Determining whether investments in information technology (IT) have an impact on ... accounted for 32.5% of all business capital equipment expenses in 1986.
The Impact of Information Technology Investment Announcements on the Market Value of the Firm Brian L. Dos Santos
Krannert Graduate School of Management Purdue University West Lafayette, Indiana 47907
Ken Peffers
School of Business-Camden Rutgers University Camden, New Jersey 08102
David C. Mauer
Graduate School of Business University of Wisconsin 1155 Observatory Drive Madison, Wisconsin 53706
Determining whether investments in information technology (IT) have an impact on firm performance has been and continues to be a major problem for information systems researchers and practitioners. Financial theory suggests that managers should make investment'decisions that maximize the value of the firm. Using event-study methodology, we provide empirical evidence on the effect of announcements of IT investments on the market value of the firm for a sample of 97 IT investments from the finance and manufacturing industries from 1981 to 1988. Over the announcement period, we find no excess returns for either the full sample or for any one of the industry subsamples. However, cross-sectional analysis reveals that the market reacts differently to announcements of innovative IT investments than to foUowup, or noninnovative investments in IT. Innovative IT investments increase firm value, while noninnovative investments do not. Furthermore, the market's reaction to announcements of innovative and noninnovative IT investments is independent of industry classification. These results indicate that, on average, IT investments are zero net present value (NPV) investments; they are worth as much as they cost. Innovative IT investments, however, increase the value of the firm. Event study—Information technology evaluation—Investment value—Information technology investments—Mariiet vaiue
1. Introduction nvestments by firms in information technologies (IT) have increased rapidly over the past three decades. According to Roach (1987), investments in IT hardware accounted for 32.5% of all business capital equipment expenses in 1986.' These investments have affected firms' products, services and internal processes. There
I
' This does not include expenditures on software and systems development. |O47-7O47/93/O4OI/0O0l/$O1.25 Copyright © 1993, The Institute of Management Sciences
Information Systems Research 4 : 1
1
Dos Santos • Peffers • Mauer
appears to be little doubt that today, information technologies (i.e., computer hardware, software and communication technologies) are extremely important to the smooth operation of most organizations. Firms in the banking and airline industries, for example, would be seriously affected if their IT-based systems failed even for a short time.^ It is not surprising, therefore, that managers and researchers have made plausible claims that investments in IT can have important strategic consequences forfirms:IT investment decisions have the potential to either improve a firm's competitive position or to allow the firm to become more vulnerable to competitive forces (Cash and Konsynski 1985, Ives and Learmonth 1984, Porter and Millar 1985). This literature suggests that IT investments have a significant impact on firm performance and, therefore, are of value to the firm. Unfortunately, empirical support for these claims consists almost exclusively of individual case studies (e.g., demons and Row 1988, Stoddard 1988). As a result, there is some doubt about whether the claimed impacts can be generalized from the individual cases to all firms. A number of recent empirical studies have suggested that IT investments do not benefit firms as much as the case studies might lead one to expect (Roach 1987, Osterman 1986, Loveman 1988, Bresnahan 1986). These studies suggest that in some industries and for the economy as a whole, money spent on IT would have been better spent in other ways. For example, after examining the productivity impact of IT in manufacturing between 1978 and 1984, Loveman (1988) concluded that the marginal dollar spent on IT would have been better spent on non-IT inputs to production. Baily and Chakrabarti (1988) concluded from their own study and studies conducted by others, that IT investments have not resulted in significant productivity gains. These studies cast doubt as to the real value of IT investments to firms. Determining whether IT investments can increase firm value poses many problems that are widely discussed in the information systems (IS) literature (Kauffman and Kriebel 1988, Strassmann 1988, DeLone and McLean 1992). Due to these problems, very few studies have attempted to link IT investments to firm performance.^ Of those that have, few have found a relationship (e.g.. Banker and Kauffman 1988, and Harris and Katz 1991). Frequently, the value of IT investments is inferred from user satisfaction, system usage, system quality, information quality, or impact on individual users (DeLone and McLean 1992), or from the direct impact of the IT application on performance of activities affected by it (Kekre and Mukhopadhyay 1993, Banker etal. 1990). When firms make IT investments, the investments resuh in some direct benefits that contribute to future cash flows. In addition, the investments may also have indirect benefits in the form of new investment opportunities for the firms. For example, investment in a new technology project may improve a firm's ability to use this new technology in future projects, thus affecting the firm's future investment opportunities (Dos Santos 1991). Financial theorists predict that managers make ^ For example, a 12-hour shut down of American Airline's SABRE System on May 12, 1989, left the airline without any infoimation about what customers had reservations on whichflightsor what seats were available. The airline was barely able to stay in operation by writing tens of thousands of tickets by hand, but many ofits customers, not to mention those of other airlines served by SABRE, were unable to reach their destinations or were late as a result of the information system failure (Salpukas 1989). ^ Kauffman and Weill (1989) and Dos Santos and Peffers (1993) provide recent reviews of these studies.
Information Systems Research 4 : 1
The Impact of Information Technology Investment
decisions that maximize the market value of the firm, where value is determined by the discounted value of future cashflowsexpected to be generated by assets already in place, plus the discounted value of investment opportunities that are expected to be available to the firm in the future. However, the value of potential future investments has been ignored, in practice and in research, because it is difficult to determine, both theoretically and practically (Mason and Merton 1985, Myers 1984). Hence, even if problems in measuring the direct benefits of IT investments are overcome, ex post determination of the effects of IT investments on firm performance tend to undervalue these investments. Adding to the problem, many direct benefits of IT investments are difficult to quantify and, therefore, are ignored (Strassman 1988). One way that this undervaluation of IT investments can be overcome is by determining how IT investments affect the value of the firm. If the net discounted cash flows that will result from an investment, the net present value (NPV), are positive, because the resulting direct and indirect benefits are expected to generate a return which is greater than the required rate of return, then the value of the firm should rise. This change in value will then be reflected in the market prices of the firm's securities. If the firm's securities are traded in an efficient market, this change in value will occur very quickly, allowing it to be observed and measured." This study addresses the question: do IT investments affect the market value of the firm? To answer this question, we analyze the impact of IT investment announcements on the common stock prices of publicly-traded firms. We use event-time methodology, which is a standard methodology in the accounting and finance literature (see, for example, Loderer and Mauer 1992). If the market responds by revaluing the firm's shares to reflect the information in the investment announcement, we can conclude that the announcement of an IT investment affects the market value of the firm. More precisely, if the market expects an IT investment to have a positive net effect on the value of the firm, then that value will be reflected in the market price of the firm's common stock. As a consequence, we can measure the market's assessment of the expected impact of IT investments on total firm value by examining stock price reactions around announcements of IT investments.' However, the market's reaction to IT investment announcements may depend on a number of factors. Industry characteristics, investment timing and the strategies of the firm are examples of factors that are thought to be among the determinants of IT investment value (Dos Santos and Peffers 1992, Cash et al. 1992). For example, in the early 1980s banks were thought to have more information-intensive processes than manufacturers (Porter and Millar 1985). As a result, IT investments may have different effects on firm value in the financial services industry than in the manufacturing industries. Therefore, this study will examine whether firm values in two major industry groups,financialservices and manufacturing, were affected differently by IT investments. " Under the semistrong form efficient markets model, for which empirical evidence is strong, market prices fully reflect publicly available information. Fama (1991) provides an excellent review of the extensive empirical research in this area. ' The event study methodology has the added advantage of being able to capture both the risk and return consequences of an IT investment since it focuses on the effect on the equity price, the change in which is an unbiased estimate of both the risk and return consequences of the proposed IT investment. In comparison, MIS studies which look at the effect of IT investment expenditures on accounting measures of performance, may not correctly capture the risk consequences of such investments.
