You are on page 1of 9

Advances in Accounting, incorporating Advances in International Accounting 25 (2009) 165173

Contents lists available at ScienceDirect

Advances in Accounting, incorporating Advances in


International Accounting
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / a d i a c

The valuation implications of human capital in transactions on and outside


the exchange
Ilanit Gavious a,, Meir Russ b,1
a
Department of Business Administration, Guilford Glazer School of Business and Management, Building 15 Room 122, The Marcus Family Campus,
Ben-Gurion University of the Negev, PO Box 653, Beer Sheva, 84105, Israel
b
Master of Management Frederick E. Baer Professor in Business, Professional Programs in Business, University of Wisconsin-Green Bay, WH460;
2420 Nicolet Drive, Green Bay, WI 54311-7001, USA

a r t i c l e

i n f o

Keywords:
Human capital
Human asset
Intellectual capital
Compensation expense
Value relevance

a b s t r a c t
We examine the valuation implications of human capital both for a broad sample of rms and for subsamples
of high-technology rms and low-technology rms. Our results suggest that the market appears to value
compensation expenses not as expenses but as if they serve as a proxy for a human asset that is omitted from
the balance sheets. The ndings are consistent with human capital comprising a more sizable portion of the
value of high-technology rms than of low-technology rms. The ndings also indicate that compensation
expenses are valued differently from other expense components of income. Markedly, despite critical
differences between investors on the exchange and those buying shares in transactions outside the exchange
(controlling interests, information asymmetry, etc.), their assessment of the enhanced value of a rm
attributable to human capital is shown to be relatively similar. The results in this study are consistent with
compensation expenses creating a valuable intangible asset, hence suggesting that reform in the accounting
treatment of these expenses is of critical importance.
2009 Elsevier Ltd. All rights reserved.

1. Introduction
Human capital comprises a sizable portion of the asset base of
companies in the 21st century knowledge economy. The changing
economy, in which there is a growing number of knowledge-based
emerging industries, triggers a need to comprehend the asset base of
companies, and how these assets contribute to value creation. A critical
aspect of that understanding is the ability to quantify the value of
intangibles. A large empirical literature provides ample evidence of the
growing importance of intangible assets (e.g., Lev, 2005) and intellectual
capital (e.g. Edvinsson & Malone, 1997) in the new, knowledge-based
economy, and the inadequacy of the existing accounting methods for
managing and measuring the value of those assets (e.g., Stewart, 1997;
Lev & Zarowin, 1999; Brown, Kim, & Lys, 1999; Francis & Schipper, 1999;
Core, Guay, & Van Buskirk, 2003).
In this study, we focus on the human aspect of the intangibles that
drive the value created by a company. Specically, we investigate the
market's assessment of human capital and the affect that this
assessment has on rm value. We inquire whether investors view
human asset-related expenses (compensation expenses) differently
from other expenses included in the rm's income statement. Generally
Accepted Accounting Principles (GAAP) mandate the immediate and

Corresponding author. Tel.: +972 8 6477538; fax: +972 8 6477691.


E-mail addresses: madaril@bgu.ac.il (I. Gavious), russm@uwgb.edu (M. Russ).
1
Tel.: +1 920 465 2757; fax: +1 920 465 2660.
0882-6110/$ see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.adiac.2009.09.004

full expensing of compensation costs because of concerns with the


reliability and objectivity of capitalization of such costs. While treated by
GAAP as a current expense, these amounts in fact provide information
about an unrecorded intangible asset (see, e.g., Bell, Landsman, Miller, &
Yeh, 2002). Writing-off investments in intangible assets leads to
depressed earnings and book values as the exaggerated costs are not
matched with their associated revenues. These accounting fundamentals earnings and book value of equity thus become less reliable and
less value relevant as investments in human capital grow.
Studies that examine the value relevance of accounting information over time support this notion. These studies show a consistent
decline in the association of stock returns with earnings and book
value of equity over the past two decades (Lev & Zarowin, 1999;
Brown et al., 1999; Francis & Schipper, 1999). Core et al. (2003) focus
on value relevance of accounting information in the New Economy
Period (19961999) and explore whether it has decreased relative to
previous periods. Differentiating between high-technology and lowtechnology rms, they nd that the ability of traditional nancial
variables to explain rm value of both groups of rms declined during
the New Economy Period. Many other value relevance studies focus
on a specic high-technology industry.2 High-technology industries
drew the attention of researchers in this context because these
2
E.g., Callen et al. (2008), Hand (2004, 2005), Ely et al. (2003), Guo et al. (2005) and
Xu et al. (2007) on biotechnology rms; Amir and Lev (1996) on cellular rms;
Trueman, Wong, and Zhang (2001), Jorion and Talmor (2001) and Rajgopal, Shevlin,
and Venkatachalam (2003) on internet rms.

166

I. Gavious, M. Russ / Advances in Accounting, incorporating Advances in International Accounting 25 (2009) 165173

industries are knowledge intensive and are characterized by large


investments in intangible assets (e.g., research and development
expenditures, intellectual property and intellectual capital). The
expensing of these investments has raised concerns that nancial
accounting may be unsuited for a changing economy (Callen,
Gavious, & Segal, 2008; p. 11).
The studies focusing on high-technology industries yield mixed
results. A lack of consensus exists, not only with regards to the value
relevance of earnings and book values, but also with regard to that of
research and development (R&D) expenditures. Notably, like compensation expenses, R&D is treated by GAAP as a current expense
rather than as an investment in an intangible asset. While many
studies document an expected positive relation between market
values and R&D (e.g., Xu, Magnan, & Andre, in press; Ely, Simko, &
Thomas, 2003; Callen et al., 2008; Guo, Lev, & Zhou, 2005; Hand, 2005;
Core et al., 2003), others show that public equity market valuations do
not necessarily increase with the magnitude of R&D expenditures. In
particular, reliable inferences regarding the valuation implication of
R&D in the case of venture-backed companies cannot be drawn (see,
e.g., Armstrong, Davila, & Foster, 2006; Gavious & Schwartz, 2008).
Noticeably, while R&D expenditures received much attention in
value relevance studies, investments in human capital have been
generally neglected, probably due to the fact that the denitions of
human capital are broad and difcult to measure (see, e.g., Swartz,
Swartz, & Firer, 2006).3 In our analyses, we utilize model specications
employed in accounting literature, that implicitly or explicitly follow
the Ohlson (1995,1999) and Feltham and Ohlson (1999) valuation
models. The model regresses market value of equity on book value of
equity, abnormal earnings and other information. Other information
modies the prediction of future protability. In their study, Core et al.
(2003) document that there is a great variation in values of hightechnology as well as low-technology rms during the New Economy
Period that remains to be explained. We hypothesize that the increase
in unexplained variation in rm values is by and large due to the
omission of a prominent value driver human capital which prior
studies show has an increasing importance to a rm's overall value and
thus should be incorporated in valuation models (see, e.g., Swartz
et al., 2006). We therefore add to the valuation model a proxy for
compensation expenses to explore the market's assessment of this
value driver and its affect on rm value.
We examine the valuation implications of human capital both for a
broad sample of rms and for subsamples of high-technology rms and
low-technology rms. We conjecture that human assets comprise a
more sizable portion of the intrinsic value of high-technology,
knowledge-intensive industries. We nd that the market appears to
value compensation expenses not as expenses but as if they serve as a
proxy for a human asset that is omitted from the balance sheets as well
as from most valuation models. Compensation expenses are positively
and highly signicantly related to rm value in the full sample
regression as well as in the subsamples. Specically, the results imply
that $1 spent on employees by a high (low)-technology rm increases
its equity value by more than $1.9 ($0.4). What is more, the addition of a
proxy for compensation expense to the valuation model signicantly
increased the regressions' power of explanation by more than 12%. The
ndings also indicate that compensation expenses are valued differently
from other expense components of income (as are R&D expenses).
Based on our ndings, we estimate a proxy for the market value derived
from human capital and nd that the market's assessment of a high
(low)-technology rm's human capital may account for up to
approximately 20% (12%) of its market value. The ndings are consistent

