FINANCE/ACCOUNTING

QUANTITATIVE METHODS/MIS

MANAGEMENT/MARKETING

Spring 1996 Vol. 3, No. 3

Vincent P. Apilado Mean-Reverting State Variables as a Factor in Mean-

Gary M. Richardson Reverting Stock Returns

Sang-Hoon Kim The Relationship Between the Price Interest Rate Risk and the Holding Period Return in a Bond Investment

Maria Kathleen Boss American and German Regulation of Insider

Cheryl A. Cruz Trading: A Comparison Cornelia Alsheimer-Barthel

Lara Preiser-Houy Assessing the Payoff from an Information Technology Infrastructure: A Multi-Phased Approach

Dorothy M. Fisher Using Data Envelopment Analysis to Evaluate Tax

Steven A. Fisher Preparation Software D. Bruce Sun

Ugur Yavas Dimensionality, Reliability and Validity of

Zeynep Bilgin SERVQUAL

JOURNAL OF BUSINESS AND MANAGEMENT

INNOVATIVE EDUCATION

Karen L. Fowler Experiential Learning in the Capstone Strategic Donna M. Scott Management

Course: Collaborative Problem Solving, the Student Live Case, and Modeling

Published jointly by the Western Decision Sciences Institute and the School of Management, California State University, Dominguez Hills

JOURNAL OF BUSINESS AND MANAGEMENTTHE OFFICIAL PUBLICATION OF THE

WESTERN DECISION SCIENCES INSTITUTE (WDSI)

The Decision Sciences Institute is a professional society dedicated to the development and application of quantitative and behavioral methods to administrative problems. Most functional areas of business are represented among the membership. Through its journals, national and regional meetings, and other activities, the Decision Sciences Institute serves as a vehicle to advance and disseminate the theory, application, pedagogy, and curriculum development of the decision sciences.

National PresidentBetty J. Whitten, University of Georgia

National President-ElectJames R. Evans, University of Cincinnati

Western Regional Officers 1995-96President, George A. Marcoulides, California State University, FullertonPresident-Elect, Thomas E. Callarman, Arizona State UniversityVice President for Programs, Richard Jenson, Utah State UniversityProgram Chair-Elect, Karen L. Fowler, University of Northern ColoradoVice President for Member Services, Marc F. Massoud, ClaremontMcKenna CollegeSecretary/Treasurer, Howard R. Toole, San Diego State University Executive Secretary, Helen Beaver, San Diego State University

JOURNAL OF BUSINESS

AND MANAGEMENT

VOL. 3, No. 3 Spring 1996

EDITORS

EDITORIAL ASSISTANT

Franklin StrierBurhan F. YavasPatty Ramirez

Editorial Offices:

JOURNAL OF BUSINESS AND MANAGEMENTSchool of ManagementCalifornia State University, Dominguez Hills1000 East Victoria StreetCarson, California 90747Phone: (310) 243-3472, (310) 243-3501Fax: (310) 516-3664, (310) 217-6964

Published jointly by Western Decision Sciences Institute (WDSI) and the School of Management, California State University, Dominguez Hills. The purpose of the JOURNAL OF BUSINESS AND MANAGEMENT is to provide a forum for the dissemination of contributions in all fields of business, management and related public policy of relevance to academics and practitioners. Original research, reports and opinion pieces are welcome. The style should emphasize exposition and clarity, and avoid technical detail and jargon.

The views expressed in articles published are those of the authors and not necessarily those of the Editors, Executive Board, Editorial Board, WDSI or California State University, Dominguez Hills. All submissions will be reviewed initially by the editors and, if judged appropriate, will be sent to knowledgeable referees for review. The authors assume responsibility for the accuracy of facts published in the articles.

Copyright 1996 WDSI and by the School of Management, California State University, Dominguez Hills. Subscriptions are $16/year. Manuscripts should be double-spaced and submitted in triplicate. Manuscripts and comments should be directed to the editors.

JOURNAL OF BUSINESS AND MANAGEMENT

Executive Board George A. Marcoulides, President, WDSI Thomas E. Callarman, President-Elect, WDSI Yoram Neumann, Vice President, University Advancement and Dean, School of Management, California State University, Dominguez

Hills Franklin Strier, Editor Burhan F. Yavas, Editor

Editorial Board Dr. Joseph R. Biggs California Polytechnic State University, San Luis Obispo Dr. Henry Brehm University of Maryland Dr. Terry E. Dielman Texas Christian University Dr. Moshe Hagigi Boston University Dr. Ronald H. Heck University of Hawaii at Manoa Dr. Richard C. Hoffman Salisbury State University, Maryland Dr. Marc T. Jones University of Otago, Dunedin, New Zealand Dr. Erdener Kaynak Pennsylvania State University Dr. Thomas Kelly State University of New York, Binghamton Dr. George R. LaNoue University of Maryland Dr. John Preble University of Delaware Dr. Arie Reichel Ben-Gurion University of the Negev, Israel Dr. Elizabeth L. Rose University of Southern California Dr. Anne S. Tsui The Hong Kong University of Science and Technology, Hong Kong Dr. Michael Useem University of Pennsylvania

Reviewer Acknowledgements

The editors of the Journal of Business and Management wish to express their appreciation to the following individuals who have reviewed manuscripts submitted for consideration in this issue of the Journal of Business and Management.

Dr. Shirley AndersonDr. Harvey ArbelaezDr. Hamdi BiliciDr. Martin BlynDr. Roger DearDr. Zvi DreznerDr. Mohamed El-BadawiDr. Chic FojtikDr. Raoul FreemanDr. Robert H. GirlingDr. Abbas HeiatDr. John KarayanDr. Richard MalamudDr. James MartinoffDr. Richard NehrbassDr. Larry PressDr. Elizabeth RoseDr. Golnaz SadriDr. Peter A. SchneiderDr. Mark G. Simkin

JOURNAL OF BUSINESS AND MANAGEMENT

TABLE OF CONTENTS

From the Editor's Desk............................................................................7

Mean-Reverting State Variables as a Factor in Mean-Reverting Stock Returns

Vincent P. Apilado and Gary M. Richardson..................................... 9

The Relationship Between the Price Interest Rate Risk and the Holding Period Return in a Bond Investment

Sang-Hoon Kim.................................................................................27

American and German Regulation of Insider Trading: A ComparisonMaria Kathleen Boss, Cheryl A. Cruz andCornelia Alsheimer-Barthel...............................................................36

Assessing the Payoff from an Information Technology Infrastructure: A Multi-Phased Approach

Lara Preiser-Houy........................................................................... 53

Using Data Envelopment Analysis to Evaluate Tax Preparation SoftwareDorothy M. Fisher, Steven A. Fisher and D. Bruce Sun....................80

Dimensionality, Reliability and Validity of SERVQUALUgur Yavas and Zeynep Bilgin..........................................................92

Experiential Learning in the Capstone Strategic Management Course: Collaborative Problem Solving, the Student Live Case, and Modeling

Karen L. Fowler and Donna M. Scott..............................................103

FROM THE EDITOR'S DESK

Articles for this issue were selected from the Western Decision Sciences Institutes 25th Annual Conference held in Seattle on April 2-6, 1996.

The theory of stock market efficiency posits that investors compete in using public information and in so doing, eliminate its value for earning additional returns. VINCENT P. APILADO and GARRY M. RICHARDSON investigate empirically the link between mean reversion in stock returns and market efficiency. Using three different methodologies (regression, variance ratios and Markov chain) on eight variables the authors find that their results are sensitive to the method used. Also, mean reversion is not unique to stock returns and may result from rational responses of investors to changing market conditions.

The interest rate risk of bond investments is divided into the price risk and the holding period return risk. SANG-HOON KIM derives a measurement for the holding period return risk in a bond investment based on the elasticity of the compound (terminal) value. The results reveal a trade off between the two types of risks: reducing or eliminating one risk can be achieved at the expense of increasing the other risk.

As the European Union (EU) gains momentum, it becomes important for investors to understand EU/US differences in approaches to regulation of insider trading. Since the most influential member of the EU is Germany, MARIA K. BOSS, CHERYL A. CRUZ and CORNELIA ALSHEIMER-BARTHEL compare and contrast the different approaches taken by Germany and the United States to regulate insider trading, with particular attention paid to the tax consequences.

An approach to assess the economic viability of infrastructure resources is presented by LARA PREISER-HOUY to contribute to our understanding of the impacts of such investments on organizational performance. The three-phase approach is aimed at helping Information

8

Systems and general managers make more informed infrastructure investment decisions.

Advances in information technology have profoundly changed the preparation of income tax returns. Many tax softwares have appeared in the marketplace. DOROTHY M. FISHER, STEVEN A. FISHER and D. BRUCE SUN use Data Envelopment Analysis to analyze and compare the performance of 18 professional tax preparation software packages. The results should help professional accountants reduce time and cost and improve their selection process.

The past decade witnessed rapid internationalization of business. Since markets in many industries are ever more integrated worldwide, methodological issues surrounding cross-national research are gaining attention. Can a research instrument prepared in one country be valid in another? UGUR YAVAS and ZEYNEP BILGIN examine the dimensionality, reliability and validity of a research measure developed in the United States (SERVQUAL) in the Turkish setting. They find that while the widely used service quality measure has acceptable reliability and validity, caution should be exercised when using the instrument.

KAREN L. FOWLER and DONNA M. SCOTT present an application of experiential learning theory to the capstone strategic management course. Traditional methods of instruction in the capstone course are compared and contrasted with the experiential approach. The article concludes with suggestions for future research which include the comparison of traditional feedback approach with the use of modeling.

FRANKLIN STRIER

BURHAN F. YAVAS

9

F

MEAN-REVERTING STATE VARIABLES AS A

FACTOR IN MEAN-REVERTING STOCK RETURNS

Vincent P. Apilado *

Gary M. Richardson **

This paper tests for mean reversion in macroeconomic and fundamental variables. We also contrast results derived from alternative methodologies. Tests of mean reversion using OLS regression, variance ratios, and Markov chain techniques are performed on S&P 500 returns, small stock returns, default premia, dividend yields, industrial production, inflation, and term premia. Findings indicate that mean reversion is not unique to stock returns. We also find that mean reversion results are highly sensitive to the methodology applied. Our findings suggest that mean reversion in stock returns may result from rational responses of investors to changing business conditions.

inancial theory suggests that if expected returns for firms are constant over time, then mean reversion in stock returns may indicate a market inefficiency. In particular, mean reversion in stock returns often is equated with systematic variation of stock returns around equilibrium values. Many studies find that stock returns are mean reverting. For example, McQueen and Thorley (1991), Fama and French (1988b), and Poterba and Summers (1988) find that stock returns are mean reverting over two to five year horizons. They also find that mean reversion is strongest in the period before World War II and is stronger for small firms than for large firms.

Many researchers, however, question the linkage between mean reversion in stock returns and market inefficiency. In particular, mean reversion in priced state variables

* Vincent P. Apilado is a Professor in the Department of Finance and Real Estate at the University of Texas at Arlington, Texas.

** Gary M. Richardson is a Professor in the Department of Business Administration at Central Washington University, Ellensburg, Washington.

Manuscript received, June, 1996, revised, July, 1996.

may cause expected returns to be mean reverting. For example, Fama and French (1988b) argue that the predictable portion of long-term returns may vary over time because equilibrium expected returns may vary over time. Therefore, even in an efficient market, stock returns may appear to be mean reverting. Unfortunately, current studies concentrate more on the symptom of mean reversion in stock returns (e.g., negative autocorrelation) than on the underlying causes of mean reversion. A fairer assessment of the market efficiency implications of mean reversion in stock returns may emerge from an analysis of state variables during changing market conditions.

With the exception of Cecchetti, Lam, and Mark (1990), there has been no at tempt to evaluate the impact of macroeconomic or fundamental forces on mean reversion in returns. These authors specify an equilibrium model and then "calibrate" it to reflect the impacts of consumption, output, or dividends on the time series behavior of returns. They evaluate the likelihood that the estimates derived from historical data could have been generated from the equilibrium process described by their model. As is well known, however, any test of market efficiency that relies on an equilibrium model is a joint test of both the model and market efficiency. Therefore, reliance on an equilibrium model may inject an additional source of error into the investigation. This study takes a simpler approach to examine mean reversion. We do not rely on equilibrium models. Instead, we propose that rational investors will elicit mean reverting stock returns if priced state variables also are mean reverting.

Fama and French (1988b) describe a mean reverting stock market as a combination of a random walk component and a stationary but mean reverting component,

R = + , (1)

where R is the market return, is the random walk component, and is a stationary but mean reverting component.

We argue that is based on the observable information contained in the fundamental and macroeconomic state variables,

Install Equation Editor and double-click here to view equation. , (2)

Where is the information set available at time t-1, x i's are the observable state variables, bi's are coefficients, and k is the number of lags. Thus, we argue that,

Rt = t + (t t-1), (3)

If we assume that investors assimilate market conditions via an examination of macroeconomic and fundamental data, then the linkage between priced state variables and stock returns can be based on the well known valuation formula which describes the stock price as the sum of the present values of all future discounted cash flows. In general, this formula can be written as:

Install Equation Editor and double-click here to view equation. , (4)

where the stock's price (P0) is the present value of all future cash flows (CF). If this is a reasonable model, then any factor that impacts either cash flows or the discount rate (k) should also affect the stock price and, consequently, returns.

Among the earlier research documenting the relationship between the state variables of this study and returns are: Chen (1991), Campbell and Shiller (1988 and 1989), Fama and French (1988), and Fama (1981) who find that dividend yields, inflation, and P/E ratios are associated with stock returns. Other studies by Chen (1991), Kaul and Seyhun (1990), Fama and French (1989), Keim and Stambaugh (1986), and Chen, Roll, and Ross (1986) find linkages between default premiums on bonds, industrial production, and term premiums on bonds and stock returns. We rely on these earlier works to identify potential variables for inclusion in this study.

To examine the efficient market implications of mean reverting stock returns we propose a two stage approach. In one stage, using three different empirical techniques, we examine the time series characteristics of both the state variables and returns for evidence of mean reverting behavior. In an additional step, we evaluate the linkage between a set of state variables and stock market returns using Granger causality tests (a form of OLS regression). Given the empirically demonstrated relationship between returns and state variables, and having demonstrated the presence of mean reversion in returns and mean reversion in the state variables, our argument that mean reverting returns can be explained by mean reverting state variables is based on the traditional interpretation of OLS regression: variability in the independent variable(s) explains or accounts for the variability in the dependent variable.

The remainder of the paper consists of four sections. Section I describes the data. Results of tests of stationarity of the variables also are presented. Section II provides the methodology. Section III presents the empirical findings and Section IV provides conclusions and implications for future research.

I. THE DATA

The data set consists of eight variables, each spanning the 1926-1991 period. For comparison with past research, we include returns on the S&P 500 and returns on an index of small capitalization stocks. The remaining six variables are default premia on bonds, dividend yields, industrial production, inflation, P/E ratios, and term premia on bonds. In each case, we consider horizons from one to five years. Default premia, dividend yields, inflation, S&P 500 returns, small stock returns, and term premia are taken from the Stocks, Bonds, Bills, and Inflation: 1992 Yearbook by Ibbotson Associates. The earnings-price ratio data are from Standard and Poor's Trade and Securities Statistics, and the industrial production series is taken from both the Survey

of Current Business and Business Statistics. The raw data for each variable are measured monthly, except for the earnings-price ratio which is measured yearly. All series are converted to an annual basis to maintain consistency with previous research.

As argued by Fama and French (1988b) and Poterba and Summers (1988), to obtain reliable results, the data series must be stationary. Before applying any of the mean reversion tests, therefore, we examine the data for stationarity using correlograms and Dickey-Fuller tests. Correlograms provide a visual presentation of the autocorrelation coefficients that are subsequently examined for various patterns (see Pindyck and Rubinfeld, 1981). The Dickey-Fuller tests, on the other hand, provide an algebraic evaluation of stationarity (see Dickey and Fuller, 1979, 1981). Of the eight variables, dividend yields and inflation are found to be non-stationary. Therefore, we use first-differences for the dividend yield and inflation series, which, upon retesting, provide the requisite stationarity.

II. METHODOLOGY

The principal methods that have been applied separately to test mean reversion in returns are OLS regression (Fama and French, 1988b), variance ratios (Poterba and Summers, 1988), and Markov chains (McQueen and Thorley, 1991). We employ all three methods in order to compare our findings with past research and to determine the sensitivity of results to each method.

A. OLS Tests for Mean Reversion

In the OLS approach, observation t is regressed on observation t-1,

Xt = + Xt-1 + t, (5)

where Xt is observation t on the variable being examined for mean reversion, observation t is based on horizon [t, t+T], and observation t-1 is based on horizon [t-T, t]. The slope coefficient, , is the OLS estimate of the autocorrelation in the X series. The series is mean reverting if the estimate of is significantly negative. Standard t-tests are used to evaluate the statistical significance of the estimate.

Unfortunately, tests using OLS make various assumptions, including normality, which may not be met. Further, Poterba and Summers (1988) find that OLS tests are more likely to reject a null hypothesis of serial independence than other tests. Ultimately, OLS based tests are viewed as relatively weak tests of mean reversion.

B. Variance Ratio Tests for Mean Reversion

The premise underlying the variance ratio test is that if a series is random, its k-horizon variance, var(k), should be k times as large as its one-period variance, var(1). Therefore, for a random series we expect to find that var(1)/[var(k)/k] = 1. As

demonstrated by Cochrane (1988), the variance ratio can be approximated by a linear combination of sample autocorrelations.

Install Equation Editor and double-click here to view equation. (6)

where k is the horizon in months and is the sample autocorrelation. A variance ratio of less than one implies negative autocorrelation and a variance ratio greater than one implies positive autocorrelation. Thus, the series is mean reverting if the variance ratio is less than one.

To test the variance ratio for statistical significance, we use a randomization approach. For each variable and for each horizon, the variance ratio, first, is calculated based on the actual data. Then, the data are randomly reordered to destroy any time dependencies and the variance ratio is recalculated. This process is repeated 10,000 times. Based on these results, we are able to evaluate the likelihood of obtaining a variance ratio (by chance) that is below the variance ratio observed from the actual series.

A sizeable ratio of repetitions to data set size is required to obtain stable randomization results. Our final data sets consist of 66 observations each, for years 1926-1991 inclusive, for a ratio slightly greater than 150 to 1 (10,000 66). Small ratios can result in unacceptably high standard deviations of results in repeat trials. Cecchetti, Lam, and Mark (1990) use 1000 repetitions on a data set of 116 observations, providing a ratio of 8.6 to 1. Our experience suggests that this is materially low.

Note that the randomization technique is related to the Monte-Carlo methodology. However, in the Monte-Carlo approach, a hypothetical distribution must be specified. If the distribution specification is incorrect, then the empirical results may be suspect. Randomization, however, does not require assumptions about the distribution of the data.

C. Markov Chain Tests for Mean Reversion

In this study, Markov chains are a simple series of ones and zeros. Based on the original data series, if an observation is above the overall average, it is assigned a value of one; if it is below the overall average, it is assigned a value of zero. If a series is random, we should be just as likely to observe a "1,1,1" or a "0, 0, 0" sequence as a "1,1,0" or a "0,0,1" sequence. However, if the market is mean reverting, then the "1,1,0" or a "0,0,1" sequence should occur more frequently. Using the randomization approach, we determine the likelihood of obtaining the observed sequences. As in the variance ratio tests, each variable is evaluated using 10,000 repetitions.

D. Granger Causality Tests

Paraphrasing Granger's (1969) findings, we can say that Y t is causing Xt if we are better able to explain time series X when utilizing the information contained in time series Y than if we ignore this information. In general, Granger causality tests are a form of OLS regression which enables the researcher to develop an understanding of the "causal" (in the Granger sense) and temporal nature of the relationship between returns and state variables. The Granger tests enable us to make inferences, not only about causality in general, but also about the direction of the causality. Further, as demonstrated by Geweke, Meese, and Dent (1983), Granger tests appear to be among the best available (from a statistical standpoint) for the investigation of causality.

The Granger test is computed by performing OLS regression on the following equations,

Install Equation Editor and double-click here to view equation. (7)

Install Equation Editor and double-click here to view equation. (8)

where Rt is the return in period t, X t is the macroeconomic/fundamental variable in period t, is a constant term, and are regression coefficients, n is the number of lags, and , are error terms.

Equations 7 and 8 are referred to respectively as restricted and unrestricted. Conceptually, if causality runs from the macro/fundamental variable to returns, then the i's in 8 will be significant and the prediction error will be smaller than that of the restricted case. Note that reverse causality (from returns to the macro/fundamental variables) can be evaluated by exchanging the equation positions of the two variables. Contemporaneous causality is examined by adding a current value of the explanatory variable (i.e., I = 0) to the lagged variable values and then reestimating the equations.

In summary, our mean reversion tests utilize various methodologies and extend the data sets of earlier research both in time and in number of variables. Further, our methods are designed to circumvent the use of potentially flawed equilibrium models and, hopefully, to enhance the rigor of standard tests associated with mean reversion. The Granger causality tests are well known and accepted and provide a vital link in our arguments about the underlying causes of mean reversion in returns.

III. RESULTS

Table I presents the results of the OLS tests. From left to right, the table presents the slope coefficient, standard error, t-value, and p-value from equation (1) for each horizon from one to five years. With the exception of term premia, the coefficients for each variable are generally negative, indicative of mean reversion. Small stock returns, default premia, dividend yields, inflation, and P/E ratios all have statistically significant

negative autocorrelation for at least one horizon. S&P 500 returns, industrial production, and term premia provide no statistically significant autocorrelation.

Table II presents the variance ratio results. From left to right, the table presents the horizon, actual variance ratio, average randomized ratio (based on 10,000 randomizations), standard deviation of the average randomized ratio, number of times the randomized ratio was greater than or less than the actual ratio, probability value based on the randomization counts, and probability value based on traditional z-scores. Findings indicate that S&P 500 returns, default premia, industrial production, inflation, and P/E ratios all exhibit significant negative autocorrelation for at least one horizon. Small stock returns, dividend yields, and term premia provide no evidence of significant autocorrelation.

In almost all cases, the average variance ratio, estimated from the randomizations of the data, is quite close to the expected value of one. This result demonstrates the efficacy of the randomization approach. In general, if the autocorrelation of a particular horizon is insignificant, the randomized p-value is greater than the p-value based on the z-score (the normal p-value), suggesting that the randomization approach is more accurate. Moreover, when the autocorrelation of a particular horizon is statistically significant, the randomized p-value is typically smaller than that of the normal p-value. There is only one instance (inflation, horizon 5) in which the randomized p-value and the normal p-value do not agree on significance. In this case, the two p-values straddle the 0.10 significance level.

Table I presents the results of the regression Xt = + Xt-1 + t, for eight macroeconomic and fundamental variables, over the 1926-1991 period. The table presents the horizon period, the regression coefficient, standard error, t-statistic, and p-value, respectively.

Table II presents results of variance ratio tests on eight macroeconomic and fundamental variables over the 1926-1991 period. The variance ratio is approximated by a linear combination of sample autocorrelations:

Install Equation Editor and double-click here to view equation.

variance ratio less (greater) than one implies negative (positive) autocorrelation. A randomization approach is used to test the variance ratio for statistical significance. Horizon refers to the time period in years. Variance Ratio is the actual calculated ratio. Avg Ratio is the average ratio obtained from 10,000 random shuffles. is the standard deviation of the average variance ratio. > and < refer to the number of times, out of 10,000 shuffles, that the randomized variance ratio was greater than or less than the actual variance ratio. Random Pval is the marginal significance level based on the randomized findings, and Normal Pval is the significance level based on the normal distribution.

Table III presents the results of the Markov chain tests. The "Actual" column contains the actual number of occurrences of a particular sequence in the unshuffled data. The "Average" column is the average number of occurrences of a sequence based on 10,000 randomized samples (of size 66) of the data. The "pGt" and "pLt" columns are the probabilities of obtaining a count greater than or less than the actual count. For example, there are twelve occurrences of the "1,1,0" sequence for the actual S&P 500 returns series. The probability of observing a count greater than twelve from the random data is only 0.59 percent. The probability of observing a count less than twelve from the random data is 96.7 percent.

While McQueen and Thorley (1991) use a likelihood ratio statistic in evaluating their Markov chain tests, our study deals only with randomization results. McQueen and Thorley note that there is an inherent small sample problem in their investigation of mean reversion. This problem leads to inconsistent findings across various test statistics (Lagrange multiplier test, Wald test, and likelihood ratio test). Instead, we focus exclusively on pure randomization results that do not rely on the knowledge of the distribution of the sample. We believe this approach is more prudent because the efficacy of the randomization approach does not depend on the sample being random or on the nature of the distribution (e.gs., Noreen, 1989, Kempthorne, 1966).

