms of inclusive financial loan 4.34 2.29 0.903
Social enterprise embeddedness
I think there is a high level of engagement and frequent communication
between farmers and social enterprises 5.03 2.339 0.956 0.954
I think there is a high level of mutual trust between farmers and social
enterprises 4.86 2.44 0.952
I think farmers can get agricultural techniques and market-related
information from social enterprises 4.34 2.29 0.899
I think farmers can get inclusive financial loan through social enterprises 4.79 1.956 0.958
Perceived benefits of ordering finance
I think ordering finance can help me to expand my production scale 5.76 1.845 0.764 0.757
I think ordering finance can help me to expand the potential market 4.78 2.088 0.737
I can develop the industry by applying inclusive financial loan and
enhance my family income by extension 4.93 1.885 0.408
I can develop the industry by applying inclusive financial loan, and I can
improve my standard of living by extension 5.02 1.827 0.809
I think inclusive finance service quality has a positive impact on my
allocation of loan to increase revenue 5.17 1.82 0.828
Perceived risks of ordering finance
If I fail to repay the loan on time, my credit will be effected 3.79 2.016 0.924 0.958
If I fail to repay the loan on time, I will experience shame and embarrassment 3.79 1.989 0.904
If others know about my inclusive finance application, they will think
I am poor 3.9 2.013 0.935
If I fail to repay the loan on time, me and my family members will be
consistently bothered by loan officer 3.84 2.092 0.971
If I fail to repay the loan on time, I will have to sell or pawn my existing
assets to repay my loan 3.82 1.793 0.892
Digital finance
I know the digital payment methods like Alipay/WeChat payment/online
banking/credit card 5.07 1.877 0.865 0.901
I have experience of using Alipay/WeChat payment/online banking/
credit card and other digital financial tools 4.78 2.088 0.981
I know the application terms of digital finance like financing tools/
investment/money transfer/settlement of transaction 4.6 2.184 0.896
Financing intention of inclusive finance
I have intention to apply for inclusive financial loan within the next
12 months 4.22 1.926 0.846 0.845 Table I.
I would recommend my relatives, friends to obtain inclusive financial loan 4.69 1.961 0.856 Summary of
I have intention to expand my application of inclusive financial loan 4.71 2.05 0.918 Cronbach’s α

5.3 Correlation analysis

Table III shows the mean, standard deviation and Pearson’s correlation coefficient between
each pair of variables in this study. The size of the value of the correlation coefficient
determines the strength of the relationship between variables (Malhotra, 2012). The negative
IMDS Category n % Coding
119,7
Gender
Male 80 50 1
Female 80 50 2
Age
1550 0–18 9 5.63 1
18–30 77 48.13 2
31–40 29 18.13 3
41–50 29 18.13 4
51–60 14 8.75 5
Above 60 2 1.25 6
Family income (yuan)
Under 1,000 49 30.63 1
1,000–3,000 53 33.13 2
3,000–5,000 32 20 3
5,000–7,000 14 8.75 4
7,000–9,000 9 5.63 5
Above 9,000 3 1.88 6
Area of living
Countryside 75 46.88 1
County 33 20.63 2
Subcity 52 32.5 3
Educational level
Primary school or below 27 16.88 1
Junior middle school 28 17.5 2
High school 17 10.63 3
Table II. Junior college 17 10.63 4
Respondents’ profile Bachelor degree or above 71 44.38 5

correlations indicate a reverse relation between two variables (Malhotra, 2009). It can be seen
that there is a significant correlation between the explanatory variables of intention of inclusive
finance and the explanatory variables of knowledge of inclusive finance and perceived benefits
and perceived risks of ordering finance. However, Pearson correlation analysis presents the
possible correlations between two variables, and the results cannot directly indicate the
existence of causality. Besides, Pearson correlation analysis cannot examine the significant
correlation of two variables accurately when multiple explanatory variables coexist in a model.
Thus, it is essential to further analyze the collected data and test the hypothesis of the study.
In addition, the correlation coefficients between the main explanatory variables are lower than
0.5, and according to VIF test, the variance expansion factor is less than 10. Therefore, there is
no multicollinearity between the main explanatory variables.

