Grincoh WP 3.04 Radosevic Yoruk

Working Paper Series
Serie 3
Knowledge, Innovation, Technology
Paper No. 3.04
A New Metrics of Technology Upgrading: The

Central And East European Countries in a
Comparative Perspective
Slavo Radošević*, Esin Yoruk*
* University College London
2015
www.grincoh.eu
This paper was funded under the FP7 project “Growth– Innovation – Competitiveness: Fostering Cohesion
in Central and Eastern Europe (GRINCOH)” under the Programme SSH.2011.2.2-1: Addressing
cohesion challenges in Central and Eastern Europe; Area 8.2.2 Regional, territorial and social cohesion.
Project Nr. 290657
Slavo Radošević s.radosevic@ucl.ac.uk
Esin Yoruk e.yoruk@ssees.ucl.ac.uk
University College London, School of Slavonic and East European Studies
http://www.ucl.ac.uk/ssees
Please cite as:

Radošević S., Yoruk A, (2015), ‘A New Metrics of Technology Upgrading: The Central And East
European Countries in a Comparative Perspective’, GRINCOH Working Paper Series, Paper No. 3.04
A New Metrics of Technology Upgrading: The Central And East European

Countries in a Comparative Perspective
Abstract
We explore the issues of measurement of technology upgrading of the economies moving from middle to high
income status. In particular, our focus is on the Central and Eastern European economies (CEE) within the
context of sample of 42 economies. Central and Eastern European countries (CEECs) are largely middle-
income economies but it is not certain whether they have achieved a threshold of technological capability
required for catching up to high-income economies status. In exploring this issue we apply theoretically
relevant and empirically grounded middle level conceptual and statistical framework based on three
dimensions: intensity and breadth of technological upgrading, and technology and knowledge exchange. As an
outcome we construct composite indicator of technology upgrading based on 35 indicators which reflects
different drivers and patterns of technology upgrading of countries at different income levels. Based on simple
statistical analysis we show that the middle income trap is present in all dimensions of technology upgrading
but their importances vary across different dimensions. A trap seems to be higher for dimensions of ‘breadth’
of technology upgrading than for index of ‘intensity’ of technology upgrading. We explore in detail positioning
of the CEE economies in the broader comparative context
Content
1. Introduction ........................................................................................................................... 2
2. A framework for measuring technology upgrading: a conceptual approach........................ 4
2.1. Intensity and types of technology upgrading (scale) ................................................................... 5
2.2. Breadth of technology upgrading: structural change, infrastructure and firms’ structure
(scope)................................................................................................................................................. 8
2.3. Interaction with global economy and technology upgrading .................................................... 11
3. Why and how composite indicator of technology upgrading? ........................................... 14
3.1 Data ............................................................................................................................................. 15
3.2 Developing measures by composite index methodology ........................................................... 16
4. Analysis of levels and profiles of technology upgrading ..................................................... 18
4.1. Probing the relationships between technology upgrading and different income groups......... 18
4.2. Ranking of countries based on index of technology upgrading and index of technology and
knowledge exchange ........................................................................................................................ 30
5. Conclusions .......................................................................................................................... 52
6. References ........................................................................................................................... 57
APPENDIX A .............................................................................................................................. 65
APPENDIX B Explanations for Composite Index Measures ...................................................... 65
APPENDIX C .............................................................................................................................. 69
APPENDIX D .............................................................................................................................. 70
APPENDIX E .............................................................................................................................. 72
APPENDIX F .............................................................................................................................. 73
1
In this report we explore the issues related to measurement of technology upgrading of the
economies moving from middle to high income status. In particular, our focus is on the central and
eastern European economies (CEE) within the context of sample of 42 economies ranging from
lower middle income to upper high income level economies.
1. Introduction
In this report we explore the issues related to measurement of technology upgrading of the
economies moving from middle to high income status. In particular, our focus is on the central and
eastern European economies (CEE) within the context of sample of 42 economies ranging from
lower middle income to upper high income level economies.
The EU has been for quite long period hailed as the ‘convergence machine’ (World Bank, 2012). With
the onset of Euro crisis of 2008 this description has become if not questioned but more seen in
differentiated light. Namely, given increasing differences in growth across the EU it seems that there
are issues with the mechanism of convergence. Underlying these issues are different views of what
lies behind increasing differences in growth across the EU. The EU28 is diverse grouping of
economies not only in terms of incomes per capita but even more in terms of technological positions
of countries in relation to technology frontier. This means that the EU28 is faced with diversity of
drivers of growth, which requires country or group of countries specific policies.
Central and Eastern European countries (CEECs) are largely middle-income economies but it is not
certain whether they have achieved a threshold of technological capability required for catching up
to high-income economies status.1 The shift from middle income to high-income is not guaranteed
or is not automatic as growth process is usually non-linear and evolves across several threshold
levels with their specific threshold requirements. In order to understand this process we need to be
open to a variety of historical experiences as well as go beyond simple explanations of growth be
they adequate institutions (Acemoglu and Robinson 2008), human capital (Glaeser et al. 2004) or
Research and Development (R&D) (OECD 2004).
There seems to be increasing differences in growth between the EU core economies and periphery
as well as polarisation in growth rates among the EU new member states (NMS). Within the EU NMS,
central Europe or Visegrad economies (Poland, Czech R, Slovakia, and Hungary) have expanded at
much higher rate than other NMS. This sub-region has exhibited features of catch-up which is the
most visible after 2008 Euro crisis which for other NMS was ‘moment of truth’ which demonstrated
vulnerability and (un) sustainability of their pre 2008 growth model. A much stronger resilience of
Central Europe is attributed to its integration into German industrial system which itself has become
more dependent on the world economy than on the EU demand (CEPI and dE, 2014). However,
there are also concerns expressed about opportunities for further growth and dangers of potential
‘middle-income trap’ or inability to generate new sources of growth. In terms of drivers of growth
1
Based on World Bank criteria only Bulgaria and Romania are middle income economies while others are in a high income
group. However, from our perspective this classification is not suitable for categorising CEECs and for understanding
middle income trap as its original purpose was to detect the least developed economies. For our purposes we classify
countries into four groups: lower middle income ($1046-$4125), upper middle income ($4126-12745), lower high
income ($12176-$30000) and upper high income ($30001>) of gross national income per capita. We take notion of
middle income in broad sense encompassing countries moving from lower middle income group to upper high income.
See Appendix 1.
2
The World Economic Forum’s Global Competitiveness Index (GCI) 2013-2014 describes the majority
of the CEECs as being in transition from an efficiency-driven stage of economic development to one
which is innovation-driven. This transition requires drivers of growth which are not only in R&D but
also in developed production capability. A perspective which we use to illuminate this transition is
technology upgrading.
Technology upgrading is a multidimensional conceptual framework which goes beyond R&D in

explaining building of technology capabilities which accompany long-term growth. This concept is
open to sensitivities of different levels of development and is empirically informed but also has some
theoretical relevance. We consider it as appreciative theorizing framework which aims to overcome
a frequent weakness of composite indicators which is that they represent “measurement without
theory” (Koopmans, 1947). A conceptual approach is based on the literature review and is
developed as part of this task in a paper by Radosevic and Yoruk (2015) as part of this deliverable.
In this paper we argue for both theory and metrics of technology upgrading. The EU has been using
the European Innovation Scoreboard as it was called in the past and now Innovation Union
Scoreboard (IUS) as the major metrics in assessing progress of all EU countries in terms of their
innovation capacity. This metrics has become so dominant that some of its either individual or
aggregate indicators have been used as policy objectives and benchmarks in measuring how
countries perform in achieving the policy aims.
The IUS (EIS) has been originally constructed to measure the EU in relation to the US and Japan in
terms of innovation activities. It has been, in a sense, victim of its own success as it has now been
used as primarily tool for measuring innovation indicators for countries and regions within the EU
and assessing how they progress in terms of achieving aims defined by the IUS indicators. This
metrics inevitably leads to imitative paths of development in the sense that countries behind
technology frontier aim to invest similar relative amounts of GERD as well as introduce similar policy
instruments that are focused on activities at S&T frontier. Lee (2012) calls this option a ‘high road’ to
development or imitation of path of high-income economies.
So, our aim is to apply theoretically relevant but empirically grounded middle level conceptual and
statistical framework which could illuminate a type of challenges which seem relevant for a large
number of low income EU and neighbouring economies in their path out of potential middle-income
trap. As an outcome of our analysis we construct composite indicator of technology upgrading,
which can complement, not replace IUS and which better reflects different drivers and patterns of
technology upgrading of countries at different income levels.
We first explain the conceptual framework that lies behind the concept of technology upgrading
(section 2). In section 3 we discuss conceptual and measurement issues that our conceptual
perspective on technology upgrading entails. We explain how it has been constructed; we describe
individual indicators as well as applied method for constructing composite indicator of technology
upgrading. In section 4 we explore key stylized facts that emerge from use of dataset that falls within
our conceptual framework. In section 5, we report and interpret results based on composite
indicator and the relationship between income and technology upgrading levels. Section 6
concludes.
3
2. A framework for measuring technology upgrading: a conceptual approach
We depart from the proposition that technology upgrading is a multidimensional process. By this we
mean that: it is based on broader understanding of innovation, which goes well beyond R&D. It is
multi-level process which means that it is micro, mezzo and macro grounded but which also means
that at its core is structural change in various dimensions: technological, industrial, organisational. It
is also an outcome of global forces (embodied in international trade and investment flows) and local
strategies (pursued by host country firms and governments).
We are fully aware that learning which ultimately leads to technology upgrading is also institutional
process. However, for analytical conveniences we need to abstract from institutional set-ups and
focus only on outcomes of the learning processes as detected through technology upgrading
activities. Complexities of multi-level at which technology accumulation takes place, of diversity of
its patterns as depicted through structural change as well as interactive nature of technology
upgrading are sufficiently complex to justify this abstraction.
We use term technology upgrading rather than industry upgrading. Industries are mixtures of
different technologies of different R&D and capital intensity levels. The notion of industry is always
context specific and should not be reduced on statistical definition of industry at whichever level of
aggregation. However, we also recognise that that there is degree of overlap between industry and
technology upgrading as some industries are based on more complex technologies than others. In
that respect, technology upgrading is about changes in technology intensity but equally about
structural change. In fact, these two are inextricably linked. These two dimensions – technology
upgrading and structural change – should be considered jointly with the way economy integrates
itself in global value chains which are today an important determinant of technology upgrading. The
key to catch-up and post-catch-up is leverage of domestic innovation efforts with global industrial or
knowledge networks. Hence, magnitude of knowledge inflows and their coupling to domestic
innovation efforts are one of three key dimensions of technology upgrading.
In a nutshell, based on literature review and at very general level we conceptualize technology
upgrading as a three-dimensional process. It consists of dimension 1 (vertical axis) which is about
intensity of technology upgrading as depicted by different types of innovation activities, of
dimension 2 (horizontal axis) which is about spread or width of technology like diversity of
technological knowledge, types of supporting infrastructure and organisational capabilities of firms
which are the main carriers of technology upgrading, and of dimension 3 (diagonal axis) which depict
knowledge inflows into economy through a variety of forms like trade, FDI and GVC. All three
dimensions have strong grounding in the respective literatures on firm level technology upgrading,
on structural change and growth, and on integration in global economy (see Radosevic and Yoruk,
2015 for overview and argument).
Figure 1 presents technology upgrading as an outcome of interaction between intensification of

different types of technology activities (axis I), structural factors and changes in this process (axis ii)
which are mediated by the way economy interacts in this process with the global economy (axis III).
Given its three dimensional nature the aggregate indicator of technology upgrading can be
calculated statistically but as we show later on it makes sense to aggregate only dimensions 1
(intensity) and 2 (breadth). A third dimension (interaction with global economy) is moderating
4
dimension, i.e. it amplify or reduces effects of technology upgrading in dependence of modes of
integration or interaction with the global economy.
Figure 1: Dimensions of technology upgrading
DIMENSION1
Intensity of technology
upgrading by types
DIMENSION 2
Width of technology upgrading
(structural features)
2.1. Intensity and types of technology upgrading (scale)

This dimension of upgrading is about acquiring different types of technology capabilities, which are
also a reflection of different technological levels of economies. Economies that operate behind
technology frontier are more likely to grow based on production capability, not technology capability
while high-income economies are more likely to grow based on technology frontier (technology
capability and R&D) activities.
Three types of capabilities (production capability, technology capability, R&D and knowledge
intensity) are present in all economies to different degrees. Their importance as drivers of growth
varies in dependence of achieved income and technology levels as well as of the structural features
of economies.
2.1.1. Production capability

Production capability is capability to produce with given level of technology at world levels of
efficiency or productivity. This requires primarily good operational efficiency. The key workforce to
operational efficiency is skilled technicians or blue-collar workers. A more complex capability is
product and process engineering, which involves improvement in existing products and processes.
This capability is largely dependent on skilled engineers. Process and product engineering are still
part of production capability as they are about incremental innovations, not changes in designs but
improvements in products and processes.
We use as proxies for production capability ISO9901 certificates, trademark applications and
assessment on the job training activities. ISO certificates are generic management standard which
indicate that there are in place businesses process which should guarantee operational efficiency
though not necessarily its improvements. However, ISO adopters have far lower organizational
5
death rates than matched firms within their industries, their sales and employment grew
substantially more rapidly post certification than at matched firms (Levin and Toffel, 2009). With the
globalization they have spread as internal mechanism of quality control and as precondition for
participating in global value chains (Stevenson and Barnes, 2001). They are also proxy for a variety of
industry specific standards which are difficult to aggregate. Terlakk and King (2006) show that they
provide a way of communicating about unobservable firm attributes, thereby generating a growth
effect for certified organizations.
Trademark applications are proxy for developed production capability but in the service sectors.
They also proxy for marketing innovation and thus suggest that firm has differentiated production
capability or brand (Millot, 2009; Mendonca et al, 2004). Baroncelli et al. (2004) et al argue that the
global distribution of trademarks is skewed toward high-income industrial countries and that they
are concentrated in research and development-intensive sectors such as pharmaceuticals, scientific
equipment, and the chemical industry. Finally, on the job training is proxy for human capital capacity
to work effectively with the given technology.
2.1.2. Technology capability

A developed technological capability indicates capacity to significantly change product and processes
through organised innovation process. There is not sharp boundary between production and
technology capability but we can assume that technology capability is about development, not
necessarily about research. A first stage of development is advanced development or prototype for
manufacture which should be distinguished from exploratory development which is about prototype
in a system (Amsden and Tschang, 2003). There is important threshold level of capability required
for firms to move from advanced development, which is development for manufacture to own
design manufacture. Production capability, process and product engineering, and advanced
development are doable within OEM enterprises while exploratory development is a feature of own
design manufacturers (ODM).
The available proxies of development activities are resident and international patents and industrial
designs. They are the right proxies because they are about development but equally patents have
well-known biases (van Zeebroeck, 2011). An important advantage of using patents is the length and
consistency of time series derived as well as the possibility to identify technological fields or
specializations using the patent classification. Also, as countries move up towards technology
frontier patenting becomes more important and is less relevant for countries behind technology
frontier where IPRs are not the major form of protection of technological knowhow. We use
resident and international (EPO and US) patents. To capture domestic technological activities
pushing the technology frontier we rely on EPO and USPTO data which reflect technological activities
relevant for competitiveness in international markets. To capture technological capability for
technological development behind the technology frontier we use direct patent applications by
residents to their respective national patent offices. In general terms (even though the patent
strategies may differ from this rule) residents will directly apply for patents in their home countries
disregarding applications abroad if their technological activities do not have a global industrial
relevance. So, resident direct patent applications to national patent offices dominantly proxy
technology effort behind the technology frontier. Countries that are behind technology frontier
should have much higher share of resident patents and their share of transnational patents is
6
marginal. However, as they move towards technology frontier their transnational patenting
increases. This pattern may be somewhat different in very large catching up economies where
domestic patenting may continue to play important role. However, their transnational patenting as
proxy of world frontier technology effort should continue to increase. We use both EPO and USPTO
data due to a geographic bias of patents. We do not use ‘international’ patents filed under the
Patent Cooperation Treaty (PCT), or applications filed simultaneously at several national offices (e.g.,
the ‘triadic families’) as these are biased towards inventions of higher value, which are often owned
by large firms. This would underestimate patenting of countries at lower income levels that do not
have many large firms (van Zeebroeck, 2011).
2.1.3. Research & development

R&D is usually considered as the major component of innovation and therefore one of the major
drivers of growth. This model is the basis of new (endogenous) growth theory (Romer, 1990; Lucas,
1988). OECD (2003) landmark study shows that there is clear positive linkage between private sector
R&D intensity and growth in per capita gross domestic product (GDP) for OECD economies.
However, there is no clear-cut relationship between public R&D activities and growth, at least in the
short term (ibid). At industry level, the effects of R&D on productivity are differentiated and
dependent on degree of concentration (high and low concentrated) and level of technology
(high/low tech)(OECD, 2003). Returns to R&D are particularly high in concentrated high tech
industries which are characterised by “creative accumulation” or large established firms with high
barriers for new innovators (ibid).
Research and development are usually treated as one category though this seems to be mainly due
to statistical convention rather than belief that research and development are indeed similar
categories. Amsden and Tschang (20003) show that the Frascati definitions are not specific enough
to allow an R&D project to be classified accurately by conventional type (basic, applied and
development). Under technological capability we capture patents as they have intended commercial
application. However, R&D has far broader aims and its links to growth and productivity is far from
straightforward. A literature conventionally accepts that R&D has two faces (Cohen and Levinthal,
1989). One is as being driver of world frontier innovation and other one is R&D as driver of imitation
activities or as factor of absorptive capacity. On that basis it is assumed that R&D is driver only of
world frontier innovation but only factor of absorptive capacity in industries that operate behind
technology frontier. Yet, Griffith et al (2004) on the example of twelve OECD countries find R&D to
be statistically and economically important in both technological catch-up and innovation. Kneller
and Stevens (2006) find opposite also on the example of twelve OECD countries and conclude that
the effect of R&D on production is primarily through its contribution to the stock of frontier
knowledge itself in each industry.
It seems that R&D plays different role in economies at different levels of development. For example,
middle income economies tend to grow more on imitation activities while transition towards high
income group requires a shift towards technology frontier activities. So, in both groups R&D plays
important but different role. In catching-up economies R&D has much more important role in terms
of absorptive capacity or capacity to effectively use knowledge from abroad in addition to its role as
driver of world frontier innovation. In catching-up EU economies technology transfer activities are
important drivers of innovation along with the non-R&D-based innovation activities. Reinstaller and
7
Fabian Unterlass (2010) show using CIS micro-data for 17 EU countries that the determinants of
successful product innovation of European innovative firms vary across countries depending on how
far they are from the technological frontier. Farther away from the technological frontier technology
transfer is more important than own R&D; close to the frontier the cooperation with universities,
own research, highly skilled personnel and intellectual property rights are very important (ibid).
In view of this evidence, the mainstream model of R&D based growth like Crépon, Duguet and
Mairesse (1998) or CDM model which establishes the link between RD, innovation and productivity
is theoretically grounded but does not capture non-R&D drivers of growth which are quite important
in catching-up EU economies. R&D remains important in the catching-up economies but its role as
the factor of absorptive capacity is much more prevalent when compared to its role as driver of
world frontier innovation. This is the main reason why we have to separate technological capability
from R&D which has much broader role in economic development.
We proxy R&D through the following indicators, both ‘hard’ and subjective: Business Enterprise
Sector expenditures as % of GDP; Research and development expenditure (% of GDP); Researchers in
R&D (per million people); Technicians in R&D (per million people); Science publications: Scientific
and technical journal articles; Science citations; and subjective assessment of Company spending on
R&D; Quality of scientific research institutions; and of University - industry collaboration.
It is important to bear in mind that that production, technology and R&D capability are not
hierarchically structured, i.e. when moving from technology capability to R&D or from development
to research does necessarily involves higher technology complexity but simply qualitatively different
set of technology or knowledge requirements. Equally, if not more important, upgrading to ’higher’
stages is not automatically more rewarding in terms of value added i.e. upgrading may not
necessarily lead to increased incomes but can simply be necessary to maintain the existing levels of
income.
2.2. Breadth of technology upgrading: structural change, infrastructure and firms’

structure (scope)
Technology upgrading is about changes in technology intensity but equally about structural change.
In fact, these two are inextricably linked. The process of technology upgrading is a process of
diversification of technological knowledge, of increased organisational diversity and specialization.
Technology diversification is expansion of technology base into a broader range of technology areas.
2.2.1. Infrastructure: human capital, physical and organisational

Technology upgrading is primarily taking place in firms but it is not only-firm level business. The
accumulation of technology capability in firms must be accompanied by an organizational and
institutional infrastructure that supports the acquisition of such capabilities. Choung et al (2014)
shows that transition from the adoption (catching-up) to the creation stage (post-catch-up) depends
on the range of infrastructures that support innovation in a country, in addition to the strategy and
resources of a single company. We consider infrastructure to be an important dimension of
structural change. Infrastructure upgrading is important element or externality of technology
upgrading. Inefficiencies in infrastructure can hinder otherwise competitive firms to upgrade
8
A human capital can be also considered a very specific type of ‘infrastructure’ or precondition for
technology upgrading. Technology embodied in new machinery and equipment will not by itself lead
to increased productivity unless there are human skills to effectively use it and improve it. Also, in
order to be effective human skills need to be part of a specific organisational and economic process
that rewards dexterity, learning and innovation (Lazonick, 2002). So, human skills unless converted
into firm specific skills will not suffice for technology upgrading. Still, a human capital acquired
through education can be considered an infrastructural precondition or input into technology
upgrading.
Another structural precondition for technology upgrading is physical infrastructure. Infrastructure

generates large externalities to firms’ transaction costs. It is both public good and an input in the
production of other intermediate inputs. Access to infrastructure services is strongly correlated with
a country’s average income (Griibler, 1990). As countries reach certain stages of economic
development the extent to which infrastructure may represent a binding constraint on development
changes. Thirty out of 32 studies of OECD countries found a positive effect of infrastructure on some
combination of output, efficiency, productivity, private investment and employment (Romp and de
Haan, 2005).
This dimension of technology upgrading is proxied by human capital indicators (average years of
schooling of those over 25y age, subjective assessment of quality of maths and science education, of
availability of research and training services, of availability of scientists and engineers. Physical
infrastructure is proxed by fixed broadband Internet subscribers (per 100 people) and by Gross Fixed
Investment as % of GDP.
2.2.2. Structural changes: technology diversification, changes in the demand and suppply
of innovation
There is not general theory of structural change but a variety of theoretical approaches of different
methodological nature that aim to explain structural shifts between three broad sectors and among
industries within these sectors (Krueger, 2008). There is a common understanding that technological
changes affect structural change in the way that industries with relatively lower rates of productivity
growth tend to shrink in terms of shares while those with higher rates of productivity growth
expand. In this way structural change promote aggregate productivity growth even if we assume
that within industries productivity growth remains stagnant. However, the empirical evidence on the
role of structural change in aggregate productivity growth escapes broad generalisations and differs
very much across different periods and countries or regions.
So, despite recognising importance of structural change we can derive very little in the way of
importance of different sectors and industries in economic growth. The whole point of structural
change driven by technology is that it changes boundaries of industries as well as the nature of
industries. Hence, using high tech, as proxy of structural change would be highly misleading as high-
tech elements permeates many low-tech sectors. Also, catching-up countries are involved
increasingly in high tech industries but at low value added segments. Similar to this, the share of ICT
industries in industry structure or in export would be highly misleading as it ignores value added
levels. So, instead of being focused on structural changes at the level of industries we prefer to focus
on more reliable trends regarding technological changes. By this we mean primarily technological
diversification as reflected in changes of structure of patenting as well as level and changes in
9
subjective perceptions of demand and supply of technologies. So, we construct Herfindahl index of
concentration to measure changes in patenting structure of WIPO, EPO and US patents. We also use
subjective perceptions of Buyer sophistication; of changes in buyer sophistication; of assessment
availability of state-of-the-art technologies as proxy of supply as well as changes in availability of
latest technologies.
2.2.3. Firm level organisational capabilities

