Corporate Reputation and Investment Performance: The US and UK Experience

Corporate Reputation and Investment Performance: The US and UK Experience Sam Y. Chung* Kristina Eneroth** Thomas Schneeweis*** * Long Island University, New York ** Lund University, Sweden *** University of Massachusetts-Amherst, MA Abstract Corporate reputation is often regarded by academics and practitioners as indicative of a firm s current and future financial performance. In this study, the relationship between a firm s equity performance and reputation ratings published in the Economist (U.K.) and Fortune (U.S.) magazine is investigated for the period of 1990-1999. On a total return basis, monthly equityreturns of high-reputation firms are shown to outperform those of low-reputation firms both in the year prior and following the 'reputation' reporting month. As for other studies, the results indicate that the size of a firm s market capitalization positively affect the firm s reputation. This study shows that, unlike previous studies, it is primarily a firm s equity performance in the pre-survey and survey period that affects the published ranking of the firm s reputation and the published ranking has little impact on the firm s future risk-adjusted equity return. JEL Classification: G11, Portfolio Choice G14, Information & Market Efficiency; Event Study G15, International Financial Market M14, Corporate Culture; Social Responsibility Key Words: Social Responsibility; Corporate Reputation; Equity performance, Investment Please send all correspondence to Sam Y. Chung at 1University Plaza, Brooklyn, NY 11201 Email: sam.chung@liu.edu Tel: (718)488-1149, Fax: (718)488-1125

Corporate Reputation and Investment Performance: The US and UK Experience Abstract Corporate reputation is often regarded by academics and practitioners as indicative of a firm s current and future financial performance. In this study, the relationship between a firm s equity performance and reputation ratings published in The Economist (U.K.) and Fortune (U.S.) magazine is investigated for the period of 1990-1999. On a total return basis, monthly equityreturns of high-reputation firms are shown to outperform those of low-reputation firms both in the year prior and following the 'reputation' reporting month. As for other studies, the results indicate that the size of a firm s market capitalization positively affect the firm s reputation. This study shows that, unlike previous studies, it is primarily a firm s equity performance in the pre-survey and survey period that affects the published ranking of the firm s reputation and the published ranking has little impact on the firm s future risk-adjusted equity return. JEL Classification: G11, Portfolio Choice G14, Information & Market Efficiency; Event Study G15, International Financial Market M14, Corporate Culture; Social Responsibility Key Words: Social Responsibility; Reputation; Firm performance, Investment 2

Corporate Reputation and Investment Performance: The US and UK Experience I. Introduction Corporate reputation is often regarded by academics and practitioners as indicative of a firm s current and future financial performance. For instance, Shefrin and Statman [1994, 1995] have presented a behavioral capital asset pricing model in which investors perception of a firm s quality may impact the firm s risk and return. Their theoretical model suggests large (small) firms may be perceived as good (bad) firms and it is consistent with corporate reputation survey evidence that large firms have superior corporate reputation in both U.K. and U.S. 1 For the United States and Britain, empirical results also exist on the correspondence between a firm s equity performance and external evaluators perceptions of the firm s qualitative attributes (e.g., quality of management, capacity to innovate). For instance, Antunovich and Laster [1998] have argued that Fortune reputation ratings are directly related to a firm's future equity performance in U.S. They report the most admired firms in the U.S. achieve high equity return performance after corporate reputation publication while the less admired firms generally underperform. 2 For other studies, however, little correlation has been found between perceived management quality and future risk-adjusted equity returns for both U.S. [Shefrin and Statman, 1994,1995; McGuire et al., 1988, 1990] and British firms, [Nanda et al., 1996]. In this paper, the relationship between a firm s published reputation rankings and its equity performance is investigated in 1) pre-survey period, post-survey period and during the survey period and 2) pre and post publication month. As in other studies, results show that: (1) 3

the 'high-rated' firms on their reputation outperform, on a total equity return basis, the 'low-rated' firms and (2) larger firms generally have a higher reputation scores than smaller firms. This study also shows that the impact of firm size is due primarily to the manner of corporate reputation survey collection and the approach taken to risk-return analysis. The results show that: (1) little relationship exists between corporate reputation ratings and a firm s future riskadjusted equity performance and (2) changes in corporate reputation ratings are related to changes in a firm s equity return performance in the pre-survey and during the survey period and, thus are not solely due to firm size. Therefore, unlike previous studies, the result in this study shows that a firm's equity return performance in the pre-survey and survey period affects published rankings of a firm's reputation qualities and that the publishing of these rankings has little impact on a firm's future risk-adjusted returns. In the next section, previous literatures about the relationship between a firm s financial performance and published corporate reputation are reviewed. The data and methodology are presented in Section III. Results are presented and discussed in Section IV. We conclude in section V. II. Literature Review Financial theory has supported the use of corporate reputation in assessing a firm s future financial performance. Firms perceived as excellent along an array of dimensions have an easier access to financial capital or have a lower cost of businesses. Managerial capacity has been cited to be of prime interest on evaluating security selection [Bodie, Z. et al., 1997]. A firm s corporate reputation quality may affect its ability to deal in cheaper implicit contracts (e.g., non-union employees) in contrast to more costly explicit (e.g., union employees) contracts [Cornell and Shapiro, 1987]. 4

