Three essays on the consequences of seismic and environmental shocks on society

Natural disasters and environmental shocks are destructive events that are growing in frequency and severity, driven also by climate change. Despite the large implications of these shocks on society, their impact has been under-investigated in several fields. Besides the direct impact of natural disasters, the exogenous nature of these shocks also provides the opportunity to disentangle individuals' and institutions' behavior and mechanisms, which otherwise would not be identifiable due to endogeneity problems. The three essays that compose this thesis contribute to fill this gap by analyzing some implications of seismic and environmental shocks. In the first essay, we analyze the response of municipalities to occurrences of earthquakes in terms of spending behavior, use of upper tier transfers and recovery, using balance sheet data of about 8000 Italian municipalities for the period 2000-2015 and the universe of earthquakes events. We find evidence of growing expenditure for about 12 years after the shocks, with asymmetric responses between earthquake-related and unconditional grants, and heterogeneous flypaper effects across the country. While in Northern municipalities expenditure tends to regress to pre-shock levels, Southern municipalities stick to higher expenditure levels when grants drop. This evidence is coupled with a faster recovery of private income and housing prices in the North. In the second essay, we analyze how the occurrence of earthquakes changes voters' behavior at municipal elections and which channels drive this change. We exploit data from municipal electoral cycles between 1993 and 2015 in Italy and apply an empirical strategy that combines propensity score matching and regression adjustment. We find that the occurrence of destructive earthquakes significantly increases the incumbent chance of being reelected and vote share in municipal elections. We run several placebo tests and robustness checks to corroborate the results. We find that incumbent mayors are rewarded for a high quality response to disaster damages that offsets the partial accountability of incumbent mayors for disaster relief since resources are sent by upper-tier governments and are not entirely converted into expenditure. Moreover, incumbents benefit from the mediatic interest of earthquake occurrence since their visibility on the media grows as compared to the main competing candidates. In the last essay, we analyze the effect of temperatures on monthly mortality rates and hospital admission rates, and investigate whether municipalities with higher social expenditure face lower mortality and hospitalizations. We use data from monthly mortality for the period 2003-2015 and the universe of hospital admissions aggregated by municipality for the period 2001-2015. We use panel data models of mortality and hospitalization rates that control for temperature level or deviation bins, with deviations being relative to municipality-specific average temperatures, income, precipitation, pollution, province-specific time trends, and municipality, month×year and province×month fixed effects. We find surges in mortality rates when extremely hot days occur independently from the adopted temperature measure. Moreover, we find that both extremely hot and cold temperatures, measured as deviations from municipality-specific mean temperatures, cause surges in hospital admission rates of the elderly both for cardiovascular and respiratory diseases, and of children for respiratory diseases. Conversely, when using temperature levels, results of hospital admissions are confounding because temperature levels do not account for local resilience to weather and offsetting behaviors. Moreover, we find evidence of a mitigating effect of social expenditure on health outcomes with the lowest-spending municipalities facing higher surges in mortality and hospital admission rates when hot or cold days occur as compared to municipalities with higher spending levels.