Firms are the main agents of innovation process and thus if we want to understand technology
upgrading paths we cannot ignore them. Yet, this structural feature of economies is by and large
ignored in understanding determinants of growth and technology upgrading. Usually, innovation is
associated to new technology based firms while the role of big business is largely ignored. The
changing interaction between large and small firms during development is very much-unexplored
topic. A quite new way of looking at the role of big business in catch-up is Lee et al (2013), who look
at the big business as one of binding constraints in growth, especially in countries that are presently
in the so-called middle-income trap.
This view rests on the following persuasive evidence. First, econometric evidence suggest that the
big businesses and not SMEs exert an independent and robust effect on economic growth. Rich
countries tend to have larger numbers of big businesses than predicted by their sizes while many
middle-income or non-members of the OECD tend to have negative residual numbers or a lesser
number of big businesses than predicted. In overall, the econometric evidence suggests that a big
business plays a more robust role in the economy than SMEs. Second, one of the crucial differences
between high- and middle-income countries might be that the latter lack the big businesses that can
compete globally. The stage of branding requires often big business, which can incur high fixed costs
for R&D and marketing. Third, excessive relative dependence on big businesses does not seem to be
good either. Lee et al (2013) explain that:
‘Increases in the relative predominance of big businesses in a nation’s economy have negative
implications for growth despite the positive link of expansion in the absolute size and number of big
businesses to national growth and GDP stability’ (Lee et al, 2013)
In overall, organisational variety in economy does matter for technology upgrading and should be
included in determinants of technology upgrading as one of structural variables. This is not to deny
the importance of SMEs but to point out that SMEs alone are insufficient as drivers of technology
upgrading. Big business plays a crucial role in structural transformation especially through diversified
business groups, which are present in all middle-income economies (Morck, Wolfenzon, and Yeung,
2005). They are carriers of usually missing organisational capabilities in middle-income economies,
which are important ingredient of technology upgrading (Khanna and Yafeh, 2007; Amsden and
Hikino, 1994).
Eastern Europe is good example of a region whose labour force post-1989 had relatively high
education levels but also low firm specific skills. In post-1989 this has been considered as great
advantage and in a few cases led to domestic innovations but with meagre economic results. A god
example is Estonian ICT industry (see Hogselius, 2005). However, a careful examination of this case
shows that individual competencies alone are not sufficient without firm specific organisational
capabilities. The inherited competencies are strongest at the level of individuals but this may be not
10
enough to develop dynamic innovation system without organisational capabilities within which
individual competencies can be harnessed. A much more known case is of the Indian software
whose success is usually attributed to individuals. Indeed, competition by MNEs in labour markets
(not competition on product markets) has induced productive efficiency among domestic firms.
However, key to use of human capital were organisational capabilities of Indian software firms which
were building through imitation and which harnessed to productive use skilled labour force
(Athreye, 2005).
This sub-component of technology upgrading is proxied by two indicators. First is a number of

Forbes 2000 companies per mn population which indicate the relative share of organisational
capabilities in economy. A second indicator is subjective assessment of firm level technology
absorption. This is to ‘pick up’ countries that do not have Forbes size firms in their economy but
where subjective assessment of firms’ technology absorption capability can be used as a substitute
2.3. Interaction with global economy and technology upgrading

A successful technology upgrading is never entirely autonomous process but is always linked to
inflow of foreign knowledge skills, which are coupled with intensive domestic technology effort
(Radosevic, 1999). A literature that documents that is far too lengthy and only a few examples will
suffice to reiterate this robust but often forgotten stylized fact (Mowery and Oxley, 1997, Kim 1997,
Amsden 2001). The emphasis is usually on one of these two elements of catching up – either on
domestic technology accumulation or on inflows of foreign knowledge through trade, FDI and
generally open economic regime.
Literature on FDI and technology upgrading or knowledge spillovers is quite numerous. A meta-
review of this literature by Bruno and Campos (2013) shows that the effect of FDI on economic
performance and growth are conditional. Firms, sectors or countries that are below certain
“thresholds” (either in terms of human capital, financial development or institutional quality) are
less likely to benefit from FDI. In overall, benefits are significantly greater in low-income than in
lower and upper middle-income countries (both at the micro and macro level). Available data
provide stronger support for differentiating the effect of FDI on growth across levels of development
rather than in terms of geographic regions.
The effects at the macro level depend upon whether recipient countries have attained minimum
levels of human capital, financial and institutional development. The effects of FDI using firm-level
data tend to find that the (micro-) effect is conditional upon the type of linkages (with backward
linkages, that is, links between the firm and its suppliers, dominating over horizontal or forward
linkages).
Indeed, FDI is a potential source of technology upgrading. Integration into the global economy and
foreign direct investment (FDI) can act as important catalysts for change but equally FDI alone are
not driver of technology upgrading. As literature suggest their effects on upgrading are highly
differentiated and dependent on indigenous technology effort. Even when countries are integrated
globally in R&D networks they do not necessarily link up with domestic manufacturing value chain
which lead to what Ernst (2014) describes on the example of Indian electronics as ‘truncation of FDI
based learning’. He explains this by fragmented Indian national innovation system in which local
11
electronics manufacturing remains disconnected from India‘s chip-design capabilities which are
integrated, instead, into global networks of innovation and production.
FDI indicators are of limited value in detecting the true knowledge that is acquired through
international industrial networks. Research on GVC is useful in that respect though it is difficult to
generalise. Different contributions show the positive and significant effects of learning through value
chains on process, product and functional upgrading up to ODM level. Yoruk (2013) shows the major
importance of both knowledge and production networks for firms’ upgrading but also that is highly
misleading to narrowing learning opportunities for upgrading to interactions with the global buyers
within GVCs. Her research in the case of Eastern Europe shows that learning by doing and learning
by exporting do not have statistically significant effect on functional upgrading. She shows that
opportunities offered by GVCs will be of little use unless firms have ability to internalise this external
knowledge through its human resources, through training and research within the firm. She also
shows that managerial upgrading is important to technology upgrading but global buyers do not
support it. This again highlights the importance of organisational capabilities or firms’ structure that
we discussed under structural dimension of technology upgrading.
A very recent example of the importance of integration into global value chains and its growth
benefits is German – Central European supply chain cluster (GCESC)(IMF, 2013; IMF2013b). The
increase in foreign value added in four major countries locations of GCESC (Czech R, Slovakia, Poland
and Hungary/CE4) appears to have led to increases in domestic value added through productivity
increases as well as by creating demand for ancillary products and services in host economies. It
seems that participation in the GCESC has led to considerable technology transfers to the CE4
countries though there is not yet clear consensus about its magnitude due to high heterogeneity
among firms in terms of fostering skills. The aspiration is to move up value added chain but equally
there are obvious weaknesses in terms of local skill accumulation, training and infrastructural
support for training, especially of firm and technology specific skills.
Similar to this example, Leitner and Steherer (2014) show that the EU13 benefit the most from
stronger trade integration. On the contrary, they do not gain from vertical specialization as they
specialize in the low-value added yielding assembly stage of the global production chain. The EU-15
countries are located higher up the value chain and they tend to gain more in output, employment
and labour productivity growth (gross output based) from more intense vertical specialization. Their
‘blessing’ is growth of exports which is quite beneficial for NMS in terms of gross output and labour
productivity but a ‘curse’ comes from the fact that the higher degree of vertical specialization does
not necessary translates into better performance of industries. On balance, they consider that the
overall effects are positive as their losses in terms of value added growth or labour productivity as
they specialize vertically are compensated by gains in terms of higher average export growth. In
overall, this suggests that vertical integration of EU13 is a mixed blessing as long as countries remain
located in low value added assembly stages of global value chains.
These examples from CEE and the conclusion from literature review on conditional effects of FDI
demonstrate that this axis of technology upgrading is not like other two. Namely, intensity of
technology upgrading axis assume that the higher are countries ranked in specific sub-components
and individual indicators the higher the potential of economy for technology upgrading and thus for
a long-term growth. The more developed production, technology and R&D capability the higher is
12
potential for technology based growth of the economy. Also, the more countries are diversified in
terms of structural features of technology upgrading the higher is their potential for technology
upgrading. The more developed is human capital and skills and physical infrastructure the more
likely is that this will positively affect technology upgrading. The more countries diversify in terms of
technological knowledge and the higher levels of assessment of their supply and demand for
technology and higher are changes in these factors the more likely is that these countries will
upgrade technologically at a higher rate,. Also, the more developed organisational and technology
absorption capabilities of firms are the more likely is that this will lead to higher pace of
technological upgrading. However, this is not necessarily the case with the third dimension which
moderates technology and knowledge exchange but does not necessarily leads to higher inflows of
knowledge and technology. Its ultimate effects will depend on interaction with other two
components.
2.3.1. Technology and knowledge exchange

Out of three dimensions of technology upgrading the interaction with global economy is probably
the most difficult to capture since technology transfer happens through capital equipment import, it
is embedded in modes like FDI, networks and subcontracting or is disembodied (licences). So, modes
of transfer by themselves cannot be taken as proxies of the real knowledge transfer that has taken
place (Radosevic, 1999). In view of that complexity we consider distinction between knowledge
imports and export via licences, FDI outflows and inflows and share of exports in complex industries.
We intentionally take this very broad category of complex industries which includes SITCRev3
categories 5, 71-79, 87 and 88 in order to avoid the problem of narrow definitions like high tech
which ignores value added difference. We assume that export in these industries is one average of
somewhat higher complexity than in other industry groups.
However, as we argued earlier the higher share of complex industries in exports may not indicate
export of technological knowledge in whatever form as countries may be integrated through vertical
specialization but at low value added segments. Also, FDI can bring knowledge and generate
spillovers but equally it may lead to insignificant or negative spillovers. Technology balance of
payment may be positive or negative unrelated to real flow of knowledge but s instrument of
transfer pricing of MNEs.
In summary, technology upgrading is three-dimensional process proxied by a variety of indicators,

which basically address three aspects (figure 2):
 Intensity of technology upgrading through various types of innovation and technology

activities
 Widening or broadening of technology upgrading through different forms of technology and
knowledge diversification
 Interaction with global economy which moderates impact of other two axes on technology
upgrading by assisting, hindering or being neutral.
13
Figure 2: Dimensions and components of technology upgrading
Intensity of • Production capability

technology • Technology capability
upgrading by types • R&D and knowledge intensity
• Technology (embodied) export

Interaction with
• Knowledge imports and exports
global economy • Foreign direct inflows and outflows
Breadth of • Infrastructure (human, physical, organizational)

technology • Structural change
upgrading • Firms’ structure
3. Why and how composite indicator of technology upgrading?

There is variety of composite indicators that measure countries’ performance in growth,
competitiveness and innovation. Examples are: the Global Competitiveness Index (GCR, 2012), the
Knowledge Economy Index (KAM, 2005; Chen and Dahlman, 2005) of the World Bank, the World
Competitiveness Report Index (WCY, 2011), Technological capability of countries and the ArCo,
(Archibugi and Coco, 2005, 2004; Archibugi et al., 2009), the UNIDO Industrial Performance
Scoreboard (UNIPS) (UNIDO 2002, 2002a), the Summary Innovation Index and the Global Innovation
Index, both of the European Commission; the Technological Activity Index of the UNIDO; the
Technological Advance Index of the UNCTAD; the Technology Achievement Index, developed by
UNDP and reported in the Human Development Report 2001, and the S&T Capacity Index (STCI)
proposed by the RAND Corporation, the High-Tech Indicators (HTI) developed at the Georgia Tech
Technology Policy and Assessment Center and reported by the National Science Foundation's
Science & Engineering Indicators.
Nasierowski and Arcelus (2000) show that similarity in ranking across different indexes are
significant. They all point to importance of innovation to economic development but differences in
their conceptual perspectives do not change significantly ranking among countries. Archibugi et al
(2009) show similar results but also show that differences in ranking cannot be substituted by single
indicator like R&D.
It is important to bear in mind that different indexes treat ‘technology’ in different ways. Some of
them cannot be taken as a direct measure of innovative performance. Indicators like Global
Competiveness Index depict the quality of the current endowment of a country (including
institutions) and among them also the technology activities as one of determinants of growth.
Moreover, GCI takes differentiated view on the role of technology across development path
14
assuming that the closer are countries towards technology frontier their growth and
competitiveness rests more on knowledge and technological activities.
A weakness of composite indicators is assumption that individual indicators are substitutes, which
thus can be aggregated and averaged. For example, poor ranking terms of ISO certificates can be
substituted by better performance in terms of science publications. Within our framework this is
possible only within the individual categories within dimensions (axis).
Our aim is not to focus on ranking per se but on different paths of technology upgrading. The
learning effect should be in showing diversity of paths and compare countries in terms of their own
upgrading paths. Ranking makes sense when comparing countries that follow similar technology
upgrading path or are at very similar stages of technology upgrading process.
Technological upgrading is a flow rather than stock concept. However, the major challenge is to find
series that are enough long to depict the changing levels of technological activities and capabilities.
Unfortunately, there is not sufficiently long data series which would enable us to construct long term
series of technology upgrading which could be regarded as flows. Instead, we are forced to measure
technology upgrading index only in terms of levels or stock. Luckily, unlike macroeconomic variables
technological capabilities are changing very slowly even during periods of deep economic crises or
high growth periods (Archibugi et al, 2009).
3.1 Data
Primary data for individual indicators (see Tables 1 and 2) have been acquired from a variety of
sources: World Bank, WEF Global Competitiveness Report, WIPO, UNESCO, UNComtrade, ISO,
Thomson NSI, Forbes and Barro-Lee dataset at national level for 42 countries (see Appendix A for a
list of selected countries). Relevance to the analysis, availability and comparability were the criteria
used to select the indicators. This has also determined the mixture of ‘hard’ and ‘subjective’
indicators. Missing data accounted for less than 5% of the data. Missing values in the data set have
been treated using several methods. Where possible, missing values were replaced with values from
an external source, e.g. from a previous round of the same survey. The remaining missing values are
treated by multiple imputation method.2 Afterwards, the data are classified into two main
categories and six components of the technology upgrading framework to create an index of
technology upgrading (ITU). We generate a separate index for technology and knowledge exchange
element (ITKE) (see Figure 1 and Tables 1 and 2).
2
Multiple imputation (MI) is a general approach that does not require a specification of parameterised likelihood for all
data. The imputation of missing data is performed with a random process that reflects uncertainty. Imputation is done
N times, to create N “complete” datasets. We have used Markov Chain Monte Carlo (MCMC) method. MCMC is a
sequence of random variables in which the distribution of the actual element depends on the value of the previous
one. It assumes that data are drawn from a multivariate normal distribution and requires assumptions that missing
values are missing at random. The theory of MCMC is most easily understood using Bayesian methodology (OECD,
2008).
15
Table 1. Categories, components and indicators of Technology Upgrading Index (ITU).
Index Categor Component Quantitative Indicators Source Year Categ Comp Cronbach’s
y ory weigh alpha
weigh t
t
1. Production 1.ISO9001 certificates pmi ISO 2007-11
A. INTENSITY AND TYPES OF TECHNOLOGY
capability 2.Trademark applications, resident pmi WB avg 1/6

3.On the job training Q.5.C WEFGCR
2012-13
2. 4.Patents resident applications to national office pmi WB 2007-11
Technology 5.Patent applications to USPTO pmi WIPO avg
1/6
capability 6.Patent applications to EPO pmi WIPO
7.Resident's industrial design count pmi WIPO
1/2 0.946
3. R&D 8.Business Enterprise Sector R&D expenditures (% of GDP UNESCO 2011
capability 9.Research and development expenditure (% of GDP)
UPGRADING (SCALE)
10.Researchers in R&D pmi WB 2010

11.Technicians in R&D pmi
1/6
12.Scientific and technical journal articles pmi
13.Science citations pmi ThomsonNSI 2007-11
Index of Technology Upgrading
14.Quality of scientific research institutions Q.12.02 WEFGCR avg

15.University - industry collaboration in R&D Q.12.04 2012-13
ITU = A + B
4. 16.Average years of schooling 25+ Barro-Lee 2010

B: BREADTH OF TECHNOLOGY UPGRADING: STRUCTURAL
Infrastructure 17.Quality of maths and science education Q.5.04 WEFGCR 2012-13

: human 18.Availability of research and training services Q.5.07 1/6
capital and 19.Availability of scientists and engineers Q.12.06
physical 20.Fixed broadband Internet subscribers (per 100 people) WB 2012
21.Gross Fixed Investment as % of GDP
5. Structural 22.Herfindahl-Hirschman Index for total national patent WIPO 2007-12
change applications avg
23.Herfindahl-Hirschman Index for patent applications to
EPO WEFGCR 2012-13
24.Herfindahl-Hirschman Index for patent applications to
1/2 0.893
USPTO
1/6
25.Buyer sophistication Q.6.16
26.Change in buyer sophistication( % change in Q. 6.16
from 2006-07 to 2012-13)
FEATURES (SCOPE)
27.Availability of state-of-the-art technologies Q.9.01

28.Change in availability of latest technologies( % change in
9.01 from 2006-07 to 2012-13)
6. Firm 29.Number of firms in Forbes 2000 pmi Forbes 2013

1/6
organisationa 30.Firm level technology absorption Q.9.02 WEFGCR 2012-13
l capabilities
Table 2. Indicators of technology and knowledge exchange index (ITKE)

Index Quantitative Indicators Source Year Cronbach’s
alpha
TECHNOLOGY AND KNOWLEDGE 31.Licensing receipts as % of GDP WB 2012
EXCHANGE 32.Licensing payments as % of GDP
33.Share of exports in complex industries in total exports UNComtrade 2008-12 avg
0.827
(SITCRev3 5 71-79 87 88)
34.Foreign direct investment, net outflows (% of GDP) WB 2007-12 avg
35.Foreign direct investment, net inflows (% of GDP)
3.2 Developing measures by composite index methodology

The individual indicators in Tables 1 and 2 have then been used to construct latent variables for ITU
and its sub-indexes Index A, Index B, Index C and ITKE by using composite index methodology.3
3
Composite indexes are widely used in economic and business statistics for benchmarking the relative progress of
countries in a variety of policy domains such as competitiveness, globalization and innovation. Even though Grupp and
Mogee (2004) criticize the approach for its vulnerability to manipulation, Archibugi et al. (2009) stress that when they
are used in the right perspective and within a sound theoretical framework they can be extremely useful tools.
16
A typical composite indicator will take the form (Freudenberg, 2003: 7):
(1) 𝐼 = ∑𝑛𝑖=1 𝑤𝑖 X 𝑖
where
I: Composite index,
Xi: Normalised variable,
wi: Weight of the Xi, ∑𝑛𝑖=1 𝑤𝑖 = 1 and 0 ≤ w ≤ 1
i: 1,…, n.
Equation (2) shows explicitly the normalisation method (Min-Max) used:
𝐽
𝑀
min max min
(2) 𝐼𝑐 = ∑ ∑ 𝑤𝑗𝑚 {(X𝑗𝑚𝑐 − X𝑗𝑚 ) | (X 𝑗𝑚 − X𝑗𝑚 )}
𝑚=1
𝑗=1
where c indicates country, j and m are indicator and component subscripts and min and max denote
the minimum and maximum values of each indicator across countries.
Based on our analytical framework, the ITU (index of technology upgrading) is composed of two
categories: intensity/type of technology upgrading and breadth of technology upgrading: structural
features. The first category is composed of three components: production capability, technology
capability and R&D and knowledge intensity based on sixteen manifest indicators. The second
category is composed of two components: human capital and physical infrastructure and structural
change based on thirteen manifest indicators.4 The sources, availability and weights for each of
these indexes and their indicators are presented in Tables 1 and 2. All indexes and sub-indexes are
estimated based on the normalisation5 of manifest indicators followed by aggregation6 of
components with equal weights given to each component. In the existing literature there are
numerous weighting methods with pros and cons. These vary from equal weighting to use of
statistical models such as factor analysis (FA)/principal component analysis (PCA) or a ‘benefit of the
doubt’ (BOD) approach which is sensitive to national priorities and weights are country specific
(OECD, 2008: 32). Based on our already developed conceptual framework, we use an equal
weighting method applied on each component. OECD (2008:31) states that “most composite
indicators rely on equal weighting, i.e. all variables are given the same weight. This essentially
implies that all variables are ‘worth’ the same in the composite, but it could also disguise the
4
See Appendix B for a detailed explanation of indicators and the formation of indexes.
5
We have used Min-Max (distance from the best and worst performers) normalization method in this research, since this is
the most compatible method with the indicators we have chosen.
6
Linear, geometric or multi-criteria aggregation might be applied. We opted for linear aggregation method where
substitution along dimensions (components) is constant, which we prefer in accordance with our conceptual
framework. It is also compatible with Min-Max method of normalization, especially when individual indicators are
measured in different units. Technical weaknesses of the aggregation convention are widely discussed in Munda and
Nardo (2009).
17
absence of a statistical or empirical basis, e.g. when there is insufficient knowledge of causal
relationships or a lack of consensus on the alternative. Moreover, if variables are grouped into
dimensions (components) and those are further aggregated into the composite, then applying equal
weighting to the variables may imply an unequal weighting of the dimension (the dimensions
grouping the larger number of variables will have higher weight). This could result in an unbalanced
structure in the composite index.” That is why we have given the same weight to each component as
based on our conceptual framework and then determined the weight of each individual indicator to
achieve a balanced structure in the composite index.
We also provide Cronbach’s alpha coefficient of reliability for each of the latent constructs (see
Tables 1 and 2) to further indicate the correlation between the selected manifest indicators
(Anderson and Gerbing, 1988). C-alpha in each case is above the 0.70 threshold of acceptable
reliability (Nunnally, 1978). This shows that they are highly likely to share common factors as
outlined by our conceptual framework and it is evidence that the indicators are measuring the same
underlying construct (OECD, 2008: 71–2). We also provide the results of Confirmatory Factor
Analysis (CFA) applied to indicators for each component separately in Appendix C. The indicators
selected for the components and main indices merge into factor groups assuring the validity of the
constructs.
In Appendix C we also show how well individual components of indexes covariate between item-
pairs, i.e. how well they are loading on the common factor. For six indexes (except production
capability) index we have very high C-alpha scores. Production capability index which is composed of
ISO9001 certificates, trademark applications per mn pop and subjective assessment of the on the job
training has low reliability score of 0.458. We think that this reflects much less conceptual
inconsistency of production capability concept but more diversity of investments in this capability at
countries of different income levels.
4. Analysis of levels and profiles of technology upgrading
4.1. Probing the relationships between technology upgrading and different income
groups
In this section we present results of construction of composite index of technology upgrading for 42
selected middle and high income economies ranked by income groups and levels of ITU index
(Tables D.1 and D.2 in Appendix D). Figure 3 shows relationship between average scores for four
income groups of countries across three indexes (intensity and breadth of technology and index of
interaction with global economy). As expected income groups are ordered in decreasing manner
across all three indexes. However, the range between income groups is the biggest in the case of
index of intensity of technology upgrading (82 percentage points between low middle income and
upper high income) and smallest in the case of index of breadth of technology upgrading (49
percentage points) with interaction index being close to the index of intensity of technology
upgrading (71 percentage points). This suggest that breadth of technology upgrading takes places at
all income levels while intensity of technology upgrading is significantly less intensive in the lower
income groups. Also, technology and knowledge exchange is significantly more intensive in the case
of upper high income group but it is much more compressed in the case of lower upper income and
higher middle income groups.
18
Figure 3. Intensity and breadth of technology and index of interaction with global economy by
income groups.
Figure 4 shows six sub-indexes that form index of technology upgrading. Again, each of sub-indexes
is ranked in decreasing order as average income per capita of the groups fall. The biggest range is in
the case of technology capability where difference between upper high income and low middle
income groups is 96% percentage points. The gap between upper high income and middle income
economies is strikingly high in this case which suggest that generation of frontier technology is
domain of only the highest income bracket group. Production and R&D capabilities are also quite
dispersed with 77 and 79 percentage points’ difference. In the case of production capability, their
ranking is quite ordered with similar distances between different income groups. It is less orderly in
the case of R&D capability as middle income group R&D capability is relatively less advanced
compared to their production capability. The breadth of technology upgrading is much less
dispersed with max-min difference for structural change being only 32 percentage points and for
infrastructure 48 percentage points. This suggest that structural change may be of different types
but it does take place at different income levels creating quite compressed distribution with small
differences between different income groups. Differences for infrastructure are comparatively also
smaller but difference between two middle income groups are quite small but significant when
compared to upper high income group. Firm level organisation capabilities are closer to dispersion of
production and R&D capability (69 percentage points). However, they are distinctively less
developed in the case of upper lower income and middle income economies when compared to
upper high income group. They are not as pronounced as in the case of technology capabilities but
they are concentrated much more in upper high income than in the lower high income group. So,
the threshold level for firm organisational capabilities is located much higher in terms of income
than for infrastructure, R&D or production capabilities.
19
Figure 4. Sub-indexes of technology upgrading by income groups.
Four scatter diagrams below show the relationship between index of technology upgrading, its two
component indexes and index of interaction with global economy (Figure 5). These relationships are
presented for the overall sample of 42 economies as well as for each of four income groups.
First, index of technology upgrading is very closely correlated to levels of income. This is not
surprising as we have selected into index those indicators that are conceptually justified and
accepted as indicators of technology upgrading and the majority of which are positively correlated to
the levels of income. However, this positive relationship is not present in each of four income
groups. In fact, the relationship is positive only in high income groups and there does not seem to
be any relationship in middle income groups. This may be expected as drivers of growth are more
related to technology capability the higher is the income level of countries. Middle income
economies grow based on factors related to production capability and availability of labour and low
labour costs. Still, it is interesting that there seems to be a kind of threshold from middle income
group to lower upper high income where this relationship changes dramatically into positive. Also,
regression line does not represent good fit for the upper middle-income economies which have
index of technology upgrading similar to the lower upper high income but this level of technology
upgrading does not seem to be converted into similar levels of income. Possibly, this indicates a kind
of technology middle income trap which we want to explore further. The same type of relationship
holds for the first component of index of technology upgrading which is index A: intensity and types
of technology upgrading. Here, we also observe strong positive relationship between index and
income levels for upper high income economies, positive but less robust for lower upper high
income and not relationship for middle income groups. Korea is the only country in our sample that
is outlier. By this we mean that it has index of technology upgrading which far exceeds its current
income per capita. Its outlier position is very strong in terms of Index of intensity of technology
upgrading.
20
Figure 5. Index of technology upgrading, its component indexes and index of interaction with
global economy by GNI per capita across income groups.
LHI UHI Total LHI UHI Total
200004000060000
200004000060000
SWE SWE SWE SWE