Swartz et al. (2006) assert that intellectual capital is omitted from the specication
of the Ohlson (1995) valuation model employed in value relevance studies. They thus
include an intellectual capital variable in the Ohlson (1995) valuation model, and nd
that it signicantly contributes to the explanation of share prices.

with human capital comprising a more sizable portion of the value of


high-technology rms relative to low-technology rms.
At the second stage of the study we investigate whether investors
outside the exchange assess the human capital differently from
investors on the exchange. This comparison is interesting because, in
transactions outside the exchange, investors have better access to inside
information in the course of the due diligence process and the negotiations that precede the transaction. As a result, these buyers in
particular those that are expected to hold controlling interests after the
transaction due to information asymmetry, should face less
information uncertainty relative to the marginal investor on the
exchange, which affects their valuation of the rm. The results indicate
that $1 spent by a high (low)-technology rm on employees increases
the price paid for the rm's equity in a merger or an acquisition by more
than $2.3 ($0.6). Additionally, the buyer's assessment of a high (low)technology rm's unrecorded human capital may account for up to
approximately 23% (14%) of the transaction price. Hence, the valuation
implications of the human asset for buyers outside the exchange and
buyers on the exchange are surprisingly similar. We conclude that
despite the important differences between the two types of buyers,
which prior literature shows have signicant value implications
(controlling interests, information uncertainty, etc.), their assessment
of the enhanced value of a rm attributable to human capital is relatively
alike.
The results in this study are consistent with compensation
expenses creating a valuable intangible asset, hence suggesting that
reform in the accounting treatment of these expenses is in place.
Specically, accounting standards should allow capitalization of all
human asset-related expenses, and determine acceptable approaches
or techniques for amortization of the contributed intangible asset.
This study contributes to the literature in a number of ways. First, it
systematically investigates the market's assessment of the human asset
on a fairly large sample of rms (3186 observations) and for subsamples of high-technology (1234 rms) and low-technology rms
(1952 rms) for the period 19932006. This differentiation is important
as the value implications of human capital are expected to signicantly
differ between fast-growing, knowledge-based industries and more
traditional industries. The two prior studies investigating this issue are
either small sample, and/or use a homogonous sample of rms from one
discipline and a certain component of employee expenses. Swartz et al.'s
(2006) study is based on 154 rm-year observations (for 19972004)
of South African rms, not differentiating between high- and lowtechnology. Bell et al. (2002) use a sample of 85 U.S. computer software
companies that reported positive earnings in 19961997, and focus on
the employee stock option component of executive compensation. We
include in our sample, prot as well as loss rms, and capture in our
estimations differences in the valuation of prots and losses. Additionally, we include in our analysis all types of compensation expense.
Second, our study is the rst to conduct a comparison of the
valuation implications of human asset-related expenses for investors
buying stocks on the exchange and those buying stocks outside
the exchange (in the course of mergers and acquisitions). Third, our
measure of compensation expenses is accessible to all types of investors
and less noisy than that used in prior studies.
The remainder of the paper is organized as follows: Section 2 outlines the literature on human capital valuation; Section 3 develops the
research design; Section 4 describes the sample and data; Section 5
presents the results and Section 6 summarizes the study.
2. Human capital valuation
Managing the human asset effectively is pivotal in high-technology as well as in more traditional industries for creating value and
achieving market power (e.g., Usoff, 2002).4 Kallunki, Karjalainen, and
4

See GAO (2003) for the public sector.

I. Gavious, M. Russ / Advances in Accounting, incorporating Advances in International Accounting 25 (2009) 165173

Martikainen (2005), in their study of six nations, found the average


value of human capital as a percent of market value to be 25.5% with
the lowest in the US (22.4%) and the highest in Switzerland (28.6%).
Drucker (1993; p. 54) states that Knowledge has become the key
economic resource and the dominant and perhaps even the only
source of competitive advantage. Mobilizing, managing and measuring the added value of the human talent might be the only way
companies can succeed in the 21st century (Bryan & Joyce, 2007).
Human capital can be dened as the knowledge that employees
take with them when they leave the rm (Meritum, 2002, p. 10) and
is functioning as the prot lever of the knowledge economy (Bontis
& Ftiz-enz, 2002, p. 224). Swartz et al. (2006) review different
denitions for intellectual capital, including, inter alia, packaged
useful knowledge (Stewart, 1997; p. 67) and combined intangible
assets which enable a company to function (Brookings, 1996; p. 12).
They choose to dene intellectual capital as the enhanced value
of a rm attributable to assets, generally of an intangible nature,
resulting from the companies organizational function, processes and
information technology networks, the competency and efciency of
its employees and its relationship with its customers (Swartz et al.,
2006; p. 72).
There are number of alternative methods to measure and to valuate
human capital. The economic (e.g. Le, Gibson, & Oxley, 2003), nancial
(e.g. Kallunki et al., 2005) and accounting (e.g. Amir & Livne, 2001)
literature suggests the classication of costs (acquisition or replacement costs), market, and income methods to human capital valuation
(e.g. Jarboe, 2007). The acquisition cost (or investment) method seems
to be the easiest to document by using objective disclosure (Chen &
Lin, 2004), if and when the appropriate management control systems
are in place, although the depreciation rates that might need to be
used, might not be straight forward. This approach is seen though, as
least preferred since it is backward oriented (Amir & Lev, 1996). The
replacement (or reproduction) cost method might compensate for
that weakness as it is more present (or future) oriented (Jarboe, 2007).
The market method seems to be the most straightforward, when and
where the appropriate and robust markets for human capital exist
(Karjalainen, 2007), but unfortunately, might not be applicable in most
cases since the markets for human capital (external or internal) are not
effective (Jarboe, 2007). The income (or value) method looks at past
streams of income and uses assumptions to forecast future streams of
income to quantify the value of the human assets (Jarboe, 2007). The
need to forecast the future (argued by some to simply be guessing the
future) is considered the major weakness of this approach.
The academic literature also identies two basic approaches to
quantify the value of human capital; topdown company wide
(e.g., Otter, 2007) and bottomup using individual employees (e.g.
Lev & Schwartz, 1971; Milost, 2007). The company wide approach
views the whole organization as a unit of analysis and seems to take
the market perspective (see, for example, Ballester, Livnat, & Sinha,
2002; Kallunki et al., 2005). The individual employee based approach
views the organization as an aggregate of individuals and seems to
allow, on top of the external perspectives, an internal point of view,
one that might enable an internal evaluation and/or benchmarking of
the organization (e.g., Chen & Lin, 2004; Milost, 2007).
Yet, there is no one widely accepted denition of human capital,
and there is little agreement about the validity and accuracy of the
value indicators for human capital (e.g. Bontis, 2001). For example,
some economist and management researchers (e.g. Becker, 1964;
Hitt, Bierman, Shimizu, & Kochhar, 2001) seem to prefer education
(e.g., number of years) as a measure of human capital. Others, seem to
prefer adjusted prot per employee (PeopleSoft, 2002), return on
skills capital employed (Sheedy-Gohil, 1996), or return on investment
of human capital (Fitz-enz, 2000), among many other measurement
approaches (e.g., Bontis, 2001).
Nevertheless, compensation or cost of employees seems to be an
acceptable indicator (see, for example, Otter, 2007 and Rosett & Shin,