As important, previous empirical research has directly addressed the correspondence between a firms financial performance and external evaluators perceptions of the qualitative attributes [see Exhibit 1]. **********Insert Exhibit 1 about here********* Some researchers have provided evidence that investing in high-reputation firms can be a profitable strategy. For instance, Clayman [1987] finds that the returns of investing in an equally weighted portfolio of 29 firms featured in the book In Search of Excellence: Lessons from America s Best-Run Corporations outperform the S&P 500 by 1.1% a year from 1981 to 1985. Antunovich and Laster [1998] have also argued that Fortune survey s reputation rankings are directly related to a firm's future equity performance. They report the most admired firms in the Fortune survey achieve high ex-post return performance (post ranking publication) while the less admired firms generally underperform. Specifically, their empirical results suggest that a portfolio of the most admired firms earns an abnormal return of 3.2 % in the year after the survey is published and 8.3 % over the three years. In contrast, the least admired firms earn a negative abnormal return of 8.6% in the nine months through the end of the year. For other studies [Shefrin and Statman, 1995, 1997; McGuire et al., 1990; Nanda et al., 1996], little correlation, however, is found between perceived management quality and future risk-adjusted security returns for either U.S. or U.K. firm. Shefrin and Statman [1995, 1997] show that the Fortune survey gives high ratings to firms with large market capitalizations and high market-to book ratios, hypothesizing that firms rated highly in the Fortune survey underperform the market. To test this hypothesis they regress one-year equity returns on the Fortune rating on value as a long-term investment, with mixed results. The reputation measure 5

has a positive coefficient in 8 years and a negative coefficient in five years; the pooled regression has a coefficient that is negative and marginally significant. Thus theoretical and empirical evidence exists on the external perception of a firm's performance across a wide variety of qualitative attributes as indicative of the ability of firms to lower costs of capital, to lower various contracting costs, to increase investor interest, and to achieve superior future financial performance. However, current reputation ratings may not necessarily be related to future equity performance. Finance theory generally accepts the position that corporate stock prices incorporate all past and current information such that unless published reputation ratings contain new information that affect firms expected risk and return, published reputation rankings should not affect future risk-adjusted equity returns. III. Data and Methodology Rankings of public perceptions of firms qualitative attributes are obtained from surveys commissioned by The Economist (U.K. firms) and Fortune magazine (U.S. firms). For U.S. and U.K. firms, these surveys have been published yearly over the time period of analysis 1990-1999. For U.K., survey participants are asked to rate companies on eight attributes; quality of management, financial soundness, quality of products and services, ability to attract, develop and retain top talent, value as a long-term investment, capacity to innovate, quality of marketing, and corporate social responsibility. For U.S., survey participants are asked to rank companies on similar attributes. 3 For both countries, a total firm quality score is computed by averaging firm ratings over the eight dimensions. Firms are ranked on a scale of 0 (poor) to 10 (excellent). 4 Monthly return data are determined for the sample firms, the S&P 500, and FTSE all share index. In addition, monthly return data for size-based market indices; Frank Russell 100, 250 and 2000 and the FTSE 100, 250 and small cap indices, are determined. For U.K., 6

approximately 200 firms are rated each year, while for the U.S., approximately 350 firms are rated. For the U.K., sixty-seven firms monthly return data are available from Datastream and those are ranked in all across the sample-period (1990-1999). No consistent industry bias is indicated from the sample firms. For the U.S., given the large number of firms, tests are conducted on a data set insured between group variance (high vs. low rankings). In each year the top ten and the bottom ten firms in total ratings are obtained for the high and low group, respectively. While the top ten firms remain relatively stable in their group, the bottom ten firms are changed such that over time a total of approximately 100 firms are listed in those groups. The portfolio size measured by the market value of equity is also determined for the highand low-rated portfolio. For instance, in the U.K. sample, the average market value of the firms equity in high-rated portfolio in the year 1990 is 5,4341 million while the low-rank portfolio is 1,486 million. The relative firm-size remains similar in future years. The firm-size difference between high- and low- portfolio is also evident for the U.S. sample. The average market value of the firms equity in the high-rated portfolio of the year 1990 (US$ 20,865 million) is approximately five times larger than that of a low-rated portfolio (US$ 5,111 million). Portfolio abnormal returns (AR) are determined by both the market-adjusted return and the risk-adjusted return basis: AR = R R where, R = return on a portfolio i in period t. R = return on a market index m in period t. SAR it it mt it mt it i CSAR = SAR, where k = 12, 11, 10... 0,..., 9, 10, 11. t we also define ARit =, σ notice that σ k= t k= 12 ik i is calculated from the prior three years data to the year before the publication. 7

In brief, ARs are determined after subtracting the market return on the appropriate market index from the returns of high-, mid-, and low-rated portfolio in each year. Standard errors from the prior three years data to the year before publication are used to determine the Standardized Abnormal Returns (SARs) for the year before and after publication. Cumulative Standardized Abnormal Returns (CSARs) are reported as means to track the relative performance of the high-, mid-, and low-rated portfolios. ARs are presented for both the pre-publication versus postpublication period and pre-survey versus post-survey period (see Exhibit 2). ********* Insert Exhibit 2 about here ******** Abnormal returns over the window of twenty-four months, one-year prior to and after the publication, measure the impact of the published reputation ranking on the equity performance of the firm. If the relative market performance affects respondents' ratings, then high (low) rated firms should find their CSARs to be rising (falling) in the survey period of the previous year. If the market incorporates the information of the reputation rankings before publication, the pattern of CSARs for the high and low rated portfolio should be similar after the publication month. In addition, if the respondents ratings affect the firm s equity performance in the months following the survey period, then high (low) rated firms should find their CSARs rising (falling) in the months following the survey period As noted earlier, a firm s reputation rating by itself may not indicate equity performance impacts, especially, if the a firm is rated in the same category (high, mid, or low) across publication years. However, for firms with an increase or decrease in rankings, the change in rankings may be due to recent market performance. For firms experiencing a dramatic shift in reputation (e.g., high to low or low to high), one may expect to see rising or falling CSARs 8

