Dugar 2021

Financial Analysts Journal
ISSN: 0015-198X (Print) 1938-3312 (Online) Journal homepage: https://www.tandfonline.com/loi/ufaj20
Equity Investing in the Age of Intangibles
Amitabh Dugar & Jacob Pozharny
To cite this article: Amitabh Dugar & Jacob Pozharny (2021) Equity Investing in the Age of
Intangibles, Financial Analysts Journal, 77:2, 21-42, DOI: 10.1080/0015198X.2021.1874726
To link to this article: https://doi.org/10.1080/0015198X.2021.1874726
© 2021 The Author(s). Published with

license by Taylor & Francis Group, LLC
Published online: 18 Mar 2021.
Submit your article to this journal
Article views: 3998
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ufaj20
Financial Analysts Journal | A Publication of CFA Institute Research
https://doi.org/10.1080/0015198X.2021.1874726
Equity Investing in the

Age of Intangibles
Amitabh Dugar and Jacob Pozharny
Amitabh Dugar is a research analyst at Bridgeway Capital Management, Houston. Jacob Pozharny is head of international equity
at Bridgeway Capital Management, Houston.
E
Expenditures on the creation of xpenditures for the creation of intangible capital have increased
intangible capital have increased, since the 1990s, but accounting standards have not adapted to
but accounting standards have this increase. Among the sample of non-US companies we studied
not kept pace. We investigated for this article, the proportion of capitalized intangible assets (exclud-
whether this has affected the ing goodwill) relative to total assets increased from 0.2% to 2.2%
value relevance of book value between 1994 and 2018, and for US companies, these assets rose
and earnings. We constructed a from 2.75% to 6.12%. Uncapitalized intangible capital expenditures,
composite measure of intangible such as those on research and development (R&D), also increased
intensity by which to classify
over the same period. For our sample of international companies, R&D
industries. The measure is based
expenses relative to total revenues rose from 1.52% to 2.20%, and for
on intangible assets capitalized on
US companies, they grew from 9.47% to 14.36% from 1994 to 2018.
the balance sheet; research and
development expenditures; and Recently, several authors investigated the impact of the increase in
sales, general, and administrative
importance of intangible assets relative to that of physical assets on
expenditures. We show that the
US value investing strategies (Arnott, Harvey, Kalesnik, and Linnainmaa
value relevance of book value and
2021; Amenc, Goltz, and Luyten 2020; Lev and Srivastava 2019).
earnings has declined for high-
intangible-intensity companies in They adjusted book-to-price ratios (B/Ps) to account for the biases
the United States and abroad, but caused by unrecorded intangible capital, but they concurred that the
for the low-intangible-intensity decline in the effectiveness of value-based investment strategies in the
group, it has remained stable in United States cannot be attributed to the structural economic changes
the United States while increasing generated by intangibles or the failure of accounting standards to
internationally. adapt to those changes. Li (2020) extended their work to certain other
countries—the United Kingdom, the countries of continental Europe,
Japan, and Asia ex Japan.
This is an Open Access article distributed

Our work does not predict whether or when the performance of US
under the terms of the Creative Commons value strategies will recover, advocate the use of one or more pre-
Attribution-NonCommercial-NoDerivatives
License (http://creativecommons.org/licenses/
ferred value measures, or recommend adjustment of valuation ratios
by-nc-nd/4.0/), which permits non-commercial to compensate for omitted intangibles in the quest for higher stock
re-use, distribution, and reproduction in any
medium, provided the original work is properly
returns. Instead, we offer a way to adapt traditional equity investment
cited, and is not altered, transformed, or built analysis to handle the effect of variations in intangible asset intensity
upon in any way.
among companies and industries and extend the US research on this
topic to the world’s 14 largest international economies—8 of the devel-
oped markets studied by Li (2020) and 6 emerging markets. We show
that the relationship between financial variables and contemporaneous
Disclosure: The authors report no
conflicts of interest.
The authors gratefully acknowledge the suggestions of Executive Editor Stephen J. Brown
and Co-Editor Steven Thorley, CFA. They also thank their colleagues at Bridgeway Capital
Management, especially Andrew Berkin (head of research) for his guidance and encour-
PL Credits: 2.0 agement and other members of the investment management team for helpful comments.
Volume 77 Number 2 © 2021 The Author(s). Published with license by Taylor & Francis Group, LLC. 21
Financial Analysts Journal | A Publication of CFA Institute
stock prices has weakened so much for high- We propose a composite measure of intangible
intangible-intensity companies in both the United intensity that captures the interindustry variation in
States and abroad that investors can no longer afford the impact on the financial statement of three types
to ignore the changes in the economic environment of intangible capital: intangible assets reported on
created by intangibles. the balance sheet (excluding goodwill), innovation
(R&D) capital, and organization (sales, general, and
Global accounting standards require companies to administrative) capital.
expense, rather than capitalize, the amounts spent
on activities that create intangible capital. This Using this composite to classify industries into high-
requirement results in a systematic and persistent and low-intangible-intensity groups, we analyzed
understatement of the book value of equity. Some the contemporaneous relationship between (1)
other value metrics that can be used to identify value stock prices and (2) book values and earnings within
stocks, such as earnings to price and cash flow to each intensity group for both US and international
price, are also affected by accounting distortions. companies. We found that within the high-intangible-
The reason is that the costs incurred for creating intensity group, the combined value relevance of
intangible capital are expensed immediately whereas book value and earnings declined between 1994 and
the corresponding revenues/cash inflows that the 2018 for both US and international companies. In
assets generate typically occur over one or more contrast, the value relevance of these variables for
future periods. The result is a mismatch between the low-intangible-intensity group remained stable
expenses and revenues on income statements.1 For in the United States but increased internationally
example, Lev and Sougiannis (1996) showed that over the same period. We show that the divergence
in the chemical and pharmaceutical industries, an in the value relevance of book value and earnings
initial outlay on R&D that is immediately expensed between the high- and low-intangible-intensity
can beneficially affect revenues and earnings for groups is greater for international companies and
up to nine years. Moreover, capitalized intangible has increased more in international markets than in
assets affect reported earnings over long periods as the US market. Our results are especially important
they are gradually amortized (expensed), although for international equity investors seeking to build
amortization practices vary by intangible type and by investment strategies that account for the impact
industry and country. of intangible intensity on valuation ratios and other
financial metrics used to assess the profitability, qual-
Arnott et al. (2021) and Amenc et al. (2020) showed ity, growth, and risk characteristics of companies.
that adjustments to book value to account for the
effect of intangibles do improve the return prediction
ability of B/P for US companies, and Li (2020) con- Motivation for the Study
firmed this finding for companies in several interna-
Our research was motivated by (1) the potential
tional markets. Nevertheless, Arnott et al. (pp. 61, 65)
effect of intangibles on equity valuation ratios that
conceded,
link book values and earnings to the intrinsic value
This improved measure of value has also of companies via the Ohlson (1995) residual income
recently suffered a large drawdown and after valuation model and (2) the possible effect of intan-
2007 has still not been as good as S/P [sales gibles on other investment metrics of importance to
to price] or E/P [earnings to price]. Perhaps equity investors. Intangibles and their inadequate
intangibles-adjusted B/P is still missing some- accounting can affect corporate value and risk fac-
thing important. . . . It will be an interesting tors, such as asset growth or debt-to-equity ratios.
topic for future research to gauge which met- Via their impact on reported earnings, intangibles
rics perform best in producing a better HML and their inadequate accounting can also affect
[high book-to-market value minus low book- profitability characteristics, such as return on equity.
to-market value] value factor or in predicting The adjustments to B/P suggested by Arnott et al.
future corporate profits and whether optimal (2021), Amenc et al. (2020), and Li (2020) provide no
settings for these metrics vary by industry,
information about this issue. Our broad examination
sector, or country.
of whether the rise of intangibles has influenced
We complement the work in these areas by explor- the relationship between fundamental financial
ing other measures of value and the variation in variables, such as book value and earnings, and
those measures among industries and countries. contemporaneous stock prices can provide insights
into these aspects.
22 Second Quarter 2021

