Are Risk-Taking Persons Less Religious?

Are Risk-Taking Persons Less Religious?

Risk-Preference, Religious Affiliation, and Religious Participation in Taiwan

Eric Y. Liu

Baylor University

ABSTRACT: Risk-preference has been identified as an important correlate of individual religiosity only in the Western societies dominated by high-risk religions such as Christianity, Islam, and Orthodox Judaism (Miller 2000; Miller and Stark 2002). This study replicates analyses of Miller (2000)—who found little significant association of risk preference and religiosity in the Buddhist Japan and Hindu India—through an examination of religious affiliation and participation in a nationally representative sample of 2,147 adults in Taiwan. From the analyses, I find supporting evidence that the relationship between risk-preference and religious affiliation is statistically insignificant among Taiwanese; however, I also find inconsistent evidence that risk-preference displays a strong, negative, and statistically significant relationship to frequency of participation in activities of religious groups in Taiwan. I argue that the general riskpreference theory has a broader application: it transcends cultures and religions, so long as they define irreligiousness as a form of risk-taking behavior. Implications of this study are discussed.

The newly advanced idea that risk preference is a strong correlate of individual religiosity is one of the controversial topics in the sociology of religion (see Miller 2000). Nevertheless, surprisingly little empirical research has examined this claim in the Eastern context, with the only exception of Miller’s (2000) analyses of the Buddhist Japan and Hindu India, based on the 1990-1993 World Values Survey (hereafter WVS). In this article, I attempt to replicate Miller’s study. I revisit the risk-preference thesis by examining religious affiliation and frequency of religious participation in a nationally representatively data set—the 2007 Taiwan Social Change Survey—for evidence that there exists a very weak or a lack of relationship between risk-preference and individual religiousness in the Eastern societies dominated by low-risk Asian religions.

THE RISK PREFERENCE THESIS

The “risk-preference” thesis claims that under some circumstances, being irreligious constitutes a form of risk-taking behavior, which can be analogous to criminal or deviant acts (Miller and Hoffmann 1995). The notion of irreligiousness as a risky career rests upon the fact that in the Western societies, the dominant religious traditions such as Christianity, Islam, and Orthodox Judaism clearly define that non-affiliation and non-participation lead to high risks of divine punishment in an afterlife (Miller and Hoffmann 1995; Miller 2000; Miller and Stark 2002). This line of thinking remains consistent with the classic studies suggesting that religion is a risk-averse strategy for reducing uncertainty in life and avoiding otherworldly punishment (e.g., Yates 1992; Malinowski 1925; Homans 1941) and rational choice theory that religious beliefs and behavior are outcomes of rational calculations of perceived costs and benefits (Stark and Bainbridge 1985, 1987).

Risk and Religion in Asian Societies

More recently, the risk-preference perspective of religion has been refined and extended to include non-Western societies, particularly the Asian nations (Miller 2000; Miller and Stark 2002). It has been argued that unlike in the West, the relationship between risk and religiosity tends to be very weak or absent in the East because the dominant Asian religions do not perceive irreligiousness as anunambiguous form of risk-taking (Miller 2000; Miller and Stark 2002). In terms of theology, Miller and Stark (2002:1412) point out that “[p]rospects of posthumous punishment are central to Christianity and Islam, as well as Orthodox Judaism. But…such notions are at most very peripheral to the major Eastern faiths.” Thus, while Christians, Muslims, and Orthodox Jews become religious out of an active state of fear of hell-fire in life after death, the motive power that causes Buddhists, Hindus, and Shintos to profess their faiths may have little to do with the apprehension of a perceived risk of otherworldly punishment (Miller 2000; Miller and Stark 2002).

Moreover, many Eastern religions such as Buddhism, Hinduism, Taoism, and popular or folk beliefsrequire no membership, remain nonexclusive, and consider religious participation mainly as individual performance rather than collective action (Miller 2000; Stark 2004; Iannaccone 1995; Stark et al 2005; Liu 2006). In view of this, Miller suggests the following:

[T]he emphasis in Eastern religious traditions is on the consequences of a person’s individual behavior (e.g., the laws of karma), and while religions offer a person guidance on proper behavior, it is ultimately the behavior itself and not the religious affiliation that is believed to be efficacious…Obviously in such a cultural setting, not joining a specific religious denomination and not participating in specific religious behaviors does not necessarily constitute risk-taking behavior (Miller 2000:7-8)

Therefore, risk-preference is expected to be unrelated to religious affiliation and participation for people living in low-risk religious societies.