March 1993
Dos Santos • Peffers • Mauer Other studies, based on economic theory, have suggested that innovative investments in IT and other technologies may result in greater rewards for investors than follow-up investments (see, for example, Mascarenhas 1992). Innovators, or first movers, may be able to obtain superior performance if they can capture favorable market positions, secure scarce resources, or benefit from learning how to restructure processes to take advantage of innovations before their competitors can imitate them. These issues are discussed in more detail later. In this study, we determine whether IT investment announcements affect firm value differently, depending on whether they are innovative or follow-up investments. For a sample of 97 IT investments over the period from 1981-1988, we find that stock price reactions to proposed IT investments are not significantly different from zero, either for the whole sample or for subsamples infinancialservices and manufacturing. These results indicate that, on the whole, IT investments have zero NPVs. That is, the average IT investment earns a rate of return which is just equal to the required rate of return, no more or less. In addition, industry classification has no affect on the value of IT investments. IT investments classified as innovative, however, result in positive changes in firm value, while the effect of other investments, including follow-up investments, is insignificantly negative. These results indicate that investments in innovative IT are expected to provide superior returns for firms. This paper is particularly noteworthy for two reasons: (1) To our knowledge, it is the first study to use a widely-accepted methodology, the event study, in research aimed at evaluating IT investments. The market-based event study methodology is extensively used in empirical research in the accounting and finance areas, is well accepted by researchers in these disciplines, and has a strong theoretical foundation. (2) It uses this methodology to examine the effeets on firm value of announcements of IT investments. The results indicate that, while IT investments do not, on the whole, systematically add value tofirms,firmsare rewarded for the risks incurred in making investments in innovative IT applications. The remainder of the paper is organized as follows. The next section outlines the sample selection procedure and describes the data set. Section three describes the methodology used to measure the stock price response to IT investment announcements. The results are presented in §4, and the last section provides a brief summary and conclusions. 2. Sample Selection Procedure and Data Description The objective of this study is to analyze the stock price behavior offirmsannouncing plans to invest in information technology. To accomplish this objective, we searched PR Newswire and PTS Prompt over the eight-year period from 1981 to 1988, for news articles about firms planning to make investments in information technology. Since these are large databases containing general business news articles, we restricted our search to investments in the financial services and manufacturing ' The choice of industrial groups is motivated by the widely-held belief that IT investments have a greater effect on firm performance in the financial services industry than in the manufacturing industry (Porter
Information Systems Research 4 : 1
The Impact of Information Technology Investment industries.* Standard Industrial Classification (SIC) codes were used to determine the type of industry.^ With this restriction, a search of the two databases produced news article titles for more than 4,500 possible IT investment announcements. The titles were reviewed to determine whether they were likely to represent news stories or news releases about specific IT investments, including purchases, agreements to purchase, or plans to buy equipment, software, or services. When there was more than one title on the same IT investment, e.g., one firm's planned purchase of computer equipment, the earliest title was retained. The sample was further reduced by discarding those titles that appeared to represent announcements by firms that were not traded on major security exchanges because of the availability of computerized daily common stock price data. These steps reduced the number of titles to 487 IT announcements by firms trading on either the NYSE, AMEX or NASDAQ stock exchanges. The full text of these 487 titles was then obtained and each news article was examined to determine whether it contained news of an IT investment. News stories about IT investments that did not appear to be the first public release of the information to the public, or which were duplicates of other announcements, were eliminated at this stage. In addition, only those announcements that dealt exclusively with IT investments were retained. For example, announcements that also contained information about other investment activity, such as mergers or acquisitions, were eliminated from further consideration. To determine the impact of new information on security prices, it is important to determine when the announcement was actually made because, in an efiicient market, the adjustment of prices to new information is almost immediate. Daily publications typically contain news that became available on the day before the publication date. Hence, for news appearing in a daily publication, we assumed that the information was available to investors on the day before the publication date, or on the publication date itself through the news article. When announcements appear in weekly, biweekly or monthly publications, it is difficult to determine the day on which an announcement was actually made. This, as we discuss later, makes it much more difficult to reject the null hypothesis of no abnormal returns. For this reason, announcements in publications with periodicity longer than one day were eliminated. After these data screens, the remaining sample included only firms that traded on either the NYSE or AMEX stock exchanges. These firms were then matched with data available on the University of Chicago's Center for Research in Security Prices (CRSP) daily common stock returns tapes. A few IT investment announcements had to be eliminated at this point because the return data necessary to use the event-study methodology (described later) was not available for the announcing firm. For the
and Millar 1985; Cash et al. 1992; Johnston and Canico 1988; Jarvenpaa and Ives 1990), and to enable comparison of our results with those studies which find modest support for such claims (Jarvenpaa and Ives 1990). ' Finance and insurancefirmshave SIC codes beginning with 60-63, and manufacturing firms have SIC codes beginning with 20-38. Firms that manufacture computer and related equipment, i.e., those with four-digit SIC code 3573, were excluded. We excluded the latterfirmsbecause the objective of the study is to determine the value of IT investments for firms that use information technology, rather than producers of the technology.
March 1993
Dos Santos • Peffers • Mauer
remaining announcements, the Wall Street Journal Index was checked for news that might contaminate the price data (e.g., dividends, earnings or other types of announcements) on the day before, the day of, and the day after the investment announcement date. Thefinalsample consists of 97 IT investment announcements by firms over the period from 1981 to 1988. Examples of abbreviated announcements are contained in Appendix A. Table 1 presents a breakdown of the announcements by industry type and the year in which the announcement appeared.* As reported there, 33 announcements concerned investments in manufacturing industries and 64 announcements were for investments in financial services industries. Of those in the manufacturing sample, the largest number of announcements (20) was by firms in the transportation equipment industry. In the finance industry group, the largest number of announcements (42) was by depository institutions. Each announcement was also carefully examined to determine whether the proposed IT investment would result in an information technology application that was innovative for the industry, and to ascertain the dollar amount of the investment expenditure, if available. Two of the authors independently categorized each investment as being either innovative or noninnovative, based on observation of claims made in the announcements. Prior to the categorization, it was agreed that the classification would be based on the following criteria. An investment would be classified as innovative if the announcement made a claim that: • the investment represented thefirstuse of a technology among firms competing in that industry, • the investment would result in a new product or service based on information technology, or • the investment would result in the development of new information technology for the industry (e.g., software with new applications). TABLE 1 Distribution of Announcements over Time
Year 1981 1982 1983 1984 1985 1986 1987 1988 Total
Manufacturing
0 3 1 3 12 6 3 5 33
Finance
Full Sample
1 3 1 0
1 6 2
20
32 17 18 18 97
II 15 13 64
3
* In the table, notice that there is a large jump in the number of announcements after 1984, with 1985 havmg the largest number of investment announcements of any year in the sample. The reason for the large jump in announcements after 1984 is that PR Newswire did not include news stories prior to 1985. We cannot explain the large number of announcements in 1985.
Information Systems Research 4 : 1
The Impact of Information Technology Investment
Announcements were to be classified as noninnovative ifthe announcement indicated that: • the firm was following investments already made by its competitors, or • the investment was intended to maintain an existing application. If an announcement could not be unambiguously classified as innovative or noninnovative from the text of the announcement, it was placed in an unclassified category. The two independent categorizations were then compared, and only when an announcement received the same classification by both authors was it given that final classification, i.e., innovative, noninnovative or unclassified. When an investment announcement was classified differently by the two authors, that announcement was placed in the unclassified category. There were only six such conflicting categorizations in the sample. In each of these six cases, one of the two authors had placed the announcement in the unclassified category (i.e., there were no instances where one classified an announcement as innovative while the other classified it as noninnovative). The result of this process was that for the full sample of 97 announcements, 25 were classified as innovative, 42 were classified as noninnovative, and 30 were unclassified. For example, in Appendix A the announcement by Inland Steel Company (January 24, 1985) is classified as innovative because it is claimed in the announcement that the investment will enable the firm to provide services that its competitors do not provide. In comparison, the announcement by Chemical Bank (November 4, 1986) is an example of a noninnovative IT investment because the investment was needed to maintain and enhance an existing system. Appendix B lists the date, source, firm and classification of the announcements in the sample. The size of an IT investment (relative to the firm's existing asset base) could determine the size of its impact on firm value. In the sample, 36 announcements included an estimate of the dollar amount of the proposed investment expenditure. Of these, frequently only a range was specified, and the cost estimates included only the costs of hardware and/or software that was to be acquired. Hence, the available estimates grossly understate the actual costs of providing a usable system because they fail to include a large portion of the costs of developing and implementing a new system. For the 36 announcements with cost estimates, the average was approximately $15 million and the median was $8 million. 3. The Statistical Methodology The impact of announcements of investments in information technologies on common stock prices is computed using event-study methods commonly employed in the accounting and finance literature (see, e.g., Patell 1976, Loderer and Mauer 1992).' The event of interest in this study is the announcement of an IT investment by a firm. If an announced IT investment contains new information, it should cause financial markets to revalue the firm. Determining whether these announcements affect a firm's stock price requires that we estimate what the firm's stock price would ' Brown and Warner (1985) provide a good discussion of the event-study methodology. Loderer and Mauer (1992) is a representative recent example of the use of this methodology. Appendix C provides an overview of the event-study methodology for IS researchers.
