اعتقادات مذهبی، نگرش قمار و نتایج بازار مالی

This study investigates whether geographic variation in religion-induced gambling norms affects aggregate market outcomes. We conjecture that gambling propensity would be stronger in regions with higher concentrations of Catholics relative to Protestants. Consistent with our conjecture, we show that in regions with higher Catholic–Protestant ratios, investors exhibit a stronger propensity to hold lottery-type stocks, broad-based employee stock option plans are more popular, the initial day return following an initial public offering is higher, and the magnitude of the negative lottery-stock premium is larger. Collectively, these results indicate that religion-induced gambling attitudes impact investors' portfolio choices, corporate decisions, and stock returns.

Gambling and speculation play an important role in financial markets. These and related activities are often associated with high levels of trading volume, high return volatility, and low average returns (e.g., Scheinkman and Xiong, 2003; Hong, Scheinkman, and Xiong, 2006; Grinblatt and Keloharju, 2009 and Dorn and Sengmueller, 2009). As gambling attains wider acceptability in society and a “lottery culture” emerges (e.g., Shiller, 2000), the influence of gambling behavior in financial markets is likely to increase and could have economically significant effects on corporate decisions and stock returns. Specifically, in market settings that superficially resemble actual gambling environments and in which skewness is a salient feature, people's gambling attitudes may influence aggregate market outcomes. For example, if the positively skewed returns of initial public offering (IPO) stocks lead investors to perceive IPOs as lotteries, their preference for lottery-like payoffs and trading behavior could generate initial overpricing (e.g., Barberis and Huang, 2008). More generally, if investors exhibit a preference for stocks with lottery features, all else equal, stocks with lottery-type characteristics would earn lower average returns.1 Similarly, the popularity of broad-based employee stock option (ESO) plans has been difficult to explain within the traditional economic framework (e.g., Oyer and Schaefer, 2004, Bergman and Jenter, 2007 and Kedia and Rajgopal, 2009). One potential explanation for this puzzle is that option grants to non-executives reflect the gambling preferences of rank and file employees (e.g., Spalt, 2009). Individuals with strong gambling preferences may find firms that offer option-based compensation plans attractive if they view stock options as “lottery tickets.”2 Some managers may even attempt to cater to those preferences. The important role of gambling in various market settings has been recognized in the recent asset pricing and corporate finance literatures. However, it has been difficult to attribute aggregate market outcomes directly to people's gambling preferences because individual-level gambling and speculative activities cannot be directly observed. In this paper, we use people's religious beliefs as a proxy for their gambling propensity and examine whether geographical variation in religious composition, particularly the variation in the ratio of Catholics to Protestants across U.S. counties, allows us to identify market-wide effects of gambling behavior. Our choice of religious composition as a proxy for gambling propensity is motivated by the observation that gambling attitudes are strongly determined by one's religious background. In particular, the Protestant and Catholic churches have very distinct views on gambling.3 A strong moral opposition to gambling and lotteries has been an integral part of the Protestant movement since its inception, and many Protestants perceive gambling as a sinful activity (e.g., Starkey, 1964, Ozment, 1991 and Ellison and Nybroten, 1999). Although individual Protestant churches vary in the intensity with which they oppose gambling, the opposition to gambling is quite general. The largest Protestant group, the Southern Baptists, is particularly strident in their censure of gambling. In contrast, the Roman Catholic Church maintains a tolerant attitude towards moderate levels of gambling and is less disapproving of gambling activities. It has even used gambling in the form of bingo and charitable gaming events as an important source of fundraising (e.g., Diaz, 2000 and Hoffman, 2000). Among other prominent religious denominations in the U.S., people of Jewish faith behave similar to Catholics and accept gambling activities more readily, while the gambling attitudes of Latter-Day Saints (Mormons) are aligned more closely with those of Protestants. The impact of these diverse viewpoints on gambling is evident in state lottery adoption policies and levels of lottery expenditures. Prior empirical research has shown that the popularity of state lotteries in a region is affected by the dominant local religion (e.g., Berry and Berry, 1990, Martin and Yandle, 1990 and Ellison and Nybroten, 1999). A few recent studies also demonstrate that religion-induced gambling attitudes carry over into financial decisions (e.g., Kumar, 2009; Doran, Jiang, and Peterson, in press). We confirm these findings using our