Table I

OLS Tests of Mean ReversionHorizon Coefficient Std Error t-value p-value 1 -0.0017 0.1266 -0.014 0.9893 2 -0.2472 0.1725 -1.433 0.1621 3 -0.1822 0.2024 -0.900 0.3794 4 -0.0221 0.2628 -0.084 0.9343 5 -0.0290 0.3111 -0.093 0.9277

Small Stock Returns

1 0.0692 0.1263 0.548 0.5857 2 -0.0977 0.1819 -0.537 0.5950 3 -0.3790 0.2106 -1.799 0.0879* 4 -0.1797 0.2664 -0.674 0.5118 5 -0.5209 0.2423 -2.150 0.0570*

Default Premium

1 -02842 0.1207 -0.355 0.0216*2 0.0155 0.1795 0.0860 0.93183 0.0715 0.2315 0.3090 0.76084 -0.0033 0.2874 -0.0120 0.99105 -0.0256 0.3239 -0.0790 0.9386

Dividend Yield (1st Difference)

1 0.0999 0.1262 0.791 0.43182 -0.2150 0.1787 -1.203 0.23863 -0.3963 0.2104 -1.883 0.0759*4 -0.5277 0.2347 -2.248 0.0426**5 -0.7246 0.2101 -3.449 0.0062**

Industrial Production

1 0.0901 0.1251 0.721 0.4739

2 -0.1772 0.1775 -0.999 0.3260 3 -0.0113 0.2293 -0.049 0.9612 4 0.0167 0.2663 0.063 0.9511 5 -0.0959 0.1821 -0.527 0.6100

Inflation (1st Difference)

1 -0.0294 0.1275 -0.230 0.8187 2 -0.5388 0.1566 -3.440 0.0018** 3 -0.7022 0.1694 -4.144 0.0006** 4 -0.1197 0.2732 -0.438 0.6686 5 -0.0647 0.2899 -0.223 0.8279

*significant at the 0.10 level, **significant at the 0.05 level

Table II

Variance Ratio Tests of Mean ReversionHorizon Variance Avg. Random Normal

Ratio Ratio > < Pval Pval

S&P 500 Returns

1 0.9663 1.0140 0.1774 6036 3964 0.3964 0.39642 0.8662 1.0232 0.2584 7091 2909 0.2909 0.27173 0.7744 1.0448 0.3320 7832 2168 0.2168 0.20764 0.5021 1.0724 0.3847 9568 432 0.0432** 0.0691*5 0.6688 1.0875 0.4337 8343 1657 0.1657 0.1672

Small Stock Returns1 0.9609 1.0132 0.1758 6125 3875 0.3875 0.38312 1.0748 1.0254 0.2510 4065 5935 0.5935 0.57793 0.7685 1.0810 0.3319 8234 1766 0.1766 0.17314 0.8483 1.0768 0.3865 6964 3036 0.3036 0.27725 0.9841 1.0871 0.4402 5502 4498 0.4498 0.4075

Default Premium1 0.5045 1.0082 0.1796 9982 18 0.0018** 0.0025*2 0.4338 1.0253 0.2559 9963 37 0.0037** 0.0104**3 0.6349 1.0585 0.3238 9170 830 0.0830* 0.0954*4 0.5605 1.0683 0.3863 9185 815 0.0815* 0.0944*5 0.6186 1.0579 0.4224 8597 1403 0.1403 0.1492

Dividen Yields (1st Difference)

1 1.3424 1.0121 0.1791 329 9671 0.9671 0.96742 1.6708 1.0193 0.2639 104 9896 0.9896 0.99323 1.2813 1.0352 0.3298 2216 7784 0.7784 0.77224 1.0872 1.0661 0.3892 4415 5585 0.5585 0.52165 1.5292 1.0874 0.4577 1654 8346 0.8346 0.8328

Industrial Production

1 1.1247 1.1090 0.1696 2557 7443 0.7443 0.73332 1.2819 1.0303 0.2490 1529 8471 0.8471 0.84383 1.0096 1.0504 0.3326 5118 4882 0.4883 0.45114 0.5098 1.0860 0.3646 9613 387 0.0388** 0.0570*5 2.0298 1.1040 0.4412 335 9665 0.9665 0.9821

Inflation (1st Difference)

1 1.1852 1.0185 0.1797 1712 8288 0.8288 0.82322 1.0040 1.0232 0.2588 5041 4959 0.4960 0.47043 0.7080 1.0564 0.3221 8623 1377 0.1378 0.13974 0.6596 1.0877 0.3706 8848 1152 0.1153 0.12405 0.5176 1.1072 0.4775 9119 881 0.0882* 0.1085

P/E Ratio

1 0.9021 1.0169 0.1825 7344 2656 0.2657 0.26472 0.6815 1.0398 0.2732 9129 871 0.872* 0.0948*3 1.0402 1.0493 0.3337 4827 5173 0.5173 0.48914 1.0548 1.0458 0.3812 4526 5474 0.5474 0.50945 0.8038 1.0997 0.4747 7071 2929 0.2930 0.2665

Term Premium

1 1.1215 1.0172 0.1790 2815 7185 0.7185 0.72002 0.9010 1.0256 0.2550 6721 3279 0.3280 0.3125

3 1.8246 1.0378 0.3308 169 9831 0.9831 0.99134 1.1724 1.0456 0.3829 3420 6580 0.6580 0.62985 1.0196 1.1445 0.4282 5758 4242 0.4243 0.3852

*significant at the 0.10 level, **significant at the 0.05 level

SPRING 1996

Table III presents the results of second order Markov chain tests for eight macroeconomic and fundamental variables. In the present context, a Markov chain is a series of 1's and 0's. A 1 (0) represents an observation greater (less) than the overall average. The sequences being studied are "0,0,1" and "1,1,0". Actual refers to the actual count of the number of sequence occurrences based on the historical data. Average is the average count based on 10,000 repetitions. "pGt", and "pLt" are the probabilities of observing a sequence count greater or less than the actual count. A relatively large "Actual" count indicates the presence of mean reversion.

Table III

Markov Chain Tests of Mean ReversionSequence Actual Average pGt pLt

S&P 500001 7 7 0.4424 0.2829110 12 8 0.0059** 0.9666

Small Stocks001 9 7001 10 8 0.0813* 0.7461

0.0922*0 0.7292Default Premium

001 5 6 0.7513 0.0689*110 11 9 0.0823* 0.7791

Dividen Yield001 12 8 0.0049** 0.9611110 8 7 0.1322 0.6376

Industrial Inflation001 9 6 0.0298** 0.8688110 10 8 0.1604 0.6293

Inflation001 9 8 0.1262 0.5404110 12 7 0.0001** 0.9977

P/E Ratio001 7 7 0.3598 0.3488110 10 8 0.1247 0.6860

Term Premium001 8 8 0.4002 0.3314110 10 8 0.0464** 0.8357

*significant at the 0.10 level, **significant at the 0.05 level

21

JOURNAL OF BUSINESS AND MANAGEMENT

S&P 500 returns, small stock returns, default premia, inflation, and term premia all have actual "1,1,0" counts that are significantly greater than expected to occur in a random series. This result indicates that the series are negatively autocorrelated. Moreover, small stock returns, dividend yields, and industrial production all have "0,0,1" counts greater than expected in a random series. This result also is an indication of negative autocorrelation.

Table IV presents a summary of the mean reversion findings by methodology, state variable, and horizon. One of the most striking findings of this paper is that, although seven out of eight variables produce evidence of statistically significant mean reversion at one or more horizons, there is no instance in which all three methodologies agree that the same variable is mean reverting at the same horizon. In fact, from Table IV, there are only two instances, default premia at horizon 1 and inflation at horizon 2, when any two methodologies agree that significant mean reversion is present. Even in those cases, there is considerable disparity between techniques with respect to the level of significance.

Table IV provides summary results for all three of the tests of mean reversion on eight macroeconomic and fundamental variables. Horizons, in years, are listed across the top row. Variables are listed down the first column. The intersection of a row and column (a cell) defines the variable that was tested and the horizon over which the test was performed. An "MC" in a particular cell indicates that the variable was found to be mean reverting, using the Markov Chain technique, at the indicated horizon. Likewise, an "OLS" indicates that the regression tests found mean reversion and "VR" indicates that the variance ratio tests found mean reversion, at the indicated horizon. An empty cell indicates that the variable was not significant, at the 0.10 level, using any of the 3 empirical techniques.

Table IV

Summary of Mean Reversion Tests

Variable/Horizon 1 2 3 4 5

S&P 500 MC VR

Small Stock OLS OLS

Default Premium OLS, VR VR VR VR

Dividend Yield MC OLS OLS OLS

Industrial Production MC VR

22

SPRING 1996

Inflation MC, OLS OLS VR

PE Ratio VR OLS

Term Premium

Table V provides summary findings for the Granger causality tests. Given the familiar nature of OLS statistics, we do not include the full set of results in this paper. As indicated in the table, dividend yields and P/E ratios are found to be significantly related to stock returns over the time period of this study. Lending support to our arguments, both dividend yields and P/E ratios are also found to be mean reverting. Although outside the scope of this paper, for future studies considering performance differences between large and small capitalization stocks, it may be interesting that returns of small stocks show evidence of contemporaneous influences with both industrial production and inflation, while large capitalization stocks show no such tendency.

Table V presents summary results for the direct, indirect, and contemporaneous Granger causality tests on both S&P 500 and small stock returns. D- Direct Granger test, I- Indirect Granger test, C- Contemporaneous Granger test. An empty cell indicates that the variable was not significantly related to returns using any of the three Granger tests. One of the above abbreviations in a cell means that significance was found at the 10 percent or better level using the indicated technique. No cell contains an "I" since no Indirect tests were significant

Table V

Summary of Granger Causality TestsS&P 500 Returns

Variable/Lag 0 1 2 3 4 5

Default Premium

Dividend Yield D

Industrial Production

Inflation

P/E Ratio D

Term Premium

Small Stock Returns

Variable/Lag 0 1 2 3 4 5

23

JOURNAL OF BUSINESS AND MANAGEMENT

Default PremiumDividend Yield DIndustrial Production CInflationP/E Ratio D

Term Premium

24

IV. SUMMARY AND CONCLUSIONS

This paper examines whether or not mean reversion in stock returns may be consistent with weak-form market efficiency. We contend that stock returns in an efficient market may be mean reverting if priced state variables also are mean reverting. Mean reversion in stock returns, then, may be a rational consequence of business conditions that are also mean reverting. To test this proposition, we first perform three separate mean reversion tests on eight variables: S&P 500 returns, small stock returns, default premia, dividend yields, industrial production, inflation, P/E ratios, and term premia. Next, we evaluate the causal and temporal aspects of the relationships between returns and the state variables using Granger causality tests.

Findings indicate that mean reversion is not unique to stock returns, extending also to important state variables. Evidence of mean reversion is found in all but one (term premium) of the state variables examined. We also find that mean reversion results are dependent on the methodology applied. For example, using a horizon of two years, S&P 500 returns, dividend yields, industrial production, and inflation are mean reverting according to Markov chain tests. According to variance ratio tests, only default risk premia and P/E ratios are mean reverting and, according to OLS tests, only inflation is mean reverting. This finding suggests that results of mean reversion derived from a single test may be misleading. Moreover, the results differ based on the horizon being examined. These findings point to the difficulty in drawing inferences on mean reversion. Further, these findings suggest the need for further research involving alternative statistical techniques and expanded categories of state variables.

Granger causality tests demonstrate that both dividends and P/E ratios are significantly related to returns of large and small capitalization stocks. Also significant, and providing crucial support for our arguments, movement in the state variables always precedes movements in returns. We interpret these findings in the traditional OLS manner: variability in the state variables, "explains" variability in returns.

Results of the mean reversion tests, in conjunction with the Granger causality tests, imply that mean reverting stock returns may be explained as rational investor responses to mean reverting state variables. These findings suggest that mean reverting stock returns are consistent with weak form efficient markets.

NOTES

1. Miller, Muthuswamy, and Whaley (1994) recently examine mean reversion in stock index basis changes and find that observed negative autocorrelations are not arbitrage-induced. Instead, they find that negative autocorrelation in basis changes arises simply because of differences in frequency of trade in stocks comprising the index.

2. The relationship between state variables and returns has also been considered within a consumption-based asset pricing framework (see Ferson, 1989) and a production-based pricing framework (see Cochrane, 1991). Although these models provide more detail about the mechanism through which state variables may impact returns, the focus of this paper is on how the state variables impact returns, not on why.

3. Earlier mean reversion studies consider returns series only. The returns series are stationary and, therefore, it has not been necessary to use first differences. Use of differences as opposed to levels does not pose a problem in this study. As discussed in Fama and French (1988b), Poterba and Summers (1988), and others, when evaluating a series for mean reversion, the focus is on a stationary yet mean reverting component in the series. First differencing assures that the series mean is stationary, but it does not preclude individual observations from systematically deviating above or below that mean (e.gs., see Pindyck and Rubinfeld (1981) or McCleary and Hay (1980)).

4 In repeated trials, the standard deviation of the test statistic was unacceptably high with ratios as low as 15:1 (1000 repetitions using 66 observations). This led to inconsistent findings of significance for the test statistic. Substantially increasing the ratio resulted in stable and consistent significance levels.

5. We also evaluated "0,1,0" and "1,0,1" sequences, however we do not report these findings since they relate to single period mean reversion and the focus of this study is on horizons of two to five years.

6. Evans, Keef, and Okunev (1994) examine data for the U.S. and the U.K. over the 1875-1975 period and find significant mean reversion in inflation and in real and nominal interest rates. Their findings for inflation are consistent with those reported in this paper for U.S. inflation during the 1926-1991 period.

7. Although the signs of the coefficient in Table I for S&P 500 returns agree with those of Fama and French (1988b), the significance levels are lower in this study. There are three possible reasons for the difference. First, Fama and French use the CRSP market index. Instead, we use the S&P 500 index, which is concentrated in larger capitalization stocks than those in the CRSP index. Second, our study includes a longer post-WWII period than that examined by Fama and French. Their sample period ends in 1985 while our sample period extends through 1991. Third, Fama

and French use overlapping observations and our study uses nonoverlapping observations.

REFERENCES

Basu, S. (1977), Investment Performance of Common Stocks in Relation to their Price-Earnings Ratios: A Test of the Efficient Markets Hypothesis. Journal of

Finance, 32(3), 663-682.Campbell, J.Y., and Shiller, R.J. (1988). Stock Prices, Earnings, and Expected

Dividends. Journal of Finance, 43(3), 661-676. , (1989). The Dividend-Price Ratio and Expectations of Future Dividends

and Discount Factors. Review of Financial Studies, 1, 195-228.Cecchetti, S.G., Lam, P. and Mark, N.C. (1990). Mean Reversion in Equilibrium Asset

Prices. American Economic Review, 80(3), 398-418.Chen, N. (1991). Financial Investment Opportunities and the Macroeconomy.

Journal of Finance, 4(3), 529-554. , Roll, R. and Ross S.A. (1986). Economic Forces and the Stock Market.

Journal of Business, 59(3), 383-403.Cochrane, J.H. (1988) How Big is the Random Walk in GNP? Journal of Political

Economy, 96(5), 893-920.Dickey, D.A. and Fuller, W. A. (1979) Distribution of the Estimators for

Autoregressive Time Series With a Unit Root. Journal of the American Statistical Association, 70(366), 427-431.

, (1981). Likelihood Ratio Statistics for Autoregressive Time Series with a Unit Root. Econometrica, 49(4), 1057-1072.

Evans, L.T., Keef, S.P., and Okunev, J. (1994). Modelling Real Interest Rates. Journal of Banking and Finance, 18, 153-165.

Fama, E.F. (1981). Stock Returns, Real Activity, Inflation, and Money. American Economic Review, 71(4), 545-565.

and French, K. R. (1988a). Dividend Yields and Expected Stock Returns. Journal of Financial Economics, 22(1), 3-25.

, (1988b). Permanent and Temporary Components of Stock Prices. Journal of Political Economy, 96(2), 246-273.

, (1989). Business Conditions and Expected Returns on Stocks and Bonds. Journal of Financial Economics, 25(1), 23-49.

Feige, E.L. and Pearce, D.K. (1979). The Casual Causal Relationship Between Money and Income: Some Caveats for Time Series Analysis. Review of Economics and Statistics, 61(3), 521-533.

Geweke, J., Meese, R. and Dent, W. (1983). Comparing Alternative Tests of Causality in Temporal Systems: Analytic Results and Experimental Evidence. Journal of Econometrics, 21, 161-194.

Granger, C.W.J. (1969). Investigating Causal Relations by Econometric Models and Cross- Spectral Methods. Econometrica, 37(3), 424-438.

Kaul, G. and Seyhun, N. H. (1990). Relative Price Variability, Real Shocks, and the Stock Market. Journal of Finance, 45(2), 479-496.

Keim, D.B. and Stambaugh, R.F. (1986). Predicting Returns in the Stock and Bond Markets. Journal of Financial Economics, 17(2), 357-390.

Kempthorne, O. (1966). Some Aspects of Experimental Inference. Journal of the American Statistical Association, 61(61), 11-34.

McCleary, R. and Hay, R.A., Jr. (1980). Applied Time Series Analysis For The Social Sciences. Beverly Hills: SAGE Publications.

McQueen, G. and Thorley, S. (1991). Are Stock Returns Predictable? A Test Using Markov Chains. Journal of Finance, 46(1), 239-263.

Miller, M.H., Muthuswamy, J. and Whaley, R.E. (1994). Mean Reversion of Standard & Poor's 500 Index Basis Changes: Arbitrage-Induced or Statistical Illusion? Journal of Finance, 49(2), 479-513.

Noreen, E.W. (1989). Computer Intensive Methods for Testing Hypotheses: An Introduction. New York: John Wiley and Sons.

Pindyck, R.S. and Rubinfeld, D.L. (1981). Econometric Models and Economic Forecasts, New York: McGraw-Hill.

Poterba, J.M. and Summers, L.H. (1988). Mean Reversion In Stock Prices, Evidence and Implications. Journal of Financial Economics, 22(1), 27-59.

Stocks, Bonds, Bills, and Inflation 1992 Yearbook, Chicago: Ibbotson Associate.

I

THE RELATIONSHIP BETWEEN THE PRICE

INTEREST RATE RISK AND THE HOLDING PERIOD

RETURN RISK IN A BOND INVESTMENT

Sang-Hoon Kim *

The interest rate risk of bond investments is divided into the price interest rate risk and the holding period return risk. The former can be measured by Macaulay's duration. However, there is no instrument for the latter. This paper derives a measurement for the latter based on an elasticity similar to the duration and provides a mathematical relationship between the two risks. The relationship derived is very useful for understanding the interest rate risk of bond investments. It also allows an easy derivation of the Fisher-Weil immunization strategy without sophisticated mathematical proof.

n 1938, Macaulay found a measurement for the price interest rate risk and called it "duration." In 1939, Hicks also independently derived a similar measurement. Thereafter, as reviewed by Weil (1973), many scholars used the duration concept to explore conditions under which the interest rate risk of bond investments can be reduced or immunized.

The bond holding period return is said to be "immunized" if the actual holding period return which will be realized by holding the bonds up to a given investment horizon is greater than or equal to the return expected from the interest rates prevailing at the time of the bond investment. A historical paper in the area of the bond immunization was published in 1971 by Fisher-Weil. They derived the famous immunization theorem (hereafter Fisher-Weil immunization strategy) based on the

* Sang-Hoon Kim is an Associate Professor of Finance in the School of Business Administration at Montclair State University, Upper Montclair, New Jersey.

Manuscript received, June, 1996, revised, September, 1996.

duration concept and showed that under certain conditions such as a flat yield curve and parallel interest rate changes, the holding period return is immunized when the bond portfolio is formed such that the duration of the portfolio is equal to the investment horizon.

Other scholars reexamined the immunization strategy. For example, Bierwag and Kaufman (1977), Bierwag (1977), Cooper (1977), Ingersoll, Skelton, and Weil (1978), Cox, Ingersoll, and Ross (1979), and Khang (1979), examined conditions under which Macaulay's duration represents the price interest rate risk and/or proposed several different forms of duration which can represent the price interest rate risk under specific assumptions along with immunization strategies. Gultekin and Rogalski (1984) performed empirical tests on the different forms of the duration measures including Macaulay's original duration and found that none of these duration measures showed any superiority in explaining the interest rate risk.

The Macaulay's duration concept has also been used to control the interest rate risk of financial institutions. For example, Samuelson (1945), Redington (1952), Wallas (1960), and Cronin (1995) showed that the interest rate risk of financial institutions can be immunized by equating the duration of the assets and liabilities.

MEASUREMENT OF THE PRICE AND REINVESTMENT RISKS

The present value of a default-risk-free bond, Vo, can be defined as:

n Ct

Vo = , (1) t=1 Rt

where n = the maturity of the bond, C t = default-risk-fre annual coupon at time t (the coupon + the face value when t = n), Rt = (1+r)t, and r = the nominal risk-free interest rate. Then, it is well known that the price interest rate risk can be measured in terms of the elasticity of the present value with respect to the interest rate factor (1+r) such that:

Vo R dVo RRVo = lim [( )] / [( )] = R0 Vo R dR Vo

R n -tRt-1Ct

= Vo t=1 R2t

n (Ct )/Rt = t , (2) t=1 Vo

= Dp

where Dp = Macaulay's duration. Note that Macaulay's duration is the same as the absolute value of the elasticity. Since the duration is an elasticity of the present value, it represents the price interest rate risk. It should be noted that if the bond is not default-risk free, the elasticity derived should include an additional term showing the impact of the change in interest rate on the coupon payments in addition to Macaulay's duration (Haugen and Wichern, 1975).

Just as the price interest rate risk of a bond can be measured in terms of an elasticity of the present value with respect to the interest rate factor (1+r), the holding period return risk of a bond can also be similarly measured in terms of an elasticity of the terminal value of the bond with respect to the same interest rate factor (1+r).

If the interim coupon incomes from a coupon bond are reinvested and compounded at the interest rate r, the terminal value, Vn, to be realized by holding the bond until its maturity, can be written as:

n Vn = CtRn-t , (3)

t=1

The elasticity of the terminal value with respect to the interest rate factor (1 + r) can be derived as:

Vn R dVn RRVn = lim [( )/( )] = R0 Vn R dR Vn

R n = [(nt) CtRn-t-1] Vn t=1

n CtRn-t n CtRn-t

= n t t=1 Vn t=1 Vn

n (CtRn-t)/Rn

= n t t=1 Vn/Rn

n Ct/Rt

= n t t=1 Vo

= n Dp , (4)

Therefore,

Dr + Dp = n, (5)

where Dr = RVn. Hereafter, D p and D r will be called "price duration" and "reinvestment duration" respectively. The relationship between the two risks is immediately available from equation (5) without any additional derivation. The implications of equation (5) can be summarized as follows:

a. If the investment horizon is the same as the bond maturity, the combined interest rate risk (the price and the holding period return risks) of a bond is always equal to the bond maturity, the investment horizon.

b. Once one risk (the price risk or the reinvestment risk) is known, the other risk (the reinvestment risk or the price risk) can be found by subtracting the known risk from the maturity of the bond.

c. The reduction or elimination of one risk can be achieved at the expense of increasing the other risk by exactly the same amount. The two risks exactly offset each other, leaving the combined interest rate risk at the same amount. It is not possible to reduce the combined duration below the investment horizon.

d. Consequently, controlling the interest rate risk of bond investments should be dealt with a matter of selecting and maintaining a desirable combination of the two risks depending on an investor's expectation on future interest rates and risk attitude.

The price and reinvestment durations are derived based on the present and terminal values which are discounted and compounded by a single discount rate. Since future interest rates are not likely to remain constant (flat yield curve), the present value, terminal value, and the durations computed by a single discount rate can not precisely represent the intended time value of money and also the price and reinvestment risks. For the true time value of money and accurate price and reinvestment risks, actual future interest rates at which the future cash flows can be reinvested should be used for the computation. However, it is practically impossible to accurately estimate such future reinvestment rates. Consequently, in practice, there is not much choice but to use the single discount rate for the durations. Nevertheless, it should be noted that just as the single discount rate is used for bond valuation, the two durations based on a single discount rate are still useful instruments for measuring and understanding overall interest rate risk of bond investments.

THE FISHER-WEIL IMMUNIZATION STRATEGY

The Fisher-Weil immunization theorem which was derived with a sophisticated mathematical proof can be easily obtained from the equation for the reinvestment duration, equation (4). It should be noted that Fisher-Weil found that for the

immunization of the holding period return risk, the price duration of a bond or a bond portfolio should be equal to the investment horizon.

According to equation (4), in order to immunize the holding period return risk, that is; in order to have zero reinvestment duration, one should choose a bond such that its price duration is equal to the maturity of the bond. This condition can be satisfied only with zero coupon bonds because the price duration of a coupon bond is always less than the bond maturity. However, it should be noted that equation (4) was derived based on the assumption that the bond will be held until its maturity. In the case when bonds are not held up to the maturity, equation (4) should be accordingly adjusted.

The adjustment can be shown as follows. Let n be the maturity of a bond and h be the investment horizon. When n > h, Vnh, the terminal value of a n-year-maturity bond at the end of investment horizon can be written as:

n nVnh = CtR(h-t) + CtR-(t-h)

t=1 t=h+1 , (6)

nVnh = CtR(h-t)

t=1 , (7)

The first part of the right side of equation (6) represents the total compound (terminal) value at the end of period h of those coupon incomes to be received up to the investment horizon. The second part of the equation represents the present value at the end of period h of the cash flows after the investment horizon.

Now, the elasticity of Vnh with respect to the interest factor R can be derived as:

dVnh R RVnh =

dR Vnh

R n = [(ht) CtRh-t-1]

Vnh t=1

n CtR(h-t)

= (ht) t=1 Vnh

n CtR(h-t)R(n-h)

= h t

t=1 VnhR(n-h)

n CtR(n-t)

= h t t=1 Vn

n CtR(n-t)R-n

= h t t=1 VnR-n

n CtR-t

= h t t=1 Vo

= h Dpn, (8)

Therefore,

Dpn + Drnh= h , (9)

n n Ct

Note that Vn = CtRn-t, Vo = , Dpn = price duration of a n-year-maturity t=1 t=1 Rt

bond, Drnh = RVnh = reinvestment duration of a n-year-maturity bond when it is held up to h.

According to equation (8), the holding period return risk becomes zero when the price duration of the bond, Dpn is equal to h, the investment horizon. This is the same immunization condition derived by Fisher-Weil. However, contrary to the complex mathematical derivation of Fisher and Weil, the immunization condition can be easily obtained from the equation for the reinvestment duration without any rigorous mathematical proof. Furthermore, the equation for the reinvestment duration provides the following implications:

First, according to equation (9), the total combined interest rate risk of the price and reinvestment risks is always equal to the investment horizon, h, again implying that the reduction or elimination of the reinvestment risk cannot be achieved without increasing the price interest rate risk by the same amount. Therefore, the Fisher-Weil immunization strategy eliminate the holding period return risk by taking the maximum price risk.

Second, the reinvestment duration, contrary to the price duration, varies depending on the length of the investment horizon. Such variation of the reinvestment duration can be easily observed from the following example. Consider a twenty year maturity bond whose price duration is twelve. If the investment horizon is 15 years, we can see from equation (9) that the reinvestment duration becomes +3. On the other hand, if the

investment horizon is 10 years, the reinvestment duration of the same bond becomes -2, a negative number. The negative reinvestment duration implies that changes in the first term of the right side of equation (6) due to changes in interest rates is less than the corresponding changes in the second term. Consequently, if reinvestment duration is zero, the changes in the first and second terms should be the same. The variation of the reinvestment risk can also be intuitively recognized by considering the following. If a long term bond is used for short term investment horizon, the reinvestment risk of the coupon incomes are limited to that short period. On the other hand, if the investment horizon is extended to the bond maturity, the reinvestment risk becomes high because all the coupons received before the maturity should be reinvested.

