The link between climate change, food security and fertility: The case of Bangladesh

positive impact on the change in TFR. In addition, we find that fertility responds negatively to GDP per capita in the short term. The error correction term reveals a long-term relationship between fertility, GDP per cap- ita, maximum temperature, mean temperature, maximum rainfall, and mean rainfall. Only the coefficients of ln(maxtemp) and ln(GDPPC) are significant, with values -7.39 and 0.499, respectively. This means that in the long-term equilibrium, the increase in maximum tempera- ture would have a positive impact on the TFR and the increase in GDP per capita would have a negative impact on the TFR. Concluding discussion Path analysis shows that climatic factors, especially maximum temperature and precipitation, can affect fertility directly and to some extent only maximum temperature indirectly through crop production, while crop production affects fertility directly and indirectly through infant mortality. But the contribution of indicators of economic development is also crucial in influ- encing fertility rates directly or indirectly. Higher GDP per capita is likely to reduce fertility rates by reducing infant mortality. Indirectly speaking, the increasing maximum temperature is likely to increase crop produc- tion; increasing crop production is likely to reduce infant mortality; decreasing infant mortal- ity contributes to the fertility decline. The positive relationship between maximum temperature and crop production is consistent with the findings from the study by Sarker et al., who also found a positive association of maximum temperature with Aus and Aman

variety in Bangladesh [53]. Crop production has a significant negative impact on the infant mortality rate. Increased food production increases food security, ensures food intake and reduces susceptibility to diseases [82], improving women’s reproductive health and increasing child survival. In other words, increased food insecurity is directly associated with increased child mortality [83, 84]. Our path analysis confirms that the infant mortality rate positively impacts the TFR, which is consistent with the literature and existing theories of the replace- ment effect [74, 75, 85, 86]. Different from findings in Winkler-Dworak’s study has that the context of sub-Saharan Africa [67], fertility decreases as food production increases, we have demonstrated a positive direct effect of crop production and a negative indirect effect of crop production (via infant mortality) on the fertility, thus resulting in a significant total impact. Moreover, the total effect of maximum temperature on TFR is significantly negative, mainly because of the negative direct effect. It means that as maximum temperature increases, TFR decreases. This is consistent with previous studies that warm weather reduces birth rates nine months later [63, 64, 87]. In assessing the short- and long-term relationship between climate change and fertility, we ran error correction models. Fertility is found to respond negatively to maximum temperature in the short run but positively in the long run. Eissler et al. found that both short- and the

long-term temperature has adverse effects on fertility [65]. The inconsistency with the findings of Eissler et al. [65] is likely because their study focuses on sub-Saharan African countries, where the climate is very different from Bangladesh. The reduction in fertility with increasing short-term maximum temperature in our study could be explained by increased migration, which in turn could affect the frequency of sexual intercourse. Carrico and Donato conclude that extreme weather conditions such as drought periods in Bangladesh are more consistently associated with increased migration of male household heads [31]. In addition, reduced birth rates could also be attributed to worse reproductive health at conception [59]. Besides, although fertility responds positively to the mean rainfall in the short term, it is insensitive to the long term. This seems to contradict the results of a recent study by Eissler et al. in 18 Sub- PLOS ONE Climate change, food security and fertility PLOS ONE | https://doi.org/10.1371/journal.pone.0258196 October 21, 2021 12 / 18