county-level measures of religious composition. In particular, we show that states with higher concentrations of Catholics relative to Protestants (i.e., higher Catholic–Protestant ratio (CPRATIO)) are more likely to legalize state lotteries and adopt them earlier. Further, at both state and county levels, we find that per capita lottery sales are higher in regions with high CPRATIO. We also show that individual investors located in high CPRATIO regions assign larger portfolio weights to lottery-type stocks (see Fig. 1) and confirm that religion-induced gambling attitudes carry over into financial decisions. Full-size image (87 K) Fig. 1. Geographical variation in religiosity and religious composition across the U.S. This figure shows the county-level religiosity (Panel A) and Catholic–Protestant ratio (Panel B) across the U.S. Each small outlined region in the figure corresponds to a county. In the top panel, darker shade indicates more religious counties, while in the bottom panel, darker shade indicates counties with higher Catholic concentration. The county-level religion data are available for years 1980, 1990, and 2000 and the figure shows the average religiosity and concentration measures for these three years. Figure options Motivated by these empirical findings, we conjecture that religion-induced heterogeneity in gambling preferences and behavior could affect economic decisions in other settings. In particular, the predominant local religion could influence local cultural values and norms and consequently affect the financial and economic decisions of individuals located in that region, even if they do not personally adhere to the dominant local faith.4 Further, these financial and economic decisions could aggregate and generate market-wide forces that can potentially influence aggregate financial market outcomes. We consider four specific economic settings in which the existing literature has suggested the possible role of gambling and examine the link between religious beliefs, gambling attitudes, and aggregate market outcomes. First, we examine the extent to which geographical heterogeneity in religious beliefs influences investors' portfolio choices. We find that the portfolio characteristics of institutional investors are influenced by the religious characteristics of the neighborhoods in which they are located. Although institutions on average tend to avoid lottery-type stocks (e.g., Kumar, 2009), institutions located in high CPRATIO regions assign larger weights to stocks with lottery features and simultaneously underweight non-lottery-type stocks. The impact of local religious norms on portfolio decisions is significant only among smaller and moderate-sized institutions. Among larger institutions with more standardized investment practices (e.g., Baker, Bradley, and Wurgler, 2011), local religious environment is not a significant determinant of institutional portfolio decisions. Further, the religion-induced differences in stock holdings are stronger among “aggressive” institutions and those institutions that trade more actively or hold concentrated portfolios containing few stocks. Over time, preference for lottery-type stocks is amplified during the end of the year when the temptation to engage in risk-seeking and gambling-type activities is likely to be stronger due to performance-based incentives (Brown, Harlow, and Starks, 1996). We conduct several tests to ensure that these results are not induced by geographical clustering of institutions in a few large counties and financial centers or due to repeated observations. Second, we investigate a corporate finance puzzle: Why do firms grant options to non-executive rank and file employees? We show that broad-based employee stock option plans, which would appeal more to employees with strong gambling preferences, are more popular in high CPRATIO regions where individuals are likely to exhibit a stronger propensity to gamble. Further, consistent with our gambling interpretation, we find that the sensitivity of the level of non-executive option grants to local religious composition is greater among high volatility firms, which would be more attractive to individuals with strong gambling preferences due to their higher skewness. These results indicate that the puzzle of broad-based employee stock option plans could at least be partially resolved within a theoretical framework that recognizes the potential link between compensation and gambling. Third, we focus on the IPO markets and test one of the key empirical predictions of the Barberis and Huang (2008) model. They conjecture that excess speculative demand of skewness-loving investors can generate overpricing in securities such as IPOs that have positively skewed returns. Consistent with this conjecture, we find that the initial day return following an initial public offering is higher for IPO firms located in high CPRATIO regions where the propensity to gamble is likely to be higher. To strengthen the link between first-day IPO return and gambling propensity of local investors, we show that the relation between initial day returns and CPRATIO is stronger in regions with higher stock market participation rates (as proxied by higher income and