In the case of a portfolio of more than one bond, the Fisher-Weil immunization condition can be satisfied more easily than in the case of the a single coupon bond. This is because the (price or reinvestment) durations of a bond portfolio are determined by the weighted average of durations of component bonds and therefore, the reinvestment durations of all component bonds are not necessarily equal to zero. Bonds with positive and negative reinvestment durations (9) can be combined to make the portfolio reinvestment duration equal to zero. On the contrary, in the case of a single coupon bond investment, it is not easy to find a bond which can satisfy the immunization condition that the price duration is equal to a given investment horizon.

SUMMARY AND CONCLUSION

The interest rate risk of bond investments is divided into the price risk and the holding period return risk. The measurement of both risks is essential for an accurate understanding of the overall interest rate risk of bond investment and investment strategies controlling the interest rate risk. The price interest rate risk can be measured by Macaulay's duration, but there is no instrument available which can measure the holding period return risk.The paper provides an equation for the holding period return risk. Just as the price interest rate risk can be measured in terms of the elasticity of the present value with respect to the interest rate factor (1+r), the holding period return risk can be measured in terms of the elasticity of the compound (terminal) value with respect to the interest rate factor. The derivation of both risks based on the elasticity with respect to the same interest rate factor reveals a mathematical relationship between the two risks. The most important implication of the relationship is that the reduction or elimination of one risk can be achieved at the expense of increasing the other risks by exactly the same amount, leaving the combined interest rate risk always at the same amount. Consequently, controlling the interest rate risk of bond investments should be dealt with a matter of selecting and maintaining a desirable combination of the two risks depending on an investor's expectation on future interest rates and risk attitude. The measurement of the both risks are also useful in that bond investment strategies such as Fisher Weil Immunization theorem can be easily derived and properly evaluated.

Both the Macaulay's price and the reinvestment durations are based on a single discount rate. Consequently, there are limitations in its application. Nevertheless, just as, in practice, a single discount rate is used for bond valuations, the two durations computed based on a single discount rate are still useful instruments in measuring and understanding the overall interest rate risk in bond investments.

REFERENCES

Bierwag, G. O. (1977). "Immunization, Duration and the Term Structure of Interest Rates." Journal of Financial and Quantitative Analysis, 12(4), 725-741.

Bierwag, G. O. & Kaufman, G. G. (1977). "Coping with the Risk of Interest-Rate Fluctuation: A Note." Journal of Business, 50(3), 364-370.

Cooper, I. A. (1977). "Asset Values, Interest-Rate Changes, and Duration." Journal of Financial and Quantitative Analysis, 12(4), 701-724.

Cox, J., Ingersoll, J. E., & Ross, S. A. (1979). "Duration and Measurement of Basis Risk." Journal of Business, 52(1), 57-61.

Cronin, D. (1995). "Irish Loan-Deposit Interest Rate Margins: A Duration-Based Approach." Applied Financial Economics, 5, 27-32.

Fisher, L., & Weil, R. L. (1971). "Coping with the Risk of Market-Rate Fluctuations: Returns to Bondholders from Naive and Optimal Strategies." Journal of Business, 44(4), 400-431.

Gultekin, N. B., & Rogalski, R. J. (1984). "Alternative Duration Specifications and the Measurement of Basis Risk: Empirical Test." Journal of Business, 57(2), 241-

264.Haugen, R. A., & Wichern, D. W. (1975). "The Intricate Relationship Between

Financial Leverage and the Stability of Stock Prices. Journal of Finance, 30(4), 1283-1292.

Hicks, J. R. (1939). Value and Capital: An Inquiry into Some Fundamental Principles of Economic Theory. London: Oxford University Press.

Ingersoll, J. E., Skelton, J., & Weil, R. L. (1978). "Duration Forty Years Later. Journal of Financial and Quantitative Analysis, 13(4), 627-650.

Khang, C. (1979). "Bond Immunization When Short-Term Rates Fluctuate More Than Long- Term Rates." Journal of Financial and Quantitative Analysis, 13(4), 1085-1090.

Macaulay, F. R. (1938). Some Theoretical Problems Suggested by the Movements of Interest Rates, Bond Yields, and Stock Prices in the U.S. since 1856 . New York: National Bureau of Economic Research Inc.

Redington, F. M. (1952). "Review of the Principle of Life-Office Valuations." Journal of the Institute of Actuaries, 78, 286-340.

Samuelson, P. A. (1945). "The Effects of Interest Rates Increases on the Banking System." American Economic Review, 35(1), 16-27.

Wallas, F. E. (1945). "Immunization." Journal of the Institute of Actuaries Students' Society, 15, 345-357.

Weil, R. L. (1973). "Macaulay's Duration: An Appreciation." Journal of Business, 46(4), 589- 592.

AMERICAN AND GERMAN REGULATION OF

INSIDER TRADING: A COMPARISON

Maria Kathleen Boss *

Cheryl A. Cruz **

Cornelia Alsheimer-Barthel ***

It is extremely important for American business people to be aware of approaches to regulation of insider trading within the European Union, particularly the approach of Germany, its single most influential member. At the same time, European Union members must be sensitive to the completely different approach to prohibited insider trading which is used by American courts. Failure of business people on either side of the Atlantic to be alert to significant differences between the two trading partners could result in criminal prosecution for actions which are perfectly legitimate in the investor's own country. Moreover, such inconsistencies can create opportunities for market failure. Given the significant and increased commercial relationship between the United States and the European Union, particularly Germany, the problem will only be exacerbated as global investment escalates.

This paper considers the differing approaches taken by the United States and Germany to regulate insider trading. It also compares and contrasts the American securities law with its German counterpart, as well as the differing tax consequences of the American and German regulation of insider trading.

* Maria Kathleen Boss is a Professor of Business Law in the School of Business and Economics at the California State University, Los Angeles, Los Angeles, CA.

** Cheryl A. Cruz is an Associate Professor of Tax and Accounting in the School of Business and Economics at the California State University, Los Angeles, Los Angeles, CA.

*** Cornelia Alsheimer-Barthel is a Professor at the Johanne Wolfgang Goethe-University in Frankfurt am Main, Germany.

Manuscript received April, 1996, revised, August, 1996.

W

e are living in an age of increasing internationalization and interdependence of world markets. The most significant trading entity for the United States is the European Union (EU). Within the EU, the most influential player is Germany. This paper addresses a specific aspect of America's crucial commercial relationship with Germany: i.e., the differing approaches taken by each country in attempting to regulate insider trading.

"Insider trading" generally refers to the improper use of material nonpublic information by persons in a position of influence in publicly traded companies. The American approach rejects a definition of what activity is prohibited and who would be considered an insider and instead relies on case law to establish culpability. The Germans, on the other hand, in 1994 passed a law which strictly defines prohibited insider trading, including who would be considered an insider.

Although Germany is still reluctant to become a dominant influence in international military policy (Dale, 1994-95a), in the economic sphere it has exerted major influence, particularly in urging the political and economic integration of the former Soviet bloc with the EU (Dale, 1994-94b). At the same time, the differences between the German and American systems continue to become increasingly more important because they create more opportunities for market failure.

A short synopsis will indicate how significant Germany and the United States are to each other's economies. In 1992 Germany was the United States's fourth largest trading partner, after Canada, Japan and Mexico ("German-American," 1993). American businesses have played an important role in the business development associated with the privatization of former East German state holdings ("German-American," 1993). However, total American direct investment in Germany declined from 40 percent in 1984 to 30 percent in 1990 as a percentage of total foreign direct investment ("German-American," 1993).

Recently there was a very serious challenge to the EU by a German attorney and politician named Manfred Brunner, who argued that the Maastricht Treaty - the master plan for the EU - would deprive German citizens of fundamental rights guaranteed by the German constitution (Gumbel, 1993). So far, however, such challenge has been unsuccessful.

The past two years have seen an acceleration of enthusiasm for the EU. Three new countries in popular elections have agreed to join the EU: Sweden, Austria, and Finland. Swedish voters strongly approved the government's recommendation to join the EU despite concern from a vocal opposition that membership in the EU would compromise Sweden's commitment to political neutrality, human equality and environmental protection (Miller, 1994). Ironically, it had been thought that a strong vote by Sweden for EU membership would positively influence the Norwegian vote, which followed two weeks later ("Sweden EU," 1994). In fact, however, Norway

rejected membership by a vote of 53 percent against and 47 percent for on November 28, 1994. That margin was also almost identical to a similar poll in 1972 ("Norwegians Vote," 1994). The Norwegian vote was also a surprise because Finland had voted on October 16, 1994 by a vote of 57 percent to 43 percent to join the EU. However, of the three scandinavian votes, Finland had been regarded as the most certain to approve joining the EU (Goldsmith, 1994).

Thus, as of January, 1995, a new phase in the process of European unification began, representing fifteen countries, eleven official languages, and 370 million people ("Neuer Abschnitt," 1995). EU officials are extremely optimistic regarding recent industrial rebounds and rise in consumer spending (Roth, 1994). Economists projected that domestic EU demand would rise by 2 percent in 1994 after a 1.8 percent contraction in 1993, with accelerated growth rates of 2.8 percent in 1995 and 3.1 percent in 1996 (Roth, 1994). Certain economies, such as Denmark, Spain, France and Ireland could advance by more than 3 percent in 1995 (Roth, 1994). Industrial spending on new equipment is expected to be better than 7 percent in 1995 in Germany, Belgium, France, Ireland, Italy and the Netherlands (Roth, 1994). Throughout Europe consumers are expected to manifest their confidence in the European economies by increased spending (Roth, 1994).

AMERICAN LAW

In this increasingly global and interdependent environment, business people on both sides of the Atlantic are going to be faced with securities regulation by countries other than their own. This paper addresses the differing approaches to securities regulation of American and German law, which represent totally opposite approaches to the regulation of insider trading. This situation may promote insecurity in those dealing with international securities markets, while at the same time may allow some fraud to escape detection. For example, in March, 1995 charges of insider trading were filed by the Security and Exchange Commission (SEC) against six men for allegedly being involved in illicit trading in shares of Norton Co. and Motel 6 L.P. (McMorris & Welsh, 1995). Two German nationals were implicated in the scheme which yielded alleged profits of at least $5.6 million for the six defendants and others whom they tipped off (McMorris & Welsh, 1995).

The American approach to regulating insider trading has been to rely on a case-by-case basis to define what is prohibited insider trading. This approach has generated considerable controversy, because persons are being prosecuted for a crime that is undefined by statute. Yet constitutional due process considerations require notice of potentially criminal action before penalties can be imposed. The opposing view, which to this point has prevailed in Congress, is that to define prohibited insider trading would allow clever attorneys to discover loopholes in the definition. A Wall Street Journal

article from eight years ago discusses these contradictory positions and remains completely valid today (Ingersoll, 1987).

Although legal analysts will dispute which of the cases decided by the U. S. Supreme Court are of most significance in determining liability for insider trading, this paper focuses on three significant cases. In each one of these cases the high court focused on a different aspect of what activity would be deemed in violation of Section 10b of the 1934 Securities Exchange Act (1934 Act) and SEC Rule 10b-5 promulgated thereunder.

Chiarella v. U.S. (1980) involved alleged wrongful insider trading by the defendant Chiarella, who worked for a financial printing company in New York City. Although the financial documents, which were printed at the company, had the names of the corporations masked, Chiarella was able to figure out from other data in the documents what companies were involved in hostile takeovers. Because Chiarella's actions took place in the environment of merger mania of the late 1970's and early 1980's, there was considerable opportunity for investors to make large sums of money based on trading on material information regarding such activity which was not available to the public at large. When it was discovered by the SEC that Chiarella had traded successfully based on information he obtained from the financial documents he was printing, action was initiated against him for wrongful use of material nonpublic information.

The question of Chiarella's liability ultimately was addressed by the U.S. Supreme Court, which held that the mere possession of material nonpublic information does not result in a duty to disclose or refrain from trading. Rather there must be the existence of some sort of fiduciary relationship and hence consequent duty between the person engaging in such trading and the corporation whose shares are being traded. Chiarella had no relationship whatsoever to any of the companies whose shares he traded and consequently did not breach any fiduciary duty to them. Thus, this case established that for purposes of liability under American securities laws a fiduciary duty must be breached before any wrongdoing can be established.

The second seminal case decided by the U.S. Supreme Court in defining prohibited insider trading involved Raymond Dirks, a securities analyst in New York City (Dirks v. SEC, 1983). Dirks did not involve himself in any trading at all. Rather the question of his liability arose out of the information he discovered when he investigated Equity Funding Corporation pursuant to a request from a former officer of the company named Secrist who was concerned regarding actions by top corporate management which Secrist believed jeopardized the very existence of Equity Funding Corporation. Dirks flew to Los Angeles and spoke to a wide variety of Equity Funding employees. Although top management denied that there were any problems, lower level employees said that the company was in serious trouble. Dirks tried to get the Wall Street Journal to report his findings, but it refused, fearing a defamation suit. When he returned to New York, Dirks advised his clients, many of whom were institutional investors with

millions of dollars in Equity Funding, to sell their shares. Shortly afterwards, the information regarding the precarious financial condition of the company became public and trading of its shares on the New York Stock Exchange was halted. Many investors lost millions of dollars in the debacle.

When the high court reviewed the question of Dirks' possible liability under Section 10b and Rule 10b-5, it primarily focused on his status as a "tippee" and the possible liability of such a tippee under the securities laws. This question was resolved based on the motivation of the "tipper" (the corporate insider with a Chiarella type obligation) in transferring material nonpublic information to a tippee (an outsider with no relationship to the corporation, such as Dirks). Because Secrist, the insider/tipper, only wished to expose wrongdoing by the corporation, he did not have an improper motivation when he shared material nonpublic information with Dirks. As a tippee Dirks inherited Secrist's motivation which was not improper. Absent improper motivation by the tipper, there can be no derivative liability to the tippee.

This case also introduced the concept of a constructive or temporary insider, to whom material nonpublic information is entrusted by corporate insiders in order for the corporate or temporary insider to perform services for the corporation. Examples would include attorneys and accountants being entrusted with secret corporate information in the course of rendering professional services. Such persons have all the fiduciary obligations of a traditional insider regarding the information they have received.

The final important case in which the U.S. Supreme Court addressed the parameters of insider trading involved someone who was not in a corporate capacity at all, but rather was a reporter for the Wall Street Journal. R. Foster Winans had the byline for the popular column in the newspaper called "Heard on the Street". His column was so successful that his predictions became self-fulfilling prophecies: i.e., if he predicted that a stock would do well, demand for the stock would result, and the stock's value would increase. Winans saw the opportunity to make profits for himself and entered into a conspiracy (along with his roommate, Carpenter)whereby he would sell the information to appear in the column prior to its publication. When the conspiracy was discovered, Winans and Carpenter were prosecuted for illicit insider trading (U.S. v. Carpenter, 1987). Although many commentators predicted that Winans would be acquitted because he had no fiduciary duty under the Chiarella standard nor tippee or constructive insider under the Dirks standard, the high court found him liable under a theory of misappropriation. By using for personal gain information which belonged to his employer, the Wall Street Journal, Winans misappropriated information.

GERMAN LAW

In sharp contrast to the United States, until recently insider trading in Germany was perfectly legal. However, there was a major change in the German approach when a new law became effective on August 1, 1994 following a 1989 Directive by the Commission of the European Community (Ascarelli, 1989). (Note that the term European Community (EC) was replaced by the term European Union (EU), connoting a closer relationship among the member countries). This law represents a major change in the German law, because it is the first time that prohibited insider trading has been defined with specificity.

Just as in the United States, there were high visibility cases involving controversial insider trading on the German stock exchange. Indeed, there was a high visibility case way back at the turn of the century in 1904 where the Reichsgericht decided a case in which someone sold mining shares fifteen minutes after he had read in the newspaper about an accident in the mine (Jüristishe, 1904). Perhaps the most important case, and the one which probably most directly led to the enactment of the new law, involved Franz Steinkühler, who was the chairman of one of Germany's most powerful trade unions ("German Union," 1993). The scandal involving Steinkühler was a classic situation whereby he made a huge profit based on information he had obtained while sitting on Daimler-Benz AG's supervisory board (equivalent to an American board of directors) ("German Union," 1993).

Unlike in America, in Germany it is common practice for top union officials to sit on boards of directors of companies where their unions are strongly represented. Steinkühler bought a large number of Mercedes AG Holding shares shortly before the holding company's shares were converted into higher priced Daimler-Benz on a one-for-one basis ("German Union," 1993). At that time insider trading was not illegal in Germany, but it was certainly frowned upon ("German Union," 1993). The situation involving Steinkühler was especially controversial, because he was very influential and well-beloved by union members. In particular he could not explain why he had bought the shares not in his own name, but in the name of his small son. That was especially suspicious (Boss & Alsheimer-Barthel, 1993).

The most significant milestone on the way to a statutory definition of prohibited insider trading in Germany was the Directive submitted by the EC. The EC Commission worked until May, 1987 on a first proposal (Amtsblatt, 1987). A second proposal followed in October of 1988 (Amtsblatt, 1988). The Directive itself was submitted on November 13, 1989 (Amtsblatt, 1989) and the EC member states were to comply by amending their laws by June 1, 1992 (George, Boss & Haraldson, 1992). The German government did not manage to comply in time, because the intention was to have the Insider Law become part of a larger framework of regulation, which included the goal of modernizing the German financial center (Finanzplatz Deutschland) (Assmann, 1994a). The Directive was prepared in response to the fact that each member state of the EU regulated insider trading differently (Assmann, 1994a). These approaches varied from utilizing a voluntary code of conduct to

imposition of criminal penalties to an absence of any regulations at all (Assmann, 1994a). As previously noted, Germany had relied on voluntary self regulation to deter insider trading (Assmann, 1994a).

The new German law (Finanzmarktforderüngsgesetz) prohibiting insider trading initiated a "new age" whereby Germany will become internationally more competitive for investors, according to German Finance Minister, Theo Waigel (o.v.: 5 Jahre Haft, 1994). In reality the new law is actually a blended law incorporating laws for bond and share trading and laws for the amendment of bond and stock regulations (börsenrechtlicher und wertpapierrechtlicher Vorschriften). One of its most important features is a new federal supervising office for bond trading (Bundesaufsichtsamt für den Wertpapierhandel). This office performs different functions, but investigating and fighting insider trading is its most important function. Therefore, the office is given a wide variety of tools to explore the activities of issuers, banks, brokers, and others involved in the stock exchange (alle Marktteilnehmer). The "Marktteilnehmer" has to inform the supervising office within one day about all transactions in shares traded on the stock exchange. The office can request details about such transactions, including customer names. Moreover, personnel from the supervising office are allowed to visit the offices of the "Marktteilnehmer" and require proof of transactions. If its suspicions are aroused, the supervising office must inform the state prosecutor (Wittich, 1995a).

The "Bundesaufsichtsamt" observes all unusual price movements and unusual/large share transactions with a specially designed computer program called Stock-Watch. This program also provides information about the date, time, quantity, and price of sales and purchases as well as the names of the banks and stock brokers involved (Wittich, 1995b).

From now on it will be forbidden to gain profit from the use of confidential information which might influence the price of shares on the German stock exchange (for example, an approaching merger). Offenders may be prosecuted and face up to five years imprisonment.

Completely contrary to the American approach, the new law contains a very broad definition of insiders. Clearly directors, members of the board or major shareholders, or those who by their activity (stockbroker, director's secretary, accountant, Betriebsrat) have knowledge of information which is not yet published and which can influence share prices would be considered insiders (Boss & Alsheimer-Barthel, 1994). Even the taxi driver, the flight attendant and the cleaning lady who obtained any information in due course of their profession are insiders. It is also clear that it is not forbidden to exploit information which was obtained through observation and analysis of the economy and the stock exchange based on publicly known information. However, the German Securities Trading Act does not provide an answer as to what is to be included in "publicly known" information (Kümpel, 1994b).

Responsibility for the new law will be assumed primarily by the major German banks and to a lesser extent by other public companies. This assignment is based on the fact that German banks have many opportunities to come in contact with inside information (Assmann, 1994b). The banks are required to install a compliance officer, who although remaining a bank employee will function as a state prosecutor (Boss & Alsheimer-Barthel, 1994). The compliance officer will match two types of information: material events which create insider knowledge and complete documentation of all security transactions by bank employees, particularly highly placed employees (Boss & Alsheimer-Barthel, 1994). The compliance officer will communicate directly with the newly established federal supervisory office to report all findings (Boss & Alsheimer-Barthel, 1994).

Under the new law a public limited company must inform the "Bundesaufsichtsamt für den Wertpapierhandel" immediately about any facts about itself as well as other companies in which it has more than a 5 percent interest (Beteiligungen), which would considerably influence share prices. Then the Bundesaufschtsamt will decide if these facts must be published or not. If the facts must be published, then the information is disseminated immediately by creating the "Bereichsöffentlichkeit" (Jürgens, 1995).

The new federal supervisory office became operational at the beginning of 1995 at the German stock exchange in Frankfurt. It will be supported by state control offices (Börsenaufsichtsbehorde der Lander) and by a trade control center (Handelsuberwachungs-stelle) at each of the eight German stock exchanges located in Frankfurt, Düsseldorf, München, Hamburg, Stuttgart, Berlin, Hannover, and Bremen (Kümpel, 1994a). It is expected that this new structure will strengthen confidence in the reliability and fairness of the German stock exchange.

TAX LAW

The tax treatment is important to the regulation of insider trading in that the amount of penalty, if any, is affected by the tax consequences of the "return of profits." If the insider obtains some type of tax benefit (i.e., deduction, credit, etc.) for the "return of profits," then there is essentially no or very little ramification for engaging in this prohibited activity. Thus, the insider or wrongdoer is restored to the status quo before any wrongdoing had occurred. On the other hand, if the insider obtains no tax benefit for the "return of profits," then he is essentially penalized in the amount of the tax paid when the prohibited activity took place and the initial profits were made. The Americans take the former approach. Since the Germans do not require the "return of profits," the tax benefit issue is not addressed unless an insider voluntarily "returns the profits." However, both countries agree that any fines and/or penalties are not deductible.

AMERICAN TAX LAW

Under the SEC rules, persons who engage in prohibited insider trading are required to return the profits generated from violation of the securities laws. This "return of profits" is normally called disgorgement. It is one of the most common forms of ancillary relief granted in SEC enforcement actions and has been ordered in a wide variety of cases (SEC v. Drexel Burnham Lambert Inc., 1993; SEC v. Texas Gulf Sulphur Co., 1971). As such, the Court in Litton Industries, Inc. v. Lehman Bros. Kuhn Loeb, Inc. (1990) found no meaningful distinction between disgorgement in insider trading cases under Section 10(b)-manipulative or deceptive devices versus Section 16(b)-short swing profits of the 1934 Act. Under both sections, disgorgement is a proper measure of damages which serves the equitable purpose of depriving a wrongdoer of ill-gotten gains. Moreover, Section 10(b), like Section 16(b) is a remedial provision (Litton v. Lehman, 1990).

The amount of disgorgement should be a reasonable approximation of profits causally connected to the violation, that is, all profits gained while in violation of the securities law (SEC v. Drexel Burnham Lambert Inc., 1993). The traditional measure of actual damages is the out-of-pocket loss measured by the difference between the fair value of what the injured party received and the fair value of what the injured party would have received had there been no fraudulent conduct (Litton v. Lehman, 1990). However, if the wrongdoer's profits are greater than the injured party's actual loss, the proper measure of damages will be the amount of the wrongdoer's profits (Litton v. Lehman, 1990). Thus, disgorgement is limited to compensatory, non-speculative damages and as such, the computation of actual damages does not include consideration of predicted or anticipated tax benefits (Torres v. Borzelleca, 1986). However, if the injured party received actual tax benefits as a result of the fraudulent transactions, the actual damage award must be reduced by the valueof that tax benefit (Austin v. Loftsgaarden, 1982).

The purpose of disgorgement is to prevent unjust enrichment by depriving a wrongdoer of ill-gotten gains. As such, it merely returns the wrongdoer to the status quo before any wrongdoing had occurred (SEC v. Lorin, 1994). However, fines, penalties, and forfeitures alter the status quo before any wrongdoing took place (SEC v. Lorin, 1994). Thus, by definition disgorgement is remedial and not punitive (SEC v. Lorin, 1994; SEC v. Blatt, 1978). Many years earlier, 1915 to be exact, the Supreme Court in Meeker v. Lehigh Valley Railroad Co. (1915) concluded that disgorgement does not constitute a fine, penalty, or forfeiture by explaining that "strictly remedial" liabilities do not fall under the "catch-all" statute because they are not "punitive". As a result, subsequent courts have held that a wrongdoer will not be subject to liability to both the SEC and private parties because such liability would be clearly punitive in its effect and would constitute an impermissible penalty assessment (SEC v. Lorin, 1994; Litton v. Lehman, 1990).

Whether disgorgement is characterized as a tax deduction depends on the facts and circumstances of each case. Most of the tax cases deal with Section 16(b) of the 1934

Act. However, based on the fact that disgorgement has the same definition in tax cases dealing with Section 10(b), any outcome of Section 16(b) tax cases can be analogized to Section 10(b) tax cases.

Early cases in this area were in disagreement as to deductibility. In 1951 the U.S. Tax Court, in the Davis, Jr. v. Commissioner (1951) case concluded that disgorgement in a Section 16(b) case was a penalty and a deduction was disallowed since it would mitigate the deterrent effect of the statute and subvert a sharply defined public policy. The sharply defined public policy was the fact that Congress considered transactions in violation of Section 16(b) to be detrimental to the public welfare and designated disgorgement of profits as a substantial deterrent to insider trading violations. Later the U.S. Tax Court, in the Marks v. Commissioner (1956) case concluded that disgorgement of insider profits under Section 16(b) was deductible as an ordinary business expense since the disgorgement was made to protect business reputation and not to frustrate a sharply defined public policy. In light of these two cases the Internal Revenue Service (IRS) issued Revenue Ruling 61-115 (1961) stating that a deduction for disgorgement of profits will not be denied on the grounds that it frustrates sharply defined public policy and whether it is deductible as an ordinary loss or a capital loss depends on the nature of the profits.