Past Research: Central Empirical Evidence

Empirical tests of the model of risk preference theory have been slim, and prior findings were overwhelmingly supportive. Based on data from the 1990-1993 WVS, Miller (2000) examined the relationship between risk preference and religiosity for five nations: the U.S., Italy, Turkey, Japan, and India. The first three are dominated by the religions that teach a high risk to irreligiousness (i.e., Protestantism, Catholicism, and Islam), and the others by the low-risk religions such as Buddhism, Shinto, and Hinduism. Results showed a statistically significant association between risk preference and the religiosity measures (i.e., religious affiliation, finding comfort in religion, religious importance, and frequency of church attendance) only for the Protestant U.S., the Catholic Italy, and the Islamic Turkey, but not for the Buddhist Japan and the Hindu India.

THE PRESENT STUDY

The goal of this article is toreexamine risk preference’s relationships to religious affiliation and participation in the Eastern context,using a nationally representative sample of adults in Taiwan. The central issue in my study is whether the findings of past researchin Japan and India based on the WVS can be replicated in the Chinesesociety. Taiwan is a Chinese society characterized by high levels of religious vitality, especially since the ending of marital law in 1987 (Katz 2003; Madsen 2007). In the past two decades, the rapid growth of religious groups has resulted in the vast majority of the Taiwanese claiming a religious affiliation (Chiu 1997). The traditional Chinese religions such as Buddhism, Taoism, and popular religions have occupied the dominant status in the religious landscape, with a huge following that represents 75-80 percent of the population in Taiwan (Chiu 1997). Overall, traditional Chinese religionsretain the polytheistic and nonexclusive nature (Chiu 1997), although some Chinese religious groups have formed semi-congregations—a result of being inspired by the Western Christian practices (Yang and Ebaugh 2001). Thus, given the claim that the Chinese traditions, which dominate the Taiwan society,are typical of low-risk religions (Miller 2000; Miller and Stark 2002),a research question emerges: Is riskpreference unrelated to religious affiliation and participation in Taiwan?

DATA & METHOD

Data

Data come from the Taiwan Social Change Survey (TSCS) 2007, Phase 5, Wave 3. As one of the largest general social survey series in the world, the TSCS joined the International Social Survey Programme (ISSP)in 2002. And the 2007 TSCS includes the core ISSP questions and topicsforTaiwan. The total sample size is 2,147. For more information about sampling and data collection, refer to

Dependent Variables: Religious Affiliation & Participation

Religious affiliationRespondents were asked: “Which of the following religious denominations do you identify with?” (1= Buddhism, 2= Taoism, 3= Popular beliefs, 4= Yiguan Dao, 5= Islam, 6= Catholicism, 7= Protestantism, 8= None, 9= Others). Table 1 presents a breakdown of religious affiliation in Taiwan. Following Miller (2000), I recoded the responses into a dichotomous variable of religious affiliation (1= affiliation, 0= non-affiliation).

[Insert Table 1 Here]

Frequency of religious participationRespondents were asked: “In the past year,how frequentlyhaveyou participated in activities at churches or other religious groups?” (1= at least once a week, 2= at least once a month, 3= several times, 4= once or twice only, 5= never). Responses were reversely coded (0= never, 4= at least once a week).

Independent Variable: Risk-Preference

The 2007 TSCS includes an item asking respondents to rate on the followingstatement: “Sometimes I like to take a risk.” (1= very true, 5= not true at all) Responses were recoded in the reversed order (Note 1).

Socio-demographic Controls

Following Miller (2000), I control for age, gender, years of education, and family income. I impute the median household income category for missing values on this item, and create a “missing income” category (1= missing, 0= non-missing) to adjust potential bias for missing values in the analyses. In addition, since marriage (Thornton, Axinn, and Hill 1992) and friendship (Stark 1996; Lofland and Stark 1965) have tremendous social influences on individuals’ religious behaviors, I also control for marriage (1= married) and frequency of contact with friends (recoded as 0= never, 4= daily).

