Divine Restraint: An experimental analysis of religious preference and intertemporal discounting

Transcription

1 Divine Restraint: An experimental analysis of religious preference and intertemporal discounting Jeremy Thornton, Brock School of Business - Samford University

2 Having, First, gained all you can, and, Secondly saved all you can, Then give all you can. John Wesley, 1840 Founder of the Methodist Church 4/11/2015 2

3 A paradox. Does religious belief make you more prosperous? 4/11/2015 3

4 Our question: Does religious belief promote delayed gratification? 4/11/2015 4

5 What is time preference? 4/11/2015 5

6 Why would we expect religion to influence time preference? 1. Religion can improve self-control (Paglieri F. Borghi, Colzato, Hommel, & Scorolli, 2013; McCullough & Willoughby, 2009) 2. Alter the relative value of material and social investments (Smith & Faris, 2005; Vaisey & Lizardo, 2010) 3. Develop skills to lower the cost of investment (Regnerus, 2003) 4/11/2015 6

7 Some other studies on religion and time preference: Religious households consider themselves more trusting, stronger bequest motives, and longer planning horizons. Oxford Economic Papers (Renneboog & Spaenjers, 2011) Christian religions are positively associated with attitudes conducive to economic growth Journal of Monetary Economics (Guiso et. al., 2003) The impact of religion on time preference varies by religious tradition. These differences were attributable to religious upbringing rather than cultural or genetic differences. Psychological Research Paglieri et. al. (2013) 4/11/2015 7

8 This paper contributes to the literature in the following ways: 1. We randomly apply religious salience through the use of a prime. <More on this later> 2. The study uses monetary payoffs (along with hypotheticals) to elicit discount rates from test subjects. 3. Using the relatively new online platform, Mechanical Turk, we are able to construct a large and diverse sample relative to previous studies. 4/11/2015 8

9 How can you test time preference? 4/11/2015 9

10 How do we get our sample? 4/11/

11 Sample: n=1000 Each subject is paid $.50 with an opportunity for increased payments. 4/11/

12 4/11/

13 What about those primes? 4/11/

14 Repeat questionnaire with opposite prime Demographic Questions Survey Flow With Pay MCQ Religious Prime n=500 No Pay MCQ Hypothetical With Pay MCQ MCQ Control Prime N=500 No Pay MCQ Hypothetical MCQ 4/11/

15 Treatments Freq. Prime Prime al w/pay Religion Control al w/pay Control Religion Hypothetical Religion Control Hypothetical Control Religion al w/o pay Religion Control al w/o pay Control Religion Total /11/

16 Religious Preference Freq. Percent Agnostic Protestant Atheist Catholic Other Christian Other Jewish Buddhist Mormon Hindu Muslim Orthodox Sikh Total /11/

17 Religious Attendance Sikh Orthodox Muslim Hindu Mormon Buddhist Jewish Other Other Christian Catholic Atheist Protestant Agnostic Never On Holidays Several x /Year Monthly 2-3x /Month 1x /week >1x /week 4/11/

18 Religious Importance Sikh Orthodox Muslim Hindu Mormon Buddhist Jewish Other Other Christian Catholic Atheist Protestant Agnostic Not Important Very Unimportant Somewhat Unimportant Somewhat Important Very Important Extremely Important 4/11/

19 Sex Education Male Female Total Some HS HS Some College College Graduate School Total 876 1,024 1,900 4/11/

20 Buddhist Hindu Other Christian Jewish Catholic Other Agnostic Atheist Mormon Protestant Sikh Muslim Orthodox Discount Rates by Religious Preference mean of k Religious Preference Mean K SD k Buddhist Hindu Other Christian Jewish Catholic Other Agnostic Atheist Mormon Protestant Sikh Muslim Orthodox /11/

21 Fraction k 4/11/

22 .1 Fraction abs_lkhat 4/11/

23 Adjusted Predictions of reg_pref with 95% CIs Linear Prediction abs_lkhat Buddhist Catholic Hindu Jewish Mormon Muslim Orthodox Protestant Sikh Agnostic Atheist Other Other Christian 4/11/

24 abs_lkhat by religious preference and religious attendance Linear Prediction abs_lkhat Buddhist Catholic Hindu Jewish Mormon Muslim Orthodox Protestant Sikh Agnostic Atheist Other Other Christian relig_attend_yes=0 relig_attend_yes=1 4/11/

25 10 abs_lkhat by Religious Preference and Importance Linear Prediction abs_lkhat Buddhist Catholic Hindu Jewish Mormon Muslim Orthodox Protestant Sikh Agnostic Atheist Other Other Christian relig_importance_yes=0 relig_importance_yes=1 4/11/

26 Margins over Religious Preference and Religious Treatment Linear Prediction abs_lkhat Buddhist Catholic Hindu Jewish Mormon Muslim Orthodox overall_treat=0 Protestant Sikh Agnostic overall_treat=1 Atheist Other Other Christian 4/11/

27 Treatment Effects over Religious Preference if Religious Attendence = High Linear Prediction abs_lkhat Buddhist Catholic Hindu Jewish Mormon Muslim Orthodox overall_treat=0 Protestant Sikh Agnostic overall_treat=1 Other Other Christian 4/11/

28 Treatment Effects over Religious Preference if Religious Importance = High Linear Prediction abs_lkhat Buddhist Catholic Hindu Jewish Mormon Muslim Orthodox overall_treat=0 Protestant Sikh Agnostic overall_treat=1 Atheist Other Other Christian 4/11/

29 : There is strong support in the literature that religious belief should be positively associated with increased self control. Other papers (which rely on surveys or smaller samples) have found relatively modest confirmation of this hypothesis. Our approach (with larger samples and an experimental payoff design) found no impact on religious belief on subjects discount rates. 4/11/

30 Jeremy Thornton Dwight Moody Beeson Chair of Business Associate Professor of Economics Brock School of Business Samford University 4/11/