167

2004) for human capital. However, one has to be careful in assuming


that this indicator captures the richness and complexity of the concept
of human capital (e.g., Boudreau & Ramstad, 2007; Bontis & Ftiz-enz,
2002).
This study explores valuation implications of human capital. Other
aspects of human capital, such as the social (or relational, e.g. Jarboe,
2007) capital, or the structural (or organizational, e.g., Stiles &
Kulvisaechana, 2003) capital which are as important (e.g. Meritum
Project, 2002), and the interaction and synergy between the different aspects referred to as organizational capital by Lev and
Radhakrishnan (2005) are beyond the scope of this research.
3. Research design
We investigate the market's perception of the economic effect
of human capital value using a model specication employed in accounting literature, that implicitly or explicitly follows the Ohlson
(1995,1999) and Feltham and Ohlson (1999) valuation models. The
model regresses market value of equity on book value of equity,
abnormal earnings and other information. Other information modies
the prediction of future protability.5
The model that we employ is:
MV = 0 + 1 BV + 2 AbE + 3 NegAbE + 4 R&D + 5 CapEx
+ 7 SalesGr + 8 Div + v

MV is market value of equity; BV is book value of equity; AbE is


abnormal earnings calculated as Et rBVt 1. E is net income before
extraordinary items. The expected rate of return on book value of
common equity, r, is determined based on two different alternatives
used in prior studies: (1) we set r at 12%, which is considered the longterm return on equities (e.g., Dechow, Hutton, & Sloan, 1999; Barth,
Beaver, Hand, & Landsman, 1999; Bell et al., 2002)6; (2) we use the
priceearnings ratio as a surrogate for the cost of capital (e.g., Fama &
French, 19927; Cheng & McNamara, 2000; Swartz et al., 2006), as this
ratio captures the risk and growth of companies. NegAbE is AbE if AbE 0,
0 otherwise. R&D is R&D expenditures; CapEx is capital expenditures;
SalesGr is annual change in sales; and Div is cash dividends.
Consistent with prior literature we restrict our sample to rms
with positive book value of equity. Beginning-of-year owners' equity
must also be positive to ensure that rms' cost of capital in calculating
abnormal earnings (rBVt 1) is positive (see also Bell et al., 2002). We
also allow for different slope coefcients on prots and losses to
account for differences in the valuation of prots and losses (e.g.,
Hayn, 1995; Basu, 1997; Collins, Maydew, & Weiss, 1997; Core et al.,
2003). Bell et al. (2002) indicate that abnormal earnings have positive
and negative coefcients, respectively, when valuation models are
estimated separately for rms with positive and negative earnings.
Thus, different valuation implications apply, not only to total prots/
losses but also to abnormal prots/losses.
R&D, CapEx, SalesGr, and Div serve as proxies for expected growth
in earnings (see, e.g., Core et al., 2003; Morck, Shleifer, & Vishny, 1988;
Demers & Lev, 2001; Swartz et al., 2006). R&D captures the expected
growth in future earnings due to investments in intangible assets.
CapEx captures the expected growth in future earnings due to
investments in tangible assets. We include sales growth and cash
dividends as additional proxies for growth in earnings (see, e.g., Core
5
Ohlson (2001) stresses the crucial role of the other information in capturing
expectations for future earnings or a rm's goodwill.
6
Bell et al. (2002) explain that As Barth et al. (1999, 208) observe, including BVt1
in the abnormal earnings equation partially relaxes the assumption that the cost of
capital associated with calculating abnormal earnings is a predetermined rate, because
current and lagged BV are highly correlated.
7
According to Fama and French (1992), cash ow to price ratio may also serve as a
surrogate for the discount rate. When we apply cash ow to price ratio we obtain the
same qualitative results (untabulated) as those obtained from the earningsprice ratio.

168

I. Gavious, M. Russ / Advances in Accounting, incorporating Advances in International Accounting 25 (2009) 165173

et al., 2003; Hand & Landsman, 1998; Swartz et al., 2006). We set
proxies for expected growth to zero when their values are missing
(e.g., Core et al., 2003; Morck et al., 1988).8
Markedly, extant value relevance research neglects to include
another prominent value driver human capital which has been
shown in the literature to have an increasing importance to a rm's
overall value. We thus add to the regression a proxy for human assetrelated expenditures, CompPr:
MV = 0 + 1 BV + 2 AbE + 3 NegAbE + 4 R&D + 5 CapEx
+ 7 SalesGr + 8 Div + 9 CompPr +

CompPr includes all types of compensation expense wages, salaries,


bonuses, stock options, employee insurance and other employee
benets. Because we do not have access to the exact expensed amounts
related to human capital, we follow the approach employed by Bell et al.
(2002) and use selling, general and administrative expense (SG&A) as a
proxy for compensation expense. SG&A excludes salaries embedded in
the Cost of Goods Sold, and on the other hand, it may include other costs
which reect unrecorded intangible assets advertising expenses and
R&D expense. In addition, SG&A includes other noncompensation
expenses (e.g., accounting expense, bad debt expense, delivery expense,
commissions, etc.). Because advertising and R&D expenses would bias
toward nding a positive valuation effect for SG&A (Bell et al., 2002), we
deduct these items from SG&A. This leaves us with a proxy that should be
negatively biased in terms of its valuation effect, due to exclusion of
salaries embedded in the Cost of Goods Sold and inclusion of current
expenses which decrease the rm's earnings and are not regarded as
investments related to innovative activities. Thus, we conjecture that the
coefcient on CompPr would be downward biased, though the exact
extent of this bias cannot be estimated.
Note that =9CompPr +; i.e., v, the residual from regression (1),
reects information about expensed amounts related to human capital
as well as random error. A signicantly positive 9 would indicate that
compensation expense has a positive incremental association with
equity value, and that this association is actually stronger and higher
than that reected by the downward biased 9. Concomitantly, a signicant increase in adjusted R2 resulting from addition of compensation
expense to the regression would imply that the part in the rm's value
that is not explained by all other value drivers stems primarily from the
human asset. This part is proxied by the residual from regression (1), v.
We will refer to this as market value derived from human capital
(MVDHC).
We examine the valuation implications of human capital both for a
broad sample of rms and for subsamples of high-technology rms and
low-technology rms. We conjecture that the ratio MVDHC/MV is higher
(lower) for high (low)-technology companies consistent with the
expectation that human capital comprises a more (less) sizable portion
of the intrinsic value of a high (low)-technology company. Concurrently,
consistent with the expectation that the role of investments in intangible assets in general, and in the human asset in particular, differs
across high-technology and low-technology industries, we predict
that the 9 coefcient from regression (2) differs across our two
subsamples. Specically, we expect a larger (smaller) 9 for the high
(low)-technology subsample.
Consistent with prior studies, we identify high-technology rms
by the likelihood that rms in the industry have signicant
unrecorded intangible assets (Core et al., 2003; see also Francis &
Schipper, 1999). High-technology rms, as per this denition, are
from industries such as computer hardware and software, pharmaceuticals and biotechnology, electronic equipment, telecommunications, etc.