Prior research evidence in US markets on how earnings reports, and annual reports to sharehold-
investments in intangible capital affect the value ers was less significant in sectors where investment
relevance of book value and earnings is inconclu- in intangible assets was relatively high, indicating
sive. Lev and Gu (2016, p. 89) showed that correla- that financial variables had lower value relevance
tions between contemporaneous stock prices and for intangible-intensive companies than for other
both book value and earnings have dropped from companies.
1950 to 2013 because of increasing investments
in intangible capital. Barth, Li, and McClure (2018) Thus, the conclusions of past research on the value
argued, however, that although the value relevance relevance of book value and earnings in US markets
of certain items, such as net income, has declined, are mixed and do not extend beyond 2012. We
the aggregate value relevance of the “accounting used evidence from more recent time periods to
amounts” they examined remained unchanged from reexamine this issue and assess whether changes in
1962 through 2014.2 Collins, Maydew, and Weiss the value relevance of book value and earnings as a
(1997) concluded that the combined value relevance result of the increase in intangible intensity in the US
of earnings and book value of US intangible-intensive market also extend to international markets.
companies did not decline over the 1953–93 period,
Past (univariate) approaches focusing on specific
but Ciftci, Darrough, and Mashruwala (2014) inferred
types of uncapitalized expenditures that create
the opposite from their analysis of a similar set of
intangible capital did not permit measurement of the
companies between 1975 and 2007. Core, Guay,
aggregate effects of corporate spending on various
and Buskirk (2003) examined the same issue for the
types of intangible capital and evaluation of interin-
1975–99 period for a broad sample of US companies
dustry differences arising from the influence of such
and for subsamples they considered to be emblem-
spending on valuation ratios or other investment
atic of the “New Economy.” They found that the
metrics.3 Some researchers (e.g., Israel, Laursen, and
explanatory power of their model deteriorated in the
Richardson 2020) have recommended intra-industry
New Economy subperiod (1995–1999) for all types
valuation of companies to account for differences
of companies. These conflicting findings from prior
in the types and amounts of intangible capital on
research indicate that the magnitude and direction
equity valuations among industries. Our composite
of the influence of intangibles on the value relevance
intangible-intensity measure offers investors an
of book value and earnings are unclear. Possibly, in
alternative way to capture the financial statement
the US market, the correlation between stock prices
effects of variations in intangible capital in various
and these two financial variables may have fluctu-
industries and compare their different effects. We
ated over time.
demonstrate that our intangible-intensity measure
International evidence about the impact of intan- can do so consistently over time and across both the
gibles on the value relevance of book value and US and international investment universes.
earnings has been extremely scarce and has gener-
Research on this topic also faces indeterminate data
ally been limited to the UK and Australian markets.
hurdles. Arnott et al. (2021) used company-level
Silva (2012) showed that in the UK market, book
estimates of intangible capital provided to them by
value is the better predictor of share price for low-
Peters and Taylor (2017) to adjust the book values
intangible-intensity industries whereas earnings
of US companies. Amenc et al. (2020) made similar
have greater efficacy in high-intangible-intensity
adjustments by using long histories of financial
industries. Goodwin and Ahmed (2006) found for the
data and certain specific data items (e.g., industry-
Australian market that during the period prior to the
specific R&D depreciation rates from the US Bureau
2005 implementation of the Australian equivalent to
of Economic Analysis and the year of a company’s
International Financial Reporting Standards, when
founding). Because of limited international research
both expensing and capitalization of intangibles
on this subject, whether enough breadth and history
were permitted, the value relevance of earnings for
of fundamental and macroeconomic data are avail-
Australian companies decreased but the decline
able to permit making such adjustments in all or most
was less pronounced for companies that recognized
international markets is unclear.4 Using financial
intangible assets (“capitalizers”). A significantly higher
statement and market data from both developed and
proportion of the capitalizers belonged to intangible-
emerging international markets, we demonstrate that
intensive industries. Fraser, Tarbert, and Tee (2009)
one can build the intangible-intensity metrics that
demonstrated that in the UK market, the share price
we propose, and these metrics are robust enough for
reaction to disclosures of interim reports, preliminary
23
Volume 77 Number 2
use in financial statement analysis and valuation of variables in the US and international markets. Hence,
both US and international stocks. we noted that cross-country differences in account-
ing standards that govern the choice of expensing
versus capitalizing expenditures to create intangible
Intangible-Intensity Metrics assets (and any change in those standards over time)
Our composite measure of intangible intensity can lead to problems in comparability. For example,
has three components: (1) total intangible assets some researchers have found changes in the value
reported on the balance sheet, excluding goodwill, relevance of capitalized intangible assets before
(2) R&D expenses, and (3) sales, general, and adminis- and after the introduction of International Financial
trative (SG&A) expenses. We discuss the reasons for Reporting Standards (IFRS), and some have found
the choice of these components and their measure- issues in subsamples of companies that made differ-
ment in the following subsections. ent financial reporting choices in regimes where both
expensing and capitalization were permitted.6
We used our composite measure to determine
the intangible intensity of companies in all indus- Third, evidence from around the globe demonstrates
tries except the Banks, Insurance, and Diversified that several types of intangible expenditures that
Financials industry groups [of the Global Industry have been capitalized and reported on the balance
Classification Standard (GICS) system] in the United sheet were value relevant, both in aggregate and
States and abroad. We excluded these three industry individually: Oliveira, Rodrigues, and Craig (2010)
groups because the metrics we used to gauge intan- examined Portuguese companies, Ritter and Wells
gible intensity are affected by their atypical financial (2006) and Dahmash, Durand, and Watson (2009)
reporting practices.5 For example, because of the studied companies in Australia, and Aboody and
nature of their business, banks bundle and report Lev (1998) evaluated the equity market effect of
several types of operating expenses in the SG&A capitalized software development costs in the
category, and globally, almost no banks or insurance United States.
companies disclose R&D expenses.
For these three reasons, we included capitalized
identifiable intangible assets in our intangible-
Identifiable Intangible Assets. We refer to a intensity composite.
company’s total capitalized intangible assets except
goodwill as identifiable intangible assets. In theory, any We excluded goodwill from our measurement of
intangible assets reported on the balance sheet are the capitalized intangible assets for two reasons.
already included in book value. Even so, we include First, our primary objective was to examine the
them as a separate component in our composite stock market effects of various forms of intangible
intangible-intensity measure for three reasons. capital investments that have gained in importance
because of the rapid transformation in corporate
First, accounting criteria for capitalization of
investment and business models since the 1990s.
expenditures that can create intangible capital are
Since goodwill is simply an accounting byproduct of
inconsistent. For example, in the United States, the
business combinations, it is unclear if it meets this
cost of internally developed patents is required to be
criterion. Second, prior evidence regarding the value
written off (i.e., expensed on the income statement),
relevance of goodwill is mixed. The findings vary not
but if the ownership of those unrecorded patents is
only for the US and international equity markets but
subsequently transferred as a result of a corporate
also for different time periods in the same market.
acquisition or merger, they must be capitalized on
The reason is that differences in the rules for writ-
the balance sheet of the acquirer at their fair value.
ing off goodwill by market and over time have led to
Therefore, ignoring capitalized intangible assets
subjective assessments of the fair value of goodwill.
could understate the aggregate level of intangible
Managerial discretion in applying goodwill valuation
intensity for companies that have grown through
rules has exacerbated the problems that affect accu-
acquisitions rather than organically. At the collec-
rate measurement of goodwill (Dahmash et al. 2009).
tive level, industries that have gone through periods
of consolidation would appear to be less intangible
R&D Expenses. US accounting standards require
intensive than other industries.
the cost of both research and development to be
Second, we intended to compare the effects of intan- expensed, but IFRS are a bit less restrictive, allowing
gible intensity on the value relevance of financial the capitalization of development costs if certain
criteria are met. In-process R&D (consisting of R&D