March 1993
Dos Santos • Peffers • Mauer have been had there been no announcement. We assume that daily common stock returns are described by the market model, which is based on the capital asset pricing model (CAPM).'" The market model is specified as follows: J^u = o!i + PiR^, + ti,,
where
(1)
Rj, = rate of return for firm / on day t; R^i = rate of return on the market portfolio on day /; a,, |8, = are market model intercept and slope parameters for firm /; and e,, = disturbance term with the usual OLS properties. The market model is estimated for each firm in the sample using 200 daily returns. The estimation period starts 201 days before the announcement date and ends 2 days before that date, defined as day 0. The estimated parameters, a, and |8,, and the realized returns on the CRSP value-weighted market index are used to predict normal returns around the event period: the day before and the day of the announcement." Prediction errors during the event period, i.e., deviations of realized returns from normal returns, are estimates of excess returns (also called abnormal returns). These estimated excess returns are unbiased estimates (expressed in return form) of changes in the market value of the firm during the event period, which are attributed to investors' reaction to the information contained in the event. Returns on the stocks in our sample and on the value-weighted market index are obtained from the CRSP Daily Stock Returns Eile. The excess return for the common stock of firm / on event day / is computed as AR,, = i?,, - (a,. + /3>^,).
(2)
The two-day excess return computed over days - 1 and 0, where - 1 is the day before the announcement in a daily periodical, and day 0 is the announcement day, is computed as CAR,= 2 AR,,.
(3)
(=-1
Thus, for a sample of A'^firms,the average two-day announcement effect is equal to CAR = -^ 2 CAR,.
(4)
The statistical significance of the excess returns for the sample is assessed by constructing a Z-statistic similar to the one developed in Loderer and Mauer (1992). Specifically, wefirstcompute the standardized excess return to stock / on event day t as '" This is the most widely used method to estimate the returns on afirm'sstock. Elton and Gruber (1987, pp. 261-279) provide an excellent discussion of the CAPM. " Most of the event studies in the finance and accounting literature use a two-day event window (for example, see Loderer and Mauer 1992 and McConnell and Muscarella 1985). A two-day event window is appropriate when the timing of the event under investigation can be accurately identified (Fama 1991). We chose to follow standard convention because we know the exact day on which the news article containing the proposed IT investment was first published. We do not use a longer window because the power of an event study to reject the null hypothesis of no abnormal returns is severely limited over longer intervals.
Information Systems Research 4 : 1
The Impact of Information Technology Investment
SAR,,= , ^ ^ " , VVar(AR)
Var(AR,) =
where
(5)
sh+-^
[ 200 and where s] is the residual return variance from the estimation of the market model over the 200 days before the announcement period; R^ is the mean return on the market index over the estimation period; and R^, is the return on the market index on day t in the estimation interval.'^ The performance of each stock over the time interval defined by [/„ ti] is measured by the cumulative standardized excess return, which is defined as
CSAR, = i
, ^^^"
.
(7)
r=(, % - t ^ + \ Assuming that daily excess returns are normal and independent through time, SAR,, and CSAR, have a student-/ distribution with 199 degrees of freedom. The Z-statistic used to assess the statistical significance of the excess return over the event interval [/,, /zl, for a sample of A'^ stocks, is therefore equal to Z = ]IN{CSAR),
where
CSAR = Y, 2 CSAR,.
(8) (9)
Under the null hypothesis of zero expected excess returns, Z is approximately unit normally distributed (see, e.g., Loderer and Mauer 1992). The event-study methodology assumes that the information contained in IT investment announcements is not anticipated by the market before the first public announcement in the press. Under this assumption, the excess returns methodology provides unbiased estimates of the market's assessment of the impact of IT investments on announcing firms' common stock prices.'^ However, if some or all of the announcements were partially anticipated by the market, the announcement period excess returns provide lower bound estimates of the impact of IT investments on common stock prices.
4. Results Table 2 reports average two-day excess returns around the announcement date of IT investments for the full sample of 97 announcements, the manufacturing and finance industry subsamples and the innovative and noninnovative IT investment '^ We assume that the return variance in the announcement period is the same as the return variance in the estimation period. This assumption was tested using the statistical procedure described in Lobo and Mahmood (1989). We found no significant difference in the variance across the two periods for the sample. '^ The joint-hypothesis problem, well known amongfinanceand accounting researchers, can be a problem in some event studies. The joint-hypothesis problem refers to the dilemma that in an event study, market efficiency (the speed of adjustment of prices to new information) must be tested jointly with an underlying asset pricing model which is used to predict normal returns (which are used to compute abnormal returns). As Fama (1991) notes (pp. 1601-1602), event studies which use daily data are not susceptible to the joint-hypothesis problem.
March 1993
Dos Santos • Peffers • Mauer TABLE 2 Average Two-Day Excess Returns Around the Announcement of IT Investments for the Full Sample, Manufacturing and Finance Industry Subsamples, and by Whether the Investment is Innovative, Not Innovative or Cannot be Classified for 97 Firms over the Time Period from 1981 to 1988°
Sample Category* Full Sample (97) Breakdown by Industry Manufacturing (33) Finance (64) Breakdown by Type of Investment Innovative (25) Not Innovative (42) Unclassified (30)
Average Excess Return (%)
Z-value"
Frequency of Positive Returns (%)
Z-value'*
0.09
0.21
47
-0.51
0.40 -0.08
0.65 -0.20
55 44
0.52 -1.00
1.03 -0.09 -0.46
1.75* -0.26 -0.90
64 43 37
1.40 -0.93 -1.46
" Two-day excess returns are computed over days - 1 and 0, where - 1 is the day before the announcement in a daily periodical, and day 0 is the announcement day. * Sample sizes are in parentheses. " Z-statistic to compute the significance of the average excess return over the two-day announcement window under the null hypothesis that the average excess return is zero. ''UP is the number offirmswith positive excess returns, and n is the sample size, this statistic is given by Z=IP- n(0.5)]/[«'"(0.5)]. Z is distributed A'(0, 1) for large samples under the null hypothesis that the expected value of f is n(0.5). * Significant at the 5% level.
classifications. In the table, notice that neither the full sample nor any one of the industry subsamples have announcement period excess returns that are significantly different from zero. Specifically, the two-day average cumulative excess return for the full sample is only 0.09%, while the corresponding average excess returns for the manufacturing and finance industry subsamples are 0.40% and -0.08%, respectively. According to the reported Z-values, these average excess returns are not statistically significantly different from zero.'"* These results indicate that investors do not perceive that IT investments within or across manufacturing and finance industries will earn rates of return greater than their costs of capital, and that, therefore, such investments are zero NPV investment projects. In particular, as is true for any planned capital expenditure, the NPV of an investment project is equal to the expected change in shareholder wealth. Accordingly, assuming a relatively efficient capital market, the moment a firm announces '•' To determine whether the market's assessment of the value of an IT investment is influenced by whether thefirmalso reports the dollar amount of the IT investment expenditure, we computed separately the average excess return for 36 announcements in which a dollar estimate of the IT expenditure was provided. For this subsample, the average excess return is also not significantly different from zero. We also estimated a regression with CAR as the dependent variable and the ratio of investment size to the market value of the equity of the firm (on day zero) as the independent variable. The coefficient estimate on size was not statistically significantly diflerent from zero at reasonable confidence levels. This is not surprising, given that the dollar estimates in the announcements comprise only a small fraction of the investment, and are not comparable across announcements.