8
Inferences remain the same when observations with missing values are extracted
from the analysis.

All analyses are repeated using transaction values as the dependent


variable, where transaction values are taken as the amount paid by
a buyer for the rm's equity in a merger or acquisition outside the
exchange. To enable a direct comparison of investors' assessment of
the human asset on the exchange and outside the exchange, the same
rm-years are used in the market value analysis as well as in the
transaction value analysis; i.e., the market value analysis is based on
observations for which we had a transaction value, and the market
value is taken based on the target stock price one week prior to the
original announcement of the transaction.9
In each regression we control for industry and year xed effects by
including dummies for industry (2 digit SIC) and year. Estimates for all
equations are in per-share form with intercepts suppressed. Included
in the scaled models a regressor: [1/number of shares outstanding
at scal year-end], to estimate the intercept for the corresponding
unscaled models.
4. Data
We obtain a sample of non-nancial rms10 from the Thomson
Financial SDC database of mergers and acquisitions. The sample
includes U.S. public-rm targets acquired by other U.S. public rms
over the time period of 19932006. The SDC database enables us to
conduct our analyses using market values as well as transaction
values. In the database, the buyer generally holds controlling interest
after the transaction. We thus eliminate rms in which the buyer does
not hold controlling interest after the transaction for the sake of
sample homogeneity. Our focus on controlling interests allows us to
make a distinction between minority-interest investors on the
exchange and majority-interest investors outside the exchange.
While the SDC database includes select nancial statement data, it
does not contain all the data we need for our tests. We thus match the
SDC data with Compustat11 and obtain all nancial statement data
from Compustat. Firms with insufcient Compustat data are excluded
from the analysis.
To mitigate the effect of outliers, we trim 0.5% of the data based
on the distribution of the dependent variable, by subsample (see also,
e.g., Collins et al., 1997; Brown et al., 1999; Core et al., 2003; DeFranco,
Gavious, Jin, & Richardson, 2009). As explained in Section 3, we
eliminate rms with negative end-of-year as well as beginning-of-year
owner's equity. The deletion procedure results in a nal full sample of
3186 observations; 1234 and 1952 observations in our high-technology
and low-technology subsamples, respectively. To deal with outliers in
the independent variables, we winsorize the independent variables at
the 1st and 99th percentiles (by subsample).
Table 1's descriptive statistics reveal that high-technology and
low-technology rms in our sample are, on average, very close in size.
This is an important attribute of our sample, as it implies that ndings
presented for the two subsamples do not derive from size differences.
The market value of equity as well as transaction values far exceed the
book value of equity, and more so in high-technology than in lowtechnology rms, consistent with practitioner and academic studies
indicating that GAAP-based values lose their relevance over time, in
particular with respect to technology based industries. Abnormal
earnings are, on average, negative, consistent with prior research (e.g.,
Barth et al., 1999; Barth, Beaver, Hand, & Landsman, 2000).12 Notably,
9
We used closing common stock price of the target for additional points in time
prior to the original announcement of the transaction and obtained the same
qualitative results.
10
Consistent with prior studies (see for example Core et al., 2003), we exclude
nancial rms from our analyses.
11
Matching the SDC data with Compustat is done using rms' ticker symbol, name
and scal year.
12
In the study of Bell et al. (2002) abnormal earnings are, on average, positive in
contrast to prior research. Bell et al. explain that this is likely attributable to restricting
their sample to rms with positive earnings and to rms from a single industry
computer software that is characterized by high growth.

I. Gavious, M. Russ / Advances in Accounting, incorporating Advances in International Accounting 25 (2009) 165173

169

Table 1
Descriptive statistics.
Pooled sample

TV
MV
BV
TV/BV
MV/BV
AbE/BV

High-technology

Low-technology

Mean

Median

S.D.

Mean

Median

S.D.

Mean

Median

S.D.

1146
773
298
7.702
5.289
0.513

184
123
62
2.870
1.833
0.066

5123
3588
1222
88.356
67.733
7.217

1148
804
235
11.635
8.829
0.964

168
116
47
3.530
2.334
0.133

4953
3713
1009
135.467
106.814
11.319

1145
753
338
5.215
3.052
0.228

199
126
74
2.580
1.641
0.043

5230
3507
1338
33.629
16.311
1.965

This table reports the descriptive statistics for our sample of 3186 (1234 high-technology and 1952 low-technology) rms. Variable denitions: TV is transaction value measured as
the amount paid by a buyer for the target rm's equity in a transaction outside the exchange; MV is market value of common shares outstanding measured based on target stock price
one week prior to the original announcement of the transaction; BV is book value of common equity as of scal year-end; AbE is abnormal earnings measured as net income before
extraordinary items and discontinued operations, minus 0.12 BVt 1.

abnormal earnings are more negative in high-technology rms than


in low-technology rms, consistent with the former being characterized by large investments in intangible assets (e.g., R&D and human
capital) which are immediately expensed in accordance with GAAP,
hence depressing current earnings.
Table 2 presents the Pearson correlations amongst the variables
used in the analyses.13 The table reveals that most of the variables are
correlated with each other in the pooled sample and in the
subsamples. Notably, transaction values and market values are highly
correlated with reported equity book value, abnormal earnings and
our proxy for compensation expense. A positive relation between rm
value and compensation expense implies that investors view the
latter as an investment i.e., an unrecorded asset rather than as an
expense. Moreover, the fact that transaction values as well as market
values are strongly positively correlated with our proxy for compensation expense strengthens our condence that the latter serves as a
good estimate of investments in human assets. It seems that the effect
of other expenses included in our proxy for compensation expense,
which is typically negatively correlated with rm value, is offset by
the effect of compensation expenses, which are positively correlated
with rm value. Hence, the positive relation between rm value and
compensation expenses may be even stronger than that presented in
Table 2.
Another important nding is that, in high (low)-technology rms,
which are characterized with higher (lower) intensity of expenditures
on R&D and human capital,14 R&D and compensation expenses are (in)
signicantly negatively correlated with abnormal earnings. The negative correlation is due to GAAP mandating the immediate expensing of
these costs; the higher the amounts invested in these intangibles, the
lower reported earnings. Despite this technical negative relation, one
can see that the equity book value is positively correlated with R&D and
compensation expenses. This may be explained by the indirect effect
of capital expenditures; the relation between investments in tangible
assets (capital expenditures) and investments in intangible assets (R&D
and compensation expenses) is signicantly positive both in hightechnology and in low-technology rms, indicating that the more a rm
invests in tangibles the more it invests in intangibles (note that this
relation is naturally stronger for high-technology rms which rely to
a larger extent on intangible assets). Thus, as investments in tangible
assets increase the book value of the rm (in contrast to treatment of
R&D and human-related compensations, expenditures on tangibles
are capitalized into assets in accordance with GAAP), investments in
intangibles are also ostensibly positively related with equity book value.

13

Spearman's nonparametric correlations between the variables (untabulated) are


qualitatively similar to Pearson's.
14
Untabulated results show that the intensity of R&D and compensation expenses is
higher (lower) in our high (low)-technology subsample.

