Journal of Market Focused Management, 2, 171–182 (1997)

c 1997 Kluwer Academic Publishers, Boston. Manufactured in The Netherlands. °

The Value of Olympic Sponsorships: Who is Capturing the Gold?* KATHLEEN ANNE FARRELL

University of Nebraska-Lincoln, Lincoln, NE 68588-0490 W. SCOTT FRAME

U.S. Treasury Department, Washington, D.C. 20220

Abstract In recent years, corporate sponsorship has become an increasingly important element of the marketing communications mix. This paper uses data from the 1996 Atlanta Summer Olympic Games to measure the value of Olympic sponsorship. Using stock return data, we find that the shareholders of sponsoring firms earn negative average abnormal returns around announcement of Olympic sponsorship agreements. This finding, consistent with an agency cost explanation of corporate investment practices, is robust to variation in a number of firm- and sponsorship-specific variables. In addition, cross-sectional analysis supports the monitoring hypothesis, as significant equity ownership by institutional investors is positively related to abnormal returns around announcement. Our results suggest that utilizing Olympic sponsorships in the marketing communications mix may not be valueenhancing. Keywords: Event Study, Sponsorship, Olympics, Communications Mix, Agency Theory.

1.

Introduction

The 1996 Atlanta Summer Olympic Games entertained some two million visitors, as well as a worldwide television audience. Access to such wide exposure, coupled with the overall positive image of the Games, provides corporations with an excellent vehicle for their products. In particular, marketing professionals see Olympic sponsorship as an opportunity to increase sales by increasing awareness of both new and existing products, as well as projecting their firm as a good corporate citizen. However, while tremendous marketing opportunities certainly exist, the costs associated with becoming an Olympic sponsor are both significant and escalating. In fact, according to Levine and Thurston (1992), the average fee for worldwide Olympic sponsorship rights jumped from $15 million to $40 million in only four years. Moreover, once acquired, these sponsorship rights provide corporations with the opportunity to spend millions more in public relations and promotion. For example, Frank (1996) reports that Coca-Cola intended to spend $300–$500 million in Olympics* The views expressed in this paper are those of the authors and not necessarily those of the U.S. Treasury Department.

172

KATHLEEN A. FARRELL AND W. S. FRAME

related marketing and sponsorships in 1996. These developments, in turn, raise significant questions concerning the value of Olympic sponsorships, as well as the motives of corporate managers in pursuing these investments. This paper applies a research tool from finance to a marketing related question. In particular, we use event study methodology and agency theory to analyze the value of Olympic sponsorships. Sponsorships, defined by Sandler and Shani (1989) as “the provision of resources (money, people, or equipment) by an organization directly to an event or activity in exchange for a direct association to the event or activity (p. 10),” are an integral element of the communications mix. According to Javalgi, Traylor, Gross, et al. (1994) and Meenaghan (1991), the emphasis on sponsorship as an element of the communications mix is growing as a result of both increases in the number of sponsoring companies and the amounts spent for sponsoring events. In fact, Javalgi, Traylor, Gross, et al. argue that few attempts have been made to understand the value and effectiveness of sponsorship or to assess the degree of success in achieving sponsorship goals. If the ultimate goal of the communications mix is to achieve the corporate objective of maximizing firm value, one would expect that Olympic sponsorships should be value enhancing. In this paper, we examine whether any net benefits to Olympic sponsorships actually exist by focusing on changes in shareholder wealth upon announcement of the sponsorship agreements. Nichols (1988), Fannin (1988), and Lehrman (1988) all discuss the value of Olympic sponsorships but do not focus on shareholder wealth. Given an efficient capital market, the net present value of all future expected benefits from an investment are incorporated into the stock price upon announcement. Under the wealth maximization hypothesis, average announcement-day abnormal stock returns for Olympic sponsors should be non-negative. In contrast, findings of a negative impact on returns from announcement of these sponsorships would support Jensen and Meckling’s (1976) theory of agency costs, whereby managers may make investments that benefit themselves at the expense of the firm’s stockholders. While academicians and practitioners have questioned the value of these sponsorships, we attempt to answer the question by using stock return data to measure the average expected net benefits. We also examine variation in these returns according to firm- and sponsorship-specific variables in order to understand what factors the market views as important in rendering its opinion on Olympic sponsorships. The paper is organized as follows. Section two provides additional information regarding the costs and benefits of Olympic sponsorships. Section three describes the data. An illustration of how security abnormal returns are calculated and the associated event study analysis are provided in section four. Section five examines cross-sectional variation in the abnormal returns. Section six summarizes.

