THE JOURNAL OF FINANCE • VOL. LV, NO. 6 • DEC.

2000

Information Asymmetry, R&D,

and Insider Gains

DAVID ABOODY and BARUCH LEV*

ABSTRACT
Although researchers have documented gains from insider trading, the sources of
private information leading to information asymmetry and insider gains have not
been comprehensively investigated. We focus on research and development ~R&D!—an
increasingly important yet poorly disclosed productive input—as a potential source
of insider gains. Our findings, for the period from 1985 to 1997 indicate that in-
sider gains in R&D-intensive firms are substantially larger than insider gains in
firms without R&D. Insiders also take advantage of information on planned changes
in R&D budgets. R&D is thus a major contributor to information asymmetry and
insider gains, raising issues concerning management compensation, incentives, and
disclosure policies.

Researchers have documented that “corporate insiders,” defined by the 1934

Securities and Exchange Act as corporate officers, directors, and owners of
10 percent or more of any equity class of securities, gain from trading in the
securities of their firms.1 However, the specific sources of information lead-
ing to insider gains in particular and to information asymmetry in general
have not been comprehensively investigated. An identification of the sources
of insiders’ information contributes to our knowledge in various ways. For
example, researchers often use proxies for the extent of information asym-
metry, such as the number of analysts following a firm, the number of com-
peting traders, or insiders’ and institutional ownership ~e.g., Stoll ~1978!,

* University of California at Los Angeles and New York University, respectively. We are
grateful for the comments and suggestions made by Yakov Amihud, Jennifer Carpenter, Ken
Garbade, Dan Givoly, Kose John, Eli Ofek, and David Ravia.
1
Estimates of the gains from insider trading vary widely: Early studies ~e.g., Lorie and
Niederhoffer ~1968!, Jaffe ~1974!, and Finnerty ~1976!! report abnormal gains ranging from 3 to
30 percent, for holding periods of up to three years. Seyhun ~1986!, using data on insider
trading reported to the SEC during 1975 through 1981, finds more modest gains to insiders:
Over 300 days subsequent to trade, the average risk-adjusted gains were 4.3 percent for stock
purchasers and 2.2 percent for sellers. Most of these gains occurred during the first 100 days
after trade. In a subsequent study, Seyhun ~1992! documents for the period from 1975 to 1989
average abnormal returns to insiders of 2.6 percent for six months after stock purchases, and
5.3 percent for six months following sales. Jeng, Metrick, and Zeckhuser ~1999!, using value-
weighted portfolio schemes for the period from 1975 to 1996, report for a one year holding
period insider gains on purchases of roughly 0.4 percent abnormal returns per month and
insignificant abnormal returns for sales.

2747
2748 The Journal of Finance

Brennan and Subrahmanyam ~1995!!. Additional information asymmetry

proxies are firm size and volume of trade ~Chari, Jagannathan, and Ofer
~1988!!, financial analysts’ forecast errors of earnings, and the volatility of
abnormal stock returns ~e.g., Krishnaswami and Subramaniam ~1999!!. Such
proxies are obviously noisy, ref lecting, in addition to information asymme-
try, numerous firm and market attributes. Identification of the major sources
or firm-specific drivers of information asymmetry will suggest more precise
~less noisy! measures of asymmetry. For policy research concerned with the
social consequences of insider gains and information asymmetry, identi-
fication of the sources of insiders’ information will suggest means ~e.g., en-
hanced disclosure of specific information! of mitigating harmful consequences.2
We investigate the insider gains issue from the perspective of a specific
source of information asymmetry—research and development ~R&D!. Invest-
ment in R&D is a major productive input in a large number of firms, par-
ticularly those operating in the technology and science-based sectors. R&D,
however, differs from other capital and financial inputs ~e.g., property, plant,
and equipment, inventory, or project financing! along several important di-
mensions related to information asymmetry. First, many R&D projects, such
as radically new drugs under development or software programs, are unique
to the developing firm, whereas most capital investments, such as com-
mercial property or airplanes, share common characteristics across firms
within an industry. Consequently, investors can derive little or no infor-
mation about the productivity and value of a firm’s R&D from observing
the R&D performance of other firms ~e.g., not much can be learned about
Merck’s drug development program from an FDA approval of a Pfizer drug!,
whereas, for example, the average store performance of one retailer provides
valuable information on the performance of other retailers. Second, while
most physical and financial assets are traded in organized markets, where
prices convey information about asset productivity and values, there are no
organized markets for R&D and hence no asset prices from which to derive
information. Third, accounting measurement and reporting rules treat R&D
differently from other investments: While these rules mandate the marking-
to-market in quarterly and annual reports of most financial investments,
and the periodic recognition of value impairment ~the decrease of market
value below cost! of physical assets, thereby providing investors with up-
dated information about changes in asset values, R&D is immediately ex-
pensed in f inancial statements, so that no information on value and
productivity changes of R&D is reported to investors.

2
See Fried ~1998! for a summary of the debate about the social consequences of insider gains
and the effectiveness of securities regulations aimed at limiting these gains. The spectrum of
opinions about social consequences ranges from viewing insider trading as desirable ~e.g., it
enhances market efficiency! to viewing insider gains as detrimental to firms ~increase cost of
capital, distort managerial incentives! and eroding investors’ confidence in the integrity of
capital markets. See also Jeng et al. ~1999! who conclude that under the current regulatory
system outsiders are not significantly disadvantaged when trading with insiders.
 Information Asymmetry, R&D, And Insider Gains 2749

Given the relative scarcity of public information about firms’ R&D activ-
ities, and the importance of these activities to the operations and profit
potential of technology and science-based companies, we hypothesize that
R&D contributes to information asymmetry between corporate insiders and
outside investors and that some of the former will exploit this asymmetry to
gain from insider trading.3 Indeed we find from comprehensive data on cor-
porate officers’ share trading from 1985 to 1997 that insider gains in firms
conducting R&D ~R&D firms! are substantially larger than insider gains in
firms with no R&D activities ~No-R&D firms!. These differences in insider
gains are both statistically and economically significant, and hold after con-
trolling for various known risk factors. We also find that investors’ reaction
to the public disclosure of insiders’ trade ~about a month, on average, after
the trade! is stronger for R&D firms than for No-R&D firms, corroborating
our hypothesis that R&D activities enhance information asymmetry, and
that this asymmetry is not eliminated by insiders’ trade and investors’ in-
formation search. We thus identify R&D as a major contributor to informa-
tion asymmetry.
The paper is organized as follows. In Section I we develop our hypothesis
and in Section II we describe the sample and summary statistics. Section III
presents the estimation equations and reports empirical findings on the as-
sociation between insider gains and firms’ R&D intensity. In Section IV we
examine investors’ reaction to the public disclosure of insider trades, and in
Section V we present robustness tests. Section VI concludes the study.