As with Miller (2000), I employ logistic regression to estimate the model for the measure of religious affiliation and ordinary least squares for the measure of frequency of religious participation.

RESULTS

Table 2 presents the results of the regression analysis for Taiwan. The first column of the table shows that risk-preference is negatively correlated with religious affiliation, but this relationship appears to be statistically insignificant, when other explanatory factors are held constant. Clearly, this finding strongly support the prediction of risk-preference theory that risk-preference is unassociated with religious affiliation in the Eastern societies dominated by low-risk religions (e.g., Miller 2000).

The second column of Table 2 displays the estimated net effect of risk-preference on frequency of religious participation. The coefficient of the measure of risk-preference seemed to be negative and statistically significant (coefficient= -.023, p<.01), after adjusting for religious affiliation and other explanatory factors. That is, individualswho like to take a risk are less likely than others to participate in activities at churches or other religious groups on a regular basis, regardless of variations in denominational and socio-demographic backgrounds. This finding seems to be inconsistent with past research showing that risk-preference is unrelated to religious attendance in Japan and India (Miller 2000).

Finally, consistent with previous findings, sociodemographic variables such as age, gender, marriage, education, and networks of friendship each have significant estimated net effects on religious affiliation or frequency of religious participation.

[Insert Table 2 Here]

CONCLUSION AND DISCUSSION

The risk-preference thesisargues that there is a very weak or lack of association between risk-preference and religiousness in the Eastern societies, where the dominant religions do not perceive non-affiliation and non-participation as a form of risk-taking behavior (Miller 2000; see also Miller and Stark 2002). This study has produced mixed support for this theoretical assertion. First, my finding strongly supports the risk-preference perspective by showing that the measure of risk-preference had little statistically significant estimated net effect on religious affiliationin Taiwan, a Chinese society traditionally dominated by a syncretic system of Buddhism, Taoism, and Chinese folk religions. Because thesedominant Chinese traditions are unchurched religions that have no membership and do not place an emphasis on exclusivity (Stark 2004), persons who choose not to belong to certain religious denomination(s) will not violate norms of the mainstream religious culture.

In addition, however,Ifound that the negative association between risk preference and frequency of religious participationwas statistically significant in Taiwan. This seems to be inconsistent with Miller’s (2000) findings in Japan and India. As I proceed to discuss this seemingly anomalous case, it should be noted that the measure of frequency of religious participation employed in this study is somewhat different from the measure of frequency of attendance at religious services used in Miller’s (2000) research based on the WVS. In my view, my measure is a more global measure—it not only taps into ritual attendance but also other religious or spiritual activities. In Chinese society, besidesformal rituals performed by priests, there also exist a wide variety ofdevotional activities—such as burning incenses in front of an altar, having pilgrims, sitting meditation, reciting mantras, praying to deities, worshiping Buddha(s), reading scriptures and so on—that can be held at templesand other religious sites (Overmyer 1986).

What might account for the negative relationship between risk preference and religious participation in Taiwan? To answer this question requires a closer look at the definition of low-risk religion (Miller 2000; Miller and Stark 2002) and the specificcultural atmosphere in the Chinese society. First, as self-evident, a low-risk religion should not be misrepresented asa non-risk religion. When Miller (2000) accounts that in the Eastern societies dominated by Asian religions the risk of non-participation is more ambiguous, he does not necessarily mean that there is no perceived risk of non–participation in those societies at all. Similarly, when Miller and Stark (2002) describe otherworldly punishment as being peripheral to the major Eastern religions, they do not claim that the religions completely rejectaperceived risk of divine punishment. As a matter of fact, as with Christianity, Islam, and Orthodox Judaism, there is always certain amount of uncertainty and risk inherent in Eastern religions (Jochim 1986; Overmyer 1986; Goodrich 1981; Gombrich 1975).