If the profit was ordinary income, then the disgorgement will give rise to a fully deductible loss (Pike v. Commissioner, 1965). If the profit was capital in nature, then the disgorgement will give rise to a capital loss deduction (Mitchell v. Commissioner, 1970; Anderson v. Commissioner, 1973; Cummings v. Commissioner, 1974; Brown v. Commissioner, 1976). These cases dealt with Section 16(b) of the 1934 Act. However, the U.S. Tax Court in Bradford v. Commissioner (1978), a Section 10(b) tax case, took a different approach using the origin-of-the-claim test and stated that the disgorgement of profits related to an acquisition of a capital asset were capital expenditures to be added to the basis of the stock purchased. The IRS agreed with this decision in Revenue Ruling 80-119 (1980), stating that the origin-of-the-claim test is the proper method of determining whether the settlement payment (disgorgement) is a deductible business expense, a capital expenditure or a combination of the two types of expenditures.

More than ten years after the origin-of-the-claim test was used to render disgorgement of profits a capital expenditure, the Tax Court introduced another theory into the Section 10(b) tax cases - Claim of Right Doctrine. In 1991 the Tax Court in Barrett v. Commissioner (1991) held that the wrongdoer was entitled to a credit under Internal Revenue Code (IRC) Section 1341(a)(5) in an amount equal to the decrease in tax attributable to the removal of short term capital gain from a prior year's gross income. Under Section 1341(a) if the taxpayer included an item in gross income in one taxable year, and in a subsequent taxable year he becomes entitled to a deduction because the item or a portion of it is no longer subject to his unrestricted use, and the amount of the deduction is in excess of $3,000, the tax for the subsequent year is

reduced by either the tax attributable to the deduction or the decrease in the tax for the prior year attributable to the removal of the item, whichever is greater. Under this rule, the taxpayer is entitled to a deduction only in the year of repayment or disgorgement. Pursuant to Section 1341(a)(2) taxpayer must show that he did not have an unrestricted right to the item. This can be done by showing that the taxpayer had a legal obligation to restore the item either through court awarded damages or an out-of-court arm's-length settlement.

Since disgorgement is a remedial remedy, the determination of whether it is deductible as an ordinary loss, capital loss or claim of right deduction will be done on a case by case basis. The IRC does not specifically address the deductibility of disgorgement of profits.

If the damage award also includes fines and/or penalties, then IRC Section 162(f) states that no deduction shall be allowed for such fines or penalties paid to a government for the violation of any law.

The wrongdoer is taxed on insider profits based on the tax rate that applies to his taxable income. The United States has a progressive tax rate, which ranges from 15 percent to 39.6 percent for individuals (IRC 1(a),(b),(c) and (d)). However, for economic reasons the American government lowered the maximum rate for long term capital gains to 28 percent in order to stimulate the economy (IRC 1(h)). In the United States, profits from insider trading will usually be characterized as capital gains, short-term or long-term depending on the holding period. Both short-term and long-term gains are included in total taxable income; however, short-term gains are subjected to a maximum tax rate of 39.6 percent, while long-term capital gains are subjected to a maximum tax rate of 28 percent.

If the disgorgement is treated as ordinary loss, it will be treated as a "for AGI (Adjusted Gross Income)" deduction (IRC 62(a)(3)); however, if it is treated as a claim of right, it will be treated as a "from AGI" deduction under miscellaneous not subject to 2 percent (IRC 67(b)(9)). Even though these deductions are different types, they both will be subjected to a maximum tax benefit of 39.6 percent. On the other hand, if the disgorgement is treated as a capital loss, it will be a "for AGI" deduction (IRC 62(a)(3)). If all capital gains and losses net to a capital loss, then the $3,000 maximum deduction (IRC 1211(b)(1) and (2)) per year will also be subjected to a maximum tax benefit of 39.6 percent.

GERMAN TAX LAW

The German tax approach, as with its securities law approach, is quite different from the American tax approach. Under the German securities law persons who engage in prohibited insider trading are not required to return the profits generated from violation of the securities laws. However, in cases where persons earned the profits by fraudulent activities, the wrongdoers usually offer to return the profits voluntarily to the court as a sign of goodwill and in an effort to decrease any possible prison sentence. Sometimes a court will order a wrongdoer to make a charitable contribution, although this does not preclude imprisonment. A third possibility is that any persons or companies which have suffered losses as a result of insider trading can sue civilly for damages. These lawsuits have a strong chance of success when the wrongdoer has already been found guilty in a criminal prosecution.

The German tax law does not allow a deduction for fines and/or penalties (Einkommensteuergesetz 4 V Nr. 8, 1996). Fines are defined as those imposed by a court or an official German authority or by an authority of the EC. In some cases where the fine is equal to the profits from insider trading, then the fine is deductible. In reality then, it is a disgorgement and not a fine, although there is no word in the German tax statute equivalent to disgorgement.

The wrongdoer is taxed on insider profits based on the tax rate that applies to his entire taxable income. Germany has a progressive tax rate, which ranges from 26 percent to 53 percent (Einkommensteuergesetz 32a, 1996). However, for political reasons the German government lowered the maximum rate for trade income to 47 percent in order to develop the German economy (Einkommensteuergesetz 32c, 1996). In Germany, profits from insider trading will usually be characterized as miscellaneous income, which is included in total taxable income and subject to the higher maximum tax rate of 53 percent. If the profits from insider trading are related to a trade or business, they are characterized as trade income and only subjected to a maximum tax rate of 47 percent.

CONCLUSION

The comparison of the American and German securities laws and cases showed just how different their approaches are in regulating insider trading. The American approach is to define prohibited insider trading through case law. The drawback in this approach is that persons will be prosecuted for a crime that is undefined by statute. On the other hand, a statute defining insider trading would allow for discovery of loopholes in the definition. Although no statute exists, the case law has defined insiders narrowly as evidenced by the "fiduciary duty" standard in the Chiarella case.

The German approach is to specifically define prohibited insider trading in the law. Completely contrary to the American approach, this new law defines insiders very broadly. For instance, the printer in the Chiarella v. U.S., an American case, would be

an insider based on German law because he obtained the nonpublic information in the due course of his profession.

With such differing approaches to the regulation of insider trading, the potential for market manipulation and failure is quite high. This problem will only be exacerbated as the United States and Germany increase their commercial relationship and their global investments escalate.

The tax consequences of the American and German regulation of insider trading are quite different. The American insider is required to return the profits ("disgorgement") in order to prevent unjust enrichment. The determination of whether the return of profits is tax deductible is dependent on case law. In contrast, the German insider is not required to return the profits and thus the question of deductibility need not be addressed. Other dissimilarities include the characterization of profits, the taxation of profits, and the tax rates. However, the most important similarity is the nondeductibility of fines and penalties.

Only time will tell how the American securities/tax case laws and the German securities/tax statutes actually will evolve given the significant international trading between the United States and Germany.

NOTES

1. The main regulations are set forth in Part 3 of Gesetz überden Wertpapierhandel.2. Translation of the German Act on Securities Trading (Securities Trading Act -

WpHG), Chapter 3, 13 (Insiders):

(1) An insider is anyone who

1. as a member of the management body (Geschäftsführungsorgan) or supervisory body (Aufsichtsorgan) or as general partner (persönlich haftender Gesellschafter) of the issuer or an enterprise associated with the issuer (verbundenes Unternehmen), or

2. by virtue of his participation in the share capital of the issuer or of an enterprise associated with the issuer or

3. by virtue of his profession or his activity or his assignment in due course has obtained knowledge of circumstances not known to the public relating to, or to one or more issuers of, insider securities which, if becoming known to the public, may materially affect the price of such securities (inside information).

(2) An assessment (analysis) which is compiled exclusively based on publicly known information shall not be deemed inside information, even if such assessment might materially affect the price of insider securities.

3. Id.4. Id.; see also, Höpt, Klaus J.: Insiderwissen und Interessenkonflikte im europaischen

und deutschen Bankrecht, in: Kübler, Friedrich Werner, Winfried/Mertens, Hans Joachim (Hrsg.): Festschrift für Theodor Heinsius, Berlin, 1991, S. 289-322, hier S. 289.

5. Internal Revenue Code Section 1222(1): Short-term capital gain means gain from sale or exchange of a capital asset held for one year or less; Internal Revenue Code Section 1222(3): Long-term capital gain means gain from sale or exchange of a capital asset held for more than one year.

REFERENCES

Anderson v. Commissioner, 480 F.2d 1304 (7th Cir. 1973).Amtsblatt EG Nr. C153 (vom 11). (1987, Juni). S. 8.Amtsblatt EG Nr. L334/30 (vom. 18). (1989, November). Amtsblatt Nr. C277 (vom. 27). (1988, Oktober). S. 13.Ascarelli, S. (1994, August 1). "Germany's Law on Insider Trading Takes Effect Today." Wall Street Journal, p. A14.Assmann, H.D. (1995a, May). "Das künftige deutsche Insiderrecht I." Die Aktiengesellschaft, S. 196-206. Assmann, H.D. (1994b, Juni). "Das künftige deutsche Insiderrecht II." Die

Aktiengesellschaft (vom. 1), S. 237-258.Austin v. Loftsgaarden, 675 F.2d 168 (8th Cir. 1982).Barrett v. Commissioner, 96 T.C. 713 (1991).Boss, M.K. & Alsheimer-Barthel, C. (1993, November). "Insider-Geschäfte An Deutschen Börsen." Speech presented at a German class at UCLA, Los Angeles.Boss, M.K. & Alsheimer-Barthel, C. (1994, April). Insiders Defined. [Interview with Dr. Hermann Reuter at Dresdner Bank, AG].Bradford v. Commissioner, 70 T.C. 584 (1978).Brown v. Commissioner, 529 F.2d 609 (10th Cir. 1976).Chiarella v. U.S., 445 U.S. 222 (U.S. Sup. Ct. 1980).Cummings v. Commissioner, 506 F.2d 449 (2nd Cir. 1974).Dale, R. (1994-95a). "EU Foreign Policy Europe's New Voice." Europe, 16.Dale, R. (1994-95b). "Uniting the Continent Europe's Greatest Challenge." Europe, 13.Davis, Jr. v. Commissioner, 17 T.C. 549 (1951).Dirks v. SEC, 463 U.S. 646 (U.S. Sup. Ct. 1983).Einkommensteuergesetz 4 V Nr. 8, zuletzt geändert durch Artikel 1 des Jahressteuergesetzes 1996.Einkommensteuergesetz 32a, zuletzt geändert durch Artikel 1 des Jahressteuergesetzes

1996.Einkommensteuergesetz 32c, zuletzt geändert durch Artikel 1 des Jahressteuergesetzes

1996.George, B.C., Boss, M.K., & Haraldson, M. (1992). "The European Community

Directive Regulating Insider Trading: Harmonization Difficulties for the US." Anglo-American Law Review, Pp. 257-285.

"German-American and EC-American Trade Relations: A Status Report." (1993, April). Fact Sheet. New York: German Information Center.

"German Union Chief Quits on Insider Trading Report." (1993, May 26). Wall Street Journal, p. A12.Gesetz über den Wertpapierhandel, Part 3.Goldsmith, C. (1994, October 17). "Voters in Finland Opt to Join the EC, Effective January." Wall Street Journal, p. A11.Gumbel, P. (1993, June 21). "Maastricht Faces German Legal Test In Suit Over Scope of Treaty's Powers." Wall Street Journal, p. A8.

Höpt, K.J. (1991). Insiderwissen und Interessenkonflikte im europaischen und deutschen Bankrecht. In Kübler, F., Werner, W. & Mertens, H.J. (Hrsg.), Festschrift für Theodor Heinsius, Berlin, S. 289-322.

Ingersoll, B. (1987, March 26). "Demand Rises for Law Defining Insider Trading To Provide More Than a Gut Feeling as a Guide." Wall Street Journal, p. 70.Internal Revenue Code Sections 1(a),(b),(c),(d) and (h).Internal Revenue Code Section 62(a)(3).Internal Revenue Code Section 67(b)(9).Internal Revenue Code Sections 1211(b)(1) and (2).Internal Revenue Code Sections 1222(1) and (3).Jürgens, U. & Rapp, V. (1995, February). "Ad-hoc-Publizitat: Ablauf und Technik." Die Bank, S. 97-100.Kümpel, S. (1994a, Mai). "Die Organleihe im Rahmen er neuen Kapitalmarktaufsicht."

Weretpapier-Mitteilungen, Festgabe für Thorwald Hellner (vom. 9), S. 35-40.

Kümpel, S. (1994b, Dezember). "Zum Begriff der Insidertatsache." Zeitschrift für Wirtschafts- und Bankrecht (vom. 3), S. 2137-2176.Litton Industries, Inc. v. Lehman Bros. Kuhn Loeb, Inc., 734 F.Supp. 1071 (S.D.N.Y. 1990).Marks v. Commissioner, 27 T.C. 464 (1956).McMorris, F.A. & Welsh, J. (1995, March 7). "SEC Files Charges of Insider Trading

Against Six Men." Wall Street Journal, p. B2.Meeker v. Lehigh Valley Railroad Co., 236 U.S. 412, 35 S.Ct. 328, 59 L.Ed. 644 (U.S.

Sup. Ct. 1915).Miller, M. (1994, November 14). "Sweden Follows Austria, Finland With Vote to Join

European Union." Los Angeles Times, p. A4.Mitchell v. Commissioner, 428 F.2d 259 (6th Cir. 1970)."Neuer Abschnitt für die EU: 15 Staaten, elf Amtssprachen, 370 Millionen Menschen,

neue Kommission." (1995, January 6). Deutschland Nachrichten, p. 1."Norwegians Vote Against Joining European Union." (1994, November 29). Wall Street

Journal, p. A21.o.V.: 5 Jahre Haft für Insidergeschafte. (1994, January). AdRem-Wirtschaft, Soziales,

Recht, S. 1.Pike v. Commissioner, 44 T.C. 787 (1965).Reichsgerichtsurteil in: Jüristische Wochenschrift. (1904). S. 167.Revenue Ruling 61-115, 1961-1 CB 46 (1961).Revenue Ruling 80-119, 1980-1 CB 40 (1980).Richtlinie des Rates (vom 13). (1989, November).Roth, T. (1994, November 28). "EU Recovery Appears to Accelerate As Industry and

Spending Rebound." Wall Street Journal, p. B7.SEC v. Blatt, 583 F.2d 1325 (5th Cir. 1978).SEC v. Drexel Burnham Lambert Inc., 837 F. Supp. 587 (S.D.N.Y. 1993).SEC v. Lorin, 869 F.Supp. 1117 (S.D.N.Y. 1994).

SEC v. Texas Gulf Sulphur Co., 446 F.2d 1301 (2d Cir. 1971), cert. denied, 404 U.S. 1005, 92 S.Ct. 562, 30 L.Ed.2d 558 (1971).

Securities Trading Act - WpHG, Chapter 3, 13."Sweden EU Vote Unpredictable." (1994, November 11). Wall Street Journal, p. A10.Torres v. Borzelleca, 641 F.Supp. 542 (E.D.Pa. 1986).U.S. v. Carpenter, 108 S. Ct. 316 (1987).Wittich, G. (1995a, March). "Neue Aufsichtsstruktur in Deutschland-Das

Bundesaufsichtsamt für den Wertpapierhandel." DTB Dialog.Wittich, G. (1995b, March). "Fallen für Jager und Gejagte." Capital, S. 207-211.

O

ASSESSING THE PAYOFF FROM AN INFORMATION

TECHNOLOGY INFRASTRUCTURE: A MULTI-

PHASED APPROACH

Lara Preiser-Houy *

The key managerial challenge associated with providing an information technology (IT) infrastructure is to be able to ascertain the economic viability of infrastructure investments. Such investments appear to be on the increase as we move into the latter half of this decade. Yet to date, very little is known about the impacts infrastructure resources have on organizational performance and the ways in which these impacts could be measured. The lack of guidelines on how to assess the infrastructure payoff presents a problem for the Information Systems (IS) managers charged with the task of providing an effective foundation for the organizational business processes.

An approach to assess the economic viability of infrastructure resources presented in this paper is aimed at helping IS and general managers make more informed decisions as to their infrastructure investments. The proposed approach consists of the three phases -- identifying infrastructure value sources, assessing the cost and value of the firm’s installed infrastructure resources, and justifying additional infrastructure expenditures. The application of the proposed approach to the infrastructure investment decisions is one of the first steps towards addressing the managerial challenge of providing a robust foundation for the firm’s business activities.

ne of the key responsibilities of an Information Systems (IS) function in the late 1990s will be the development and maintenance of a technological highway that enables multifunctional systems to support the networked organizational society (Rockart, 1992). This technological highway is an information technology (IT) infrastructure. In fact, a survey of top IS managers suggests that developing and managing IT infrastructure resources is one of the ten critical tasks facing the IS function in the

* Lara Preiser-Houy is an Assistant Professor at the California State Polytechnic University, Pomona, CA.

Manuscript received May, 1996, revised July, 1996.

1990s (Niederman et at., 1991). However, the management of infrastructure resources is quite a challenge in light of the changing technological and business environments. The key challenges associated with building and sustaining an effective IT infrastructure are technical as well as managerial.

On a technical level, one of the challenges is to provide functional, efficient and easily maintainable IT infrastructure (Davenport and Linder, 1994). Furthermore, managers are faced with the challenge of matching infrastructure’s technical sophistication to the role infrastructure resources play in an organization (Weill et al., 1993). On a managerial level, the key challenges are to identify the benefits derived from infrastructure investments, to measure such benefits, and to justify additional infrastructure expenditures (McKay and Brockway, 1989; Weill, 1992a).

Preliminary results of an empirical investigation of infrastructure investments show not only that such expenditures represent a large component of an overall IT budget (i.e., 46 percent of total IT expenditures), but that the infrastructure investments appear to be increasing at the rate of 9.5 percent per annum (Weill et al., 1995). Yet to date, the literature provides very few guidelines on how to measure the value of such investments. One of the possible reasons for the lack of empirical investigations in this area is the difficulty of measuring the infrastructure benefits. The questions of how to justify flexibility or the future ability of an organization to respond quickly to competitive thrusts require answers, yet such answers are hard to come by. Hence, the current practices of measuring the value of an IT infrastructure appear to be more like an art that depends on gut feel and experience rather than on science. This lack of reliable measures as to infrastructure value makes it difficult for managers to justify additional investments in infrastructure. Yet, the increasing rate of infrastructure-related expenditures points to the need to find meaningful ways to identify and measure the value derived from infrastructure resources.

Purpose Of This Paper

The purpose of this paper is to present a multi-phased approach to assessing the economic viability of IT infrastructure investments. Specifically, the paper addresses three important questions related to the managerial challenge of providing an IT infrastructure. These questions are: (1) How does IT infrastructure provide value? (2) How does the firm assess the business value of its firm’s installed infrastructure resources? and (3) How does the firm justify additional investments in an IT infrastructure?

This paper first provides a comprehensive definition of an IT infrastructure. Next, it discusses the ways in which IT infrastructure may provide value to an organization. Then, it discusses a method to assess the business value of the firm’s existing infrastructure resources. Finally, it presents three methods for justifying additional infrastructure expenditures.

DEFINITION OF AN INFORMATION TECHNOLOGY INFRASTRUCTURE

The topic of an IT infrastructure is not a new one. For example, in the IT planning literature, at least five different authors discuss IT infrastructure in articles published between 1970 and 1984 (Boynton and Zmud, 1987).

1. Weill (1992a) defines an IT infrastructure as the base foundation of IT capability budgeted for and provided by the information systems function and shared across multiple business units or functional areas. The IT capability includes both the technical and managerial expertise required to provide reliable services (p. 8).

2. Davenport and Linder (1994) also suggest that an IT infrastructure consists of both technical and human components which are shared throughout the organization (Table 1).

3. Hoffman (1994) views an IT infrastructure as a set of technical and human resources that enable the organizational IT activities to take place.

Building upon the definitions of an IT infrastructure provided by Weill (1992a), Davenport and Linder (1994), and Hoffman (1994), an IT infrastructure may be conceptualized as an integrated set of technical components coupled with human skills and expertise to form a shared foundation for enabling IT-related activities that support an organization at the firm and business unit levels. Figure 1 depicts this conceptual view of an IT infrastructure. Table 2 provides a more detailed view of infrastructure components.

Table 1Davenport and Linder’s (1994) Five-Layer Model of an IT Infrastructure

IT Infrastructure Layer Technical Component Human Component

Core Technologies Wires and Boxes Basic Technical Skills

Technical Functionality Platforms and Utilities General Computing and Networking Skills

Business Applications Common Applications Application Development and Use Skills

Business Information Shared Data Information Management and Use Skills

Business Process Support IT Enablement of Business Processes Process Design and Execution Skills

Summary

An IT infrastructure is an enabling foundation for the firm and business unit IT activities of an organization. Such activities may, for example, include construction and operation of object-oriented application systems, provision of architectural standards for data, information, platforms and communication networks, as well as the provision of a secure IT environment throughout the firm. The IT activities enabled by the technical and human components of an IT infrastructure lead to the development and operation of applications that in turn provide business capabilities such as, interactive order-processing or materials planning and acquisitions.

The conceptual framework of an IT infrastructure presented in Figure 1 suggests that infrastructure resources impact organizational performance indirectly by enabling applications that provide business value. Thus, before the benefits derived from an IT infrastructure can be measured, it is important to understand the process of how infrastructure creates value. The next section addresses this topic in more detail.

VALUE PROVIDED BY AN IT INFRASTRUCTURE

Issues / Problems with Measuring the Infrastructure Value

IT infrastructure expenditures appear to grow from year to year as evidenced by preliminary research conducted by Weill et al. (1995) . Yet empirical studies of Chief Information Officers suggest that business leaders are not quite satisfied with the currently used practices to assess the actual payoff from their investments in technology (Wilson, 1993). To date, empirical findings of IT payoffs suggest that little is known about the actual effects IT capital investments have on organizational performance (Brynsjolfsson and Hitt, 1993). Unfortunately, even less is known about the payoffs from infrastructure investments, since empirical studies of IT value do not differentiate between infrastructure and other types of IT assets. Possible reasons for the lack of studies in this area is the difficulty of measuring the payoff derived from infrastructure resources. Measuring the value of an IT infrastructure is a difficult task because of the following:

IT infrastructure generates indirect, qualitative and at times uncertain benefits that can not always be easily measured using the traditional capital budgeting techniques (Weill, 1992a). For example, the benefits like increased flexibility, availability of real-time information to geographically-dispersed business entities, better architectural standards, just to list a few, are difficult to qualify and tie directly to the firms performance. Moreover, since the infrastructure investments are far removed from the performance benefits, many other factors like economy, industry structure, and firms competitive position, may impact the relationship between the infrastructure investment and the firms performance (Weill, 1992a).

IT infrastructure provides value to many different communities of interest (i.e., functions, departments, business units) and thus supports many different business purposes (McKay and Brockway, 1989). Moreover, different stakeholders may view the value derived from an IT infrastructure from different perspectives. For example, business unit developers may perceive firm-wide data architectural standards as constraining and potentially not useful, while corporate developers may argue that the realistic value of such standards has a positive value (i.e., the integrity and accuracy of the firm-wide database is 90 percent due to the data architectural standards as opposed to 50 percent accuracy before the standards were implemented).

The above problems suggest that measuring the infrastructure value is a difficult task. To date, few researchers and practitioners have succeeded in identifying the locus of infrastructure impacts and measuring such impacts. Does the lack of evidence as to the positive contribution of an IT infrastructure to organizational performance mean that infrastructure expenditures do not pay off? Brynsjolfsson (1993) once remarked that the shortfall of evidence as to the positive contribution of IT to productivity is not necessarily the evidence of a shortfall in IT productivity. Mismeasurement of IT inputs and outputs may be at the root of the IT productivity paradox dilemma. Similarly, the lack of evidence as to the positive contribution of infrastructure to organizational performance is not necessarily the evidence that infrastructure does not pay off. To uncover the payoff from infrastructure investments, it is important to first understand how an IT infrastructure creates value in an organization, and then to attempt measuring or estimating that value. In the next sub-section, a process model of how infrastructure creates value will be presented.

Toward a Model of an IT Infrastructure Value

The basic aim of the firm’s IT infrastructure is to enable the IT activities of an organization. These IT activities ultimately result in planning, building and operating many different applications that support and enhance current operations as well as enable the organization to achieve its strategies. There appear to be five key sources of value provided by an IT infrastructure in enabling the firm’s IT activities. These sources of value are as follows:

economy of scale and leveraging of existing IT investments (McKay and Brockway, 1989; Weill et al., 1993).

--economies of scale are derived from improved technology utilization, elimination of redundancy, and the standardization of the IT use and management practices.

Figure 1Conceptual View of an IT Infrastructure

Table 2

Detailed View of IT Infrastructure Components

IT Infrastructure Components Description

Technology Base Client/Server ArchitecturesLocal and Wire Area NetworksWireless TechnologyMainframes/Minis/MicrosSystems and Communications Software

Shared IT Services Applications Development Services(e.g., CASE, 4gls, object tools, project management tools, expert system shells)Data Warehouse ServicesNetwork Services

Standards/Security Architectural Standards for Hardware, Networks, Applications, Data and Information

Technical (internal and external) security provisions for networks, data, and applications

IT Human Skills and Expertise

Data ManagementApplication DevelopmentNetwork AdministrationStandards SettingTraining

SPRING 1996

foundation to build and run business applications that support current strategic intent of an organization (Hoffman, 1994; Weill et al., 1993).

flexibility to quickly respond to competitive trusts and changing business conditions (Duncan, 1995; Davenport and Linder, 1994; Hoffman, 1994; Weill et al., 1993).

architectural standardization throughout the organization (Davenport and Linder, 1994; Hoffman, 1994; Weill et al., 1993).

secure IT environment throughout the organization (Davenport and Linder, 1994; Hoffman, 1994; Weill et al., 1993).

Empirical observations of infrastructure used in organizations suggest that the expected benefits (i.e., value) derived from infrastructure, the level of infrastructure investment, and the approaches to justify such investments are contingent upon the role infrastructure plays in the organizational context (Weill et al., 1993). Furthermore, the more long-term the role is, the more difficult it is to quantify and assess the value derived from infrastructure resources. Weill et al. (1993) suggest three different views of infrastructure role that may be useful in categorizing infrastructure investments. These roles are utility, dependent and enabling. (Weill et al., 1993) characterize each role as follows:

Organizations that view an IT infrastructure as a utility consider infrastructure investments to be an administrative overhead expense and justify such expense on the basis of cost savings that result from economies of scale.