I. Information Asymmetry, R&D, and Insider Gains

All corporate investments create information asymmetries because man-
agers can continually observe changes in investment productivity on an in-
dividual asset basis ~e.g., aircraft utilization—load factor—at the route level!,
whereas outsiders obtain only highly aggregated information on investment
productivity at discrete points of time. The extent of information asymmetry
associated with R&D, however, is larger than that associated with tangible
~e.g., property, plant, and equipment! and financial investments because of
the relative uniqueness ~idiosyncrasy! of R&D. Thus, for example, a failure
of a drug under development to pass Phase I clinical tests, or of a software
program to successfully complete an alpha ~technological feasibility! test in
a particular company are unique events not shared by other pharmaceutical
or software companies. In contrast, a downturn in demand for commercial

3
A perspective on the importance of R&D as a productive input can be gained from the
fact that in 1997 total R&D spending in the United States was $210 billion, compared with
$215 billion invested by manufacturing firms in property, plant, and equipment. The increas-
ing importance of R&D is also ref lected by its faster growth rate compared with other major
inputs. For example, over the period from 1970 to 1997, the average annual growth rate of R&D
was 8.0 percent, whereas the growth rate of capital investment ~property, plant, and equip-
ment! was 6.8 percent ~see Economic Report of the President, 1997, and National Science
Foundation0SRS: Research and Development Performance by Sector!.
2750 The Journal of Finance

properties, for example, will exert a strong common effect on the property
values of all real estate companies operating in a given geographical region.
Similarly, interest-rate changes will affect systematically the values of bond
and stock portfolios of companies. Thus, we argue, the relative uniqueness of
R&D investments makes it difficult for outsiders to learn about the produc-
tivity and value of a given firm’s R&D from the performance and products of
other firms in the industry, thereby contributing to information asymmetry.4
The absence of organized markets in R&D further contributes to informa-
tion asymmetry. Whereas investors can derive considerable information from
prices of traded tangible and financial assets concerning their values at the
firm level ~e.g., inferring from changes in commodity prices about swings in
values of firms’ inventories!, there is no direct price-based information on
firm-specific changes in the value and productivity of R&D.5 Some informa-
tion on R&D can, of course, be inferred from stock prices of R&D-intensive
companies, yet such information is noisy, given the multiple activities of
R&D firms ~e.g., manufacturing, services!.
Accounting rules exacerbate the information asymmetries associated with
R&D. Most financial assets have to be marked-to-market in quarterly and
annual financial reports, and impairments in the values of tangible assets
~i.e., when expected future benefits fall short of book values! have to be
routinely reported in financial statements. Similarly, inventories and ac-
counts receivable have to be written down in financial reports to market
values. Thus, investors are periodically informed about changes in the val-
ues of most tangible and financial assets. In contrast, R&D expenditures are
uniformly expensed in financial reports and therefore no information is re-
quired to be provided to outsiders about changes in the productivity and
value of R&D. Even major R&D events, such as when a drug under devel-
opment successfully passes clinical tests, are not routinely reported to
investors.
Empirical evidence is consistent with a relatively large information asym-
metry associated with R&D. For example, Barth, Kasznik, and McNichols
~1998! report that analyst coverage ~number of analysts following a firm! is
significantly larger for firms intensive in R&D relative to firms with lower
or no R&D, presumably because of the private information concerning R&D

4
The uniqueness of R&D is widely recognized in economics and finance research. Thus, for
example, Titman and Wessels ~1988, p. 5! postulate that asset uniqueness is a determinant of
corporate capital structure and measure uniqueness by R&D intensity, arguing that R&D “mea-
sures uniqueness because firms that sell products with close substitutes are likely to do less
research and development since their innovations can be more easily duplicated. In addition,
successful research and development projects lead to new products that differ from those ex-
isting in the market.”
5
Griliches ~1995, p. 77! notes: “A piece of equipment is sold and can be resold at a market
price. The results of research and development investments are by and large not sold directly
. . . the lack of direct measures of research and development output introduces an inescapable
layer of inexactitude and randomness into our formulation.” Such randomness and inexactitude
are obviously less severe to insiders than to outsiders.
 Information Asymmetry, R&D, And Insider Gains 2751

activities. Furthermore, analysts’ efforts and, presumably, costs of analyzing

firms vary positively with R&D intensity.6 Similarly, Tasker ~1998! reports
that R&D-intensive companies conduct more conference calls with analysts
than low ~or no! R&D firms, implying a stronger investors’ demand for in-
formation about the R&D activities of firms.7
Is all the private R&D-related information shared in a timely manner
with outsiders through the information search of analysts and investors, so
that in equilibrium there are no substantial information asymmetries left?
Kyle’s ~1985! model of a single informed trader with many uninformed ~noise!
traders and a market maker addresses this question and indicates that in
equilibrium: ~1! “The informed trader trades in such a way that his private
information is incorporated into prices gradually.” ~p. 1316, emphasis ours!,
and ~2! “not all information is incorporated into prices by the end of trading.”
~p. 1326!. The main conclusion of the model is that, while much of the in-
sider’s information gets gradually incorporated in prices through his0her
trades, the “insider makes positive profits by exploiting his monopoly power
optimally in a dynamic context. . .” ~p. 1315, emphasis ours!.8
Particularly relevant to our hypothesis concerning R&D-related informa-
tion asymmetry and the consequent insider gains is Kyle’s ~1985! conclusion
that insiders’ profits are proportional to ~S 0 su !102 , where S 0 is the variance
of the liquidation value of the risky asset. We know from Kothari, Laguerre,
and Leone ~1998! that: “. . . R&D investments generate more uncertain fu-
ture benefits than investment in tangible assets. Specifically, in a regres-
sion of future earnings variability on investment in R&D, PP&E @property,
plant, and equipment# and other determinants of earnings variability like
firm size and leverage, we find that the coefficient on R&D is about three
times as large as that on PP&E” ~p. 3!. Thus, the documented larger variabil-
ity of earnings and, by implication, firms’ liquidation value associated with
R&D should, according to Kyle ~1985!, enhance the gain of insiders in R&D
firms, relative to insiders in No-R&D firms with lower variability of earnings.

