Being a Good Samaritan or just a politician? Empirical evidence of disaster assistance Jeroen Klomp Netherlands Defence Academy & Wageningen University and Research The Netherlands
Introduction Since 1970 there have been about 10,000 natural disasters worldwide, affecting more than 7 billion people, and causing over the $2 trillion in estimated damages (EM-DAT, 2017). Although natural disasters happen beyond government control, the responses to these events are in the hands of elected officials. According to the theory of retrospective voting, the electorate will hold politicians responsible for the humanitarian and economic losses. A large number of empirical studies find clear support for the idea that voters punish the government when a catastrophe strikes. However, a number of subsequent studies have also found that voters may not react only to the disaster itself, but also to the government's actions taken in response to the catastrophe. Natural disasters might therefore create a window of opportunity that can be used for political motives as political parties like to gratify the needs of their particular constituencies.
Introduction The evidence in the empirical studies up till now are foremost based on local or regional data. These studies are unable to explain whether, and if so to which extent, the existence of political ideological cycles in public disaster support are related to differences in the political system in place. Different dimensions of political ideology Partisan cycles: left vs. right wing political parties; Special interest politics. The aim of this study is to explain cross-country differences in public disaster support caused by political ideology by relating them to the country-specific political context. i.e., upcoming elections, political system, electoral system.
Theoretical background Partisan cycle Left wing government parties provide more support: Low income voters need more assistance as they are often uninsured and live in disaster prone areas; More in favour of redistribution to reduce inequality => Low income voters vote more often on left wing parties (health, education, sanitation). Special interest politics Based on the disaster vulnerability of their main constituency; Disaster prone industries or living in climate affected regions. Variation among political factors Partisan cycles - left vs. right wing Political ideology Special interest politics Elections upcoming (within one year) + + Electoral system (majority vs. proportional) + + Political system (presidential vs parliamentary) + +
Data and methodology The economic consequences of natural disasters are the product of hazard (exogenous), exposure (endogenous) and vulnerability (endogenous). Focus on the hazard part -> reducing endogeneity of the results. Geophysical -> threshold on the Richter scale (earthquakes) and VIE index (volcano eruptions). Meteorological -> threshold based on the wind speed. Hydrological -> severity based on the product of the interval probability in years and the duration. Climatic -> at least three subsequent months have rainfall below fifty percent of the long-run average monthly mean, or if at least five months within a year have rainfall below fifty percent of the long-run monthly mean.. Disaster count variable: sum of events in a country-year divided by land size.
Data and methodology Geophysical Meteorological Hydrological Climatic Frequency in the dataset 10% 26% 41% 23% Duration Minutes Hours to days Weeks to months Weeks to months Geographical scope Local Local to regional Regional Regional Predictability - +/- + + Preparation possibilities - - + +/- Population affected - +/- +/- + Physical damage ++ +/++ + -
Data and methodology Political ideology Partisan cycle Variable running from: (1) full right wing cabinet, (3) centre cabinet to (5) full left wing cabinet. Special interest politics dummy taking the value one when At least one religious political party in cabinet; At least one rural political party in cabinet; At least one nationalistic political party in cabinet.
Data and methodology Lack of cross country data on disaster assistance. An indirect approach Panel model ln spending = α + γ ln spending + β x + η disaster + µ pol + it i it n k it j it it ϕ( disaster pol ) + δ + ε About 85 democratic countries between 1985-2013. Control variables: real GDP per capita; growth rate of real GDP per capita, inflation, age-dependency ratio; coalition dummy; checks and balances; foreign aid assistance; trade openness; monetary union dummy. Estimated using system-gmm. it it t it ln spending disaster = η+ ϕpol
Results Standardized coefficients for a median sized country Table 1: Natural disaster support and partisan cycles Dependent variable: ln Central government spending per capita All countries OECD non-oecd (1) (2) (3) (4) Natural disasters 0.010 * 0.010 * 0.006 * 0.015 ** (0.005) (0.005) (0.004) (0.007) Partisan 0.018 0.018 0.019 0.021 (0.014) (0.023) (0.035) (0.034) Religious political party 0.024 0.020 0.028 (0.026) (0.022) (0.045) Nationalistic political party 0.037 0.035 0.041 (0.028) (0.033) (0.034) Rural political party 0.009 0.010 0.011 (0.007) (0.009) (0.016) Natural disasters partisan 0.006 ** 0.005 * 0.003 ** 0.006 ** (0.002) (0.003) (0.001) (0.002) Natural disasters Religious political party 0.007 0.006 0.006 (0.012) (0.006) (0.009) Natural disasters Nationalistic political party 0.009 ** 0.006 0.010 * (0.004) (0.008) (0.006) Natural disasters Rural political party 0.003 * 0.002 * 0.004 * (0.002) (0.001) (0.002) Controls YES YES YES YES Observations 1888 1534 692 842 Countries 84 75 33 42 Sargan test (p-value) 0.457 0.674 0.699 0.692 Arellano-Bond AR(2) (p-value) 0.609 0.833 0.709 0.687 Note: **/* Indicating significance levels of respectively 5 and 10 percent. Bootstrapped standard errors are shown between brackets.