2.

Olympic Sponsorship Costs and Benefits

In considering the value of Olympic sponsorship, firms must examine both the direct and indirect costs of acquiring and exploiting sponsorship rights, as well as measure the effectiveness of these investments. This section highlights the primary costs and benefits associated with Olympic sponsorship.

OLYMPIC SPONSORSHIPS

173

Direct Costs In 1991, The Atlanta Committee for the Olympic Games (ACOG) entered into a formal agreement with the United States Olympic Committee (USOC) to form Atlanta Centennial Olympic Games Properties (ACOP) in order to jointly market Olympic emblems and designations in the United States. In order to maximize its sponsorship revenues, ACOG created two categories distinguished by both cost and marketing rights: Centennial Olympic Games Partners and 1996 Olympic Games Sponsors. Corporations becoming Centennial Olympic Games Partners paid approximately $40 million (in cash and/or services) to market their products within the United States using such Olympic symbols as the torch and rings. Olympic Games Sponsors paid roughly $20 million for similar (but restricted) rights. For example, these sponsors were unable to use “Centennial” designations and were also subject to additional marketing restrictions. ACOG also received additional sponsorship revenue from the International Olympic Committee (IOC), in which members obtain worldwide marketing rights for about $40 million. Nieroth (1995) and Bennett (1994) argue that the conventional wisdom among marketing professionals is that, to successfully exploit Olympic sponsorships, firms must carefully plan how to leverage the sponsorship rights to achieve maximum visibility, goodwill, and (ultimately) increased sales. Anecdotal evidence suggests that long-term Olympic partnerships may be more value-enhancing due to the learning curve associated with optimally tying products to the Games, as well as the accumulation of goodwill. In fact, some worldwide sponsors (such as Coca-Cola, Eastman-Kodak, and Visa International) have had long-standing relationships with the Olympic Games and are continuously promoting their involvement. In essence, these firm’s have integrated Olympic sponsorship into their long-term marketing strategies. Indirect Costs Perhaps the greatest marketing fear among Olympic sponsors is that their customers do not recognize their status. For example, during the 1994 Winter Olympic Games in Lillehammer, Trenc (1994) reports survey results finding that 55% of respondents could not name three official U.S. Olympic sponsors. In this same survey, Coca-Cola was correctly identified as a sponsor by 17.7% of respondents, while their rival (and non-sponsor) Pepsi-Cola received the third highest response rate of 7.5%. Consumers may become confused due to either advertising clutter from all the various Olympic sponsors or competitors alluding to the Olympic Games through various media channels. The latter phenomena, where a sponsors’ direct competitors run advertising campaigns to confuse consumers as to the identity of the official sponsor, is referred to as either “ambush” or “parasite” marketing. For example, Levine and Thurston (1992) report that during the 1992 Winter Olympic Games in Albertville, Federal Express bought television time during the Games and successfully convinced 61% of viewers that they (and not the U.S. Postal Service) were the official sponsor. Sandler and Shani (1989) investigate the effectiveness of Olympic sponsorship in the presence of ambush marketing to determine whether official sponsors achieve consumer awareness of their status. Using consumer survey data from the 1988 Calgary Winter