USA USA USA USA
AUSJPN AUSJPN AUS
BELJPN
DEU AUS
BELJPN
DEU
BELDEU BELDEU IRL IRL
GBR GBR
IRL GBR IRL GBR ITA ITA
ITA ITA ESP ESP
ESP ESP GRCPRT
SVN KOR GRCPRT KOR
SVN
KOR KOR CZE CZE
GNI per capita in 2012

GRC SVN GRC SVN EST EST
PRT
CZE PRT
CZE RUSPOL CHL RUSPOL
HUN
BRA
KAZTUR
MEX
CHL
MYS
CHLEST CHLEST ROU
BGR
PER ZAFCHNBLR
RUS
POL RUS
POL
HUN
BRA
TUR ALB
MDA IDN
PHLTHA
JOR
UKR
MAR
KAZMEX MYS GHN
VNM IND
0
ROU
ZAF
BGR
PER BLR
CHN
ALBTHA
JOR
UKR
IDN
MAR
PHL
MDA
GHN
VNM
IND
0
0 10 20 30 10 20 30 40
LMI UMI LMI UMI
200004000060000
200004000060000
KAZHUN
ROU
BGR
BRA
TUR
MEX
ZAF MYS
IDN PER
ALB CHNBLR
THA
JOR
HUN
BRA MDA PHLUKR
GHN
VNM MAR
IND
0
KAZ TUR
MEX
ROU MYS
ZAFCHN
BGR
PER
ALBTHA
JOR BLR
UKR
IDN
MAR
PHL
MDA
GHN
VNM
IND
0
10 20 30 40 10 20 30 40
0 10 20 30 0 10 20 30
Index B : Breadth of technology upgrading_structural features
Index A : intensity and types of technology upgrading
Fitted values gnipc
Fitted values gnipc
Graphs by incomegroup
LHI UHI Total

200004000060000
LHI UHI Total

200004000060000
SWE SWE
USA USA SWE SWE
AUS JPN
DEU AUS JPN
DEU USA USA
BEL BEL AUS
JPN AUS
JPN
IRL
GBR IRL
GBR DEU BEL DEU BEL
ITA ITA GBR IRL GBR IRL
ESP ESP ITA ITA
GRCSVN KOR GRCSVN KOR ESP ESP
PRT PRT
CZE CZE KOR

GRCSVN GRCSVNKOR
EST EST PRT PRT

POLCHL
RUS RUS
KAZ POLCHL
HUN
BRA
TUR MYS
MEX
CZE
EST CZE
EST
ROU
BGR ZAFCHN
BLR CHL
RUS
POL RUS
BRA CHL
POL HUN
PER
ALB THA
JOR
UKR
IDN
PHL
MAR TUR
KAZ
MEX
ROUMYS
MDA
GHN
VNM IND ZAF
BLR
PER BGR
CHN
0
THA
ALB
UKR
IDN
MAR
PHL JOR
MDA
GHN
IND VNM
0
20 40 60 80 0 50 100
LMI UMI LMI UMI

200004000060000
200004000060000
BRA
TUR
MYS HUN
KAZ HUNMYS
BRA
TUR
MEX KAZ
ZAFMEX
ROU
BLRBGR
ROU
BGR
PER ZAFCHN
BLR UKR
IDN
MAR
PHL
PER CHN
THA
ALB JOR
UKR
IDN
MDAPHL
MAR ALB THA
JOR MDA
GHN
INDVNM
0
GHN
VNM IND
0
0 50 100 0 50 100
20 40 60 80 20 40 60 80
Index C : Index of interaction with global economy
Index of technology upgrading
Fitted values gnipc
Fitted values gnipc Graphs by incomegroup
The relationship between Index B: breadth of technology upgrading which is about structural
features of technology upgrading and income levels is also very similar to Index A. However, the
relationship is less positive. In particular this is due to much stronger structural change in middle
income group which is not converted in levels of income. Especially, upper middle income group
weakens regression fit and thus indicates a kind of structural change specific to middle income trap.
By this we mean that intensive structural change has not yet been converted into appropriate levels
of income.
Unlike previous indexes Index C: interaction with global economy which is based on proxies of
knowledge and technology exchanges has significant and positive relationship with income levels
but very low R2. Also, regression is not robust due to several ‘outliers’ (Ireland, Belgium and
Hungary). However, the positive relationship which we could observe in the high income group is
21
also much less present in the case of this index. As before, we do not see any clear relationship
between knowledge and technology exchange intensity and middle income levels group. This
suggest that the interaction with the global economy in terms of technology and knowledge
exchange is very much country and not income specific. A strong technology and knowledge inflow
may operate as substitute or as complement to own technology efforts recorded in other three
indexes. In that respect, it would be mistake to integrate this index into index of technology
upgrading as this index is moderating variable which can either magnify or reduce the effects of own
technology upgrading. The reason for this is that indicators of technology and knowledge exchange
are neither measuring the true scale of knowledge inflow/outflow nor indicate whether imported
knowledge has been absorbed. FDI can be extensive but contribute little to technology transfer.
Licensing payments are very often proxies for the extent of transfer pricing (which can explain why
Ireland, Hungary and Belgium are outliers) rather than true proxies of knowledge flows. Share of
complex industries in exports is good proxy for integration in global value chains but this does not
tell us about the value added levels of export. For these reasons we consider this index to be
moderating variable which can either magnify or reduce or compensate the effects of technology
upgrading on growth.
Figure 6. Sub-indexes of intensity and types of technology upgrading by GNI per capita across
income groups.
200004000060000
200004000060000
SWE SWE
USA USA SWE SWE
AUS
JPN AUS
JPN USA USA
DEU BEL DEU BEL AUS
JPN AUS
JPN
IRL GBR IRL GBR DEU
BEL DEU
BEL
ITA ITA IRL GBR IRL GBR
ESP ESP ITA ITA
GRC SVN PRT KOR GRC SVN PRT KOR ESP ESP
CZE
EST CHL CZE
EST CHL GRC KOR
SVN GRC KOR
SVN
RUS POL RUSMEX

KAZ POL
HUN
BRA
TUR MYS PRT
CZE PRT
CZE
PER ROU
ZAF
BGR
CHN BLR CHL EST CHL EST
ALB
UKR THA
JOR
IDN POL
RUS POL
RUS
BRAHUN
GHNMAR
MDA
VNM PHL
IND KAZTUR
MEX MYS
0
ROU ZAF
BGRBLR
PERCHN
THA
ALB
MAR JOR
UKR
IDN
PHL
MDA
GHN IND
VNM
0
0 5 10
0 5 10 15
LMI UMI
200004000060000
LMI UMI
200004000060000
HUN
KAZROU BRA
MEXTUR MYS
PER
ALB ZAF
BGR
THA
JOR CHN BLR
UKR
MDA
VNM
GHN IDN
MAR
PHL
IND
0
TURBRA
KAZ
MEX MYSHUN
ROUZPER
BGR AF
BLR
CHN
THA
ALBJOR
0 5 10 0 5 10 MDA
GHNIUKR
IDN
MAR
PHLND
VNM
0
Index : production capability

0 5 10 15 0 5 10 15
Fitted values gnipc Index : R&D and knowledge intensity
Fitted values gnipc
LHI UHI Total

200004000060000
BEL
AUS
SWE
USA
JPN
DEU BEL
AUS
SWE
USA
JPN
DEU
In continuation, we explore the relationship
IRL
GBR IRL
GBR
ITA ITA
GRC
ESP
KOR GRC
ESP
KOR between six sub-indexes of technology
SVN
PRT SVN
PRT
CZE
EST CZE
EST
CHL
RUS
POL CHL
RUS
POL
HUN
BRA
MYS
MEX
KAZ
ROUTUR
ZAF
BGR
BLR
PER CHN
upgrading and income levels (Figure 6).
THA
JOR
ALB
UKR
IDN
PHL
MAR
MDA
GHN
IND
VNM
0
0 5 10 15
LMI UMI Production capability is the key precondition

200004000060000
for further technology upgrading and good

HUN
BRA
MEX
KAZ
MYSTUR
ROU
ZAF
production capability is an important
BGR
BLR
PER
THA
ALB
JORCHN
UKR
IDN
PHL
MAR
MDA
IND
VNM
GHN
0
0 5 10 15 0 5 10 15 precondition for further growth of middle

Index 2 : technology capability
Fitted values gnipc
income economies. High income economies
require also high levels of production
capability but in the context of pervasive
presence of global value chains they can
relocate manufacturing to low costs locations.
Index of production capability is positively

related to income levels but this relationship
22
is (as would be expected) much less robust Index of R&D capability shows positive and
due to strong production capabilities in significant relationship with income levels of
middle income economies. Unlike for the countries. As expected, this relationship is
overall index, the relationship between strong and positive in the case of high income
income level and index of production groups but much less clear or with not
capability is much weaker for upper high- relationship in the case of middle income
income economies. In fact, the relationship is group. Unlike with previous indexes we
much stronger for lower upper income and observe here relatively stronger R&D
there is even some positive relationship for capability of lower upper income economies
upper middle income economies. than would be expected. So, instead of middle
income trap we observe lower upper income
This is also index where regression fit is much R&D capability trap. When compared to
weaker due to very strong production technology capability R&D index shows much
capabilities of middle income economies stronger capabilities of middle income
which is at levels of upper high income economies and lower upper high income
economies. In that respect, the gap between group. This is largely due to dual face of R&D
the level of capability and income levels is the which operates not only as factor of
biggest which denotes a kind of production technology frontier activities but also as factor
capability middle income trap. of absorptive capacity.
Index of technology capability show positive Quality of infrastructure (human and physical)
relationship with income levels but this and its upgrading is one of important
relationship is actually significant only in the structural features of technology upgrading
case of upper high income economies. Middle (Ozawa, 2009) (see Figure 7). Index of
income economies have very low levels of infrastructure is positively related to income
technology capability which in this case is levels and positive for upper high income and
largely about disembodied patent based much less for lower high income and
knowledge. The relationship between completely flat for middle income countries.
technology capability and income levels would The levels of infrastructures of middle income
be highly positive for lower upper middle economies are very often equal to those of
income group if it was not for Korea which is low high income group while Korea’s
again ‘outlier’. Given its income level Korea infrastructure index is the second highest. So
has exceptionally high, in fact the highest it seems that the middle income trap is among
index of technology capability. A seemingly the highest in the case of infrastructure.
positive relationship between income and Educated young population, developed R&D
production capability for middle income and physical infrastructures do not necessarily
group turns into not relationship for convert into proportionate levels of income.
technology capability.
23
Figure 7. Sub-indexes of breadth of technology upgrading by GNI per capita across income groups
200004000060000
200004000060000
SWE SWE SWE SWE
USA USA USA USA
AUSDEU
JPN AUSDEU
JPN
BEL BEL DEUAUS JPN
BEL DEUAUS JPN
BEL
IRL GBR IRL GBR IRL IRL
ITA ITA GBR GBR
ESP ESP ITA ITA
KOR KOR ESP ESP
PRTGRC
SVN PRTGRC
SVN GRC KOR GRC KOR
SVN SVN

CZEEST CZEEST PRT PRT
CHL
RUS
POL BRATUR CHL
RUS
POL
HUNMYS CZE
EST CZEEST
ZAF KAZ
MEX ROU CHN
BGR RUSPOL CHL RUSPOLHUN CHL
BRA
PERPHL
ALBTHA
IDN JOR
UKR BLR KAZ
ROU MEXTURMYS
VNMMDA
GHN M AR
IND BGR PER
CHN ZAF BLR
0
ALB THA
UKR
IDN
PHLJOR
GHNMAR
MDA
VNM IND
0
4 6 8 10 12 0 5 10 15
LMI UMI LMI UMI
200004000060000
200004000060000
BRATUR
MEX
KAZ HUNMYS
ZAF ROU CHN
BGR
PER ALBTHA BLR HUN
PHL
MDA
GHN IDN
M AR UKR JOR KAZ BRA
MEXTURMYS
VNM IND ROU
THAZAF
0
BGR PER
CHN
ALB JORBLR
MDA
VNM UKR
IDN
GHNMARPIND
HL
0
4 6 8 10 12 4 6 8 10 12 0 5 10 15 0 5 10 15
Index 4 : infrastructure_human capital and physical and organisational Index 6 : firm level capabilities
Fitted values gnipc Fitted values gnipc
LHI UHI Total

200004000060000
SWE SWE
USA USA
AUS
DEUJPN AUS
DEUJPN
BEL BEL
IRL GBR IRL GBR
ITA ITA
ESP ESP
SVNGRCKOR SVNGRCKOR
PRT PRT
ESTCZECHL ESTCZECHL
RUS
POL HUNRUS
POL BRA
KAZ TURMYS
ROU MEX
ZAF
PERBGR
BLR CHN
ALB
GHN VNM PHL THA
JOR
IDN
MDA IND UKR
MAR
0
4 6 8 10 12
LMI UMI
200004000060000
HUN
KAZ BRA
TURMYS
ALB BGRTHA MEX
ROU
BLR
PER
JOR ZAF
CHN
IDN
PHL
GHN VNMMDA IND UKR
MAR
0
4 6 8 10 12 4 6 8 10 12
Index 5 : lstructural change
Fitted values gnipc
Index of structural change is the one where the middle income trap seems the biggest. Relationship with
income is not linear but closer to exponential which suggest that there is very intensive structural change
taking place in middle income economies. This change is both on supply side as recorded by diversification of
patent groups and levels and changes in the availability of the latest technologies as well as on demand side
as recorded by levels and changes in the levels of buyers sophistication. Still, intensive structural change is
just one of components of technology upgrading which needs to be accompanied by increased intensity of
technology upgrading activities (production, technology and RD capabilities).
An important dimension of structural change in technology upgrading is the firms’ organisational

capabilities. These are proxied by relative number of Forbes2000 firms per mn pop and by subjective
assessment of firms’ absorption capabilities. In overall, the relationship is positive and significant as well as
for high income countries. This suggests that technology upgrading capabilities are not sufficient without
organisational capabilities or complementary assets (Teece, 1986). The relationship for middle income
groups is almost flat which suggest the large diversity of organisational capabilities at that income levels but
also for some upper middle income countries there seems to be organisational capabilities middle income
trap, i.e. strong organisational capabilities are not reflected in their income levels. Equally, there are lower
upper income economies like Italy which do not have organisational capabilities as would be expected given
their income levels.
In continuation, through a series of OLS regressions we test econometrically robustness of the relationship
between indexes and income levels (Table 3). However, we add dummy variable for different income groups
so that we can detect what visually seems present – technology related middle income trap phenomenon.
24
Model 1 shows results of regression of Index of technology upgrading on income levels with separate
dummies for each of four income groups. We are interested whether dummies are significant. If dummies
are negative that would indicate the presence of income trap for that particular group or some unspecified
reason that the level of index for that group is not corresponding to the income level for the whole sample.
If positive, it would indicate the group specific factor which makes the income level to be higher than
expected given the technology upgrading level of the group. Results confirm what we could visually observe.
First, dummy for aggregate middle income group is significantly negative suggesting the existence of
technology middle income trap. Dummy for upper high income group is also significant but positive
suggesting a kind of upper high income bonus or level of income which does not reflect the level of
technology upgrading as captured by our index. This may be expected as our index is aimed to measure
progression from middle income to high income rather than explain differences among the high income
economies. Dummy variables for middle income subgroups are also negative but not significant which
suggest that the middle income phenomenon is not easily detectable and is diffused across several
components of index of technology upgrading. In addition, dummy for lower high income group is also
negative but highly not significant suggesting that the phenomenon of middle income trap may be also much
more dispersed across income groups and should not necessarily correspond to middle income groups as
defined in this paper. However, inclusion of dummies in OLS shows that technology upgrading is closely
correlated to income levels as R2 is very high.7 It is the highest in regression with the upper high income
group which may be expected given the increasing role that technology plays in growth as countries incomes
rise.
OLS regression with index of technology and knowledge exchange carries much less explanatory power as R2
drops to much lower levels. However, in all models coefficient on index is significant and dummies for lower
middle income and upper high income are as in the previous model. Dummy for middle income group is
negative suggesting that the levels of technology and knowledge exchange for this group are lower than
would have been expected given the relationship. Also, dummy for upper high income group is positive
suggesting that the level of knowledge exchange is higher than would have been expected given the
relationship. Dummies for middle income subgroups show that only upper middle income group has
significant negative dummy which may indicate that the issue of middle income trap with respect to
technology and knowledge exchange is strongly present in this groups rather than in the lower middle
income group. Also, coefficient for dummy for lower high income group is positive but not significant. Also,
R2 is much higher for model with middle income dummy and with upper high income dummy which again
suggest that for the middle income group the gap in terms of technology and knowledge exchange is quite
significant. In other words, the issue of how these economies relate their technology upgrading activities to
interaction with the global economy is one of important factors in explaining their lower than expected
levels of income.
7
Results of regressions without dummies are presented in Appendix E.
25
Table 3. OLS regressions for Index of technology Upgrading (ITU) and Index of technology and knowledge
exchange (ITKE).
Model1 Model 2
1.0 1.1 1.2 1.3 1.4 1.5 2.0 2.1 2.2 2.3 2.4 2.5
ITU 956.7 732 936 905 963 557 478.8
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.001)
ITKE 247 381 437 491 41
(0.01) (0.006) (0.001) (0.001) (0.611)
constant -16603 -3456 -15545 -13166 -16400 -6466 7531.4 24901 12249 13199 6215 9449
(0.000) (0.483) (0.000) (0.000) (0.000) (0.010) (0.041) (0.000) (0.003) (0.001) (0.118) (0.000)
Dummy MI -9327 -22796

(0.003) (0.000)
Dummy LMI -1680 -13917
(0.606) (0.016)
Dummy UMI -4410 -13399
(0.076) (0.003)
Dummy LHI -1833 4475
(0.491) (0.397)
Dummy UHI 19471 34133
(0.000) (0.000)
Number of 42 42 42 42 42 42 41 42 42 42 42 42
observations
F-test sig. 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
2
R 0.81 0.85 0.81 0.83 0.81 0.92 0.25 0.69 0.35 0.40 0.26 0.80
2
Adjusted R 0.81 0.84 0.80 0.82 0.80 0.91 0.23 0.67 0.32 0.37 0.22 0.79
Note: in brackets p value for t test.
Table 4 shows OLS results for Index A: Intensity of technology upgrading which is composed of three sub-
indexes of production, technology and R&D capability. Results for Index of intensity of technology upgrading
are quite similar to the overall index of technology upgrading. Dummy for middle income groups is
significantly negative and positive for upper high income group. Dummies for other subgroups are negative
but not significant. Signs for index of production capability dummies are same as for the Index of intensity of
technology upgrading except for dummy for upper middle income group where coefficient is negative and
significant at 5% level. This suggests that middle income production capability trap is located within the
middle income group but more likely within the upper middle subgroup. In other words, given their
production capability levels this group records the level of income which is below expected. Regression with
income and technology capability shows negative and significant dummy coefficient for aggregate middle
income group and for lower middle income group. So, the issue of technology capability middle income trap
seems to be more relevant for lower middle income group. For lower upper income group negative and
insignificant coefficient for production capability remains insignificant but positive for technology capability.
Regression for R&D capability generates similar results as for production capability but with two important
differences. First, coefficient for middle income group is again negative but is significant only at 5% level.
This suggests that R&D plays important role in absorptive capacity of middle income economies. It explains
why results for R&D capability are more similar to production capability than to technology capability.
Regression with Index B: Breadth of technology upgrading shows that dummy for middle income group is
significant and negative. However, it is also significant and negative for upper middle income group which
suggest that middle income trap is more related to structural change issues in the subgroup of upper middle
income economies. This index is positive for high income group but is significant only for upper high income
group. Index with infrastructure proxies also shows negative and significant dummy for middle income group
26
but without clear indication of location of trap between two middle income subgroups. The coefficient for
high income group dummy is positive and significant for upper high income group but not significant and
negative for lower upper high income group. We get identical results in terms of signs and significance for
structural change index. Middle income trap here is also largely located in upper middle income group.
Finally, regression with the firm level organisation capabilities shows very strong middle income trap for
both middle income groups though much stronger in upper middle income group. Also, dummies for both
high income groups are positive but significantly only for upper high income subgroup.
A series of OLS regressions with middle income dummy enable us to compare coefficients on middle income
dummies and get ranking of sizes of these coefficients (Table 4). The higher the coefficient at similar level of
robustness (t-test) of dummy variable the higher the middle incomes trap in that respective dimension of
technology upgrading. Data below show dummy coefficients for middle income group ranked in descending
order of size.
Table 4. Comparison of MI dummies across OLS regression models.