assets acquired in business combinations or asset expenses, Angelopoulos, Giamouridis, and Vlismas
acquisition transactions) can also be capitalized. (2012) showed that intangible organization capital is
Thus, we can capture the effect of in-process R&D helpful in predicting stock returns for US companies.
on financial statements in our first intangible capital Comparable results for international companies,
metric—identifiable intangible assets—which was dis- however, are sparse. Tronconi and Marzetti (2011)
cussed previously. Although US accounting rules that reported a positive link between an SG&A-based
require R&D costs to be expensed have remained measure of organization capital and certain financial
consistent since 1974, international accounting guid- performance metrics for European companies.
ance on this subject exhibits considerable variation
and has continued to evolve.7 Some countries have Because the validity of SG&A expenses as a
been edging into greater conformity with the United proxy for intangible organization capital has been
States at an uneven pace as they move to IFRS. confirmed by multiple studies, we used the same
approach. Note that Arnott et al. (2021) and Amenc
Research evidence from the United States and et al. (2020), however, considered only 30% of the
abroad indicates that R&D expenditures create total SG&A expenses reported on income statements
intangible innovation capital that is reflected in to be capitalizable intangible assets. Our research
equity market values. This research includes that of was unaffected by this design choice because we
Ahmed and Falk (2006), who examined Australian used SG&A expenses to rank and classify companies
companies; a study of companies in France, Germany, according to their organization capital rather than
the United Kingdom, and the United States by Zhao attempt to assess its value relevance or adjust book
(2002); and findings by Smith, Percy, and Richardson values by the amount of unrecorded organization
(2001) from studying the Australian and Canadian capital.
markets. Lev and Sougiannis (1996) estimated the
R&D capital of a sample of more than 800 US manu- Other Types of Intangible Capital. Two
facturing companies, of which about half belonged other financial statement items—advertising
to five high-intangible-intensity industries (chemicals expenses and labor costs—have also been posited to
and pharmaceuticals, machinery and computer create intangible capital. Advertising expenses are
hardware, electrical and electronics, transportation considered to be a gauge of intangible brand capital,
vehicles, and scientific instruments). The authors and labor costs are regarded as an indicator of intan-
demonstrated that adjusting earnings and book gible human capital. We did not include these items
values of companies for the capitalized value of R&D in our composite measure of intangible capital for the
makes those variables more value relevant. following reasons.
SG&A Expenses. SG&A expenses have been Both US and international evidence corroborating
used by several authors to proxy for another type the value relevance of advertising expenditures
of intangible company capital, namely, organization for UK companies (e.g., Shah and Akbar 2008;
capital. Lev and Radhakrishnan (2005) characterize Shah, Stark, and Akbar 2009) is weak.8 Moreover,
organization capital as the set of “unique systems and Govindarajan, Rajgopal, Srivastava, and Wang (2019)
processes employed in the investment, production, showed that in the United States, advertising has
and sales activities of the enterprise, along with the stayed constant since the 1980s at very low levels
incentives and compensation systems governing its compared with expenditures on other forms of
human resources” (p. 73). They used annual sales intangible capital, such as R&D.9 Finally, advertis-
and general and administrative expenses reported ing expenses are a subcomponent of sales and
in income statements to estimate changes in com- marketing expenses, which are included in the
panies’ organization capital and showed that such aggregate SG&A expenses figure usually reported
changes explain differences between the market in income statements. Because we used aggregate
values and book values of US companies. SG&A expenses to represent organizational capital
in our intangible-intensity composite, either sales
Using SG&A expenses to estimate the stock of and marketing expenses or, ideally, advertising
organization capital for a sample of US companies, expenses should be excluded from the aggregate
Eisfeldt and Papanikolaou (2013) concluded that SG&A expenses figure to avoid double counting of
companies with a high ratio of organization capital advertising expenses in the composite. Most compa-
to book assets exhibit higher annual average market nies, however, do not disclose either of these items
returns. Using an industry-relative measure of SG&A separately. Because convincing evidence about the
25
Volume 77 Number 2
value relevance of intangible brand capital is lacking small and loss-making companies in our sample
and in light of the practical difficulties related to its because prior research (Darrough and Ye 2007;
measurement, we did not include it in our composite Collins et al. 1997; Joos and Plesko 2005) indicated
measure of intangible intensity. that such companies are often persistently unprofit-
able entities that tend to invest more heavily in R&D
Prior evidence supporting the value relevance of activities that create intangible capital than do larger
human capital includes Angelopoulos et al. (2012), and profitable companies.
who reported that long–short portfolios based on
an industry-relative human capital measure provide For each company, we computed three metrics of
statistically significant risk-adjusted returns for only intangible intensity:11 (1) total intangible assets,
the first year after portfolio formation, and Pantzalis excluding goodwill, relative to total assets, (2) R&D
and Park (2009), who found that arbitrage portfolios expenses relative to total revenues, and (3) SG&A
based on a market valuation measure of intangible expenses relative to total revenues.12 Figure A1 and
human capital provide excess returns only for small Figure A2 in Appendix A show changes in the data
companies. Edmans (2011) considered employee availability for these metrics over time in, respec-
satisfaction to be a type of intangible asset and tively, the US universe and international universe.
demonstrated that for a limited number of companies Because the first of our three intangibility metrics is
(the “100 Best Companies to Work for in America”), derived from balance sheet items and the other two
it is positively correlated with shareholder returns. from income statement items, the presentations in
He acknowledged that the companies in his sample Figures A1 and A2 are structured accordingly.
are unusually large and exhibit notably better earn-
ings performance than other companies. In fact, not Except for some narrowly focused studies in
all the companies in his sample are publicly traded, Australia and the United Kingdom, past research on
which further limits data availability for comparable this topic has focused on US companies, so our work
studies. A common theme underlying all studies in adds to this literature by reporting on the relative
this area is that, unlike R&D and SG&A expenses that availability of data to construct the aforementioned
link intuitively to, respectively, innovation capital and three metrics of intangible intensity in both the
organization capital, investments in human capital United States and 14 other countries.13 Note that on
are notoriously difficult to measure, which prompts average, the information required to compute capital-
researchers to use indirect, output-based estimates. ized intangible assets was available for 52% (67%)
Because investments in intangible human capital of the companies in the US (international) universe
assets are difficult to quantify and data to estimate that reported total assets and 33% (32%) of all US
such investments are hard to obtain in all 15 of the (international) companies that disclosed informa-
markets we studied, we did not include such invest- tion about goodwill. SG&A expenses were available
ments in our intangible-intensity composite, but for 97% (95%) of US (international) companies; data
these assets remain a topic for future investigation. on R&D expenses were available for 31% (30%) of
US (international) companies that reported total
revenues.14
Data and Methodology Most of the previous work on this subject has
Our sample consisted of companies based in coun- focused on specific types of intangible capital
tries ranked by 2018 GDP among the top 15 in the (primarily, innovation capital created by R&D activi-
world according to the World Bank (2019).10 We ties) one at a time. A drawback of such univariate
obtained the requisite financial and market data approaches is that innovation capital dominates in
for these companies from the Standard & Poor’s certain industries, such as pharmaceuticals, because
Xpressfeed database. We used data for fiscal years it is widespread and of large magnitude, but in other
between 1994 and 2018 because that database industries, different types of intangible capital may
is sparsely populated before 1994, especially for be more significant and value relevant than R&D.
international companies. Investors who prefer to hold broadly diversified
portfolios rather than a narrow selection of compa-
For each year, we included companies that reported
nies from specific industries can gain comprehensive
the required financial data (described later) for an
insights about the effects of intangible capital invest-
annual financial reporting period that ended during
ment by using our intangible-intensity composite to
that year. For our value-relevance tests, we used
make investment decisions for all types of compa-
stock price data up to the end of 2019. We retained
nies. By aggregating the impact of the main types

of intangible capital that the literature has linked intensity in the United States and internationally in
to stock prices and returns, the composite enables a similar fashion—an important consideration for
investors to classify and compare companies belong- investors who wish to use the composite to con-
ing to different industries on common ground.15 We struct global investment strategies or compare factor
expected the availability of financial data for compu- performance among investment universes.
tation of each of our three intangible capital metrics
to vary according to the nature of a company’s In addition to assessing cross-universe consistency,
business, which might, in turn, depend on its industry we also evaluated the consistency of the annual
membership.16 intangible-intensity ranks across industries over time
within each investment universe for each of the three
For every year during our sample period, we intangible-intensity metrics and for the intangible-
computed the median intangible intensity for intensity composite. We, again, relied on Kendall’s
all companies within each of 21 four-digit GICS concordance statistic to compare the relative annual
industry groups (which excludes Banks, Insurance, intangible-intensity ranks for the 21 industries over
and Diversified Financials) for each of our three the 25-year sample period (with slight exceptions
intangible-intensity metrics. Next, we ranked the for the intensity of R&D expenses because sufficient
21 industries annually according to median intangible data were lacking for certain industries in early
intensity, measured independently for each of the years). The results, provided in Table 3, show that
three metrics. Finally, we combined every industry’s for each of three types of intangible capital and for
annual rank on the three intangible-intensity metrics the intangible-intensity composite, relative industry
to obtain its equally weighted composite intangible- ranks have remained stable over time at a statistical
intensity rank for that year.17 Thus, we calculated a level of confidence exceeding 99%. Because the pace
set of 21 annual composite intangible industry ranks of evolution of intangible intensity among industries
for each of the 25 years in our sample period. and various types of intangible capital varies, this
finding is important. The time-series persistence
Applying this procedure independently to the US of our composite measure of intangible intensity
and international universes, we obtained two sets provides assurance that investment strategies based
of ranks. In Table 1, we present the 25-year aver- on the choice or weighting of factors that drive
age composite intensity rank of each industry for investment returns according to intangible intensity
the two investment universes. We used these are likely to be stable and replicable.
average composite ranks to classify the 10 lowest-
ranked industries into the “low-intangible-intensity
category” and the remaining 11 industries into the Combined Value Relevance
“high-intangible-intensity category.” The order of
industries that resulted is remarkably similar for the of Book Value and Earnings
two universes. Indeed, except for two differences We used our composite intangible-intensity mea-
(Energy and Retailing), the set of high- and low- sure to study the effect of investments in intangible
intangible-intensity industries in the United States capital on the value relevance of book value and
and abroad is identical. earnings, which are often used to construct valuation
ratios, as well as other financial metrics that inves-
To evaluate the consistency of the US and inter- tors use to evaluate the profitability, quality, growth,
national composite intangible-intensity rankings, and risk characteristics of companies. Prior studies
we calculated and present in Table 2 the Kendall’s of this issue have defined intangible intensity in an
coefficient of concordance—a W-statistic with a χ2 ad hoc manner, typically with a focus on intangible
distribution (Zar 1999)—for each of the 25 years in innovation capital created by R&D and ignoring the
our sample.18 Note that for all years, the W-statistics identifiable intangible assets reported on the balance
are highly significant. When we divided our 25-year sheet. The reason is that most previous research-
study period into two subsamples, 1994–2006 and ers adopted the following definition of intangible
2007–2018, we found similar results for the two intensity, which was initially proposed by Collins
subperiods that were also consistent with the full- et al. (1997, p. 51, footnote 16):
sample results. For investors, the implication of the
findings reported in Table 2 is that the composite Note that intangible intensity does not refer
measure of intangible intensity that we propose is to the presence of large amounts of recorded
built on pervasive intangibility metrics. The composite intangibles because the concerns raised in
can be used to classify industries by their intangible the literature relate more to unrecorded
27
Volume 77 Number 2
Table 1. Average Composite Intangible-Intensity Ranks, 1994–2018
United States International
Average Average
Composite Intangible Composite
Industry Rank Intensity Industry Rank
Utilities 3.84 Low Real Estate 4.20

Real Estate 4.13 Low Transportation 4.29
Energy 4.49 Low Automobiles & Components 6.13
Transportation 4.78 Low Capital Goods 6.91
Materials 5.89 Low Utilities 7.04
Food & Staples Retailing 6.10 Low Materials 7.07
Automobiles & Components 7.44 Low Consumer Durables & Apparel 7.39
Capital Goods 8.99 Low Food, Beverage & Tobacco 7.63
Food, Beverage & Tobacco 9.92 Low Food & Staples Retailing 8.84
Consumer Durables & Apparel 10.12 Low Retailing 9.40
Retailing 10.81 High Semiconductors & Semiconductor 10.41
Equipment
Commercial & Professional Services 11.45 High Energy 11.00
Consumer Services 11.51 High Technology Hardware & Equipment 11.04
Semiconductors & Semiconductor 12.32 High Commercial & Professional Services 11.85
Equipment
Technology Hardware & Equipment 13.43 High Consumer Services 12.04
Household & Personal Products 14.55 High Household & Personal Products 16.47
Health Care Equipment & Services 16.29 High Telecommunication Services 16.67
Telecommunication Services 16.64 High Media & Entertainment 16.77
Media & Entertainment 18.56 High Software & Services 17.64
Software & Services 18.71 High Health Care Equipment & Services 17.68
Pharmaceuticals, Biotechnology & 19.85 High Pharmaceuticals, Biotechnology & 19.61
Life Sciences Life Sciences
Notes: The average composite intangible-intensity ranks shown were computed as the average of annual composite intangible-
intensity ranks. For each industry, the annual composite intangible-intensity rank was the average of its intensity ranks for the
following three intangible-intensity metrics: (1) Intangible Assets (excluding Goodwill)/Total Assets, (2) Research and Development
Expenses/Total Revenues, and (3) Sales, General & Administrative Expenses/Total Revenues. For each metric, ranks were annually
assigned to industries based on the relative industry median values of that metric.
intangibles. Consequently, we define compa- and three-digit SIC codes 282 (plastics and
nies as intangible intensive when their produc- synthetic materials), 283 (drugs), and 357
tion functions likely contain large amounts of (computer and office equipment).
unrecorded intangibles. We recognize that any
such classification is somewhat ad hoc. We Since the time of the Collins et al. (1997) study,
define intangible intensive as being compa- however, the relative importance of other types of
nies in the two-digit SIC codes 48 (electronic intangible capital, especially organization capital, has
components and accessories), 73 (business grown and additional intangible-intensive industries,
services), and 87 (engineering, account- such as Media & Entertainment, with new types of
ing, R&D and management related services) intangible capital, like subscriber lists, have emerged.