6
Analysts’ efforts were proxied by Barth et al. ~1998! by the number of other firms followed
by a given firm’s analysts. Assuming that analysts have a common capacity limit and expend
efforts up to that capacity, the smaller the number of firms an analyst follows, the larger, on
average, the efforts spent analyzing those firms.
7
Tasker ~1998b! also reports that the majority of questions raised by analysts in conference
calls involve R&D-related issues, such as the content of the company’s product pipeline.
8
Kyle’s ~1985! predictions are sensitive to the number of informed traders. For example,
Back, Cao, and Willard ~1999! conclude: “Competition may or may not lead to greater ‘effi-
ciency’ of prices. Whether the market price ref lects private information more quickly when
there are competing informed traders depends on how the information is distributed among the
agents. Indeed, in this model, it is never the case that the market is always more efficient when
information is distributed among competing traders than when the information is possessed by
a single trader” ~p. 30!. “A somewhat surprising result is that, beyond some date, the market
would have learned more from a monopolist informed trader than from competing traders,
regardless of the correlation of the competitors’ signals” ~p. 2!. “The relatively large amount of
private information remaining near the end of trading leads to an extreme adverse selection
problem” ~p. 3!.
2752 The Journal of Finance

Kyle’s second ~1989! model is particularly relevant to our case, because it

allows uninformed investors to acquire private information. Among the con-
clusions of this model are: ~1! “thus, while uninformed speculators break
even on average, informed speculators make money ‘at the expense’ of noise
traders . . .” ~p. 337!, ~2! “with imperfect competition, prices never reveal
more than one-half the private precision of informed speculators.” ~p. 334!,
and ~3! “in order for the large market model to reveal any private informa-
tion, private information must be cheap enough so that a large number of
speculators find it profitable to purchase it.” ~p. 344!. Thus, even with en-
dogenous information acquisition, insiders are expected to gain from insider
trading.
Given the importance of R&D in firms’ productivity and growth, it is rea-
sonable to expect investors and analysts to acquire private R&D-related in-
formation from managers, as the evidence discussed above indicates. However,
because such information is costly, requiring among other things a signifi-
cant investment in scientific knowledge ~e.g., understanding genome re-
search in biochemistry! and a considerable time investment ~e.g., analyzing
financial reports, participating in conference calls!, optimal information ac-
quisition by outsiders will generally fall short of completely exhausting in-
siders’ information. Stated differently, in equilibrium the marginal value of
acquired information will equal marginal cost, but this does not necessarily
imply that the total R&D information possessed by managers will be quickly
incorporated in prices, leaving managers with no opportunities to gainfully
trade on inside information. Indeed, our evidence ~Sec. IV! indicates a sig-
nificant market reaction to the disclosure of insiders’ trades ~25 days, on
average, subsequent to the actual trades!, a reaction that is more pro-
nounced for R&D than for No-R&D companies, implying that almost a month
after trade by insiders and the extensive information search by outsiders,
prices still did not fully ref lect all of insiders’ private information.
We accordingly hypothesize in this study that the R&D activities of firms
create unique information asymmetries and that officers of R&D companies
will gainfully exploit these asymmetries by trading the shares of their
firms.

II. Data Sources and Sample Characteristics

The insider trading data analyzed in this study were obtained from CDA0
Investnet. The database contains all purchase and sale transactions made
by insiders and reported to the SEC from January 1985 through December
1997.9 We focus on 323,949 open market purchase and sale transactions that
were conducted by officers of firms, as officers are a priori presumed to have

9
The enforcement of the Securities and Exchange Act of 1934 was considerably strengthened
by the Insider Trading Sanctions Act of 1984 ~see Bainbridge ~1998!!, providing a reason to
start the sample period in 1985.
 Information Asymmetry, R&D, And Insider Gains 2753

inside information on R&D.10 We delete 732 duplicate transactions and change

of control transactions where the number of shares exchanged exceeded 20
percent of the total shares outstanding. We also delete 2,334 companies with
31,417 officer transactions that could not be located on the CRSP database,
and 1,459 firms with 29,897 transactions that were not found on COM-
PUSTAT. Finally, we delete 8,865 insider transactions because of missing
data on R&D in COMPUSTAT, resulting in a final sample of 253,038 trans-
actions related to 10,013 firms.
Table I provides sample statistics by type of insider transaction and year.
Of the 10,013 sample firms, 3,818 are classified as R&D firms and 6,195 are
classified as No-R&D firms; the latter are those for which COMPUSTAT
does not report any R&D expenditures during the period from 1985 to 1997.
Consistent with prior studies, the total number of sale transactions ~165,949;
81,539 for R&D and 84,410 for No-R&D firms! is almost twice the number of
purchase transactions ~87,089!. The difference between purchase and sale
transactions ~number and value!, however, is substantially larger for R&D
firms ~23,008 vs. 81,539! than for No-R&D firms ~64,081 vs. 84,410!, ref lect-
ing the pervasiveness of stock options and awards ~included in sales but not
in purchases! in R&D companies. The number and volume of insider trans-
actions has increased continuously during the period from 1985 to 1997, and
the increase in transaction values for R&D firms has been proportionately
larger than for No-R&D firms. The three transaction measures—number of
transactions, number of shares, and total value of transactions—are highly
correlated. In the tests reported below we use the number of shares trans-
acted. Replication of the tests with dollar value of transactions yielded very
similar results to those derived from the number of shares.
Panel B of Table I reports mean and median market-adjusted returns ~raw
return minus the return on a value-weighted NYSE0AMEX0Nasdaq index!
for three intervals subsequent to insider transactions: From transaction date
to 1 day before the transaction’s filing with the SEC ~an average of 25 days
in our sample! and 6 and 12 months following the transaction. The excess
return data indicate that, on average, insider share purchases were followed
by positive returns in each interval, whereas share sales were followed by
negative returns. Thus, insiders in both R&D and No-R&D companies tend
to buy shares ahead of good news and sell ahead of bad news. However, both
the mean and median excess returns of R&D firms are significantly higher
for purchase and lower for sale transactions than returns of firms without
R&D activities. Thus, for example, the mean market-adjusted return from
transaction to SEC filing date of insider purchases in R&D firms was 3.0
percent versus 0.9 percent for share purchases in No-R&D firms. Note that

10
The database we use includes several types of transactions we did not consider because it
is not clear to what extent they are motivated by inside information. These transactions in-
clude: Acquisition of stocks through company plans, gifts of stocks, acquisitions through divi-
dend reinvestment plans and under employee benefit plans. Examples of nonofficer insiders we
exclude from the sample are various trustees and owners without managerial capacity.
 2754
Table I
Distribution of Insider Transactions and Market-Adjusted Returns, January 1985–December 1997
All means and medians in Panel B are significantly different from zero at the one percent level and significantly different between R&D and
No-R&D firms at the five percent level. Market adjusted returns are the raw returns minus the return on a value weighted NYSE0AMEX0
Nasdaq index. No-R&D firms are those for which COMPUSTAT does not report R&D expenditures in any quarter during 1985 to 1997.