Dummy MI Coef. Std. Err. t P>|t| [95% Conf. Interval]
Index C -22795.7 3076.517 -7.41 0.000 -29018.54 -16572.86
Index 5 -20455.11 3400.798 -6.01 0.000 -27333.87 -13576.34
Index 6 -17819.95 2624.515 -6.79 0.000 -23128.53 -12511.36
Index 2 -16621.22 2980.479 -5.58 0.000 -22649.81 -10592.63
Index 4 -15424.98 3769.616 -4.09 0.000 -23049.74 -7800.209
Index 1 -14563.4 3965.638 -3.67 0.001 -22584.66 -6542.135
Index 3 - 8210.954 3286.42 -2.50 0.017 -14858.36 -1563.543
The biggest coefficient is in regression with index C: technology and knowledge exchange followed by index
of structural change. The highest index on middle income dummy for index of technology and knowledge
exchange suggests that given the level of knowledge and technology exchange income of middle income
countries are lower by 22795 USD per capita. This means that middle income economies are not benefiting
from being engaged in global technology and knowledge exchange as much as they should compared to
other income groups. In nutshell, this suggests that they should better use their existing levels of technology
and knowledge exchange, i.e. make them complementary to their own technology generation and
absorption.
27
Table 5. OLS regressions for Index A: Intensity and types of technology upgrading and its sub-categories.
Model 3 Model 4 Model 5 Model 6
3.0 3.1 3.2 3.3 3.4 3.5 4.0 4.1 4.2 4.3 4.4 4.5 5.0 5.1 5.2 5.3 5.4 5.5 6.0 6.1 6.2 6.3 6.4 6.5
Index A 1676.5 1314 1624 1608 1703 984
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Index1: Production Capability 4666.9 2728 4557 4263 4874 2266
(0.000) (0.001) (0.000) (0.000) (0.000) (0.000)
Index2: Technology capability 4200 2583 3810 3835 4192 1717
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Index3: R&D capability 3939 3085 3789 3803 4021 2295
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
constant 3247.9 5823 -2102 -1342 -2805 1197 -5646.4 12001 -4827 -1019 -5405 884 9586.8 21788 12415 12406 9193 8656 -4231.1 5009 -2797 -2618 -3770 642
(0.100) (0.143) (0.385) (0.596) (0.158) (0.396) (0.084) (0.037) (0.260) (0.782) (0.094) (0.646) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.044) (0.236) (0.267) (0.335) (0.070) (0.659)
Dummy MI -8311 -14563 -16621 -8211

(0.012) (0.001) (0.000) (0.017)
Dummy LMI -2566 -1416 -10926 -3171
(0.421) (0.765) (0.008) (0.324)
Dummy UMI -2955 -7309 -5932 -2397
(0.247) (0.029) (0.101) (0.362)
Dummy LHI -3250 -5361 1707 -3839
(0.219) (0.152) (0.655) (0.153)
Dummy UHI 19508 25439 26563 19848
(0.000) (0.000) (0.000) (0.000)
Number of observations 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42
F-test sig. 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
R2 0.81 0.84 0.82 0.82 0.82 0.92 0.64 0.73 0.64 0.68 0.65 0.89 0.61 0.78 0.67 0.63 0.61 0.85 0.81 0.83 0.81 0.81 0.82 0.92
Adjusted R2 0.81 0.83 0.81 0.81 0.81 0.92 0.63 0.72 0.62 0.66 0.63 0.88 0.60 0.77 0.65 0.61 0.59 0.85 8.80 0.82 0.80 0.80 0.81 0.92
28
Table 6. OLS regressions for Index B: Breadth of technology upgrading and its sub-categories.
Model7 Model 8 Model 9 Model 10
7.0 7.1 7.2 7.3 7.4 7.5 8.0 8.1 8.2 8.3 8.4 8.5 9.0 9.1 9.2 9.3 9.4 9.5 9.1 10.1 10.2 10.3 10.4 10.5
IndexB: 1951.5 1307 1862 1801 1950 951
Breadth of (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
technology
upgrading:
structural
features
Index4: 5058.6 2895 4640 4630 5184 2306
Infrastructu (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
re: Human
and physical
Index5: 5234.5 2445 4475 4801 5264 2061
Structural (0.000) (0.007) (0.000) (0.000) (0.000) (0.001)
change
Index6: Firm 3538.6 2261 3182 3229 3594 1248
organisation (0.000) (0.000) (0.000) (0.000) (0.000) (0.005)
al
capabilities
constant -27892 -5327 -25179 -21736 -27936 -9578 -21897 3461 -17432 -16466 -22009 -5956 -34144 4658 -25199 -25499 -34207 -9188 -2263 14683 1452 2945 -3777 4681
(0.000) (0.400) (0.000) (0.000) (0.000) (0.021) (0.000) (0.651) (0.006) (0.009) (0.000) (0.097) (0.002) (0.637) (0.044) (0.010) (0.002) (0.116) (0.478) (0.000) (0.683) (0.387) (0.265) (0.038)
Dummy MI -13752 -15425 -20455 -17820

(0.000) (0.000) (0.000) (0.000)
Dummy LMI -3349 -6275 -7525 -8967
(0.405) (0.159) (0.196) (0.044)
Dummy -7412 -5824 -12204 -9703
UMI (0.012) (0.103) (0.002) (0.005)
Dummy LHI 310 -3631 -945 5056
(0.925) (0.343) (0.842) (0.198)
Dummy UHI 23454 26001 30061 27719
(0.000) (0.000) (0.000) (0.000)
Number of 41 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42
observation
s
F-test sig. 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
R2 0.71 0.81 0.71 0.75 0.71 0.88 0.62 0.73 0.63 0.64 0.62 0.87 0.41 0.69 0.43 0.54 0.41 0.85 0.58 0.81 0.62 0.66 0.59 0.84
Adjusted R2 0.70 0.80 0.70 0.74 0.69 0.87 0.61 0.72 0.62 0.62 0.60 0.87 0.39 0.68 0.40 0.51 0.38 0.84 0.57 0.80 0.60 0.64 0.57 0.83
29
The second biggest coefficient on middle income dummy is with index of structural change. Despite
their intensive technology upgrading through structural change middle income economies record
levels of income lower by 20455 units when compared to other income groups. This may not be
surprising given evidence on structural change and growth discussed earlier in the paper.
Technology and knowledge exchange and structural change are therefore two the most significant
dimensions of technology upgrading where middle income trap is visible.
We can label next group of variables as of medium importance in terms of middle income trap.
Organisational firm level capability, technology capability, infrastructure and R&D capability are
dimensions of technology upgrading where middle income trap is of medium importance.
Coefficients on middle income dummies for these dimensions of technology upgrading are in
between 17819 (organisational firm level capabilities) and 14563 (production capability). Finally, the
smallest dummy coefficient is for R&D capability which is significant only at 5% level.
In overall, middle income trap is present in all dimensions of technology upgrading but its
importance vary across different dimensions. A trap seems to be higher for dimensions of breadth of
technology than for index of intensity of technology upgrading. Index of intensity of technology
upgrading reflects cumulative technology capability while index of breadth of technology upgrading
refers to structural, infrastructural and organisational features of economies. These latter are
subject to various market and system failures and are outcomes of variety of non-technological
factors the most important of which seem to be political economy of a specific economy.
4.2. Ranking of countries based on index of technology upgrading and index of

technology and knowledge exchange
In this section we first explore ranking of selected 42 economies based on index of technology
upgrading and related indexes based on income groups. For this comparison we use box plot
diagrams (in Figures 8 to 15) as convenient way to present spread of countries of different income
levels across nine indexes. Second, we specifically explore positioning of the central and eastern
European economies on the map of technology upgrading. Third, multidimensional indicators give us
opportunity to explore profiles of technology upgrading and various pairs and groups of countries.
4.2.1. Ranking of countries based on technology upgrading indexes

Ranking of countries based on index of technology upgrading shows fairly good ordering in terms of
income per capita as has been confirmed by our exploratory OLS analysis in the previous section.
However, boxplots diagrams show clearly those countries which are either below or above their
income group.
Index of technology upgrading does not indicate speed or pace of upgrading as that would require
long time series which are simply not available for the type indicators that are required for
measuring technology upgrading. However, index represents proxy of potential for technology
upgrading. In that respect, a country that is low in terms of income per capita but has high index of
technology upgrading suggest that this country has good potential for growth based on technology
upgrading.
The highest ranked countries in terms of index of technology upgrading are Sweden, Korea, Japan
and Germany. Korea has very high index given its current level of income and thus has very good
30
potential for further growth based on technology. Given its level of income, Italy has modest
potential for further technology upgrading. Similarly Russia has very low potential for technology
based growth given its income level. Although Poland has been fast growing CEE economy in the last
20 years it seems that its potential for technology based growth does not look so bright. Of course,
Poland may grow based on its production capability, and on-non technological factors for some time
but it seems that drivers for technology based growth are far from where they should be given its
recent high growth.
Figure 8. Box plot diagram for Index of technology upgrading by income groups.
Index of technology upgrading

80
SWE
KOR
60
40
itu
ITA
20
0
LMI UMI LHI UHI
China’s ranking in terms of index of technology upgrading is well above its income per capita which
suggests a room for further growth based on technology. For example, in terms of index of
technology upgrading China is above Hungary which has GNIpc of $12K when compared to Chinese
$5.7Kpc. On the other hand, it is quite surprising to see that Belarus whose GNIpc is $6.4K is slightly
ahead of China. However, we should bear in mind that the majority of indicators are used on per
capita basis which clearly underestimate differences in the absolute size of economies and nonlinear
effects that this may generate. The issue of absolute vs relative size is also striking when comparing
India and Ukraine with their respective GNI pc of $1550 and $3640 respectively but with Ukraine
being only slightly ahead in terms of index of technology upgrading even though the majority of
indicators are used in per capita terms. However, the biggest surprises emerge when we compare
two economies which are currently in widely disparate situations. Polish economy has been growing
at higher rates for a long period while Greece has been experiencing the biggest downfall in its
recent economic history. Yet, in terms of index of technology upgrading or potential for further
growth based on technology Greece is slightly ahead of Poland. We should bear in mind that still in
2012 Greece’s GNIpc was almost double of Poland ($23.6K vs $12.67K) so relatively in terms of index
of technology upgrading Poland is doing comparatively much better and thus has potential for
31
further growth based on technology. On the other hand, Greece’s GNIpc is just $1K lower than
Korean which is second economy in our sample in terms of index of technology upgrading,
immediately after Sweden. This obviously reflects widely different structures of these two
economies, like in terms of share of services and the role of industry, but also widely different
potential for further economic growth based on technology.
In exploring these seeming puzzles, our analysis could extend well into analysis of each individual of
42 economies. However, in each of these cases it is important to bear in mind that index of
technology upgrading is measuring potential for growth based on technology broadly defined which
includes not only R&D but also production capability and several major structural features of
economies. Thus the purpose of index is not to forecast economic growth but only component of
growth which is technology related. A variety of other factors which are non-technology related can
influence growth, especially in a short and medium term and this should be kept in mind when
interpreting results.
In continuation we will not analyse two subcomponents of index of technology upgrading – intensity
and breadth – which we will analyse only in relation to the CEE economies in the next section. It
suffices here to point only to two issues. First, there are much bigger differences in ranking among
countries in terms of intensity of technology upgrading than in terms of breadth index or
infrastructure, structural change and firm level organisation capabilities. The ordering of countries
within specific income groups is much more regular i.e related to income per capita for intensity of
technology upgrading than for breadth of upgrading. This is especially the case from upper middle
income group upwards and less for lower middle income group. This may be expected as technology
and its cumulative features play much more important role in income generation the closer is the
country to technology frontier. Second, a much stronger convergence in terms of breadth of or
scope of technology upgrading is due to smaller differences in infrastructure but especially due to
intensive process of structural change which takes place at different levels at different income
groups. So, much of differences in terms of breadth of technology upgrading are largely due to
remaining third component - differences in terms of firm level organisational capabilities. It is with
respect to firm level organisation capabilities that we have on average the biggest differences within
(not between) different income groups.
A third component of technology upgrading is the interaction with the global economy or
technology and knowledge exchange (see Figure 9). However, this component is not part of index of
technology upgrading. Unlike scale and scope of technology upgrading components which
cumulatively individually and as a pair contribute to index of technology upgrading this component
does not add to index in cumulative manner. In other words, increased technology and knowledge
exchange with the global economy does not necessarily increases potential for growth based on
technology. This is for two reasons. First, proxies for technology and knowledge exchange do not
necessarily measure true flows of knowledge and technology. An inflow and outflow of FDI relative
to GDP very poorly correlates to the true to ‘amount’ of technological knowledge transferred.
Receipts and payments in technology balance are also poor proxy of true knowledge transferred and
they are far too often influenced by transfer pricing practices of MNEs. Also, structure of exports by
technological levels hides big differences in value added levels which are much closer proxy of
technological knowledge but which are not widely available. Second, even when proxies reflect well
‘the amount of true knowledge transferred’ that does not mean that this knowledge has been
32
absorbed in the economy. When it is absorbed it is actually already reflected in index of intensity of
technology upgrading in activities of country in terms of production, technology or R&D capability.
So, in a way this would be double counting. This same applies when knowledge is exported it is
already embodied in index of technology upgrading. Imported knowledge is never simply added or
substituted for the lacking domestic knowledge but it instead interacts with domestic technology
effort creating virtuous or vicious circles in terms of growth and technology upgrading.
Figure 9. Box plot for index of interaction with global economy across income groups
Index of interaction with global economy

100
IRL
80
HUN
60
BEL
indexc
40
20
VNM
GRE
0
LMI UMI LHI UHI
In view of this it becomes clear why intensity of technology and knowledge exchange does not form
part of index of technology upgrading but instead operates as moderating index, i.e. as index which
either amplify or reduce or is neutral in effects on index of technology upgrading on growth. This
also explains why the correlation between income levels and intensity of technology and knowledge
exchange index is much different from indexes of intensity of technology upgrading. This link is quite
weak, both overall and within four income groups and with three major outliers (Ireland, Hungary
and Belgium). These outliers are largely due to high share of FDI, both in inflows and outflows, and
high share of technology balance payments and receipts which reflect transfer pricing related to high
levels of FDI. Still, even if we ignore outliers the intensity of technology and knowledge exchange is
on average related to income levels of income groups but not within groups. Fourth outlier is Greece
which despite being low high income economy is on 40th place in terms of index of technology and
knowledge exchange in sample of 42 economies. This is largely due to very low share of inward FDI
and related to that a low share of technology payments in GDP (see Appendix D).
In continuation we briefly comment each of six indexes that form index of technology upgrading.
Production capability index (see Figure 10) is composed of ISO9001 certificates pmi 2007-11 avg,
Trademark applications, resident pmi 2007-11 avg and subjective assessment of On the job training
2012-13. These are proxies of capability to produce at required level of quality, not proxies to
generate innovations of any kind. However, this is actually the important area of technological
activity especially for middle income economies that do not compete on technology. In overall, there
is good correlation with income levels but the most interesting results are about differences within
specific income groups. China’s production capability index is below Belarus and Malaysia and
33
above Hungary. Obviously being ‘factory of the world’ with its capacity to meet production
requirements of foreign buyers does not yet apply to the whole China. A dual economy syndrome
seems obvious in the case of Hungary who with two times higher GNIpc compared to China and
being among the three most intensive economies in terms of ‘technology and knowledge exchange’
has quite low production capability index. An interesting picture emerges when we compare Russia,
Poland and Greece. In terms of production capability index Greece is even slightly ahead of Russia
which indicates very low level of on the shop floor and service quality and training activities in
Russia. Poland is well above Russia though with similar GNIpc. However, in terms of production
capability Poland is still trailing behind its Central European peers like Estonia, Slovenia and Czech
Republic. Within upper high income group Italy is above Ireland in terms of index of production
capability. Although Italy is not host to FDI it has the biggest per capita number of ISO
Certificates9001 and surprisingly small for high income economy (but similar to Ireland) number of
trademarks pc. On the other hand, Korea is at the top in terms of trademarks per capita which
indicates progress towards more non-production activities in the value chain.
Figure 10. Box plot for index of production capability across income groups.
Index of production capability
BEL
10
index1
IRL
5
0
LMI UMI LHI UHI
Figure 11. Box plot for index of technology capability across income groups.
Index of technology capability
15
KOR
10
index2
TUR
CHN
0
LMI UMI LHI UHI
Technology capability (see Figure 11) is heavily concentrated in high income economies. This is the
component of index which is the most dispersed and which shows that technology generation is
present only much later in development. Among the upper middle income economies Turkey and
34
China are very active given their income levels. Among lower high income group Korea is not only
absolute leader abut also at the top of technology capability index. This once again indicates Korean
high potential for technology based growth. Belgium, being location for many MNEs, is top country
in terms of production capability but it ranks much lower among the highest income group in terms
of technology capability, but still above Ireland and UK.
R&D capability (see Figure 12) ranking is not following technology capability rankings which suggest
that R&D in its absorptive capacity function plays important in countries at all income levels. As data
for this index are the most available we use nine indicators to ensure good approximation of this
capability. These are hard and subjective data but they are all highly correlated. There is clear
ordering of the income groups by R&D capability but with interesting variations within the specific
groups. Among the low-middle income group Vietnam ranks highly despite its very low technology
and production capability. Among upper middle income group China and partly Peru rank high given
their income levels of around $5kpc. Hungary which has GNIpc of $12k pc is actually ranked lower
than would have been expected. However, given their income levels the biggest underperformers in
terms of R&D capability are Chile, Poland, Russia and Greece and among upper high income group
Italy. Sweden is by far the top performer. Korea and Slovenia are also very high performers given
their income levels. However, similar performance in terms of R&D capability between Korea and
Slovenia is strikingly different in terms of technology capability where Korea is the top performer. In
the case of technology capability, Slovenia is still at the top of the CEE range.
Figure 12. Boxplot for index of R&D capability across income groups.
Index of R&D capability
15
SWE
10
index3
ITA
5
0
LMI UMI LHI UHI
Composite indicator of human and physical infrastructure (see Figure 13) is composed of proxies for
schooling, quality of education, R&D infrastructure, Internet diffusion and physical investment.
Differences in ‘infrastructure endowments’ are much smaller across different income groups i.e
distribution of index is quite compressed. This means that countries at middle income level invest
beyond of what would be expected given their income levels. In upper middle income group Belarus
has level of infrastructure as upper high income group. Korea has level of infrastructure which puts it
at second place overall, behind Belgium. The major under investments in infrastructure are in South
Africa which despite being upper middle income country has the lowest index of infrastructure in the
sample. The biggest contributors to compressed distribution are two human capital indicators:
average years of schooling and quality of maths and science education.
35
Figure 13. Box plot for index of human and physical infrastructure
Index of human and physical infrastructure
KOR
12
10
index4
ITA
8
6
4
LMI UMI LHI UHI
Figure 14 Boxplot for index of structural change.

Index of structural change
12
10
index5
8
6
4
LMI UMI LHI UHI
Structural change index (see Figure 14) is composed of proxies of technology diversification
(concentration) of domestic resident, US and EPO patenting and subjective assessments of levels and
changes in demand and supply. Level of diversification and concentration in patenting structure
should reflect levels of transition trends towards increasing diversification in number of patenting
categories of countries transiting from middle to high income status. Changes of upper high income
group may go in both directions, towards diversification or concentration. So, in that respect
direction of diversification at the income level is indeterminate. However, we believe that use of
levels of concentration/diversification is a fair reflection of levels of technological upgrading.8 9 In
addition to patents which have their weaknesses as proxy for countries behind technology frontier
we also use subjective assessment of levels and changes of quality of demand and supply. Levels of
both these proxies are highly correlated to income per capita. However, in order to measure
8
In the next version of this working paper we will add also changes in the concentration/diversification of three types of
patents (WIPO residents patents, USPTO and EPO).
9
Correlation coefficient between levels of GNIpc and HH index of concentration is – 0.3, - 0.45 and -0.38 for WIPO, EPO
and USPTPO respectively. The higher the income the more dispersed is the patent portfolio of countries.
36
structural change we combine levels with changes in assessments of demand supply of technology.
As expected middle income economies have significantly lower levels of buyer sophistication and
lower availability of state of the art technologies. However, they have significant catching up
opportunities and hence they lead to in terms of changes in demand and supply of technology. In
2006-2013 period there was large decline in levels of buyers’ sophistication due to faltering growth
caused by global financial crisis 2008. So, only a few countries had increase in assessment of buyer
sophistication. Still, top ten economies in terms of changes in buyer sophistication (demand for
innovation) are middle income economies. The same trend is present in the case of changes in the
availability of the latest technologies where 14 out of 16 economies are middle income.
Firm level organisational capabilities are an attempt to capture organisational side of technological
upgrading. Technology generation is by and large the privilege of large firms and hence we use
Forbes2000 firms per mn population. We intentionally did not want to make distinction and
separate financial (banks) from financial firms as banks are in many countries agents of technology
upgrading through direct or indirect firm ownership. However, there are nine economies in our
sample which do not have Forbes2000 firms and hence we needed to have additional indicator
which would capture organisational capabilities of firms for technology upgrading. The closest
indicator for that aim is subjective assessment of firm level technology absorption. Absorption is not
only function of technological capabilities and skills but it is also an issue of organisational
capabilities. Assumption behind is that competent firms will be better absorbers, not necessarily
better technology generators. So, we combine these two indicators which complement each other
as firm level technology absorption indicator may proxy this capability in sector of medium and small
sized firms. In fact, correlation coefficient between scores for Forbes2000 and firm level technology
absorption is relatively high 0.64. A top two countries in terms of firm level organisational
capabilities are Sweden and Ireland. Sweden is top in terms of subjective assessment of firm level
technology absorption and Ireland in terms of Forbes2000 per mn population. Among individual
income groups it is noticeable a very low ranking of Italy in the group of upper high income
economies largely due to low firm level technology absorption assessment. Among lower high
income group Russia is extremely low placed largely due to the lowest ranking in terms of firm level
technology absorption. So, while Russia is important foreign outward investor among BRICS its
organisational capabilities for technology absorption at firm level are pulling it very much down.
Among this group Korea is ranked well above the average due to high rankings on both dimensions.
Poland’s ranking is in terms of Forbes 2000 similar to Russian but it is better ranked in overall due to
better firm level technology absorption capability which is however still among the bottom group of
countries. Greece is also a specific case as it ranks very high in terms of Forbes2000 due to large
number of big public companies and banks given its size but its ranking in firm level technology
absorption pull it also down in the overall score. China is evenly scored in terms of both Forbes2000
per mn pop and firm level absorptive capacity. In per capita terms, Belarus ranks higher than China
although its firms are not recorded in Forbes2000. So, its ranking is derived entirely from subjective
assessment which is quite high. This is obviously not the most satisfactory as many of subjective
assessments like for Jordan, Malaysia or South Africa are difficult to explain. Any subjective
assessment is highly contextual and should be understood in the context of the existing
technological levels of the country and local demand for technology. For example, ranking of Jordan
who has GNI pc of $4.6K and Spain with $29Kpc are 5.15 and 4.23 respectively. These are
assessments of their technology absorption capacity but relative to the existing level of demand for
37
technology in the country, not some absolute external metrics. Similar to other structural change
indicators these should be understood as potential for technology upgrading taking into account the
current existing levels of technology capability of the economy.
In overall, ranking of countries based on components of index of technology upgrading are quite
revealing as they show potential for technology based growth which sometimes may widely differ
from the recent growth figures or current macroeconomics forecasts. In that respect, index of
technology upgrading may serve as very useful complement in understanding long term drivers of
growth. There are much bigger differences in ranking among countries in terms of intensity of
technology upgrading than in terms of breadth index. This reflects cumulative nature of technology
especially the closer is the country to technology frontier.
Figure 15 Box plot for firm organisational capabilities.