Table 2. Consistency of US and International Composite Intangible-

Intensity Ranks, 1994–2018
Year W-statistic χ2 p-Value
1994 0.86 34.34 0.02

1995 0.82 32.85 0.04
1996 0.79 31.50 0.05
1997 0.80 31.99 0.04
1998 0.85 34.15 0.03
1999 0.88 35.23 0.02
2000 0.93 37.20 0.01
2001 0.90 35.95 0.02
2002 0.90 35.80 0.02
2003 0.90 36.16 0.01
2004 0.92 36.83 0.01
2005 0.94 37.41 0.01
2006 0.92 36.94 0.01
2007 0.90 36.15 0.01
2008 0.94 37.64 0.01
2009 0.94 37.58 0.01
2010 0.94 37.59 0.01
2011 0.94 37.66 0.01
2012 0.96 38.32 0.01
2013 0.93 37.38 0.01
2014 0.97 38.75 0.01
2015 0.94 37.51 0.01
2016 0.97 38.74 0.01
2017 0.95 38.05 0.01
2018 0.97 38.62 0.01
Notes: For each of the three intangible-intensity metrics, ranks were annually assigned to indus-
tries according to the relative industry median values of that metric. Then, an annual composite
intangible-intensity rank was computed for each industry as the average of the three intangible-
intensity ranks. Consistency of the composite intangible-intensity ranks between universes was
evaluated annually on the basis of Kendall’s coefficient of concordance. The null hypothesis was
that the composite intangible-intensity ranks of industries would have no correlation across the
US and international universes (W = 0). The alternate hypothesis was that the composite intangi-
ble-intensity ranks of industries were identical across the US and international universes (W = 1).
The p-values for each year are based on observations for 21 (four-digit GICS) industry groups.
Moreover, as Lev and Gu (2016) showed, corporate a misleading or inconsistent intangible intensity–
investment in intangible assets has increased so based classification of industries. Therefore, rank-
much faster than the investment in tangible assets ing and classifying industries by a broader set of
that since the mid-1990s, it has overtaken invest- intangible-intensity metrics is warranted. Doing so
ment in tangible assets. As discussed in the sec- may lead to conclusions that differ from prior work
tion “Data and Methodology,” ignoring intangibles about interindustry variations in the combined value
already recorded on the balance sheet may produce relevance of book value and earnings.
29
Volume 77 Number 2
Table 3. Consistency of Intangible-Intensity Ranks over Time, 1994–2018
Intangible Intangible
Assets R&D SG&A Assets R&D SG&A
(ex goodwill) Expenses Expenses Composite (ex goodwill) Expenses Expenses Composite
W-statistic 0.88 0.90 0.96 0.94 0.79 0.91 0.92 0.89
χ2 442.37 406.20 481.84 470.88 394.69 365.63 462.75 443.40
p-Value 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Notes: For each investment universe (US and international), this table shows the consistency of annual industry ranks for the three
intangible-intensity metrics and the consistency of annual composite intangible-intensity ranks based on the Kendall’s coefficient
of concordance. For each of the 21 industry groups shown in Table 1, the annual composite intangible-intensity rank is the average
of the intensity ranks for these three intangible-intensity metrics. For each metric, ranks were annually assigned to industries
according to the relative industry median values of that metric. For intensity of R&D expenses in the international universe,
24 observations were available for the Consumer Services, Real Estate, and Telecommunication Services industry groups and
19 observations were available for the Food & Staples Retailing industry group. For intensity of R&D expenses in the US universe,
22 observations were available for the Food & Staples Retailing and Transportation industry groups.
Oddly, the limited international research in this area net income and book value. We made the following
has also relied on the Collins et al. (1997) categoriza- hypotheses: (1) If investments in intangible capital
tion of intangible-intensive industries, although the affect the value relevance of financial statements
system was initially conceived for the US universe. of companies in the high-intangible-intensity group
Hence, the implications of using our proposed more unfavorably than in the low-intangible-intensity
composite intangible-intensity measure to study the group, the combined R2 of book value and earnings
effects of intangible intensity on the value relevance should be lower for the high-intangible-intensity
of book value and earnings in the international uni- group, and (2) if the adverse effect of intangible
verse are unknown and deserve investigation. capital investment on the value relevance of earnings
and book value for high-intangible-intensity compa-
Following previous research, we used regression nies has intensified over time, regression R2 values
analysis to investigate the impact of intangible for that group should be shown to decline gradually
intensity on the value relevance of earnings and book over time and should slide below that for the low-
value. For each investment universe, we regressed intangible-intensity group.
contemporaneous share price on net income per
share and book value per share for companies in The R2 values obtained from our annual regres-
each intangible-intensity category (based on the sions are plotted in Figure 1 for the US universe and
ranks of industry groups in Table 1). For all sample Figure 2 for the international universe, where the
companies, we obtained book values, net income, dashed lines are the linear trends. These data tend
and the outstanding number of shares for each fiscal to support our hypothesis, although the inference
year between 1994 and 2018. We also extracted, is weaker for US companies than for international
using the filing dates for the reports provided by companies.
Xpressfeed, the month-end share price for the month
in which the financial report containing book value Overall, our findings are consistent with the results
and net income became publicly available. To match of comparable analyses conducted by Ciftci et al.
book value and net income with contemporaneous (2014) and Core et al. (2003) for US companies. First,
share prices, we excluded observations for which the in Figures 1 and 2, we detect a declining linear trend
month-end date of the share price was more than six for the value relevance of earnings and book value
months beyond the end of the annual fiscal period among companies in the high-intangible-intensity
covered by the financial report. group in both the US and international universes over
the sample period. Second, we note a sharp drop
We estimated all regressions annually and com- in the value relevance of these financial variables
puted the R2 values for each regression; higher R2 for both groups during the 1995–99 so-called New
values denote more combined value relevance for Economy period in both investment universes.19

Figure 1. R2 Values R2

from Annual Regressions 0.65
of Share Price on Book
0.55
Value per Share and Net
Income per Share: US 0.45
Universe, 1995–2019
0.35
0.25
0.15
1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 2019
Low Intangible Intensity High Intangible Intensity

Linear Trend
Thus, our objective and comprehensive methodol- To evaluate the statistical significance of the change
ogy for classifying industries into low- or high- in R2 values over time, we conducted two types of
intangible-intensity categories leads to conclusions tests. First, to determine the trend of annual R2 val-
that are comparable to the conclusions from previous ues, we computed the Theil–Sen20 slope of each of
research for companies in the United States. As for the four sets of 25 (for each year) R2 values (i.e., for
the international universe, we believe our study is the two intangible-intensity groups in each invest-
the first to document that similar relationships exist ment universe) and the related z-statistic for each
between intangible intensity and the value rel- slope estimate. These results are reported in Table 4.
evance of earnings and book values for international
companies. For the US universe, Table 4 shows that the trend
was strongly negative and statistically significant
Nevertheless, we also note some differences in for the high-intangible-intensity group and slightly
results from analogous past research on US com- positive but insignificant for the low-intangible-
panies. Figure 1 indicates that the value relevance intensity group. The 95% confidence intervals for
of book value and earnings for companies in the the two trend estimates do overlap slightly.21 Taken
low-intangible-intensity group increased from 2009 together, these findings imply that in the United
on; this relatively recent period was not included in States, the combined value relevance of book value
the Ciftci et al. (2014) and Core et al. (2003) studies. and earnings has decreased over time for compa-
For companies in the high-intangible-intensity group, nies in the high-intangible-intensity group but this
a similar upswing is visible beginning in 2014. decrease has not occurred for companies in the
Figure 2. R2 Values R2

from Annual Regressions 0.85
of Share Price on Book
0.75
Value per Share and
Net Income per Share: 0.65
International Universe,
1995–2019 0.55
0.45
0.35
1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 2019
Low Intangible Intensity High Intangible Intensity

Linear Trend
31
Volume 77 Number 2
Table 4. Trend of R2 Values from Regressions of Share Price on Book Value per Share
and Net Income per Share, 1995–2019
Low Intangible High Intangible Low Intangible High Intangible