5. Results
Tables 3 and 4 present the results from estimating valuation
Eqs. (1) and (2), respectively, with expected rate of return on book
value of equity, r, set at 12%. We perform the sensitivity analyses for
our results with different estimates of r; rst, we employ different
rates of return in the range of 1014% and second, we employ r
derived from the rms' PE ratio. The results obtained (untabulated)
are qualitatively similar to those presented in the tables, indicating
that our ndings are robust to the approach applied for estimating r.
Table 3 reveals that the signs of the estimated coefcients in the
full sample regression are consistent with prior research and are statistically signicant. The coefcient on book value of equity (1.357,
p: 0.000) signicantly exceeds 1. The Ohlson (1995) model predicts that
the book value of equity coefcient should equal 1, however the model
does not account for accounting conservatism which results in
systematic understatement of equity book value and, hence, systematic
overstatement of its valuation coefcient (Bell et al., 2002; p. 984). This
overstatement is even more pronounced in high-technology industries
than in low-technology. The coefcient on abnormal earnings in prot
years is signicantly positive (8.203, p = 0.000) and the coefcient on
abnormal earnings in loss years is insignicantly negative (sum of
coefcients on AbE and NegAbE, 0.154, p = 0.367).15 This nding
suggests that after controlling for other explanatory variables, investors
expect that a current ability to generate positive abnormal earnings
precedes higher future cash ows. On the other hand, negative abnormal earnings are value-irrelevant; in this case, investors form their
expectations for the future based on alternative proxies for future
earnings (or cash ow) growth. When we run the regression for rms
with negative abnormal earnings only, we observe signicant changes
in two coefcients on proxies for future earnings growth; the coefcient
on sales growth as well as its signicance increase (0.676, p = 0.000)
whereas the coefcient on dividends loses its signicance (0.791,
p = 0.300). This result implies that the ability to generate higher sales
becomes more crucial in the market's assessment of a rm's future
performance when the rm fails to generate current abnormal earnings.
On the other hand, dividends are irrelevant as such rms generally avoid
paying dividends.
For the high-technology and low-technology subsamples, we nd
the coefcients are all of the same sign as those in the full sample.
Additionally, all coefcients are statistically signicant with the exception of sales growth in low-technology rms. A possible explanation for
the difference in the value relevance of sales for high-technology and

15
Prior studies that use reported earnings rather than abnormal earnings in their
specications document a negative coefcient on earnings in loss years (see, for
example, Core et al., 2003). When we run our regressions substituting abnormal with
reported earnings, we obtain a negative coefcient on negative earnings, consistent
with prior research.

170

I. Gavious, M. Russ / Advances in Accounting, incorporating Advances in International Accounting 25 (2009) 165173

Table 2
Pearson correlations among variables: transaction value, market value, abnormal earnings, R&D, capital expenditures, sales growth, dividends, and proxy for compensation expenses.
Variable

TV

MV

BV

AbE

R&D

CapEx

1.000
0.522
0.479
0.158
0.227
0.145
0.280
0.347

1.000
0.105
0.005
0.279
0.168
0.391
0.528

1.000
0.078
0.000
0.165
0.077
0.011

1.000
0.156
0.060
0.028
0.071

1.000
0.119
0.153
0.222

Panel B: high-technology (n = 1234)


TV
1.000
MV
0.921
1.000
BV
0.485
0.463
AbE
0.432
0.458
R&D
0.166
0.170
CapEx
0.268
0.220
SalesGr
0.209
0.174
Div
0.233
0.255
CompPr
0.423
0.383

1.000
0.139
0.128
0.316
0.098
0.228
0.671

1.000
0.097
0.087
0.091
0.156
0.065

1.000
0.217
0. 109
0.031
0.075

Panel C: low-technology (n = 1952)


TV
1.000
MV
0.861
1.000
BV
0.603
0.587
AbE
0.444
0.491
R&D
0.164
0.178
CapEx
0.310
0.239
SalesGr
0.149
0.142
Div
0.361
0.316
CompPr
0.428
0.364

1.000
0.136
0.058
0.250
0.151
0.396
0.467

1.000
0.005
0.020
0.180
0.040
0.033

1.000
0.095
0.045
0.019
0.084

Panel A: pooled sample (n = 3186)


TV
1.000
MV
0.882
BV
0.559
AbE
0.446
R&D
0.137
CapEx
0.300
SalesGr
0.167
Div
0.325
CompPr
0.418

SalesGr

Div

CompPr

1.000
0.046
0.262

1.000
0.148

1.000

1.000
0.091
0.085
0.321

1.000
0.025
0.269

1.000
0.208

1.000

1.000
0.111
0.150
0.184

1.000
0.032
0.170

1.000
0.094

1.000

TV, MV, BV, and AbE are as dened in Table 1. R&D is abnormal earnings measured as net income before extraordinary items and discontinued operations, minus 0.12 BVt 1; R&D
is research and development expenditures; CapEx is capital expenditures; SalesGr is annual change in sales; Div is cash dividends; and CompPr is a proxy for compensation expenses
estimated as SG&A expenses, minus advertizing and R&D expenses (if included in SG&A). , , and indicate signicance levels of 10%, 5%, and 1%, respectively.

low-technology rms relates to higher volatility and market risks that


characterize the former. Due to these risks, high-technology rms end
up reporting losses more frequently than do low-technology rms
(52.2% compared with 27.9%, respectively). Higher intensity of investments in intangible assets, which are immediately expensed in
accordance with GAAP requirements, further contributes to depression
of earnings reported by high-technology rms. As shown above, the
ability to generate revenues is perceived as a more crucial indicator for
Table 3
Market value regressions excluding a proxy for compensation expense.

Constant
BV
AbE
NegAbE
R&D
CapEx
SalesGr
Div
Adj. R2
# of obs.

All rms

High-technology rms

Low-technology rms

2.401
(0.000)
1.357
(0.000)
8.203
(0.000)
8.357
(0.000)
4.724
(0.000)
0.519
(0.000)
0.296
(0.021)
1.379
(0.031)
0.745
3186

5.075
(0.005)
1.718
(0.000)
10.078
(0.000)
10.143
(0.000)
2.948
(0.004)
0.887
(0.000)
2.052
(0.000)
6.852
(0.001)
0.729
1234

1.902
(0.000)
1.296
(0.000)
7.692
(0.000)
7.097
(0.000)
2.503
(0.023)
0.464
(0.000)
0.106
(0.384)
1.555
(0.009)
0.755
1952

See Tables 1 and 2 for denitions of all variables. NegAbE is AbE if AbE 0, 0 otherwise.
Estimates for all equations are in per-share form with intercepts suppressed. The
regressor Constant is the inverse of number of shares outstanding at scal year-end
(included in the scaled model to estimate the intercept for the corresponding unscaled
model). P-values are in parentheses.

future performance and survivorship when the rm reports negative


abnormal earnings, hence sales growth is more (less or not) relevant for
high (low)-technology stock pricing. The magnitudes of coefcients on
the other explanatory variables also differ signicantly across the two
subsamples (larger for the high-technology subsample consistent with
prior research), indicating that the two groups of rms have different
valuation characteristics.
In Table 4, the results from estimating regression model (2) are
displayed. We focus on the coefcient on our proxy for compensation
expenses (9). Inferences from coefcients on all other explanatory
variables are consistent with those obtained from regression model (1).
The results show that 9 is positive and highly signicant in the full
sample regression as well as in the subsamples. This nding indicates
that compensation expenses are valued differently from other expense
components of income (as are R&D expenses). As expected, 9 is
signicantly larger (smaller) for high (low)-technology rms. In the full
sample, 9 is approximately 1 while in high (low)-technology rms it is
approximately 1.9 (0.4). As CompPr is a noisy measure of compensation
expenses that leads to a downward biased 9 with respect to the actual
valuation effect of compensation expenses, the incremental association
of the actual amounts that a high (low)-technology rm invests in its
human capital should be larger than 1.9 (0.4). I.e., $1 spent by a high
(low)-technology rm on employees increases its equity value by more
than $1.9 ($0.4). What is more, the addition of a proxy for compensation
expense to the valuation model signicantly increased the regressions'
power of explanation by more than 12%; adjusted R2 increased from
72.9% (75.5%) to 85.8% (87.7%) in the high (low)-technology subsample.
Bell et al. (2002), for a sample of 85 protable computer software
companies, nd that while the market appears to value employee
stock options expense not as an expense but as an intangible asset, it
appears to negatively value SG&A expense. Notably, Bell et al. do not
exclude either R&D or advertising expenses from SG&A. Hence, their

I. Gavious, M. Russ / Advances in Accounting, incorporating Advances in International Accounting 25 (2009) 165173
Table 4
Market value regressions including a proxy for compensation expense.