174

KATHLEEN A. FARRELL AND W. S. FRAME

Olympic Games, they find that on average, ambush marketers were unable to convince consumers that they were official sponsors. In fact, they found no difference in consumer awareness between ambushers and a control group of non-sponsor, non-ambusher firms. They note, however, that in only four of seven product categories examined were the official sponsors most often correctly identified. According to Thomas (1996), one of the most effective marketing aspects of Olympic sponsorship are the hospitality packages for entertaining clients since sponsors have access to the best hotel rooms and event seating. It could be argued, however, that such benefits are simply a managerial perk, rather than a significant vehicle for increased profitability. In fact, Javalgi, Traylor, Gross, et al. (1994) claim that the individuals responsible for corporate sponsorship may often be reluctant to examine its effects because of possible career risk. In other words, sponsorship may be fulfilling personal objectives as opposed to conventional communications mix objectives. According to financial economists, such behavior is most likely to occur in those organizations in which the interests of owners and managers are not aligned, resulting in agency costs. Such agency costs generally exist in firms with both a relatively low concentration of insider ownership, coupled with a low concentration of ownership by large, outside blockholders that are often institutional investors. Benefits While the significant costs associated with Olympic sponsorship are quite clear, the benefits derived from the increased media exposure and positive image-building are more opaque. According to Meenaghan (1991), there are five main methods of measuring sponsorship effectiveness. The five methods include measuring the level of media coverage/exposure gained, measuring the communications effectiveness of sponsorship involvement, measuring the sales effectiveness of sponsorship, monitoring guest feedback, and cost-benefit analysis. According to Thomas (1996), many of the Olympic sponsors, as well as private consulting firms, measure sponsorship effectiveness by either sales figures or consumer awareness surveys. Both methods are highly problematic. For instance, Meenaghan (1991) notes that to properly interpret sales figures, they must be adjusted for non-sponsorship influences, such as seasonal trends, macroeconomic influences, and other firm-specific variables. Also, increased consumer awareness of a product or service does not necessarily imply greater profitability. Javalgi, Traylor, Gross, et al. (1994) cite the enhancement of corporate and brand images as the primary communications objectives characterizing most sponsorship activities. In fact, they find that corporate sponsorship is image-enhancing only if the company has a good image prior to sponsorship and that such endeavors may exacerbate a negative image if consumers hold prior negative perceptions. Since Olympic sponsorships include both direct costs, indirect costs, and benefits, a measure of the net benefit is essential in determining the effectiveness of the sponsorship. We employ event study methodology to measure shareholders’ assessment of various management decisions. If markets are assumed to be semi-strong form efficient, then investors should impound the new information immediately and unbiasedly into the stock price upon the announcement of the event. Thus, to the extent that an Olympic sponsorship is viewed by investors as a value enhancing (reducing) activity, stock price reaction on or around the

175

OLYMPIC SPONSORSHIPS

announcement date of the sponsorships should be non-negative (negative), on average. We attempt to measure the announcement effects using the stock price and announcement date data described in the next section. 3.

Data

The primary data employed in this paper are daily stock returns of firm’s having announced Olympic sponsorship agreements prior to the 1996 Atlanta Summer Olympic Games. Announcement dates are specified as one day prior to publication in the Wall Street Journal. We define day zero to be the day before the Wall Street Journal article is printed since the announcement may be made prior to the article appearing in print. However, to the extent that announcements occur after the closing of the stock exchange, the first day of trading on the new information may occur on day +1. We obtained a complete listing of sponsorship categories and participating firms from ACOG. We include only Olympic sponsors that had publicly-traded common stock on either the NYSE, AMEX, or the NASDAQ exchange at the time of the sponsorship announcement. As a result, both foreign and privately-held domestic companies are excluded from the analysis. The final sample includes 26 announcements. A list of the sample firms and their announcement dates can be found in Table 1. Cross-sectional data concerning firm size, ownership structure, and the percent of outstanding equity held by either insiders or large outside blockholders, are collected from Compact Disclosure. For each sponsor, the firm size and ownership structure data are collected as of January 1st of the year in which the announcement is made. The daily stock return data are obtained from the Center for Research in Security Prices (CRSP). Both the individual return series and the CRSP value-weighted index used in this paper are scaled by 100. 4.