during the pre-survey or survey period relative to firms staying in the same portfolio across the sample years. IV. Results and Discussion Corporate Reputation and Market Performance Previous research [McGuire et al, 1988, 1990; Nanda et al. 1996; Shefran and Stateman, 1994,1995] indicate that a firm-size (market capitalization) may be a primary determinant of corporate reputation ratings. Similarly, Antunovich and Laster [1998] suggest that firm size may be a primary determinant in firm quality assessments and subsequent return performance based on this quality assessment. In Exhibit 3A (relative performance in the year before the publication) and 3B (relative performance the year after the publication), the return and correlation results show the close correspondence between market indices based on firm size and portfolios based on corporate reputation. For the U.K., the average correlation between firm-size and the corporate reputation score for the sample period is 0.47. 5 For the high-, mid-, and low-rated portfolios, the average corporate reputation rate and firm-size over the sample period are as follows; high rank (7.45; 5,341 million), mid rank (6.29; 2,302 million), and low rank (4.99; 1,486 million). Given the relationship between corporate reputation rate and firm-size, it is expected that the return of the high-ranked portfolio has its highest correlation with the return of the FTSE 100 index. And the return of the mid-ranked portfolio has a higher correlation with the return of the FTSE 250 (mid-cap) index than that of the FTSE Small-Cap index, while the low-ranked portfolio has a higher correlation with the FTSE Small-Cap index than with any other FTSE indices (see the correlation tables). As important, the performance of the different FTSE indices reflects the performance of the U.K. high-, mid-, and low-ranked portfolios. For instance, over 9

the period 1990-1999, as shown in Exhibit 3B, the average monthly returns for the largecap/small-cap FTSE index (1.0% / 0.7%) and high/low reputation portfolio (1.1% / 0.9%) reflect the relatively higher performance of both the large cap and the associated large firms (highranked firms in reputation) in the months following the published rankings. Similarly, for the period 1990-1999, the differential performance between the FTSE small-cap /large-cap index (1.8% / 0.2%) as well as the low-rank/high-rank reputation portfolio (1.4% / 1%) reflects the traditional small cap effect (higher performance of small firms for the first quarter of the year). When the relative risk (standard deviation) is adjusted, however, the performance of the two portfolios is not significantly different. Of importance is that in contrast to the first quarter of the year, during the survey months (especially in the pre-publication year) as shown in Exhibit 3A, the large-cap index and highranked firms (0.9% / 1.2%) outperform the small-cap index and low-ranked firms (0.1% / - 1.0%). Thus during the survey period, large firms generally outperform small firms. The results are consistent with respondents view that the better return performance of the larger firms is indicative of higher quality. Thus survey ratings are reflected in the higher correlation between large-cap/small-cap market indices and high-rank/low-rank portfolios in the following year. ************Insert Exhibit 3A and 3B about here********* In Exhibit 4A and 4B, the similar return and correlation result also shows the close correspondence between corporate reputation and firm size for the U.S. firms. For the U.S., the overall correlation between firm-size and the reputation score is 0.52. For the high-, mid-, and low-ranked portfolios, the representative corporate reputation score and firm size are as follows; high rank (8.25/ US$20,865 million), mid rank (7.39/ US$16,740 million) and low rank (5.62/ 10

US$5,111 million). Given the relationship between corporate reputation score and firm-size, it is also expected that the returns of the high-ranked portfolio have the highest correlation (0.88 / 0.89) with the Russell 200 in the pre and post publication years and the returns of the midranked portfolio have the highest correlation (0.92 / 0.91) with the mid-cap Russell, while the low-ranked portfolio has a higher correlation with Russell small-cap index (0.74 / 0.82). As important, the performance of the different Russell indices reflects the performance of the U.S. high-, mid-, and low-ranked portfolios. For the period between January, 1990 and March, 1997, as shown in Exhibit 4B, the average monthly returns for the large/small cap Russell index (1.3% / 1.1%) and high/low ranked portfolio (1.54% / 0.4%) reflect the relatively higher return of both the large-cap and the associated high-ranked portfolio in the months following the publication. In contrast, the differential performance between the Russell small/large cap index (2.3% / 1.5%) and the low/high ranked portfolio (2.9% / 1.3%) in the first quarter of the year also reflects the traditional 'small cap effect' (higher performance of small firms for the first quarter of the year). ********Insert Exhibit 4A and 4B about here********** Again, contrast to the first quarter of the year, during the survey months of the prepublication year, as shown in Exhibit 4A, the large-cap index and high-ranked firms (1.1%/1.8%) outperform the small-cap index and low-ranked portfolio (0.5% /-0.01%). Thus during the survey period, larger firms outperform smaller firms, which is the consistent result with the U.K. analysis. Similarly, the reported superior performance of the high-rank portfolio relative to the low-rank portfolio after the publication month may be due solely to the seasonal pattern of the firm-size return differential. For the six month period (April-September) following the 11