Intensity Intensity Intensity Intensity
Theil–Sen’s slope 0.003 –0.006 0.007 –0.001

z-Statistic 1.10 –1.94* 4.65** –0.30
Notes: Shown are the nonparametric estimates of the linear trend (Theil–Sen’s slope) of the 25 annual R2 values obtained from the
cross-sectional regression of share price on book value per share and net income per share within each intangible-intensity group.
This slope is the median of the slopes of all lines through pairs of points (i.e., pairs of R2 values) in the sample. For each of the two
investment universes, the Theil–Sen slope was estimated separately for the high- and low-intangible-intensity groups.
*Significant at the 5% level.
**Significant at the 1% level.
low-intangible-intensity group. From Figure 1, a clear provided evidence that in the more permissive pre-
divergence in value relevance for high- and low- IFRS regime, capitalization was informative to inves-
intangible-intensity industries is evident in the US tors for companies that were intangible intensive.
universe after 2008, but the magnitude of the differ- IFRS adoption compelled international companies to
ence between the two groups fluctuates over time. hew more closely to US accounting provisions, which
tend to prohibit capitalization of intangibles. Scaling
For the international universe, the z-statistic for the back the capitalization option may have caused the
trend of the combined value relevance of earnings informativeness of book value and earnings to drop
and book value is negative but statistically insignifi- after implementation of IFRS, especially for highly
cant for the high-intangible-intensity group but posi- intangible intensive international companies.
tive and significant for the low-intangible-intensity
group. Furthermore, note that the 95% confidence Figure 1 and Figure 2 also highlight the differences
intervals for the trend of R2 values for the high- and between the high- and low-intangible-intensity
low-intangible-intensity groups do not overlap.22 The groups in how the combined value relevance of book
nonoverlapping confidence intervals allow us to infer value and earnings changed during and immediately
that the difference between the slight downward after the New Economy period. For both US and
trend for the high-intangible-intensity group and the international companies, the value relevance of
upward trend for the low-intangible-intensity group book value and earnings fell more sharply for high-
is statistically meaningful. intangible-intensity industries than low-intangible-
intensity industries during the dot-com bubble of
The R2 values for the international universe plotted 1995–1999, but as noted by Core et al. (2003) and
in Figure 2 show a steady decline beginning in 2006 Ciftci et al. (2014), the cause may have been tempo-
in the combined value relevance of book values and rary overoptimism about companies that represented
earnings for the high-intangible-intensity compa- the New Economy. Furthermore, for international
nies but a gradual increase in value relevance for companies, the extraordinary increase in intangible
low-intangible-intensity companies. This increasing investment in the mid- and late 1990s seems to have
divergence between the two groups for international reversed course in 2000, leading to a correspondingly
companies may be the result of standardization of greater rebound in value relevance of book value
accounting policies governing the capitalization of and earnings for high-intangible-intensity industries.
intangibles after the widespread adoption of IFRS For US companies, the period of excessive optimism
in 2005. In the pre-IFRS period, legacy accounting seems to have been longer, the decline in value
standards in several countries—notably, Australia, relevance more gradual, and the subsequent rebound
France, and the United Kingdom—permitted both more muted and occurring over a shorter period.
capitalization and expensing of the costs incurred to
create intangible capital assets. Goodwin and Ahmed To gain assurance that these empirical findings were
(2006) and Oswald, Simpson, and Zarowin (2017) not driven by a few industries in either of the two

intangible-intensity groups or by systematic interin- CV−Pgt = the coefficient of variation of share price
dustry differences in the relationship between the for each year and intensity group
fundamental financial variables and stock prices, we
conducted an additional test. We regressed the con- CV−BVPS gt = the coefficient of variation of book
temporaneous share price on net income per share value per share for each year and intensity group
and book value per share for companies in each of egt = the regression error for each year and inten-
the 21 industries and estimated the time trend of sity group
the 25 annual R2 values for each industry. For each
investment universe, we then calculated the correla- Ciftci et al. (2014) and Brown, Lo, and Lys (1999)
tion between the 21 industry time trends and the emphasized that R2 values are not comparable across
corresponding 25-year average composite intangible regressions conducted on subsamples of compa-
intensities for the 21 industries. We found that nies because of differences in scale. To control for
for the US (international) universe, the correlation such differences, they recommended that certain
between the time trend of R2 values and composite additional independent variables be included when
intangible intensity across all 21 industries was –0.59 combining or comparing regression results for
(–0.62). Both correlations are significant at the 99% different samples of companies or time periods.
level of confidence. They confirm the existence of In fact, Ciftci et al.’s replication of the Collins et al.
a strong negative relationship between intangible (1997) study that included controls for differences in
intensity and the value relevance of book value and scale led to conclusions about changes in the value
earnings for industries around the world. relevance of book value and earnings over time that
were opposite to those of Collins et al. Therefore, we
In addition to analyzing the trend of R2 values for included the two independent scale control variables
each intangible-intensity group, we used the meth- in our panel regressions, the coefficient of variation
odology in Ciftci et al. (2014) to test for differences of share price and the coefficient of variation of
in the combined value relevance of book value and book value per share, as suggested by Brown et al.
earnings between the high- and low-intangible- Our results for regressions that did not include the
intensity groups. For each investment universe, scale control variables (not reported for brevity) are,
we estimated the following panel regression of the however, qualitatively similar.
25 annual R2 values for both intangible-intensity
groups together. We used dummy variables to desig- Regression results for both the US and international
nate time, intangible intensity, and the interaction of universes are reported in Table 5. The coefficient b3
time and intangible intensity, and we included certain for the variable INT−Dh × TIMEt is of particular interest
scale control variables. because it captures the difference in slopes between
the low- and high-intangible-intensity groups. For
The regression was both the US and the international universes, this
regression coefficient is negative and statistically sig-
R2gt = a + b1TIME + b2INT_ Dh + b3INT_ Dh × TIMEt nificant, at confidence levels exceeding, respectively,
(1)
99% and 93%. This finding indicates that for US and
+ b4CV_ Pgt + b5CV_ BVPS gt + egt ,
international companies, the slope of R2 values rep-
resenting the combined efficacy of book value and
where earnings in explaining contemporaneous share prices
has been dropping over time for the high-intangible-
R2gt = the R2 for the regression of share price on
intensity group relative to the low-intangible-
earnings and book value for each year, t, and
intensity group.23 This trend is visually depicted in
intensity group, g
Figure 1 and Figure 2 and confirms our previous
TIME = a variable with values between 1 and analysis of the nonparametric trend of R2 values.24
25 depending on the year of the regression
Finally, from Figures 1 and 2, we also conclude that
INT−Dh = a variable with the value of 1 if the R2 was the declining linear trend of the (combined) value
for an observation in the high-intangible-intensity relevance of book value and earnings for high-
group and 0 otherwise intangible-intensity companies has been slightly
INT−Dh × TIMEt = the value of INT−Dh multiplied by greater in the United States than internationally.
TIME Over the full period of our study, the R2 of the
regression fell by about 30% (from 0.55 to 0.385)
33
Volume 77 Number 2
Table 5. Trend of R2 Values from Regression of Share Price on Book

Value per Share and Net Income per Share, 1995–2019
Coefficient p-Value Coefficient p-Value
a (intercept) 0.992 0.00 1.661 0.00

TIME 0.006 0.00 0.018 0.00
INT_Dh 0.102 0.00 0.093 0.01
INT_Dh × TIMEt –0.004 0.06 –0.014 0.00
CV_Pgt –0.278 0.03 –0.139 0.72
CV_BVPSgt –0.208 0.00 –0.627 0.09
Note: This table shows the regression coefficients and their corresponding p-values from the
cross-sectional regression given in Equation 1.
for the high-intangible-intensity group of US compa- To investigate the value relevance of earnings and
nies as opposed to a 22% drop (from 0.75 to 0.585) book value, we proposed and validated a composite
for the high-intangible-intensity group of interna- measure of intangible intensity that captures the
tional companies. financial statement impact of three types of intangi-
ble capital: intangible assets reported on the balance
To our knowledge, the findings we report here for sheet (excluding goodwill), innovation capital cre-
international companies constitute a new contri- ated by R&D expenditures, and organization capital
bution to the literature because no past studies resulting from SG&A expenses. We first showed
have examined how intangible intensity affects the that our composite intangible-intensity measure is
relationship between financial statement variables consistent over time and across the US and interna-
and stock prices in multiple countries. tional investment universes in its ability to rank and
classify industries by their intangible intensity. We
then analyzed the contemporaneous relationship
Conclusions between stock price and the two financial variables
Earnings and book value are of interest to investors of interest, book value per share and net income per
because these variables underlie two correspond- share, for US and non-US companies. We formed
ing valuation ratios—earnings to price and book to these subsamples on the basis of the intangible
price—that are the basis of popular value investing intensity of the industry to which the companies
and other types of investment strategies based on belong. As we hypothesized, we did find a decline in
the profitability, quality, growth, and risk charac- the combined value relevance of earnings and book
teristics of companies. However, the efficacy of value of companies in the high-intangible-intensity
value-investing strategies has fallen precipitously in group in both the US and international universes, but
recent years. A possible reason is that the volume we did not find such a decline for companies in the
and variety of corporate expenditures on activities low-intangible-intensity group.
that create intangible capital have increased, albeit
unevenly, over time and for different industries but Our approach to this issue differs from that of
financial reporting standards did not accommodate Arnott et al. (2021), Amenc et al. (2020), and Li
such structural economic changes. Our primary (2020), who attempted to adjust book values for the
conclusion is that intangible capital intensity is, in impact of unaccounted intangible capital, and from
fact, related to changes in the value relevance of that of Angelopoulos et al. (2012), who estimated
earnings and book value, as reflected in the power of industry-relative intangible intensity for companies
these financial variables to explain contemporaneous in each industry. Although such methods can com-
movements and cross-sectional variation in stock pensate for biases in valuation metrics that result
price in the 1994–2019 time period for the compa- from inadequate accounting of intangible capital,
nies in our global sample. company-level estimates of intangible capital may

be volatile and fraught with measurement error. companies, in that the divergence between the low-
The industry-level methodology for gauging intan- and high-intangible-intensity groups of industries is
gible intensity that we used in this study can mitigate greater in the international arena and has continued
both problems. Moreover, industry-level measures to increase over time. For investors who aim to build
of intangible intensity can capture macroeconomic and use value investing or other types of strategies
aspects of intangible intensity—such as industry that rely on book values and earnings, the implication
concentration (see Crouzet and Eberly 2019) and is that such strategies may benefit from taking varia-
product/market competition (see Gu 2016)—that tions in intangible intensity into account. Our primary
industry-relative estimates of the intangible intensity objective in this study, however, was to measure
of individual companies are unable to incorporate. intangible intensity and establish that it is relevant
Nevertheless, we acknowledge that all these alterna- for investors as a step toward building a robust and
tives are imperfect ways to address this important consistent investment framework. Therefore, we did
but complex issue. not attempt to investigate whether value investors
can enhance the return prediction ability of their val-
Our conclusions held for both US and international uation models or whether equity investors in general
companies in the largest 14 economies of the world. can improve their assessments of the profitability,
Importantly, our conclusions about the impact of quality, growth, and risk of companies by accounting
intangible intensity on the value relevance of earn- for cross-sectional variations in intangible intensity.
ings and book value are stronger for international We leave these issues to future research.
Appendix A. Details of Sample Data