5.4 Data analysis

The model contains multiple explanatory variables, including both direct and indirect
effects composed of moderating and mediating effects. Therefore, in this paper, multiple
hierarchical regression analysis was applied to test the conceptual model. The verification
analysis was divided into three major steps: the first step verified the impact of the
knowledge of inclusive finance on perceived benefits and perceived risk of ordering finance,
and the moderating role of social enterprise embeddedness.
Second, the mediated moderation financing intention of inclusive finance integrated in
the first part of the model was examined to analyze the knowledge of inclusive finance’s
Variables 1 2 3 4 5 6 7 8 9 10

1. Financing intention of inclusive finance 1

2. The knowledge of inclusive finance 0.234*** 1
3. Perceived benefits 0.618*** 0.283*** 1
4. Perceived risks −0.050 0.164** 0.034 1
5. Social Enterprise embeddedness 0.066 0.301*** 0.143* 0.069 1
6. Digital finance 0.429*** 0.277*** 0.440*** 0.030 0.090 1
7. Gender 0.022 0.126 0.017 −0.200*** 0.118 −0.078 1
8. Age −0.004 −0.386*** −0.039 −0.187** −0.414*** 0.033 0.033 1
9. Monthly income −0.057 0.153* 0.000 −0.171** 0.302*** −0.058 0.050 0.231*** 1
10. Living area −0.156 0.373*** −0.068 −0.027 0.444*** −0.155** 0.192** −0.178** 0.262*** 1
11. Educational level −0.135* 0.544*** −0.053 0.079 0.644*** −0.155* 0.138* −0.437*** 0.193** 0.603***
Notes: *,**,***Correlation is significant at p o0.1, p o 0.05, p o0.01, respectively
intention
financing

1551
rural farmers’
Analysis of

Correlation matrix

and descriptive
Table III.

for constructs

information
IMDS impact on the financing intention of inclusive finance. Further, it examined whether
119,7 the moderating impact of social enterprise embeddedness was mediated by the
intermediaries, namely, perceived benefits and perceived risks. Third, the verification of
moderated mediation was performed. Therefore, BOOTSTRAP was applied to testify
whether perceived benefits and perceived risks of ordering finance were moderated by
digital finance in the mediation between the knowledge of inclusive finance and financing
1552 intention of inclusive finance.
The first step was to test the impact of the knowledge of inclusive finance on perceived
benefits and perceived risks of ordering finance, and the moderating effect of social
enterprise embeddedness. The regression results are shown in Table IV.
In total, three models were designed, and perceived benefits and perceived risk of
ordering finance were identified as explanatory variables. Model 1 included all control
variables and main explanatory variables. Model 2 was based on the integration of Model 1
and modifying variables. Model 3 was designed on the basis of the integration of Model 2
and other interactive items. Model 1 indicated that there was a significant positive
correlation between the knowledge of inclusive finance and perceived benefits and risks of
ordering finance. This positive correlations indicated that the farmer household with more
knowledge of inclusive finance tended to have greater benefits and risk perception of their
financial decision. Model 3a showed a positive coefficient correlation between the knowledge
of inclusive finance and social enterprise embeddedness, which indicated the cooperation of
social enterprise strength farmer households’ perceived benefits of ordering finance.
However, Model 3b showed a significant negative correlation between the knowledge of
inclusive finance and social enterprise embeddedness, which indicated the mediating impact
of social enterprise embeddedness. However, such mediating impact of social enterprise
decreased farmer households’ risk perceptions of ordering finance.
The second step was to verify the mediating impact of perceived benefits and perceived
risks of ordering finance. The regression results are shown in Table V.
In total, four models were designed to examine the intention of inclusive finance as the
explained variable. Model 4 included all control variables and major explanatory variables.
Model 5 added social enterprise embeddedness as a moderating variable on the basis
of Model 4. Model 6 added perceived benefits of ordering finance on the basis of Model 5.
Model 7 added perceived risks ordering finance on the basis of Model 5. Model 4 showed
significant positive correlations between the knowledge of inclusive finance and the
intention of inclusive finance, which indicated that the farmers’ financing intention of
inclusive finance would be enhanced with their increasing knowledge level of inclusive
finance. In both Models 6 and 7, the coefficients of perceived benefits and perceived risks
were significant, which meant that the correlation between knowledge of inclusive finance
and financing intention of inclusive finance was mediated by perceived benefits and
perceived risks.
However, the impact of the knowledge of inclusive finance was not significant in Model 6,
which indicated that perceived benefits had an absolute mediating impact. Model 7 illustrated
a significance of the knowledge of inclusive finance, which indicated the significant
meditating role of perceived risks of ordering finance as an intermediary variable.
The third step was to verify the moderating role of digital finance. BOOTSTRAP was
applied for further examination of the collected data with the sample size of 5,000 and
90 percent confidence interval; perceived benefits of ordering finance had a mediating impact
on the correlation between social enterprise embeddedness and the financing intention of
inclusive finance. This result further confirmed the related conclusion from the second step.
Meanwhile, the moderating impact of digital finance on perceived benefits of ordering
finance’s mediating role was tested by, respectively, adding and subtracting standard
deviation on the mean value to distinguish three levels of digital finance application.
 Perceived benefits Perceived risks
Variables Model 1a Model 2a Model 3a Model 1b Model 2b Model 3b