reputation ranking publication, as shown Exhibit 3B and 4B, the large-cap FTSE/Frank Russell indices (0.8%/0.9%) outperform their low-cap FTSE/Frank Russell indices (0.2%/0.3%). Similarly for the U.K. and U.S. the high-rank portfolio (1.0%/1.1%) outperforms their corresponding low-rank portfolio (0.1%/-0.4%) during the six months period. Thus, to the degree that the reputation ratings reflect firm-size, portfolio performance after the reputation ranking publication may be less a result of reputation effects than firm size effects. More appropriately, after adjusting relative-risk (standard deviation), the actual performance of the two portfolios, large-cap/small-cap or high-rank/low-rank, is not significantly different in either country. In brief, the relationship of the portfolio returns for the high, mid, and low rated firms in the U.K. and the U.S. is consistent with previous results; that is a firm-size alone may be seen as a basis for the relative firm ratings and their equity performance in the post reporting period. Moreover, there is evidence that for both the U.S. and the U.K., the high-rank (large-cap) portfolio outperforms the low-rank (small-cap) portfolio. However, the variability in return and the lack of consistent yearly pattern make any long-term investment policy questionable 6. Moreover, the seasonal pattern of the return performance of small/large firms may provide further insight as to the impact of a firm s equity performance on the firm's corporate reputation. Since large firms consistently outperform medium and small firms during the second and third quarter or the survey period (July September), to the degree that a firm's equity performance affects respondents' perception of the firm's quality, larger firms may dominate smaller firms in their reputation ratings of the following year s publication. Corporate Reputation and Risk-Adjusted Equity Performance 12

In the previous section, results indicate the consistency between the corporate reputation and the equity performance especially in regard to a firm's size (market capitalization). In Exhibit 5 and 6, we describe the risk-adjusted returns (Cumulative Standardized Abnormal Returns) of high, medium and low ranked firms across the U.K. and U.S. for the entire sample period (pooled sample results). 7 *******Insert Exhibit 5A&B and 6A&B about here******** Results are presented relative to both the publication month and the survey period. In Exhibit 5A, results show that for the U.K. high-rank portfolio the CSARs increases (3.13) while the low-rank portfolio decreases (-5.94) in the year before the publication month. In the year after publication, the CSAR of the high rank portfolio increases on average (0.30) while that of the low rank portfolio decreases (-0.57). Neither the increase in CSARs of the high rank portfolio or the decrease in CSARs of the low ranked portfolio is significant in the year after publication. The difference between the two CSARs is statistically larger in the pre-publication period than in the year following publication. This is an indicative of the inability of reported scores to obtain abnormal returns on risk-adjusted basis for U.S. firms. Similarly, for U.S. firms in Exhibit 6A, results show that for the high rank portfolio the CSARs increases (3.21) in the year before the publication date while the low rank portfolio decreases (-4.59) in the year before publication. In the year after publication the CSAR of the high rank portfolio increases on average (1.62) while that of the low rank portfolio decreases (- 3.45). Again, in the year after publication, neither the increase in CSARs of the high rank portfolio or the decrease in CSARs of the low ranked portfolio is significant. The difference 13

between the two CSARs is also statistically greater in the pre-publication period than in the postpublication period. This is again indicative of the inability of reported scores to obtain excess risk-adjusted returns and the fact that the equity return performance during the pre-publication period may be a primary driver for the respondents perceptions of the firms reputation. Changes in Firms Rankings and Securities Performance In the previous section, the risk-adjusted performance for the each portfolio is compared. The results could be impacted by firms having no change in reputation ratings from one year (year t-1) to the next year (year t). If survey respondents are less willing to change the ratings of a previously high or low ranked firm, the reputation classification may simply be affected by the current reputation ranking of the firm. This section reports the CSARs for the four U.K. and U.S. portfolios based on changes in reputation rankings from year t-1 to year t; 1) High-rank to High-rank portfolio (HH), 2) Low rank to High-rank portfolio (LH), and 3) High-rank to Low-rank portfolio (HL) and Low rank to Low rank portfolio (LL). If a current market performance has more impact to the respondents perception for a firm s reputation than the general market characteristics does, the CSARs of the increasing firms in their rankings (LH portfolio) should outperform those of decreasing firms in their rankings (HL portfolio) both before and during survey period. Similarly, relative to LL (HH) portfolios, the HL (LH) portfolios should have a greater decrease (increase) in CSARs. In other words, if a firm s pre- and current market performance affects the respondents perception of a firm s quality, then large shift in their reputation rankings should have unusual market performance preand during the survey period. Results for both the U.K. and U.S. are reported in Exhibit 7 and 8. 8 14

*********Insert Exhibit 7 A&B and 8A&B about here******** Results are presented relative to both the publication date and the survey period. In Exhibit 7A, results show that for the U.K. the CSAR of decreasing firms in their rankings (HL portfolio) falls (-2.88%) in the year before publication while that of increasing firms in their rankings (LH portfolio) increases (3.14%). Similar results are presented for the U.S. firms. The U.S. survey respondents are also affected by market performance over the survey period. In Exhibit 8A, results show that for the U.S. the CSAR of the HL portfolio decreases (-0.79%) in the year before publication in comparison to the CSAR of the HH portfolio (3.30%). V. Conclusion In this paper, the impact of corporate reputation is shown to be due primarily to the manner of reputation survey collection and the approach taken to risk-return performance analysis. Therefore, unlike previous studies, the result in this study shows that a firm's equity market performance in the pre-survey and survey period primarily affects published ratings of a firm's reputation and the publishing of these ratings has little impact on a firm's future riskadjusted returns. Empirical results indicate that, for the time period of analysis (1990-1999), large firms tend to outperform small firms in the second and third quarters of the year and small firms tend to outperform large firms in the first quarter of the year. As a result, respondents generally witness higher equity returns for large firms relative to small firms during the survey period (July-September). Consequently, if a firm's equity performance affects respondents perception of the firm's quality, large firms may dominate small firms in reported rankings of the following 15