Organizations that view an IT infrastructure as a dependent resource consider infrastructure investments to be a business expense and justify such expenses on the basis of how well they enable the current strategy of the firm.

Organizations that view an IT infrastructure as an enabling resource consider infrastructure expenditures to be business investments and justify such investments on the basis of how well they enable current and future flexibility.

The empirical investigation of the relationship between investments in IT and firm performance in the valve manufacturing industry suggests that the effectiveness with which IT expenditures are converted into successful IT use, and not the magnitude of IT spending, is the key to deriving value from such expenditures (Weill, 1992b). Similarly, the results of a preliminary study of IT payoffs in the banking industry show a negative association of IT spending with firm performance and positive association of the accumulated IT assets (i.e., percentage of banking functions computerized) with the performance of large banks (Markus and Soh, 1993). Thus, it appears that the firms benefiting most from their IT spending are the ones that have effective internal processes of converting IT expenditures into a portfolio of applications that provide business functionality to the firm.

63

JOURNAL OF BUSINESS AND MANAGEMENT

The literature reviewed and synthesized so far indicates that the sources of infrastructure value are varied and are contingent upon the way an organization views the role of infrastructure resources. Moreover, the effectiveness with which such expenditures are converted into IT use contributes to the value of infrastructure assets. But how are the infrastructure expenditures actually converted into IT impacts that ultimately lead to improved organizational performance? A process theory model developed by Soh and Markus (1995) may help provide an answer to this question.

Soh and Markus’s model is depicted in Figure 2 and described below. It consists of the three distinct, yet tightly coupled processes by which IT creates business value in an organization. Each of these processes is comprised of elements that lead to the transformation of the original IT expenditures into improved organizational performance.

As depicted in Figure 2, the IT Conversion Process transforms IT expenditures into IT assets. The activity of managing IT resources influences the effectiveness with which IT expenditures are converted into IT assets which in turn are comprised of business applications, IT infrastructure and user skills (Soh and Markus, 1995). The model, however, does not differentiate between infrastructure-related expenditures and other types of IT investments.

Furthermore, the IT Use Process transforms IT assets into IT impacts. Among such impacts are redesigned business processes, improved decision-making at all levels of the organizational hierarchy, and new products and services. This transformation is moderated by the appropriateness of technology use in light of the organizational directives and culture. The model, however, does not differentiate between infrastructure-related IT impacts and other types of impacts.

Finally, the Competitive Process transforms IT impacts into improved organizational performance (i.e., productivity, market share). The industry structure, economic conditions and firm’s competitive position are the key factors that moderate and influence that transformation.

The process model of IT business value proposed by Soh and Markus (1995) is quite complex and thus difficult to test. That is why the authors suggest to break the model into three sub-processes for the purpose of empirical validation.

The empirical observations of Weill et al. (1993), and the framework of IT value developed by Soh and Markus (1995) provide a foundation upon which the model of how infrastructure expenditures create business value may be developed. This model is

64

SPRING 1996

depicted in Figure 3. The key characteristics of the proposed Infrastructure value model are as follows:

65

Figure 2How IT Creates Business Value: A Process Theory

Soh and Markus (1995)

Figure 3Conceptual Model of How Infrastructure Provides Value

SPRING 1996

IT expenditures are subdivided into two categories: infrastructure-related expenditures (e.g., hardware, systems software, communication links, etc.) and non-infrastructure-related expenditures (e.g., non-share application systems).

IT assets are subdivided into two categories: IT infrastructure (as defined above in the section titled Definition of IT Infrastructure) and business applications coupled with user skills.

The Infrastructure Support process converts the infrastructure assets into business applications. Such conversion takes place when the technology infrastructure enables the IT activities that lead to the development and implementation of business applications. The sources of infrastructure value are economy of scale, support of current strategies, flexibility, architectural standards and secure IT environment. Moreover, the process of converting infrastructure assets into business applications is contingent upon the role of an infrastructure in the organization and the effectiveness with which the infrastructure expenditures are converted into business applications. For example, undercapacity of infrastructure and lack of standards, may negatively influence the processes of building, operating and using business applications.

Business applications / user skills, and not the IT infrastructure component of IT assets, are transformed into IT impacts.

The model depicted in Figure 3 posits that infrastructure expenditures are converted into infrastructure resources. These resources in turn enable the IT activities of building, operating and using business applications that support the firm and business-unit activities of an organization. Furthermore, business applications and user skills are then transformed into IT impacts that lead to ultimate business value as captured by financial and productivity statistics. The Infrastructure Support process is contingent upon the effectiveness with which the infrastructure assets are converted into business applications and the role played by the infrastructure in an organization.

Summary

In this section I have described how IT infrastructure expenditures create business value in an organization. An IT infrastructure provides value to an organization by enabling the development of applications which support business activities of the firm. The sources of infrastructure value are contingent upon the role infrastructure plays in the organization (i.e., utility, dependent, enabling) and the effectiveness with which the infrastructure assets are converted into an applications portfolio. Once the sources of infrastructure value are identified, the next step is to assess the payoff from the existing infrastructure resources.

67

JOURNAL OF BUSINESS AND MANAGEMENT

A METHOD TO ASSESS THE COST AND VALUE OF THE FIRMS INSTALLED IT INFRASTRUCTURE RESOURCES

In assessing the payoff from the IT assets, management often looks for answers to the following questions: (1) How well are we doing today?, (2) How well are we positioned to compete in the future? (Wilson, 1993). The answer to the first question requires an understanding of how much the company is currently investing in various infrastructure components and the sources of value for each component. The answer to the second question depends on the level of current infrastructure investments and the role IT infrastructure plays in the organization. If the management determines that the level of current infrastructure investments is relatively low vis-à-vis the role infrastructure plays in the organization, the current infrastructure investment strategy may need to be re-evaluated. For example, if a firm with the enabling view of IT infrastructure spends only 2 percent of the total infrastructure investments to support future flexibility, the management would need to develop strategies to justify additional infrastructure-related investments.

One way to assess the cost and value of the firm’s installed IT infrastructure resources is to identify how much money was spent on each infrastructure component and to allocate the cost of each component to the various purposes served by the infrastructure in supporting the overall IT activities of the organization (Hoffman, 1994).

The purposes served by an IT infrastructure vary from company to company and depend on how the firm views the sources of infrastructure value. While some organizations may view the key source of infrastructure value to be in supporting the current strategy (i.e., enhancement and operation of current applications), other organizations may view future flexibility as the key source of value to be derived from the infrastructure resources. It is also likely that an IT infrastructure could serves different purposes in the same organization. If that is the case, the senior IS management may need to prioritize the importance of each purpose in light of the company’s strategy, and allocate infrastructure expenditures in accordance with the company’s goals and objectives.

Table 3 presents an example of infrastructure expenditures allocated to various purposes served by an IT infrastructure in creating value for a hypothetical manufacturing organization with annual revenue of $100 million. The first column of Table 3 details the individual components of an IT infrastructure. The second and third columns list the estimated budget amounts and the percentages of the total budget associated with each infrastructure component. The remaining columns detail the costs of each component allocated to five different purposes served by an IT infrastructure.

68

SPRING 1996

The purposes served by an IT infrastructure could be associated with the sources of infrastructure value as delineated in the previous section. For example, the purpose of applications support (Table 3, column 3) may be associated with the following sources of infrastructure value -- economies of scale and support of current strategy.

Table 3

Infrastructure Expenditures by an IT Infrastructure Purpose

(Thousands of Dollars) (source: Hoffman, 1994, p. 75)

Infrastructure Components

Budget Amounts

Percentage Support Applications

System Securit

y

Disaster Recovery

Support Flexibility

Support Standards

Operations Computers $1,250 25% $1,000 $125 $50 $75 Telecomm. 500 10 375 25 50 25 25 Data Admin. 150 3 75 30 15 15 15 Help Desk 100 2 90 10Total Operations $2,000 40 $1,540 $180 $115 $115 $50ApplicationDevelopment Maintenance $1,000 20 1,000 New Systems Marketing 450 9 270 45 45 45 45 Manufacturing 350 7 210 35 53 18 35 Human Res. 150 3 90 15 15 15 15 R& D 50 1 40 10Total new systems

$1,000 20 $610 $105 $113 $78 $95

Total application development

$2,000 40 $1,610 $105 $113 $78 $95

IT planning 250 5IT administration 500 10 350 50 50 50Training 250 5 175 75Total budget $5,000 100 $3,675 $335 $277 $317 $395% budget 100% 73% 7% 6% 6% 8%

Once this allocation process is complete, the management may be in a better to position to determine if adequate investments are undertaken to sustain the key sources of value to be derived from an IT infrastructure. For example, let us assume that a firm considers the provision of flexibility to accommodate changing business practices to be the key source of infrastructure value. The flexibility qualities of an IT infrastructure

69

JOURNAL OF BUSINESS AND MANAGEMENT

may be conceptualized as the compatibility of hardware resources, connectivity of networking resources, modularity of data and reuse of the application development tools and methodologies (Duncan, 1995). Such flexibility may enable the organization to quickly build new business applications, or to add new networking nodes to geographically dispersed business units to enable seamless transfer of critical information. If the review of the current infrastructure expenditures shows that only 6 percent of the total IT budget is spent for flexibility, the managers may either need to re-evaluate their expectations of infra- structure purpose or to prepare a justification for additional investments in infrastructure assets.

METHODS TO JUSTIFY ADDITIONAL INVESTMENTS IN AN IT INFRASTRUCTURE

Infrastructure investments, just like the public investments in the areas of flood control and public health, are characterized by large initial expenditures, payoff uncertainty, risk, and many intangible benefits allocated to a variety of stakeholders over a long period of time (Toraskar and Joglekar, 1993). Weill’s (1992a) empirical observations of IT infrastructure use and management practices suggest that organizations vary in their approaches to justify additional infrastructure expenditures. The approaches range from traditional capital budgeting techniques to gut feel that infrastructure assets are critical to the long-term well-being of the organization.

Weill (1992a) suggests that some organizations used discounted cash flow techniques to justify their infrastructure expenditures. The expenditures that returned a positive discounted cash flow were approved while the ones with a negative cash flow were rejected. On the other hand, other organizations justified their infrastructure investments on the basis that such investments were required to achieve current or future strategies. Such justifications relied less on tangible, quantitative estimates of infrastructure value (e.g., reduced cost of selected IT services, fewer errors per line of code, etc.) and more on qualitative, intangible sources of value (e.g., future ability of an IS organization to build specific business applications, improved flexibility through reuse and sharing of resources).

The approach to justify IT infrastructure investments as well as the level of that investment and the expected benefits derived from such investment depend on the role IT infrastructure plays in the organization (Weill et al., 1993). For example, an organization that views IT infrastructure as a utility resource aimed to providing cost savings via economies of scale, may be better suited to use the net present value techniques to justify infrastructure investments since the inflows/ outflows of money

70

SPRING 1996

and discounted rate could be reasonably estimated. On the other hand, the organization that views IT infrastructure as enabling resource aimed at providing flexibility to quickly respond to changing business needs, would be better off using other justification techniques since it may not be possible to quantify ahead of time the benefits to be derived from infrastructure investments.

Next, the methods to justify the investments in IT infrastructure resources will be presented for each of the three roles infrastructure plays in the organization - utility, dependent and enabling.

Investment Justification Method for the Utility Infrastructure

Organizations in which IT infrastructure plays a utility role consider infrastructure investments to be an administrative overhead expense and justify such expense on the basis of cost savings that result from economies of scale (Weill et al., 1993). Once the cost savings of the IT activities that result from the proposed infrastructure investments are identified, the management may use the net present value on discounted cash flow method to justify infrastructure investments.

Net Present Value Method (NPV)

Net Present Value is a traditional capital investment technique used by many organizations to justify investments in manufacturing and other types of capital equipment (Weston and Brigham, 1985). The NPV formula is shown in Figure 4. Managers may implement the NPV technique to justify infrastructure-related investments by deriving the present value of the net cash flows associated with such investments (e.g., cost savings that result from economies of scale), discounting such cash flows at the appropriate cost of capital, and subtracting from the resulting amount the cost of infrastructure resources. The NPV method is useful in the infrastructure justification decisions only if the inflows and outflows of money and the discount rate could be reasonably estimated.

Pros and Cons of the Net Present Value Method

The key advantage of using the NPV method to justify infrastructure investments is that this method provides quantitative estimates of the investment’s value. Such estimates are appealing to managers who may be more comfortable with financial measures of capital performance than intangible, qualitative and gut feel measures. The relative ease of use is yet another advantage of this method.

The key disadvantage of using the NPV method to justify infrastructure investments is related to the accuracy of such method. Some elements of an IT infrastructure (e.g.,

71

n NPV = [ CF1/ (1 + k ) 1 + CF2/ (1+ k)2 + ... CFn/(1+ k)n] - I = CF t / (1 + k) t - I.

t=1where CF1...CFn are net cash flowsk is the appropriate discount rate (i.e., project's

cost of capital)I is the initial cost of the projectn is the project's expected life

JOURNAL OF BUSINESS AND MANAGEMENT

IS staff expertise) are difficult to capitalize (Kriebel, 1993). Yet even if the investment costs are capitalized, the NPV estimates may still lack accuracy. The reasons is that infrastructure investments often pay off in ways that are very difficult to quantify. For example, quantifying improvements in software development productivity attributed to the proposed integrated computer aided software engineering (ICASE) tool is difficult at best. Some investment expenditures may yield a negative return on investment, yet be critical for the long-term competitive stance of an organization.

Investment Justification Method for the Dependent Infrastructure

Organizations in which IT infrastructure plays a dependent role consider infrastructure investments to be a business expense and tend to justify such expense on the basis of how well the infrastructure resources enable the current strategy of the firm (Weill et al., 1993). So the sources of value derived from a dependent infrastructure include, but are not limited to the following: expected benefits from applications that support current strategies, architectural standardization, and secure IT environment. While some of these benefits are more or less tangible (i.e., quality of systems as measured by errors per line of code), others are not (i.e., architectural standardization). A broadly-defined Cost-Benefit Analysis Methodology (Toraskar and Joglekar, 1993) may be applicable to justify the infrastructure investments that have both tangible and intangible benefits.

Figure 4

Net Present Value Formula

(source: Weston and Brigham, 1985, p. 339)

Cost-Benefit Analysis (CBA) Methodology

CBA methodology has been applied for many years to justify public investment decisions in areas such as flood control and national defense (Toraskar and Joglekar, 1993). Such investment decisions are complex, uncertain and translate into tangible as well as intangible benefits that accrue to many different stakeholders. Toraskar and Joglekar (1993) suggest that strategic IT investment decisions have very similar characteristics with the public investment decisions. Decisions to invest in the IT infrastructure that enable current strategy are also complex, involve tangible as well as intangible benefits, and affect many different stakeholders.

72

SPRING 1996

The CBA approach to assess cost/benefit of IT infrastructure investments is summarized in Figure 5. This approach includes five phases as recommended by Toraskar and Joglekar (1993):

identification of changes associated with IT infrastructure investments (e.g., reduction in costs associated with providing a data administration IT service; improved integration of business applications)

measurement of tangible changes (e.g., reduction of software errors per 100 lines of code) and comprehensive description of changes that can not easily be measured (e.g., improved transportability of applications across multiple platforms)

explicit valuation of both tangible and intangible changesadjustment of explicit values for timing and uncertainty of their occurrence (e.g., discount rates and risk premiums)

final assessment of infrastructure investment alternatives

Pros and Cons of the CBA Method

The key advantages of using the CBA method to justify IT investments are as follows (Toraskar and Joglekar, 1993):

the method measures intangible as well as tangible costs and benefits, the method takes into consideration the multi-dimensional nature of infrastructure

investments; such investments affect many different stakeholder groups (e.g., IS professionals who provide IT-related services and clients who use business applications and networking services), and

the method provides a comprehensive methodological base that includes tools, techniques and guidelines on when to use various tools and techniques.

The CBA method, however, is not without faults. The IT investment literature criticizes this method for its emphasis on quantification of intangibles, disregard for the nature of strategic IT investments that are risky and uncertain, ignorance of non-economic issues (i.e., quality of life), focus on efficiency not effectiveness, and a bias towards investments that have a short-term payoff (Toraskar and Joglekar, 1993). Some of the above limitations may be avoided by broadening the tool kit of alternative techniques. Moreover, being cognizant of the above problems is the first step towards avoiding some of them as the IS managers attempt to apply the CBA method to justify investments into dependent IT infrastructure.

Investment Justification Method for the Enabling Infrastructure

73

JOURNAL OF BUSINESS AND MANAGEMENT

Organizations in which IT infrastructure plays an enabling role consider infrastructure expenditures to be a business investment (Weill et al., 1993). Furthermore, they justify such an

74

Figure 5Flowchart of the Cost-Benefit Analysis Methodology

JOURNAL OF BUSINESS AND MANAGEMENT

investment on the basis of how well it enables the current and future flexibility to achieve the strategic aims of the firm. Infrastructure investments today may often be necessary for future investments in transactional, informational and strategic applications (Weill, 1992b). Thus, investing into enabling infrastructure resources is like investing into organizational capabilities that may be required in the future to sustain or gain competitive advantage. But how could the firm justify investments in an IT infrastructure that provides an option to invest in future applications that in turn would provide value to the organization in the long run? Clearly, there is a need to value such an option.

Kambil et al. (1993) propose a framework based on the financial options theory to value the IT investments real option, that is an opportunity to implement strategies for future growth. This framework could be applied to the IT infrastructure context. The framework suggests that the value of an infrastructure investment option would be in enabling future value-adding IT activities to take place.

What is a real option and how could it be valued? Kambil et al. (1993) define a real option as follows:

... implicit contracts to exercise new information systems based business strategies during the life-time of a specific investment, or to expand or adapt existing projects and strategies to changing environmental contingencies (p. 165).

The formula to estimate the value of a real option provided by an IT infrastructure investment is shown in Appendix A. To illustrate the above framework an example of a hypothetical company faced with the need to justify an IT infrastructure investment will be presented. The company, XYZ Unlimited, has yearly revenues of approximately $10 million and specializes in manufacturing of office supplies. The Chief Executive Officer of the company recognizes the need to reduce the time it takes for his organization to respond to market needs, and to shorten new product development cycle. He wants the development of new business systems that would enable his strategy. Thus, there is an increasing pressure on XYZ’s IS organization to develop future business applications that keep pace with changing organizational needs.

The company’s Chief Information Officer (CIO), in reviewing the way business applications are developed, finds that the traditional software development process employed by his organization often leads to budget/time schedule overruns and failed systems. Furthermore, he believes that an integrated computer-aided software engineering environment (i.e., ICASE tools and methodologies) could improve the timeliness of the future systems development effort, and the effectiveness as well as the

76

SPRING 1996

flexibility of the newly developed software. Thus, the CIO is faced with the need to justify the investment in the IT infrastructure (i.e., ICASE tool, methodology, training) that would enable a more effective development of the future business applications.

The estimate to acquire an option on future business applications enabled by the investments in the ICASE environment will be derived following the steps suggested by Kambil et al. (1993). These steps are detailed in Appendix B. The estimates derived in Appendix B suggest that the value of an option ($113,251) exceeds the present value of the initial investment Ip = $100,000. Since the infrastructure investment appears to add positive value to the firm, the CIO would now be in a better position to justify the investment in the ICASE environment.

Pros and Cons of the Real Option Method

The key advantage of using the real option framework to justify future infrastructure investments is that such a framework provides a systematic approach to estimate and quantify the value of yet uncertain future projects made possible by the initial investment in an IT infrastructure. Moreover, the real option approach enables the organization to more effectively align its business, IT and financial strategies aimed at increasing the overall value for the organization (Kambil et al., 1993).

The downside of applying the real option method to justify infrastructure investments is that this method appears to be quite complex. Moreover, the effective use of this method requires fairly robust estimates of the expected cash flows and the cost of exercising the real option (Kambil et al., 1993). It may be difficult to appropriately estimate the cash flows from infrastructure investments since IT infrastructure enables many future projects and provides value to many different stakeholders.

Summary

This section has presented three different methods to justify future investments in IT infrastructure -- net present value, cost-benefit analysis and the real option. While each of the three justification methods has its advantages and limitations, they nevertheless are viable approaches to estimate the tangible as well as the intangible nature of infrastructure expenditures.

CONCLUSION

77

JOURNAL OF BUSINESS AND MANAGEMENT

Investments in IT infrastructure resources are usually large, uncertain, and risky. Moreover, the benefits derived from infrastructure are indirect, largely qualitative, and accrue to many different stakeholders. Thus, it is difficult to tie infrastructure investments directly to organizational performance. Consequently, the difficulty with ascertaining the infrastructure value makes it even more difficult for the managers to justify additional infrastructure investments. The three-phased approach to assess the economic viability of infrastructure resources presented in this paper is aimed at helping IS and general managers make more informed decisions as to their infrastructure investments. The proposed approach consists of the three interdependent phases. The first phase deals with identifying the sources of infrastructure value. The second phase deals with assessing the cost and value of the firm’s installed infrastructure resources. The third phase deals with justifying additional investments in an IT infrastructure.

The key managerial challenge associated with developing and sustaining a robust foundation for the firm’s business activities is assessing infrastructure payoff. The application of the payoff assessment approach proposed in this paper is one of the first step towards addressing this challenge.

78

C = { Cu [(r-d) / (u-d)] + Cd [(u-r) / (u-d)] } / r

where:

Cu = max[ 0, u*S - K ] present value of the call option based on the optimistic scenario (note: optimistic and pessimistic scenarios are derived based on he identification of project risks)

APPENDIX A

Formula to Estimate the Value of a Real Option Provided by an IT Infrastructure Investment

APPENDIX B

Steps to Estimate the Value of Acquiring an Option on Future Business Applications Enabled by the Investments in the Integrated Computer Aided

Software Engineering (ICASE) Environment

Step 1 Derive the traditional net present value estimate for the project of developing future business applications using the ICASE environment

Step 1a Estimate Project Costs

the ICASE tool, methodology, training (i.e., IT infrastructure) would cost $100,000; application programming for the new business systems would cost $120,000; Io = $220K

Step 1b Identify Project Risk

the key organizational risk associated with this project is the acceptance of the new ICASE environment by the software developers. Empirical literature suggests that the implementation of CASE tools is often met by the software developers with resistance and even sabotage (Orlikowski, 1989).

Step 1c Estimate Cash Flows Under Different Scenarios

optimistic scenario is that the ICASE environment will be successfully accepted by the software developers of the XYZ Unlimited organization (40 percent probability); estimated perpetual annuity in savings from this project are $80,000; these savings will be realized beginning one year from the initial investment.

pessimistic scenario is that the ICASE environment will not be successfully accepted by the software developers (60 percent probability); estimated perpetual annuity in savings will be only $10,000; these savings will be realized beginning one and a half year from the initial investment

Step 1d Estimate the Cost of Capital

the cost of capital is assumed to be 20 percent as suggested by Kambil et al. (1993)

Step 1e Estimate Traditional Net Present Value

Formula for the present value of the perpetual annuity that begins a year from now is PV = Annual cash flow / r (Kambil et al., 1993, p. 178)

Formula for the expected present value of the project is as follows (Kambil et al., 1993, p. 169)

E(PV) = (PV optimistic) * optimistic probability + (PV pessimistic) * pessimistic probability

Formula for the net Present Value of the project is:NPV = E(PV) - Io

The following calculations are based on the above formulas:

PV (optimistic scenario) = $ 80,000 / 20% = $400,000PV (pessimistic scenario) = $ 10,000 / 20% = $

E(PV) = $400,000 * 0.4 + $50,000 * 0.6 = $ 190,000NPV = E(PV) - Io = $190,000 - $220,000 = - $ 30,000

Step 1f Conclusions from the NPV Analysis

The proposed investment has a negative net present value. Based on this estimate, the project should not be authorized. The next step, however, is to estimate the value of acquiring an option on future business applications by investing in IT infrastructure. If the value of this option exceeds the present value of the initial investment, the CIO may be in a better position to justify infrastructure investment.

Step 2 Estimation of Project Cost

Initial project investments (i.e., infrastructure investments) is $100K; thus Ip = 100K

The cost to expand the project a year from now (i.e., building new applications, implementation, training, conversion, etc.) is $120K

Step 3 Defining the Option

Investing in the ICASE environment (i.e., IT infrastructure) may be compared to obtaining a single period call option on the new business systems development project

Cu = max[0, u*S - K] = max[0, 480K - 120K] = $360KCd = max[0, d*S - K] = max[0, 60K - 120K] = $0u*S = (1 + 1.2) * 400,000 = $480Kd*S = (1 + 1.2) * 50,000 = $60K u = 2.53; d = .32; r = 1.05 (risk-free rate of return)

Step 4 Estimating the Option Value

C = { 360K[(1.05 - .32) / (2.53 - .32)] + 0K [(2.53 - 1.05) / (2.53 - .32)]} / 1.05 = $113,251

Step 5 Investment Decision

REFERENCES

Boynton, A.C. & Zmud, R.W. (1987). Information Technology Planning in the 1990’s: Directions for Practice and Research, MIS Quarterly, 11(1), 100-115.

Brynjolfsson, E. (1993). The Productivity Paradox of Information Technology, Communications of the ACM, 36(12), 66-77.

Brynjolfsson, E. & Hitt, L. (1993). Is Information Systems Spending Productive? New Evidence and New Results. In DeGross, J.I. , Bostrom, R. P. & Robey. D.

(Eds.). Pp. 47-64. Proceedings of the Fourteenth International Conference on Information Systems, Orlando, Florida.

Davenport, T. & Linder, J. (1994). Information Technology Infrastructure: The New Competitive Weapon? Working Paper, Ernrst & Young, 1-26.

Duncan, N. B. (1995). Capturing Flexibility of Information Technology Infrastructure: A Study of Resource Characteristics and their Measure. Journal of Management Information Systems, 12(2), 37-57.

Hoffman, G.M. (1994). Technology Payoff: How to Profit with Empowered Workers in the Information Age, Burr Ridge: Irwin Professional Publishing.