Constructing Abnormal Returns

In this section, we discuss how to calculate abnormal security returns for our sample firms around the Olympic sponsorship announcement. For any security, an abnormal return is defined as the difference between the daily realized return and its expected value. This expected value is based on information prior to the announcement by estimating the market model. Brown and Warner (1985) provide a comprehensive outline of the event study methodology and related issues. In this study, abnormal returns are constructed in the following manner. First, the market model is estimated for each sample security by regressing the individual returns (Ri ) on the market return (Rm ) over a sample ranging from 250 trading days to 10 trading days prior to each Olympic sponsorship announcement, or: Rit = ai + bi Rmt + ²it

²it − N (0, σ 2 )

(1)

where i = 1, . . . , N sample firms (N = 26) and t = 1, . . . , T daily stock return observations (T = 241) from the pre-event sample. Second, abnormal returns (AR’s) for the

176

KATHLEEN A. FARRELL AND W. S. FRAME

Table 1. Olympic sponsorship announcements. Olympic Sponsorship Categories

Announcement Date

Worldwide Olympic Sponsors Coca-Cola Company∗ Eastman Kodak Company∗ Bausch and Lomb∗ Xerox∗ Time-Warner, Inc. (Sports Illustrated)∗

02/07/92 07/30/92 04/07/93 07/07/93 02/18/93

Centennial Olympic Games Partners NationsBank Sara Lee Corp. (Champion) The Home Depot International Business Machines (IBM) Anheuser-Busch Companies (Budweiser)∗ McDonald’s Corp.∗ AT&T∗ Delta Airlines Motorola, Inc.∗

03/04/92 07/13/92 08/21/92 09/17/92 04/13/93 10/14/93 01/11/94 07/19/94 05/24/95

Sponsors Sensormatic Electronics Corp. Bell South Corp. Southern Company (Georgia Power) Scientific Atlanta Borg-Warner Security Corp. King World Productions (Jeopardy!) General Motors Avon Products Texaco, Inc.∗ International Paper Co. The Dial Corporation Textron

03/11/94 09/19/94 12/22/93 11/29/94 11/09/94 01/24/95 03/09/95 04/26/95 08/22/95 09/14/95 11/29/95 12/11/95

∗ Repeat

Sponsor

period around the announcement date (i.e. the event window) are defined to be the daily realized return for each security net of the estimated expected return (conditional on the realized market return), or: A Ri T +k = Ri T +k − aˆ i − bˆi RmT +k

(2)

where k = 1, . . . , K out-of-sample daily returns and the estimated intercept and slope coefficients are from equation (1). We define the event window to be five trading days prior to and five trading days following the announcement, or 11 trading days. Third, we standardize the announcement-period abnormal returns by their respective standard deviations to prevent securities with large variances from dominating our hypothesis tests: S A Ri T +k = A Ri T +k /σˆ i

(3)

Under the null hypotheses, for any given day in the event window, a sample of standardized abnormal returns (SAR’s) can be characterized by the standard normal distribution.

177

OLYMPIC SPONSORSHIPS

Table 2. Mean Standardized Abnormal Returns SAR(Event Date)

Mean

Z-Statistic

Mean

T-Statistic

SAR(−5) SAR(−4) SAR(−3) SAR(−2) SAR(−1) SAR(0) SAR(1) SAR(2) SAR(3) SAR(4) SAR(5)

−0.2301 0.4250∗∗ −0.1020 0.2271 −0.1433 0.0070 −0.3273∗ −0.4284∗∗ −0.1290 0.1426 −0.0649

−1.1734 2.1670 −0.5201 1.1579 −0.7308 0.0357 −1.6690 −2.1845 −0.6578 0.7269 −0.3309

−0.2301 0.4250∗∗ −0.1020 0.2271 −0.1433 0.0070 −0.3273 −0.4284∗ −0.1290 0.1426 −0.0649

−1.0978 2.0070 −0.5320 1.0747 −0.7920 0.0448 −1.5392 −1.8713 −0.6925 0.9152 −0.3961

CSAR(Event Dates) CSAR(0, 2) CSAR(−5, 5)

−0.4323∗∗ −0.1880

−2.2041 −0.9586

−0.4323∗ −0.1880

−1.8643 −1.1564

∗ Statistically

significant at the 10% probability level. significant at the 5% probability level.