year. Moreover, given no new information over the following year, respondents may use the last year's ranking as a 'naive' basis for the next year's rating. The result of this study also indicates that a firm's equity performance in the survey period affects its reputation rating; that is, firms which perform poorly (well) in their equity performance in the survey period decline (rise) in their reputation rankings. This result is also indicative of non-naive respondents who use the equity market performance as a basis for reputation ratings. Thus firm size is not a sole determinant of reputation ranking. Lastly, little relationship between firms risk-adjusted equity performance and their published rankings is also indicative of the lack of a market reaction to the information of reputation rankings. It is important to realize that information about individuals view about a firm should be immediately reflected into the today's equity price. Only future unexpected changes in firms' corporate activities would affect future equity prices. Investors or corporate managers who use reputation rankings as a basis for future investment or as indicative of future risk-adjusted performance may only be capturing the expected returns underlying the fundamental risk and return patterns of the firm. 16

References Antonovich, P. and D. Laster, 1998, Do Investors Mistake a Good Company for a Good Investment?, Federal Reserve Board, New York Working Paper. Bodie, Z., A. Kane and A. Markus., 1997, Investment, 3 rd Edition. Irwin. Branch, B. and T. Schneeweis., 1980, Capital Market Efficiency in Fixed Income Securities, Review of Business and Economic Research, 34-42. Clayman, M., 1987, In search of Excellence: The Investors Viewpoint, Financial Analyst s Journal, 54-63. Cornell, B.and Shapiro, A., 1987, Corporate Stakeholders and Corporate Finance, Financial Management 16, 5-14. America's Most Admired Corporations, Fortune, 1990-1999. U.K. Most Admired Firms, The Economist, 1990-1999. McGuire, J., T. Schneeweis and B. Branch., 1990, Perceptions of Firm Quality, Journal of Management 16, 167-180. McGuire, J., J. Naroff, and Thomas Schneeweis., 1988, Effect of Cabinet Appointments on Shareholder Wealth, Academy of Management Journal, 201-212. McGuire, J., A. Sundgren, and T. Schneeweis., 1988, Corporate Social Responsibility and Firm Financial Performance, Academy of Management Journal 31, 854-872. Nanda, S., T Schneeweis, and K. Eneroth., 1996, Corporate Performance and Firm Perception: The British Experience, European Financial Management 2, 197-221. Shefrin, H. and M. Statman., 1994, Behavioral Capital Asset Pricing Theory, Journal of Financial and Quantitative Analysis 29, 323-349. Shefrin, H. and M. Statman., 1995, Making Sense of Beta, Size, and Book-to-Market, The Journal of Portfolio Management 21, 26-34. Shefrin, H. and M. Statman., 1997, Behavioral Portfolio Theory, Santa Clara University, working paper. Solt, M. and H. Shefrin., 1989, Good Companies, Bad Stocks, The Journal of Portfolio Management, 39-44. 17

1 Empirical result suggests both the Fortune and The Economist survey tend to give high ratings to firms with large market capitalizations (firm size). 2 For the U.S., our results differ from Antunovich and Laster in that the survey period is assumed to be in the third quarter of the year and the publication month in the first quarter. For the U.S. over the period of analysis the actual month of publication differs from January to March. However, in all cases, the survey was conducted in the previous year. In addition, we report on the basis of changes in ranking in contrast to ex-post portfolio performance. It should also be noted that this study concentrates on the statistical significance of the risk adjusted performance. Lastly, the issue of window dressing is not directly analyzed in this study. To the degree that window dressing exits, research has concentrated on fund managers reaction to the published information. Given that the published information is in the first quarter of the year, any window dressing impacts (buy best/sell worst) should be found in the quarter immediately following the publication of the survey. 3 For the U.S. the measured attributes are similar; that is, ability to attract, develop and keep people; innovativeness; financial soundness; community and environmental responsibility, use of corporate assets; value as a long term investment, quality of management, and quality of products. 4 To our knowledge, the Economist and Fortune magazines are the most comprehensive and widely circulated surveys of firms qualitative attributes. 5 Corrlations between firm-size and corporate reputation for the each individual year are also measured and available from authors. 6 Authors also conduct several statistical tests for the seasonal return-patterns of the large/small firms and high/low reputation firms and the results are available upon request. 7 Results for the individual years for both the U.S. and the U.K. are available from authors. 8 Results for the individual years for both the U.S. and the U.K. are also available from authors. 18

