Natural disasters can bring out the best in people, even between those who mistrusted one another before calamity struck. This column investigates how a natural disaster affected residents’ perceptions of each other in a region long divided by religious conflict. Using household survey data from Central Sulawesi, Indonesia, before and after the 2018 Sulawesi earthquake, it finds that victims expressed higher expectations for support from other religious groups going forward, suggesting that the experience of intergroup cooperation during the 2018 disaster raised expectations for cooperation during future emergencies.
Natural disasters may improve our perception of and trust in others through the experience of mutual aid in the post-disaster period (Cassar et al. 2017). For example, in the recovery process from the Indian Ocean tsunami in 2004, the long-lasting military conflict between the Indonesian government and the Free Aceh Movement was resolved. Anecdotal evidence in Pandya (2006) suggests the tsunami played a major role in the resolution of the conflict. However, this is not always the case. The same tsunami escalated the confrontation between the Sri Lankan government and the Tamil Tigers in Sri Lanka, possibly because the Tamil Tigers did not allow any other organisations to provide post-disaster support directly to victims (Beardsley and McQuinn 2009).
Consistent with these two opposing anecdotes, empirical studies on the effect that natural disasters have on perceptions of others show mixed results. On the one hand, Andrabi and Das (2017) found that after an earthquake in Pakistan in 2005, Pakistanis’ trust in Westerners improved through receipt of Western aid. Cassar et al. (2017) also found that the 2004 tsunami increased trust between neighbours in rural Thailand, as measured by a trust game. On the other hand, according to Chantarat et al. (2016, 2019), rural households that experienced severe floods in Cambodia and Thailand were less likely to expect help from others in future disasters. These results suggest that the effects of natural disasters on perceptions of others can be positive or negative, depending on various factors.
Our recent paper (Kashiwagi and Todo 2021) revisits this issue, using the case of the 2018 Sulawesi earthquake, which caused 4,340 deaths and $1.45 billion in economic damages (EM-DAT 2021). This is a notable case because Central Sulawesi province, the epicentre of the earthquake, is a region of religious conflict between Muslims (78% of the province’s total population) and Christians (17%). Serious communal violence in the region from 1998 to 2001 killed between 300 and 800 people (Aragon 2001) and has not completely ceased; terrorist attacks and killings still occur occasionally (Beech et al. 2018). We examine whether the earthquake improved or deteriorated perceptions of others from various religious groups that suffered from severe conflicts in the pre-disaster period. The effect of disasters on inter-group perceptions among conflicting groups has not been studied in the empirical literature.
Our analysis is based on our post-disaster survey of 4,154 cocoa farmers near the epicentre, including those severely affected and those not severely affected by the earthquake, as shown in the figure below. One of our survey questions asked each respondent, “When you are in trouble such as food or water shortage after a natural disaster, who can you ask for help?”, with multiple choice answers of Muslim/Christian organisations and Muslims/Christians living in the same village. Combining the response to this question with information about the respondent’s own religion, we create dummy variables that indicate whether they expect intergroup emergency support from organisations of other religions and from villagers of other religions.
To examine the effect of the earthquake, our key independent variable is the distance from the Palu-Koro fault line that caused it. As the figure shows, the distance from the fault is closely related to the level of damage to houses from the earthquake. In addition, because the distance is not significantly correlated with observed household characteristics, and the distance from another neighbouring fault line to each household is not significantly correlated with the dummies for expectation of inter-group support, we presume that the distance from the Palu-Koro fault line is not correlated with the error term.
Then, we find that the distance from the fault line is negatively correlated with the measure of perceptions of villagers from other religions. A log decrease in the distance from the fault line by one standard deviation results in an increase in the dummy for expectation of inter-group support by 0.14. Because the mean of the expectation dummy is 0.20, the effect of the earthquake is quite large. In other words, as damage caused by the earthquake increased (the distance from the fault line is smaller), people were more likely to expect support from those of other religions, or post-earthquake perceptions of people from different religions became substantially better. In addition, damage caused by the earthquake was positively correlated with whether farmers were able to borrow money from neighbours from other religions when necessary. This finding implies that inter-group social networks expanded after the earthquake.
Moreover, we examine the mechanism of the positive effect the earthquake had on inter-group perceptions. For this purpose, the benchmark regression of the perception variable on the distance from the fault line incorporates additional independent variables that indicate whether the respondents or their neighbours received post-disaster support from organisations from other religions. We find that the effect of the distance from the fault line becomes smaller and less significant, while the variables for receiving inter-religion support have a positive and highly significant effect. These results imply that inter-group perceptions improved after the earthquake due to receiving support across religious groups, which is consistent with empirical evidence of Andrabi and Das (2017).
However, the effect of the earthquake was heterogeneous across farmers. For example, farmers’ perceptions of those from other religions deteriorated after the earthquake rather than improving when they lost family members, relatives, or close friends. The finding suggests that a natural disaster is most likely to negatively affect inter-group perceptions if damage caused by the disaster is substantial.
Our findings indicate that perceptions of others, including people from opposing groups, can be improved through interactions involving mutual aid; though natural disasters cause harm, they can provide an opportunity for such interactions. However, without such mutual aid after a disaster, inter-group perceptions could deteriorate. Therefore, the results of our paper suggest that post-disaster support by governments and other organisations should not focus on a particular group of victims but should be provided to a wide range of groups in a manner that promotes mutual aid across groups. This implication can also be applied to the COVID-19 pandemic, thereby calling for international cooperation to tackle the pandemic as well as the anti-globalisation and anti-foreigner sentiments arising from it.
Editor’s note: The main research on which this column is based (Kashiwagi and Todo 2021) first appeared as a Discussion Paper of the Research Institute of Economy, Trade and Industry (RIETI) of Japan.
This article first appeared on www.VoxEU.org on December 21, 2021. Reproduced with permission.