Constant
BV
AbE
NegAbE
R&D
CapEx
SalesGr
Div
CompPr
Adj. R2
# of obs.

All rms

High-technology rms

Low-technology rms

2.163
(0.000)
1.425
(0.000)
8.460
(0.000)
8.324
(0.000)
4.681
(0.000)
0.518
(0.000)
0.394
(0.003)
1.109
(0.085)
0.988
(0.000)
0.866
3186

4.284
(0.026)
1.790
(0.000)
10.289
(0.000)
10.386
(0.000)
2.470
(0.004)
0.993
(0.000)
2.026
(0.000)
7.065
(0.001)
1.884
(0.001)
0.858
1234

1.814
(0.000)
1.322
(0.000)
7.625
(0.000)
7.028
(0.000)
3.006
(0.021)
0.440
(0.000)
0.150
(0.234)
1.401
(0.021)
0.395
(0.001)
0.877
1952

See Tables 1 and 2 for denitions of all variables. Estimates for all equations are in per-share
form with intercepts suppressed. The regressor Constant is the inverse of number of
shares outstanding at scal year-end (included in the scaled model to estimate the
intercept for the corresponding unscaled model). P-values are in parentheses.

SG&A proxy for compensation expenses is much noisier than our


CompPr. Additionally, sample rms in Bell et al.'s study are from a
single industry computer software and are limited to protable
rms. Thus, the generalizability of their inferences could be limited.
Our ndings, based on a measure for compensation expenses that
downward biases the actual valuation effect of a rm's investments
in human assets, lead us to conjecture that the term 9CompPr + it
primarily reects equity value derived from human assets. This term
is the part in a rm's value that is not explained by all the other value
drivers included in regression (1), i.e., the market value derived from
human capital (MVDHC). In the regressions presented in Table 3, the
estimated proxy for MVDHC in high-technology rms is, on average,
$3.168 per-share. The average share price is $15.529, and thus the
ratio MVDHC/MV is 3.168/15.529 = 0.204, implying that the market's
assessment of a high-technology rm's unrecorded human assets may
account for up to approximately 20% of its market value.16 For lowtechnology rms, the ratio MVDHC/MV is 2.070/17.395 = 0.119. The
same qualitative results are obtained using medians instead of
averages. The ndings are consistent with human capital comprising
a more sizable portion of the value of high-technology rms relative
to low-technology rms.
The results thus far imply that the market appears to value
compensation expenses as if they serve as a proxy for an intangible
asset that is omitted from the balance sheets as well as from most
valuation models, such as that specied in Eq. (1). We now turn to
investigate whether investors outside the exchange assess the human
asset differently from investors on the exchange. To analyze the
assessment of human capital by investors outside the exchange and
their perception of the economic effect of compensation expenses on
rm value, we use transaction values (TV) as the dependent variable in
our valuation models. Specically, we replace market values with the
sale price of the rms' equity in a merger or acquisition that has taken
place outside the exchange. The results from estimating valuation
Eqs. (1) and (2) with transaction values are presented in Tables 5 and 6,
respectively. Inferences from regressions with transaction values are
consistent with those obtained for market values; the coefcients are
16
We calculate ratio of averages rather than average of ratios to avoid loss of
observations (MVDHC/MV cannot be calculated when MVDHC the regression
residual takes on a negative value). Notwithstanding, inferences are generally
similar when applying average of ratios.

171

Table 5
Transaction value regressions excluding a proxy for compensation expense.

Constant
BV
AbE
NegAbE
R&D
CapEx
SalesGr
Div
Adj. R2
# of obs.

All rms

High-technology rms

Low-technology rms

1.584
(0.317)
2.044
(0.000)
12.870
(0.000)
13.095
(0.000)
5.748
(0.000)
1.421
(0.000)
0.733
(0.000)
4.401
(0.000)
0.785
3,186

0.221
(0.933)
2.473
(0.000)
13.977
(0.000)
14.238
(0.000)
3.573
(0.005)
1.667
(0.000)
3.861
(0.000)
8.164
(0.008)
0.749
1,234

3.004
(0.122)
1.970
(0.000)
12.505
(0.000)
12.345
(0.000)
2.175
(0.296)
1.343
(0.000)
0.323
(0.099)
4.979
(0.000)
0.801
1,952

See Tables 1 and 2 for denitions of all variables. Estimates for all equations are in per-share
form with intercepts suppressed. The regressor Constant is the inverse of number of
shares outstanding at scal year-end (included in the scaled model to estimate the
intercept for the corresponding unscaled model). P-values are in parentheses.

with the same sign, however larger, possibly because the sale price
includes in it a control premium that is not priced on the exchange. One
exception is the coefcient on R&D in low-technology rms, which is
smaller (larger) in the regression of transaction (market) values.
Table 6 shows that 9 is positive and highly signicant in the full
sample regression as well as in the subsamples. Again, 9 is signicantly
larger (smaller) for high (low)-technology rms. In the full sample, 9 is
1.315 compared with 0.988 in the market value regression. For high
(low)-technology rms, 9 is 2.309 (0.605) compared with 1.884
(0.395) in the market value regression. The results are robust to cost of
capital estimating approach. Thus, each dollar that a rm spends on its
employees whether a high-technology or low-technology is worth
more in the eyes of a buyer outside the exchange because this buyer is
about to control the rm and thus will directly be affected by the quality
and the interests of its employees. Indeed, the transaction value regressions have a higher power of explanation than the market
regressions. The addition of a proxy for compensation expense to the
Table 6
Transaction value regressions including a proxy for compensation expense.

Constant
BV
AbE
NegAbE
R&D
CapEx
SalesGr
Div
CompPr
Adj. R2
# of obs.

All rms

High-technology rms

Low-technology rms

1.080
(0.510)
2.046
(0.000)
12.676
(0.000)
12.863
(0.000)
5.686
(0.000)
1.392
(0.000)
0.741
(0.000)
4.284
(0.000)
1.315
(0.005)
0.916
3186

0.994
(0.723)
2.538
(0.000)
14.233
(0.000)
14.542
(0.000)
3.522
(0.006)
1.828
(0.000)
3.829
(0.000)
8.366
(0.006)
2.309
(0.002)
0.886
1234

2.625
(0.200)
1.902
(0.000)
12.013
(0.000)
11.708
(0.000)
1.635
(0.434)
1.279
(0.000)
0.211
(0.297)
5.153
(0.000)
0.605
(0.003)
0.937
1952

See Tables 1 and 2 for denitions of all variables. Estimates for all equations are in per-share
form with intercepts suppressed. The regressor Constant is the inverse of number of
shares outstanding at scal year-end (included in the scaled model to estimate the
intercept for the corresponding unscaled model). P-values are in parentheses.