'0
Informatioti Systems Research 4 : 1
The Impact of Information Technology Investment
plans to undertake an investment project, the market price of the firm's stock should reflect an unbiased estimate of the project's expected NPV. Finding a zero average excess return in our sample, therefore, implies that the average IT investment project has a zero expected NPV. These results are not surprising given recent empirical evidence on the impact of IT investments. Most studies report mixed support for the anticipated impacts of IT investments. For example, Venkatraman and Zaheer (1990) compared the performance of insurance agencies whose operations were electronically integrated with those of an insurance carrier, with a matched set of agencies who were not so integrated. The carrier firm decided to develop a proprietary network linking the firm and its agents. The results were mixed. Integration resulted in no significant effects on efficiency. The researchers observed positive effects, however, regarding agents' ability to generate new business. Dos Santos and Peffers (1993) review recent studies of the impact of investments in specific IT applications, as well as the impact of aggregate IT investments by firms, and Kauffman and Weill (1989) review earlier studies. None of the reviewed studies were able to link investments in specific IT applications to overall firm performance and, therefore, to the value of the firm. In the literature, it is often suggested that top level managers know little about the likely profitability of IT investments (Dos Santos 1991, Kauffman and Kriebel 1988, Strassmann 1988). The price of a firm's shares and, therefore, any adjustments in that price, is determined by demand and supply in the financial market. The event-study method assumes that any new information relevant to pricing those shares is impounded in an unbiased manner into the market price. It may be unreasonable to assume that the market knows more than does the insider/manager of the firm. Therefore, if as suggested in the literature, top level managers know little about the likely profitability of a proposed IT investment, it is probably a sure bet that the market does not have better information. That managers do not know whether some IT investments will generate positive future returns, is consistent with our finding of zero average abnormal returns (i.e., zero NPV) in our full sample of 97 investment announcements. The question that remains to be answered is: which projects, if any, are positive NPV projects? The zero average returns for the full sample obscure the market's assessment of innovative IT investments. As can be seen in Table 2, the 25 IT investment announcements that are classified as innovative have a statistically significant average excess return of 1.03%, and over 64% of the announcements have positive excess returns. In comparison, the 42 noninnovative IT investment announcements have a negative, but statistically insignificant, average excess return of —0.09%. Notice that the 30 unclassified investments also have predominantly negative excess returns. It is a well-accepted notion in economics that one way for a firm to earn higher returns is for it to seize new lucrative opportunities early (Schumpeter 1950), i.e., the early bird gets the worm. This suggests that early adopters of new IT applications earn economic rents and, therefore, would be expected to have positive NPV investments. Technological leaders may reap 'first-mover' benefits, such as reputations for technical leadership, preemption of market positions, favorable access to scarce resources, and cost or differentiation learning curve advantages, which may result in competitive advantages for the firm (Porter 1985). For industrial and consumer products, empirical studies indicate that early movers do reap most of the benefits of innova-
March 1993
11
Dos Santos • Peffers • Mauer tions (e.g., Urban et al. 1986, Lieberman and Montgomery 1988, Lielien and Yoon 1990, Mascarenhas 1992). Yet, although it has been claimed that only innovative IT applications result in competitive advantages (Keen 1988, Porter and Millar 1985), support for these claims is provided mostly by case studies (e.g., Clemons and Row 1988, Stoddard 1988). An exception is a recent study of the impact of ATMs in the banking industry, indicating that first-movers gained most of the benefits (Dos Santos and Peffers 1992).'^ Given that economic rents earned by innovators are predicted by theory and supported by empirical evidence for different types of innovations, it is not surprising that announcements of innovative IT investments are perceived by financial markets to add value to firms. The innovative IT investments were not disproportionately represented in either the manufacturing or finance industry subsamples. In particular, 10 such investment announcements were in the manufacturing industry subsample, and 15 were in the finance industry subsample. The two-day average excess returns for these two subsamples of innovative investments are 1.28% and 0.87%, respectively, and are not significantly different from each other. In contrast, the noninnovative investments were predominantly represented in the finance industry, which has 32 of the 42 IT announcements classified as noninnovative.'* This is not surprising since heavy investment in IT was common in the finance and insurance industries prior to 1981. Therefore, there were likely fewer opportunities for innovative IT applications in those industries. In contrast, heavy investments in IT are a more recent phenomenon in manufacturing industries. To determine the joint effect of IT investment type (innovative or noninnovative) and industry on the market's reaction to IT investment announcements, we estimate the following regression models: CAR, = 00 + e,INNOV,. + 02IND, + e,,
(10)
CAR, = (9o + e,INNl, + e2lNN2i + OjIND, + e,.
(11)
These two specifications model investment classification and industry group with indicator variables. In both, CAR, is the two-day average excess return for the rth firm's IT investment announcement. In (10), INNOV, is set equal to 1 if the rth firm's investment is classified as innovative and 0 otherwise; IND, is set equal to 1 if the /th announcing firm is in the manufacturing industry and 0 otherwise. The regression in (10) is estimated with the 67 announcements that could be classified as either innovative or noninnovative. In (11), two sets of indicator variables are employed for investment type, INNI, and INN2,. Of these, INNI, is set equal to 1 if the /th firm's investment is innovative and 0 otherwise; and INN2, is set equal to 1 if the rth firm's investment is noninnovative and 0 otherwise. The regression in (11) is estimated with the full sample of 97 announcements. " The study investigates the effects of early investments in automated teller machine technology (ATM) involving a sample of over 3,000 banks. The results indicate that many early investors (i.e., banks that invested in ATM technology between 1971 and 1979) were able to achieve comparative gains in market share and income (Dos Santos and Peffers 1992). " The distribution of the unclassified IT investment announcements is approximately evenly distributed across the manufacturing and finance industry subsamples.
12
Information Systems Research 4 : 1
The Impact of Information Technology Investment
tn (10), the intercept measures the average announcement eiFect for noninnovative investments in the finance and insurance industries, and in (11), the intercept measures the average announcetnent effect for unclassified investments in the finance and insurance industries. The investment type slope coefficients then measure the marginal impact on announcement period returns provided by the innovative indicator variable in the equation. We estimate both Equations (10) and (11) with and without the industry indicator variable (IND) to examine the marginal explanatory power of industry classification on announcement period returns. Estimation results are reported in Table 3, wherein the four columns correspond to regressions (10) and (11) with and without the industry indicator variable. In models 1 and 2 (regression (10)), observe that innovative investments are greeted by the market with a statistically significantly larger announcement effect than noninnovative investments. In addition, note that after controlling for type of investment, the industry indicator variable does not have significant additional explanatory power. Indeed, the inclusion of the industry indicator variable (Model 2) actually reduces the explanatory power of the regression; the F-static is no longer significant. Models 3 and 4, which also include the unclassified irivestment-type group, yield essentially the same conclusions. Observe, however, that the difference between the announcement effects for noninnovative investments and unclassified investments (the coefficient estimate on ttN2) is not significantly different from zero. One of the interesting results of this study is that innovative IT investments by firms in either of the two industry groups can be valuable. While firms in the finance and insurance industries may spend a larger portion of their revenues on IT than do firms in the manufacturing group, IT investments that increase the value of the firm can be found in either industry group. More importantly, it is the innovative IT investments that increase firm value, while follow-up investments are, at best, zero NPV investments.'^ These are importantfindingsfor two reasons. Thefirsthas to do with the risks of IT investments. Innovative IT applications are very risky investments. The costs can be high and the benefits are difficult to identify in advance. Moreover, innovative IT applications are not protected by patent or copyright laws, and frequently are easily copied by competitors. Consequently, it is difficult to justify investments in innovative IT applications. The evidence that investments in innovative IT applications increase firm value, supports the efforts of managers who have been espousing investments in innovative IT applications to gain competitive advantages. The second reason has to do with the opportunities that are available to use IT in innovative ways. Each new IT that is introduced providesfirmswith opportunities to use the technology to change the way the firm conducts its business. In the 1960s and 1970s, the technologies that were available provided firms in the finance and insurance industries with many valuable investment opportunities because IT enabled firms to reduce costs by automating clerical activities. More recently, however, new ITs have given firms in the manufacturing industries opportunities to use IT to change design and manufacturing practices, and to change the way in which these firms are linked to their customers and suppliers. Therefore, ITs have provided firms " It should be noted that ifthe size of the innovative projects in our sample, relative to the market value of the firm, were much larger than the noninnovative and unclassified projects, it could be that size alone (rather than project characteristics) might explain the results.