Gender 0.026 (0.125) 0.018 (0.088) 0.027 (0.135) −0.769*** (−2.697) −0.778*** (−2.756) −0.789*** (2.830)
Age 0.019 (0.173) −0.039 (−0.355) −0.058 (−0.539) −0.092 (−0.617) −0.159 (−1.048) −0.137 (0.911)
Monthly income −0.004 (−0.048) 0.093 (0.959) 0.081 (0.857) −0.250** (−2.021) −0.138 (−1.027) −0.124 (0.938)
Living area −0.139 (−0.944) −0.133 (−0.918) −0.159 (−1.113) −0.099 (−0.486) −0.092 (−0.456) −0.062 (0.311)
Educational level −0.202** (−2.145) −0.136 (−1.403) −0.096 (−0.997) 0.048 (0.369) 0.125 (0.932) 0.078 (0.582)
Knowledge of inclusive finance 0.397*** (4.902) 0.730*** (4.641) 0.147 (0.552) 0.226** (2.020) 0.611*** (2.797) 1.298*** (3.479)
Social enterprise embeddedness −0.308** (−2.455) −0.629*** (−3.662) −0.356** (−2.044) 0.023 (0.097)
Knowledge of Inclusive finance×social
enterprise embeddedness 0.104*** (2.667) −0.123** (−2.256)
F-value 4.372*** 4.731*** 5.203*** 3.403*** 3.574*** 3.848***
R2 0.146 0.179 0.216 0.118 0.141 0.169
Notes: *,**,***Correlation is significant at p o0.1, p o 0.05, p o0.01, respectively
intention
financing

1553
rural farmers’
Analysis of

Table IV.

and correlation matrix

Descriptive statistics

for intermediary
variables
IMDS Financing intention of inclusive finance
119,7 Variables Model 4 Model 5 Model 6 Model 7

Gender 0.110 (0.420) 0.105 (0.414) 0.087 (0.401) 0.001 (0.005)

Age 0.053 （0.384) −0.052 (−0.383) −0.014 (−0.121) −0.070 (−0.517)
Monthly income −0.062 (−0.544) 0.080 (0.668) 0.027 (0.259) 0.064 (0.535)
Living area −0.294 (−1.575) −0.305* (−1.684) −0.200 (−1.284) −0.313* (−1.741)
1554 Educational level −0.297** (−2.491) −0.161 (−1.325) −0.098 (0.939) −0.151 (1.249)
Knowledge of inclusive
finance 0.511*** (4.996) 0.565* (1.670) 0.468 (1.612) 0.735** (2.106)
Social enterprise
embeddedness −0.735*** (−3.373) −0.320 (1.644) −0.732*** (−3.383)
Knowledge of inclusive
finance×social enterprise
embeddedness 0.083* (1.682) 0.014 (0.332) 0.067 (1.344)
Perceived benefits 0.659*** (7.449)
Table V.
Descriptive statistics Perceived risks −0.131* (−1.788)
and regression matrix F-value 5.208 5.702 13.063 5.498
for independent R2 0.17 0.232 0.439 0.248
variable Notes: *,**,***Correlation is significant at po 0.1, p o0.05, po 0.01, respectively

Also, different levels of digital financial utilization were identified accordingly. The
BOOTSTRAP test result showed positive coefficient, and the confidence interval was (0.0006,
0.0221), and the confidence interval did not contain 0, indicating the significant modifying
impact of digital finance. In other words, the digital finance had a significant moderating
impact on the correlation between the knowledge of inclusive finance, perceived benefits of
ordering finance and financing intention of inclusive finance. The BOOTSTRAP was applied
to examine the moderating impact of digital finance on perceived risks’ meditating impact as
an intermediary variable. Same methods were applied to examine the moderating impact of
digital finance on perceived benefits’ mediating impact. However, the moderating impact
of digital finance was not significant on the correlation between perceived risks and financing
intention, which indicated that the correlation between perceived risks of ordering finance and
financing intention of inclusive finance was not significantly moderated by digital finance.
One possible explanation is that digital finance has not been largely applied in rural areas.
Moreover, digital finance provides medium of payment and it still has safety supervision
concerns. The main role of digital payment is to improve time efficiency and decrease
transaction cost. Due to the incomplete digital finance regulations, it is reasonable that digital
finance cannot decrease consumers’ risk perceptions. At the same time, with the impact of
social enterprise embeddedness, perceived risk’s adverse impact on financing intention has
already weakened. Therefore, there is a limited room for digital finance to further weaken the
association between perceived risks and financing intention.