172

I. Gavious, M. Russ / Advances in Accounting, incorporating Advances in International Accounting 25 (2009) 165173

valuation model signicantly increased the power of explanation of the


transaction value regressions by more than 13% (adjusted R2 increased
from 74.9% (80.1%) to 88.6% (93.7%) for the high (low)-technology
subsample).
The estimated proxy for transaction value derived from human
capital (TVDHC) is, on average, $5.523 per-share for high-technology
rms. The average transaction share price is $24.015, and thus the
ratio TVDHC/TV is 5.523/24.015 = 0.230, implying that approximately
23% of the price paid for a high-technology rm in a merger or an
acquisition derives from the buyer's assessment of its human assets.
For low-technology rms, the ratio TVDHC/TV is 4.131/29.295 0.141.
The same qualitative results are obtained using medians instead of
averages. Again, the ndings are consistent with human assets
comprising a more sizable portion of the value of high-technology
rms relative to low-technology rms. When comparing the valuation
of human capital on- and outside the exchange, the average share
price derived from human capital in an M&A transaction is about 75%
(100%) higher than the average market price derived from human
capital in a high (low)- technology rm. However, when taken as a
portion of total share price, this portion albeit higher in transactions
outside the exchange is quite close to that in the market values.
This nding implies that, although the two types of buyers differ in
imperative aspects, such as controlling interests and access to inside
information, their assessment of the enhanced value of a rm attributable to human capital is quite similar.
6. Summary and conclusions
This paper examines the valuation implications of human capital
both for a broad sample of rms and for subsamples of high-technology
rms and low-technology rms. We nd that, as hypothesized, the
market appears to value compensation expenses as if they serve as a
proxy for an unrecognized intangible asset. Compensation expenses
are positively and highly signicantly related to rm value in the full
sample regression as well as in the subsamples. Furthermore, the results
support the notion that human capital comprises a more sizable portion
of the value of high-technology rms than of low-technology rms.
When comparing the assessment of human capital by investors on
and outside the exchange, we nd that the two types of buyers assess
the enhanced value of a rm attributable to human capital similarly,
despite the important differences between these investors, which prior
literature shows have signicant value implications (controlling
interests, information asymmetry, etc.).
The evidence presented in the paper is consistent with compensation expenses creating a valuable intangible asset. Markedly, investors
seem to see through the deciencies of GAAP, and treat employee
compensation, that is immediately expensed in income statements,
as a value-increasing investment. This suggests that reform in the
accounting treatment of compensation expenses is in place. Specically,
accounting standards should allow capitalization of all human assetrelated expenses, and determine acceptable approaches or techniques
for amortization of the contributed intangible asset. This might be even
more important for the high-technology, fast-growing, knowledgebased industries where the value implications of human capital are
signicantly higher than in more traditional industries.
The current research attempts to estimate the relationship
between compensation expenses and value created in two different
types of industries and in the context of mergers and acquisitions.
Future research should further explore the exact type of relationship
between compensation and value created. For example, a curvilinear
relation would indicate that more is better up to a certain point (e.g.,
Ester & Kerkhofs, 2007).
Another question of interest is the rate of depreciation on human
capital (e.g. Kallunki et al., 2005), specically, assessing different
rates of depreciation for different environments and industries. For
example, in high (low)-technology rms, the rate of depreciation

should be higher (lower), since life of knowledge (e.g. Russ & Jones,
2008) and product life cycles (e.g. Gardner, Johnson, Lee, & Wilkinson,
2000) in the former are shorter.
Lastly, Schmidt and Dharan (2004) offered a typology of intellectual
assets' exploratory models while recommending moving human capital
valuation studies toward a more rigorous approach to research. Within
their framework, our study would be identied as an association model,
which is the rst step in understanding the phenomenon. We join their
call for a need to move the research of human capital valuation towards
models that are more detailed and descriptive, and as such would
provide more managerial insights in regard to creating and managing
human assets more effectively and efciently.
References
Amir, E., & Lev, B. (1996). Value-relevance of nonnancial information: The wireless
communications industry. Journal of Accounting and Economics, 22, 330.
Amir, E., & Livne, G. (2001). Accounting for human capital when labor mobility is
restricted.UK: London Business School Accounting Subject Area, no. 002.
Armstrong, C., Davila, A., & Foster, G. (2006). Venture-backed private equity valuation
and nancial statement information. Review of Accounting Studies, 11(1), 119154.
Ballester, M., Livnat, J., & Sinha, N. (2002). Labor costs and investments in human
capital. Journal of Accounting, Auditing and Finance, 17(4), 351373.
Barth, M. E., Beaver, W. H., Hand, J. R. M., & Landsman, W. R. (1999). Accruals, cash ows
and equity values. Review of Accounting Studies, 4, 205229.
Barth, M.E., Beaver, W.H., Hand, J.R.M. and Landsman, W.R. (2000). Accrual components,
earnings forecasting, and equity values, working paper, The University of North
Carolina at Chapel Hill, NC, and Stanford University, Stanford, CA.
Basu, S. (1997). The conservatism principle and the asymmetric timeliness of earnings.
Journal of Accounting and Economics, 24, 337.
Becker, G. S. (1964). Human capital. New York: National Bureau of Economic Research.
Bell, T. B., Landsman, W. R., Miller, B. G. L., & Yeh, S. (2002). The valuation implications of
employee stock option accounting for protable computer software rms. Accounting
Review, 77(4), 971996.
Bontis, N. (2001). Assessing knowledge assets: A review of the models used to measure
intellectual capital. International Journal of Management Reviews, 3(1), 4160.
Bontis, N., & Ftiz-enz, J. (2002). Intellectual capital ROI: A causal map of human capital
antecedents and consequents. Journal of Intellectual Capital, 3(3), 223247.
Boudreau, J. W., & Ramstad, P. M. (2007). Beyond HR: The new science of human capital.
Harvard Business School Press, June, 19, 2007.
Brookings, A. (1996). Intellectual capital: Core assets from the third millennium enterprise.
London: Thomson.
Brown, S., Kim, L., & Lys, T. (1999). Use of R2 in accounting research: Measuring changes
in value relevance over the last four decades. Journal of Accounting and Economics,
28, 83115.
Bryan, L. L., & Joyce, C. I. (2007). Mobilizing minds: Creating wealth from talent in the 21stcentury organization. New York: McGraw-Hill.
Callen, J. L., Gavious, I., & Segal, D. (2008). The complementary relationship between
nancial and non-nancial information in the biotechnology industry and the degree of
investor sophistication. Working paper, University of Toronto.
Chen, H. M., & Lin, K. J. (2004). The role of human capital cost in accounting. Journal of
Intellectual Capital, 5(1), 116130.
Cheng, C. S. A., & McNamara, R. (2000). The valuation accuracy of the priceearnings
and pricebook benchmark valuation methods. Review of Quantitative Finance and
Accounting, 15, 349370.
Collins, D. W., Maydew, E. L., & Weiss, I. S. (1997). Changes in the value-relevance of
earnings and book values over the past forty years. Journal of Accounting and Economics,
24, 3967.
Core, J. E., Guay, W. R., & Van Buskirk, A. (2003). Market valuations in the new economy: An
investigation of what has changed. Journal of Accounting and Economics, 34, 4367.
Dechow, P., Hutton, A., & Sloan, R. (1999). An empirical assessment of the residual
income valuation model. Journal of Accounting and Economics, 26, 134.
DeFranco, G., Gavious, I., Jin, J., & Richardson, G. D. (2009). Do private company targets that
hire Big4 auditors receive higher proceeds? Working paper, University of Toronto.
Demers, E., & Lev, B. (2001). A rude awakening: Internet shakeout in 2000. Review of
Accounting Studies, 6, 331359.
Drucker, P. E. (1993). Post-capitalist society. New York: Harper Business.
Edvinsson, L., & Malone, M. S. (1997). Intellectual capital: realizing your company's true
value by nding its hidden brainpower. New York: Harper Business.
Ely, K., Simko, P. J., & Thomas, L. G. (2003). The usefulness of biotechnology rms drug
development status in the evaluation of research and development costs. Journal of
Accounting, Auditing and Finance, 18, 163196.
Ester, P., & Kerkhofs, M. (2007). Human capital and the older worker: The need for solid
indicators. Paper presented at the second OECD world forum on Statistics, knowledge
and policy measuring and fostering the progress of societies Istanbul, Turkey, 2730
June 2007.
Fama, E. F., & French, K. R. (1992). The cross-section of expected stock return. Journal of
Finance, 47, 427465.
Feltham, G., & Ohlson, J. A. (1999). Residual earnings valuation with risk and stochastic
interest rates. Accounting Review, 74, 165184.
Fitz-enz, J. (2000). The ROI of human capital: Measuring the economic value of employee
performance. New York: Amacom.