Kambil, A., Henderson, J. & Mohsenzadeh, H. (1993). Strategic Management of Information Technology Investments: An Options Perspective. In Banker, R.D., Kauffman, R. J. & Mahmood, M. A. (Eds.). Strategic Information Technology Management: Perspectives on Organizational Growth and Competitive Advantage. Pp. 161-178, Harrisburg: Idea Group Publishing.

Kriebel, C.H. (1993). Formal Models in Research on IT Investment Evaluation. In Banker, R.D., Kauffman, R. J. & Mahmood, M. A. (Eds.). Strategic Information Technology Management: Perspectives on Organizational Growth and Competitive Advantage. Pp. 25-31, Harrisburg: Idea Group Publishing.

Markus, M.L. & Soh, C. (1993). Banking on Information Technology: Converting IT Spending into Firm Performance. In Banker, R.D., Kauffman, R. J.& Mahmood, M. A. (Eds.). Strategic Information Technology Management: Perspectives on Organizational Growth and Competitive Advantage. Pp. 375-403, Harrisburg: Idea Group Publishing.

McKay, D.T. & Brockway, D.W. (1989). Building I/T Infrastructure for the 1990s, Stage by Stage, 9(3), 1-11.

Niederman, F., Brancheau, J.C. & Wetherbe, J.C. (1991). Information Systems Management Issues for the 1990s, MIS Quarterly, 15(4), 475-495.

Orlikowksi, W.J. (1989). Division Among the Ranks: The Social Implications of Case Tools For Systems Developers. In DeGross, J.I., Henderson, J. & Konsynski, B. (Eds.). Pp. 199-210, Proceedings of theTenth International Conference on Information Systems, Boston, Massachusetts.

Rockart, J.F. (1992). The Line Takes the Leadership - IS Management in a Wired Society, Sloan Management Review, 33(4), 47-54.Soh, C. and Markus, M. L. (1995). How IT Creates Business Value: A Process

Theory Synthesis. In DeGross, J. I., Ariav, G., Beath, C., Hoyer, R. & Kemerer,

C. (Eds.). Proceedings of the Sixteenth International Conference on Information Systems, Pp. 29-41, Amsterdam, The Netherlands.

Toraskar, K. & Joglekar, P. (1993). Applying Cost Benefit Analysis Methodology for Information Technology Investment Decisions. In Banker, R.D., Kauffman,

R. J.& Mahmood, M. A. (Eds.). Strategic Information Technology Management: Perspectives on Organizational Growth and Competitive Advantage. Pp. 119-142, Harrisburg: Idea Group Publishing.

Weill, P., Broadbent, M., Butler, C. & Soh, C. (1995). An Exploration of Firm-Wide Information Technology Infrastructure Investment and Services. In DeGross,

J. I., Ariav, G., Beath, C., Hoyer, R. & Kemerer, C. (Eds.). Proceedings of the Sixteenth International Conference on Information Systems, Amsterdam, The Netherlands.

Weill, P., Broadbent, M. & St. Clair, D. (1993). IT Value and the Role of IT Infrastructure Investments, Working Paper, 1-26.

Weill, P. (1992a). The Role and Value of Information Technology Infrastructure: Some Empirical Observations. Working Paper, #240, Center for Information Systems Research, Sloan School of Management, Massachusetts Institute of Technology, 1-33.

Weill, P. (1992b). The Relationship Between Investment in Information Technology and Firm Performance: A Study of the Valve Manufacturing Sector, Information Systems Research, 3(4), 307-331.

Weston, J.F. & Brigham, E. F. (1985). Essentials of Management Finance, Seventh Edition, New York: CBS College Publishing.

Wilson, D. D. (1993). Assessing the Impact of Information Technology on Organizational Performance. In Banker, R.D., Kauffman, R. J.& Mahmood, M. A. (Eds.). Pp. 471-514, Strategic Information Technology Management: Perspectives on Organizational Growth and Competitive Advantage . Harrisburg: Idea Group Publishing.

R

USING DATA ENVELOPMENT ANALYSIS TO

EVALUATE TAX PREPARATION SOFTWARE

Dorothy M. Fisher *

Steven A. Fisher **

D. Bruce Sun ***

This study uses Data Envelopment Analysis (DEA) to analyze and compare the performance of 18 professional tax preparation software packages. The DEA model connects price and setup cost (inputs) with operating capabilities (outputs) to evaluate the relative performance of individual software packages. DEA does not require a set of preassigned weights for inputs and outputs and, thus, overcomes the deficiency introduced by using arbitrary weights. The findings of this study provide professional accountants with nonsubjective assessments of professional tax preparation software packages.

ecent advances in information technology are having a profound impact on the preparation of income tax returns. Tax preparation software is replacing paper and pencil as the primary means for preparing individual and corporate tax returns. Tax software enables the professional accountant to input data quickly and accurately, with immediate returns, analyses, and electronic filing.

* Dorothy M. Fisher is a Professor of Computer Information Systems in the School of Business at California State University, Dominguez Hills, Carson, CA.

** Steven A. Fisher is a Professor of Accountancy at California State University, Long Beach, Long Beach, CA.

***D. Bruce Sun is a Professor of Information Systems at California State University, Long Beach, Long Beach, CA.

Manuscript received April, 1996, revised, August, 1996.

The evolution of tax preparation software has lead to a wide range of products. At the low-end is software with limited capabilities and capacity which allow for little customization. Such software may be adequate for the narrow needs of individuals and small firm users who have relatively few transactions and simple tax problems. However, many professional accountants require high-end tax software that possesses the capabilities, capacity, and flexibility necessary to meet complex tax problems of individuals and businesses (Courtney and Flippen, 1995).

An important task of the accountant in preparing for the tax season is to select the best tax preparation software. With more than 100 tax software packages on the market, selection is not an easy task. The professional accountant must match client needs with appropriate software. Frequently, the best software is not readily apparent. The tax preparer must gather relevant information from vendors and other sources and compare the relative merits of different software packages (Cohn, 1995).

Oftentimes, the software selection decision depends on the professional accountant's judgment. However, individual tax preparers have difficulty in properly processing numerous attributes of many competing software packages. Additionally, individual biases may affect the decision. Furthermore, the tax preparer is confronted with rapidly changing products, which means that the decision must be frequently reevaluated. Therefore, it is not surprising that many software decisions are subjective and influenced by the strength of the sales pitch (Stearman, 1992).

As a consequence, the professional accountant frequently seeks independent objective assessments of the relative merits of tax preparation software packages. One problem with these assessments is that they rely on arbitrary weightings of the relative costs and capabilities of the competing packages. This paper presents a DEA evaluation of 18 leading tax preparation software packages. DEA is advantageous in that it does not require preassigned weights for package costs and capabilities and, thus, overcomes the deficiency introduced by using arbitrary weights. The results should be useful to professional accountants in identifying the best package to meet their needs.

TAX PREPARATION SOFTWARE

Professional accountants have three alternatives available for preparing income tax returns: manual preparation, service bureaus, and in-house tax preparation software. A recent survey of professional accountants indicates that in-house tax preparation software has gained 80.6 percent share of the market, while manual preparation and service bureaus have the remaining share (Nelson and Langer, 1994). This is surprising since PC-based tax preparation software was not available to the professional accountant until the 1980's. The advent of personal computing has seen tax software evolve into complete systems for tax preparation, planning, and analysis. Software is available for the preparation of Federal as well as state income tax returns.

Furthermore, in recent years the cost of tax preparation software has dropped significantly, making it feasible even for the smallest accounting firms (Langer, 1995).

Tax preparation software must be developed within the framework of the Internal Revenue Code and various state income tax regulations. All software packages possess generalized capabilities for preparing basic individual and business tax returns. However, this does not mean that all software packages are alike. Packages differ in the extent of their detailed operating capabilities, in providing complete sets of returns, and in tax planning and analysis (Stearman, 1993).

More specifically, operating capabilities that distinguish tax software packages include electronic filing, installation and setup, analysis, processing, on-line help, and printing. Other relevant software capabilities are the number of individual, estates, and business forms included and the number of states available. These distinguishing capabilities are the subject of analysis in this study (Stearman, 1995).

Tax software must be updated annually due to changing income tax regulations and technological advances. As a consequence, the professional accountant must review his tax software selection every year. The professional accountant must continually keep abreast of new products and the latest developments in the tax software field.

DATA ENVELOPMENT ANALYSIS

In general, surveys of tax preparation software packages use arbitrary weights for individual criteria to arrive at a weighted score to measure the overall performance for each package (Giorgis, 1993; Marshall, 1993). Furthermore, the selection of criteria used to evaluate tax software packages is subjective. Potential deficiencies in an existing set of weights include bias and inconsistency with organization or user objectives.

DEA is a nonparametric methodology. It requires neither an explicit formulation of the underlying functional relationships nor preassigned weights for outputs and inputs in evaluating a production operation in a multiple-output, multiple-input setting (Banker et al., 1994; Schefczyk, 1993; Sun et al., 1993). Therefore, DEA can avoid certain theoretical and computational problems.

In DEA convention, a production operation using m inputs to produce s outputs is called a DMU (decision making unit). A DMU has discretion in using an input -mix to produce an output-mix. From a variety of DEA models, we chose the BCC (Banker, Charnes, and Cooper) model as formulated below (Banker et al., 1984) to evaluate tax preparation software packages.

Maximize VP = T yo

Subject to T yj - Txj < 0, j = 1, ... n T xo = 1 (1) T, T > 0

where n is the number of DMUs; xj is the input vector; yj is the output vector; DMU o is the DMU currently being evaluated; and correspond to x o and yo respectively. In this study, outputs are operating capabilities and inputs are price and setup cost. The DEA model evaluates each tax package relative to the peer group one at time.

The dual of the above linear programming problem (2) takes the following form

Minimize VD =

Subject to Install Equation Editor and double-click here to view equation. yj j - S+ = yo

Install Equation Editor and double-click here to view equation. xjj - x o - S-= 0 (2)

Install Equation Editor and double-click here to view equation. j = 1,

j > 1

S+, S- 0

where S+ is the output slacks; S- is the input slacks; is a coefficient vector for DMUs. Obviously there exist optimal solutions for (1) and (2), and VD <VP< 1.

Definition of Efficiency: Let (*, *) denote an optimal solution to (1). DMUo is said to be efficient if T yo

= 1 where * > 0 and * > 0. Alternatively, the efficiency of DMUo can be measured in terms of the dual problem (2). DMUo is efficient if * = 1, S+* = 0, and S-* = 0 where (*, *, S+*, S-*) is an optimal solution to problem (2). For an efficient performance, DMUos optimal inputs and outputs should be ( x o*, y o*) where x o* = *x o - S-* and y o* = yo + S+*. Therefore, the input wastes are (1- *) xo - S-* and corresponding output shortfalls are S+*.

From the Definition of Efficiency, when *= 1, S+* =S-*= 0, then x o* = x o, and y o* = y o, i.e., optimal values equal observed values. Otherwise, DMUo can improve its productivity by eliminating the input wastes and/or increasing the outputs. DEA provides the direction for improvement by pinpointing the specific components of deficient production.

Table 1

Tax Preparation Software Packages Evaluated

Program Vendor Phone

A-Plus-Tax Arthur Andersen (800) 872-1040

CPASoftware CPA Software (904) 434-2685

Digitax Cold River Software (800) 432-1065

LMS/Tax SCS/Compute (800) 488-0779

Lacerte Lacerte Software (800) 765-7777

Package EX Exac Tax (800) 352-3638

PencilPushers Damirus (800) 370-2500

Professional Tax System TAASC (918) 493-6500

ProSeries Intuit (800) 934-1040

ProSystem fx CCH (800) 457-7639

RAM Ram Software (800) 888-6217

Tax Machine SCS/Compute (800) 326-1040

Tax/Pack Professional Alpine Data (800) 525-1040

Tax Relief Micro Vision (800) 829-7354

TaxSimple TaxSimple (800) 323-2662

TaxWorks Laser Systems (800) 230-2322

Ultra Tax Creative Solutions (800) 968-8900

Veritax Cold River Software (800) 837-4829

JOURNAL OF BUSINESS AND MANAGEMENT

DATA

Eighteen high-end professional tax preparation software packages were chosen for analysis. Although not a complete survey of all packages on the market, those included in this survey represent the leading software packages in terms of sales volume. The Journal of Accountancy is the primary source of information on the tax packages. Table 1 presents a listing of the packages included in this study (Stearman, 1995).

The general areas of operating capabilities and characteristics captured from each of the eighteen tax software packages include general features, electronic filing, analysis, processing, printing, states available, individual forms available, business forms available, and installation and setup. These operating capabilities are supported by 478 detailed operating functions which are the primary focus of analysis. The individual detailed operating functions are too numerous to list in this article. Therefore, examples are presented for just two capabilities. The general features capability is concerned with whether or not a package has operating functions for preparing estimated payment vouchers, amended returns, and net operating loss schedules for Federal as well state returns. Electronic filing capability concerns whether or not a package has operating functions for providing direct filing using a third party, refund anticipation loans, state electronic filing, and electronic filing for business packages.

Furthermore, a complete package enables the preparation of all Federal and state tax forms. However, many packages lack operating functions for providing full sets of Federal tax forms and for preparing returns for all states that levy an income tax. Some packages permit the preparation of returns only for states with relatively high populations and significant economic activity. For example, California and New York are covered by all the packages, while North Dakota and Montana are provided for in relatively few packages. Hence, tax software packages are frequently aimed only at the high sales volume markets.

For each package, detailed operating functions were evaluated through a series of yes-no response-type questions to determine whether or not individual packages possess certain operating features and capabilities. For example, with respect to Electronic Filing, one question is Direct filing capabilities? For the Analysis capability, one question is Tax planning supported? These and similar questions provide the basis for differentiating among the software packages.

A tally of the yes responses was made for each of the capabilities for the individual software packages. This tally is the basis for evaluating the relative

90

SPRING 1996

performance capabilities of the packages. The greater the number of yes responses, the greater the performance capability of the software package.

PERFORMANCE EVALUATIONS

For evaluation of software performance, price and setup cost are designated as inputs while operating capabilities represent outputs. Since tax software can be purchased in individual modules, separate analyses are made for (1) individual tax returns-1040, (2) business returns, (3) business and estate returns, (4) individual and business returns, and (5) a bundled package of individual, business, and estate returns. Outputs are reflected in the operating capabilities of electronic filing, analysis, processing, on-line help, printing, number of states available-1040, number of states available-business, number of states available-estates, forms-1040, forms supported, and forms-business.

Individual modules of a package are evaluated relative to their peer group. The price of a module is associated with a differing combination of outputs. For example, a complete tax preparation package includes all operating capabilities, while a business returns module excludes capabilities related to individual and estate returns. Thus, the analyses provide for an evaluation of the relative performance capabilities of the individual modules as well as the package as a whole.

To compute DEA efficiency scores, the DEA model (1) presented in the section on Data Envelopment Analysis is used. Based on the input-output mix used in the computation, the model generates the optimal multiplier vector (or the vector of shadow prices) for DMUo to reach an optimal level of performance. The optimal multiplier vector is actually a set of individual weights assigned by DMU o to the inputs and outputs it is using. Comparing this set of weights with the optimal multipliers of other DMUs shows the relative strengths and weaknesses of DMUo in terms of possible input/output slacks.

The DEA efficiency scores for individual and combined sets of modules are presented in Table 2. The results for 1040 module indicate that ProSeries, Tax/Pack Professional, TAXSIMPLE, and Veritax have efficiency scores of 1. These packages provide professional accountants with the greatest value per dollar in preparing individual tax returns. Other 1040 tax software modules are deficient in one or more operating capabilities relative to these efficient modules. Similarly, the most efficient business returns modules are Professional Tax Systems, RAM, Tax/Pack Professional, Tax Relief, and TaxWorks. The results for the combined 1040 and business returns, indicate that the packages with efficiency scores of 1 are Professional Tax System, Tax/Pack Professional, Tax Relief, TaxWorks, and Ultra Tax. Furthermore, the most

91

JOURNAL OF BUSINESS AND MANAGEMENT

efficient packages for the combined business and estates modules include Lacerte, Package EX, Professional Tax System, ProSeries, Tax Relief, TAXSIMPLE, and TaxWorks.

92

Table 2

Ranking Based on DEA Efficiency Scores

Tax Software 1040 Business 1040 & Business Business & Estates Complete Package

Package Efficiency Ranking Efficiency Ranking Efficiency Ranking Efficiency Ranking Efficiency Ranking

Anderson 0.3277 13 0.4628 12 0.4407 12 0.5757 7 0.5815 8CPA Software 0.5356 8 0.8657 2 0.7946 5 n.a. n.a. n.a. n.a. Digitax 0.2299 15 0.7031 9 0.4731 11 0.9527 2 0.8202 5LMS/Tax 0.5182 9 0.6288 11 0.6324 9 0.7235 6 0.7151 7Lacerte 0.4889 11 0.861 3 0.7683 6 1.0000 1 1.0000 1Package EX 0.5407 7 0.7598 7 0.6906 8 1.0000 1 0.9595 2Pencil Pushers 0.6189 5 0.7537 8 0.7624 7 0.9163 3 0.8798 4Professional 0.6233 4 1.0000 1 1.0000 1 1.0000 1 1.0000 1ProSeries 1.0000 1 0.7708 6 0.8673 3 1.0000 1 1.0000 1ProSystem 0.3058 14 0.4425 13 0.4530 15 0.5097 8 0.5503 9RAM 0.5465 6 1.0000 1 0.8402 4 n.a. n.a. n.a. n.a. Tax Machine 0.3703 12 0.6408 10 0.5804 10 0.8304 5 0.7395 6Tax/Pack Prof. 1.0000 1 1.0000 1 1.0000 1 n.a. n.a. n.a. n.a.Tax Relief 0.9284 2 1.0000 1 1.0000 1 1.0000 1 1.0000 1TAXSIMPLE 1.0000 1 0.8547 4 0.9905 2 1.0000 1 1.0000 1TaxWorks 0.7002 3 1.0000 1 1.0000 1 1.0000 1 1.0000 1Ultra Tax 0.4977 10 0.8332 5 1.0000 1 0.9146 4 0.9472 3Veritax 1.0000 1 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

JOURNAL OF BUSINESS AND MANAGEMENT

It is interesting to note that the Ultra Tax-1040 module and the Ultra Tax-business module individually are deficient relative to the cheaper modules. However, Ultra Tax's combined module for 1040 and business returns is able to achieve a perfect efficiency score because the price of the combined module is competitive. This information is valuable to software vendors. The implication is that to be competitive, Ultra tax should lower its price for the 1040 module and the business module but not for the combined 1040 and business module.

The results show that for the complete packages, Lacerte, Professional Tax System, ProSeries, Tax Relief, TAXSIMPLE, and TaxWorks have efficiency scores of 1. These software packages provide the professional accountant with greatest value per dollar. This evaluation is particularly significant since most professional accountants prepare a combination of individual, business, and estate returns and, therefore, are normally in the market for a complete tax software package. Further analysis of the efficient complete packages indicates that oftentimes their individual modules are deficient relative to other modules. For example, Lacerte's complete package achieved a 1.00 efficiency score while its 1040 and business returns modules are deficient. Another example is ProSeries which has perfect efficiency score for the complete package and for its 1040 module but does not attain efficiency for its business module. The implication for professional accountants is that the synergistic effects of the complete packages overcome the deficiencies in the individual modules.

Furthermore, the DEA results for the complete packages are generally in agreement with a survey of tax preparers attending recent American Institute of Certified Public Accountants tax conferences. The survey asked the participants to rate their overall satisfaction with their tax software packages. The findings indicate agreement that ProSystem, Tax Relief, Tax Machine, Lacerte, ProSeries, TaxWorks, and Professional Tax System are the leading packages currently on the market (Steed, 1995).

In addition to an overall efficiency score for each module of a package, DEA generates slacks, S+ and S- in the dual problem (2). The slack for an operating capability is the difference between the value measured and the value which is considered efficient. Slacks identify input wastes and output shortfalls and provide estimated quantities of improvement in individual operating capabilities for a package necessary to attain an efficient performance. For instance, as indicated in Table 2 all evaluated Tax Relief modules attain efficient performance except for the 1040 module which has an efficiency score of 0.9284. Detailed results for Tax Relief's 1040 module are presented in Table 3A. These results indicate that Tax Relief's 1040 module is deficient in the areas of general features, electronic filing, processing, states available, individual forms, and installation & setup.

94

SPRING 1996

Another example is TaxWorks with an efficiency score of 1 for all its evaluated modules except the 1040 module. The TaxWorks 1040 module has an efficiency score of 0.7002. The slacks for the TaxWorks 1040 module are presented in Table 3B. This module is deficient relative to other 1040 modules in terms of its general features, analysis, printing, individual forms, maximum forms and installation and setup. However, Tax/Pack Professional achieves efficiency scores of 1 for all of it modules on an individual basis and on a combined basis. As a consequence, it has no slack for any operating capabilities. This information is valuable to the professional accountants in choosing a product, while the software company can use the same information to improve its product.

CONCLUSION

Tax preparation software has become a critical factor as professional accountants take advantage of rapidly changing information technology. A professional accountant's tax preparation software strategy is a significant decision about information technology, productivity, and client satisfaction. The selection is a multiple criteria decision-making process that matches the client's requirements with the appropriate software package.

The DEA model has been used to connect price and setup cost operating capabilities of tax preparation software packages for the purpose of evaluating the performance of individual packages relative to their peer group. DEA does not require a set of preassigned weights for inputs and outputs, thereby overcoming the deficiency introduced by using arbitrary or subjective assessments. By using nonsubjective assessments of tax software packages, DEA provides tax preparers with unbiased evaluations for selection. Even though the DEA evaluations identify several packages that are efficient, the professional accountant's judgment is still critical in the final selection. With numerous packages available, the findings of this study should help professional accountants reduce time and cost and improve their selection process.

95

Table 3A

Slacks for Tax Relief, 1040 Module

Input or Output Value Measured Value if Efficient SlackGeneral Features 60.00 62.00 2.00Electronic Filing 30.00 38.00 8.00Analysis 90.00 90.00 0.00Processing 30.00 36.00 6.00Printing 110.00 110.00 0.00States Available 38.00 53.20 15.20Individual Forms 61.00 83.00 22.00Max Forms 40.00 72.00 32.00Installation & Setup 20.00 36.00 16.00Price 775.00 775.00 0.00

Table 3B

Slacks for Taxworks , 1040 Module

Input or Output Value Measured Value if Efficient SlackGeneral Features 60.00 76.90 16.90Electronic Filing 50.00 50.00 0.00Analysis 70.00 94.20 24.20Processing 30.00 30.00 0.00Printing 90.00 130.50 40.80States Available 40.00 40.00 0.00Individual Forms 59.00 87.50 28.50Max Forms 60.00 80.80 20.80Installation & Setup

20.00 40.40 20.40

Price 1097.00 1097.00 0.00

SPRING 1996

REFERENCES

Banker, R.D., Charnes A., & Cooper, W. W. (1984). Models for Estimating Technical and Scale Efficiencies in Data Envelopment Analysis. Management Science, 30(9), 35-44.

Caves, D.W., Christensen, L. R., & Diewert, L. E. (1982). The Economic Theory of Index Numbers and the Measurement of Input, Output, and Productivity. Econometrica, 50(6), 1393-1414.

Charnes, A., Cooper, W. W., Huang, Z., & Sun, D. B. (1990). Polyhedral Cone-Ratio DEA Models with Illustrative Application to Large Commercial Banks. Journal of Econometrics, 46, 73-91.

Charnes, A., Cooper, W. W., Wei, Q. L., & Huang, Z. M. (1991). Cone Ratio Data Envelopment Analysis and Multi-Objective Programming. International

Journal of System Sciences, 20, 1099-1118.Cohn, M. (1995). 1040 Tax Prep Update. Accounting Technology, 11(2), 18-23.Courtney, H. M. and Flippen, C. L.(1995). A Shopper's Guide to Accounting

Software. Journal of Accountancy, 179(8), 37-59. Fisher, D. M, and Sun, D. B. (1996). Lan-Based E-mail: Software Evaluation.

Journal of Computer Information Systems, 34(2), 21-25.Langer, J. (1995). Practitioners Ask for (and Get) More Tax Applications. Taxation

for Accountants, 55(4), 212-232.Mahmood, M. A., Pettingell, K. J., Shaskevich, A. I. (1996). Measuring Productivity of

Software Projects: A Data Envelopment Analysis Approach. Decision Sciences, 27(7), 57-80.

Schefczyk, M. (1993). Operational Performance of Airlines: An Extension of Traditional Measurement Paradigms. Strategic Management Journal, 14, 301-317.

Sherman, H. D., Ladino, G. (1995). Managing Bank Productivity Using Data Envelopment Analysis (DEA). Interfaces, 25(2), 60-73.

Stearman, S. W., A Review of the Leading Tax Programs. Journal of Accountancy, 176(4), 1992, 57-81.

------------, What's New in Tax Software. Journal of Accountancy, 177(4), 50-66.------------, (1994). Spotlight on Tax Software. Journal of Accountancy, 178(4), 49-74.------------, (1995). Tax Software Buyers' Guide. Journal of Accountancy, 180(4), 51-81.Sun, D. B., and Gong, L. G. (1993). Performance Evaluation of New Production

Operations. Applications of Management Science, 7, 99-113.

97

A

DIMENSIONALITY, RELIABILITY AND VALIDITY OF

SERVQUAL

Ugur Yavas *

Zeynep Bilgin **

This article examines selected psychometric properties of SERVQUAL in an international setting. An empirical study conducted among Turkish students reveals that SERVQUAL demonstrates acceptable reliability and validity but needs internal structure refinement, since different dimensions load on the same factors.

deliberate attempt to study services marketing dates back to the mid-1960s (Rathmell, 1966). However, the interest on the topic has gained considerable momentum within the past decade, undoubtedly owing to Parasuraman, Zeithaml and Berry's (1985, 1988) seminal work on service quality. In today's competitive markets, businesses seek profitable ways to differentiate themselves and to gain a competitive edge over their rivals. Delivery of high service quality to customers offers firms an opportunity to distinguish themselves in crowded markets. Unlike goods quality, which can be measured objectively, service quality is abstract and elusive. In the absence of objective measures, firms must rely on consumers' perceptions of service quality to determine their own relative strengths and weaknesses, and to set up priorities. Hence, development of valid instruments to measure service quality is imperative.