Exhibit 1 Selected Research Regarding a Relationship between a Firm's Corporate Reputation and Financial Performance Author Title Sample Major Finding Antunovich, P. and Do Investors Mistake a Good Company Fortune Survey, *A portfolio of the most admired firms earns a significant abnormal return in D. Laster, 1998 for a Good Investment? 1983-1995 the year after the survey is published (well admired firms are not overpriced). published by *The least admired firms earn a significantly negative abnormal return after AMAC DataBook the publication (the timing of returns to least admired firms provides (13 years) evidence of window dressing). Shefrin, H. and Behavioral Porfolio Theory Fortune Survey, *Fortune survey gives high rating to firms with large market capitalizations and M. Statman, 1997 1997 high market-to-book ratios. (1 year) *Firms rated highly in the survey underperform the market. Nanda, S., T. Schneeweis Corporate Performance and Firm The Economist *Differences may exist between US and UK in the use of qualitative survey data and K. Eneroth, 1996 Perception: The British Experience 1989, 1991, on a firm's strategic attributes as a forecast of a firm's future quantitative and 1992 performance measures. (3 years) *For small firms, certain qualitative factors may be of importance in forecasting accounting and security market returns. Shefrin, H. and Making Sense of Beta, Size and Fortune Survey, *Investors tend to indifferent to the stock's beta when ranking a company. M. Statman, 1995 Book-to-Market 1994 *The qualitative measure (e.g., value as a long-term investment) has a negative (1 year) relationship with the standard deviation of return on a security. McGuire, J., T. Schneeweis Perception of Firm Quality: Fortune Survey, *Financial measures of both risk and return influence perceptions of firm quality. and B. Branch, 1990 A Cause or Result of Firm Performance 1983 *Perceptions of firm quality more closely related to prior financial performance (1 year) than to subsequent financial performance. Solt, M. and H. Shefrin, 1989 Good Companies, Bad Stocks Fortune Survey, *When analyzing managerial performance, investors review both qualitative 1987 factors as well as financial performance such as profit margin. (1 year) McGuire, J., A. Sundgren Corporate Social Responsibility Fortune Survey, *Financial performance influence variables of social responsibility more and T. Schneeweis, 1988 and Firm Financial Performance 1983 and 1986 than the reverse. (2 years) *Reduction of firm risk is an important benefit of social responsibility. Clayman, M., 1987 In Search of Excellence: U.S. firms based *U.S.'s best-run companies beat the S&P 500 by 1.1% a year from 1981 to 1985. The Investors' Viewpoint on market/book * A portfolio of "unexcellent" S&P 500 firms with allow growth, low profitability, and P/E ratio. and low market-to-book ratios outperforms the index by 12.4 percent a year.

Exhibit 2 Time Table: Survey vs. Publication Estimation window 3 years 1 st quarter 3 rd quarter Year t-4 Year t-1 Publication Survey period Year t

Exhibit 3A: Size Effects and Firm Image - U.K. Results for Pre-Publication Year (1989-1995) Average Monthly Returns*: from 1/89 to 12/ 95 FTSE100 FTSE250 FTSE Small High Rank Mid-Rank Low Rank FTSE100-FTSE small High Rank-Low Rank Jan.-Mar. 0.90% 1.80% 2.20% 1.40% 1.40% 0.90% -1.30% 0.50% Apr.-Sept. 0.90% 0.50% 0.10% 1.20% 0.20% -1.00% -0.10% 2.20% Oct.-Dec 2.10% 1.60% 0.20% 0.90% 0.80% -0.30% 1.90% 1.20% All month 1.20% 1.10% 0.70% 1.20% 0.70% -0.40% 0.50% 1.60% Average Monthly Standard Deviations*:from 1/89 to 12/95 FTSE100 FTSE250 FTSE Small High Rank Mid-Rank Low Rank FTSE100-FTSE small High Rank-Low Rank Jan.-Mar. 4.80% 5.20% 5.60% 4.70% 4.70% 6.30% -0.80% -1.60% Apr.-Sept. 4.70% 5.50% 5.00% 4.10% 4.10% 5.70% -0.30% -1.60% Oct.-Dec 3.70% 4.20% 4.30% 2.80% 3.60% 4.80% -0.60% -2.00% All month 4.50% 5.10% 5.00% 3.90% 4.20% 5.60% -0.50% -1.70% Return/Risk Ratio: 1/89-12/95 FTSE100 FTSE250 FTSE Small High Rank Mid-Rank Low Rank FTSE100-FTSE small High Rank-Low Rank Jan.-Mar. 0.19 0.35 0.39 0.30 0.30 0.14-0.21 0.16 Apr.-Sept. 0.19 0.09 0.02 0.29 0.05-0.18 0.17 0.47 Oct.-Dec 0.57 0.38 0.05 0.32 0.22-0.06 0.52 0.38 All month 0.27 0.22 0.14 0.31 0.17-0.07-1.00-0.94 Correlation Table: 1/89-12/95 FTSE100 FTSE250 FTSE SmallHigh Rank Mid-Rank Low Rank FTSE100 1.00 FTSE250 0.91 1.00 FTSE Small 0.77 0.92 1.00 High Rank 0.81 0.60 0.51 1.00 Mid-Rank 0.73 0.80 0.69 0.84 1.00 Low Rank 0.74 0.79 0.75 0.55 0.79 1.00 Exhibit 3B: Size Effects and Firm Image - U.K. Results for Post-Publication Year (1990-1996) Average Monthly Returns*: from 1/90 to 12/ 96 FTSE100 FTSE250 FTSE Small High Rank Mid-Rank Low Rank FTSE100-FTSE small High Rank-Low Rank Jan.-Mar. 0.20% 1.40% 1.80% 1.00% 1.20% 1.40% -1.60% -0.40% Apr.-Sept. 0.80% 0.40% 0.20% 1.00% 0.60% 0.10% -0.40% 0.90% Oct.-Dec 2.10% 1.70% 0.70% 1.30% 1.60% 2.30% 1.40% -1.00% All month 1.00% 1.00% 0.70% 1.10% 1.00% 0.90% 0.30% 0.20% Average Monthly Standard Deviations*:from 1/90 to 12/96 FTSE100 FTSE250 FTSE Small High Rank Mid-Rank Low Rank FTSE100-FTSE small High Rank-Low Rank Jan.-Mar. 3.70% 4.60% 5.30% 3.30% 3.80% 5.00% -1.60% -1.70% Apr.-Sept. 4.60% 5.50% 5.20% 4.20% 4.90% 6.50% 0.30% -2.30% Oct.-Dec 2.90% 3.40% 3.20% 2.50% 3.60% 4.30% -0.30% -1.80% All month 4.00% 4.80% 4.80% 3.60% 4.30% 5.70% -0.80% -2.10% Return/Risk Ratio: 1/90-12/96 FTSE100 FTSE250 FTSE Small High Rank Mid-Rank Low Rank FTSE100-FTSE small High Rank-Low Rank Jan.-Mar. 0.05 0.30 0.34 0.30 0.32 0.28-0.29 0.02 Apr.-Sept. 0.17 0.07 0.04 0.24 0.12 0.02 0.14 0.22 Oct.-Dec 0.72 0.50 0.22 0.52 0.44 0.53 0.51-0.01 All month 0.25 0.21 0.15 0.31 0.23 0.16-0.38-0.10 Correlation Table: 1/89-12/95 FTSE100 FTSE250 FTSE SmallHigh Rank Mid-Rank Low Rank FTSE100 1.00 FTSE250 0.90 1.00 FTSE Small 0.75 0.92 1.00 High Rank 0.84 0.74 0.68 1.00 Mid-Rank 0.81 0.87 0.75 0.85 1.00 Low Rank 0.71 0.80 0.76 0.57 0.85 1.00 *Average monthly returns and standard deviations are calculated by monthly stock return data from Datastream over the period.