March 1993
13
Dos Santos • Peffers • Mauer TABLE 3 Estimated Coefficients from Regressing IT Investment Announcement Excess Returns on Type of Investment and Industry Variables The dependent variable is the two-day IT investment announcement excess return for the /th firm in the sample. The various explanatory variables are defined as follows. In Models 1 and 2, INNOV is a 0/1 indicator variable which is set equal to 1 if the ith firm's IT investment announcement is classified as innovative. These models include only IT investment announcements which could be classified as either innovative or noninnovative. In models 3 and 4, INNI and INN2 are 0/1 indicator variables which are set equal to 1 if the rth firm's IT investment announcement is classified as innovative or noninnovative, respectively. These models also include IT investment announcements which could not be classified as either innovative or noninnovative. Finally, IND is a 0/1 indicator variable which is set equal to 1 if the /th firm is in the manufacturing industry, r-values are in parentheses. Independent Variables
Model 1
Model 2
Model 3
Model 4
Intercept
-0.0009 (-0.24) 0.0112 (1.88)*
-0.0018 (-0.46) 0.0105 (1.73)*
-0.0046 (-1.00)
-0.0067 (-1.28)
INNOV INNI
0.0150 (2.18)* 0.0038 (0.62)
INN2 IND R' F-statistic Sample Size
0.051 3.52* 67
0.0040 (0.62) 0.057 1.93 67
0.051 2.54** 97
0.0t51 (2.20)* 0.0047 (0.76) 0.0047 (0.85) 0.058 1.93 97
* Significant at the 5% level. ** Significant at the 10% level.
in different industries with opportunities to make firm value-enhancing IT investments. Our results indicate that IT investments are not necessarily more important to firms in one industry, but that the value of IT investments may differ across different industries. It should be noted that the announcements included in our sample are voluntarily disclosed by firms. As a result, the sample may be biased in that it contains only those IT investments that firms wish to make known. Analysis of voluntary announcements has both a cost and a benefit. On the benefit side, there is presumably more information in a voluntary disclosure than in a mandatory disclosure. Of course, the cost is that our sample might be biased in the sense that it may contain only those announcements of IT investments which are profitable. The other half of an unbiased sample would be those firms who contemplated an IT investment, but realized after some study that such an expenditure would not be profitable and, therefore, did not announce their intentions to the market.'* Given this potential bias, we might '* This potential bias is not unique to IT investment announcements. In fact, most event studies in the finance literature involve voluntary announcements offinancingand/or investment policy decisions. The primary reason for the emphasis on voluntarily disclosed information is that there appears to be little new information in announcements that a firm is required to make. The small group of analysts that closely follows a firm (and has a major influence on the price of the firm's shares) is apparently able to obtain this information prior to the formal announcement. Therefore, the price of thefirm'sshares usually reflects the information in regular mandatory announcements prior to the formal announcement.
14
Information Systems Research 4 : 1
The Impact of Information Technology Investment expect a positive average abnormal return in the full sample of IT investments. That we find a zero average return instead, is evidence against such a bias in our sample. 5. Summary and Conclusions This paper presented the results of a study of the effects of announcements of IT investments on announcing firms' common stock pdces. Our results indicated that over the announcement period, there were no excess returns for either the full sample or for any one of the industry subsamples: namely, firms in the manufacturing and financial industries. Cross-sectional analysis revealed, however, that the market reacts differently to innovative IT investments in either industry, than to follow-up or other noninnovative IT investments. These results suggest that the value of IT investments to firms can vary, and that certain types of IT investments are of value to the firm. Investors apparently have concluded that innovative IT investments are valuable, while catch-up (or imitative) investments are, at best, zero NPV investments. Consequently, these results provide empirical support for the case study evidence linking innovative IT investments with competitive advantages for firms. Further testing to support thisfindingis necessary. In general, future studies should seek to refine the identification of characteristics of IT investments that are positively valued by financial markets. Do other characteristics of the investment, the firm, or the industry systematically affect the impact of IT investments on the market value of the firm? For example, does past experience in developing innovative applications matter? Further study may also be useful to determine whether the pattern of positive excess returns that we observe for innovative IT investments holds for other industries and for other time periods in the finance and manufacturing industry subsamples. Our results suggest that we should observe positive excess returns for innovative IT investments in any industry or period. Managers are expected to make decisions that increase the value of the firm. However, providing evidence that information technology investments are of value to the firm has proven to be extremely difficult. Most studies that attempt to determine the value of IT investments do not use widely accepted measures of investment value. For example, many studies focus on individual users and groups that use the system, by employing as dependent variables, measures such as user satisfaction and system use (DeLone and McLean 1992). User satisfaction is an appealing measure since IT professionals frequently have difficulty delivering systems that satisfy users, and it is often necessary to have satisfied users if there is to be a good chance that a positive return on the investment will be obtained. Having satisfied users, however, does not necessarily mean that the investment will increase firm value. For example, giving managers the information they want will typically lead to satisfied managers, but may have no impact on outcomes (Ackoff 1967). System use also is an appealing measure since an unused system will provide no return on investment. The fact that a system is used, however, does not necessarily mean that it will add value (Ackoff 1967). Other studies determine the impact of IT investments on the outcome of internal activities and processes, using dependent variable measures such as material costs, operating expenses and product defects (Harris and Katz 1991, Banker et al. 1990, Kekre and Mukhopadhyay 1993). The use of such measures has the advantage that the IT investment is expected to directly affect the performance of these activities and processes. The disadvantage is that these measures are difficult to directly link to changes in firm value. Therefore, a study that links IT investments to changes in the
March t993
15
Dos Santos • Peffers • Mauer value of the firm, albeit a more difficult task and not always possible, is desirable to provide more convincing evidence of IT investment value. Linking IT investments to firm performance is difficult because many factors other than the investment affect firm performance, and separating the effects of IT investments from other effects is extremely difficult. The event-study methodology can enable researchers to circumvent this problem by allowing for a direct estimate of the impact of an IT investment on firm value. Such estimates are clearly as good as (if not better than) other measures of IT investment performance that have been used in the extant literature. The event-study methodology could be useful in addressing other interesting management issues involving classes of investments or circumstances in which investments are made. Many case studies extol the virtues of specific kinds of IT investments (e.g., airline reservation systems), suggesting that they are valuable strategic investments for firms. If so, we would expect favorable (unfavorable) public announcements concerning these investments to positively (negatively) affect firms' security prices. In addition, acquisition offirmsin the IS services industry byfirmsin other industries is becoming increasingly common (e.g.. General Motors' decision to acquire EDS). Are these investments valuable? And, are these acquisitions more valuable than acquisitions of firms that are not in the IS services industry? Similar questions arise regarding decisions byfirmsto disinvest IT by outsourcing portions of their IT services (e.g., Kodak's decision to divest a portion of its information services department). The impact of these decisions on firm value should be of great interest to managers of information resources as well as to researchers interested in information systems management issues.* Acknowledgments. We wish to thank the Associate Editor and three anonymous reviewers for very helpful comments and suggestions. We also appreciate comments from Gerald Lobo, Claudio Loderer and Dennis Sheehan on earlier versions of the paper. This paper has benefited from presentations at the Third Workshop on Information Systems and Economics, New York University, December 1991, and seminars at the University of California at Irvine, Notre Dame University and Rutgers University. An earlier version of this paper was titled "The Value of Investments in Information Technology: An Event Study." * Jin Whang, Associate Editor. This paper was received on April 6, 1992, and has been with the authors 2j months for 2 revisions. Appendix A Representative Announcements of IT Investments in the Sample Announcement Date Source January 9, 1985 PR Newswire
16
Announcements (abbreviated) Coldwell Banker is installing computers that will let its realty agents obtain mortgage funds electronically. The system will enable mortgage lenders in any part of the country to make their money available through computer terminals in broker's offices. The realty agents and home buyers will be able to scan a number of loan programs within minutes, as well as determine how large a loan the consumer will qualify for. This technology will also let the buyer apply for a loan electronically and allow the realty agent to track progress of the application. Coldwell Banker will be thefirstto
Classification: Innovative/ noninnovative (reason) Innovative (new product/ service based on IT)
Information Systems Research 4 : 1
The Impact of Information Technology Investment Appendix A (cont'd)
Representative Announcements of IT Investments in the Sample Announcement Date Source
January 24, 1985 PR Newswire
March 28, 1985 American Banker
May 30, 1985 PR Newswire
June 3, 1985 PR Newswire
August 21, 1985 PR Newswire
October 14, 1985 American Metal Market
October 21, 1985 PR Newswire
March 1993
Announcements (abbreviated) provide these services through their own IS department and process the mortgages through their own Residential Mortgage Services subsidiary. Inland Steel Company announced today that they will establish the first fully integrated electronic customer communication system in the domestic steel industry. Serving the needs of both steel customers and processors, the new system will permit Inland to transmit to them all of the information associated with the purchase and delivery ofa steel order, including: price and delivery quotes, order entry, order change, materialrelease,order status, invoicing, shipment notification and test results. "Establishing this new communication network is part of a larger Inland initiative to improve our service to customers," said Robert J. Darnell, president. "The evolving steel market demands that we meet customers' needs for shorter lead times and fuller information of the status of their orders." Seattle First National Bank will buy 1,(KX) Apple Macintosh computers through 3/86. The purchase will include a number of Fat Macs, which have 512 kilobytes of RAM, as well as regular Macs, which have 128 kilobytes of RAM. The latter will be used largely for access to larger computers, while the former will also be used to provide access to spreadsheet, graphics and word processing to those who need it most. The Macs will be linked via the bank's computer network. Marine Midland Bank has signed a multi-million dollar contract with SEI Corporation that will make Marine thefirstbank to use SEI's Portfolio Management System.... "This is the first time that a commercial package is available that provides investment managers sophisticated decision support tools integrated with the enormous processing power of a trust accounting system.... This integrated system will make Marine an industry leader in the provision of investment products and services to the individual and institutional markets." According to Bruce S. Eagleson, senior vice president of First Fidelity Bank, "The use of digital image technology for check processing is thefirstsignificant breakthrough in 25 years. Banks currently use manually operated ink encoders and high-speed reader sorters, which were introduced around 1960.... This step will lead to other innovations utilizing laser optical disks, electronic reading of check amounts, image statements, and other technological advances." DuPont Co. has ordered a CRAY-1 computer system, the first supercomputer to be installed in the chemical industry. The system will be installed at DuPont's Experimental Station near Wilmington, Del., and will be used in a number of research areas, including life sciences, electronics, catalysis and polymer sciences. Chrysler will install automated tool management systems in its engine plant for $150 million. The systems will make use of an IBM computer system that is compatible with other computers in the Trenton plan. The systems will be capable of communicating directly with other computers in the plant. They will provide automatic tool offset information by downloading tool layouts and dimensions to a host computer. Mellon Bank will test a Bell Atlantic prototype network that should substantially reduce transaction costs for p>oint-of-sale credit card authorizations, the two companies announced today. Mellon's trial marks thefirstuse of basic packet-switching technology for pointof-sale credit transactions. "Financial institutions currently pay for an information-access system as though it were in use 100 percent of the time a user is on-line."