This research was supported by a Non-Instructional Assignment and a Research Development Committee grant awarded to Dr. Yavas by East Tennessee State University.

* Ugur Yavas is a Professor of Marketing at East Tennessee State University, Johnson City, Tennessee.

** Zepnep Bilgin is an Associate professor of Marketing at Marmara University, Istanbul, Turkey.

Manuscript received, April, 1996, revised, July, 1996.

98

Parasuraman, Zeithaml and Berry's (1985) initial work toward development of a service quality measure was based on in-depth interviews with executives and twelve focus groups with consumers in four service industries (credit card, banking, brokerage, and repair services). This pioneering effort identified ten determinants of service quality including access, communication, competence, courtesy, credibility, reliability, responsiveness, security, tangibles and understanding/knowing customers. In subsequent work, by using an iterative procedure via factor analysis of a ninety-seven-item questionnaire, Parasuraman, Zeithaml and Berry (1988) refined the ten determinants into SERVQUAL, an instrument specifically designed to measure service quality.

In SERVQUAL, the initial ten determinants were consolidated into five: tangibles (TANG)--physical evidence of the service; Reliability (RELI)--consistency of performance and dependability; Responsiveness (RESP)--willingness and readiness of employees to provide service; Assurance (ASSU)--confidence communicated by the service provider, and Empathy (EMPA)--service provider's efforts to understand the customer's needs and then to individualize the service delivery. Assurance encompassed the prior five determinants of communication, credibility, security, competence and credibility, and EMPA captured the former access and knowing/understanding the consumer dimensions. Parasuraman, Zeithaml and Berry proposed that SERVQUAL was an adaptable instrument which could fit any organization's needs in measuring service quality. The considerable enthusiasm spawned by SERVQUAL is evidenced by numerous works which examined its psychometric properties and/or used it in applied settings (see, for example: Cronin and Taylor, 1992, 1994; Carman, 1990; Gagliano and Hathcote, 1994; Brown and Swartz, 1989).

The purpose of this article is to examine dimensionality, reliability and validity of the wide-ranging SERVQUAL instrument in the Turkish setting. Specifically, by using a sample of Turkish college students as its database, this study expands upon the research by Yavas and Arsan (1995) and Akan (1995) who examined psychometric properties of SERVQUAL among bank employees and consumers in Turkey.

METHOD

Sample

Data for the study were collected from undergraduate students attending Bogazici and Marmara universities located in Istanbul. Students in both schools completed the questionnaire in a self-administered manner during regular class periods. The medium of instruction at Bogazici University is English. Marmara University, in its separate divisions, provides instruction in Turkish and English. To eliminate the cross-linguistic

99

equivalence problem associated with questionnaire translation (Aulakh and Kotabe, 1993), the surveys in the latter institution were administered only to the students attending the English division. Usable responses were obtained from 292 students. The sample was almost evenly divided in gender--female (51.7 percent) and male (48.3 percent). Students were primarily business/economics majors (61.3 percent).

Use of students as surrogates of other populations and generalizability of results obtained from student samples are often questioned (Burnett and Dunne, 1986). However, evidence shows that both in domestic and international studies, use of student subjects is appropriate when the objective is to assess psychometric properties of instruments (Dunne, Lund and Luchsinger, 1980; Yavas, 1994).

Measurement

The final form of SERVQUAL is comprised of 22 items (Parasuraman, Zeithaml and Berry, 1988). The breakdown of the items by dimension is as follows:

TANG (4)

RELI (5)

RESP (4)

ASSU (4)

EMPA (5)

In this study, service quality was operationalized by using these 22 items. Specifically, for each item, the subjects were asked to evaluate the performances of their bank, doctor, hair stylist (barber) and post office on seven-point scales ranging from "Much worse than I expected" to "Much better than I expected." This specific scale was borrowed from Brown, Churchill and Peter (1993) whose research indicated that it is more efficient than the disconfirmation procedure used in the original SERVQUAL measure.

RESULTS

Dimensionality

To investigate the purported dimensionality of the SERVQUAL instrument (i.e., TANG, RELI, RESP, ASSU and EMPA), two sets of factor analyses were run. First, for each service setting, 22 SERVQUAL items were subjected to exploratory factor analysis. None of these analyses yielded the five-structure solution posited by

100

Parasuraman, Zeithaml and Berry (1988). Instead, three factors with eigenvalues greater than one emerged in the cases of bank, barber and post office settings and a two-factor solution was obtained in the doctor setting. In the next step, factor analyses were run for each setting by restricting the number of factors to five. In other words, the factor program was forced to generate five-factor solutions. Tables 1, 2, 3 and 4 present these results for the bank, doctor, barber and post office settings, respectively.

As can be seen from Table 1, nine items had their highest loadings on Factor 1 which accounted for 45.8 percent of the variance in the data. While four of the five EMPA items loaded on this factor, the factor also had items from the other domains. Four RELI items along with two items from the RESP domain were at the root of Factor 2. Two of the four ASSU items loaded on Factor 3 and three of the four TANG items loaded on Factor 4 (see Table 1). Of all the domains, TANG appears to be the only one which is distinct from the others. As shown in Tables 2, 3 and 4, four items comprising this domain load on distinct factors. Otherwise, there are no clear-cut patterns in the compositions of other factors. For instance, Factor 1 in Table 2 includes one RELI, three RESP, four ASSU and four EMPA items. Six items with highest loadings on Factor 2 (Table 3) belong to three different domains, RELI, RESP and ASSU. The same is true for the second factor of the post office setting where seven items comprising the factor come from these three domains (see Table 4).

The preceding discussion suggests that the dimensionality of SERVQUAL instrument is suspect. The instrument does not decompose into the purported five-structure solution. Even when the analysis is forced to conform to a five-factor structure, composition of the factors are different than those intended in the original formulation.

Table 1

Factor Analysis: Banka,b

Domain FACTOR 1 FACTOR 2 FACTOR 3 FACTOR 4 FACTOR 5

TANG 1 - - 3 -

RELI 1 4 - - -

RESP 2 2 - - -

ASSU 1 - 2 - 1

EMPA 4 - - - 1

Percent of Variance

45.8 7.7 6.5 4.4 3.9

101

Domain FACTOR 1 FACTOR 2 FACTOR 3 FACTOR 4 FACTOR 5

TANG 1 - - 3 -

RELI 1 4 - - -

RESP 2 2 - - -

ASSU 1 - 2 - 1

EMPA 4 - - - 1

Percent of Variance

45.8 7.7 6.5 4.4 3.9

a Results are based on varimax rotated matrixb Numbers across each domain indicate the number of items from that particular domain

which have their highest loadings on that factor

Table 2

Factor Analysis: Doctora,b

Domain FACTOR 1 FACTOR 2 FACTOR 3 FACTOR 4 FACTOR 5

TANG - 4 - - -

RELI 1 - 4 - -

RESP 3 - 1 - -

ASSU 4 - - - -

EMPA 4 - - 1 -

Percent of Variance

67.8 5.8 4.2 2.5 2.2

a Results are based on varimax rotated matrixb Numbers across each domain indicate the number of items from that particular domain

which have their highest loadings on that factor

Table 3

Factor Analysis: Barbera,b

102

Domain FACTOR 1 FACTOR 2 FACTOR 3 FACTOR 4 FACTOR 5

TANG - - - 4 -

RELI - 2 3 - -

RESP - 2 2 - -

ASSU 2 2 - - -

EMPA 5 - - - -

Percent of Variance

56.4 9.3 6.6 3.1 2.8

a Results are based on varimax rotated matrixb Numbers across each domain indicate the number of items from that particular domain

which have their highest loadings on that factor

103

JOURNAL OF BUSINESS AND MANAGEMENT

Table 4

Factor Analysis: Post Officea,b

Domain FACTOR 1 FACTOR 2 FACTOR 3 FACTOR 4 FACTOR 5

TANG - - 4 - -

RELI 1 4 - - -

RESP 2 2 - - -

ASSU 2 1 - 1 -

EMPA 4 - - - 1

Percent of Variance

51.8 7.9 6.6 4.3 3.4

a Results are based on varimax rotated matrixb Numbers across each domain indicate the number of items from that particular domain

which have their highest loadings on that factor

The adverse dimensionality finding, however, is not surprising in factor analysis of complex structures. Furthermore, this study is not unique in its failure to generate the neat five-factor structure posited by Parasuraman, Zeithaml and Berry (1988). Cronin and Taylor's (1992) confirmatory factor analysis of the SERVQUAL items resulted in a unidimensional solution. The five-factor structure they obtained from oblique factor rotation had a poor fit. Finn and Lamb (1991) and Spreng and Singh (1993) used confirmatory factor analysis and their five-factor solutions had poor fits. Brensinger and Lambert (1990) generated a five-factor solution but only four factors had eigenvalues exceeding one. Babakus and Boller's (1992) analysis of 22 SERVQUAL items produced a two-factor structure. In one of the earlier studies conducted in Turkey, Yavas and Arsan (1995) obtained a five-factor solution, yet the decomposition of the items into the five dimensions was different than the one purported in SERVQUAL. The number of factors extracted by Akan (1995) in her study of Turkish consumers ranged from seven to twelve.

Reliability

To assess the reliability of the five dimensions of SERVQUAL as well as the overall instrument itself across settings, coefficient alphas were computed. Coefficient alpha indicates reliability in terms of the internal consistency of items relating to a multi-item

104

SPRING 1996

scale (Peterson, 1994). Reliability coefficients of .70 or higher are deemed acceptably high (Nunnally, 1978). The internal consistency reliability coefficients at the dimension level ranged between .75 and .91 (TANG), .86 and .92 (RELI), .83 and .93 (RESP), .80 and .93 (ASSU), and .86 and .94 (EMPA). Furthermore, the coefficient alphas for the overall instrument across all four settings were well above Nunnally's (1978) guidelines; they ranged between .94 (bank) and .97 (doctor).

To further examine the internal consistency reliability of the measures, inter-domain and domain-total correlations were compared to the Guilford and Fruchter (1973) guidelines. These guidelines suggest that inter-domain correlations should fall within .10 - .60 bounds and domain-total correlations should range between .30 and .80. An examination of the results revealed that of the fifteen pairwise correlations computed for each service setting, doctor had the most and bank the least number of correlation coefficients falling outside of the suggested range. Yet even in the case of bank, two-thirds of the coefficients were out-of-bounds. In this as well as the other three settings, coefficients falling outside of the guidelines exceeded the upper bound.

An inspection of the results at the domain level showed that of all five domains, tangibles most conformed to the Guilford-Fruchter (1973) guidelines. Of the twenty inter-domain and domain-total correlations computed for tangibles across all four settings, fourteen were within the guidelines. With all correlations falling within bounds, TANG's best performances were in the cases of bank and barber settings.

Validity

Convergent Validity. Confirmation of the existence of a construct indicated by correlations of independent measures of the construct provides evidence for convergent validity (Jaccard, Brinberg and Ackerman, 1986). To test for convergent validity, for each setting, a single-item direct service quality measure was included in the survey. To obtain this measure, respondents were asked to rate the overall quality of services they received from their banks, doctors, barbers/hair stylists and post-offices on a six point scale ranging from excellent to terrible. These single-items measures were then correlated with their multiple-item counterparts (i.e., the overall service quality scores obtained for each respondent by summing their answers across 22 SERVQUAL items). All the correlations were significant at .0001 level of significance and they ranged from a high of .87 (doctor) to a low of .64 (bank). These results attest to the convergent validity of the instrument.

Nomological Validity. When a construct of interest is related to other constructs assessing a different but conceptually related construct by an established body of theory or according to a priori expectations (Peter, 1981), confirmation of the relationship

105

JOURNAL OF BUSINESS AND MANAGEMENT

predicted by the theory/expectations is evidence of nomological validity. Past writings (O'Connor, Shewchuk and Bowers, 1991; Boulding et al., 1993; Bitner, 1990; Brown and Swartz, 1989) show that service quality is significantly related to consumer satisfaction, recommending, continued patronage/repeat purchase, complaining and intentions to switch.

In this study, nomological validity of SERVQUAL was assessed by relating it to four behavioral measures. These were consumer satisfaction, intention to continue to do business/patronize the service provider, recommending it to friends and complaint behavior. Satisfaction was operationalized by a five-item measure where each item was measured on a 6-point scale ranging from "extremely satisfied" to "extremely dissatisfied." Intention to patronize was measured on a 5-point "very likely" to "very unlikely" scale. Again a 5-point scale ranging from "definitely would recommend" to "definitely would not recommend" was employed in measuring the recommendation behavior. Complaint behavior was measured in terms of frequency of complaints via a 5-point scale. The anchor points of this scale were "always" and "never."

It was hypothesized that higher levels of service quality sentiments would result in higher levels of consumer satisfaction, higher likelihoods of continuing to patronize the particular service provider, higher probabilities of recommending that provider to friends and lower frequencies of complaint. Because of the scoring system used, the signs of the coefficients were expected to be positive with respect to the first three hypotheses and negative for the fourth one.

The results reported in Table 5 lend support to these expectations. All the correlation coefficients were significant at .05 or better level of significance. Furthermore the coefficients had the expected signs. These results suggest that SERVQUAL demonstrates reasonable nomological validity.

CONCLUSIONS

The past decade or so was characterized by rapid internationalization of business. As a result, markets in many industries are becoming increasingly integrated worldwide. Such developments stimulate interest in international research in general, and methodological issues surrounding cross-national research in particular. Measurement equivalence is one of the methodological issues which is gaining increased attention. A common a priori assumption in international marketing research has been that measures developed in one culture will be universally applicable to other cultures. Typically, emic measures developed in the United States are transferred to other cultures without any modification. It appears that by virtue of precedence, cross-

106

SPRING 1996

validity of measures has been endorsed. However, such assumptions and endorsements are now being challenged.

This study examined dimensionality, reliability and validity of a wide-ranging measure developed in the United States (SERVQUAL), in the Turkish setting. The results of the study show that SERVQUAL instrument demonstrates acceptable reliability and validity when tested in a different culture. While this is encouraging, results also suggest that caution should be exercised when using the instrument. It appears that the internal structure of the scale is not crystallized. Items purporting to measure different dimensions tend to load on the same factor, and many items contribute little to the assessment of different dimensions. Hence, further refinement of the scale is in order.

Table 5

Nomological Validity of SERVQUAL Across Settingsa

Measure Bank Doctor Barber Post Office

Continue to Patronize .43 .78 .68 .39

Satisfaction .77 .91 .87 .81

Recommend to Friends .65 .82 .69 .57

Complain -.45 -.62 -.42 -.44

a All the correlation coefficients reported in the table are significant at .05 or better level of significance

To achieve this, Churchill's (1979) paradigm for developing better measures of marketing constructs, recently illustrated by Webster (1993) in a service context, can be followed. Briefly, the procedure includes: a) specifying the domain of the construct, b) generating sample items that may tap the construct, c) collecting data on the measures, d) purifying the measure via coefficient alpha and factor analysis, e) collecting additional data to further assess reliability/validity and f) developing norms. These steps are essential to refine the SERVQUAL measure in the U.S. setting where the instrument suffers from dimensionality problem. However, the systematic procedure is imperative for the Turkish setting not only to address the dimensionality issue but also to determine the scalar and conceptual equivalence of the SERVQUAL measure.

107

REFERENCES

Akan, P. (1995). Dimensions of Service Quality: Expectations of Turkish Consumers from Services. Istanbul: Bogazici University.

Aulakh, P.S. & Kotabe, M. (1993). "An Assessment of Theoretical and Methodological Developments in International Marketing." Journal of International

Marketing, 1(2), 5-28.Babakus, E. & Boller, G.W. (1992). "An Empirical Assessment of the SERVQUAL

Scale." Journal of Business Research, 24(3), 253-268.Bitner, M.J. (1990). "Evaluating Service Encounters: The Effects of Physical

Surroundings and Employee Responses." Journal of Marketing, 54(2), 69-82.Boulding, W., Kalra, A., Staelin, R. & Zeithaml, V.A. (1993). "A Dynamic Process

Model of Service Quality: From Expectations to Behavioral Intentions." Journal of Marketing Research, 30(1), 7-27.

Brensinger, R. & Lambert, D.M. (1990). "Can the SERVQUAL Scale be Generalized to Business-to-Business Services?" In Bearden, W. et al. (Eds.), Enhancing

Knowledge Development in Marketing (p. 289). Chicago: American Marketing Association.

Brown, S.W. & Swartz, A. (1989). "A Gap Analysis of Professional Service Quality." Journal of Marketing, 53(2), 92-98.

Brown, T.J., Churchill, G.A. & Peter, J.P. (1993). "Improving the Measurement of Service Quality." Journal of Retailing, 69(1), 127-139.

Burnett, J.J. & Dunne, P.M. (1986). "An Appraisal of the Use of Student Subjects in Marketing Research." Journal of Business Research, 14(4), 329-343.

Carman, J.M. (1990). "Consumer Perceptions of Service Quality: An Assessment of the SERVQUAL Dimensions." Journal of Retailing, 66(1), 33-55.

Churchill, G.A. (1979). "A Paradigm for Developing Better Measures of Marketing Constructs." Journal of Marketing Research, 16(1), 64-73.

Cronin, J.J. & Taylor, S.A. (1992). "Measuring Service Quality: A Reexamination and Extension." Journal of Marketing, 56(3), 55-68.

Cronin, J.J. & Taylor, S.A. (1994). "SERVPERF versus SERVQUAL: Reconciling Performance-Based and Perceptions-Minus-Expectations Measurement of

Service Quality." Journal of Marketing, 58(1), 125-131.Dunne, P., Lund, D. & Luchsinger, V. (1980). "The Use of Students as Surrogates: A

Jeopardy Game?" In Bagozzi, R.P. et al. (Eds.). Marketing in the 80's: Changes and Challenges (Pp. 312-315). Chicago: American Marketing Association.

Finn, D.W. & Lamb, C.W. (1991). "An Evaluation of the SERVQUAL Scale in a Retailing Setting." In Holman, R. and Solomon, M.R. (Eds.). Advances in Consumer Research (Pp. 483-490). Provo: Association for Consumer Research.

Gagliano, K.B. & Hathcote, J. (1994). "Customer Expectations and Perceptions of Service Quality in Retail Apparel Specialty Stores." Journal of Services Marketing, 8(1), 60-69.

Guilford, J.P. & Fruchter, B. (1973). Fundamental Statistics in Psychology and Education. New York: McGraw-Hill Book Company.

Jaccard, J., Brinberg, D. & Ackerman, L. (1986). "Assessing Attribute Importance: A Comparison of Six Methods." Journal of Consumer Research, 12(4), 463-468.

Nunnally, J.C. (1978). Psychometric Theory. New York: McGraw-Hill.O'Connor, S.J., Shewchuk, R.M. & Bowers, M.R. (1991). "A Model of Service Quality

Perceptions and Health Care Consumer Behavior." Journal of Hospital Marketing, 6(1), 69-92.

Parasuraman, A., Zeithaml, V.A. & Berry, L.L. (1985). "A Conceptual Model of Service Quality and Its Implications for Future Research." Journal of Marketing, 49(4), 41-50.

Parasuraman, A., Zeithaml, V.A. & Berry, L.L. (1988). "SERVQUAL: A Multiple-Item Scale for Measuring Consumer Perceptions of Service Quality." Journal of Retailing, 64(1), 12-37.

Peter, J.P. (1981). "Construct Validity: A Review of Basic Issues and Marketing Practices." Journal of Marketing Research, 18(2), 133-145.

Peterson, R.A. (1994). "A Meta-analysis of Cronbach's Coefficient Alpha." Journal of Consumer Research, 21(2), 381-391.

Rathmell, J.M (1966). "What is Meant by Services?" Journal of Marketing, 30(4), 32-36.Spreng, R.A. & Singh, A.K. (1993). "An Empirical Assessment of the SERVQUAL

Scale and the Relationship between Service Quality and Satisfaction." In Cravens, D.W. and Dickson, P. (Eds.). Enhancing Knowledge Development in Marketing (Pp. 1-6). Chicago: American Marketing Association.

Webster, C. (1993). "Refinement of the Marketing Culture Scale and the Relationship between Marketing Culture and Profitability of a Service Firm." Journal of Business Research, 26(2), 111-131.

Yavas, U. & Arsan, N. (1995). "An Assessment of Selected Psychometric Properties of a Service Quality Scale: An International Study (Bank Employees)." Journal of Customer Service in Marketing and Management, 1(4), 105-116.

Yavas, U. (1994). "Research Note: Students as Subjects in Advertising and Marketing Research." International Marketing Review, 11(4), 35-43.

T

EXPERIENTIAL LEARNING IN THE CAPSTONE STRATEGIC

MANAGEMENT COURSE: COLLABORATIVE PROBLEM

SOLVING, THE STUDENT LIVE CASE, AND MODELING

Karen L. Fowler *

Donna M. Scott **

This paper presents the application of experiential learning theory to the capstone strategic management course. The history of the capstone course is reviewed and the primary objectives for requiring this course discussed. Traditional methods of instruction for teaching the capstone course are presented, along with a discussion of the limitations of these approaches. The development of the Student Live Case is presented and advantages of this teaching approach discussed. The use of modeling to facilitate the experiential approach is described and student perceptions on the use of this technique are presented. Future research designed to compare the use of modeling with the traditional feedback approach is suggested.

oday's business organizations expect college graduates to possess competencies that go beyond mere content knowledge. Because most businesses now find themselves competing in a rapidly changing global marketplace, they need graduates who not only know their content areas well, but also are skilled in innovative problem solving, research capabilities, and self-empowered learning.

The emerging needs of today's organizations place new demands on institutions of higher education and are forcing educators to modify both what is taught and how

* Karen L. Fowler is a Professor of Management in the College of Business Administration at the University of Northern Colorado, Greeley, CO.

** Donna M. Scott is a Professor of Professional Studies at Mesa State College, Grand Junction, CO.

Manuscript received April, 1996, revised, July, 1996.

students are taught. The traditional educational practice of having faculty impart information to students is being replaced by educational methods that emphasize the learning process and promote learning partnerships between instructor and learner (Bradford, 1993; Weil & McGill, 1989). This article describes an instructional approach that combines experiential learning with modeling, an approach that was used by the authors in teaching the undergraduate capstone strategic management (policy) course. The history and evolution of the strategic management course is briefly reviewed, along with the traditional methods of instructing the course. The authors' experiential learning approach is presented by describing how the components of collaborative problem solving, the student live case, and modeling were integrated into the course.

HISTORY AND EVOLUTION OF THE CAPSTONE COURSE

According to Christensen, Andrews, and Bower (1978), the business policy course originated at Harvard Business School in 1911 when a small group of faculty first developed course materials designed for managers who were studying business. The role of the course has been to culminate and integrate the various functional and tool areas of instruction. Traditionally, students have been given business case studies that require application of many areas of content knowledge to successfully perform the assigned analysis. The course has long been a required course for colleges of business accredited by the American Assembly of Collegiate Schools of Business (Schendel & Hofer, 1979).

Prior to World War II, the primary focus of the course was the integration of the various functional areas (accounting, finance, marketing, and management) as they pertained to daily operating decisions. At most colleges and universities, the name of the course was Business Policy. Following World War II, however, the emphasis of the course began to move toward concern with the organization's external environment. Massive social, political, legal, technological, and economic changes confronted American companies. The course was restructured to include an analysis of the organization's external environment as well as an analysis of the organization's internal environment based on an integrated view of the functional areas. Many texts in the 1970s were titled Business Policy and Strategic Management. Strategic management can be defined as the process of matching or fitting an organization with its external environment in the most advantageous way possible (Digman, 1995). By the early 1990s, many textbook authors had dropped the use of Business Policy in their titles and had opted for titles such as Strategic Management. To date, the objective of the course remains the analysis of the firm's complex external environment, followed by the development of strategies that will enable the firm to compete successfully. An inherent component of this strategic process is an integrated assessment of the various functional areas within the organization.

TRADITIONAL METHODS OF INSTRUCTION

The long-standing tradition for teaching this capstone course has been through the use of business case studies. Students, either individually or in teams, have been required to analyze cases written about real or fictitious organizations. The use of the case approach remains very popular today, and most strategic management texts include a collection of cases.

In the early-to mid-1980s, a number of simulations which utilized mainframe computers began to appear. The computer simulation exercises challenged students to make a number of successive operating decisions during the semester. Typically, students would compete with one another in teams, with each team representing a separate competitor within a fictitious or simulated industry. A number of strategic management computer simulations are now available to run on personal computers.

There are, however, some problems associated with the use of both the case method and the simulation method in the capstone course. When using the case method, students are often limited to the information and data presented in the case. Second, the lead time necessary for getting a case published and into the students' hands is very long. Often the case data are several years old before students read the case. Strategic management is very future oriented, but the use of the case method relies on archival data from years past. Third, there may not be sufficient material in the case to expose students to the entire strategic management process. Fourth, instructors typically have all students read the same case information, providing no opportunity for intragroup interdependence. In reality, organizational decision makers possess some relevant information but not all; they are often dependent on others who have needed information or varying perspectives. Fifth, the instructor usually assigns the specific cases to be analyzed, leaving no opportunity for choice on the part of students, thus reinforcing the teacher-as-expert role.

With respect to computer simulation games, problems exist as well. Experience has shown that technical problems in running the program often create frustrations for both instructor and students alike. Second, many students tend to focus their efforts on trying to figure out the computer program in order to "win the game," rather than on using sound analysis and decision-making skills. Third, the simulations are nearly always fictitious situations that the students tend to relate to as a game, not as a matter deserving serious thought. Fourth, many of the simulations have been marketed as strategic simulations, but in reality the decisions required by students are operational in nature. Fifth, these simulations tend to produce very clear outcomes in response to each set of student decisions, a situation perhaps not representative of the ambiguities present in real industries and organizations.

These and other characteristics of the case and simulation methods do not represent the complexities of strategic management as it realistically exists in today's dynamic industries. The purpose of this paper is to present an alternative teaching methodology that has been used very successfully by the authors in the classrooman approach which more closely mirrors the strategic management process faced by managers. Additionally, and equally important, this experiential approach shifts the student's focus from specific outcomes to the learning process, a skill deemed critical for success in our current work environment. Finally, our approach redistributes the responsibility for learning, creating an equal and reciprocal partnership between instructor and student.