5.5 Discussion
5.5.1 Implications for theory. This research has developed an empirical inclusive finance
model to investigate and to enhance farmers’ attitudes toward ordering finance and to also
strengthen financing intentions of inclusive finance in rural areas of China. This research
study proposes two types of attitudes toward ordering finance, namely, perceived benefits
and perceived risks. In particular, it analyzed relevant data and provided insights into
farmers’ perception of inclusive financial service in rural areas of China. Also, perceived
risks of ordering finance hinder consumers from participating in inclusive finance,
whereas perceived benefits urge consumers to actively participate in inclusive financial
activities. Therefore, these constructs and its indicators can be applied readily to other
segments of the BOP market. This study also proposes a mechanism for transforming Analysis of
rural framers’ attitudes and intentions of inclusive finance. rural farmers’
This conceptual model would potentially contribute to researchers interested in financing
investigating the financing intention of inclusive financial services relating to rural
population and the other consumers from the BOP market. Additionally, this model also
intention
can be applied to other financial products as they share similar characteristics. The
integration of social enterprise embeddedness and digital finance as modifying factors is 1555
the uniqueness of this research conceptual model. It provides a revolutionized prospective
for researchers to get new consumer insights in today’s digital age. In this conceptual
model, knowledge of inclusive finance is considered to be an antecedent, and social
enterprise embeddedness and digital finance are considered to be moderators of other
variables. Thus, the participation of social enterprise can enhance farmers’ favorable
attitudes toward ordering finance and can significantly stimulate their financing
intentions in rural areas. Additionally, digital finance can significantly strengthen the
positive correlation between perceived benefits of ordering finance and financing
intention of inclusive finance.
Therefore, these constructs can be incorporated into the research model, which would
better predict attitudes and intentions about inclusive finance in the context of the BOP
market. The further discussion suggests that this conceptual model potentially contributes
to research relating to behavioral finance, which has gained popularity among researchers
and managers in the marketing field in recent years. Furthermore, by incorporating social
enterprise and digital finance service, this model can be utilized for other financial products
and services used by consumers from other market segments.
5.5.2 Implications for practice. The results revealed a significant correlation between the
financing intention of inclusive finance and the explanatory variables of knowledge of
inclusive finance and perceived benefits and perceived risks of ordering finance. Knowledge
enhancement of perceived benefits was consistent with the previous findings of scholars
(Brau and Woller, 2004; Turvey and Kong, 2010). However, the knowledge of inclusive
finance also increased farmers’ risk perception of ordering finance. This finding was
inconsistent with (Li et al., 2011; Jebarajakirthy et al., 2015). One possible explanation is that
farmers who understand components of inclusive finance will evaluate the unfavorable
consequences of a financial decision thoroughly and tend to make more rational decisions,
as compared to their counterparts. Furthermore, farmers with sufficient knowledge about
inclusive finance might experience blissful ignorance effect, which means they know the
performance of their financial product and it is not easy for them to rationalize the
risk-related outcomes (Michael, 2009).
It is interesting to note that in rural China, modifying construct social enterprise
embeddedness can strengthen the perceived benefits of ordering finance and decrease
farmers’ risk perceptions. The mediating impact of social enterprise embeddedness has a
significant negative impact on knowledge of inclusive finance. It is possible that social
enterprises provide farmers a holistic financial service that shortens unnecessary and
reluctant procedures to improve the time efficiency and decrease transaction cost. Thus, it
is reasonable to make an assumption that the decrease of transaction cost might lead to
less concerns toward perceived risks of ordering finance. China has made significant
contributions in terms of implementation of agricultural development and social
enterprises. However, the cooperation between social enterprise and inclusive financial
institutions is not very common in rural China due to the functional differences between
two organizations. The study provides possible mechanism for them to cooperate at both
institutional and functional levels. Further, the inclusive financial institution managers
should consider implementing digital financial service, as it can help them to improve
IMDS accessibility, efficiency and quality of their services. The local government should set up
119,7 related regulations that can accelerate the process of co-operations between social
enterprise and inclusive financial institutions.