Classification: Innovative/ noninnovative (reason)
Innovative (first use in an industry)
Noninnovative
Innovative (first use in an industry)
Innovative (first use in an industry)
Innovative (first use in an industry)
Unclassified
Innovative (first use in an industry)
17
Dos Santos • Peffers • Mauer Appendix A (cont'd) Representative Announcements of IT Investments in the Sample Announcement Date Source December 5, 1985 American Metal Market
March 17, 1986 American Metal Market
August 25, 1986 PR Newswire
October 22, 1986 American Banker
November 4, 1986 American Banker
November 12, 1986 PR Newswire
December 15, 1986 American Metal Market
January 13, 1987 Minn. Star & Tribune
January 14, 1987 Wall Street Journal
April 9, 1987 PR Newswire
18
Announcements (abbreviated) Inland Steel let a $ 1 million plus contract to Scientific Services for a computer software system for the basic oxygen furnace shop at its • Indiana Harbor plant. Scientific systems said the software system will track, monitor and support all steel making operations in of the works' 2 BOF shops, and will assist operators in controlling furnace charges and temperatures. Dana let a $4.5 million order to Westinghouse Electric to install a computer integrated manufacturing system and automated electrical induction heat treating cell to produce constant-velocity joints for four wheel drive cars at a plant in Columbia SC, to open in early 1986. Westinghouse said the CIM system will automate Dana's full manufacturing process, from order entry to finished product, but will be simple to use and readily expandable. Bank One has entered into an agreement with Electronic Data Systems Corp. (EDS) to develop an integrated financial system for large financial institutions. "This is a major development, combining the expertise of a leading computer services company and a progressive, innovative bank to meet the specialized processing needs of thefinancialindustry." The system is designed with a focus on the current and future needs of the larger financial institutions. Donald L. McWhorter, president of Bank One, said, "We will create a product that will become the major thrust for financial institutions' data processing in the 199O's." EF Hutton is investing $20 million in 10,000 new NCR personal computer workstations to link its account executives. With AWE, an account executive can fetch real-time and historic data on commodities, options, stock, bonds and mutual funds, customizing it by color, layout, and format. The user can also send individual or broadcast messages to others on the network, get customer account information, research reports, Dow Jones news, and information about Hutton products. Chemical Bank has ordered 100 ATMs from Omron Financial Systems. Chemical had been using ATMs made by Olivetti, but Olivetti is planning to exit the ATM business in the US. Chemical will add new locations to its ATM network because the customer lines are too long, while replacing its old Docutel machines at existing locations. Fidelity Bank, N.A., has signed a contract with Data Architects, Inc. for the installation of that firm's proprietary software product, BESS. Fidelity Bank will be the 22nd major financial institution to install BESS. The BESS product is a fully integrated, multicurrency, automated wireroom and electronic funds transfer system with message switching capabilities. Pratt and Whitney Government Products has let a $12.5 million contract for a Cray X-MP/28 computer system to Cray Research. The supercomputer will be used on jet engine design projects, including the National Aerospace Plane propulsion system. Norwest will develop a data processing system for its 200 banks and branches vrith Electronic Data Systems. The system will give Norwest access to its customers' accounts at all its branches and will assist in the transition to full branch banking if the move is approved by the state legislature. Merrill Lynch has contracted with Automatic Data Processing to develop a computerized financial information system. The quote system will be delivered to all branches over a two-year period beginning in 1988. Automatic Data Processing has over 55,000 quotation terminals in use by over 1,000 clients. Lockheed Missiles and Space Co. acquired a CRAY X-MP/24 computer system to replace a CRAY-1 S/1000 computer system
Classification: Innovative/ noninnovative (reason) Unclassified
Unclassified
Innovative (investment in development of new technology)
Unclassified
Noninnovative
Noninnovative
Unclassified
Unclassified
Noninnovative
Noninnovative
Information Systems Research 4 : 1
The Impact of Information Technology Investment Appendix A {cont'd) Representative Announcements of IT Investments in the Sample Announcement Date Source
July 1, 1987 PR Newswire
August 26, 1987 American Banker
May 31, 1988 PR Newswire
August 10, 1988 American Banker September 21, 1988 PR Newswire
October 27, 1988 PR Newswire
Announcements (abbreviated) that had been in use since 1983. The new system will be used for design and analysis of missiles and spacecraft. Citibank, N.A., today signed an agreement with Cognitive Systems, Inc. to have the firm provide services in connection with the development of Artificial Intelligence (AI) software. Cognitive Systems specializes in the development of knowledge based, natural language software systems that allow people to communicate with computers in everyday language. The company is a recognized leader in thefieldof AI. American Express will test Hypercom's POS terminals in 10/87. Some 2,000-5,000 terminals may be ordered to replace the present card authorization terminals Some analysts are speculating that American Express may be planning new services and applications. According to A. Irato, American Express VP of transaction services, the company is exploring a variety of payment and information services. American Express (AE) is thefirstto offer Latin American on-line authorizations through service 800 toll-free network. Service 800 and AE have reached an agreement whereby AE will become the first company to order local toll-free numbers in Latin America for on-line credit card authorizations. Real time voice or data tollfree access to major credit card authorization centers provides an economic advantage to both the customer and credit card companies. "It saves time, increases sales and minimizes fraud." Bank One Indianapolis let a $ 1.6 million contract for 47 ATMs from NCR. The new machines will replace Docutel ATMs at the bank's branches and remote locations. Chrysler Motors will buy a CRAY X-MP/14se computer system. The system will be installed at Chryslers engineering center and will be used for computer-aided engineering applications such as structural analysis, design optimization, etc., in support of the design and testing of Chrysler's next generation vehicles. Currently 13 of the world's automotive manufacturers have installed similar systems. CONVEX Computer Corporation will install C210 and C120 supercomputers at Ford Motor Co. Ford will be using the supercomputers for engineering analysis to achieve higher quality and more cost-effective new car design. "Using CONVEX supercomputer capabilities, we're able to perform CAE in less time and at less cost," said Dr. Howard Crabb, manager of CAE functions at Ford.
Classification: Innovative/ noninnovative (reason)
Innovative (investment in development of new technology)
Innovative (new product/ service based on IT)
Innovative (first use in an industry)
Noninnovative Noninnovative
Noninnovative
Appendix B Sample Events No. 1 2 3 4 5 6 7 8 9 10
Date
Source
Firm
Innovative
04/10/81 03/22/82 05/07/82 06/09/82 07/22/82 11/03/82 11/03/82 03/15/83 09/26/83 09/06/84
Walt Street Journal Wall Street Journal Wall Street Journal American Banker New York Times Wall Street Journal Wall Street Journal Wall Street Journal Wall Street Journal New York Times
Aetna Life and Casualty Co. First Interstate Bancorp. Crocker National Corp. Citicorp New York Titnes Co. Ford Motor Company General Motors Corp. Ford Motor Co. Boeing Co. Chrysler Corp.