Index of firm organisational capabilities
15
KOR
10
index6
5
0
LMI UMI LHI UHI
4.2.2. Positioning of the central and east european economies on the map of technology
upgrading
In this section we explore, in Figures 16 and 17, positioning of the CEE countries in relation to other
economies in the sample. As detailed descriptive analysis of ranking can be quite exhaustive we
report here only several major conclusions.
First, CEECs are spread widely along index of technology upgrading which reflects partly their income
per capita levels but also different potential for further growth. The leading two CEE economies in
this respect are Czech R and Estonia. Slovenia, whose income ($22.8) is well above both Czech R
($18) Estonia ($16) is ranked below which suggest that its potential for further growth is not really
there to similar extent. Hungary and Poland have similar income per capita ($12k) but widely
different current growth rates. Based on index of technology upgrading Poland does not have really
developed technology drivers of growth. However, Russia which has similar income per capita as
Poland seems to be even with the fewer bases for technology based growth.
38
Figure 16 Ranking of Index of technology upgrading and its component index of intensity and
types of technology upgrading by selected 42 countries.
INDEX OF TECHNOLOGY INDEX A: INTENSITY AND TYPES

OF TECHNOLOGY UPGRADING
UPGRADING (SCALE)
0 20 40 60 80 100 0 10 20 30 40 50
Sweden Korea, Rep.
Korea, Rep. Sweden
Japan Germany
Germany Japan
Belgium Belgium
United States Austria
Austria United States
United Kingdom United Kingdom
Ireland Ireland
Spain Slovenia
Portugal Italy
Czech Republic Spain
Estonia Czech Republic
Italy Portugal
Malaysia Estonia
Slovenia Malaysia
Belarus Belarus
Chile China
China Chile
Hungary Hungary
Greece Turkey
Turkey Brazil
Brazil Poland
South Africa South Africa
Poland Greece
Jordan Bulgaria
Mexico Russian Federation
Ukraine Mexico
Thailand Jordan
India Peru
Morocco Romania
Indonesia Thailand
Russian Federation Indonesia
Bulgaria India
Philippines Ukraine
Romania Philippines
Kazakhstan Morocco
Peru Vietnam
Albania Kazakhstan
Moldova Albania
Vietnam Ghana
Ghana Moldova
39
Figure 17. Ranking of Index of breadth of technology upgrading and index of interaction with
global economy by selected 42 countries.
INDEX OF INTERACTION WITH
INDEX B: BREADTH OF
GLOBAL ECONOMY
TECHNOLOGY UPGRADING: 0 10 20 30 40 50 60 70 80 90 100
STRUCTURAL FEATURES (SCOPE)
Ireland
0 10 20 30 40 50 Hungary
Sweden Belgium
Japan Sweden
Ireland United…
United States Austria
Korea, Rep. Estonia
Belgium Jordan
United Kingdom United States
Austria Japan
Germany Germany
Malaysia Korea, Rep.
Belarus
Spain
Portugal
Bulgaria
Spain
Slovenia
Estonia
Chile Malaysia
China Czech…
Czech Republic China
Greece Chile
Jordan Vietnam
Italy Thailand
Ukraine Mexico
Slovenia Portugal
South Africa Italy
Mexico Poland
Turkey Turkey
Brazil Kazakhstan
Hungary Romania
Thailand
Albania
Morocco
Belarus
Poland
Ukraine
India
Russian…
Indonesia
Philippines Moldova
Kazakhstan Philippines
Russian Federation India
Romania Brazil
Bulgaria Morocco
Peru South Africa
Albania Ghana
Moldova Greece
Vietnam Peru
Ghana Indonesia
40
Second, a comparison of the CEECs shows on average higher ranking in terms of index of intensity of
technology than index of breadth of technology upgrading. This is quite important feature of the CEE
as it shows that they are not structurally dynamic economies. Why this is so can be depicted from
rankings in specific dimension of technology upgrading. CEECs are by far the lowest ranked in terms
of index of structural change which is about diversification of technological knowledge and changes
in demand and supply of technology. They are the most compressed and well ranked group in terms
of infrastructure but also group that has very weak firm level organisational capabilities.
Third, important structural feature of the CEE technology upgrading is their openness in terms of
technology and knowledge flows. For example, Hungary is the second ranked overall in terms of
Index of interaction with global economy and Estonia 7th. So, the key issue for their technology
upgrading whether knowledge inflows in these economies operate as substitute or complements to
their own technology activities. All three European CIS economies (Russia, Ukraine and Belarus) are
behind the CEE in terms of technology and knowledge exchange which indicates quite different
growth model. However, Belarus is quite distinct from both Russia and Ukraine as it is ranked quite
high in terms of breadth and intensity of technology upgrading.
Fourth, among indexes of intensity of technology upgrading CEE ranks the best in terms of index of
production capability (see Figure 18). This is not surprising as it is production, not technology
capability (Figure 18), that drove growth in this region during 1990s and 20000s (see Kravtsova and
Radosevic). The most interesting case in this respect is Hungary which is important location for
German and Austrian FDI but which ranks quite low on this index when compared to Czech R and
even Poland. ISO900 certificates are one of three indicators used to construct this index and Hungary
ranks on high 5th place immediately below Czech R and Slovenia. However, in terms of trademarks
per capita and in subjective assessment of on the job training Hungary scores very low. This suggest
that dual economy syndrome is quite pronounced in Hungary where part of economy is
internationally integrated as witnessed by the top position in terms of technology and knowledge
exchange index while the rest of the economy is doing very poorly in terms of technology as it is
ranked very low in terms of firm specific capabilities like trademarks and on the job training. It is
interesting to draw in this respect comparison with Italy which has similar profile of production
capability as Hungary but ranks much lower in terms of index of interaction with global economy.
Italy has much higher number of ISO9900 certificates than Hungary as it is ranked as the top
economy in this respect but has quite surprisingly low number of trademarks per capita despite
known Italian fashion brands and brands in several other sectors like cars, kitchen appliances, and
subjective assessment on the job training which is not much above that of Hungary. Unlike Hungary,
Italy scores very well on index of production capability, also high on index of technology capability, is
similar to Hungary in terms of R&D capability and better on two of three indexes of breadth of
technology upgrading. So, Italy represent good example that interaction with global economy is not
necessary for technology upgrading while Hungary shows limits of ‘hyperintegrationst’ model of
development (see Scepanovic).
41
Figure 18 Ranking of Indexes of production capability, technology capability and R&D capability by
selected 42 countries.
INDEX 1: Production INDEX 2: Technology INDEX 3: R&D

capability Capability capability
0 5 10 15
0 5 10 15 0 5 10 15
Sweden
Belgium Korea, Rep. Austria
Korea, Rep. Japan Belgium
Sweden Germany Germany
Germany Sweden United Kingdom
Austria United States Korea, Rep.
Italy Austria Japan
Japan Italy Slovenia
Czech Republic Belgium United States
Spain United Kingdom Ireland
United Kingdom Ireland Estonia
Chile Spain Czech Republic
Portugal Slovenia Portugal
United States Turkey Spain
Belarus China
Hungary
Malaysia Czech Republic
Italy
Estonia Portugal
China
Ireland Greece
Malaysia
Slovenia Estonia
Brazil
Turkey Hungary
Russian…
China Poland
Ghana Greece
Poland
Moldova Belarus
Brazil
Bulgaria Bulgaria Poland
South Africa Morocco Peru
Hungary Russian… Vietnam
Mexico Belarus South Africa
Romania Ukraine Jordan
Thailand Romania Chile
Philippines Malaysia Ukraine
Indonesia Thailand India
Jordan Kazakhstan Turkey
Greece South Africa Bulgaria
India Brazil Mexico
Russian… Chile Indonesia
Morocco Mexico Thailand
Kazakhstan Vietnam Romania
Peru Indonesia Philippines
Ukraine Jordan Kazakhstan
Albania India Albania
Ghana Philippines Morocco
Moldova Albania Ghana
Vietnam Peru
Moldova
Fifth, CEE ranks much better in terms of production than technology capability especially in terms of
distance to frontier economies. Czech R is 8th ranked in this respect which shows that it is emerging
as at least in the EU context as important production location. CEECs are well behind frontier
economies in terms of technology capability but are comparable to countries of similar income level.
In terms of R&D capability they are quite divided between central Europe which ranks relatively well
and Eastern Europe (Romania and Bulgaria) which are well behind with Poland being in intermediate
42
position. Three European CIS economies (Russia, Belarus and Ukraine) are below the CEECs in terms
of technology capability which suggest big difference between domestic behind frontier technology
effort and world frontier activities. This is largely due to low number of EPO and USPTO but not
domestic resident patents. Russia and Belarus are both better ranked in terms of R&D capability
while Russia and Ukraine are very low in terms of production capability. Belarus again represents
very specific case as it is ranked very high in terms of production capability, below Czech R and
above all other CEE countries.
Sixth, we already pointed out that in terms of breadth of technology upgrading CEECs countries are
ranked behind their positions in terms of intensity of technology upgrading. We use three indexes to
describe breadth or scope of technology upgrading: infrastructure, structural change and firm
organisational capabilities (Figure 19). CEECs are well ranked in terms of human capital and physical
infrastructure but much less well in terms of structural change. As a group they are well placed in
terms of schooling years but not so well in quality of math and science (except Estonia and Slovenia).
These suggest that ‘quantity’ in education is much less issue when compared to ‘quality’. They are
comparatively lagging behind in terms of quality and availability of RD infrastructure and Internet
and investments. It is quite surprising to see that subjective assessments by business community of
availability of research and training services and availability of scientists and engineers for some
CEECs are in the lowest quarter of our sample (cf. Romania, Bulgaria). Similarly assessments are in
this respect in similar group for Ukraine and Russia, countries that had in the past extensive and
developed RTD infrastructure. It seems that as with the firm specific skills these are new factors that
constrain growth and are not any more assets. They are well ranked in terms of Internet (fixed
broadband subscribers) as CEECs are in the top half of distribution. It is very interesting that Belarus
ranks in this respect ahead of Estonia which is usually taken as exemplar of IT diffusion in the region.
We use indexes of diversification/concentration of patent categories as proxy for degree of

structural change of technological knowledge. Assumption is that the higher the income of economy
the more diversified is its technological knowledge across patent categories. The correlation
between WIPO resident patents, EPO and USPTO patents and GNIpc is 0.31, 0.46 and 0.38
respectively suggesting moderate relationship. We also use proxies for levels of demand for
technology (assessment of buyer sophistication) and supply of technology (the availability of the
latest technologies) which are much strongly correlated to income with correlations of 0.61 and 0.71
respectively. We also use changes in assessments of demand and supply which shows potential for
catching as changes in demand and supply are negatively correlated to income, i.e. richer economy
have slower changes due to expected convergence effects as they are already closer to the frontier
in that respect while laggards have more room for catching up. As we are not sure about direction of
structural change for upper high income group in this version of the paper we do not use changes in
concentration/dispersion of technological knowledge. However, based on Lee(2013) we should
expect that as countries move from middle to high income their knowledge base should diversify.
CEEC are scattered across in terms of dispersion of patent categories but their ordering does not
follow size of the countries but differences in knowledge bases. As expected their foreign patenting
is slightly more concentrated than domestic. It is quite surprising that in terms of buyer
sophistication CEECs are all located in the second half of the sample with Hungary, Slovenia and
Estonia having slightly lower or equal assessment of buyer sophistication as Romania and Bulgaria.
As our data are for 2012-13 this ranking largely reflects very depressed demand in these economies
due to post2008 crisis. This is confirmed by similar very low ranking in terms of changes in buyer
43
sophistication in between 2012/13 and 2006/07which has quite deteriorated reflecting decline in
purchasing power and thus decline in demand for technology.
Figure 19. Ranking of indexes of human and physical infrastructure, structural change and firm
organisational capabilities.
INDEX 4: Human and INDEX 5: Structural INDEX 6: Firm

physical infrastructure change indicators organisational
0 5 10 15 capabilities
0 5 10 15
Japan 0 5 10 15
Belgium
Belgium Sweden
Korea, Rep.
Sweden Ireland
Japan
United Kingdom Japan
Sweden
Portugal United States
United States
Austria Korea, Rep.
Germany
Brazil Austria
United Kingdom
South Africa United…
Austria
Germany Germany
Belarus
United States Belarus
Estonia
China Belgium
Ireland
Ireland Malaysia
Malaysia
Mexico Portugal
Czech Republic
Ukraine Jordan
Spain South Africa
Korea, Rep.
China Chile
Chile
Slovenia Spain
Malaysia
Hungary Estonia
Italy
Greece Turkey
Spain
Italy India
Morocco
Jordan Brazil
Greece
Poland Philippines
Czech Republic
Chile Thailand
Turkey
Ukraine Czech Republic
Thailand
Portugal Greece
Poland
Russian… Hungary
Russian…
Romania Indonesia
Slovenia
India Mexico
Estonia
Bulgaria China
Bulgaria
Indonesia Ukraine
Belarus
Thailand Morocco
Kazakhstan Peru
Morocco
Philippines Italy
Kazakhstan
Hungary Slovenia
Mexico
Indonesia Albania
Turkey
Romania Kazakhstan
Albania
Peru Poland
Moldova
Jordan Ghana
Philippines
India Romania
Ghana
Moldova Bulgaria
Brazil
Albania Vietnam
Vietnam
Vietnam Moldova
Peru
Ghana Russian…
South Africa
Their ranking in terms of availability of the state of the art technologies is higher than assessment of
buyer sophistication which reflects opportunities to import technologies. However, their ranking still
places the majority of them (with exception of Slovenia and Czech R) in the second half of
distribution. They are ranked relatively better in terms of changes in this respect between 2006 and
44
2013 which also reflects greater catching up opportunities when compared to technology frontier
economies.
Firm level organisational capabilities are proxied by firm structure (Forbes2000 per mn pop) and
subjective assessment of firms’ absorption capacities (firm level technology absorption). This
unusual combination of indicators actually reflects similar underlying phenomena -capabilities of
firms to organise technology generation and absorption. Correlation coefficient between
Forbes2000 ranking and subjective assessment of firms level of absorption of technology is
surprisingly high 0.62. Also, correlation coefficient between GNIpc and Forbes2000 per mn pop is
very high 0.76 while correlation with subjective assessment of firm absorption capabilities is also
high 0.63.
So, in terms of index of firms’ organisational capabilities the CEE region is placed quite low which in
addition to weak structural change ranks them low on breadth of technological upgrading index
despite higher ranking in terms of infrastructure. All CEE countries except Estonia are in this respect
placed in the second half of distribution. It is quite striking to see that Russia is in this respect the
lowest ranked economy despite having several very large global firms like Gazprom which is the
biggest company in the world and oil majors and large metal mining companies. It is also striking to
see that Belarus is very highly ranked which is actually estimated figure since Belarus is not part of
WEF reporting system and its companies are not included in Forbes list although the country has
many Soviet era large companies. So, in this respect we should take Belarus ranking with a high
degree of caution. Still, low ranking of the CEECs indicate that the important structural feature for
technology upgrading – large and small firms – capable to generate and absorb technology are
important constraint to growth. In terms of Forbes2000 firms per mn pop Poland and Hungary are
placed just above middle of distribution and Poland ranking is identical to Russia. However, in cases
of Poland and Russia high ranking in terms of number of firms that potentially should be the major
organisers of innovation process and technology generation in economy is undermined by very low
ranking in terms of technology absorption or inability to employ organisational capabilities to
effectively absorb technology.10 Some CEECs (Romania and Bulgaria) do not have Forbes2000 firms
which is further indication of their weak organisational capabilities. Other CEECs are quite small
(Slovenia, Estonia) which may explain the absence of their firms in Forbes2000 list.11
4.2.3. Comparative analysis of technology upgrading profiles

In this section we explore technology upgrading through upgrading profiles which are composed of
two indexes of technology upgrading and index of interaction with global economy and then through
six indexes that form index of intensity of technology upgrading and index of breadth of technology
upgrading.
Our sample of 42 economies can be organised in individual pairs but also we compare by grouping
countries into the following regions.: EU North (Sweden, Austria, Germany, Belgium, Ireland and
United Kingdom), EU South (Italy, Spain, Greece, Portugal), EU Central Europe (Slovenia, Czech
10
Greece seems to be in very similar situation in this respect which is good reflection of its competitiveness weaknesses
notwithstanding the crucial role of macroeconomic constraints for its growth nowadays. Also, Italy has similar
structural feature: good presence of large firms but low technology absorption capability of economy.
11
Unfortunately Forbes3000 list is not available.
45
Republic, Estonia, Poland, Hungary), EU Eastern Europe (Romani and Bulgaria), European CIS (Russia,
Ukraine and Belarus), and BRICS (Brazil, India, China, Russia).
EU North and the USA are strikingly similar in intensity and breadth of technology upgrading. These
similarities are visible through six dimensions of technology upgrading where in four out of six
dimensions they are pretty similar (Figure 20). It is only in relation to technology capability that the
US is strongly ahead of the EU North which is ahead in terms of production capability. However, EU
North is much more engaged in international exchange of knowledge and technology when
compared to the US.
This similarity should be understood in the context of index which is based on normalized values so
that USA and EU North are similar only in the context of 42 countries in the sample. They are not
necessarily similar if we were to compare them in absolute terms or only with each other. So, their
similarity is strong only in relation to other countries and groupings of our sample.
Figure 20 Comparison of indexes for EU North and the USA.
Intensity of
technology
Production
upgrading
capability
Firm org. Technology

capabilities capability
Interaction Breadth of EU North EU North
with global technology
economy upgrading United States United States
Structural R&Dcapabil
change ity
ITU: Index Human,

of physical
technology infstr.
upgrading
Figure 21 Comparison of indexes for EU North and Japan.
Intensity of
technology Production
upgrading capability

economy upgrading Japan Japan
Structural R&D
change capability
ITU: Index Human,

of physical
technology infstr.
upgrading
46
EU North is quite similar to Japan in terms of indexes of technology upgrading though Japan’s
indexes are 9 percentages points higher (Figure 21). However, in terms of technology and knowledge
exchange EU North is much more engaged in global economy than Japan. Although in aggregate
similar in terms of index of technology upgrading there are important structural differences between
Japan and EU North. Japan’s technology capability is much higher than EU North and its firms’
organisational capabilities and infrastructure are somewhat at higher level. However, EU North is
ranked higher in terms of R&D capability.
With the 2008 Global financial crisis which was followed by Eurozone financial crisis a deep rift has
opened between EU North and EU South economies. While this may seem at first glance as only
macroeconomic issues our data indicate that the basis for this is deeply rooted in differences in
technology upgrading between EU North and EU South (Figure 22). Index of technology upgrading
of EU South is 66% of EU North. This is composed of much bigger gap of 43 percentage points at
index of technology intensity and smaller gap of 27 percentage points at index of breadth of
technology upgrading. However, the gap in terms of intensity of technology and knowledge
exchange is much wider and basically shows that EU South is very much outside of technology flows
when compared to the EU North. Regarding specific dimensions of index of technology upgrading
there is large gap in terms of technology capability (37% of EU North), firms’ organisational
capabilities (48%) and R&D capability (52%). EU South is quite close to EU North in terms of
structural change of technological upgrading while it still lags in infrastructure and production
capability.
Figure 22 Comparison of indexes for EU North and EU South.
Intensity of Production
technology capability
upgrading

economy upgrading EU South EU South
Structural R&D
change capability
ITU: Index
of Human,
technology physical
upgrading infstr.
With the new EU member states from Central and Eastern Europe the EU is quite heterogonous
entity in terms of its economies being at different income levels and levels of technology upgrading.
Central and East European countries are also divided on more advanced Central European and
Eastern Europe which include Bulgaria and Romania.
Differences between EU North and EU Central Europe in indexes across all four dimensions of
technology upgrading is quite balanced and is around 61% (Index of technology upgrading) to 64%
(Interaction with global economy) of EU North level (Figure 23). The gap in terms of intensity is
higher than in terms of breadth of technology upgrading. However, these differences seem to be
quite even and thus indicate that two technology systems are mutually balanced. However, this
hides very wide gap in terms of technology capability (19% of EU North) and firms organisational
capabilities (37% of EU North). Differences in terms of production capabilities and R&D capabilities
47
are even (65% and 62% respectively) and identical in terms of infrastructure and structural change
(79%). This balance may be explained by the effects of integration of central Europe into EU
production and technology networks. This is especially so given that in terms of technology and
knowledge exchange CE is proportionally behind as in index of technology upgrading.
Figure 23 Comparison of indexes for EU North and EU Central Europe.
Intensity of
technology
upgrading Production
capability

EU Central capabilities capability EU Central
Interaction Breadth of Europe Europe
economy upgrading EU North EU North
Structural R&D
change capability
Human,
ITU: Index physical
of infstr.
technology
upgrading
Figure 24 Comparison of indexes for EU South and EU Central Europe.
Intensity of
technology
upgrading Production
capability

EU South capabilities capability
Interaction Breadth of EU South
economy upgrading EU Central EU Central Europe
Structural R&D
Europe
change capability
Human,
ITU: Index physical
of infstr.
technology
upgrading
Central Europe and South EU can be considered as peripheries in relation to the EU North and hence
it is interesting to explore what are differences between them. EU CE is slightly behind the EU South
in terms of two indexes of technology upgrading (92% in terms of index of technology upgrading)
(Figure 24). However, differences are striking in terms of index of technology and knowledge
exchange where Central Europe is by 83% percentage points more integrated into EU and global
networks when compared to the EU South. Their profiles of technology upgrading in aggregate
seems similar but in terms of profiles are quite different. EU Central Europe lags significantly in
technology capability (51%) and firms’ organisational capabilities (77%). Lagging in terms of
production capability (88%) and in terms of structural changes (85%) is much smaller. However, in
terms of R&D capability Central Europe is ahead by 19 percentage points and 8 points in terms of
human capital and physical infrastructure.
Our previous analysis shows that Central Europe in terms of technology upgrading is grouped into
two subgroups: central Europe (Poland, Czech R, Estonia, Slovenia, Hungary) and Eastern Europe
(Romania and Bulgaria). Radial graphs below shows that lagging behind of Eastern Europe is quite
48
substantial (Figure 25). The biggest lag is in terms of intensity of technology upgrading (47% of CE)
and less in terms of breadth (74%) with the overall index being at 64% and index of interaction with
Global economy at 61%. The profile of this lagging shows that it is concentrated in low firm level
organisational capabilities (33%), R&D capability (only 34% of CE), technology capability (42%) and
production capability (65%). So, apart from technology and demand structure and infrastructure all
other dimensions represent heavy constraints on technology upgrading.
Figure 25 Comparison of indexes for EU Eastern Europe and EU Central Europe.
Intensity of
technology Production
upgrading capability
Firm level Technology