6. Conclusion
The inclusive finance is important for improving the living standards and eliminating the
1556 poverty of poor farmers, since they need financial access to improve agricultural
productions and further develop entrepreneurial activities in the rural areas. However, the
development of inclusive finance in rural China is full of various challenges along with
opportunities, as it has not gained enough popularity due to the misconceptions from
different stakeholders. Especially, the lack of inclusive financial knowledge and irrational
risk perceptions prevent farmers to get access from inclusive financial institutions.
Therefore, this paper investigates rural farmers’ level of inclusive financial knowledge,
perceived benefits and risk perceptions on obtaining ordering finance and financing
intention of inclusive finance in the context of BOP market in rural China.
The study formed an inclusive financial intention model on the basis of the TPB and the
theory of BOP. In this model, two driving factors, social enterprise embeddedness and
digital finance, emerge as moderating factors. To analyze the collected data, one sample
t-test, multiple regression analysis, and BOOTSTRAP method are applied. The results of
this paper can be directly applied by inclusive financial institutions, related government
departments and social enterprises to optimize inclusive financial services.
Our study contributes via the following findings: it suggests that knowledge of inclusive
finance has a positive significant impact on the perceived benefits of ordering finance and
intention of inclusive finance, which indicates that a systematic inclusive finance educational
project is needed to enhance rural farmers’ understanding of inclusive finance and its
components in BOP market in China to form farmers’ positive attitudes and to further
stimulate their financing intentions. Moreover, the social enterprise embeddedness has a
positive significant moderating impact on the associations between knowledge of inclusive
finance and the perceived benefits, and it can significantly reduce risk perception. The result
reveals that it is crucial to promote holistic social enterprise organizations to develop inclusive
finance in the BOP market of China, as the service attributes of social enterprise can fill the
existing supply‒demand gap and improve the service accessibility, quality and efficiency.
Furthermore, perceived benefits of ordering finance can positively enhance rural farmers’
financing intention of inclusive finance, whereas perceived risks can negatively influence
the financing intention. Thus, it is crucial to leverage the service attributes of both inclusive
financial institutions and social enterprise to adjust farmers’ risk perceptions and
improve perceived benefits. Finally, digital finance as a modifying factor can significantly
strengthen the positive correlations between perceived benefits of ordering finance and
financing intention of inclusive finance. Therefore, the efficiency of digital finance can greatly
reduce the existing transaction cost of rural farmers and generate positive perceptions toward
inclusive financial services.
In addition, with the development of information technology and the acceleration of
digital process, the digital finance innovation model has further stimulated the potential of
social enterprises to play a greater role in poverty alleviation. It should be noted that, in spite
of those successful social enterprise models, a further development of internal and external
environment of social enterprises is of the essence. Hence, it is necessary to improve the
institutional basis for developing social enterprises, which includes clarifying the legal
status of social enterprises, strengthening the government’s policy support and establishing
the evaluation system for social enterprises. Moreover, the government needs an innovative
management system by streamlining government departments and delegating authorities,
reducing reluctant positions and enhancing benefits. It is also crucial to reduce entry
barriers for social enterprises and broaden business distributions to further promote the Analysis of
social enterprise participation. rural farmers’
Additionally, cultivating the public welfare mindset of social enterprise managers can financing
accelerate the pace of social enterprise participation. To enable social enterprises to fulfill
their social missions, it is important for their policymakers to strengthen their altruistic intention
values. One possible way is by organizing specific training for managers, which involves
public welfare activities in the agricultural chain with corresponding performance appraisal 1557
standards. Public welfare mindset-oriented training can enhance managers’ real
understanding of public welfare concept and can strengthen farmers’ knowledge of social
enterprise and inclusive finance by extension. Thus, it is essential for social enterprise,
government, higher educational institutions and research institutes to accelerate their
cooperation to continuously gain farmers’ trust and the recognition of social enterprises’
value. Furthermore, adopting digital finance is crucial in terms of enhancing the service
quality by reducing overall cost of rural farmers. The limited educational level of rural
farmers encourages inclusive financial institutions to consistently revolutionize their digital
finance services to best serve the need of the consumers from the BOP market. Therefore, it
is essential to integrate the design of inclusive financial products and digital finance to meet
rural farmers’ financial need in today’s digital world.

7. Limitation and future research direction

First, due to the time and budget constraint, a limited amount of empirical data was
collected, and the size of the sample might be relatively insufficient to generalize China rural
farmers’ perception and intention of inclusive finance. Besides, the age ratio of this study is
unbalanced, as considerable amount of data were collected from youths aged between 18
and 30. Also, age was not proposed as a control variable. Thus, there is a need for further
studies to increase the size of the sample and balance respondents’ age ratios to conclude the
socio-demographic findings. Moreover, the survey was collected by the researcher, who
might have inevitably influenced the respondents’ response to some extent. This study
investigates rural farmers’ attitude and intentions at one point in time; therefore, testing of
the conceptual model with longitudinal data is essential in the future research.
Second, a unique conceptual model was proposed on the basis of the context of BOP
market in China. Researchers can apply the conceptual model to test rural farmers’ intention
of inclusive finance after providing innovative agricultural and entrepreneurial projects.
In future studies, innovation-related factors could be incorporated into the conceptual model.
Besides, the impact of inclusive finance on the performance of farmers’ agricultural orders
can be further examined by conducting multiple empirical researches.
Third, the measurement of embeddedness of social enterprise and digital finance were
developed by applying simple methods, which can be further investigated and improved.
Previous research suggested that the embeddedness of social enterprise factors include two
types, namely, relational embeddedness and structural embeddedness (Xie et al., 2017).
Thus, it is essential to develop more detailed and comprehensive scales and dimensions for
measuring the modifying impact of both relational embeddedness and structural
embeddedness of social enterprises.