Figure A1. Availability A. Number of Companies for Which Balance Sheet Items Were Available
of Data Items Required Number of Companies
for Computation of 12,000
Intangible-Intensity 10,000
Metrics by Year: 8,000
US Companies, 6,000
1994–2018 4,000
2,000
0
1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
Total Assets Intangible Assets Goodwill
B. Number of Companies for Which Income Statement Items Were Available

Number of Companies
12,000
10,000
8,000
6,000
4,000
2,000
0
1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
Revenues SG&A Expenses R&D Expenses
35
Volume 77 Number 2
Figure A2. Availability A. Number of Companies for Which Balance Sheet Items Were Available
of Data Items Required Number of Companies
Metrics by Year: 16,000
International 12,000
Companies, 8,000
1994–2018
4,000
0
1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
B. Number of Companies for Which Income Statement Items Were Available

Number of Companies
24,000
20,000
16,000
12,000
8,000
4,000
0
1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
Table A1. Availability of Data Items Required for Computation of Intangible-Intensity

Metrics by Country, 1994–2018 (in company-year units)
Total Total Intangible R&D SG&A

Assets Revenues Assets Goodwill Expenses Expenses
United States 236,008 245,275 123,180 78,034 76,188 237,294
Developed international markets

Australia 34,762 34,862 16,176 9,879 5,377 30,563
Canada 66,162 66,565 20,965 10,814 9,149 63,752
France 15,539 15,578 13,461 11,093 3,400 13,161
Germany 19,225 19,299 15,873 10,318 5,150 17,187
Italy 5,924 5,939 5,377 4,472 1,044 5,214
Japan 78,754 78,806 73,902 22,710 38,347 78,290
Spain 4,210 4,215 3,580 2,316 661 3,848
UK 47,491 47,594 26,504 19,097 8,402 43,837
Total developed 272,067 272,858 175,838 90,699 71,530 255,852
(continued)

Table A1. Availability of Data Items Required for Computation of Intangible-Intensity

Metrics by Country, 1994–2018 (in company-year units) (continued)
Total Total Intangible R&D SG&A

Assets Revenues Assets Goodwill Expenses Expenses
Emerging international markets

Brazil 7,636 7,695 4,437 2,067 875 7,557
China 40,663 40,727 38,612 14,358 19,780 40,585
India 55,537 56,056 25,199 11,037 10,827 54,693
Mexico 2,400 2,419 1,537 1,026 87 2,311
Russia 3,551 3,561 2,624 702 392 3,221
South Korea 29,476 29,514 26,894 11,256 18,475 29,160
Total emerging 139,263 139,972 99,303 40,446 50,436 137,527
Total international 411,330 412,830 275,141 131,145 121,966 393,379
Notes: The three intangible-intensity metrics are intangible assets (excluding goodwill)/total assets, R&D expenses/total revenues,
and SG&A expenses/total revenues. Intangible-intensity metrics were computed for each of 21 industry groups, which excludes
Banks, Diversified Financials, and Insurance. Of 123,180 (275,141) company-years in the US (international) universe in which intan-
gible assets were reported, 20,069 (18,905) were for companies in the excluded industries. Of 78,034 (131,145) company-years
in the US (international) universe in which goodwill was reported, 13,102 (10,433) were for companies in the excluded industries.
Of 76,188 (121,966) company-years in the US (international) universe in which R&D expenses were reported, 1,151 (1,499) were
for companies in the excluded industries. Of 237,294 (393,379) company-years in the US (international) universe in which SG&A
expenses were reported, 62,498 (43,226) were for companies in the excluded industries.
Figure A3. Availability A. Number of Company-Years for Which Balance Sheet Items Were Available
of Data Items Required Number of Company-Years

16,000
Metrics by Industry: 12,000
US Companies, 10,000
8,000
1994–2018 6,000
4,000
2,000
0
ar me Sc rgy
so ns s
m nte ipm s
Se ces
ve ns Re Re s
Fo A s
at s
r s
l G re
ch qu L E ials
ge e E il
m To les
om l s
Se es
l G rt
Re o
Co erc tain ent
er rvi t
o u te
il
du co
um Se en
er ra ce
Co / E Equ vice
ec U tor
e
d
M ood
ed re Se e
ra um al ta
ta
od ut
na po
ta a
/ F r D sta
ic
M wa nt/ ienc
oo
on c
Se od/ rab
pi w
. H ip ife ne
er
/P T rvi
ns ial m
ic ba
i
rv
c
Ca So
m r
ia /
Te h E , &
T el
lt ch
ld
d
Be Co
ho
ea e
H iot
se
B
ou
s,
H
al
c
eu
ac
m
ar
Ph
B. Number of Company-Years for Which Income Statement Items Were(continued)

Available
Number of Company-Years
18,000
16,000
14,000
12,000
Volume 77 Number 2 37
10,000
8,000
6,000
.
s
B
lt
Be
ch
ou
ea
s,
Te
H
al
H
c
eu
ac
m
ar
Ph
Figure A3. Availability B. Number of Company-Years for Which Income Statement Items Were Available
16,000
US Companies, 1994– 10,000
8,000
2018 (continued)
6,000
4,000
2,000
0
ar me Sc rgy
U ces
m nte ipm s
Se ces
ve ns Re Re s
Fo A s
at s
r s
l G re
ch u L E ials
ge e E il
m To les
Se ors
so ns s
l G rt
Re o
Co erc tain ent
er rvi t
o u te
il
co du o
um Se en
Co / E Equ vice
d
M ood
ed re Se e
er ra e
ra um al ta
ta
od ut
le n cc
na po
ta a
/ F r D sta
ic
M wa nt/ ienc
/P T li
oo
Se od/ rab
m ct
pi w
. H ip ife ne
er
i
ns ial m
Te ico ba
rv
rv
Ca So
m r
ia /
Te h E , &
h
ld
d
Be Co
c
q
ho
ea e
H ot
se
Bi
lt
ou
s,
H
al
c
eu
ac
m
ar
Ph
Figure A4. Availability A. Number of Company-Years for Which Balance Sheet Items Were Available

70,000
Intangible-Intensity
60,000
International 40,000
Companies, 1994–2018 30,000
20,000
10,000
0
ge R uip gy
l G es
an s
rv s
t/ Au s
ds
ur s
ns al e s
er Go ls
w s
e/ E are
m er il
Se to
m U ort
se Te F duc s
ld co d rs
Bi ia o Es t
ec En To te
ie t
so er il
Li in o
Sc en
s, ed Fo eal en
Tr vice
Se ce
e
D od
So ble
e
ta
er S ta
& rta cc
um tal ria
ho le oo to
ot / d/ ta
na vic
ic
nc
on li
oo
r
sp
/P m Re
Eq ne
er vi
fe m
m
h, te ba
a
ns pi te
r
Co Ca Ma
c
i
en
Co rc
i
u
e
ar
pm
Se
m
al M /
dw
ui
Co
r
Eq
ra
a
.H
ve
lth
Be
ch
ou
ea
Te
H
H
c
eu
ac
m
ar
Ph
(continued)
B. Number of Company-Years for Which Income Statement Items Were Available
Number of Company-Years
80,000
70,000
60,000
50,000
40,000
38 30,000 Second Quarter 2021
20,000
10,000
ot
h
.H
al M
ve
se
lth
Bi
Be
ch
ou
ea
s,
Te
H
H
c
eu
ac
m
ar
Ph
Figure A4. Availability B. Number of Company-Years for Which Income Statement Items Were Available
60,000
Metrics by Industry:
50,000
International
40,000
Companies, 1994–2018 30,000
(continued) 20,000
10,000
0
ge R uip gy
l G es
an s
rv s
t/ Au s
ds
ur s
ia Re s
er Go ls
w s
e/ E are
m er il
Se to
m U ort
se Te F duc s
ld co d rs
Bi ia o Es t
so er il
ec En To te
ie t
Li in o
Sc en
s, ed Fo al en
Tr vice
Se ce
e
D od
So ble
e
s u l S ta
er S ta
& rta cc
um tal ria
ho le oo to
ot / d/ ta
na vic
ic
nc
on li
oo
r
sp
/P m Re
Eq ne
er vi
fe m
m
h, te ba
a
ns pi te
r
Co Ca Ma
ic
en
Co erc
ar
pm
Se
m
/
n
dw
ui
Co
ar
Eq
ra
.H
M
e
lth
v
Be
ch
ou
ea
Te
H
al
H
c
eu
ac
m
ar
Ph
Editor’s Note acknowledgments. John Adams and one anonymous

