Alpaslan Akay

1
Do Religious Rituals Promote Cooperation and
Punishment?-Evidence from Ramadan

• 2007-07-01
• Alpaslan Akay
• Department of Economics, Göteborg University,
  Sweden
• Gokhan Karabulut
• Department of Economics, Istanbul University,
  Turkey
• Peter Martinsson
• Department of Economics, Göteborg University,
  Sweden

2
Introduction

• Religion influences our daily life in many
  perspectives from our legal system to our
  calendar
• We are also affected by several religious rituals
  such as Easter, Christmas and Ramadan and by many
  others
• The effects and functions of religion and
  religious rituals have been discussed since
  Durkheim (1912), Geertz (1966) and Weber (1930)
• It has recently been argued that one of the
  primary functions of religion and religious
  rituals is to increase solidarity and cohesion in
  a society implying that pro-social behaviour of
  the individuals such as cooperation and trust
  should be promoted (Sosis and Ruffle (2003, 2004,
  2005) Iannaccone (1992, 1994))
• Then, it is crucial to know whether religion and
  religious rituals can help to solve collective
  action problems in a society

3
Introduction

• Ramadan is the one of the most celebrated
  religious rituals in the world followed partly or
  fully by more than 1.5 billion Muslims
• Ramadan imposes a social shock on ordinary daily
  activities. It affects what people dress, talk
  etc. during one month
• Ramadan contains the fifth pillar of Islam, which
  is fasting (sawm or siyam). Fasting requires
  strict avoidance of fluids and nourishment from
  dawn (sahur) to sunset (iftar) and it is
  obligatory for every Muslim (Quran, 2183-84)
• Islamic scholars indicate that the aim of fasting
  is to improve the ability of empathy for the
  poor, sharing, kindness, generosity, forgiveness
  and self-control (Buhkari, 1987)

4
Hypothesis

• H0 No effect of Ramadan month on the
  cooperation and punishment levels
• H0 No effect of high degree of participation
  into fasting on cooperation and punishment levels

5
Experimental Design

• Two treatments Ramandan and Non-Ramadan
• The Ramadan experiment is conducted in Leylat
  al-Qadr (The Night of Power) which is the 27th
  day of Ramadan. The non-Ramadan experiment is
  conducted after one month of Ramadan.
• Each has two linear public good experiments, one
  without, and one with punishment possibility.
• The designs are the same as Gächter et al.
  (2004), and Gächter and Herrmann (2006).
• We use 90 undergraduate students in Istanbul
  University, Turkey, which is a secular country.
  30 groups each with three anonymous subjects.
• We elicit the subjects beliefs about how much
  the others on average have contributed. We reward
  correct guesses by applying the same monetary
  incentive scheme as in Gächter and Renner (2006).
  

6
Results 1
Table 1a. Contributions in the public goods experiment Table 1a. Contributions in the public goods experiment Table 1a. Contributions in the public goods experiment Table 1a. Contributions in the public goods experiment Table 1a. Contributions in the public goods experiment Table 1a. Contributions in the public goods experiment
Sub-groups Sub-groups Part I (without punishment) Part I (without punishment) Part II (with punishment) Part II (with punishment)
Sub-groups Sub-groups Realized contribution Expected contribution Realized contribution Expected contribution
Ramadan experiment Full-fasters 8.10 9.82 10.0 11.41
Ramadan experiment Non full-fasters 7.42 9.22 8.56 10.12
Ramadan experiment Total 7.63 9.49 9.14 10.68

Non-Ramadan experiment Full-fasters 7.56 9.56 9.79 11.38
Non-Ramadan experiment Non full-fasters 10.91 11.23 11.37 12.26
Non-Ramadan experiment Total 9.46 10.46 10.58 11.88

7
Results 2
Table 1b. P-values from the test of no contribution differences between different sub-groups Table 1b. P-values from the test of no contribution differences between different sub-groups Table 1b. P-values from the test of no contribution differences between different sub-groups Table 1b. P-values from the test of no contribution differences between different sub-groups Table 1b. P-values from the test of no contribution differences between different sub-groups
Sub-groups Part I (without punishment) Part I (without punishment) Part II (with punishment) Part II (with punishment)
Sub-groups Realized contribution Expected contribution Realized contribution Expected contribution
H0 No difference between Ramadan and Non-Ramadan (in general) H0 No difference between Ramadan and Non-Ramadan (in general) H0 No difference between Ramadan and Non-Ramadan (in general) H0 No difference between Ramadan and Non-Ramadan (in general) H0 No difference between Ramadan and Non-Ramadan (in general)
Ramadan vs. Non-Ramadan 0.01 0.03 0.05 0.004
H0 No difference between Full-Fasters vs. Non-Full Fasters (in general) H0 No difference between Full-Fasters vs. Non-Full Fasters (in general) H0 No difference between Full-Fasters vs. Non-Full Fasters (in general) H0 No difference between Full-Fasters vs. Non-Full Fasters (in general) H0 No difference between Full-Fasters vs. Non-Full Fasters (in general)
Full-Fasters vs. Non-Full Fasters 0.16 0.51 0.78 0.99

H0 No difference between full-fasters and non-full fasters within experiments H0 No difference between full-fasters and non-full fasters within experiments H0 No difference between full-fasters and non-full fasters within experiments H0 No difference between full-fasters and non-full fasters within experiments H0 No difference between full-fasters and non-full fasters within experiments
Ramadan 0.35 0.33 0.10 0.21
Non-Ramadan 0.004 0.04 0.29 0.19