References
Ajzen, I. (1991), “The theory of planned behavior”, Organizational Behavior and Human Decision
Processes, Vol. 50 No. 2, pp. 179-211.
Ajzen, I. (1996), “The social psychology of decision making”, in Higgins, E.T. and Kruglanski, A.W.
(Eds), Social Psychology: Handbook of Basic Principles, Guilford Press, New York, pp. 297-325.
IMDS Albert, P. and Ren, C. (2001), “Microfinance with Chinese characteristics”, World Development,
119,7 Vol. 29 No. 1, pp. 39-62, available at: http://0-dx.doi.org.wam.leeds.ac.uk/10.1016/S0305-750
X(00)00087-5
Anderson, J.L., Markides, C. and Kupp, M. (2010), “The last frontier: market creation in conflict zones,
deep rural areas, and urban slums”, California Management Review, Vol. 52 No. 4, pp. 6-28.
Anita, L.Z., Hanmer-Lloyd, S., Ward, P. and Goode, M.M.H. (2008), “Perceived risk and Chinese
1558 consumers’ internet banking services adoption”, The International Journal of Bank Marketing,
Vol. 26 No. 7, pp. 505-525.
Antonides, G., Groot, I.M. and Raaij, W.F. (2011), “Mental budgeting and the management of household
finance”, Journal of Economic Psychology, Vol. 32 No. 4, pp. 546-555.
Ashraf, M.A. and Noor, M.S.I. (2010), “Microfinance customer (borrower) experience towards the
effectiveness of MFIs in Bangladesh: an exploratory analysis”, ABAC Journal, Vol. 30 No. 2,
pp. 12-25.
Barki, E. and Parente, J. (2010), “Consumer behaviour of the base of the pyramid market in Brazil”,
Greener Management International, Vol. 56 No. 13, pp. 11-23.
Brau, J.C. and Woller, G.M. (2004), “Microfinance: a comprehensive review of the existing literature”,
Journal of Entrepreneurial Finance, Vol. 9 No. 1, pp. 1-27.
Bruett, T.C. (2007), “Kiva, and prosper.com: how disintermediation and the internet are changing
microfinance”, Community Development Investment Review, Vol. 3 No. 2, pp. 44-50.
Byerly, R.T. (2014), “The social contract, social enterprise, and business model innovation”, Systematic
Botany, Vol. 4 No. 4, pp. 325-343.
Chan, S. (2010), “Microfinance in China: an overview”, Microfinance Focus, July 28, available at:
www.microfinancefocus.com/news/2010/07/28/microfinance-in-china-an-overview/ (accessed
January 21, 2011).
Chandon, P., Wansink, B. and Laurent, G. (2000), “A benefit congruency framework of sales promotion
effectiveness”, Journal of Marketing, Vol. 64 No. 4, pp. 65-81.
Cheng, E. (2007), “The demand for microcredit as a determinant for microfinance outreach-evidence
from China”, Savings and Development, Vol. 31 No. 3, pp. 307-334.
Chien, Y.W. and Devaney, S.A. (2001), “The effects of credit attitude and socioeconomic factors on
credit card and installment debt”, Journal of Consumer Affairs, Vol. 35 No. 1, pp. 162-179.
Churchill, G.A. and Iacobucci, D. (2002), Marketing Research: Methodological Foundations, Harcourt
College Publishers, Dryden Press, NY.
Dees, J.G. (1998), “Enterprising nonprofits”, Harvard Business Review, Vol. 76 No. 1, pp. 54-69.
Diacon, S. and Ennew, C. (2001), “Consumer perceptions of financial risk. Geneva papers on risk and
insurance”, Issues and Practice, Vol. 26 No. 3, pp. 389-409.
Doherty, B., Haugh, H. and Lyon, F. (2014), “Social enterprises as hybrid organizations: a review
and research agenda”, International Journal of Management Reviews, Vol. 16 No. 4, pp. 417-436.
Epstein, M.J. and Yuthas, K. (2011), “The critical role of trust in microfinance success: identifying
problems and solutions”, Journal of Developmental Entrepreneurship, Vol. 16 No. 4, pp. 477-497.
Gärling, T., Kirchler, E., Lewis, A. and Van Raaij, F. (2009), “Psychology, financial decision making, and
financial crises”, Psychological Science in the Public Interest, Vol. 10 No. 1, pp. 1-47.
Godwin, D.D. (1997), “Dynamics of households’ income, debt, and attitudes toward credit, 1983–1989”,
Journal of Consumer Affairs, Vol. 31 No. 2, pp. 303-325.
Irwin, W. (2014), “Social enterprise – an international phenomenon?”, Institute for Small Business &
Entrepreneurship (ISBE), September 26, available at: www.isbe.org.uk/emsocialenterprise
(accessed September 30, 2014).
Jacoby, J. and Kaplan, L.B. (1972), “The components of perceived risk”, Proceedings of the 3rd Annual
Conference of the Association for Consumer Research, Association for Consumer Research,