reviewer were the reviewers for this article.
This article was externally reviewed using our double-blind
peer-review process. When the article was accepted for Submitted 19 October 2020
publication, the authors thanked the reviewers in their Accepted 30 December 2020 by Stephen J. Brown
Notes
1. The mismatch may be more acute for business entities that the Xpressfeed database, 63,602 (46,160) company-year
are in the early stages of their life cycle, when spending on observations were for companies in the Banks, Insurance,
activities that create intangible assets (e.g., research and and Diversified Financials industry groups.
development or customer acquisition) is high.
6. In a pre-IFRS reporting regime, certain countries (includ-
2. In addition to net income and book value, they studied ing Australia, the United Kingdom, and France) permitted
cash flow from operations, cash, total assets, intangible both expensing and capitalization of R&D expenditures.
assets, sales, sales growth, R&D expenses, advertising Oswald, Simpson, and Zarowin (2017) found differences in
expenses, cost of goods sold, capital expenditures, other (1) the value relevance of the capitalized versus expensed
comprehensive income, and special items. development costs in a pre-IFRS regime and (2) changes
in the value relevance of R&D expenditures before and
3. Intangible capital items may include computerized after IFRS adoption for UK companies that switched from
information, innovation (including both scientific R&D expensing to capitalization. Jaafar (2011) showed that the
and nonscientific discovery and development; Corrado, adoption of Australian-equivalent IFRS led to an increase
Hulten, and Sichel 2005), human resources (Pantzalis in the value relevance of identifiable intangible assets.
and Park 2009), organizational competencies (Lev and
Radhakrishnan 2005), customer franchises (Bonacchi, 7. For example, Lee and Lee (2020) stated that prior to 1999
Kolev, and Lev 2015), and brand values (Barth, Clement, in South Korea, R&D expenditures were classified as either
Foster, and Kasznik 1998). ordinary or extraordinary depending on the characteristics
of the activities; R&D expenditures that occurred in the
4. Li (2020) argued and showed that, at least in developed ordinary course of business were expensed, whereas those
international markets, book value can be successfully not meeting this criterion were capitalized.
adjusted without relying on the complex procedures
suggested by Peters and Taylor (2017). 8. Most prior studies on this subject (e.g., Bublitz and
Ettredge 1989) have shown that the life of brand value
5. Of the 236,008 (411,330) company-year observations in assets created by advertising expenditures is no more than
our US (international) sample for which ubiquitous finan- one to two years.
cial statement items, such as total assets, were available in
39
Volume 77 Number 2
9. In unreported results, we found that for our sample of enable assignment of more appropriate (unequal) weights
US companies, advertising expenditures dropped from to different sources of intangible capital.
about 3.6% to 1.6% of total revenues while R&D expen-
ditures rose from about 9.5% to more than 14.0% in the 18. We required at least three companies in an industry for
1994–2018 period. estimation of the median intangible intensity, and we
required that intangible intensity medians for at least two
10. In descending order of GDP, the top 15 countries in the of the three metrics be available for computation of the
world are the United States, China, Japan, Germany, the composite intangible intensity in any year.
United Kingdom, France, India, Italy, Brazil, Canada, Russia,
South Korea, Australia, Spain, and Mexico. The complete 19. According to Core et al. (2003), this period was marked by
list by GDP ranking is available from the World Bank several unusual economic developments, including high
(2019). stock market returns, high valuations, and increased pro-
ductivity driven by the declining price of computing power
11. Xpressfeed reports R&D expenses and SG&A expenses and investments in information technology and modern
as separate components of the income statement item manufacturing facilities that benefit from information
under “Other Operating Expenses.” Our three metrics had technology.
to have nonnegative values, and their intensity could not
be computed if the scaling variable (total assets or total 20. The Theil–Sen estimator (Theil 1950; Sen 1968) is a
revenues) was missing or zero. Such cases (amounting to nonparametric technique for estimating a linear trend by
fewer than 0.4% of all available company-year observa- choosing the median of the slopes of all lines through pairs
tions) were treated as data errors and excluded from of points in the sample. This procedure produces a (statis-
our sample. tically) efficient estimator that is insensitive to outliers. It
can be significantly more accurate than a nonrobust simple
12. Relative rankings for intangible intensity based on alterna- linear (least-squares) regression for skewed and heteroske-
tive measures (total assets or total expenses) to scale R&D dastic data.
expenses and SG&A expenses were similar and are not
reported here for brevity. 21. For the US universe, the 95% confidence intervals for,
respectively, the high- and low-intangible-intensity groups
13. Table A1 in Appendix A provides some summary informa- are (–0.011, 0.000) and (–0.002, 0.007). For the interna-
tion about the availability of such data separately for tional universe, the 95% confidence intervals are (–0.005,
developed and emerging countries. 0.004) and (0.005, 0.012).
14. In light of our focus on interindustry differences in intan- 22. See previous note.
gible intensity, we provide, in Figure A3 and Figure A4
in Appendix A, information about the availability of the 23. We repeated the analyses in Table 4 and Table 5 for a full
requisite data at the industry level for, respectively, the “global” sample of companies. For this test, we combined
United States and other countries. US and international companies in the high-intangible-
intensity groups, added an indicator variable to distinguish
15. We recognize that this approach is imperfect. Within the whether a particular company belonged to the US or
four-digit GICS classifications that we used, intangible international universe, and ran our primary annual cross-
intensity can vary at the subindustry level; for example, sectional regression of stock price on book value and earn-
within the Utilities sector, the wind and solar power sec- ings for this “global” sample of high-intangible-intensity
tors are likely to differ from those that rely on fossil fuels companies. We obtained 25 R2 values. We repeated the
and nuclear energy. A more granular industry classification same procedure for the low-intangible-intensity compa-
approach would yield additional insights but would come at nies. The Theil–Sen’s slopes (z-statistics) for the high- and
the expense of reduced sample sizes at the industry level. low-intangible-intensity groups were, respectively, –0.00
(0.26) and 0.007 (3.10), and the coefficient (t-statistic) for
16. Figures A3 and A4 confirm this conjecture. INT−Dh × TIMEt was –0.007 (–3.49).
17. We acknowledge that our assumption that all sources 24. The coefficient on the TIME variable is positive and
of intangible capital are equally important contributors significant for both the US and international regressions,
to intangible intensity is subjective. The three metrics indicating that the combined value relevance of earnings
of intangible capital have different useful lives, and they and book value has increased for companies in the low-
differ in the amount and timing of the cash flows they intangible-intensity group for the time period and sample
generate. Accurate measurement of these attributes would of companies included in our study.
References
Aboody, David, and Baruch Lev. 1998. “The Value Relevance Ahmed, Kamran, and Haim Falk. 2006. “The Value Relevance
of Intangibles: The Case of Software Capitalization.” Journal of of Management’s Research and Development Reporting
Accounting Research 36 (Supplement: Studies on Enhancing the Choice: Evidence from Australia.” Journal of Accounting and
Financial Reporting Model): 161–91. Public Policy 25 (3): 231–64.