∗∗ Statistically

Hypothesis Testing In testing the wealth maximization hypothesis, we conduct two primary tests. First, we test the null hypothesis that Olympic sponsorship announcements have no impact on expected returns, or that mean standardized abnormal returns are zero. Second, we conduct crosssectional regressions of abnormal returns on potentially important explanatory variables to evaluate variation in these returns. Mean Abnormal Returns Table 2 presents the estimated mean standardized abnormal returns for each of the 11 days in the event window, as well as for series’ of days. The test statistic used is the product of each day’s mean standardized abnormal return and the square root of the number of sample firms, or: Z = S A R T +k ∗ N 1/2

(4)

Table 2 illustrates that both one and two days following Olympic sponsorship announcement the sponsors earn negative average abnormal returns. The results indicate that the market does not perceive these investments to be value-enhancing. Such a delayed reaction to these announcements may be due to the subsequent release of information pertaining to specific (or not so specific) promotional activities. Also, average standardized abnormal returns four days prior to announcement (day −4) are positive and statistically significant at the 5% probability level. These results are robust to changes in the estimation period from 241 days to 100 days and in the event window (−10 to 10 days). We also use the CRSP equally-weighted index in the market model using our 241 day estimation period and find

178

KATHLEEN A. FARRELL AND W. S. FRAME

that this reduces our significance level of the abnormal return on day +2 to 10 percent. We conclude that our results are robust to alternative specifications of the estimation period and the choice of the market index. The Z-statistic (4) tests the joint hypothesis that neither the mean nor variance are affected by the event. However, if we wish to only test the null hypothesis that mean abnormal returns are zero, we may allow for the possibility that there is event-induced variance. Under the single null hypothesis that SAR’s have a zero mean, the following t-test suggested by Boehmer, Musumeci, and Poulsen (1991) is used: v u N u X (S A Ri T +k − S A R T +k )2 T = S A R T +k /t1/N (N − 1)

(5)

i=1

In Table 2, using the t-test outlined by statistic (5), we find negative abnormal returns (at the 10% significance level) two days following announcement and positive returns four days prior. In short, the increased variance around the event date affects our inferences by lowering the significance levels of the mean abnormal returns following the sponsorship announcement. Given these results, we further examine the cumulative effect of Olympic sponsorship announcements over a number of days in the event window. We calculate cumulative standardized abnormal returns (CSAR’s), which are simply the sum of mean SAR’s divided by the square root of the number of days over a predetermined interval. We test the null hypothesis of zero abnormal performance for the full 11-day event window, as well as for a 3-day interval between the announcement day through two days following. The mean 3-day CSAR is negative and statistically significant using both test statistics. In results not reported, we further divide the sample into repeat sponsors and new sponsors. Ten firms are repeat sponsors and the remaining sixteen firms are new sponsors. We again calculate the mean 3-day CSAR between the announcement day through two days following. We find negative mean abnormal returns for both groups but are unable to reject the null hypothesis that the two groups are different. These results raise further questions as to the value of Olympic sponsorships. Specifically, we are interested in two issues: (1) is there variation in the value of Olympic sponsorships according to sponsor group?; and (2) do certain sponsors make value-reducing investments due to the presence of agency costs, or the misalignment of managerial and shareholder interests? 5.

Cross-Sectional Variation in Abnormal Returns

Next, we use the estimated security abnormal returns as the dependent variable in crosssectional regressions to test several hypotheses. The independent variables of interest include the natural logarithm of the sponsors’ total assets (ASSETS), a set of dummy variables representing each sponsorship category (GROUP), whether the firm had been a sponsor of the 1992 Summer Olympics (REPEAT), the percentage of insider ownership (INSIDE), and the percentage of ownership by large, outside blockholders (OUTSIDE).