Iowa city, Iowa, April 10-11.
Jebarajakirthy, C. and Lobo, A.C. (2014), “War affected youth as consumers of microcredit: an Analysis of
application and extension of the theory of planned behaviour”, Journal of Retailing & Consumer rural farmers’
Services, Vol. 21 No. 3, pp. 239-248.
financing
Jebarajakirthy, C., Lobo, A. and Hewege, C. (2015), “Enhancing youth’s attitudes towards microcredit in
the bottom of the pyramid markets”, International Journal of Consumer Studies, Vol. 39 No. 2, intention
pp. 180-192.
Jose, S., Rugimbana, R. and Gatfield, T. (2012), “Consumer responses to CSR driven microfinance 1559
strategy of banks-an empirical investigation based on India”, International Journal of Business
and Management, Vol. 7 No. 21, pp. 1-14.
Karnani, A.G. (2006), “Fortune at the bottom of the pyramid: a mirage”, Social Science Electronic
Publishing, Vol. 1 No. 2, pp. 85-100.
Kidwell, B. and Turrisi, R. (2004), “An examination of college student money management tendencies”,
Journal of Economic Psychology, Vol. 25 No. 5, pp. 601-616.
Kinsey, J. and McAlister, R. (1981), “Consumer knowledge of the costs of open-end credit”, The Journal
of Consumer Affairs, Vol. 15 No. 2, pp. 249-270.
Koh, F., Phoon, K.F. and Ha, C.D. (2018), “Chapter 15-digital financial inclusion in South East Asia”, in
Chuen, D.L.K. and Deng, R. (Eds), Handbook of Blockchain Digital, Finance & Inclusion, Vol. 2,
Academic Press, pp. 387-403.
Lee, M. (2009), “Factors influencing the adoption of internet banking: an integration of TAM and TPB
with perceived risk and perceived benefit”, Electronic Commerce Research and Applications,
Vol. 8 No. 3, pp. 130-141.
Li, X., Gan, C. and Hu, B. (2011), “Accessibility to microcredit by Chinese rural households”, Journal of
Asian Economics, Vol. 22 No. 3, pp. 235-246.
Liu, Y.X., Qi, L.I. and Business, S.O. (2016), “The development and supervision of internet finance: a
perspective of transaction cost economics”, Journal of Wuxi Institute of Commerce, Vol. 16 No. 2,
pp. 7-12.
Liu, M.T., Brock, J.L., Shi, G.C., Chu, R. and Tseng, T.-H. (2013), “Perceived benefits, perceived risk, and
trust: influences on consumers’ group buying behaviour”, Asia Pacific Journal of Marketing and
Logistics, Vol. 25 No. 2, pp. 225-248.
Luarn, P. and Lin, H.H. (2005), “Toward an understanding of the behavioral intention to use mobile
banking”, Computers in Human Behavior, Vol. 21 No. 6, pp. 873-891.
Malhotra, N.K. (2004), Marketing Research: An Applied Orientation, 4th ed., Prentice Hall, Harlow.
Malhotra, N.K. (2009), Basic Marketing Research: A Decision-Making Approach, 3th ed., Prentice Hall,
Upper Saddle River, NJ.
Malhotra, N.K. (2012), Basic Marketing Research: Integration of Social Media, 4th ed., Pearson
Education, NJ.
Mayer, J.D. and Salovey, P. (1995), “Emotional intelligence and the construction and regulation of
feelings”, Applied and Preventive Psychology, Vol. 4 No. 3, pp. 197-208.
Michael, R.S. (2009), Consumer Behavior: Buying, Having, and Being, 8th ed., Pearson Education, Upper
Saddle River, NJ.
Montano, D.E. and Kasprzyk, D. (2008), “Theory of reasoned action, theory of planned behavior, and
the integrated behavioral model”, in Glanz, K., Rimer, B.K. and Viswanath, K. (Eds), Health
Behavior and Health Education: Theory, Research, and Practice, 4th ed., Jossey-Bass, San
Francisco, CA, pp. 67-96.
Moseley, S.F. and Rogers, E. (2004), “Diffusion of innovations theory: its utility and value in public
health”, Journal of Health Communication, Vol. 9 No. S1, pp. 149-151.
Owens, J. (2013), “Offering digital financial services to promote financial inclusion: lesson we’ve
learned”, Innovation, Vol. 8 Nos 1-2, pp. 279-290.
IMDS Pitta, D.A., Guesalaga, R. and Marshall, P. (2008), “The quest for the fortune at the bottom
119,7 of the pyramid: potential and challenges”, Journal of Consumer Marketing, Vol. 25 No. 7,
pp. 393-401.
Prahalad, C.K. (2010), The Fortune at the Bottom of the Pyramid: Eradicating Poverty through Profits,
Wharton School Publishing, Upper Saddle River, NJ.
Prahalad, C.K. (2012), “Bottom of the pyramid as a source of breakthrough innovations”, Journal of
1560 Product Innovation Management, Vol. 29 No. 1, pp. 6-12.
Preacher, K.J. and Hayes, A.F. (2004), “SPSS and SAS procedures for estimating indirect effects in