Panel A: Transaction Data

Jan. 1985–Dec. 1989 Jan. 1990–Dec. 1993 Jan. 1994–Dec. 1997 1985–1997

R&D No-R&D R&D No-R&D R&D No-R&D R&D No-R&D

No. of transactions

The Journal of Finance

No. of purchases 7,027 17,124 7,004 21,794 8,977 25,163 23,008 64,081
No. of sales 18,255 21,585 28,018 29,150 35,266 33,675 81,539 84,410
Total no. of transactions 25,282 38,709 35,022 50,944 44,243 58,838 104,547 148,491
No. of shares ~millions!
No. of shares purchased 51.1 150.4 110.4 246.7 150.5 329.3 312.0 726.4
No. of shares sold 222.0 383.7 392.6 567.6 572.4 869.9 1,187.0 1,821.2
Total no. of shares traded 273.1 534.1 503.0 814.3 722.9 1,199.2 1,499.0 2,547.6
Value of transactions ~$ millions!
Val. of shares purchased 366.7 2,257.8 311.4 1,599.2 699.8 2,855.7 1,377.9 6,712.7
Val. of shares sold 3,386.2 5,213.6 8,544.8 10,725.4 16,190.5 18,489.4 28,121.5 34,428.4
Total val. of shares traded 3,752.9 7,471.4 8,856.2 12,324.6 16,890.3 21,345.1 29,499.4 41,141.1

Panel B: Return Data

R&D Firms No-R&D Firms
Purchases Sales Purchases Sales
Mean Median Mean Median Mean Median Mean Median
Market-adjusted return from transaction to SEC filing date 3.0% 0.7% 20.5% 20.2% 0.9% 0.07% 20.1% 20.1%
Market-adjusted return over 6 months following the 9.61% 1.0% 24.1% 27.1% 3.56% 0.2% 23.8% 24.6%
transaction date
Market-adjusted return over 12 months following the 8.56% 1.93% 210.0% 214.0% 2.47% 0.36% 28.6% 29.2%
transaction date
 Information Asymmetry, R&D, And Insider Gains 2755

the 3.0 percent excess return for R&D firms is economically large, given that
the mean interval between transaction and filing dates is 25 days only. For
insider sales, the mean market-adjusted return over the transaction-filing
interval for R&D firms was 20.5 percent versus 20.1 percent for No-R&D
firms.
The post-transaction returns generally increase from the first interval ~trans-
action to SEC filing date! to the second interval ~6 months following trade!,
and further increase for the third interval ~12 months following trade!. Dur-
ing six months following trade, excess mean returns from share purchases
for R&D firms were 9.61 percent versus 3.56 percent for No-R&D firms. All
the differences between mean ~median! returns of R&D versus No-R&D firms
in Table I are statistically significant at the 5 percent level.
The return data in Table I are consistent with our hypothesis: Insider
gains in R&D-intensive companies are substantially higher than insider gains
in No-R&D companies. However, the underlying individual transaction data
are not independent ~there are multiple transactions per firm!, and the re-
turns over the 6 and 12 months intervals are overlapping. Moreover, firm
attributes ~e.g., risk, size! related to R&D may affect the documented re-
turns, in addition to the hypothesized information asymmetry differences.
Accordingly, the return data in Table I should be viewed as descriptive and
tentative, and we proceed below to aggregate insider transactions by firms,
considering only no overlapping return intervals and controlling for known
risk factors.

III. Insider Gains and R&D

We wish to examine the association between insider gains and informa-
tion asymmetry as proxied by the firm’s R&D intensity. To accomplish this
we construct four monthly calendar-time portfolios conditional on the firms’
R&D activities and the type of insiders’ transactions ~purchase or sale!. The
four portfolios are: ~1! RDp , for firms engaged in R&D whose insiders were
“net purchasers” of shares in a given month during the period from 1985 to
1997 ~a net purchaser firm-month is one where the number of shares pur-
chased by the firm’s officers during the month exceeded the number of shares
sold!; ~2! RDs , for firms engaged in R&D whose insiders were “net sellers” of
shares ~i.e., number of shares sold by insiders during the month exceeded
purchases!; ~3! NORDp , for firms without R&D whose insiders were net pur-
chasers; and ~4! NORDs , for firms without R&D whose insiders were net
sellers of shares.11
We calculate returns for each of the four portfolios as follows. For each
calendar month ~January 1985 through November 1997, a total of 155 months!
we compute a firm-specific mean raw return from the transaction dates of
insiders’ trades to one day prior to the filing date of those transactions with

11
In the large majority of sample firm-month cases, insider trades were either all sales or all
purchases.
2756 The Journal of Finance

the SEC. These firm-specific transaction-to-reporting returns are averages

over all the individual insider trades that occurred during the month. We
then compute calendar-time equally weighted portfolio returns over all the
firms with insider transactions in a given month, classified into the four
portfolios described above, namely firms with and without R&D, where in-
siders during the month were net stock purchasers or net sellers.12 We thus
focus on portfolio returns over the transaction-to-SEC filing period, during
which information asymmetry is presumably large. The public disclosure of
insiders’ trade substantially decreases information asymmetry, as evidenced
by investors’ reaction to this information ~Sec. IV!.
To examine the extent to which insiders in R&D firms gain more than
those in No-R&D firms, we employ an intercept test using the three-factor
model of Fama and French ~1993!. The dependent variable is the differ-
ence between calendar-time portfolio returns of R&D and No-R&D firms
~RDpt 2 NORDpt !. For example, for our first month, January 1985, RDpt is
the average return for all R&D firms whose officers were net purchasers of
shares during January 1985, over the transaction-to-filing interval. NORDpt
is the average return for all firms without R&D whose officers were net
purchasers of shares during January 1985, over the transaction-to-filing in-
terval. The regression is run over 155 observations—individual months dur-
ing January 1985 to November 1997. The independent variables are the
three factors: Market return, size, and book-to-market. The regression is
thus:

RDpt 2 NORDpt 5 ap 1 bp ~R mt 2 R ft ! 1 d SMBt 1 sp HML t 1 ep , ~1!

where

RDpt 2 NORDpt 5 return from going long on a portfolio of firms that en-
gage in R&D and short on a portfolio of firms that do
not engage in R&D, in months where insiders were net
purchasers of shares. The return interval is between
transaction date and one day prior to SEC filing date—
25 days on average.
R mt 2 R ft 5 the market excess return in month t.
SMBt 5 the difference between month t return on a value-
weighted portfolio of small stocks and one of large stocks.
HML t 5 the difference between month t return on a value-
weighted portfolio of high book-to-market stocks and
one of low book-to-market stocks.13

12
We replicate our tests with value-weighted portfolio returns and obtain very similar re-
sults to those of the equally-weighted returns.
13
The construction of these variables is described in Fama and French ~1993!. We thank Ken
French for providing us the data for the independent variables in equation ~1!.
 Information Asymmetry, R&D, And Insider Gains 2757

Table II
Returns from Going Long on Insider Trading in R&D Firms
and Short on No-R&D Firms
Panel A presents mean and median percentage raw returns earned on portfolios formed as
follows: For each month between January 1985 and November 1997 we calculate, for each
sample firm, the mean raw return from the insider transaction date to one day prior to the SEC
filing date, over all insider transactions during the month. We calculate the mean returns
separately for firms with and without R&D, and for firms where insiders were net purchasers
during the month ~i.e., share purchases exceeded sales! and firms where insiders were net
sellers during the month. # . 0 indicates number of months ~out of 155! when the mean return
was positive. In Panel B, the intercept ~a! of the Fama-French three-factor model in equation
~1! is presented. It is the estimated intercept from a time series regression ~155 observations!
of the portfolio returns of firms with R&D minus the portfolio return of firms without R&D
~RD 2 NORD! on the market excess return ~RM 2 RF!, a zero-investment size portfolio ~SMB!,
and a zero-investment book-to-market portfolio ~HML!.