Exhibit 4A: Size Effects and Firm Image - U.S. Results for Pre-Publication Year (1989-1995) Average Monthly Returns*: from 1/89 to 12/ 95 FR200 FR MID FR2000 High Rank Mid-Rank Low Rank FR200-FR2000 High Rank-Low Rank Jan.-Mar. 1.00% 1.60% 2.00% 0.70% 1.50% 2.60% -1.00% -1.90% Apr.-Sept. 1.10% 0.90% 0.50% 1.80% 0.50% -0.01% 0.00% 1.81% Oct.-Dec 1.60% 1.40% 1.10% 2.00% 1.00% -0.60% 0.50% 2.60% All month 1.20% 1.10% 1.00% 1.60% 0.90% -0.10% 0.20% 1.70% Average Monthly Standard Deviations*:from 1/89 to 12/95 FR200 FR MID FR2000 High Rank Mid-Rank Low Rank FR200-FR2000 High Rank-Low Rank Jan.-Mar. 3.20% 3.60% 4.70% 3.70% 3.20% 6.00% -1.50% -2.30% Apr.-Sept. 3.40% 3.90% 4.20% 4.00% 3.60% 5.20% -0.80% -1.20% Oct.-Dec 3.10% 4.10% 4.40% 3.00% 4.10% 6.10% -1.30% -3.10% All month 3.30% 3.90% 4.40% 3.70% 3.60% 5.80% -1.10% -2.10% Return/Risk Ratio: 1/89-12/95 FR200 FR MID FR2000 High Rank Mid-Rank Low Rank FR200-FR2000 High Rank-Low Rank Jan.-Mar. 0.31 0.44 0.43 0.19 0.47 0.43-0.11-0.24 Apr.-Sept. 0.32 0.23 0.12 0.45 0.14 0.00 0.20 0.45 Oct.-Dec 0.52 0.34 0.25 0.67 0.24-0.10 0.27 0.77 All month 0.36 0.28 0.23 0.43 0.25-0.02-0.18-0.81 Correlation Table: 1/89-12/95 FR200 FR MID FR2000 High Rank Mid-Rank Low Rank Frank Russell 200 1.00 Frank Russell MID 0.90 1.00 Frank Russell 2000 0.76 0.92 1.00 High Rank 0.88 0.69 0.59 1.00 Mid-Rank 0.81 0.92 0.71 0.70 1.00 Low Rank 0.58 0.73 0.74 0.35 0.74 1.00 Exhibit 4B: Size Effects and Firm Image - U.S. Results for Post-Publication Year (1990-1996) Average Monthly Returns*: from 1/90 to 12/ 96 FR200 FR MID FR2000 High Rank Mid-Rank Low Rank FR200-FR2000 High Rank-Low Rank Jan.-Mar. 1.50% 2.10% 2.30% 1.30% 2.00% 2.90% -0.80% -1.60% Apr.-Sept. 0.90% 0.60% 0.30% 1.10% 0.20% -0.40% 0.10% 1.50% Oct.-Dec 1.90% 1.70% 1.50% 2.40% 1.20% -0.30% 0.40% 2.70% All month 1.30% 1.30% 1.10% 1.50% 0.90% 0.40% 0.20% 1.10% Average Monthly Standard Deviations*:from 1/90 to 12/96 FR200 FR MID FR2000 High Rank Mid-Rank Low Rank FR200-FR2000 High Rank-Low Rank Jan.-Mar. 2.90% 2.80% 4.10% 3.30% 2.70% 5.00% -1.20% -1.70% Apr.-Sept. 3.30% 4.00% 4.60% 3.40% 3.40% 4.70% -1.30% -1.30% Oct.-Dec 3.40% 4.10% 4.20% 3.20% 4.10% 5.50% -0.80% -2.30% All month 3.20% 3.70% 4.40% 3.30% 3.50% 5.10% -1.20% -1.80% Return/Risk Ratio: 1/90-12/96 FR200 FR MID FR2000 High Rank Mid-Rank Low Rank FR200-FR2000 High Rank-Low Rank Jan.-Mar. 0.52 0.75 0.56 0.39 0.74 0.58-0.04-0.19 Apr.-Sept. 0.27 0.15 0.07 0.32 0.06-0.09 0.21 0.41 Oct.-Dec 0.56 0.41 0.36 0.75 0.29-0.05 0.20 0.80 All month 0.41 0.35 0.25 0.45 0.26 0.08-0.17-0.61 Correlation Table: 1/90-12/96 FR200 FR MID FR2000 High Rank Mid-Rank Low Rank Frank Russell 200 1.00 Frank Russell MID 0.88 1.00 Frank Russell 2000 0.72 0.92 1.00 High Rank 0.89 0.73 0.60 1.00 Mid-Rank 0.87 0.91 0.73 0.74 1.00 Low Rank 0.58 0.73 0.82 0.55 0.79 1.00 *Average monthly returns and standard deviations are calculated by monthly stock return data from Datastream over the period.