EU Central capabilities capability EU Central
Interaction Breadth of Europe Europe
with glocal technology
economy upgrading EU Eastern EU Eastern
Europe Structural R&D Europe
changes capability
ITU: Index Human &

of physical
technology infstr.
upgrading
Figure 26 Comparison of indexes for EU Eastern Europe and Turkey.
Turkey is close neighbour of EU Eastern Europe, it is potential EU member and represent important
trading partner. A comparison with Turkey shows that Eastern Europe is lagging behind in terms of
indexes of technology upgrading but has stronger interaction in terms of technology and knowledge
exchange (Figure 26).
A comparison of technology upgrading profiles shows very deep lag of Eastern Europe in firm
organisational capabilities and technology capability, smaller lag in R&D and production capability
but better infrastructure and technology and demand structure. So, if Turkey were to join the EU it
would not lead to further diversification of the EU in terms of technology upgrading profile but more
towards homogenisation.
In the final section we compare four regions of the EU which in terms of technology upgrading
operate as four distinct groups (Figure 27). In terms of index of technology upgrading they are
ordered from EU North, South, Central and East with difference between South and Central being
relatively small. Index of intensity of technology upgrading is similarly distributed as well as breadth
of technology upgrading but at higher levels. With respect to interaction with global economy both
49
EU Central and East are more engaged than EU South. This raises the question to which extent are
current economic issues of the EU South related to their weak technology and knowledge exchange.
Ordering – North, South, central and East – is also present in terms of production capability which is
quite unexpected given that Central Europe has become important production location within the
EU. This suggests that there is significant scope for improvements in production capability in the EU
Central. Gap of EU East is also particularly strong in this respect. In terms of technology capability EU
has very strong core – periphery features as the gap between North and three other regions is quite
substantial. This is also present in terms of R&D capability but in this case EU Central is ahead of EU
South with the East trailing much behind.
Differences within the EU in terms of human capital and physical infrastructure are much less
pronounced when compared to different capabilities or index of technology intensity. Central EU is
again ahead of the EU South with East trailing visibly behind. Ordering in terms of level and speed of
structural changes in technology and demand is again North, South, Central and East but with much
smaller differences. This suggests that in the last period there has been intensive structural change
in regions but with the catching up regions actually trailing behind in terms of technology and
demand structure. However, the biggest inter-regional differences are in terms of firms’
organisational capabilities where ordering is as expected but differences between North and other
regions are substantial. This structural feature is about big differences in power of enterprises to
generate and absorb technology and those differences are not directly related to the technology
activities.12
After the intra-EU comparison we compare BRICS with CEE countries: where CEEs stand in terms of
technology upgrading compared to BRICS? The Central EU is slightly ahead of China in terms of
overall index while the EU East is at the bottom of BRICS just slightly below Russia (Figure 28).
Central EU is ahead of BRICS and China in terms of breadth of technology upgrading largely on the
account of better human and physical infrastructure but behind China in terms of intensity of
technology upgrading. The East EU is again at the bottom in terms of breadth of upgrading but
ahead only of India in terms of intensity of technology upgrading. Both the EU Central and the EU
East are ranked high in terms of technology and knowledge exchange with the EU Central leading
the pack while the EU East is third after China. Brazil, India, and Russia have growth models which
so far have not been extensively reliant of technology and knowledge exchange despite the buzz
about emerging markets MNEs. Both BRICS and EU Central and East rank very low in terms of
technology capability with China being leader and the EU Central behind. It is quite surprising that
Central Europe ranks the highest in terms of production capability and even more in terms of R&D
capability while East EU is fourth in terms of production capability but ranks last in terms of R&D
capability. Infrastructure is factor of advantage for EU Central and East though EU Central is slightly
behind China with EU East being fifth. Both regions are in the medium positions in terms of
structural changes. There is a big difference between EU Central and East in terms of firms’
organisational capabilities. EU East ranks above Russia which reflects very poor capabilities of firms
to absorb technology. Central Europe is ranked third in this respect behind India, Brazil but ahead of
China.
12
Correlation coefficient between countries index of technology capabilities and firms’ organisational capabilities in the
overall sample is moderately high 0.68.
50
Figure 28 Comparison of indexes for BRICS and CEE.
60,0 80,0
50,0 70,0
60,0
40,0 50,0
30,0 40,0
30,0
20,0
20,0
10,0 10,0
0,0 0,0
Intensity of Breadth of ITU: Index of Interaction with Production Technology R&D and Human and Structural Firm level
technology technology technology global economy capability capability knowledge physical change capabilities
upgrading upgrading upgrading intensity infstr.
EU Central Europe China Brazil EU Central Europe China EU Eastern Europe
India Russian Federation EU Eastern Europe Brazil India Russian Federation
Finally, we compare the EU regions with the Korea and China (Figure 29). Korea is second ranked
economy in terms of technology upgrading while China is a newly emerged major global competitor.
As second ranked overall Korea is ahead of the EU North in terms of overall index as well as in
intensity of technology upgrading and very close to the EU North in terms of breadth of technology
upgrading. As Korea follows its specific growth model its technology and knowledge exchange
intensity is well below the EU North and the EU Central but ahead of other EU regions and China.
China is ranked in this grouping very slightly behind EU Central and South but well above the EU
East. This ranking is similar in terms of intensity of technology upgrading but China is ahead of the
Central EU in terms of structural features of technology upgrading. China is ranked similar to EU East
and South in terms of technology and knowledge exchange.
Figure 29 Comparison of indexes across EU regions, Korea and China.

80,0 80,0
70,0
60,0 60,0
50,0
40,0 40,0
30,0
20,0 20,0
10,0
0,0 0,0
Intensity of Breadth of ITU: Index of Interaction with Production Technology R&D and Human and Structural Firm level
technology technology technology global economy capability capability knowledge physical change capabilities
upgrading upgrading upgrading intensity infstr.
Korea, Rep. EU North EU South Korea, Rep. EU North EU Central Europe
EU Central Europe China EU Eastern Europe EU South China EU Eastern Europe
Korea is ahead of the EU North in terms of production capability, infrastructure and is quite well
ahead in terms of technology capability. In this latter respect, its relative technology intensity is just
at different level when compared to the EU North. This probably reflects Korea’s specific technology
structure which is around electronics related big MNEs which are heavily involved in patenting at
world level. The EU North is slightly ahead in terms of structural features of technology upgrading
and in terms of firms’ organisational capabilities. China is in all six dimensions of technology
upgrading index either at the lower end or in the middle of the EU regions distribution. It is only
above the EU East in terms of production capability, technology, R&D capability, and firms’
organisational capabilities. It is very close to the EU North in terms of structural features of
technology upgrading but close or above the EU Central and South in infrastructure. In nutshell,
given its size and global reach its ranking which is within the EU regional ranges represent the major
global change in terms of technology upgrading.
51
5. Conclusions
This analysis is about the relationship between levels of development and technology upgrading
from the perspective of broadly defined middle income economies, with special reference to Central
and East European economies.
We approach to technology upgrading as a multidimensional process. Index of technology

upgrading is conceptually based on two dimensions, intensity of technology upgrading and breadth
or structural features of technology upgrading, which jointly form the index. A third dimension is
index of technology and knowledge exchange which proxy for interaction with global economy and
which does not form the index but operates as moderating variable and helps to understand how
economies go about this third dimension of technology upgrading. This index interacts with the
index of technology upgrading but does not directly contribute to it. Reasons for this are related to
variable role that technology and knowledge exchange can play in relation to domestic technological
activities which are primary drivers of technology upgrading.
A measurement of technology upgrading is based on 35 indicators of both ‘hard’ and ‘soft’ nature
and they are applied to a balanced sample of 42 countries ranging from low middle income to upper
high income (see Annex). We apply composite index methodology and check for robustness of
groupings. Indicators have been selected based on their conceptual relevance, on their availability
and relationship to income levels. Hence, there is strong and positive relationship between index of
technology upgrading and income level distribution. However, our interest lies in variability between
income levels and index within individual income groups, across different sub-indexes and across
different regional groupings. Our aim is to explore changing drivers of growth across different
income levels and to undertake analysis of technology upgrading of the CEE countries in comparative
global context. This report is the first step in this analysis which is entirely descriptive based on
tables, scatter diagrams, radial diagrams and simple OLS econometrics. This means that our results
should be largely seen as tentative, hypothesis creating and exploratory rather than hypothesis
testing type.
1. The relationship between index of technology upgrading and income subgroups is positive only
in high income groups and there does not seem to be any relationship in middle income groups.
This may be expected as drivers of growth are more related to technology capability the higher is the
income level of countries. Middle income economies grow based on factors related to production
capability and availability of labour and low labour costs. There seems to be a kind of threshold from
middle income group to lower upper high income where the relationship between income level and
technology upgrading changes dramatically into positive.
In regressions with dummy for aggregate middle income group the regression fit is significantly
negative suggesting the existence of technology middle income trap. Dummy variables for middle
income subgroups are also negative but not significant which suggest that the middle income
phenomenon is not easily detectable and is diffused across several components of index of
technology upgrading. In addition, dummy for lower high income group is also negative but highly
insignificant suggesting that the phenomenon of middle income trap may be also much more
dispersed across income groups and should not necessarily correspond to middle income groups
proper as defined in this paper.
52
A broadly defined middle income trap is present in all dimensions of technology upgrading but its
importance varies across different dimensions. A trap seems to be higher for dimensions of breadth
of technology upgrading than for index of intensity of technology upgrading. Index of intensity of
technology upgrading reflects cumulative technology capability while index of breadth of technology
upgrading refers to structural, infrastructural and organisational features of economies. These latter
are subject to various market and system failures and are outcomes of variety of non-technological
factors the most important of which seem to be political economy of a specific economy which
underpins many of structural features of the economies.
Index of technology upgrading is composed of index A which measures intensity and types of
technology upgrading (production, technology and R&D capability) and index B which measures
structural features of the economies (infrastructure, structural change, firms’ organisational
capabilities)
2. Index A: intensity and types of technology upgrading is strongly and positively correlated for
upper high income economies but is less robust for lower upper high income and there is not
relationship for middle income groups. This would be expected as intensity of technology activities
rises with the income and is the strongest in the case of index of technology capability where
difference between upper high income and low middle income groups is 96% percentage points on
100% scale. This gap demonstrates that generation of technology is domain of only the highest
income bracket group.
Index of production capability is positively related to income levels but this relationship (as would
be expected) is much less robust due to strong production capabilities of middle income
economies whose production capability levels are at levels of upper high income economies.
These two indexes – technology and production capability - reflect global dispersion of production
capabilities and concentration of technology generation activities.
R&D capabilities are much stronger in middle income economies and lower upper high income
group than would be expected based on their income levels. This is largely due to dual face of R&D
which operates not only as factor of technology frontier activities but also as factor of absorptive
capacity. This explains why results for R&D capability are more similar to production capability than
to technology capability.
There are much bigger differences in ranking among countries in terms of intensity of technology
upgrading than in terms of index of breadth of technology upgrading or infrastructure, structural
change and firm level organisation capabilities. The ordering of countries within specific income
groups is much more regular i.e related to income per capita for intensity of technology upgrading
than for breadth of upgrading. This may be expected as technology and its cumulative features play
much more important role in income generation the closer is the country to technology frontier.
3. The breadth of technology upgrading is much less dispersed when compared to intensity of
technology upgrading with max-min difference for structural change being only 32 percentage
points and for infrastructure 48 percentage points. This suggest that structural change may be of
different types but it does take place at different income levels creating quite compressed
distribution with small differences between different income groups. A much stronger
convergence in terms of breadth of or scope of technology upgrading is due to smaller differences in
53
infrastructure but especially due to intensive process of structural change which takes place at
different levels at different income groups. So, much of differences in terms of breadth of
technology upgrading are largely due to remaining third component - differences in terms of firm
level organisational capabilities. It is with respect to firm level organisation capabilities that we have
on average the biggest differences within (not between) different income groups.
The relationship of breadth of technology upgrading to income is less positive than with respect to
intensity of technology upgrading. It seems that there is stronger structural change in middle income
group which is not reflected in levels of income. Also, regression with Index B: Breadth of technology
upgrading shows that dummy for upper middle income group is significant and negative which
suggest that middle income trap is more related to structural change issues in this subgroup. This
change is both on supply side as recorded by diversification of patent groups and levels and changes
in the availability of the latest technologies as well as on demand side as recorded by levels and
changes in the levels of buyers sophistication. Still, intensive structural change is just one of
components of technology upgrading which needs to be accompanied by increased intensity of
technology upgrading activities (production, technology and RD capabilities).
The levels of infrastructures of middle income economies are very often equal to those of low high
income group. So it seems that the middle income trap is high in the case of infrastructure i.e
infrastructure is not reflected in income levels. Educated young population, developed R&D and
physical infrastructures do not necessarily convert into proportionate levels of income.
One of major constraining structural feature of technology upgrading are firms’ organisational
capabilities which are located much higher in terms of income than for infrastructure, R&D or
production capabilities. The relationship for middle income groups is almost flat which suggest the
large diversity of organisational capabilities at that income level.
4. Technology and knowledge exchange component of technology upgrading does not add to index
in cumulative manner. In other words, increased technology and knowledge exchange with the
global economy does not necessarily increases potential for growth based on technology.
Technology and knowledge exchange is significantly more intensive in the case of upper high income
group but it is much more compressed in the case of lower upper income and higher middle income
groups. The relationship between Index C: interaction with global economy which is based on
proxies of knowledge and technology exchanges is not robust due to several ‘outliers’ (Ireland,
Belgium, Hungary, Greece) but also there is much weaker relationship when compared to index of
technology upgrading. There is not clear relationship between knowledge and technology exchange
intensity and middle income levels group. The interaction with the global economy in terms of
technology and knowledge exchange is very much country and not income specific. A strong
technology and knowledge inflow may operate as substitute or as complement to own technology
efforts recorded in other three indexes.
Dummy for middle income group is negative suggesting that the levels of technology and knowledge
exchange for this group are lower than would have been expected given the relationship. Dummies
for middle income subgroups show that only upper middle income group has significant negative
dummy which may indicate that the issue of middle income trap with respect to technology and
knowledge exchange is strongly present in this group rather than in the lower middle income group.
54
The highest index on middle income dummy is for index of technology and knowledge exchange.
This suggest that middle income economies are not benefiting from being engaged in global
technology and knowledge exchange as much as they should compared to other income groups. In
nutshell, this suggests that they should better use their existing levels of technology and knowledge
exchange ie. make them complementary to their own technology generation and absorption.
5. Index of technology upgrading represents proxy of potential for technology upgrading. In that
respect, a country that is low in terms of income per capita but has high index of technology
upgrading suggest that this country has good potential for growth based on technology upgrading.
The highest ranked countries in terms of index of technology upgrading are Sweden, Korea, Japan
and Germany. Korea has very high index given its current level of income and thus has very good
potential for further growth based on technology. Given its level of income, Italy has modest
potential for further technology upgrading. Similarly Russia has very low potential for technology
based growth given its income level. Poland has been fast growing CEE economy in the last 20 years
but its potential for technology based growth is moderate. Poland may grow based on its
production capability, and on-non technological factors but it seems that drivers for technology
based growth are not yet in place. China’s ranking in terms of index of technology upgrading is well
above its income per capita which suggests a room for further growth based on technology.
6. CEECs are spread widely along index of technology upgrading which reflects partly their income
per capita levels but also different potential for further growth. The leading two CEE economies in
this respect are Czech R and Estonia. Slovenia, whose income ($22.8) is well above both Czech R
($18) and Estonia ($16) is ranked below which suggest that its potential for further growth is not
really there to a similar extent. Hungary and Poland have similar income per capita ($12k) but
widely different current growth rates. Based on index of technology upgrading Poland does not have
really developed technology drivers of growth.
When compared to other countries in the sample the CEECs shows on average higher ranking in
terms of index of intensity of technology than index of breadth of technology upgrading. This is quite
important feature of the CEE as it shows that they are not structurally dynamic economies. Why this
is so can be depicted from rankings in specific dimensions of technology upgrading. CEECs are by
far the lowest ranked in terms of index of structural change which is about diversification of
technological knowledge and changes in demand and supply of technology. They are the most
compressed and well ranked group in terms of infrastructure but also group that has very weak firm
level organisational capabilities
An important structural feature of the CEE technology upgrading is their openness in terms of
technology and knowledge flows. For example, Hungary is the second ranked overall in terms of
Index of interaction with global economy and Estonia 7th. So, the key issue for their technology
upgrading is whether knowledge inflows in these economies operate as substitute or complements
to their own technology activities.
Among indexes of intensity of technology upgrading CEE ranks the best in terms of index of
production capability confirms the important role of central Europe as production location in
enlarged EU. Czech R is in overall 8th ranked which shows that it is emerging as at least in the EU
context as important production location. CEECs are well behind frontier economies in terms of
55
technology capability but are comparable to countries of similar income level. In terms of R&D
capability they are quite divided between central Europe which ranks relatively well and Eastern
Europe (Romania and Bulgaria) which are well behind with Poland being in intermediate position.
Among structural features of the CEECs technology upgrading they rank the best in terms of human
capital and physical infrastructure but much less well in terms of structural change. As a group they
are well placed in terms of schooling years but not so well in quality of math and science (except
Estonia and Slovenia). These suggest that ‘quantity’ in education is much less issue when compared
to ‘quality’. They are comparatively lagging behind in terms of quality and availability of RD
infrastructure and Internet and investments. Also, in terms of buyer sophistication CEECs are all
located in the second half of the sample. Also, in terms of index of firms’ organisational capabilities
the CEE region is placed quite low which in addition to weak structural change ranks them low on
breadth of technological upgrading index despite higher ranking in terms of infrastructure.
7. A comparative analysis of technology upgrading profiles shows that there is a quite significant
EU North – South technology upgrading gap. Index of technology upgrading of EU South is 66% of
EU North. The gap in terms of intensity of technology and knowledge exchange is much wider and
basically shows that EU South is very much outside of technology flows when compared to the EU
North. Differences between EU North and EU Central Europe in indexes across three dimensions of
technology upgrading is quite balanced and is around 61% to 64% of EU North level. However, this
hides very wide gap of EU CE in terms of technology capability and firms’ organisational capabilities.
Central Europe and South EU can be considered as peripheries in relation to the EU North. EU CE is
slightly behind the EU South in terms of two indexes of technology upgrading (92% in terms of index
of technology upgrading). However, differences are striking in terms of index of technology and
knowledge exchange where Central Europe is by 83% percentage points more integrated into EU
and global networks when compared to the EU South. Their profiles of technology upgrading in
aggregate seems similar but in terms of profiles are quite different. EU Central Europe lags
significantly in technology capability (51%) and firms’ organisational capabilities (77%). Lagging in
other dimension of breadth of technology upgrading is much smaller. However, in terms of R&D
capability Central Europe is ahead by 19 percentage points and 8 points in terms of human capital
and physical infrastructure. In the EU Central and EU South context the lagging behind of Eastern
Europe is quite substantial. It is of that extent that its reference is Central Europe, not EU North. The
biggest lag is in terms of intensity of technology upgrading (47% of CE) and less in terms of breadth
(74%) with the overall index being at 64% and index of interaction with Global economy at 61%.
In terms of index of technology upgrading index EU regions are ordered from EU North, South,
Central and East with difference between South and Central being relatively small. With respect to
interaction with global economy both EU Central and East are more engaged than EU South. This
raises the question to which extent are current economic issues of the EU South related to their
weak technology and knowledge exchange.
Ordering – North, South, Central and East – is also present in terms of production capability which
suggests that there is significant scope for improvements in production capability in the EU Central.
Gap of EU East is also particularly strong in this respect. In terms of technology capability EU has
very strong core – periphery features as the gap between North and three other regions is quite
56
substantial. This is also present in terms of R&D capability but in this case EU Central is ahead of EU
South with the East trailing much behind.
Differences within the EU in terms of human capital and physical infrastructure are much less
pronounced when compared to different capabilities of index of technology intensity. However, the
biggest inter-regional differences are in terms of firms’ organisational capabilities where ordering is
as expected but differences between North and other regions are substantial. This structural feature
is about big differences in power of enterprises to generate and absorb technology.
When compared to BRICS Central EU is slightly ahead of China in terms of overall index while EU
East is at the bottom of BRICS just slightly below Russia. Both EU Central and EU East are ranked
high in terms of technology and knowledge exchange with the EU Central leading the pack while EU
East is third after China.
A comparison of the four EU regions with Korea and China shows that Korea is ahead of EU North in
terms of overall index as well as in intensity of technology upgrading and very close to EU North in
terms of breadth of technology upgrading. As Korea follows its specific growth model its
technology and knowledge exchange intensity is well below EU North and EU Central but ahead of
other EU regions and China. China is ranked in this grouping very slightly behind EU Central and
South but well above the EU East.
8. In overall, our conceptual attempt to depict technological upgrading of middle income economies
with special reference to the CEECs has demonstrated its relevance and analytical value. We believe
that the conceptual approach has proven resilient and able to withstand the test of battery of
indicators that we have tested in the initial stage of research. As would be expected the choice and
selection of indicators will remain an issue for further work but the majority of indicators has shown
their relevance and link to our conceptual model. Analytical value of this composite indicator is
quite relevant as it focuses metrics on dimensions of technology upgrading that matter for growth
of broadly defined middle income economies.
6. References
Abernathy, W.J. and Utterback, J.M. (1978) Patterns of Industrial Innovation. Technology
Review.vol.8, iss.7, p.40-47.
Acemoglu Daron and James A Robinsn (2008) Why Nations Fail. The origins of power, prosperity,
and poverty, Crown Publsihers, New York.
Acemoglu, Daron, Fabrizio Zilibotti and Philippe Aghion (2006) Journal of the European Economic
Association, Vol. 4, No. 1 (Mar., 2006), pp. 37-74
Aghion, Philippe, Ufuk Akcigit and Peter Howitt (2013) What Do We Learn From Schumpeterian
Growth Theory? NBER Working Paper Series No 18824, February
Akamatsu K.(1962): A historical pattern of economic growth in developing countries. Journal of