THE EXPERIENTIAL APPROACH

Experience as a Vehicle for Learning

Drawing on the works of John Dewey, Kurt Lewin, and Jean Piaget, David Kolb offers the most-often cited definition of experiential learning (Kolb, 1984). Kolb defines experiential learning as "the process whereby knowledge is created through the transformation of experience" (Kolb, 1984). Kolb points out that this definition emphasizes several critical aspects of the learning process with respect to the experiential approach. First, the process of adaptation and learning is emphasized rather than content or outcomes. Second, knowledge is a transformation process that is continuously created and recreated. Third, learning transforms experience in both its objective and subjective forms.

Learner-centered aims of experiential learning methods include an atmosphere of trust in the classroom whereby curiosity and the desire to learn are nourished, a participatory mode of decision making in all aspects of learning, experiences that help students build confidence and self-esteem, and the excitement of intellectual and emotional discoveryall of which encourage students to become life-long learners (Rogers, 1983). Experiential learning methods focus on stimulating learner motivation to acquire the skills that will allow them to become assertive, adaptable, proactive, and effective communicators; in short, competent individuals who know how to find relevant information and apply it (Henry, 1989). A distinction can and should be made between learning from experience and experiential learning. Brah and Hoy (1989) suggest that the two are not synonymous. Learning from experience relies on starting from personal experience. Experiential learning uses experience as a vehicle for learning.

The role of the facilitator in experiential learning is to provide a helpful structure which enables learners to clarify their expectations, develop plans to meet their goals, and draw on resources which are available. Most forms of experiential learning are based on some notion of freedom and autonomy for the learner (Boud, 1989).

Experiential Learning in the Capstone Course

The idea for creating an experiential approach in the capstone strategic management course originated as a result of our experiences with the previously discussed limitations of the case and simulation methods, from our recognition of the need to focus attention on the process of learning, and from our desire to shift the emphasis from specific problems to more complex issues. As Hutton (1989) states, a problem is associated with a specific answer or solution, whereas an issue, or situation, or focus of concern more readily accommodates ongoing change over time. Hutton believes the experiential approach helps students learn how to make sense out of complexity and act constructively with whatever information is available to them. Hutton's approach for experiential learning relies on the process of making judgements as opposed to the outcome expected.

Lau and Shani's Behavior in Organizations: An Experiential Approach (1992, 1995) also served as a motivation. It made sense to teach business students about organizational behavior by involving them in behavioral situations which exemplified the theories and concepts being taught. Students have repeatedly indicated their strong preference for the experiential organizational behavior approach over traditional lecture and case classes.

The capstone course presented a dilemma in terms of providing an experiential approach, especially for undergraduate business students. Most undergraduate students do not have first-hand experience with strategic management. Environmental analysis, industry structure, mission statements, corporate diversification, turnaround strategies, networks, strategic alliances, and functional area integration are content topics learned only from textbooks, with little opportunity for actual experience by students. The challenge was to find a way to involve the students in their strategic management learning experiences in a way that would emphasize the learning process.

Group and Interpersonal Competencies

Given the emphasis in today's organizations on working in teams, Bradford (1993) contends that group-centered teaching is especially relevant in management courses. Students should increase their specific subject content knowledge as well as their group and interpersonal competencies. Bradford believes, however, that the traditional form of instructor-to-student lecture still prevails in most business school classrooms.

An extensive amount of literature has been generated with respect to groups, cooperative learning, and collaborative problem solving. Our interest in these areas stems from the primary integration objective of the capstone strategic management course, our desire to improve students' teamwork skills, and our desire to emphasize the process of learning.

According to Johnson and Johnson (1975), learning environments can be structured competitively, individualistically, or cooperatively. In competitive learning situations,

students' goal achievements are negatively correlated. Students seek to improve their individual performance at the expense of their classmates. In an individualistic learning structure, students seek to improve their individual performance without regard to the performance of others. Self-paced learning and programmed instruction are examples. In a cooperative learning situation, students seek to improve their individual performance as well as the performance of those to whom they are cooperatively linked. The cooperative learning situation represents the type of work environment college graduates will encounter when they begin their careers.

Numerous empirical studies have shown a wide range of positive outcomes from cooperative learning and collaborative problem solving (Johnson & Johnson, 1975, 1978, 1983, 1993; Johnson, Johnson, & Maruyama, 1983; Johnson, Maruyama, Johnson, Nelson & Skon, 1981; Sharan, 1980; Slavin, 1977). Johnson and Johnson (1993) contend there are at least seven positive instructional outcomes from cooperative learning: higher achievement, greater motivation, better attitudes toward subject area and professor, more positive relationships with other students, higher levels of self-esteem and psychological health, greater cognitive and emotional perspective taking, and better interpersonal and small-group skills.

Boud (1989) also discusses the appropriateness of the group-centered approach in experiential learning, pointing out that much learning occurs from the interactions among group members. Individuals pursue their own learning needs within the context of the group, relying on others for feedback and support. Emphasis on democratic decision making and the consideration of different points of view are characteristic of the group approach.

Collaborative Problem Solving in the Capstone Course

Collaborative problem solving is especially appropriate in the capstone strategic management course. The course objective is to enhance the student's understanding of the organization's complex relationship with its external environment and the interrelatedness of the various functional areas within the organization. Students are taught that strategic management is a continuing process by which attempts are made to match or fit an organization with its external environment in a way that will enable the organization to survive and prosper (Digman, 1990, 1995). However, organizations do not make strategic decisionspeople in organizations make decisions. It is imperative that students learn how to work together to effectively accomplish this complex process. Organizational members with diverse backgrounds and various agendas must work together to analyze future trends, assess perceived organizational strengths and weaknesses, and develop what they feel will be the best game plan for the future.

Development of the Live Case

The label "Live Case" or "Living Case" was suggested to the first author by Allen Wedell, Professor of Marketing, Colorado State University, during discussions of various teaching methodologies. According to Learned (1991), in living cases, students are teamed with local businesses to develop strategic plans for the businesses. This concept is quite similar to the Small Business Institute course found in many business schools. In the SBI course, students are teamed with local businesses to develop strategic plans for the businesses or to solve the business client's problems. Wynd and Wedell (1992) define a live case somewhat differently. Clients bring their problems into the classroom in person and interact with students to define the problems, determine alternatives, and agree on recommendations.

The capstone strategic management course is a separate course from the Small Business Institute course at many business schools. The strategic management course tends to focus more on industry structures and their associated characteristics, global competitiveness, and strategies for those competitors within given industries. It would be impossible for most undergraduate or graduate business students to maintain ongoing contact for an entire semester with the CEOs and other line managers of the major corporations they are studying. Thus a modified form of the live case has been created. This adaptation of the live case will be referred to as the Student Live Case (SLC). The SLC concept described below has been extended from earlier ideas developed by Robert E. Jones, Professor of Management, University of Wyoming.

As mentioned previously, the capstone strategic management course is required for all business majors in schools accredited by the American Assembly of Collegiate Schools of Business. Normally, there is a fair representation of the various business majors in each class: accounting, computer information systems, finance, general business, management, and marketing. In order to create the Student Live Case approach, the instructor assigns students to teams based primarily on the students' majors. An attempt is made to provide representation from each functional major. In the real world, strategic decisions are affected by every functional area within the organization. This form of team structure forces the finance major to collaborate with the marketing major, the accounting major with the management major, and so on. Team size is usually four to six students, depending on total class size.

The team works together for an entire semester and must complete a major strategic analysis term project. At the beginning of the semester, the team is allowed to choose any industry and company within that industry for extensive analysis during the semester. Instructor guidance is recommended to help students avoid selecting highly fragmented industries or extensively diversified companies that would be difficult to analyze in the course of one semester. The team is then responsible for researching and collecting all data relevant to their SLC project, applying all concepts learned during the semester, and presenting a complete written executive summary of their strategic analysis to the professor near the end of the semester. A strategic management text, such as Digman (1995), Hunger and Wheelen (1996), Pearce and Robinson (1994), or

Thompson and Strickland (1996), covering the basic strategic management process is used to supply the conceptual foundation for the course. Examples illustrating the application of each text concept are presented as the semester progresses to encourage transfer to the students' analyses of their Live Cases. Class time is provided for team meetings, thus enabling the instructor to facilitate application of the concepts learned. Providing time in class for team meetings also helps circumvent the common student complaint that the diverse schedules of their team members make it almost impossible for them to meet outside of class.

Advantages of the Student Live Case Approach

The advantages of utilizing the SLC are numerous. First, the students feel actively involved in a number of ways. The team has the freedom to choose which industry and company it wants to analyze. Often this decision takes a week or two as students struggle to find an industry and company that interests every team member. Once consensus is reached, however, all members tend to move forward with a commitment that is shared. Students cannot blame the instructor for the choice of text cases, they cannot complain about how the text case was written or how old the case data are, and they cannot claim that there was insufficient information on which to base recommendations. Text concepts can be covered while teams move toward consensus on choosing their companies.

Second, the students improve their research abilities. Students must discover ways to find the information they need. Students with weak library research skills can be led to the primary information sources through handouts listing the call letters of various business and economics holdings in the library, as well as CD-ROM and online index services that are available to students. Many colleges of business or their main libraries also subscribe to computer-readable sources of information that may be useful. Standard and Poor's PC Plus and Bridge are popular sources for corporate data. Online access to a wide variety of electronic databases can be gained through CompuServe, Prodigy, Lexis/Nexis, America Online, Delphi, DIALOG, and many other sources on the Internet. Additionally, most students write or telephone the companies and request information.

Third, the students become more skilled at sorting through information and determining what is applicable. In both the case and simulation methods, students often complain that there is too little information provided for sufficient analysis. In the Student Live Case, students usually gather too much information and are then faced with the dilemma of what is useful. This information overload is a much better representation of real-world decision environments. Experience has shown that it is desirable to set a page limit for the executive summaries. Thirty pages, double spaced is a good target. A "no page limit" policy has resulted in 100-page summaries replete with student rambling. No outline is provided for the executive summary, forcing students to decide what information is critical to their analysis. Experience has shown that students

prefer to be told exactly what the professor wants, but this is not representative of the job environment they will face upon graduation. Students are encouraged to explore a number of complex issues and present a range of alternatives rather than a single recommendation. Further comment on this desire for structure is made later in this paper.

Fourth, the use of the SLC requires the students to actually work through the strategic management process rather than just read about it. This allows students to experience all of the challenges associated with collaborative team efforts. Dealing with conflict resolution, personality conflicts, and individual contribution differences are powerful learning experiences for students. Jalajas and Sutton (1993) offer an entertaining and useful article for instructors who plan to use student groups. The title of their article is "Feuds in Student Groups: Coping with Whiners, Martyrs, Saboteurs, Bullies, and Deadbeats."

Fifth, from the instructor's perspective, the Student Live Case term projects are much more interesting to read and grade than traditional cases or simulation reports. Most professors are pressed for time to keep up with what is happening in the global marketplace, and we would all love to have more time to read Business Week and The Wall Street Journal. The term projects result in interesting student accounts of the current issues facing a variety of real industries and companies. Students feel more responsible for their share of the instructor-student partnership, as they will actually be providing information and analyses of benefit to the instructor.

Finally, students have a finished product they can show to prospective employers when they are interviewing. Students typically show a great deal of pride in their finished projects and are pleased when they have an opportunity to tell employers about their research and analyses capabilities.

FACILITATING THE EXPERIENTIAL APPROACH THROUGH MODELING

Two observations led the authors to believe that there was a need for incorporating model reports into the experiential learning process described above. First, anyone who has ever taught management is familiar with the frequent student comment, "How do we know if we're doing it right?" as if management were a perfect science analogous to mathematics. Students continually ask for an outline of what the professor wants and how it should be presented. Second, in past semesters, copies of previous strategic management student projects were made available for student guidance, but the students had to check them out of the reserve section of the library. Typically, the better students were the only students to take advantage of this learning opportunity; the average and poorer students would not make the effort to check out and read the prior term projects. These observations led the authors to explore further the use of modeling to facilitate the experiential approach.

Modeling as a Vehicle for Learning

Bandura's Social Learning Theory is a behavioral theory that takes into account intrapersonal and environmental determinants. Social learning refers to the idea that individuals acquire much of their behavior by observing and imitating others in a social context. People's behavior and the environment influence each other (Steers & Porter, 1991). Cameron and Whetten (1993) have suggested that Social Learning Theory is an excellent approach for teaching management skills to business students. These well-known researchers argue convincingly that students must be taught to develop critical management skills in the classroom. Learning about management is not sufficient, and it is too late for the students when they have to learn to manage on the job. "Social learning theory focuses on changing behavior through the modeling process" (Cameron & Whetten, 1993).

Modeling is a component of the social learning theory framework (Steers & Porter, 1991; Wood & Bandura, 1989). Individuals can expand their knowledge and skills by observing the behavior of others and the consequences of that behavior (Bandura, 1986; Rosenthal & Zimmerman, 1978). An overview of ways that modeling can be used to develop intellectual, social, and behavioral competencies can be found in Steers and Porter (1991).

Jernstedt (1986) extends the application of modeling to experiential learning. Vicarious learning techniques such as handouts, recalled personal experiences, instructor modeling, and demonstrations serve as examples of others' behavior. Jernstedt has studied the impact of these vicarious learning methods and reports an increase in student achievement and motivation.

Modeling in the Experiential Approach

Rather than providing previous student projects for library reserve checkout, the authors prepared model strategic analysis reports on four different corporations for use in team assignments. The reports were prepared specifically to model the type of research and level of reporting expected of the students in their Student Live Case term projects. Each report contained relevant information about the company's external environment, the industry, the company's history and mission, product mix, competitive position, and strategies. The reports were typed using appropriate writing style, headings, paragraphs, referencing, etc. The students were told that the model reports had been prepared for their reference in completing their SLC projects. Each model report was sufficiently different in approach and content to show the students that strategic management can take various formats for reporting issues and analyses. The idea was to demonstrate for students that there is not just one right way to present their findings and analysis.

Team members were given a copy of the first model report and were then required to meet in class one week later to discuss it. The students were told that they were to be prepared to identify and analyze the company's strengths, weaknesses, opportunities, threats, strategic issues, and strategic recommendations, but that they were not to write out any responses in advance.

Each team met in a separate room, without the instructor, during their collaborative problem-solving sessions. The professor was available nearby, however, to answer any questions as to what was expected of the group. The instructor was often asked specific content questions regarding the model report, but in responding always made an effort to facilitate group learning rather than offering a definitive answer. The instructor must be prepared to say, "What do you and your other team members think?" At the end of the class period, each team handed in a handwritten summary of their analysis. The second model report was then handed out and the process repeated until teams had read and analyzed four model reports during the middle portion of the semester.

It is important to note that in this modeling approach, the turned-in collaborative responses were not graded immediately and returned to the students. The responses were held by the instructor until the end of the semester. Students were told that the responses would be graded when all four had been turned in, and, in fact, the responses were not graded until after the students had turned in their term SLC projects. The intent was to have students learn how to prepare their SLC projects from the modeled reports and their successive collaborative experiences, rather than by any form of feedback from the instructor.

As expected, many students asked repeatedly for immediate feedback, offering the previously cited comment, "How do we know if we are doing it right?" It was explained to the students that strategic management is not a perfect science and that the instructor was more concerned that they learn and understand the strategic management process from their collaborative efforts than specific content. In the real world, the CEO does not "grade" the strategy team's report and say, "Your team got an F. Do it over." Making complex decisions involves judgement, and learning the process of making judgements is a key task for professional education (Hutton, 1989). However, since the problem-solving assignments represented a moderate portion of the students' final grades (an incentive deemed necessary for committed student participation), students still begged for feedback, albeit unsuccessfully. This apparent need for feedback is often a topic of discussion among faculty members and is discussed again later in this paper.

Our approach to modeling incorporates several elements suggested by Wood and Bandura (1989). These researchers contend that modeling has been used very successfully to develop intellectual, social, and behavioral competencies, provided three components are present (Bandura, 1986, 1988; Wood & Bandura, 1989). First, the appropriate skills are modeled to convey the desired competencies. The students in our

research were provided with modeled strategic analysis reports. The second component involves guided skills mastery. Our students were provided structured opportunities to practice their new skills throughout the semester. The third component requires a transfer program whereby the newly learned skills are actually tried out in a job situation. In our research, the students had to apply their new skills in order to complete their Student Live Case term projects.

Student Response to Modeling

By the end of the fourth model report team discussion, there appeared to be a significant level of improvement both in terms of collaborative processes and content analysis. Comments such as "at first we were clueless" and "we could have done better" seem to indicate some improvement over the four meetingsat least from the students' perspectives.

In order to assess student perceptions regarding the use of the modeling technique, a questionnaire was administered to the students. The results indicated that the students felt very strongly about the usefulness of the modeling approach (see attached exhibit). The modeled reports seemed to provide an example of what students needed to do regarding their SLC term project and showed them that there is not just one right way to present their findings and analyses. The individual student comments suggest that the modeled reports helped students in other ways as well. Several comments indicated that the modeled reports served as a confidence builder and motivator.

DIRECTIONS FOR FUTURE RESEARCH

As mentioned previously, students continually asked for immediate feedback on their collaborative problem-solving responses. Many college courses are taught with the traditional feedback technique, thus conditioning students to expect feedback on all their class work. Experience has shown that students exhibit a low tolerance for ambiguity and lack of clear structure from the professor. Many teaching evaluation forms include an item on whether or not assignments were graded and returned promptly to students. This student expectation of clearly structured tasks and immediate feedback is not representative of today's work environment. We believe Hutton's (1989) suggestion of "habitual expectation" probably applies here. As Bradford (1993) points out, however, the traditional methods of instruction may no longer be appropriate. We may actually be doing a disservice to our students if we lead them to believe they will find a clear sense of direction with immediate feedback once they enter the work place. This observation is particularly important for undergraduate students. Graduate students with work experience may more readily discern the differences between the classroom and the work environment.

Of further interest are questions concerning the use of the feedback technique as compared with the modeling approach. We are currently extending this research project by analyzing the application of the traditional feedback approach in the capstone course. In the future, we hope to make direct comparisons between the use of the modeling and feedback methods within the experiential approach. The impact of the findings may affect both students and instructors alike in terms of motivation, achievement, satisfaction, student-professor relationships, and instructor assessment.

Fowlers figure 1 of 3

Fowlers figure 2 of 3

Fowlers figure 3 of 3

JOURNAL OF BUSINESS AND MANAGEMENT

REFERENCES

Bandura, A. (1986). Social Foundations of Thought and Action: A Social Cognitive Theory.

Englewood Cliffs: Prentice-Hall.Bandura, A. (1988). Organizational Applications of Social Cognitive Theory. Australian

Journal of Management, 13, 137-164.Boud, D. (1989). Some Competing Traditions in Experiential Learning. In S. W. Weil & I.

McGill (Eds.), Making Sense of Experiential Learning (Pp. 38-49). Philadelphia: Society for Research into Higher Education and Open University Press.

Bradford, D. L. (1993). Learning in Groups. In C. Vance (Ed.), Mastering Management Education (Pp. 173-178). Newbury Park: Sage.Brah, A. & Hoy, J. (1989). Experiential Learning: A New Orthodoxy? In S. W. Weil & I.

McGill (Eds.), Making Sense of Experiential Learning (Pp. 70-77). Philadelphia: Society for Research into Higher Education and Open University Press.

Cameron, K. S. & Whetten, D. A. (1993). A Model for Teaching Management Skills. In C. Vance (Ed.), Mastering Management Education (Pp. 160-172). Newbury Park:

Sage.Christensen, C. R., Andrews, K. R., & Bower, J. L. (1978). Business Policy. Homewood: Irwin. Digman, L. A. (1990). Strategic Management. Homewood: BPI/Irwin.Digman, L. A. (1995). Strategic Management. DAME Publications.Henry, J. (1989). Meaning and Practice in Experiential Learning. In S. W. Weil & I.

McGill (Eds.), Making Sense of Experiential Learning (Pp. 25-37). Philadelphia: Society for Research into Higher Education and Open University Press.

Hunger, J. D., & Wheelen, T. L. (1996). Strategic Management. Reading, Massachusetts: Addison Wesley.Hutton, M. (1989). Learning From Action: A Conceptual Framework. In S. W. Weil & I.

McGill (Eds.), Making Sense of Experiential Learning (Pp. 50-59). Philadelphia: Society for Research into Higher Education and Open University Press.

Jalajas, D. S. & Sutton, R. I. (1993). Feuds in Student Groups: Coping With Whiners, Martyrs, Saboteurs, Bullies, and Deadbeats. In C. Vance (Ed.), Mastering

Management Education (Pp. 217-227). Newbury Park: Sage.

126

SPRING 1996

Jernstedt, G. C. (1986). Experiential Components in Academic Courses. In R. Kraft & J. Kielsmeier (Eds.), Experiential Education and the Schools (Pp. 109-117).

Boulder: Association for Experiential Education.Johnson, D. W. & Johnson, R. (1975). Learning Together and Alone: Cooperation,

Competition, and Individualization. Englewood Cliffs: Prentice-Hall.Johnson, D. W. & Johnson, R. (1978). Cooperative, Competitive and Individualistic

Learning. Journal of Research and Development in Education, 13, 3-15.Johnson, D. W. & Johnson, R. (1983). The Socialization and Achievement Crises: Are

Cooperative Learning Experiences the Solution? Applied Social Psychology Annual, 4, 119-164.

Johnson, D. W. & Johnson, R. T. (1993). Structuring Groups for Cooperative Learning. In C. Vance (Ed.), Mastering Management Education (Pp. 187-198). Newbury Park:

Sage.Johnson, D. W., Johnson, R., & Maruyama, G. (1983). Interdependence and Interpersonal

Attraction Among Heterogeneous and Homogeneous Individuals: A TheoreticalFormulation and Meta-Analysis of the Research. Review of Educational Research, 53, 5-54.

Johnson, D. W., Maruyama, G., Johnson, R., Nelson, D., & Skon, L. (1981). The Effects of Cooperative, Competitive, and Individualistic Goal Structures on Achievement: A

Meta-Analysis. Psychological Bulletin, 89, 47-62.Kolb, D. A. (1984). Experiential Learning. Englewood Cliffs: Prentice-Hall.Lau, J. B. & Shani, A. B. (1992). Behavior in Organizations: An Experiential Approach. Homewood: Irwin.Learned, K. E. (1991). The Use of Living Cases in Teaching Business Policy.Journal of Management Education, 15, 113-120.Pearce, J. A., & Robinson, R. B. (1994). Competitive Strategy. Burr Ridge: Irwin.Rogers, C. (1983). Freedom to Learn for the 80's. Columbus: Merrill.Rosenthal, T. L. & Zimmerman, B. J. (1978). Social Learning and Cognition. New York: Academic Press. Schendel, D. E. & Hofer, C. W. (1979). Strategic Management. Boston: Little, Brown and Co.Sharan, S. (1980). Cooperative Learning in Small Groups: Recent Methods and Effects on

127

JOURNAL OF BUSINESS AND MANAGEMENT

Achievement, Attitudes and Ethnic Relations. Review of Educational Research, 50, 241-171.

Shani, A. B. & Lau, J. B. (1995). Behavior in Organizations: An Experiential Approach. Homewood: Irwin.Slavin, R. E. (1977). A Student Team Approach to Teaching Adolescents With Special Emotional and Behavioral Needs. Psychology in the Schools, 14, 77-84.Steers, R. M. & Porter, L. W. (1991). Motivation and Work Behavior. New York:

McGraw-Hill. Thompson, A. A., & Strickland, A. J. (1996). Strategy Formulation and

Implementation. Homewood: Irwin.Weil, S. W., & McGill, I. (1989). Making Sense of Experiential Learning.

Philadelphia: Society for Research into Higher Education and Open University Press.

Wood, R. & Bandura, A. (1989). Social Cognitive Theory of Organizational Management. Academy of Management Review, 14, 361-384.Wynd, W. R. & Wedell, A. J. (1992). A Live Case in the Classroom. Proceedings of the Annual Conference of the Midwest Society for Case Research, 53-57.

CALL FOR PAPERS

JOURNAL OF BUSINESS AND MANAGEMENTThe JOURNAL OF BUSINESS AND MANAGEMENT is soliciting papers for its upcoming issue. Its purpose is to provide a forum for the dissemination of fresh ideas and research in all areas of business, management and public policy which would be of interest to business persons, public officials and academics. The JOURNAL invites both empirical and theoretical articles. Reports on research and opinion pieces as well as original manuscripts are welcome. Book reviewers are requested to submit articles on recent publications in related topics. Only articles not previously published or currently under review elsewhere can be considered. All submissions are blind refereed. It should not be assumed that readers are completely familiar with the concepts and terminology of the specific subject under study. Directness and clarity of presentation are desired.

Manuscripts must be typed and double-spaced. Three copies should be submitted and become the property of the JOURNAL. The length of the paper should not exceed 20 pages, excluding notes, references, tables, figures and appendices. Notes and references

128

SPRING 1996

should appear at the end of the article. Each chart, figure or table should be camera ready, that is, an original on a separate page with instructions indicating the placement in the article. A separate cover sheet indicating the title of the manuscript, as well as the name, affiliation, position and address of the author should accompany each submission. To facilitate blind review, the body of the text should include the title, but not the author's name. Include also an abstract of not more than 100 words summarizing the article.

The views expressed in published articles are those of the authors and not necessarily those of the Editors or the Editorial Review Board. Responsibility for the correctness of quotations and citations rests with the author.

The JOURNAL OF BUSINESS AND MANAGEMENT is a bi-annual publication of the Western Decision Sciences Institute and the School of Management, California State University, Dominguez Hills.

Submit Papers to:

Burhan F. Yavas, EditorJOURNAL OF BUSINESS AND MANAGEMENTSchool of ManagementCalifornia State University, Dominguez HillsCarson, CA 90747

129

Subscription to the

JOURNAL OF BUSINESS AND MANAGEMENT

The JOURNAL OF BUSINESS AND MANAGEMENT is published bi-annually by the Western Decision Sciences Institute and the School of Management, California State University Dominguez Hills.

The annual subscription fee is $16.00. To receive the

JOURNAL OF BUSINESS AND MANAGEMENT, please complete the form below, and make your check or money order payable to the CSUDH Foundation.

Dean's Office

School of Management

California State University Dominguez Hills

Carson, CA 90747

Name

Organization

Address

Please make your check or money order payable in the amount of $16.00 payable to the CSUDH Foundation.