179

OLYMPIC SPONSORSHIPS

Table 3. Sample statistics. Variable1 ASSETS($000) INSIDE OUTSIDE GROUP1 GROUP2 GROUP3 REPEAT

Mean

Maximum

Minimum

$26,163,073 6.41 59.10 0.19 0.35 0.46 0.38

$188,200,900 45.58 85.35 1.00 1.00 1.00 1.00

$524,210 0.02 27.07 0.00 0.00 0.00 0.00

1 The

variable definitions are as follows: ASSETS = the natural logarithm of the sponsors’ total assets INSIDE = the percentage of insider ownership of the firm OUTSIDE = the percentage of ownership of the firm by large, outside blockholders GROUP# = a set of dummy variables representing each Olympic sponsorship category REPEAT = a firm that was a sponsor of the 1992 Summer Olympics

Summary statistics for each variable are provided in Table 3. Inclusion of each sponsors’ total assets controls for larger firms absorbing the costs of sponsorship more easily relative to smaller firms. Therefore, we expect a positive relationship between ASSETS and abnormal returns. Because information on the actual prices paid for Olympic sponsorships are not publicly available, we include dummy variables to indicate sponsorship level to control for differences in the prices paid and benefits received. As mentioned previously, long-term sponsorships, such as the worldwide TOP agreements (GROUP1), may be more successful and therefore, value-enhancing. Thus, we expect to find a positive coefficient estimate for GROUP1. The sign of the estimated coefficient for GROUP2, Centennial Olympic Games Partner, may be negative due to the fact that this category seems to couple the high price of the TOP sponsorships with marketing rights similar to those of Sponsor (GROUP3). A dummy variable for this third category is omitted to avoid collinearity and the constant term captures the effect of this group. Repeat sponsors should have learned from their prior experiences although the market may anticipate well in advance their intention to maintain their Olympic ties. Therefore, we expect to find a positive (but perhaps insignificant) coefficient on REPEAT. Firm ownership acts as an incentive for corporate managers to align their interests with those of shareholders, thereby mitigating agency costs. However, the monotonicity of the relationship between increased managerial ownership and firm value has been questioned. In particular, Morck, Shleifer and Vishny (1988) argue that over “low” ranges of managerial ownership a conflict-of-interest exists, whereby managers’ personal preferences are not aligned with those of shareholder’s. This effect is thought to be mitigated over some “intermediate” range of managerial ownership. However, “high” levels of managerial ownership may result in an entrenchment effect due to the increasing ability of managers to make decisions without significant shareholder pressure to maximize firm value. Following Morck, Shleifer, and Vishny (1988), we examine the relationship between managerial ownership and abnormal returns by constructing the following three variables based on the

180

KATHLEEN A. FARRELL AND W. S. FRAME

percent of equity ownership by insiders: INSIDE1 = managerial ownership level if managerial ownership level < 0.05 = 0.05 if managerial ownership > 0.05 INSIDE2 = 0 if managerial ownership < 0.05 = managerial ownership level minus 0.05 if 0.05 < managerial ownership level < 0.25 INSIDE3 = 0 if managerial ownership level < 0.25 = managerial ownership level minus 0.25 if managerial ownership > 0.25 We expect the estimated coefficient on INSIDE1 to be positive since these individuals will be moving towards the intermediate level of ownership and thus, be approaching conflict resolution. The estimated coefficient in INSIDE3 should be negative due to the entrenchment effect. The coefficient on INSIDE2 is indeterminate depending on which (if either) effect dominates over that range. One could argue that agency theory predicts that the greater the percentage of institutional ownership of a firm, the greater the firm value. Institutions serve to reduce free-riding by creating shareholder coalitions with sufficient power and a significant wealth incentive to actively monitor managerial actions. To the extent that institutional ownership reduces agency costs in the firm by fulfilling a monitoring role, firm value is enhanced. Thus, we expect a positive relationship between OUTSIDE and abnormal returns. We estimate our regressions using weighted least squares (WLS), where the weights are defined as the individual security standard deviations. In light of our previous results, we estimate the following model for the announcement day, each of the two days following announcement, as well as for the 3-day (0, 2) cumulative abnormal returns discussed previously: A Ri T +k = d0 + d1 ASSETSi + d2 GROUP1i + d3 GROUP2i + d4 REPEAT +d5 INSIDE1i + d6 INSIDE2i + d6 INSIDE3i + d8 OUTSIDEi +²i T +k ²i T +k − N (0, σ 2 )