*Exhibit 5A: CSAR of Pre & Post Publication (U.K.) High Rank Mid Rank Low Rank 1 year Pre-Publication 3.13-1.86-5.94 1year Post-Publication 0.30 0.07-0.57 CSAR: Before & After the Event (U.K.) CSAR(%) 4.00 2.00 0.00-2.00-4.00-6.00-8.00 High Rank Mid Rank 1 year Pre-Publication 1year Post-Publication Low Rank **Exhibit 5B: Pre-Survey, Survey, Post-Survey Period (U.K.) High Rank Mid Rank Low Rank Jan-June 2.35-2.06-3.06 July-Sept -0.35 0.13-1.05 Oct-Dec 0.98-0.40-1.65 CSAR: Pre,On & Post Survey Period (U.K.) CSAR(%) 3.00 2.00 1.00 0.00-1.00-2.00-3.00-4.00 High Rank Mid Rank Jan-June July-Sept Oct-Dec Low Rank Cumulative Standarized Abnormal Return (CSAR) is calculated by: R = it R CSAR mt σ i For more information, refer to page 8. * UK Most Admired Companies' rankings are published on January of each year on average. ** UK Most Admired Companies are surveyed between July and September on average.

*Exhibit 6A: CSAR of Pre & Post Publication (U.S.) High Rank Mid Rank Low Rank 1 year Pre-Publication 3.21-3.81-4.59 1year Post-Publication 1.62-3.61-3.45 CSAR: Before & After the Event (U.S.) CSAR(%) 4.00 2.00 0.00-2.00-4.00-6.00 1 year Pre-Publication 1year Post-Publication High Rank Mid Rank Low Rank **Exhibit 6B: Pre-Survey, Survey, Post-Survey Period (U.S.) High Rank Mid Rank Low Rank Jan-June 0.66-1.58-2.05 July-Sept 0.79-1.53-1.84 Oct-Dec 1.18-1.09-1.80 CSAR: Pre,On & Post Survey Period (U.S.) CSAR(%) 1.50 1.00 0.50 0.00-0.50-1.00-1.50-2.00-2.50 Jan-June July-Sept Oct-Dec High Rank Mid Rank Low Rank Cumulative Standarized Abnormal Return (CSAR) is calculated by: R = it R CSAR mt σ i For more information, refer to page 8. * US Most Admired Companies' rankings are published on January of each year on average. ** US Most Admired Companies are surveyed between July and September on average.

*Exhibit 7A: CSAR of Pre & Post Publication (U.K.) HH LH HL LL 1 year Pre-Publication 1.88 3.14-2.88-1.46 1year Post-Publication 1.85 1.51-1.56-1.86 CSAR: Before & After the Event (U.K.) CSAR(%) 4.00 3.00 2.00 1.00 0.00-1.00-2.00-3.00-4.00 1 year Pre-Publication 1year Post-Publication HH LH HL LL **Exhibit 7B: Pre-Survey, Survey, Post-Survey Period (U.K.) HH LH HL LL Jan-June 0.65 2.66-2.52-1.07 July-Sept -0.21 0.10-0.76-0.75 Oct-Dec 1.51 0.59-0.55-0.39 CSAR: Pre,On & Post Survey Period (U.K.) CSAR(%) 3.00 2.00 1.00 0.00-1.00-2.00-3.00 Jan-June July-Sept Oct-Dec HH LH HL LL Cumulative Standarized Abnormal Return (CSAR) is calculated by: R = it R CSAR mt σ i * US Most Admired Companies' rankings are published on January of each year on average. ** US Most Admired Companies are surveyed between July and September on average.

*Exhibit 8A: CSAR of Pre & Post Publication (U.S.) HH LH HL LL 1 year Pre-Publication 3.50 3.30-0.79-1.92 1year Post-Publication 0.92-1.62-0.80 1.91 CSAR: Before & After the Event (U.S.) CSAR(%) 4.00 3.00 2.00 1.00 0.00-1.00-2.00-3.00 1 year Pre-Publication 1year Post-Publication HH LH HL LL **Exhibit 8B: Pre-Survey, Survey, Post-Survey Period (U.S.) HH LH HL LL Jan-June 2.51 0.31-0.11 0.19 July-Sept 1.08-0.84-1.46-1.06 Oct-Dec 0.76-0.09-0.27-2.82 CSAR: Pre,On & Post Survey Period (U.S.) CSAR(%) 3.00 2.00 1.00 0.00-1.00-2.00-3.00-4.00 Jan-June July-Sept Oct-Dec HH LH HL LL Cumulative Standarized Abnormal Return (CSAR) is calculated by: R = it R CSAR mt σ i * US Most Admired Companies' rankings are published on January of each year on average. ** US Most Admired Companies are surveyed between July and September on average.