India is electing more and more politicians that are, or have been, accused of criminal activity. This column asks whether accused – and potentially corrupt – politicians influence economic activity by focusing on data from road construction, which is often tied to corruption. The evidence tentatively suggests that caste-based and patronage politics favours lower quality politicians, who in turn deliver benefits to specific groups. This generally leads to lower aggregate growth.
They don’t have GDP data so look at intensity of night lights. They show how electing an accused politician leads to lower activity:
Using information on the charges filed against candidates, we estimate the causal effect of electing criminally accused politicians to the State Assembly on the subsequent economic activity in their constituency. In particular, we focus on elections in 20 Indian states during the 2004 to 2008 period. Since economic data are not systematically available for constituencies, we rely on satellite data on the intensity of night-lights. These data have been increasingly used to proxy for economic growth, as studies find a strong relationship between GDP and night-light intensity at the sub-national level (Bleakley and Lin 2012, Henderson et al. 2012, Hodler and Rashky 2014, Storeygard 2014). Figures 1 (a) and (b) show the strong relation between growth and night lights. Figure 1 (a) shows India in 1992 at the time of the economic reforms, while Figure 1 (b) shows India after the resultant high levels of growth – not only has the intensity of night lights increased, but new areas are lit due to the spread of electricity and economic gains.
Figure 1 (a). Night-time light output in India in 1992
Figure 1 (b). Night-time light output in India in 2009
Nice bit of graphs
So do we see accused politicians impacting economic activity?
We answer this question by comparing constituencies that elect criminally accused politicians with those that elect non-accused politicians in similarly close elections. We find that the election of an accused politician leads, on average, to roughly a 22 percentage point lower yearly growth in the intensity of night lights. Based on conversions between GDP and night lights, this is roughly 5.61% to 5.86% GDP growth per year (as compared to the 6% otherwise).
Overall, these results highlight the high aggregate economic costs of electing lower quality politicians (i.e. criminally accused) and point to likely significant individual costs in foregone access to public services.
…..While electing accused politicians leads to a 22 percentage point lower yearly growth of night lights, this effect is not instantaneous. When we examine the accumulation of these costs, we find that the effects only appear in the later years of the politician’s term. There is no apparent effect in the initial years. We believe that this is explained by the need for politicians to collaborate with local bureaucrats to engage in corrupt activity (Iyer and Mani 2012). After elections, bureaucrats frequently change positions so it takes a certain amount of time for corruption politicians and bureaucrats to identify each other. Additionally, the effects of neglected public infrastructure, such as roads, may take some time to slow down economic activity.
While it is difficult to pinpoint the exact causes, the evidence suggests that the local context matters. When we compare the effects of electing accused politicians, we find that these costs increase in states with high corruption levels and lower levels of development, and those that plausibly have weaker institutions. This variation across states suggests that the local context plays an important role in limiting the detrimental effect of electing accused politicians. More broadly, since high corruption levels, low levels of development, and weak institutions frequently occur in the same places, it is difficult to separate their effects.