(6)

where i = 1, . . . , N sample firms (N = 26), t = 1, . . . , T daily stock return observations (T = 241) from the pre-event sample and k = 1, . . . , K out-of-sample daily returns. Table 4 presents results from the four cross-sectional regressions. Columns 1–3 provide estimates of the model for the three individual days (day0, day1, and day2), while column 4 presents results for the cumulative 3-day return (0, 2). Equations for the announcement-day returns as well as those for one-day following provide little support for any of our crosssectional hypotheses concerning the sources of abnormal returns. In contrast, the results for two days following announcement and the 3-day cumulative return provide some insight. Specifically, we find that negative abnormal returns persist even after controlling for several firm- and sponsorship-specific variables—as the constant term (i.e. the conditional mean) is negative and statistically significant. This suggests that utilizing Olympic sponsorships in the marketing communications mix may not be value-enhancing. In addition, we find a

181

OLYMPIC SPONSORSHIPS

Table 4. Cross-sectional regressions using weighted least squares. Variable1 CONSTANT ASSETS GROUP1 GROUP2 REPEAT INSIDE1 INSIDE2 INSIDE3 OUTSIDE R-Squared

AR(0)

AR(1)

AR(2)

CAR(0,2)

−2.8228 (−0.9334) 0.1207 (0.7758) −0.6522 (−0.9899) 0.4022 (0.7720) 0.2362 (0.4447) 0.0244 (0.1521) 0.0954 (1.1041) −0.1942 (−1.6078) 0.0097 (0.6210)

−2.3015 (−0.5179) 0.0269 (0.1177) −0.9151 (−0.9453) 0.1692 (0.2210) 0.1635 (0.2095) −0.0408 (−0.1728) 0.1111 (0.8746) −0.0518 (−0.2918) 0.0239 (1.0406)

−8.9423∗∗ (−2.0827) 0.3240 (1.4672) −0.4121 (−0.4406) 0.2123 (0.2870) 0.0362 (0.0480) −0.1864 (−0.8179) −0.0177 (−0.1446) 0.0134 (0.0779) 0.0614∗∗ (2.7656)

−14.0666∗ (−1.8639) 0.4716 (1.2149) −1.9794 (−1.2039) 0.7837 (0.6027) 0.4360 (0.3289) −0.2027 (−0.5061) 0.1887 (0.8751) −0.2326 (−0.7718) 0.0950∗∗ (2.4349)

0.4543

0.4516

0.3353

0.2944

T-statistics appear in parentheses. ∗ Statistically significant at the 10% probability level. ∗∗ Statistically significant at the 5% probability level. 1 The variable definitions are as follows: ASSETS = the natural logarithm of the sponsors’ total assets INSIDE = the percentage of insider ownership of the firm OUTSIDE = the percentage of ownership of the firm by large, outside blockholders GROUP# = a set of dummy variables representing each sponsorship category REPEAT = a firm that was a sponsor of the 1992 Summer Olympics

positive relationship between significant ownership by large, outside blockholders and security abnormal returns. This finding is consistent with the monitoring hypothesis, whereby these shareholders are able to wield sufficient power to discipline managers. 6.