H0 No difference between full-fasters across experiments H0 No difference between full-fasters across experiments H0 No difference between full-fasters across experiments H0 No difference between full-fasters across experiments H0 No difference between full-fasters across experiments
Full-Fasters 0.61 0.78 0.96 0.50
H0 No difference between non-full fasters across experiments H0 No difference between non-full fasters across experiments H0 No difference between non-full fasters across experiments H0 No difference between non-full fasters across experiments H0 No difference between non-full fasters across experiments
Non-Full Fasters 0.0003 0.002 0.005 0.001
Note. All tests are based on Mann-Whitney-U tests (two-sided) Note. All tests are based on Mann-Whitney-U tests (two-sided) Note. All tests are based on Mann-Whitney-U tests (two-sided) Note. All tests are based on Mann-Whitney-U tests (two-sided) Note. All tests are based on Mann-Whitney-U tests (two-sided)
8
Results 3
Table 2a. Punishment in the public goods experiment Table 2a. Punishment in the public goods experiment Table 2a. Punishment in the public goods experiment Table 2a. Punishment in the public goods experiment Table 2a. Punishment in the public goods experiment Table 2a. Punishment in the public goods experiment
Sub-groups Sub-groups Average punishment Average punishment Proportion of Punishers Average punishment points awarded if
Sub-groups Sub-groups Realized Expected Proportion of Punishers Average punishment points awarded if
Ramadan experiment Full-fasters 1.18 2.38 0.44 2.60
Ramadan experiment Non full-fasters 1.77 1.91 0.46 3.72
Ramadan experiment Total 1.57 2.15 0.45 3.33

Non-Ramadan experiment Full-fasters 0.82 2.04 0.28 2.81
Non-Ramadan experiment Non full-fasters 1.19 1.85 0.38 3.21
Non-Ramadan experiment Total 1.03 1.90 0.33 3.08
9
Results 4
Table 2b. P-values from the test of no difference between different sub-groups Table 2b. P-values from the test of no difference between different sub-groups Table 2b. P-values from the test of no difference between different sub-groups Table 2b. P-values from the test of no difference between different sub-groups Table 2b. P-values from the test of no difference between different sub-groups
Sub-groups Average punishment Average punishment Proportion of punishers Average punishment points awarded if
Sub-groups Realized Expected Proportion of punishers Average punishment points awarded if
H0 No difference between Ramadan and Non-Ramadan (in general) H0 No difference between Ramadan and Non-Ramadan (in general) H0 No difference between Ramadan and Non-Ramadan (in general) H0 No difference between Ramadan and Non-Ramadan (in general) H0 No difference between Ramadan and Non-Ramadan (in general)
Ramadan vs. Non-Ramadan 0.08a 0.53a 0.05b 0.43a
H0 No difference between Full-Fasters vs. Non-Full Fasters (in general) H0 No difference between Full-Fasters vs. Non-Full Fasters (in general) H0 No difference between Full-Fasters vs. Non-Full Fasters (in general) H0 No difference between Full-Fasters vs. Non-Full Fasters (in general) H0 No difference between Full-Fasters vs. Non-Full Fasters (in general)
Full-Fasters vs. Non-Full Fasters 0.10a 0.18a 0.48b 0.04a

H0 No difference between full-fasters and non-full fasters within experiments H0 No difference between full-fasters and non-full fasters within experiments H0 No difference between full-fasters and non-full fasters within experiments H0 No difference between full-fasters and non-full fasters within experiments H0 No difference between full-fasters and non-full fasters within experiments
Ramadan 0.17a 0.44a 0.28b 0.04a
Non-Ramadan 0.26a 0.29a 0.34b 0.53a

H0 No difference between full-fasters across experiments H0 No difference between full-fasters across experiments H0 No difference between full-fasters across experiments H0 No difference between full-fasters across experiments H0 No difference between full-fasters across experiments
Full-Fasters 0.40a 0.90a 0.03b 0.98a
H0 No difference between non-full fasters across experiments H0 No difference between non-full fasters across experiments H0 No difference between non-full fasters across experiments H0 No difference between non-full fasters across experiments H0 No difference between non-full fasters across experiments
Non-Full Fasters 0.18a 0.68a 0.58b 0.27a
Note. a Based on Mann-Whitney-U tests (two-sided) , b based on chi two-sample test of equal proportions in the groups punished zero, punished one and punished two members respectively. Note. a Based on Mann-Whitney-U tests (two-sided) , b based on chi two-sample test of equal proportions in the groups punished zero, punished one and punished two members respectively. Note. a Based on Mann-Whitney-U tests (two-sided) , b based on chi two-sample test of equal proportions in the groups punished zero, punished one and punished two members respectively. Note. a Based on Mann-Whitney-U tests (two-sided) , b based on chi two-sample test of equal proportions in the groups punished zero, punished one and punished two members respectively. Note. a Based on Mann-Whitney-U tests (two-sided) , b based on chi two-sample test of equal proportions in the groups punished zero, punished one and punished two members respectively.
10
Conclusions

• Ramadan month influence the cooperation behavior
  significantly
• negative but its effect on punishment
  behavior is insignifcant,
• In general, there is no difference in
  contribution and punishment levels between
  full-fasters and non full-fasters,
• In Ramadan, the full-fasters contribute an
  insiginifcantly higher amount
• than non full-fasters,
• In non-Ramadan, the non full-fasters contribute
  significantly higher than full-fasters,
• The sub-group which change their behavior is non
  full-fasters.