Not Innovative Unclassified Unclassified Unclassified Unclassified Not Innovative Innovative Not Innovative Innovative Unclassified
March 1993
19
Dos Santos • Peffers • Mauer Appendix B (cont'd) Sample Events No.
Date
Source
Firm
Innovative
II 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
10/05/84 12/12/84 01/09/85 01/18/85 01/24/85 02/25/85 02/27/85 03/05/85 03/28/85 04/25/85 05/15/85 05/30/85 06/03/85 06/12/85 07/01/85 07/09/85 07/15/85 08/01/85 08/21/85 09/11/85 09/12/85 09/24/85 09/26/85 10/07/85 10/11/85 10/14/85 10/17/85 10/18/85 10/21/85 11/12/85 12/02/85 12/04/85 12/05/85 12/09/85 01/08/86 03/05/86 03/17/86 03/17/86 04/03/86 05/08/86 07/07/86 08/20/86 08/25/86 10/02/86 10/22/86 11/04/86 11/10/86 11/12/86 12/08/86 12/15/86 12/31/86 01/08/87 01/13/87 01/14/87 01/28/87
Wall Street Journal Wall Street Journal PR Newswire PR Newswire PR Newswire PR Newswire PR Newswire PR Newswire American Banker PR Newswire American Banker PR Newswire PR Newswire PR Newswire American Metal Mkt PR Newswire American Metal Mkt Wall Street Journal PR Newswire PR Newswire New York Times PR Newswire PR Newswire PR Newswire PR Newswire American Metal Mkt New York Times PR Newswire PR Newswire PR Newswire American Metal Mkt PR Newswire American Metal Mkt American Metal Mkt PR Newswire American Banker American Metal Mkt American Metal Mkt PR Newswire PR Newswire PR Newswire PR Newswire PR Newswire PR Newswire American Banker American Banker American Metal Mkt PR Newswire PR Newswire American Metal Mkt American Banker American Banker Minneapolis Star Wall Street Journal American Banker
Chevron Corporation U S X Corp. Sears Roebuck & Co. Citicorp Inland Steel Indus. Inc. Security Pacific Corp. First Pennsylvania Corp. Mobil Corp. Bank America Corp. Sears Roebuck & Co. Bank America Corp. Marine Midland Banks Inc. First Fidelity Bancorp Mobil Corp. Boeing Co. Chase Manhattan Corp. Xerox Corp. Bank America Corp. DuPont De Nemours & Co. American Motors Corp. Merrill Lynch & Co. Inc. Bank America Corp. Exxon Corp. DuPont De Nemours & Co. American Express Co. Chrysler Corp. ITT Corp. Manufacturers Hanover Mellon Bank Corporation Briggs & Stratton Corp. Ford Motor Co. Lockheed Corp. Inland Steel Industries Inc. Grumman Corp. Bell Industries, Inc. Citicorp Chrysler Corp. Dana Corp. Exxon Corp. Bristol Myers Co. CIGNA Corp. Manufacturers Hanover Bane One Corp. Chemical Bank Corp. Hutten EF Group Inc. Chemical Bank Corp. Cincinnati Milacron Inc. First Fidelity Bancorp Citicorp United Technologies Corp. Citicorp Citicorp Nonvest Corp. Merrill Lynch & Co. Inc. Chemical Bank Corp.
Unclassified Innovative Innovative Unclassified Innovative Not Innovative Not Innovative Innovative Not Innovative Not Innovative Not Innovative Innovative Innovative Not Innovative Innovative Unclassified Innovative Not Innovative Innovative Not Innovative Not Innovative Unclassified Unclassified Unclassified Unclassified Unclassified Not Innovative Innovative Innovative Not Innovative Unclassified Unclassified Unclassified Not Innovative Not Innovative Unclassified Unclassified Unclassified Not Innovative Innovative Innovative Not Innovative Innovative Not Innovative Unclassified Not Innovative Unclassified Not Innovative Innovative Unclassified Unclassified Not Innovative Unclassified Not Innovative Not Innovative
20
Information Systems Research 4 : 1
The Impact of Information Technology Investment Appendix B (cont'd) Sample Events No.
Date
Source
Firm
Innovative
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
01/29/87 02/18/87 03/18/87 03/30/87 04/09/87 04/13/87 07/01/87 07/15/87 08/14/87 08/25/87 08/26/87 09/01/87 10/05/87 11/11/87 01/29/88 02/04/88 03/23/88 04/04/88 04/18/88 05/31/88 07/13/88 07/26/88 08/10/88 08/10/88 08/18/88 09/21/88 09/22/88 10/05/88 10/05/88 10/05/88 10/27/88 12/14/88
Wall Street Journal PR Newswire PR Newswire PR Newswire PR Newswire PR Newswire PR Newswire American Banker PR Newswire PR Newswire American Banker PR Newswire PR Newswire PR Newswire PR Newswire PR Newswire Greensboro News PR Newswire PR Newswire PR Newswire PR Newswire PR Newswire American Banker American Banker Greenville News PR Newswire PR Newswire PR Newswire PR Newswire American Banker PR Newswire News Release
Citicorp General Motors Corp. Atlantic Richfield Co. Bankers TR N Y Corp. Lockheed Corp. General Dynamics Corp. Citicorp Bank America Corp. Citicorp Chubb Corp. American Express Corp. Citicorp General Motors Corp. Morgan JP c& Co. Inc. Chrysler Corp. Ford Motor Co. First Union Corp. United Jersey Banks Norwest Corp. American Express Co. American General Corp. Nynex Corp. Bane One Corp. Security Pacific Corp. First Union Corp. Chrysler Corp. Citicorp Chrysler Corp. First Union Corp. First VA Banks Inc. Ford Motor Co. Citicorp
Not Innovative Unclassified Innovative Unclassified Not Innovative Unclassified Innovative Not Innovative Not Innovative Not Innovative Innovative Unclassified Innovative Not Innovative Innovative Not Innovative Not Innovative Not Innovative Not Innovative Innovative Unclassified Innovative Not Innovative Not Innovative Innovative Not Innovative Not Innovative Not Innovative Not Innovative Not Innovative Not Innovative Unclassified
Appendix C An Overview of the Event-study Methodology An event study seeks to determine the effect on common stock value of an announcement (the "event"). An announcement is information supplied to the market (typically published in the press, e.g., The Wall Street Journal) by the managers of a firm, analysts who have investigated the firm, or from other individuals or agents who follow the industry, the market, or general economic conditions. The information may regard investment decisions, such as those examined in this paper, or any other information relevant to the firm's success or failure, e.g., dividend and earnings announcements, actions taken by afirm'scompetitors, etc. The^r^; step in an event study is to identify the event, in this case, announcements of IT investments. The second step is to define an event period. In many applications this is centered on the announcement date, which is designated as day 0 in event time. If the event of interest can be accurately identified, the event period usually encompasses only the day before (day - 1 in event time) and the date of the announcement. The reason the event period also includes the day before the announcement is that news published in daily periodicals usually is publicly known the day before (e.g., a firm announces that it will outsource its operations to IBM on Monday and the papers report the news on Tuesday). Note that although day 0 is the date the announcement is made by a firm in event time, the actual calendar date of the event under investigation may differ across firms in the sample. The third step is to compute a predicted (or normal) stock return, /?„, for each day in the event period for each firm. The predicted stock return is the rate of return which stockholders would earn if the event did not take place. The most widely used technique for computing predicted returns is the market model, which explicitly takes account of the risk of the firm's stock relative to the market. The market model is
March 1993
21
Dos Santos • Peffers • Mauer computed for each firm by estimating a time-series regression of the firm's returns against returns on the market over a period of time which does not include the event period (e.g., day -201 to day - 2 in event time). Formally, Equation (1) in the paper specifies the relationship between the return for firm / on day;, Rj,, and the return on the market on day /, /?„,. The predicted return for a firm on day nn the event period is then computed as R, = a, + ft/?™,, where a, and ft are the estimated market model regression coefficients, and /?„, is the return on the market for an actual day in the event period. The fourth step is to compute the excess (or abnormal) rate of return, AR,,, for each day in the event period for each firm. The excess return is the actual rate of return on that day minus the predicted return, AR,, = Rj, - Rj,. As such, it measures the effect of the announcement on the market value of the firm on day; in the event period. For each firm in the sample, the excess returns are summed over the event period to produce a cumulative excess return, specified by Equation (3) in the paper. The fifth step is to organize and group these cumulative abnormal returns according to the study objectives. In this study the groups were: the full sample offirms,a breakdown of the sample by type of industry, and a breakdown of the sample by type of investment. The cumulative abnormal returns are then averaged across the various groups to produce measures of the average effect of the announcement on the market value of the firm. Equation (4) in the paper illustrates the calculation. Standard statistical tests can be performed to determine the significance of the cumulative abnormal returns for the full sample, or any subsample. (See §3 of the paper for details.) In addition, by specifying appropriate regression models, cross-sectional analysis can be conducted to determine how various factors affect CAR. With CAR as the dependent variable, one can determine the effects of different independent variables on CAR (e.g.. Equations (10) and (11)). Such cross-sectional analysis could enable IS researchers to determine how various factors affect the market's evaluation of different IT decisions.