Panel A: Univariate Returns

Insider Purchases Insider Sales

Mean Median #.0 Mean Median #.0

Firms with R&D 5.49%*** 4.64%*** 136 21.26%*** 21.17%*** 57

Firms without R&D 4.57%*** 3.82%*** 133 20.66%*** 20.44%** 67
RDt 2 NORDt 0.92%*** 1.66%*** 96 20.60%*** 20.70%*** 53

Panel B: Three-factor Model

a RMt 2 RFt SMBt HMLt Adjusted R 2

Insider Purchases
RDt 2 NORDt 0.011** 20.129 0.491** 20.364 0.056
t-statistic ~2.40! ~21.01! ~2.65! ~21.56!
Insider Sales
RDt 2 NORDt 20.005* 0.088 0.232** 20.072 0.043
t-statistic ~21.85! ~1.24! ~2.24! ~20.55!

*, **, and *** denote significance at the 10, 5, and 1 percent levels, respectively.

An identical regression to equation ~1! was run for portfolios of insider sales:
RDst 2 NORDst .
Panel A of Table II presents the univariate raw returns for the four port-
folios examined. Consistent with prior research, we observe for all trades
positive stock returns subsequent to insiders’ share purchases and negative
returns subsequent to insiders’ sales. As hypothesized above, both the mean
and median returns on the R&D portfolios are significantly higher than
returns on the No-R&D portfolios when insiders purchased shares and lower
when they sold shares. Investing long in a portfolio of R&D firms whose
insiders purchased shares and short in a portfolio of No-R&D firms whose
insiders purchased shares ~RDt 2 NORDt ! yields a mean ~median! return of
0.92 percent ~1.66 percent! over an average of 25 days between transaction
2758 The Journal of Finance

and SEC filing dates.14 The insider sales portfolios yield a differential mean
~median! return of 20.60 percent ~20.70 percent! over the same interval in
favor of R&D firms.
Panel B of Table II presents estimates from the Fama-French three-factor
model in equation ~1!.15 As hypothesized, the estimated intercepts from time-
series regressions of the difference in return between firms with R&D and
those without R&D ~RD 2 NORD! on the three systematic factors are sig-
nificantly positive when insiders purchased shares ~0.011, t 5 2.40!, and
significantly negative when insiders sold shares ~20.005, t 5 21.85!. The
estimated intercepts are close to the univariate returns in Panel A: 0.0092
for purchases and 20.006 for sales. Given that the average interval over
which these returns were gained was 25 days, the estimated gains are eco-
nomically meaningful as well, particularly for stock purchasers.
An additional test is performed to gain insight into the association be-
tween R&D intensity and insider gains. First, we divide the sample into
three groups: Firms without R&D expenditures, firms with R&D intensity
~R&D over sales! below the sample median ~low R&D!, and firms with R&D
intensity above the sample median ~high R&D!. We then run the three-factor
model in equation ~1! for return differences between Low and No-R&D, High
and No-R&D, and High and Low R&D portfolios. These regressions are run
separately for net purchasers and net sellers of shares. Following are the
estimated a coefficients and t-values from these regressions:

A. Insider Purchases a coefficient t-value

Low R&D minus No-R&D 0.002 0.388
High R&D minus No-R&D 0.021 3.446
High R&D minus Low R&D 0.019 2.997
B. Insider Sales a coefficient t-value
Low R&D minus No-R&D 20.004 21.643
High R&D minus No-R&D 20.006 21.783
High R& D minus Low R&D 20.002 20.510

The estimated a coefficients indicate that for insider purchases of shares,

the differential gain between R&D and No-R&D firms is mainly due to high
R&D-intensity firms, as indicated by the estimated a for High R&D minus
No-R&D ~0.021!, which is 10 times larger than the estimated a of Low R&D
minus No-R&D ~0.002!. For insider sales, the increase in insider gains in
R&D firms ~relative to No-R&D! is more monotonic in R&D intensity, as
indicated by an estimated a of 20.004 for Low R&D minus No-R&D and a of
20.006 for High R&D minus No-R&D firms.

14
Such an investment strategy cannot, of course, be implemented by outsiders who are not
informed about insiders’ transactions in real time.
15
We obtain similar results with a four-factor model, where the fourth factor is a return
momentum ~see Carhart ~1997!!.
 Information Asymmetry, R&D, And Insider Gains 2759

The analysis reported in this section thus indicates that insiders in R&D
firms gain from trade in their firms’ shares significantly more than insiders
in No-R&D firms, and that for share purchases the gain differential is mainly
attributed to firms with high ~above median! R&D intensity.

IV. Investors’ Reaction to the Disclosure of Insider Trades

Section 16~a! of the 1934 Securities Act requires insiders to report their
preceding month’s trades to the SEC no later than the tenth day of each
month. An examination of investors’ reaction to the public disclosure of in-
siders’ trades offers an opportunity to investigate several important issues.
In particular:
~1! Is R&D a significant driver ~cause! of information asymmetry? If, as
hypothesized above, R&D contributes to information asymmetry, it is
reasonable to expect investors to react more strongly to the disclosure
of insider trades in R&D firms than in No-R&D firms. The reason:
Given the relative scarcity of public information about firms’ research
and product development activities, insider trades in R&D companies
convey, on average, more information to the market than insider trades
in No-R&D firms.
~2! Is insiders’ private information fully and quickly revealed to investors
by financial analysts’ search efforts and by insiders’ own trade activ-
ities? In the case of such efficient information revelation, there should
be no unusual investor reaction to the public disclosure of insiders’
trades. If, however, insiders’ private information is not fully revealed
to the market ~thereby creating gain opportunities!, the public disclo-
sure of their trades should trigger investor reaction.
Our test results ~below! indicate that R&D activities do contribute to in-
formation asymmetry and that not all of insiders’ private information is
revealed before the public disclosure of their trades. Table III presents raw
returns for three disclosure intervals: The day the trade information was
filed with the SEC ~day 0!; 16 the cumulative return over the day of filing
and the subsequent trading day ~0, 11!; and the three-day cumulative re-
turn beginning with the day of filing ~0, 11, 12!. We include days 11 and
12 in the analysis because the SEC may disclose insiders’ filings after trad-
ing hours of day 0 or on the following day.
As reported in Table III, for each of the three disclosure intervals the
mean returns are positive when insider purchases were disclosed and neg-
ative when insider sales were disclosed, for both R&D and No-R&D firms.
Thus, consistent with Kyle ~1985, 1989!, despite analysts’ search activities
and insiders’ own trades, not all of insiders’ private information is revealed
to investors prior to trade disclosure. However, the reaction to disclosure of
insider trades in R&D firms is significantly stronger than the reaction to

16
This date is defined in our database as the date when the filing was received by the SEC.
2760 The Journal of Finance

Table III
Investors’ Reaction to the Disclosure of Insider Trades
Day 0 is the date when the filing on insider transactions was received by the SEC. Days ~0, 11!
refers to the cumulative raw return over day 0 and the following trading day. Days ~0, 11, 12!
is the three-day return. There are 29,689 sale transactions for No-R&D firms and 24,166 sale
transactions for firms with R&D. There are 24,961 purchase transactions for No-R&D firms
and 9,262 purchase transactions for R&D firms. The total number of insider transactions here
is smaller than in Table I, because multiple transactions per firm reported to the SEC on the
same day are considered here as one transaction. All mean and median returns are signifi-
cantly different from zero at the one percent level. # . 0 indicates the percentage of individual
returns that are positive.