I. Gavious, M. Russ / Advances in Accounting, incorporating Advances in International Accounting 25 (2009) 165173
Francis, J., & Schipper, K. (1999). Have nancial statements lost their relevance? Journal
of Accounting Research, 37, 320352.
GAO (2003). High-risk series: Strategic human capital management (GAO-03-120).
Available at http://www.gao.gov/pas/2003/d03120.pdf, Downloaded January 18,
2009.
Gardner, D. M., Johnson, F., Lee, M., & Wilkinson, I. (2000). A contingency approach to
marketing high technology products. European Journal of Marketing, 34(9/10),
10531077.
Gavious, I., & Schwartz, D. (2008). Market valuations of start-up ventures around the
technology bubble. Working paper, Ben-Gurion University.
Guo, R., Lev, B., & Zhou, N. (2005). The valuation of biotech IPOs. Journal of Accounting,
Auditing & Finance, 20(4), 423459.
Hand, J. R. M. (2004). The market valuation of biotechnology rms and biotechnology
R and D. In J. McCahery, & L. Renneboog (Eds.), Venture capital contracting and
the valuation of high-technology rms Oxford, UK: Oxford University Press.
Hand, J. R. M. (2005). The value relevance of nancial statements in the venture capital
market. Accounting Review, 80, 613648.
Hand, J., & Landsman, W. (1998). Testing the Ohlson model: V or not V, that is the question.
Working Paper Chapel Hill, NC: University of North Carolina and Chapel Hill.
Hayn, C. (1995). The information content of losses. Journal of Accounting and Economics,
20, 125153.
Hitt, M. A., Bierman, L., Shimizu, K., & Kochhar, R. (2001). Direct and moderating effect
of human capital on strategy and performance in profesional service rms: A
resource-based perspective. Academy of Management Journal, 44(1), 1328.
Jarboe, K.P. (2007). Measuring intangibles: A summary of recent activity, presented to the
Innovation vital signs workshop April 2627, 2007, Athena Alliance, working paper
no. 02. Available at http://www.usinnovation.org/pdf/ASTRA_MeasuringIntangibles807.
pdf, downloaded January 15, 2009.
Jorion, P., & Talmor, E. (2001). Value relevance of nancial and nonnancial information
in emerging industries: The changing role of web trafc data. London Business
School Accounting Subject Area, No. 021.
Kallunki, J. -P., Karjalainen, P., & Martikainen, M. (2005). Investments in human capital in
different institutional environments. Advances in International Accounting, 18, 121140.
Karjalainen, P. (2007). Valuation of intangible assets in different nancial environments.
Academic Dissertation, presented to the Faculty of Economics and Business Administration of the University of Oulu, on May 25th, 2007, Available at http://herkules.oulu./
isbn9789514284403/isbn9789514284403.pdf. Downloaded January 22, 2009.
Le, T., Gibson, J., & Oxley, L. (2003). Cost and income-based measures of human capital.
Journal of Economic Surveys, 17(3), 271307.
Lev, B. (2005). Intangible assets: Concepts and measurements, encyclopedia of social
measurement, Vol. 2. (pp. 299305): Elsevier Inc Available at http://pages.stern.nyu.
edu/~blev/docs/Intangible%20Assets%20Concepts%20and%20Measurements.pdf,
Downloaded March 21, 2008.
Lev, B., & Radhakrishnan, S. (2005). The valuation of organization capital. In C. Corrado,
J. Haltiwanger, & D. Sichel (Eds.), Measuring capital in the new economy Chicago:
University of Chicago Press.
Lev, B., & Schwartz, A. (1971). On the use of the economic concept of human capital in
nancial statements. Accounting Review, 46, 103111.

173

Lev, B., & Zarowin, P. (1999). The boundaries of nancial reporting and how to extend
them. Journal of Accounting Research, 37, 353385.
MERITUM Project (2002). Guidelines for managing and reporting on intangibles (intellectual
capital report).MERITUM Project Brussels: European Commission Available at http://
www.pnbukh.com/site/les/pdf_ler/MERITUM_Guidelines.pdf, Downloaded January
16, 2009.
Milost, F. (2007). A dynamic monetary model for evaluating employees. Journal of
Intellectual Capital, 8(1), 124138.
Morck, R., Shleifer, A., & Vishny, R. (1988). Management ownership and market valuation:
An empirical analysis. Journal of Financial Economics, 20, 293315.
Ohlson, J. A. (1995). Earnings, book value and dividends in security valuation. Contemporary Accounting Research, 11, 661687.
Ohlson, J. A. (1999). On transitory earnings. Review of Accounting Studies, 4, 145162.
Ohlson, J. A. (2001). Earnings, book values and dividends in equity valuation: An
empirical perspective. Contemporary Accounting Research, 18(1), 107120.
Otter, J. (2007). Putting the people component of the business entity on the balance sheet.
Available at http://donaldhtaylor.les.wordpress.com/2007/03/puttinghumanassetsonthebalancesheet.pdf, Downloaded January 15, 2009.
PeopleSoft (2002). Maximizing human capital. White Paper Series, January. (pp. ) Available
at http://www.csupomona.edu/~jcbassett/gba562/misc/Human_Capital_White_Paper.
pdf, Downloaded January 18, 2009.
Rajgopal, S., Shevlin, T., & Venkatachalam, M. (2003). Does the stock market fully
appreciate the implications of leading indicators for future earnings? Evidence
from order backlog. Review of Accounting Studies, 8, 461492.
Rosett, J., & Shin, Y. (2004). Does labor disclosure provide useful information? The case of
analysts' forecasts. Working papers and publications in nancial economics; Claremont
McKenna College 2004-02, April, Available at http://www.claremontmckenna.edu/
fei/papers/pdf/2004-02.pdf , Downloaded January 8, 2009.
Russ, M., & Jones, J. K. (2008). Regional economic development indicators for a knowledgebased economy in a knowledge deprived region. Journal of Regional Analysis and Policy,
38(2), 189205.
Schmidt, J. A., & Dharan, B. G. (2004). Intangible assets and the science of people
management. Working Paper Available at http://www.ruf.rice.edu/~bala/les/SchmidtDharan_Intangibles_040504-draft.pdf, downloaded January 8, 2009.
Sheedy-Gohil, K. (1996). Putting the asset value of skills on the balance sheet. Managerial Auditing Journal, 11(7), 1620.
Stewart, T. (1997). Intellectual capital: The wealth of organizations. New York: Doubleday.
Stiles, P., & Kulvisaechana, S. (2003). Human capital and performance: A literature review.
Available at www.accountingforpeople.gov.uk.
Swartz, G. E., Swartz, N. -P., & Firer, S. (2006). An empirical examination of the value
relevance of intellectual capital using the Ohlson (1995) valuation model. Meditari
Accountancy Research, 14(2), 6781.
Trueman, B., Wong, F., & Zhang, X. (2001). Back to basics: Forecasting the revenues of
internet rms. Review of Accounting Studies, 6, 305329.
Usoff, C. (2002). The importance of intellectual capital and its effect on performance
measurement systems. Managerial Auditing Journal, 17(1), 915.
Xu, B., Magnan, M., & Andre, P. (2007). The stock market valuation of R&D information in
biotech rms. Contemporary Accounting Research, 24(4), 12911318.

You might also like