Developing Economies, 1(1):3–25, March–August.
Amsden, Alice H. (2001) The Rise Of ‘‘The Rest’’. Challenges to the West from Late-Industrializing
Economies, Oxford University Press.
57
Amsden, A. H. and Hikino,T. (1994) , ‘Project execution capability, organizational know-how and
conglomerate corporate growth in late industrialization’, Industrial and Corporate Change 3, pp.
111-148.
Amsden Alice H. and F. Ted Tschang (2003) A new approach to assessing the technological
complexity of different categories of R&D (with examples from Singapore), Research Policy 32 (2003)
553–572
Anderson, J.C., Gerbing, D.W., 1988. Structural equation modeling in practice: a review of
recommended two-stage approach. Psychological Bulletin 103 (3), 411–423.
Archibugi Daniele and Alberto Coco (2005) Measuring technological capabilities at the country level:
A survey and a menu for choice, Research Policy 34 (2005) 175–194
Archibugi, D., Denni, M. and Filippetti, A. (2009) ‘The technological capabilities of nations: A review
of the synthetic indicators’. Technological Forecasting and Social Change, Vol. 76, No. 7, pp. 917-31.
Athreye, Suma (2005) “The Indian Software industry and its evolving service capability”; Industrial
and Corporate Change, Vol. 14(3): 393-418.
Baroncelli,E. Fink, C. and Smarzynska Javorcik, B. (2004) The Global Distribution of Trademarks:
Some Stylized Facts World Bank Policy Research Working Paper 3270, April 2004
Beck T. Demirguc_Kunt A, Levine R (2005) SMEs, growth and poverty: Cross-country evidence,
Journal of Economic Growth, 10(3):199-229.
Bell, M. and K. Pavitt (1993) “Technological Accumulation and Industrial Growth: Contrasts between
developed and developing countries”, Industrial and Corporate Change, 2(2): 157-210.
Bell, Martin and Keith Pavitt (1997) Technological accumulation and industrial growth: contrasts
between developed and developing countries, in Daniele Archibugi, Jonathan Michie, (eds),
Technology, Globalisation and Economic Performance,
Berend Iván T. and György Ránki (1982) The European Periphery and Industrialization, 1780–1914,
New York: Cambridge University Press
Bruno Randolph L. and Nauro F. Campos (2013) Reexamining the Conditional Effect of Foreign Direct
Investment, IZA Discussion Paper No. 7458, June 2013. Available at
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2287068
Bruno, R., E. Douarin, J. Koosteleva and S. Radosevic (forthcoming) Technology Choices and Growth:
Testing New Structural Economics in Transition Economies, Journal of Economic Policy Reform.
CEPI and DE (2014) Central Europe fit for the future. Visegrad group ten years after EU accession,
The Central European Policy Institute (CEPI), Bratislava, and demosEUROPA (dE) – Centre for
European Strategy, Report by the High Level Reflection Group, Published in January 2014 in
Bratislava and Warsaw
Chandler, Alfred D. Jr (1994) Scale and Scope. The Dynamics of Industrial Capitalism, Harvard
University Press.
Chenery Hollis and Moises Syrquin (1975), Patterns of development, 1950-1970, (World Bank)
Oxford University Press, Oxford.
58
Choung, Jae-Yong, Hye-Ran Hwang and Wichin Song (2014) Transitions of Innovation Activities in
Latecomer Countries: An Exploratory Case Study of South Korea, World Development Vol. 54, pp.
156–167, 2014
Cohen, W., and D. Levinthal, “Innovation and Learning: Two Faces of R&D,” Economic Journal 99
(1989), 569–596.
Crépon, B., E. Duguet and J. Mairesse (1998), “Research and Development, Innovation and
Productivity: An Econometric Analysis at the Firm Level”, Economics of Innovation and New
Technology, 7(2), 115-158.
Dahlman, C., Ross-Larsen, B. and Westphal, L.E. (1987), Managing technological development:
Lessons from the Newly Industrializing Countries, World Development, Vol.15, No. 6, 759-775.
Dosi, Giovanni, Alfonso Gambardella, Marco Grazzi, Luigi Orsenigo (2008) Technological Revolutions
and the Evolution of Industrial Structures: Assessing the Impact of New Technologies upon the Size
and Boundaries of Firms, Capitalism and Society, Volume 3, Issue 1 2008 Article 6:1-47
Dutrenit, Gabriela (2000) Learning and Knowledge Management in the Firm: from Knowledge
Accumulation to Strategic Capabilities, Edward Elgar: Aldershot, England.
Easterly, William (2001) The Elusive Quest for Growth. Economists’ Adventures and Misadventures in
the Tropics, MIT Press, Cambridge, Mass.
Ernst, D. (1997) “From Partial to Systemic Globalization: International Production Networks in the
Electronics Industry”, BRIE Working Paper 98.
Ernst, D. (1998) “Catching-Up, Crisis and Industrial Upgrading. Evolutionary Aspects of Technological
Learning in Korea`s Electronics Industry “, Asia Pacific Journal of Management, Volume 15 (2): 247-
283.
Ernst, D. (2001) “Global Production networks and Industrial Upgrading – A Knowledge Centred
Approach”, East-West Center Working Papers, Economics Series, No. 25.
Ernst, Dieter (2008) Asia’s “upgrading through innovation” strategies and global innovation
networks: an extension of Sanjaya Lall’s research agenda, Transnational Corporations, Volume 17,
Number 3, December 2008
Ernst, D. (2008) “Can Chinese IT Firms Develop Innovative Capabilities Within Global Knowledge
Networks?”, East-West Center Working Papers, Economics Series, 94.
Ernst Dieter (2013) Industrial Upgrading through Low-Cost and Fast Innovation— Taiwan’s
Experience, East-West Center working paper, Honolulu, Economics Series, No. 133, September 2013
Ernst Dieter (2014) Upgrading India's electronics manufacturing industry : regulatory reform and
industrial policy, East West Centre, Honolulu,
Ernst, D. and Kim, L. (2002) Global production networks, knowledge diffusion and local capability
formation. Research Policy, 31: 1417-29.
European Competitiveness Report 2011, Chapter 2 ‘Convergence of knowledge intensive sectors

and the EU’s external competitiveness’ (further ECR 2011)
Foray, Dominique (2015), Smart specialization. Opportunities and challenges for regional innovation
policy, Routledge.
59
Freeman, Chris (2002) Continental, national and sub national innovation systems – complementarity
and economic growth, Research Policy 31: 191-211.
Freeman Chris and Francisco Louçã (2001) As Time Goes By From the Industrial Revolutions to the
Information Revolution, Oxford University Press.
Freudenberg, M., 2003. Composite indicators of country performance: a critical assessment. STI
Working Paper 2003/16, OECD.
Furman Jeffrey L. and Richard Hayes (2004) Catching up or standing still? National innovative
productivity among ‘follower’ countries, 1978–1999 Research Policy 33 (2004) 1329–1354
Glaeser, Edward L., Rafael La Porta, Florencio Lopez-de-Silanes and Andrei Shleifer. "Do Institutions
Cause Growth?," Journal of Economic Growth, 2004, v9(3,Sep), 271-303.
Gort Michael and Steve Klepper (1982) Time Paths in the Diffusion of Product Innovations, The
Economic Journal, Vol. 92, No. 367 (Sep., 1982), pp. 630-653
Griibler, A. (1990) The Rise and Fall of Infrastructures. Dynamics of Evolution and Technological
Change in Transport, Physica-Verlag Heidelberg
Grupp, H., Mogee, M.E., 2004. Indicators for national science and technology policy: how robust are
composite indicators? Research Policy 33, 1373–1384.
Hanushek Eric A. and Ludger Wößmann (2007) Education Quality and Economic Growth, World
Bank, Washington.
Hanushek, E., Economic Growth in Developing Countries: The Role of Human Capital, Economics of
Education Review (2013), http://dx.doi.org/10.1016/j.econedurev.2013.04.005
Hausmann, R., J. Hwang, and D. Rodrik. 2007a. “What You Export Matters.” Journal of Economic
Growth 12:1-25.
Hausmann, R. and B. Klinger. 2006. “Structural Transformation and Patterns of Comparative

Advantage.” CID Working Paper No. 128. Cambridge, MA: Center for International Development,
Harvard University.
Hidalgo, C., and R. Hausmann. 2009. “The Building Blocks of Economic Complexity.” Proceedings of
the National Academy of Sciences 106(26): 10570–10575.
Hidalgo, C., B. Klinger, A. L. Barabasi, and R. Hausmann. 2007b. “The Product Space Conditions the
Development of Nations.” Science 317: 482-487.
Hobday, M., H. Rush., J. Bessant (2004) Approaching the innovation frontier in Korea: the transition
phase to leadership, Research Policy 33:1433-1457.
Hogselius, Per The (2005) Dynamics of Innovation in Eastern Europe: Lessons from Estonia, Edward
Elgar, Cheltenham
Hobday, M. (1995) “East Asian Latecomer Firms: Learning the Technology of Electronics, World
Development, 23(7): 1171-1193.
Humphrey, J. and Schmitz, H. (2004) ’Governance in global value chains’, in H. Schmitz (ed.) Local
Enterprises in the Global Economy. Issues of Governance and Upgrading, Edward Elgar: Cheltenham,
UK, 95-109.
60
Imbis Jean and Roman Wacziarg, Stages of diversification, American Economic Review 2003. Vol. 93,
No. 1, pp. 63-86.
IMF (2013) “German-Central European Supply Chain-Cluster Report”, IMF Country Report 13/263,
August 20th, 2013
IMF (2013b) Trade Linkages, Balance Sheets, and Spillovers: The Germany-Central European Supply
Chain, Prepared by Selim Elekdag and Dirk Muir, IMF Working Paper, European Department, October
2013
Jorgenson Dale W. and Marcel P. Timmer (2011) Structural Change in Advanced Nations: A New Set
of Stylised Facts, Scand. J. of Economics 113(1), 1–29, 2011
Khanna, T. and Yafeh, Y. (2007) Business Groups in Emerging Markets: Paragons or Parasites? Journal
of Economic Literature, June, Vol. XLV, No. 2, pp. 331-372
Kim, L. (1980) Stages of development of industrial technology in a less developed country: a model,
Research Policy 9(3):254-277
Kim, L. (1997) Imitation to Innovation: The Dynamics of Korea's Technological Learning, Boston,
Klepper, S. and E. Graddy (1990), The Evolution of New Industries and the Determinants of Marke
Structure,' Rand Journal of Economics, 21, 27—44
Klepper, Steven (1997) Industry Life Cycles, Industrial and Corporate Change, Vol., 6, No.1: 145-181.
Knell Mark (2008) Product-Embodied Technological Diffusion and Intersectoral Linkages in Europe,
28 April, Innovation Watch – SYSTEMATIC, NIFUSTEP
Koopmans ,T.C. (1947) Measurement without theory, Rev. Econ. Stat. 29 (3), 161–172.
Kravtsova,V. and Radosevic,S. (2011) Are systems of innovation in Eastern Europe efficient?
Economic Systems, doi:10.1016/j.ecosys.2011.04.005
Krueger, Jens J. (2008) Productivity and structural change: a review of the literature, Journal of
Economic Surveys (2008) Vol. 22, No. 2, pp. 330–363
Harvard Business School Press.
Kuznets, S. (1930) Secular Movements in Production and Prices: Their nature and their bearing upon
cyclical fluctuations. Houghton Mifflin, Boston and New York, repr. Kelley, New York, 1967.
Lall, S. (1992), Technological capabilities and industrialization, Research Policy, Vol. 20, No. 2, 165-
186
Langlois Richard N.(2003) Chandler in a Larger Frame: Markets, Transaction Costs, and
Organizational Form in History, Business and Economic History, Vol.1 (On line)
http://www.thebhc.org/publications/BEHonline/2003/Langlois.pdf
Lazonick William (2002) “The Theory of Innovative Enterprise”, In Malcolm Warner, ed.,
International Encyclopaedia of Business and Management, Thomson Learning, 3055-3076.
Lee Keun, B Y Km, Y Y Park, E Sanidas (2013) Big business and economic growth: identifying a binding
constraint for growth with country panel analysis, Journal of Comparative Economics, 41 (2):561-82
61
Lee, Keun (2013) Schumpeterian analysis of economic catch-up, Knowledge, Path-creation and the
middle-income trap, Cambridge University Press. .
Leitner, S.M. and Stehrer, R. (2014) Trade integration, production fragmentation and performance in
Europe - blessing or curse? A comparative analysis of the New Member States and the EU-15, The
Vienna Institute for International Economic Studies, GRINCOH project, WP5, Task 2, P2.5. Final
version: May 30th, 2014
Levine David I. and Michael W. Toffel (2009) Quality Management and Job Quality: How the ISO
9001 Standard for Quality Management Systems Affects Employees and Employers, Harvard
Business School Working Paper 09-18.
Lin Justin Yifu (2011) From Flying Geese to Leading Dragons New Opportunities and Strategies for
Structural Transformation in Developing Countries, Policy Research Working Paper 5702, World Bank
Lin Justin Yifu (2012a) From Flying Geese To Leading Dragons: New Opportunities and Strategies for
Structural Transformation in Developing Countries, Global Policy Volume 3 . Issue 4 . November
2012: 397-407
Lin, Justin Yifu (2012b) New Structural Economics. A Framework for Rethinking Development and
Policy. The World Bank, Washington D.C.
Lin Justin Yifu & David Rosenblatt (2012c): Shifting patterns of economic growth and rethinking
development, Journal of Economic Policy Reform, 15:3, 171-194
Lucas, R.E. (1988), “On the Mechanics of Economic Development”, Journal of Monetary Economics,
Vol. 22, pp. 3-42.
Majcen, Boris, Slavo Radosevic, Matija Rojec (2009) Nature and determinants of productivity growth
of foreign subsidiaries in Central and East European countries, Economic Systems, 33 (2009) 168–
184
Malerba Franco and Luigi Orsenigo (1994) The Dynamics and Evolution of Industries, Working Paper
WP-94- 120, November 1994, The International Institute for Applied Systems Analysis
Marshall, A. (1890) Principles of Economics. C.W. Guillebaud (ed), 9th edn, vol.1, Macmillan/Royal
Economic Society, London, 1961.
Mendonca, Sandro, Tiago Santos Pereira and Manuel Mira Godhino (2004) Trademarks as an
indicator of innovation and industrial change, Paper presented at the DRUID Summer Conference
Industrial Dynamics, Innovation and Development, Elsinore, Denmark, June 14-16, 2004
Millot Valentine (2009) Trademarks as an Indicator of Product and Marketing Innovations, STI
Working Paper 2009/6, OECD
Morck,R., Wolfenzon,D. and Yeung, B. (2005) Corporate Governance, Economic Entrenchment, and
Growth, Journal of Economic Literature, Vol. XLIII (September 2005), pp. 655–720
Mowery, D.C. and J. Oxley (1997) “Inward Technology Transfer and Competitiveness: The Role of
National Innovation Systems”, in Technology, Globalisation and Economic Performance, D. Archibugi
and J. Michie (eds.), Cambridge University Press.
Munda, G., Nardo, M., 2009. Non-compensatory/Nonlinear composite indicators for ranking
countries: a defensible setting. Applied Economics 41 (12), 1513–1523.
62
Nasierowski, W. and Arcelus, F.J. (2000) On Perceptions of Technical Efficiency of the Basis of the
Innovation Union Scoreboard Nelson Richard R. and Sidney G. Winter (1985), An Evolutionary Theory
of Economic Change, Belknap Press.
Nunnally, J., 1978. Psychometric Theory. McGraw-Hill, New York.
OECD (20003) The Sources of Economic Growth in OECD countries, OECD, Paris
OECD, 2008. Handbook on Constructing Composite Indicators: Methodology and User Guide. OECD,
Paris.
Ozawa, Terutomo (2009) The Rise of Asia, The ‘Flying Geese’ theory of tandem growth and regional
agglomeration, Edward Elgar, Cheltenham.
Peneder Michael (2003), Industrial structure and aggregate growth, Structural Change and Economic
Dynamics, 14 (2003): 427-448
Perez Carlota (2010) Technological revolutions and techno-economic paradigms, Cambridge Journal
of Economics 34: 185–202
Prud’homme, Rémy. 2005. « Infrastructure and Development», in: Bourguignon, François & Boris
Pleskovic, ed. 2005. Lessons of Experience (Proceedings of the 2004 Annual Bank conference on
Development Economics). Washington: The World Bank and Oxford University Press, pp. 153-181.
Radelet, Steven and Jeffrey Sachs (1997) Asia’s Re-emergence, Foreign Affairs, November/December
Radosevic, S. (1999) International technology transfer and catch-up in economic development,

Edward Elgar, Cheltenham.
Radosevic, S. and D. E. Yoruk, ‘The growth of enterprise through entrepreneurship and network
alignment’, In Radosevic, S. and B. Sadowski (Eds.)(2004), International industrial networks and
industrial restructuring in central Europe, Russia and Ukraine, Kluwer, 2004, pp.109-130
Radosevic Slavo and Anna Kaderabkova (eds) (2011) Challenges for European Innovation Policy:
Cohesion and Excellence from a Schumpeterian Perspective, Edward Elgar, Cheltenham
Radosevic, S. and Yoruk, D.E. 2004, “The Growth of Enterprise Through Entrepreneurship and
Network Alignment” in S. Radosevic and B.M. Sadowski (eds), International Industrial Networks and
Industrial Restructuring in Central and Eastern Europe, Kluwer, Dordrecht, the Netherlands.
Radosevic, S. and Yoruk, E. (2015) Why do we need theory and metrics of technology upgrading?
GRINCOH Deliverable.
Reinstaller, A. and Unterlass F., 2011. Sectoral Innovation Modes and the State of Economic
Development: Implications for Innovation Policy in the New Member States. In: Radosevic S.,
Kaderabkova, A. (eds.). Innovation Policy in Multi-Tier Europe. The Specific Challenges of the Central
And Eastern European Countries. Aldershot: Edward Elgar Publ.
Romer, P.M. (1990), “Endogenous Technological Change”, Journal of Political Economy, Vol. 98, No.
5, pp. 71-102
Romp, W. and J. de Haan (2005), ‘Public capital and economic growth: a critical survey’, EIB Papers,
10(1)
Rostow, W.W. (1960) The stages of economic growth: A non-communist manifest. Cambridge
University Press, Cambridge; 2nd enlarged edn, 1971.
63
Sandven, Tore, Keith Smith and Aris Kaloudis (2005), Structural change, growth and innovation: the
roles of medium and low tech industries, 1980-2000, In H. Kreinsen et al (eds), Low-tech Innovation
in the Knowledge Economy, Frankfurt: P. Lane. Jan 29 2005
Seong, Jieun, Jongseon Kim and Yejin Cho (2014) Korea’s Post Catch-up Policy Experiment: Focusing
on the Transition towards a Sustainable Socio-technical System, Paper prepared for thematic issue of
AJTM
Stevenson H. Thomas and Frank C. Barnes (2001) Fourteen Years of IS0 9000: Impact, Criticisms,
Costs, and Benefits, Business Horizons, May/June: 45-51
Suarez Fernando and James Utterback (1995) Dominant designs and the survival of firms, Strategic
Management Journal, Vol. 16, 415-430.
Teece, David J. 1986, Profiting from technological innovation: Implications for integration,
collaboration, licensing and public policy. Research Policy 15 (6): 285-305.
Terlakk Ann and Andrew A. King (2006) The effect of certification with the ISO9000 Quality
Management Standard: A signalling approach, Journal of Economic Behavior & Organization Vol. 60
(2006) 579–602
UNCTAD (2005) World Investment Report, UN, Geneva
Van Zeebroeck, N. (2011) The puzzle of patent value indicators, Economics of Innovation and New
Technology, 20:1, 33-62, DOI: 10.1080/10438590903038256Von Tunzelmann, G.N. (1995)
Technology and Industrial Progress: The Foundations of Economic Growth, Edward Elgar, Aldershot
Yoruk, D.E. (2013) Firm-level upgrading in low- and medium-technology industries in emerging
markets: the role of learning networks, PhD thesis, University of Sussex, December, SPRU.
64
APPENDIX A
Selected countries by the level of income per capita in 2013.
Lower Middle Income Upper Middle Income Lower High Income Upper High Income
(GNI pc atlas method (GNI pc atlas method (GNI pc atlas method (GNI pc atlas method
$1046-4125) $4126-12745) $12176-30000) $30001- )
Ghana Albania Chile Austria
India Belarus Czech Republic Belgium
Indonesia Brazil Estonia Germany
Moldova Bulgaria Greece Ireland
Morocco China Korea Italy
Philippines Hungary Poland Japan
Ukraine Jordan Portugal Norway
Vietnam Kazakhstan Russia Sweden
Malaysia Slovenia UK
Mexico Spain USA
Peru
Romania
South Africa
Thailand
Turkey
APPENDIX B Explanations for Composite Index Measures

1. Index of technology upgrading (ITU)
According to our conceptual approach, ITU is the outcome of the interaction of two components:
type and intensity of technology upgrading (INDEX A) and breadth of technology upgrading
(structural features) (INDEX B) each of which are formed by selected indicators.
ITU = INDEX A + INDEX B
where
ITU = Index of technological upgrading
INDEX A = Index of types and intensity of technology upgrading
INDEX B = Index of technology upgrading breadth (structural features)
1.1 Index of types and intensity of technology upgrading (INDEX A)

Type and intensity of technology upgrading lie in the production and technology capabilities and
skills of enterprises and the population, investments and outputs in new knowledge creation and
generation, and the extent of R&D activities. They are essential to technology upgrading as without
them product and process innovations could not be developed.
65
Index A is composed of three components: Production capability (Index 1), technology capability
(Index 2) and R&D and knowledge intensity (Index 3). Accordingly:
Index A = Index 1 + Index 2 + Index 3
Production capability (Index 1) intends to capture the rate of activities and output in relation to
production activity. It is composed of three indicators:
1. ISO9001 certificates (per million inhabitants) is taken from International Organization for
Standardization (ISO) and average of 2007-2011 values is used in the analysis.
2. Trademark applications, resident (per million inhabitants) is taken from World Intellectual
Property Organization (WIPO) and average of 2007-2011 values is used in the analysis.
3. On the job training is taken from World Economic Forum Global Competitiveness Report
(WEF GCR) Question 5.C and 2012-13 value is used in the analysis.
Technology capability (Index 2) is built on measuring technology generation capabilities, mainly

patents. It is composed of four indicators drawn from World Intellectual Property Organization
(WIPO) and averages of 2007-2011 values are used in the analysis:
1. Patents resident applications to national office (per million inhabitants)

2. Patent applications to USPTO (per million inhabitants)
3. Patent applications to EPO (per million inhabitants)
4. Resident's industrial design count (per million inhabitants)
R&D and knowledge and intensity (Index 3) intends to capture the knowledge developed by
investments in R&D as well as the influence of capabilities embodied in people, i.e. R&D personnel,
scientists and their publication outputs. It draws on eight indicators:
1. Business Enterprise Sector R&D expenditures (% of GDP) is taken from UNESCO for the year
2011.
2. Research and development expenditure (% of GDP) is taken from UNESCO for the year 2011.
3. Researchers in R&D (per million inhabitants) is taken from World Bank for year 2010.
4. Technicians in R&D (per million inhabitants) is taken from World Bank for year 2010.
5. Scientific and technical journal articles (per million inhabitants) is taken from World Bank
and average of 2007-2011 values is used in the analysis.
6. Science citations (per million inhabitants) is taken from Thomson National Science Indicators
and average of 2007-2011 values is used in the analysis.
7. Quality of scientific research institutions is taken from WEFGCR Question 12.02 for year
2012-13. It is based on the question: How would you assess the quality of scientific research
institutions in your country? [1 = very poor; 7 = the best in their field internationally]
8. University - industry collaboration in R&D is taken from WEFGCR Question 12.04 for year
2012-13. It is based on the question: To what extent do business and universities collaborate
on research and development (R&D) in your country? [1 = do not collaborate at all; 7 =
collaborate extensively]
1.2 Index of technology upgrading in structural features (INDEX B)

Breadth of technology upgrading lies in structural features and changes in these structural features.
Structural features are based on the human capital, physical capital and organizational issues.
66
Index B is composed of three components: Human capital, physical and organizational infrastructure
(Index 4), structural change (Index 5) and firm level capabilities (Index 6). Accordingly:
Index B = Index 4 + Index 5 + Index 6
Human capital, physical and organizational infrastructure (Index 4) is built on measuring the
influence of capabilities embodied in people through the wider population with education,
the respond to skills demand, the extent people exploit available infrastructural
technologies and the level of fixed investment. Accordingly, it is composed of six manifest
indicators:
1. Average years of schooling for ages 25+ is taken from Barro-Lee database for the latest
available year 2010.
2. Quality of maths and science education institutions is taken from WEFGCR Question 5.04 for
year 2012-13. It is based on the question: How would you assess the quality of math and
science education in your country’s schools? [1 = poor; 7 = excellent – among the best in the
world]
3. Availability of specialized research and training services is taken from WEFGCR Question 5.07
for year 2012-13. It is based on the question: In your country, to what extent are high-
quality, specialized training services available? [1 = not available; 7 = widely available]
4. Availability of scientists and engineers is taken from WEFGCR Question 12.06 for year 2012-
13. It is based on the question: To what extent are scientists and engineers available in your
country? [1 = not at all; 7 = widely available]
5. Fixed broadband Internet subscribers (per 100 people) is taken from World Bank for year
2012.
6. Gross Fixed Investment as % of GDP is taken from World Bank for year 2012.
Structural change (index 5) intends to capture over time changes in technology capability, demand
structure and level of available technologies. It comprises seven indicators. The first three indicators
uses patent data from WIPO, USPTO and EPO to calculate Herfindahl-Hirschman Index. By this, we
aim to assess the level of diversification by technology field in patenting structure of the countries.
The formula for Herfindahl-Hirschman index calculation is as below:
𝑛
𝐻 = ∑ 𝑠𝑖2
𝑖=1
Where si is the share of patents of a country in a specific technology field. Index is calculated for
each of the 42 countries based on WIPO technology classification (see Appendix F for a list of 35
technology fields). The same method is applied to calculate Herfindahl-Hirschman index for national
patent applications (Indicator 22), applications to EPO (Indicator 23) and applications to USPTO
(Indicator 24).
1. Herfindahl-Hirschman Index for total national patent applications.