Amenc, Noel, Felix Goltz, and B. Luyten. 2020. “Intangible Eisfeldt, Andrea L., and Dimitris Papanikolaou. 2013.
Capital and the Value Factor: Has Your Value Definition Just “Organization Capital and the Cross-Section of Expected
Expired?” Journal of Portfolio Management 46 (7): 83–99. Returns.” Journal of Finance 68 (4): 1365–406.
Angelopoulos, Georgios, Daniel Giamouridis, and Orestes Fraser, Ian, Heather Tarbert, and Kai Hong Tee. 2009. “Do the
Vlismas. 2012. “Inferring the Value of Intangible Assets.” Financial Statements of Intangible-Intensive Companies Hold
Working paper (1 March). Available at https://ssrn.com/ Less Information Content for Investors?” Applied Financial
abstract=1786285. Economics 19 (17): 1433–38.
Arnott, R. D., C. R. Harvey, V. Kalesnik, and J. T. Linnainmaa. Goodwin, John, and Kamran Ahmed. 2006. “Longitudinal
2021. “Reports of Value’s Death May Be Greatly Exaggerated.” Value Relevance of Earnings and Intangible Assets: Evidence
Financial Analysts Journal 77 (1): 44–67. from Australian Firms.” Journal of International Accounting,
Auditing and Taxation 15 (1): 72–91.
Barth, Mary, Michael B. Clement, George Foster, and Ron
Kasznik. 1998. “Brand Values and Capital Market Valuation.” Govindarajan, Vijay, Shivaram Rajgopal, Anup
Review of Accounting Studies 3 (1–2): 41–68. Srivastava, and Ye Wang. 2019. “R&D Spending Has
Dramatically Surpassed Advertising Spending.” Harvard
Barth, Mary E., Ken Li, and Charles G. McClure. 2018. Business Review (20 May). https://hbr.org/2019/05/
“Evolution in Value Relevance of Accounting Information.” rd-spending-has-dramatically-surpassed-advertising-spending.
Working paper, Stanford Graduate School of Business (May).
Gu, Lifeng. 2016. “Product Market Competition, R&D
Bonacchi, Massimiliano, Kalin Kolev, and Baruch Lev. 2015. Investment, and Stock Returns.” Journal of Financial Economics
“Customer Franchise—A Hidden, Yet Crucial, Asset.” 119 (2): 441–55.
Contemporary Accounting Research 32 (3): 1024–49.
Israel, R., K. Laursen, and Scott A. Richardson. 2020. “Is
Brown, S., K. Lo, and T. Lys. 1999. “Use of R2 in Accounting (Systematic) Value Investing Dead?” Working paper, AQR
Research: Measuring Changes in Value Relevance over the Capital Management (8 May).
Last Four Decades.” Journal of Accounting and Economics
28 (2): 83–115. Jaafar, Hartini. 2011. “Accounting for Intangible Assets, Firm
Life Cycle and the Value Relevance of Intangible Assets.”
Bublitz, B., and M. Ettredge. 1989. “The Information Doctoral dissertation, University of Tasmania.
in Discretionary Outlays: Advertising, Research and
Development.” Accounting Review 64 (1): 108–24. Joos, Peter, and George A. Plesko. 2005. “Valuing Loss Firms.”
Accounting Review 80 (3): 847–70.
Ciftci, Mustafa, Masako Darrough, and Raj Mashruwala. 2014.
“Value Relevance of Accounting Information for Intangible- Lee, Jangwook, and Jiyoon Lee. 2020. “Mispricing or Risk
Intensive Industries and the Impact of Scale: The US Premium? An Explanation of the R&D-to-Market Anomaly.”
Evidence.” European Accounting Review 23 (2): 199–226. Working paper (7 October). Available at https://ssrn.com/
abstract=3706541.
Collins, Daniel W., Edward L. Maydew, and Ira S. Weiss. 1997.
“Changes in the Value-Relevance of Earnings and Book Values Lev, Baruch, and Feng Gu. 2016. The End of Accounting and the
over the Past Forty Years.” Journal of Accounting and Economics Path Forward for Investors and Managers. Hoboken, NJ: John
24 (1): 39–67. Wiley & Sons.
Core, John E., Wayne R. Guay, and Andrew Van Buskirk. 2003. Lev, Baruch, and S. Radhakrishnan. 2005. “The Valuation of
“Market Valuations in the New Economy: An Investigation Organization Capital.” In Measuring Capital in a New Economy,
of What Has Changed.” Journal of Accounting and Economics edited by John C. Haltiwanger, Carol Corrado, and Daniel E.
34 (1–3): 43–67. Sichel, 73–99. Chicago: University of Chicago Press.
Corrado, C. A., C. R. Hulten, and D. Sichel. 2005. “Measuring Lev, Baruch, and Theodore Sougiannis. 1996. “The
Capital and Technology: An Expanded Framework.” In Capitalization, Amortization, and Value Relevance of R&D
Measuring Capital in a New Economy, edited by John C. Outlays.” Journal of Accounting and Economics 21 (1): 107–38.
Haltiwanger, Carol Corrado, and Daniel E. Sichel, 11–45.
Chicago: University of Chicago Press. Lev, Baruch, and Anup Srivastava. 2019. “Explaining the
Demise of Value Investing.” Working paper, NYU Stern School
Crouzet, Nicolas, and Janice C. Eberly. 2019. “Understanding of Business (25 October).
Weak Capital Investment: The Role of Market Concentration
and Intangibles.” NBER Working Paper 25869 (May). Li, Feifei. 2020. “Intangibles: The Missing Ingredient in Book
Value.” Working paper, Research Affiliates (2 November).
Dahmash, Firas N., Robert B. Durand, and John Watson. 2009. Available at https://ssrn.com/abstract=3686595.
“The Value Relevance and Reliability of Reported Goodwill and
Identifiable Intangible Assets.” British Accounting Review 41 (2): Ohlson, J. A. 1995. “Earnings, Book Values and Dividends in
120–37. Equity Valuation.” Contemporary Accounting Research 11 (2):
661–87.
Darrough, Masako, and Jianming Ye. 2007. “Valuation of Loss
Firms in a Knowledge-Based Economy.” Review of Accounting Oliveira, Lídia, Lúcia L. Rodrigues, and Russell Craig. 2010.
Studies 12 (1): 61–93. “Intangible Assets and Value Relevance: Evidence from the
Portuguese Stock Exchange.” British Accounting Review 42 (4):
Edmans, Alex. 2011. “Does the Stock Market Fully Value 241–52.
Intangibles? Employee Satisfaction and Equity Prices.”
Journal of Financial Economics 101 (3): 621–40. Oswald, Dennis, Ana Simpson, and Paul Zarowin. 2017. “The
Information Benefits of R&D Capitalization.” Working paper
(13 April). Available at https://ssrn.com/abstract=2952388.
41
Volume 77 Number 2
Pantzalis, C., and J. C. Park. 2009. “Equity Market Valuation Smith, Dean T., Majella Percy, and Gordon D. Richardson.
of Human Capital and Stock Returns.” Journal of Banking & 2001. “Discretionary Capitalization of R&D: Evidence
Finance 33 (9): 1610–23. on the Usefulness in an Australian and Canadian Context.”
Advances in International Accounting 14: 15–46.
Peters, R., and L. Taylor. 2017. “Intangible Capital and
the Investment-q Relation.” Journal of Financial Economics Theil, H. 1950. “A Rank-Invariant Method of Linear and
123 (2): 251–72. Polynomial Regression Analysis. I, II, III.” Proceedings of
the Nederl Akad Wetensch Indag Math 53: 386–92, 521–25,
Ritter, A., and A. Wells. 2006. “Identifiable Intangible Asset 1397–412.
Disclosures, Stock Prices and Future Earnings.” Accounting and
Finance 46 (5): 843–63. Tronconi, Claudia, and Giuseppe Vittucci Marzetti. 2011.
“Organization Capital and Firm Performance. Empirical
Sen, Pranab Kumar. 1968. “Estimates of the Regression Evidence for European Firms.” Economics Letters 112 (2):
Coefficient Based on Kendall’s Tau.” Journal of the American 141–43.
Statistical Association 63 (324): 1379–89.
World Bank. 2019. “Data Catalog: 2018 GDP Ranking.”
Shah, Syed Zulﬁqar Ali, and Saeed Akbar. 2008. “Value https://datacatalog.worldbank.org/dataset/gdp-ranking.
Relevance of Advertising Expenditure: A Review of the Accessed 17 December 2019.
Literature.” International Journal of Management Reviews
10 (4): 303–25. Zar, J. H. 1999. Biostatistical Analysis. Fourth edition.
Upper Saddle River, NJ: Prentice-Hall.
Shah, Syed Zulﬁqar Ali, Andrew Stark, and Saeed Akbar.
2009. “The Value Relevance of Major Media Advertising Zhao, Ronald. 2002. “Relative Value Relevance of
Expenditures: Some U.K. Evidence.” International Journal of R&D Reporting: An International Comparison” Journal
Accounting 44 (2): 187–206. of International Financial Management and Accounting
13 (2): 153–74.
Silva, Rita A. 2012. “Equity Valuation Using Accounting
Numbers in High and Low Intangible-Intensive Industries.”
Dissertation, Católica Lisbon School of Business and
Economics, Lisbon, Portugal.

Dugar 2021

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

Dugar 2021

Uploaded by

Copyright:

Available Formats

Financial Analysts Journal

ISSN: 0015-198X (Print) 1938-3312 (Online) Journal homepage: https://www.tandfonline.com/loi/ufaj20

Equity Investing in the Age of Intangibles

Amitabh Dugar & Jacob Pozharny

To link to this article: https://doi.org/10.1080/0015198X.2021.1874726

© 2021 The Author(s). Published with

Published online: 18 Mar 2021.

Submit your article to this journal

Article views: 3998

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

Equity Investing in the

This is an Open Access article distributed

22  Second Quarter 2021

24  Second Quarter 2021

26  Second Quarter 2021

Table 1. Average Composite Intangible-Intensity Ranks, 1994–2018

United States International

Utilities 3.84 Low Real Estate 4.20

28  Second Quarter 2021

Table 2. Consistency of US and International Composite Intangible-

Year W-statistic χ2 p-Value

1994 0.86 34.34 0.02

Table 3. Consistency of Intangible-Intensity Ranks over Time, 1994–2018

United States International

30  Second Quarter 2021

Figure 1. R2 Values R2

Low Intangible Intensity High Intangible Intensity

Figure 2. R2 Values R2

Low Intangible Intensity High Intangible Intensity

United States International

Low Intangible High Intangible Low Intangible High Intangible

Theil–Sen’s slope 0.003 –0.006 0.007 –0.001

32  Second Quarter 2021

Table 5. Trend of R2 Values from Regression of Share Price on Book

United States International

Coefficient p-Value Coefficient p-Value

a (intercept) 0.992 0.00 1.661 0.00

34  Second Quarter 2021

Appendix A. Details of Sample Data

for Computation of 12,000

Total Assets Intangible Assets Goodwill

B. Number of Companies for Which Income Statement Items Were Available

Revenues SG&A Expenses R&D Expenses

for Computation of 24,000

Total Assets Intangible Assets Goodwill

B. Number of Companies for Which Income Statement Items Were Available

Revenues SG&A Expenses R&D Expenses

Table A1. Availability of Data Items Required for Computation of Intangible-Intensity

Total Total Intangible R&D SG&A

United States 236,008 245,275 123,180 78,034 76,188 237,294

Developed international markets

36  Second Quarter 2021

Table A1. Availability of Data Items Required for Computation of Intangible-Intensity

Total Total Intangible R&D SG&A

Emerging international markets

for Computation of 18,000

Total Assets Intangible Assets Goodwill

B. Number of Company-Years for Which Income Statement Items Were(continued)

Total Assets Intangible Assets Goodwill

Revenues SG&A Expenses R&D Expenses

for Computation of 80,000

22 Second Quarter 2021

24 Second Quarter 2021

26 Second Quarter 2021

Table 1. Average Composite Intangible-Intensity Ranks, 1994–2018

28 Second Quarter 2021

Table 2. Consistency of US and International Composite Intangible-

Table 3. Consistency of Intangible-Intensity Ranks over Time, 1994–2018

30 Second Quarter 2021

32 Second Quarter 2021

Table 5. Trend of R2 Values from Regression of Share Price on Book

34 Second Quarter 2021

Table A1. Availability of Data Items Required for Computation of Intangible-Intensity

36 Second Quarter 2021

Table A1. Availability of Data Items Required for Computation of Intangible-Intensity

40 Second Quarter 2021

42 Second Quarter 2021