Insider Purchases Insider Sales

Mean Median #.0 Mean Median #.0

Day 0
Firms without R&D 0.15%*** 0.00%* 35.6% 20.10%* 0.00%*** 39.3%
Firms with R&D 0.31%*** 0.01%* 55.5% 20.15%* 20.01*** 38.4%
Days ~0, 11!
Firms without R&D 0.24%*** 0.00%* 44.3% 20.19%*** 0.00%*** 45.1%
Firms with R&D 0.40%*** 0.01%* 55.1% 20.34%*** 20.01%*** 44.4%
Days ~0, 11, 12!
Firms without R&D 0.39%*** 0.00%* 47.7% 20.28%*** 0.00%*** 46.9%
Firms with R&D 0.58%*** 0.02%* 58.8% 20.47%*** 20.01%*** 42.2%

* and *** denote that the means or medians of the R&D and No-R&D firms are significantly
different from each other at the 10 and 1 percent levels, respectively.

firms without R&D. When insiders’ purchases of shares were disclosed, the
mean market reaction on day 0 for R&D firms was 0.31 percent versus 0.15
percent for No-R&D firms ~the difference is significant at the 1 percent level!.
When insiders’ sales of shares were disclosed, the mean market reaction on
day 0 for R&D firms was 20.15 percent versus 20.10 percent for No-R&D
firms.17 The reaction to insider trade disclosure increases with the interval
length, indicating that not all of the filed information is disclosed to inves-
tors on the filing date.18
The volume of trade surrounding the disclosure of insider transactions is
another indication of information revelation. We find that for R&D firms
where insiders sold shares the mean ~median! volume of shares traded ~de-
f lated by outstanding shares! on day 0 was 0.74 percent ~0.34 percent!, whereas
for No-R&D firms the mean ~median! volumes were 0.41 percent ~0.19 per-

17
Repeating the analysis of Table III using value-weighted mean returns slightly increases
the significance of the results, whereas using size-adjusted returns slightly decreases the sig-
nificance of the results.
18
We also calculate the correlation between the firms’ R&D intensity ~R&D to sales!, as a
proxy for information asymmetry, and the market reaction to trade disclosure on day 0. The
correlation coefficient is 0.039 ~ p-value of 0.001! for insider purchases and 20.019 ~ p-value of
0.001! for insider sales.
 Information Asymmetry, R&D, And Insider Gains 2761

cent!. The differences between those means ~medians! are statistically sig-
nificant at the 1 percent level. For insider purchases in R&D firms, the
mean ~median! volume of shares traded ~def lated by outstanding shares! on
day 0 was 0.44 percent ~0.28 percent!, whereas for No-R&D firms, the mean
~median! volume was 0.25 percent ~0.10 percent!. The differences between
those means ~medians! are also statistically significant at the one percent
level. We document similar, yet larger differences in volume of trade be-
tween R&D and No-R&D firms as we expand the interval to one and two
days following the SEC filing date.
Based on the evidence, we conclude that investors react more strongly to
the public disclosure of insider trades in R&D firms compared with firms
without R&D activities, consistent with the hypothesis that R&D contrib-
utes to information asymmetry. Stated differently, there is more private in-
formation in R&D companies than in companies without R&D and observing
insider trades ~even with a time lag! is one means of narrowing the infor-
mation gap. Furthermore, investors’ significant reaction to insider trade dis-
closure, made on average 25 days after the actual trade, indicates that not
all of the insiders’ private information is revealed through their own trade or
analysts’ search activities.

V. Robustness Tests
A. Trade Intensity and R&D
If R&D enhances information asymmetry and managers exploit this asym-
metry we would expect heavier insider trade activity in R&D firms than in
firms without R&D, as the formers’ officers attempt to gain from unique
private information. The heavier activity of insiders in R&D firms may be
ref lected in more frequent trade and0or more intensive trade ~relative to shares
outstanding!. Indeed we find that insiders in R&D firms trade relatively
frequently: The mean ~median! number of insider trades for R&D firms over
the sample period was 17.1 ~11! per firm, compared with 15.3 ~9! for No-
R&D firms.19 We also find that the intensity of insider trades, measured in
various ways, is higher for R&D firms than for firms without R&D. For
example, when we measure trade intensity by the ratio of the number of
shares traded by insiders in each transaction to the firm’s outstanding shares,
we find the mean ~median! of that ratio for R&D firms to be 0.07 percent
~0.03 percent!, whereas for firms without R&D it is 0.06 percent ~0.02 per-
cent!; the differences for both mean and median are statistically significant
at the one percent level. Furthermore, the correlation between R&D inten-
sity and insider trade intensity is 0.112 ~ p-value of 0.0001!. When insider
trade intensity is measured as the ratio of the number of shares traded by
an insider to the individual’s total share holdings, we find the mean ~me-

19
The means and medians mentioned here are significantly different at the one percent
level between R&D and No-R&D firms.
2762 The Journal of Finance

dian! intensity in R&D firms to be 22.4 percent ~8.5 percent! versus 21.5 per-
cent ~6.5 percent! in No-R&D firms ~the differences are once more significant
at the one percent level!. The correlation between this intensity measure
and the firm’s R&D intensity is 0.066, significant at the 1 percent level.20
The heavier trade activity of insiders in R&D firms documented here raises
the question whether our findings in Section III ~R&D officers gain more
from inside trading than their counterparts in No-R&D firms! are due to
trade intensity differences, rather than to information asymmetry. To exam-
ine this question we construct a sample of insider trades matched by trade
size and run the three-factor regressions ~1! on this sample. Specifically, for
each sample month ~1985 to 1997! we rank the combined R&D and No-R&D
samples ~separately for stock purchases and sales! by dollar value of trans-
action, and select adjacent R&D and No-R&D trades, where the difference in
trade value was smaller than 10 percent. These pairs of value-matched trans-
actions are inputs into the three-factor regression.21
The regression estimates of these trade-size-controlled insider transac-
tions are somewhat stronger than those in Panel B of Table II. The intercept
estimate for insider purchases is 0.011 ~t 5 2.00!, nearly identical to the a
estimate in Table II, and the intercept estimate for insider sales is 20.007,
t 5 22.27 ~compared with 20.005, t 5 21.85 in Table II!. We conclude, there-
fore, that the difference in insider gains between R&D and No-R&D firms is
not induced by differences in size ~value! of trades; rather managers of R&D
firms exploit the relatively large information asymmetry associated with R&D.