References Ackoff, R. L., "Management Misinformation Systems," Management Sci., 14, 4 (1967), 147-156. Baily, M. N. and A. Chakrabarti, Innovation and the Productivity Crisis, Brookings Institution, Washington, DC, 1988. Banker, R. D. and R. J. Kauffman, "Strategic Contributions of Information Technology: An Empirical Study of ATM Networks," Proceedings ofthe Ninth International Conferencefor Information Systems, Minneapolis, MN, 1988, 141-150. , and R. C. Morey, "Measuring Gains in Operational Efficiency from Information Technology: A Study of the Positran Deployment at Hardee's Inc.," / Management Information Systems, 7, 2 (1990), 29-54. Bresnahan, T. F., "Measuring the Spillovers from Technical Advance: Mainframe Computers in Financial Services," American Economic Rev., 76, 4 (1986), 742-755. Brown, S. J. and J. B. Warner, "Using Daily Stock Returns: The Case of Event Studies," J. Financial Economics, 14(1985), 3-31. Cash, J. I., Jr. and B. Konsynski, "IS Redraws Competitive Boundaries," Harvard Business Rev., MarchApril (1985), 134-142. , F. W. McFarlan and J. L. McKenney, Corporate Information Systems Management: The Issues Facing Senior Executives, Third Edition, Irwin, Homewood, IL, 1992. demons, E. K. and M. C. Row, "McKesson Drug Company: A Case Study of Economost-A Strategic Information System," / Management Information Systems, 5, 1 (Summer 1988), 36-50. DeLone, W. H. and E. R. McLean, "Information Systems Success: The Quest for the Dependent Variable," Information Systems Res., 3, 1 (March 1992), 60-95. Dos Santos, B. L., "Justifying Investments in New Information Technologies," J. Management Information Systems, 1, 4 (Spring 1991), 71-89. and K. Peffers, "Rewards to Investors in Innovative Information Technology Applications: A Study of First Movers and Early Followers in ATMs," Working Paper, Krannert Graduate School of Management, Purdue University, West Lafayette, IN 47907, 1992. and , "Firm Level Performance Effects: A Framework for Information Technology Evaluation Research," in Strategic Information Technology Management: Perspectives on Organizational Growth and Competitive Advantage, R. D. Banker, R. J. Kauffman and M. A. Mahmood (Eds.), Idea Group Publishing, Harrisburg, PA, 1993, 515-546.
22
Information Systems Research 4 : 1
The Impact of Information Technology Investment Elton, E. J. and M. J. Gruber, Modern Portfolio Theory and Investment Analysis. John Wiley & Sons, New York, 1987. Fama, E. F., "Efficient Capital Markets: U" Journal of Finance, 46, 5 (December 1991), 1575-1617. Harris, S. E. and J. L. Katz, "Firm Size and the Information Technology Investment Intensity of Life Insurers," MIS Quarterly, 15 (1991), 333-352. Ives, B. and G. P. Learmonth, "The Information System as a Competitive Weapon," Communications of the ACM, 27, 12 (December 1984), 1193-1201. Jarvenpaa, S. K. and B. Ives, "Information Technology and Corporate Strategy: A View from the Top," Information Systems Research, 1,4(1990), 351-376.
Johnston, H. R. and S. R. Carrico, "Developing Capabilities to Use Information Strategically," MIS Quarterly, 12, 1 (March 1988), 153-165. Kauffman, R. J. and C. H. Kriebel, "Measuring and Modeling the Business Value of IT," in Measuring Business Value ofInformation Technologies, ICIT Research Study Team #2 (Ed.), ICIT Press, Washington, DC, 1988, 93-120. and P. Weill, "An Evaluative Framework for Research on the Performance Effects of Information Technology Investment," in Proceedings of the Tenth International Conference on Information Systems, Boston, MA, 1989, 377-388. Keen, P. G. W., Competing in Time Using Telecommunications for Competitive Advantage, Ballinger, Cambridge, MA, 1988. Kekre, S. and T. Mukhopadhyay, "Impact of Electronic Data Interchange Technology on Quality Improvement and Inventory Reduction Programs: A Field Study," / Manufacturing and Operations Management (1993), in press. Lieberman, M. B. and D. B. Montgomery, "First Mover Advantages," Strategic Management J.,9(l 988), 41-58. Lilien, G. I. and E. Yoon, "The Timing of Competitive Market Entry: An Exploratory Study of New Industrial Products," Management Sci., 36, 5 (1990), 568-585. Lobo, G. J. and A. A. W. Mahmood, "Relationship Between Different Amounts of Prior Information and Security Return Variability," / Accounting Research, 27, 1 (Spring 1989), 116-134. Loderer, C. and D. C. Mauer, "Corporate Dividends and Seasoned Equity Issues: An Empirical Investigation," J. Finance, 47, 1 (March 1992), 201-225. Loveman, G. W., "An Assessment of the Productivity Impact of Information Technologies," Working Paper, Department of Economics, Massachusetts Institute of Technology, Cambridge, MA, June 1988. Mascarenhas, B., "First Mover Effects in Multiple Dynamic Markets," Strategic Management J 13 (March 1992), 237-243. Mason, S. P. and R. C. Merton, "The Role of Contingent Claims Analysis in Corporate Finance," in Recent Advances in Corporate Finance, E. I. Altman and M. G. Subramanyam (Eds.), Richard D. Irwin, Homewood, IL, 1985. McConnell, J. J. and C. J. Muscarella, "Corporate Capital Expenditure Decisions and the Market Value of the Firm,"/ Financial Economics, 14 (1985), 399-422. Myers, S. C, "Finance Theory and Financial Strategy," Interfaces, 14, 1 (Jan-Feb. 1984), 126-137. Osterman, P., "The Impact of Computers on the Employment of Clerks and Managers," Industrial and Labor Relations Rev., 39, 2 (January 1986), 175-186. Patell, J. M., "Corporate Forecasts of Earnings per Share and Stock Price Behavior: Empirical Tests," Journal ofAccounting Res., Autumn (1976), 246-276. Porter, M. E. and V. E. Millar, "How Information Gives You Competitive Advantage," Harvard Business /Jev, July-August (1985), 149-160. Porter, M. E., Competitive Advantage, The Free Press, New York, NY, 1985. Roach, S. S., America's Technology Dilemma: A Profile of the Information Fconomy, Special Economic Study, Morgan Stanley and Co., April, 1987. Salpukas, A., "Computer Chaos for Air Travelers," The New York Times. May 13, 1989. Schumpeter, J. A., Capitalism, Socialism, and Democracy. Harper, New York, 1950. Stoddard, D., Otisline, Harvard Business School Case Study, #9-186-304, Cambridge, MA, 1988. Strassmann, P. A., "Management Productivity as an IT Measure," in Measuring Business Value of Information Technologies. ICIT Research Study Team #2 (Ed.), ICIT Press, Washington, DC, 1988, 93120. Urban, G. L., T. Carter, S. Gaskin and Z. Mucha, "Market Share Rewards to Pioneering Brands: An Empirical Analysis and Strategic Implications," Management Sci., 32, 6 (1986), 645-659. Venkatraman, N. and A. Zaheer, "Electronic Integration and Strategic Advantage: A Quasi-Experimental Study in the Insurance Industry," Information Systems Res.. 1, 4 (1990), 377-393.
March 1993
23