simple mediation models”, Behavior Research Methods, Instruments, & Computers, Vol. 36 No. 4,
pp. 717-731.
Radcliffe, D. and Voorhies, R. (2012), Digital Pathway to Financial Inclusion, Bill & Melinda Gates
Foundation, Social Science Electronic Publishing, NY.
Russell, J.A. and Barrett, L.F. (1999), “Core affect, prototypical emotional episodes, and other things
called emotion: dissecting the elephant”, Journal of Personality and Social Psychology, Vol. 76
No. 5, pp. 805-819.
Sharif, N. (2000), “Developing sustainable livelihoods: the case of poor female youth in sub-Saharan
Africa”, Journal of Sustainable Development in Africa, Vol. 2 No. 1, pp. 1-24.
Smith, J.R., Terry, D.J., Manstead, A.S.R., Louis, W.R., Kotterman, D. and Wolfs, J. (2008), “The
attitude-behavior relationship in consumer conduct: the role of norms, past behavior, and
self-identity”, The Journal of Social Psychology, Vol. 148 No. 3, pp. 311-334.
Szajna, B. (2010), “Empirical evaluation of the revised technology acceptance model”, Management
Science, Vol. 42 No. 1, pp. 85-92, (Seminar participation. Journal of Family and Economic
Issues, 29, 23–40).
Tang, L. (2009), “Study on China’s sustainable development of rural micro-credit”, International Journal
of Business and Management, Vol. 4 No. 8, pp. 128-131.
Turvey, C.G. and Kong, R. (2010), “Informal lending amongst friends and relatives: can microcredit
compete in rural China?”, China Economic Review, Vol. 21 No. 4, pp. 544-556.
Varghese, A. (2001), “Women and microcredit in rural Bangladesh: an anthropological study of
Grameen bank lending: women and microcredit in rural Bangladesh: an anthropological study
of Grameen bank lending”, American Anthropologist., Vol. 103 No. 1, p. 258.
Wang, H., Alon, I. and Kimble, C. (2015), “Dialogue in the dark: shedding light on the development of
social enterprises in China”, Global Business & Organizational Excellence, Vol. 34 No. 4,
pp. 60-69.
Wang, Y. (2004), “Financing difficulties and structural characteristics of SMEs in China”, China &
World Economy, Vol. 12 No. 2, pp. 34-49.
Weber, E.U. (2004), “Perception matters: psychophysics for economists”, in Carrillo, J. and Brocas, I.
(Eds), The Psychology of Economic Decisions, Vol. 2, Oxford University Press, pp. 163-176.
Weisberg, J., Te’Eni, D. and Arman, L. (2011), “Past purchase and intention to purchase in e-commerce:
the mediation of social presence and trust”, Internet Research, Vol. 21 No. 1, pp. 82-96.
Wen, Z. and Ye, B. (2014), “Intermediate effect analysis: methodological and model development”,
Progress in Psychological Science, Vol. 22 No. 5, pp. 731-745.
Wickramasinghe, V. and Gurugamage, A. (2012), “Effects of social demographic attributes, knowledge
about credit cards and perceived lifestyle outcomes on credit card usage”, International Journal
of Consumer Studies, Vol. 36 No. 1, pp. 80-89.
WMFG (2010), “Microfinance in China”, WMFG, March, Geneva, available at: www.microfinancegateway.
org/sites/default/files/mfg-en-paper-microfinance-in-china-mar-2010.pdf
Xie, J., Liu, L., Liang, L., et al. (2017), “Coordination for value networks of agricultural social
enterprises: an empirical analysis from the perspective of social embeddedness”, Journal of
Finance & Economics, Vol. 43 No. 10, pp. 83-96.
Zhang, W., Li, H., Ishida, S. and Park, E. (2010), “China’s non-governmental microcredit practice: Analysis of
history and challenges”, Journal of Family and Economic Issues, Vol. 31 No. 3, pp. 280-296. rural farmers’
Zhu, L.Y. and Meeks, C.B. (1994), “Effects of low income families’ ability and willingness to use financing
consumer credit on subsequent outstanding credit balances”, The Journal of Consumer Affairs,
Vol. 28 No. 2, pp. 403-422. intention

1561
Further reading
World Bank (2001), “World development report 2000/2001: attacking poverty”, Oxford University
Press, ©World Bank, New York, NY, available at: https://openknowledge.worldbank.org/
handle/10986/11856License: CCBY3.0IGO (accessed November 23, 2000).

Appendix. Survey questionnaire

IMDS
119,7

1562
 Analysis of
rural farmers’
financing
intention

1563

Corresponding author
Jiaping Xie can be contacted at: jiaping@sufe.edu.cn

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com