2. Herfindahl-Hirschman Index for patent applications to EPO
3. Herfindahl-Hirschman Index for patent applications to USPTO
67
4. Buyer sophistication is taken from WEFGCR Question 6.16 for year 2012-13. It is based on
the question: In your country, how do buyers make purchasing decisions? [1 = based solely
on the lowest price; 7 = based on a sophisticated analysis of performance attributes]
5. Change in buyer sophistication( % change in Q. 6.16 from 2006-07 to 2012-13)
6. Availability of state-of-the-art technologies is taken from WEFGCR Question 9.01 for year
2012-13. It is based on the question: To what extent are the latest technologies available in
your country? [1 = not available; 7 = widely available]
7. Change in availability of latest technologies( % change in 9.01 from 2006-07 to 2012-13)
Firm level capabilities (Index 6) has two manifest indicators:
1. .Number of firms in Forbes 2000 (per million inhabitants)

2. Firm level technology absorption is taken from WEFGCR Question 9.02 for year 2012-13. It is
based on the question: To what extent do businesses in your country absorb new
technology? [1 = not at all; 7 = aggressively absorb]
2. Index of technology and knowledge exchange (ITKE)
This index intends to capture the influence of global interactions of countries y which knowledge
flows take place. We assess the impact of such interactions as complementary to technology
upgrading. The index comprises five manifest indicators:
1. Licensing receipts as % of GDP is taken from World Bank for year 2012.
2. Licensing payments as % of GDP is taken from World Bank for year 2012.
3. Share of exports in complex industries in total exports (SITCRev3 5 71-79 87 88). Data for this
indicator have been extracted from UNComtrade for years 2008-12 average. We calculated
the share of exports in total exports for each country, particularly in SITC Rev.3 sectors 5 -
Chemicals and related products, n.e.s.; 71 to 75 Machinery (Power generating machines,
special industrial machinery, metalworking machinery, general industrial machinery, n.e.s,
office machines); 76 - telecommunications equipment; 78-79 transport equipment (road
vehicles, other transport equipment); 87-88 electrical and optical (scientific equipment,
n..e.s., photo apparatus n.e.s., clocks).
4. Foreign direct investment, net outflows (% of GDP) is taken from World Bank for year 2007-
2012 average.
5. Foreign direct investment, net inflows (% of GDP) is taken from World Bank for year 2007-
2012 average.
68
APPENDIX C
Factor analysis results.
Ind C Compon Quantitative Indicators One Cumu Chi2 Cronb One Cumu Chi(2)
ex a ent factor lative (sig.) ach’s factor lative (sig.)
t soluti expla Alpha solution expla
e on for natio for natio
g comp n of category n of
o onent Facto (CFA) Facto
r (CFA) r r
y
1. 1.ISO9001 certificates pmi 0.28 - 7.94 0.458 0.92 0.69 729.90
Producti 2.Trademark applications, resident pmi 0.55 (0.0473) 0.76 (0.0000)
A. INTENSITY AND TYPES OF TECHNOLOGY UPGRADING (SCALE)
on 3.On the job training Q.5.C 0.47 0.86

capabilit
y
2. 4.Patents resident applications to national office 0.83 0.87 97.2 0.837 0.75
Technol pmi (0.0000)
ogy 5.Patent applications to USPTO pmi 0.89 0.90
capabilit 6.Patent applications to EPO pmi 0.71 0.85
y 7.Resident's industrial design count pmi 0.65 0.83
3. R&D 8.Business Enterprise Sector R&D expenditures (% 0.87 0.86 383.24 0.946 0.78
and of GDP (0.0000)
knowled 9.Research and development expenditure (% of 0.74 0.40
ge GDP)
intensity 10.Researchers in R&D pmi 0.86 0.60
11.Technicians in R&D pmi 0.79 0.72
12.Scientific and technical journal articles pmi 0.94 0.60
13.Science citations pmi 0.90 0.78
14.Quality of scientific research institutions 0.85 0.85
Q.12.02
Index of Technology Upgrading
15.University - industry collaboration Q.12.04 0.76 0.57
4. 16.Average years of schooling 25+ 0.66 0.78 95.53 0.765 0.51 0.61 293.27
ITU = A + B
Infrastru 17.Quality of maths and science education Q.5.04 0.51 (0.0000) 0.37 (0.0000)
B: BREADTH OF TECHNOLOGY UPGRADING: STRUCTURAL FEATURES (SCOPE)
cture: 18.Availability of research and training services 0.68 0.85

human Q.5.07
capital 19.Availability of scientists and engineers Q.12.06 0.57 0.66
and 20.Fixed broadband Internet subscribers (per 100 0.89 0.82
physical people)
and 21.Gross Fixed Investment as % of GDP -0.38 -0.37
organisa
tional
5. 22.Herfindahl-Hirschman Index for total national 0.72 0.85 93.55 0.782 -0.59
Structur patent applications (0.0000)
al 23.Herfindahl-Hirschman Index for patent 0.27 -0.41
change applications to EPO 0.81 0.61
24.Herfindahl-Hirschman Index for patent
applications to USPTO -0.53 -0.61
25.Buyer sophistication Q.6.16 0.63 0.80
26.Change in buyer sophistication( % change in Q.
6.16 from 2006-07 to 2012-13) -0.65 -0.49
27.Availability of state-of-the-art technologies
Q.9.01 0.55 -0.53
28.Change in availability of latest technologies( %
change in 9.01 from 2006-07 to 2012-13)
0.75
6. Firm 29.Number of firms in Forbes 2000 pmi - - - - 0.79
level 30.Firm level technology absorption Q.9.02
capabilit
ies
7. 31.Licensing receipts as % of GDP 0.86 0.83 115.43
GLOBAL
INTERACTION
Technol 32.Licensing payments as % of GDP 0.76 (0.0000)

ogy and 33.Share of exports in complex industries in total 0.52
knowled exports (SITCRev3 5 71-79 87 88)
ECONOMY
ge 34.Foreign direct investment, net outflows (% of 0.84

exchang GDP)
WITH
e 35.Foreign direct investment, net inflows (% of 0.63

C:
GDP)
Note: Confirmatory Factor Analysis (CFA) is based on principal factors method in STATA. The communalities are estimated
using the squared multiple correlation coefficients.
69
APPENDIX D
Table D.1. Index of technological upgrading, its sub-indexes and Index of interaction with global
economy for 42 selected countries.
4. B: BREADTH OF
A. INTENSITY Infrastructure: TECHNOLOGY
AND TYPES OF human capital UPGRADING: INDEX OF
TECHNOLOGY 1. 2. 3. R&D and TECHNOLOGY and physical 5. Structural STRUCTURAL INTERACTION
UPGRADING Production Technology knowledge UPGRADING and change 6. Firm level FEATURES WITH GLOBAL
INDEX capability capability intensity (SCALE) organisational indicators capabilities (SCOPE) ECONOMY
Income
Country Name Group ITU INDEX 1 INDEX 2 INDEX 3 INDEX A INDEX 4 INDEX 5 INDEX 6 INDEX B INDEX C
Sweden UHI 70.3 9.4 7.5 14.9 31.7 11.9 12.2 14.5 38.6 36.0
Japan UHI 65.4 8.1 10.1 10.6 28.8 11.9 12.4 12.3 36.6 24.4
Germany UHI 60.8 8.9 9.1 11.4 29.4 11.4 11.6 8.4 31.4 23.5
Belgium UHI 58.9 10.9 3.6 11.5 26.0 12.5 12.3 8.2 32.9 57.1
United States UHI 58.5 7.0 7.0 10.5 24.4 11.8 11.4 11.0 34.1 24.7
Austria UHI 57.8 8.7 5.2 11.6 25.5 10.4 11.8 10.1 32.3 27.5
United Kingdom UHI 54.8 7.8 2.9 11.3 22.0 10.7 12.2 9.9 32.8 29.1
Ireland UHI 52.9 6.0 2.7 9.3 18.1 9.3 11.3 14.3 34.9 100.0
Italy UHI 39.9 8.1 3.8 5.7 17.6 8.1 11.0 3.3 22.3 17.1
Korea, Rep. LHI 67.1 9.8 12.2 11.3 33.3 12.3 11.2 10.3 33.8 22.3
Spain LHI 43.0 8.0 2.1 7.2 17.2 9.0 10.9 5.9 25.8 21.7
Portugal LHI 42.5 7.4 0.9 7.4 15.7 7.3 12.1 7.4 26.8 17.3
Czech Republic LHI 40.8 8.1 1.0 7.9 16.9 9.0 10.1 4.8 23.9 19.6
Estonia LHI 40.1 6.3 0.8 8.1 15.1 9.9 9.2 5.9 25.0 25.0
Slovenia LHI 39.2 5.5 1.7 10.5 17.7 8.8 9.4 3.2 21.4 20.4
Chile LHI 35.5 7.7 0.1 3.1 10.9 7.5 11.1 6.0 24.6 18.5
Greece LHI 30.8 2.7 0.8 4.1 7.6 8.2 10.5 4.5 23.3 7.1
Poland LHI 28.8 4.3 0.7 3.8 8.8 7.7 9.9 2.3 20.0 17.1
Russian Federation LHI 24.1 2.4 0.4 4.1 6.9 7.3 9.4 0.5 17.2 11.1
Malaysia UMI 39.4 6.6 0.2 4.7 11.5 9.3 11.0 7.6 27.9 20.4
Belarus UMI 38.6 6.6 0.3 4.1 11.1 10.4 8.8 8.3 27.5 12.6
China UMI 34.9 4.4 1.4 5.1 10.9 9.0 11.3 3.7 24.0 19.0
Hungary UMI 31.6 3.9 0.7 6.0 10.6 8.3 8.5 4.2 21.0 62.6
Turkey UMI 30.2 4.7 1.7 2.8 9.1 5.4 10.1 5.6 21.1 15.8
Brazil UMI 30.0 4.3 0.1 4.6 9.0 4.3 11.7 5.1 21.1 9.8
South Africa UMI 29.1 4.0 0.1 3.6 7.7 3.4 11.6 6.4 21.4 9.5
Jordan UMI 28.6 2.9 0.0 3.1 6.1 8.0 8.2 6.4 22.6 25.0
Mexico UMI 27.6 3.8 0.1 2.6 6.5 5.7 11.2 4.1 21.1 17.7
Thailand UMI 26.6 3.2 0.2 2.3 5.7 6.1 10.1 4.8 20.9 17.9
Bulgaria UMI 23.7 4.1 0.5 2.8 7.3 6.3 8.9 1.1 16.3 21.2
Romania UMI 22.4 3.3 0.3 2.1 5.7 6.8 8.3 1.6 16.7 13.9
Kazakhstan UMI 21.8 2.1 0.2 1.9 4.2 6.0 8.6 2.9 17.6 15.0
Peru UMI 21.2 2.1 0.0 3.7 5.8 3.9 8.2 3.3 15.4 6.9
Albania UMI 17.8 1.5 0.0 1.7 3.2 5.3 6.1 3.2 14.6 13.1
Ukraine LMI 27.1 1.8 0.3 2.9 5.1 7.3 11.2 3.5 22.1 12.5
India LMI 25.3 2.5 0.0 2.8 5.4 6.7 8.1 5.1 19.9 9.9
Morocco LMI 24.8 2.4 0.4 1.7 4.5 6.0 10.8 3.5 20.3 9.8
Indonesia LMI 24.2 2.9 0.0 2.5 5.5 6.1 8.4 4.2 18.7 6.4
Philippines LMI 23.3 3.0 0.0 2.0 5.0 4.7 8.6 5.0 18.3 10.3
Moldova LMI 15.2 1.0 0.6 0.9 2.5 5.1 6.5 1.0 12.7 10.7
Vietnam LMI 15.0 0.6 0.1 3.6 4.2 4.2 5.4 1.1 10.7 18.2
Ghana LMI 13.4 1.1 0.6 1.3 3.0 4.7 4.0 1.6 10.4 9.1
70
Table D.2. Index of technological upgrading, its sub-indexes and Index of interaction with global
economy ranks for 42 selected countries.
4.
A. INTENSITY Infrastructure: B: BREADTH OF
AND TYPES OF human capital TECHNOLOGY
TECHNOLOGY 1. 2. 3. R&D and TECHNOLOGY and physical 5. Structural UPGRADING: C: INTERACTION
UPGRADING Production Technology knowledge UPGRADING and change 6. Firm level STRUCTURAL WITH GLOBAL
INDEX capability capability intensity (SCALE) organisational indicators capabilities FEATURES (SCOPE) ECONOMY
Income
Country Name Group ITU ITU rank INDEX 1 INDEX 2 INDEX 3 INDEX A INDEX 4 INDEX 5 INDEX 6 INDEX B INDEX C
Sweden UHI 70.3 1 3 4 1 2 4 3 1 1 4
Korea, Rep. LHI 67.1 2 2 1 6 1 2 15 5 5 12
Japan UHI 65.4 3 7 2 7 4 3 1 3 2 10
Germany UHI 60.8 4 4 3 4 3 6 9 8 9 11
Belgium UHI 58.9 5 1 8 3 5 1 2 10 6 3
United States UHI 58.5 6 13 5 9 7 5 10 4 4 9
Austria UHI 57.8 7 5 6 2 6 8 6 6 8 6
United Kingdom UHI 54.8 8 10 9 5 8 7 4 7 7 5
Ireland UHI 52.9 9 17 10 10 9 11 12 2 3 1
Spain LHI 43.0 10 9 11 14 12 14 19 16 13 13
Portugal LHI 42.5 11 12 16 13 14 24 5 12 12 23
Czech Republic LHI 40.8 12 8 15 12 13 13 22 23 17 17
Estonia LHI 40.1 13 16 18 11 15 10 28 17 14 7
Italy UHI 39.9 14 6 7 16 11 19 18 32 20 24
Malaysia UMI 39.4 15 15 29 18 16 12 17 11 10 16
Slovenia LHI 39.2 16 18 12 8 10 16 27 33 22 15
Belarus UMI 38.6 17 14 26 22 17 9 30 9 11 30
Chile LHI 35.5 18 11 34 28 19 22 16 15 15 19
China UMI 34.9 19 20 14 17 18 15 11 28 16 18
Hungary UMI 31.6 20 25 19 15 20 17 33 25 27 2
Greece LHI 30.8 21 32 17 21 25 18 21 24 18 40
Turkey UMI 30.2 22 19 13 31 21 34 23 18 25 26
Brazil UMI 30.0 23 22 33 19 22 39 7 20 26 36
South Africa UMI 29.1 24 24 32 26 24 42 8 14 23 38
Poland LHI 28.8 25 21 20 23 23 21 25 36 30 25
Jordan UMI 28.6 26 31 38 27 29 20 37 13 19 8
Mexico UMI 27.6 27 26 35 33 28 33 13 27 24 22
Ukraine LMI 27.1 28 38 27 29 35 23 14 29 21 31
Thailand UMI 26.6 29 28 30 35 32 30 24 22 28 21
India LMI 25.3 30 33 39 30 34 27 38 19 31 35
Morocco LMI 24.8 31 35 24 40 37 31 20 30 29 37
Indonesia LMI 24.2 32 30 37 34 33 29 34 26 32 42
Russian Federation LHI 24.1 33 34 25 20 27 25 26 42 35 32
Bulgaria UMI 23.7 34 23 23 32 26 28 29 39 37 14
Philippines LMI 23.3 35 29 40 37 36 37 32 21 33 34
Romania UMI 22.4 36 27 28 36 31 26 35 38 36 28
Kazakhstan UMI 21.8 37 36 31 38 39 32 31 35 34 27
Peru UMI 21.2 38 37 42 24 30 41 36 31 38 41
Albania UMI 17.8 39 39 41 39 40 35 40 34 39 29
Moldova LMI 15.2 40 41 22 42 42 36 39 41 40 33
Vietnam LMI 15.0 41 42 36 25 38 40 41 40 41 20
Ghana LMI 13.4 42 40 21 41 41 38 42 37 42 39
71
APPENDIX E
Table E.1. OLS regressions for indexes without dummies.
Model11a Model12a Model13a Model14a Model15a Model16a Model17a Model18a Model19a Model 20a
ITU 956.7
(0.000)
Index 1 4666.9
(0.000)
Index 2 4200
(0.000)
Index 3 3939
(0.000)
Index A 1676.5
(0.000)
Index 4 5058.6
(0.000)
Index 5 5234.5
(0.000)
Index 6 3538.6
(0.000)
Index B 1951.5
(0.000)
Index C: ITKE 478.8
(0.001)
Constant -16603 -5646.4 9586.8 -4231.1 -3247.9 -21896.8 -34143.8 -2262.5 -27892.7 7531.4
(0.000) (0.084) (0.000) (0.044) (0.100) (0.000) (0.002) (0.478) (0.000) (0.041)
No of 42 42 42 42 42 42 42 42 42 42
observations
F-test sig. 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
R2 0.81 0.64 0.61 0.81 0.81 0.62 0.41 0.58 0.71 0.25
Adjusted R2 0.81 0.63 0.60 0.80 0.81 0.61 0.39 0.57 0.70 0.23
Table E.2. OLS regressions for indexes without dummies and without outlier countries Ireland,
Hungary and Belgium.
Model11b Model12b Model13b Model14b Model15b Model16b Model17b Model18b Model19b Model20b
ITU 931.9
(0.000)
Index 1 4607.2
(0.000)
Index 2 4063
(0.000)
Index 3 3851
(0.000)
Index A 1628.9
(0.000)
Index 4 4905.6
(0.000)
Index 5 4881.1
(0.000)
Index 6 3609.6
(0.000)
Index B 1910.1
(0.000)
Index C: 1657.2
ITKE (0.000)
Constant -15964.9 -5774.7 8872 -3849.5 -3045.1 -20977 -31468.6 -2864.9 -27119.5 11810.3
(0.000) (0.088) (0.000) (0.072) (0.125) (0.000) (0.004) (0.388) (0.000) (0.021)
No of 39 39 39 39 39 39 39 39 39 39
observations
F-test sig. 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
R2 0.79 0.62 0.63 0.79 0.81 0.59 0.38 0.56 0.68 0.51
Adjusted R2 0.79 0.61 0.62 0.79 0.80 0.58 0.36 0.55 0.67 0.50
72
APPENDIX F
WIPO patent technology classification.
1 - Electrical machinery, apparatus, energy
2 - Audio-visual technology
3 - Telecommunications
4 - Digital communication
5 - Basic communication processes
6 - Computer technology
7 - IT methods for management
8 - Semiconductors
9 - Optics
10 - Measurement
11 - Analysis of biological materials
12 - Control
13 - Medical technology
14 - Organic fine chemistry
15 - Biotechnology
16 - Pharmaceuticals
17 - Macromolecular chemistry, polymers
18 - Food chemistry
19 - Basic materials chemistry
20 - Materials, metallurgy
21 - Surface technology, coating
22 - Micro-structural and nano-technology
23 - Chemical engineering
24 - Environmental technology
25 - Handling
26 - Machine tools
27 - Engines, pumps, turbines
28 - Textile and paper machines
29 - Other special machines
30 - Thermal processes and apparatus
31 - Mechanical elements
32 - Transport
33 - Furniture, games
34 - Other consumer goods
35 - Civil engineering
73

Grincoh WP 3.04 Radosevic Yoruk

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Grincoh WP 3.04 Radosevic Yoruk

Uploaded by

Copyright:

Available Formats

Working Paper Series

Paper No. 3.04

A New Metrics of Technology Upgrading: The

Slavo Radošević*, Esin Yoruk*

* University College London

Please cite as:

A New Metrics of Technology Upgrading: The Central And East European

Technology upgrading is a multidimensional conceptual framework which goes beyond R&D in

Figure 1 presents technology upgrading as an outcome of interaction between intensification of

Figure 1: Dimensions of technology upgrading

2.1. Intensity and types of technology upgrading (scale)

2.1.1. Production capability

2.1.2. Technology capability

2.1.3. Research & development

2.2. Breadth of technology upgrading: structural change, infrastructure and firms’

2.2.1. Infrastructure: human capital, physical and organisational

Another structural precondition for technology upgrading is physical infrastructure. Infrastructure

2.2.3. Firm level organisational capabilities

This sub-component of technology upgrading is proxied by two indicators. First is a number of

2.3. Interaction with global economy and technology upgrading

2.3.1. Technology and knowledge exchange

In summary, technology upgrading is three-dimensional process proxied by a variety of indicators,

 Intensity of technology upgrading through various types of innovation and technology

Intensity of • Production capability

• Technology (embodied) export

Breadth of • Infrastructure (human, physical, organizational)

3. Why and how composite indicator of technology upgrading?

capability 2.Trademark applications, resident pmi WB avg 1/6

10.Researchers in R&D pmi WB 2010

14.Quality of scientific research institutions Q.12.02 WEFGCR avg

4. 16.Average years of schooling 25+ Barro-Lee 2010

Infrastructure 17.Quality of maths and science education Q.5.04 WEFGCR 2012-13

27.Availability of state-of-the-art technologies Q.9.01

6. Firm 29.Number of firms in Forbes 2000 pmi Forbes 2013

Table 2. Indicators of technology and knowledge exchange index (ITKE)

3.2 Developing measures by composite index methodology

Equation (2) shows explicitly the normalisation method (Min-Max) used:

4. Analysis of levels and profiles of technology upgrading

LHI UHI Total LHI UHI Total

SWE SWE SWE SWE

GNI per capita in 2012

LMI UMI LMI UMI

LHI UHI Total

LHI UHI Total

CZE CZE KOR

EST EST PRT PRT

LMI UMI LMI UMI

RUS POL RUSMEX

Index : production capability

LHI UHI Total

LMI UMI Production capability is the key precondition

for further technology upgrading and good

0 5 10 15 0 5 10 15 precondition for further growth of middle

Index of production capability is positively

GNI per capita in 2012

LHI UHI Total

An important dimension of structural change in technology upgrading is the firms’ organisational

Dummy MI -9327 -22796

Table 4. Comparison of MI dummies across OLS regression models.

Index C -22795.7 3076.517 -7.41 0.000 -29018.54 -16572.86

Index 5 -20455.11 3400.798 -6.01 0.000 -27333.87 -13576.34

Index 6 -17819.95 2624.515 -6.79 0.000 -23128.53 -12511.36

Index 2 -16621.22 2980.479 -5.58 0.000 -22649.81 -10592.63

Slavo Radošević, Esin Yoruk