B. Insider Gains and Changes in R&D Expenditures

Research indicates that changes in firms’ R&D expenditures, once dis-
closed, are positively associated with stock returns ~e.g., Chan, Kesinger,
and Martin ~1992!, Lev and Sougiannis ~1996!, and Aboody and Lev ~1998!!.
Insiders, aware of changes in R&D budgets ahead of investors, are therefore
expected to act on such information, increasing share purchases prior to
disclosure of R&D increases and selling more than usual ahead of disclosure
of R&D decreases. This conjecture is examined here.
Firm-specific changes in R&D expenditures are computed as the percent-
age change in a given quarter’s R&D relative to R&D expenditures in the
preceding quarter ~there is no seasonality in quarterly R&D expenditures
and therefore adjacent quarters are used in the change computation!. Dur-
ing the sample period ~1985 to 1997!, quarterly R&D expenditures increased

20
We also find that the mean ~median! number of shares traded in each insider transaction
was 7,102 ~4,000! shares for firms with R&D activities and 5,693 shares ~2,000 shares! for firms
without R&D expenditures. Similarly, the value of shares traded was significantly higher for
firms with R&D than for No-R&D firms: The mean ~median! value of shares traded in each
transaction was $164,764 ~$50,000! for firms with R&D versus $140,399 ~$24,200! for firms
without R&D expenditures. Moreover, the number and value of trades is higher in R&D firms
than in No-R&D firms for both sale and purchase transactions.
21
The value match was close: The mean ~median! transaction value is $215,648 ~$48,000! for
R&D firms and $214,526 ~$47,787! for No-R&D firms.
 Information Asymmetry, R&D, And Insider Gains 2763

~decreased! in 63.1 percent ~36.9 percent! of the sample cases. Because many
of the quarterly R&D changes are rather small ~R&D time series are gen-
erally stable!, we rank the firm-specific R&D changes by size and focus the
analysis on the upper ~relatively large R&D increases! and lower ~large R&D
decreases! quartiles.
For firms in the upper quartile of R&D changes, insiders purchased over
the sample period a total of 40.0 million shares and sold a total of 91.4
million shares ahead of the disclosure of R&D increases in quarterly re-
ports.22 This purchase-to-sales ratio of 0.44 is substantially larger than the
overall sample purchase-to-sales ratio ~Table I! for R&D companies—0.26
~312 to 1,187 million shares!—suggesting enhanced insider purchases prior
to an R&D increase announcement. For the bottom quartile of firms ranked
by R&D changes ~relatively large R&D decreases!, insiders purchased 9.6
million shares and sold 138.9 million shares ahead of the quarterly report.
The ratio of purchase-to-sales ahead of R&D decreases—0.07—is substan-
tially smaller than the sample mean ~0.26!. It appears, therefore, that in-
siders increase share purchases ahead of disclosing R&D increases and enhance
the sale of shares ahead of R&D decreases.23
We also perform a portfolio analysis conditional on the direction of R&D
changes. Specifically, we form calendar-time portfolios for each sample month
~1985 to 1997, 155 months in total! of firms that have experienced an in-
crease in R&D and those whose R&D has decreased. We do this separately
for firms where insiders were predominantly purchasing shares during the
month ~RDPt ! and firms in which insiders were predominantly sellers of
shares ~RDSt !. For each firm, an average return ~over all insider transac-
tions during the month! is computed from the date of an insider transaction
to one day prior to filing with the SEC, similarly to the returns analyzed in
Table II. The Fama-French three-factor model is used to estimate excess
returns ~a!, where the dependent variable is the difference between returns
from share purchases and returns from the sale of shares ~RDPt 2 RDSt !,
for firms that increased R&D expenditures ~i.e., going long on firms in which
insiders were net purchasers and short on firms in which insiders were net
sellers!. For firms that have decreased R&D, the dependent variable is the
difference between returns from insider sales minus returns from insider
purchases ~RDSt 2 RDPt !.
Panel B of Table IV presents estimates of the three-factor model, condi-
tional on foreknowledge ~prior to public disclosure! of the change in R&D
expenditures. It is evident from the a coefficients that for R&D increases

22
The computation of shares purchased and sold by insiders is made for every fiscal quarter
t, in 1985 to 1997. The public disclosure of the change in R&D expenditures in quarter t,
relative to t 2 1, was made in quarter t financial report, released in quarter t 1 1.
23
The mean ~median! number of shares per insider transaction is also consistent with the
direction of R&D changes. For R&D increases, the mean ~median! number of shares in purchase
transactions were 17,019 ~5,000! versus 15,912 ~2,500! for sale transactions. For R&D de-
creases, the mean ~median! number of shares in purchase transactions were 10,003 ~1,000!
versus 29,000 ~15,000! for sale transactions.
2764 The Journal of Finance

Table IV
Returns Earned on Portfolios Formed on the Basis
of Changes in R&D Expenditures
Panel A presents percentage raw returns earned on portfolios formed as follows: For each month
between January 1985 and November 1997 we calculate, for each firm engaged in R&D, the
raw return for each insider transaction from the transaction date to one day prior to the SEC
filing date. We then calculate firm-specific mean ~median! returns for firms in which insiders
were net purchasers and for firms in which insiders were net sellers of shares. We compute
returns separately for firms that increased R&D and those that decreased R&D expenditures.
These mean ~median! returns are reported in Panel A. # . 0 refers to number of months ~out
of 155! in which returns were positive. In Panel B, the estimated coefficients of the three-factor
Fama-French regressions are presented. The dependent variable for R&D increases is the dif-
ference in excess returns ~over the interval between transaction and filing date! between a
portfolio of firms in which insiders were net purchasers of shares ~RDPt ! and a portfolio of
firms in which insiders were net sellers of shares ~RDSt !. For R&D decreases, the dependent
variable is reversed: RDSt 2 RDPt .

Panel A: Univariate Returns

Insider Purchases Insider Sales

Mean Median #.0 Mean Median #.0

Increase in R&D 5.55%*** 4.74%*** 129 1.53%*** 1.37%*** 98

Decrease in R&D 0.88%*** 0.70%* 88 21.28%*** 20.60%*** 68

Panel B: Three-factor Model

a RMt 2 RFt SMBt HML t Adjusted R 2

R&D increases
RDPt 2 RDSt 0.042*** 20.211 0.135 0.023 0.006
t-statistic ~8.33! ~21.59! ~0.70! ~0.09!
R&D decreases
RDSt 2 RDPt 20.022*** 0.053 0.034 0.034 0.001
t-statistic ~25.23! ~0.48! ~0.21! ~0.17!

* and *** denotes significance at the 10 and 1 percent levels, respectively.