higher education levels) and stronger local bias. In these areas, local investors are more likely to trade local IPOs and, thus, more likely to play a marginal price-setting role. Collectively, our IPO results indicate that the puzzling phenomenon of IPO underpricing is at least partially influenced by the gambling behavior of local investors. In the last part of the paper, we study the effect of gambling on stock returns in a broader market setting. Specifically, we investigate the pricing of stocks with lottery-type characteristics. This exercise is also motivated by the theoretical predictions in Barberis and Huang (2008), who conjecture that securities with lottery features are expected to earn lower average returns because investors are willing to accept lower average returns for a tiny probability of a large potential gain. Consistent with their conjecture, we find that lottery-type stocks earn lower average returns. In addition, consistent with our gambling hypothesis, we find that the magnitude of the negative lottery-stock premium is stronger in regions with high CPRATIO. Our empirical findings are robust to a large number of variations to the baseline specifications. In particular, when we use the Rajan and Zingales (1998) method to account for unobserved heterogeneity at the state and county levels, we obtain results that are qualitatively similar to our baseline results. Overall, our empirical results indicate that religion-induced gambling norms influence gambling preferences, individual-level economic decisions, and aggregate market-level outcomes. Both corporate policies and asset prices are influenced by religion-induced local gambling propensity. These results complement recent evidence in Hilary and Hui (2009) and show that religion influences financial market outcomes not only through the risk aversion channel but also through its effect on the skewness and gambling preferences of individuals.5 Further, while previous studies indicate that religion could influence economic growth (Barro and McCleary, 2003) and the level of investor protection in a country (Stulz and Williamson, 2003), our results highlight the importance of religious composition at a more disaggregate individual and firm level. In broader terms, our empirical evidence contributes to the emerging literature in economics that examines the interplay between culture and economic outcomes (e.g., Guiso et al., 2003 and Guiso et al., 2006). Because religion is a key cultural attribute, our results indicate that through its impact on people's gambling attitudes, cultural shifts can influence aggregate financial market outcomes. The rest of the paper is organized as follows. In the next section, we summarize our key testable hypotheses. We describe our main data sources in Section 3 and motivate the choice of our gambling proxy in Section 4. We present our main empirical results in Section 5 and conclude in Section 6 with a brief summary.

We use religious background as a proxy for gambling propensity and examine whether geographical variation in religion-induced gambling norms affects aggregate market outcomes. We focus on four distinct economic settings in which the existing literature has suggested a role for gambling and speculative behavior. Our results indicate that in regions with higher concentrations of Catholics relative to Protestants, institutions hold larger lottery-type stock portfolios, non-executive employees receive larger stock option grants, the initial day IPO return is higher, and the magnitude of the negative lottery-stock premium is higher. These seemingly unrelated findings are driven by a common gambling-based mechanism. Operating through this gambling channel, religion influences investors' portfolio choices, corporate decisions, and stock returns. The consistency in the relation between local religious composition and aggregate market outcomes in multiple settings provides strong support to our gambling-related hypotheses and highlights the important role of gambling in understanding aggregate market outcomes. In broader terms, our empirical evidence contributes to the emerging literature in economics that examines the interplay between culture and economic outcomes. Because religion is one of the key cultural attributes, our results indicate that through an influence on gambling attitudes, cultural norms may impact financial markets more strongly than previously believed. Our study focuses on economic settings in which the existing literature has already suggested a possible role of gambling and speculation. However, the remarkable similarities in our results across different economic settings indicate that religious beliefs might be important in other economic settings and could have an even stronger influence on financial markets. For example, recent studies in corporate finance indicate that managerial incentives affect corporate policies. It is likely that when offered similar contracts, the religious background of a manager determines her response to performance-based incentives. Therefore, through its effect on ethics and values, differences in religious beliefs could have a significant effect on corporate policies, including capital structure choices.