Buddhism here serves as a ready example. Buddhist belief in karma and rebirth defines that bad deeds will be punished and good deeds will be rewarded. Nevertheless, sentient beings, even with good deeds, will inevitably live a grievous life, so long as they are reborn in the endless life-cycle, being enslaved by cravings and ignorance. The way, the only way, to achieve eternal salvation and escape from life’s tortures and misery in rebirths is to practice Buddha’s eightfold noble path. Thus, those who do not practice or delay the practices put themselves at high risk. Moreover, although the notion of laws of karma and rebirth is not prominent in the basic Buddhist doctrines, it allows Buddhism to be legitimized as a moral agent of the reward and punishment in societies where it flourishes (Gombrich 1975). Thus, while the idea of karma and rebirth can be easily dispensed with in the Buddhist texts, it may still have significant social consequences when the majority of members of a society believe it (Gombrich 1975). In the sociological sense, Buddhism provides a means of social control of individuals’ deviant behaviors (Gombrich 1975).

When it comes to the specific cultural context of Chinese society,Goodrich (1981) observes, thoughts and behaviors contrary to the Confucian, Buddhist, or Taoist teachings are considered as sins and crimes; and being irreligious is a crime that leads to severe punishments in life after death. Along with killers, liars, drunkards, meat-eaters, cheaters, hookers, thieves, and other convicts, “[t]hose who have omitted to pray, who do not perform the rituals, who do not distribute printed prayers, who do not repent, non-worshipers, and skeptics, all go to Hell.” (Goodrich 1981:106) Upon arriving in Hell (s), according to specific sentences, the sinners will end up being burned in flames, butchered and hunted, boiled in oil or water, finding themselves in burning hot iron beds, having molten metal poured down their throats, their backs plowed, tongues torn out with hot iron pincers, their skin stripped off, or facing other kinds of cruelpunishment (Goodrich 1981). Not only non-participation but also improper practices will result in harsh penalties—for instance, if a priest omits a few words when he reads a religious scripture, he will be thrown into Hell and forced to read all that he left out in a light that flickers during his prison life (Goodrich 1981). Ultimately, it is out of the question of whether irreligiousness is perceived as a form of risk-taking in the Chinese society; rather, it is a matter ofto what degree conforming to Chinesereligious norms constitutes risk-taking behavior.

This study is limited to Taiwan, a Chinese enclave. Although the findings of this research seem adequate to suggest that the original risk-preference claim that there exist null correlations between risk preferences and religiosity in Eastern societies may be too broad, future research should further explore this topic in other Eastern societies. For instance, scholars may draw on nationally representative datasets to reexamine how risk-preference relates to religiosity in Indian and Japan, which are dominated by Hinduism and Shinto/Buddhism, respectively. For China scholars, more attention should be given to mainland China, where religions struggle in the Marxist-Atheist monopoly.

In sum, risk-preference’s relationship to religiosity may be a general mechanism that transcends cultures and societies, despite that the strength of the relationship may vary widely with the degree to which irreligiousness is defined as a form of risky behavior across religio-cultural settings.

NOTES

1. According to Miller and Hoffmann (1995), this risk measure reflects a general risk orientation in the form of adventure-seeking. The reason risk is conceptualized in terms of adventure-seeking is that risk preference theory of religion is centrally concerned with the question of what causes people to be irreligious, instead of asking why people are religious (Miller and Hoffmann 1995). Of course, risk can also be conceptualized in terms of investment strategy to avoid future loss (Iannaccone 1995; Miller 2000); but this line of reasoning seems to deviate from the original risk-preference argument that being irreligious is analogous to criminal and deviant behaviors. In this study I stick to the initial idea that irreligiousness is a form of risk-taking behavior.

REFERENCE

Freese, Jeremy. 2004. “Risk Preferences and Gender Differences in Religiousness:

Evidence from the World Values Survey.” Review of Religious Research 46:88-91

Iannaccone, Laurence R. 1995. “Risk, Rationality, and Religious Portfolios” Economic

Inquiry 33: 285-295.

Ch'en, Kenneth. 1964. Buddhism in China: A Historical Survey. Princeton University Press.

Eckel, Catherine C., and Philip J. Grossman, “Men, Women and Risk Aversion: Experimental Evidence.” Forthcoming in Handbook of Experimental Results, edited by C. Plott and V. Smith. New York, Elsevier. 2008