~top row of Panel B!, going long on firms in which insiders were net pur-
chasers of shares and short on firms where insiders were net sellers yielded
an excess return of 4.2 percent over an average of 25 days between trans-
action and SEC filing date. For firms that have decreased R&D, the opposite
strategy would have yielded an excess return of 2.2 percent. Both excess
returns are highly statistically significant.
Note that the estimated excess returns ~a! in Table IV ~4.2 percent and
22.2 percent! are substantially larger than those in Table II ~1.1 percent and
20.5 percent!, indicating extra gains obtained by insiders possessing private
information about future changes in R&D budgets, relative to the gains based
on the “average information” possessed by insiders.
Finally, Panel A of Table IV presents mean ~median! returns to insider
trading in the expected direction ~e.g., purchasing during an R&D increase!,
as well as to “contrarians”—insiders that purchased shares ahead of an R&D
 Information Asymmetry, R&D, And Insider Gains 2765

decrease announcement and those that sold shares ahead of an R&D increase.
The contrarians’ mean gain from purchasing shares during a period of an R&D
decrease ~0.88 percent! is substantially lower than that of insiders who pur-
chased during R&D increases ~mean return of 5.55 percent!. The contrarians
who sold shares during R&D increases saw share prices gain 1.53 percent, on
average, after they sold their shares. We can only speculate that those con-
trarians were motivated by liquidity or portfolio diversification needs.

VI. Concluding Remarks

We examine insiders’ gain from trade focusing on a specific source of in-
formation asymmetry—firms’ R&D activities. R&D is unique ~firm-specific!
relative to other forms of capital. It is not traded in organized markets and
disclosure in corporate reports about the productivity and value of R&D
activities is deficient relative to the disclosure of tangible and financial as-
sets. Accordingly, we hypothesize, R&D activities contribute substantially to
the information asymmetry between managers and investors, and the for-
mer will tend to exploit this asymmetry to gain from insider trade.
We corroborate the hypothesis by providing evidence that insider gains in
R&D-intensive companies are significantly larger than insider gains in firms
not engaged in R&D. The R&D-related gains are both statistically and eco-
nomically meaningful. We also find that investors’ reaction to the public
disclosure of insider trades is significantly stronger for R&D companies than
for No-R&D companies, implying a larger information asymmetry in R&D
firms, and that the R&D-related private information is not fully revealed prior
to the public disclosure of insiders’ trade. Finally, we report that insiders ap-
pear to time their transactions according to the direction of change in R&D ex-
penditures, which is known to trigger investor reaction upon disclosure.
In this study we do not join the debate about the social consequences of
insider trading. Some consider insider trading beneficial to market effi-
ciency, and the loss to outside investors negligible, given the large volume
differences between inside and outside investors ~Jeng et al. ~1999!!. Others,
including Congress and the SEC ~which have, over the last six decades,
constructed an elaborate system of civil and criminal laws designed to re-
strict insider trading and the consequent profits!, are obviously concerned
with the consequences of insider trading. Such concerns extend beyond the
direct loss to outside investors, to include, for example, adverse effects on
managers’ incentives ~e.g., Fried ~1998!!. For those concerned with the con-
sequences of insider gains, our study points at an important and fast-
increasing source of private information leading to such gains—firms’ R&D
activities. Improved disclosure about R&D operations, such as the capital-
ization of development costs when products successfully pass technological
feasibility tests ~e.g., Aboody and Lev ~1998! for software companies! and the
timely release of information about planned changes in R&D expenditures
may be considered as means for mitigating the R&D-related information
asymmetry and the consequent insider gains.
2766 The Journal of Finance

REFERENCES
Aboody, David, and Baruch Lev, 1998, The value-relevance of intangibles: The case of software
capitalization, Journal of Accounting Research 36 supplement, 161–191.
Back, Kerry, Henry Cao, and Gregory Willard, 1999, Imperfect competition among informed
traders, Working paper, Washington University.
Bainbridge, Stephen, 1998, Insider trading, Encyclopedia of Law & Economics, forthcoming.
Barth, Mary, Ron Kasznik, and Maureen McNichols, 1998, Analyst coverage and intangible
assets, Working paper, Stanford University.
Brennan, Michael, and Avanidhar Subrahmanyam, 1995, Investment analysis and price infor-
mation in securities markets, Journal of Financial Economics 38, 361–381.
Carhart, Mark, 1997, On persistence in mutual fund performance, Journal of Finance 52, 57–92.
Chan, Su, John Kesinger, and John Martin, 1992, The market rewards promising R&D and
punishes the rest, Journal of Applied Corporate Finance 5, 59–66.
Chari, V. V., Ravi Jagannathan, and Aharon Ofer, 1988, Seasonalities in securities returns: The
case of earnings announcements, Journal of Financial Economics 21, 101–121.
Fama, Eugene, and Kenneth French, 1993, Common risk factors in returns on stocks and bonds,
Journal of Financial Economics 33, 3–56.
Finnerty, Joseph, 1976, Insiders and market efficiency, Journal of Finance 31, 1141–1148.
Fried, Jesse, 1998, Reducing the profitability of corporate insider trading through pretrading
disclosure, Southern California Law Review 71, 303–392.
Griliches, Zvi, 1995, R&D and productivity: Econometric results and measurement issues, in
Paul Stoneman, ed.: Handbook of the Economics of Innovation and Technological Change
~Blackwell, Oxford!.
Jaffe, Jeffrey, 1974, Special information and insider trading, Journal of Business 47, 410–428.
Jeng, Leslie, Andrew Metrick, and Richard Zeckhuser, 1999, The profits to insider trading: A
performance-evaluation perspective, Working paper 6913, National Bureau of Economic
Research.
Kothari, S. P., Ted Laguerre, and Andrew Leone, 1998, Capitalization versus expensing: Evi-
dence on the uncertainty of future earnings from current investment in PP&E and R&D,
Working paper, University of Rochester.
Krishnaswami, Sudha, and Venkat Subramaniam, 1999, Information asymmetry, valuation,
and the corporate spin-off decision, Journal of Financial Economics 53, 73–112.
Kyle, Albert, 1985, Continuous auctions and insider trading, Econometrica 6, 1315–1335.
Kyle, Albert, 1989, Informed speculation with imperfect competition, Review of Economic Stud-
ies 56, 317–356.
Lev, Baruch, and Theodore Sougiannis, 1996, The capitalization, amortization and value-
relevance of R&D, Journal of Accounting and Economics 21, 107–138.
Lorie, James, and Victor Niederhoffer, 1968, Predictive and statistical properties of insider
trading, Journal of Law and Economics 11, 35–51.
Seyhun, Nejat, 1986, Insiders’ profits, costs of trading, and market efficiency, Journal of Fi-
nancial Economics 16, 189–212.
Seyhun, Nejat, 1992, The effectiveness of the insider-trading sanctions, Journal of Law and
Economics 35, 149–182.
Stoll, Hans, 1978, The pricing of security dealer services: An empirical study of NASDAQ stocks,
Journal of Finance 33, 1153–1172.
Tasker, Sarah, 1998a, Bridging the information gap: Quarterly conference calls as a medium for
voluntary disclosure, Review of Accounting Studies 3, 137–167.
Tasker, Sarah, 1998b, Technology company conference calls: A small sample study, Journal of
Financial Statement Analysis 4, 6–14.
Titman, Sheridan, and Roberto Wessels, 1988, The determinants of capital structure choice,
Journal of Finance 43, 1–19.