Conclusions

In recent years, corporate sponsorship has become an increasingly important element of the marketing communications mix. This paper uses data from the 1996 Atlanta Summer Olympic Games to measure the value of Olympic sponsorship. We find that the shareholders of sponsoring firms earn negative average abnormal returns around announcement of Olympic sponsorship agreements. This finding, consistent with an agency cost explanation of corporate investment practices, is robust to variation in a number of firm- and sponsorship-specific variables. In addition, cross-sectional analysis supports the monitoring hypothesis, as significant equity ownership by institutional investors is positively related to abnormal returns around announcement. In short, our results suggest that utilizing Olympic sponsorships in the marketing communications mix may not be value-enhancing.

182

KATHLEEN A. FARRELL AND W. S. FRAME

Acknowledgments The authors thank Annette Poulsen and an anonymous referee for helpful suggestions and Martin Horn (DDB Needham) for data on past sponsors. References Bennett, John (1994). “Shopping for Sponsorships? Integration is Paramount,” Brandweek, February 14, p. 22. Boehmer, Ekkehart, Jim Musumeci, and Annette Poulsen (1991). “Event Study Methodology Under Conditions of Event Induced Variance,” Journal of Financial Economics, December, pp. 253–272. Brown, Stephen, and Jerold Warner (1995). “Using Daily Stock Returns: The Case of Event Studies,” Journal of Financial Economics, March, pp. 3–31. Fannin, Rebecca (1988). “Gold Rings or Smoke Rings?” Marketing and Media Decisions, September, pp. 64–70. Frank, Robert (1996). “Coca-Cola Ads for Games Star Olympic Fans,” Wall Street Journal, March 6, pp. B1, B10. Jafalgi, Rajshekhar, Mark Traylor, Andrew Gross, et al. (1994). “Awareness of Sponsorship and Corporate Image: An Empirical Investigation,” Journal of Advertising, December, pp. 47–58. Jensen, Michael, and William Meckling (1976). “Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure,” Journal of Financial Economics, October, pp. 305–360. Lehrman, Celia (1988). “The Gold Standard,” Public Relations Journal, July/August, pp. 20–24. Levine, Jonathan, and Karen Thurston (1992). “The Real Marathon: Signing Olympic Sponsors,” Business Week, August 3, p. 30. Meenaghan, Tony (1991). “The Role of Sponsorship in the Marketing Communications Mix,” International Journal of Advertising, pp. 35–47. Morck, Randall, Andrei Shleifer, and Robert Vishny (1988). “Management Ownership and Market Valuation: An Empirical Analysis,” Journal of Financial Economics, January/March, pp. 293–316. Nichols, Don (1988). “Do Olympic Tie-In’s Really Work?” Incentive, September, pp. 127–130. Nieroth, Alex (1995). “Success Takes Strategic, Olympic Effort,” Brandweek, August 21, p. 22. Sandler, Dennis, and David Shani (1989). “Olympic Sponsorship versus Ambush Marketing: Who Gets the Gold?” Journal of Advertising Research, August/September, pp. 9–14. Thomas, Emory (1996). “The Bottom Line,” Wall Street Journal, July 19, p. R14. Trenc, Milan (1994). “Can You Name Three Official U.S. Olympic Sponsors?” Adweek, February 21, p. 14.

Kathleen Farrell has been an assistant professor of finance at the University of Nebraska-Lincoln since 1993. Her primary teaching responsibilities have been in Corporate Finance and Commercial Banking. Her research interests include executive compensation, executive turnover, corporate governance and other agency theory issues. She is a member of the American Finance Association, American Economic Association, and the Financial Management Association. She earned her Ph.D. in finance at the University of Georgia in 1994. Prior to entering academia, Kathleen was an auditor for Peat, Marwick, Main.

W. Scott Frame is a financial economist in the U.S. Treasury’s Office of Financial Institutions Policy in Washington, D.C. Prior to moving to the Treasury in 1996, he was an economic analyst at the Federal Reserve Bank of Atlanta. He earned his Ph.D. in economics in 1996 from the University of Georgia specializing in financial economics, industrial organization, and econometrics. He has previously published research concerning the origins of antitrust, the application of antitrust laws to the banking industry, and the competitive effects of geographic regulation in banking.