How can the quality of judges’ decisions be improved and what factors influence them? NES Professor Sultan Mehmood suggests an unusual answer to this topical question. Together with co-authors Avner Seror (Aix-Marseille School of Economics) and Daniel L. Chen (Toulouse School of Economics), he examined millions of cases and thousands of judges across the Muslim world to show the effect of Ramadan observance, one of the major rituals in Muslim countries, on judicial decisions. The quantitative analysis concludes that Ramadan observance, that is associated with austerity in food and drink, paradoxically, makes judges more lenient. In this column for GURU, Sultan Mehmood talks about this paper.
Rituals are a feature of all known human societies. Whether associated with a religion, region or culture, their symbolism informs purposeful behavior. Therefore, it is not surprising that rituals have the capacity to affect our decisions, and studying the long-run impacts of traditions and rituals has long been of interest to social scientists and policymakers. In the paper (Forthcoming in Nature Human Behavior), we demonstrated these impacts by considering the example of judicial decisions.
Judicial decision-making is supposed to be immune from extraneous factors, including rituals or how hungry you are. However, behavioral biases resulting from seemingly unrelated factors such as litigants’ names or birthdates, the weather or time on the day of the decision, the political climate have been shown to have a negative effect on the quality of the judgements and make judges harsher. This then warrants that the observance of religious rituals by judges should also have an impact on the process. But how exactly does it work and is this impact positive or negative?
In our paper, we take the case of Ramadan, a religious ritual practiced by over a billion Muslims worldwide each year, to study its impact on criminal sentencing for half a century of data from India and Pakistan, which together comprise a quarter of the world’s population. Ramadan requires all practicing adult Muslims to fast from dawn to sunset as well as to focus on prayer, reflection and demonstrating self-control in moral decision-making.
We conceptualize two channels that may govern ritual’s impact on decision-making. On the one hand, physiological or nutrient deprivation of the fasting ritual can negatively affect the quality of judgments. On the other hand, better decision making may result from exerting greater effort to do the right thing. Thus, theoretically, the effect of Ramadan fasting will vary from case to case: for some judicial cases, the negative physiological deprivation effect will dominate, while the positive, better decision-making effect may dominate in others. As a result, the net effect of Ramadan fasting ritual is theoretically ambiguous, depending on the proportion of cases experiencing either positive or negative effect.
Why did we choose this particular religious ritual and why study judicial decisions? Criminal sentencing decisions provide close to an ideal setting with which to observe better decision-making seen through the lens of errors (decision reversals) and downstream consequences (recidivism or reoffense). We use micro-data from courts and the variation in the Ramadan calendar to overcome three empirical challenges that have precluded rigorous empirical analysis of rituals.
First, since the daily length of Ramadan fasting varies by geographic latitude, it provides us a source of variation in ritual intensity at the spatial level, with fasting intensity varying up to two hours on the same day. Second, because the exact month for Ramadan changes across years according to the lunar calendar, we can use variation in ritual intensity separate from seasonality (the calendar season). Third, since cases are randomly assigned, similar decisions are made by different individuals during the ritual period (the decisions of non-Muslim judges serve as a placebo group to compare with those of Muslim judges in Ramadan). In sum, our daily data frame provides a new setting to examine the effects of ritual intensity. We can disentangle the extensive margin effect of general societal shifts around the ritual season from the intensive margin effect of the ritual.
Our investigation of data reveals that the Ramadan month itself sees a 40% increase in acquittals compared to other months. In addition, Muslim judges are about 10% more likely to acquit with each additional hour of Ramadan fasting (Figure1).
Figure 1: The Ramadan Jump for Muslim and No Effect for Non-Muslim Judges
Note: The above graph plots the coefficients in our baseline regression with Ramadan Hours (t), and coefficients on day light hours during preceding and subsequent Islamic calendar months. Specifically, we also plot coefficients on Jumada al Akhirah Hours (t-3), Rajab Hours (t-2), Shaban Hours (t-1), Shawwal Hours (t+1), Dhul Kada Hours (t+2), Dhul Hijja Hours (t+3).
At the same time, the higher acquittal rate is not, however, linked to poorer quality of judgements. On the contrary, cases decided in Ramadan are also about 25 percentage points less likely to be reversed in higher courts, compared to decisions made in other months (Fig. 2). And an hour increase in Ramadan fasting reduces the likelihood that decisions are reversed in higher courts by 4%.
Figure 2: Decisions of Muslim and Non-Muslim Judges Reversed in the High Court
Note: The figures represent average overturned decisions in High Courts that were previously decided in Ramadan and those that were decided in non-Ramadan months in the Indian Lower Courts.
The higher acquittals may well be a result of physiological depletion such as disrupted sleep or nutrient deprivation. However, several patterns in data suggest this is unlikely: the reversal of decisions and recidivism or reoffense rate in decisions meted out in Ramadan is smaller by about 5%. We also observe no impact of ritual intensity on antipathy towards non-Muslim litigants, nor do we find any evidence of Ramadan impacting judges’ productivity: daily case-load and days to case resolution remain unaffected.
The effects observed are particularly pronounced for violent crimes where the accused can remain in prison for life or face similarly harsh sentences. The effects are also more pronounced for judges with fewer years of experience and potentially have more room for improvement. Reassuringly, we find no effect of Ramadan on rulings by non-Muslim judges, which consists of a much larger sample and since cases are assigned randomly across Muslim and non-Muslim judges, therefore the results are unlikely to be a statistical artifact.
Our study suggests a potentially new direction for improving the effectiveness of the justice system. The results are policy-relevant in at least two ways. First, ritual intensity appears to improve decision-making and there is a possible recognition of this improved decision-making by litigants who appeal less because of their reduced likelihood of being overturned and lowers recidivism rates of decisions in Ramadan. Second, long delays until final resolution for defendants in the criminal justice system have been a policy concern in many developing countries, and ritual-induced acquittals may help reduce overcrowding of jails and courts.
The paper provides empirical evidence that rituals affect judicial decision-making and could improve decision quality. Viewed more broadly, once the channels for a rituals’ impact are identified, they can be harnessed to complement conventional approaches to reduce cognitive load, while enhancing the efficiency of the justice system.