Results
Result 1: Results Full left-wing governments spend about 2.5 percent more on public disaster support compared to a complete right-wing cabinet. Cabinets consisting at least of one nationalistic (+0.9 percent) or rural (+0.3 percent) political party provide more disaster support than other coalitions. The impact of partisan cycles declines when a country is more often hit within one year.
Results Result 2: Specific disaster effects Partisan Rural Religous Nationalistic Hydrological s+ s+ ns ns Meteorological s+ ns ns ns Geophysical s- ns ns s+ Climatic ns s+ ns ns
Conditional effect Results m ln spending = α + γ ln spending + β x + λ z + η disaster + µ pol it i it j k it j m it it m m m + κ ( z disaster ) + φ ( z pol ) + ϕ ( disaster pol ) + θ ( z disaster pol ) + δ + ε it it it it it it it it it t it it ln spending disaster m m = η+ ϕpol + κ z + θ ( z pol ) it it it
Results Table 4: Political mechanisms Election year Dependent variable: ln Central government spending per capita No election Presidential Parliamentary Proportional Majority year (1) (2) (3) (4) (5) (6) Natural disasters 0.020 ** 0.006 * 0.008 ** 0.006 * 0.007 * 0.006 * (0.006) (0.003) (0.003) (0.003) (0.004) (0.003) Partisan 0.015 0.012 0.011 * 0.011 0.012 0.010 (0.010) (0.012) (0.006) (0.009) (0.024) (0.019) Religious political party 0.019 0.028 0.026 0.020 0.027 0.021 (0.018) (0.042) (0.021) (0.033) (0.020) (0.022) Nationalistic political party 0.041 0.036 0.033 0.031 0.035 0.034 (0.038) (0.047) (0.044) (0.028) (0.028) (0.045) Rural political party 0.010 0.010 0.010 0.007 0.008 0.011 (0.008) (0.011) (0.007) (0.012) (0.006) (0.018) Natural disasters Partisan 0.021 ** 0.004 * 0.021 ** 0.006 * 0.004 * 0.015 ** (0.007) (0.002) (0.010) (0.003) (0.002) (0.005) Natural disasters Religious political party 0.008 0.007 0.008 0.006 0.006 0.007 (0.005) (0.005) (0.005) (0.006) (0.008) (0.012) Natural disasters Nationalistic political party 0.025 ** 0.007 * 0.018 ** 0.009 * 0.006 * 0.014 ** (0.008) (0.004) (0.004) (0.005) (0.003) (0.005) Natural disasters Rural political party 0.004 ** 0.003 * 0.004 ** 0.003 * 0.003 0.003 ** (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) Controls YES YES YES Sargan test (p-value) 0.425 0.355 0.643 Arellano-Bond AR(2) (p-value) 0.519 0.438 0.737 Note: **/* Indicating significance levels of respectively 5 and 10 percent. Bootstrapped standard errors are shown between brackets.
Result 3: Results Governments provide more political ideological disaster motivated support when elections are upcoming. Governments provide more political ideological disaster motivated support in presidential and majoritarian elections systems.
This study explores: Conclusions If there are political ideology cycles in public support after a natural disaster. And if yes, if the existence relies on certain elements of the political system in place. The main findings indicate that indeed public support after a natural disaster that is subject to political cycles. However the impact is to a certain extent conditional on the political system, electoral rules, and when elections are upcoming.
Thank you for your attention Questions?