When the political capital of a state is relatively isolated from the rest of the population, then people may care less about state politics. Newspapers won’t cover state politics as much either, so in general accountability will be low. This will then allow state politicians to be more corrupt.
Other negative effects may exist as well. For instance, money may be more important in politics if the capital’s isolated. And in the end, this may all lead to poorer government performance such as a low provision of public goods. These are the hypotheses discussed in this post.
First, they studied correlations between corruption and the isolation of the capital. The main measure of state-level corruption was federal convictions for corruption-related crimes per capita, averaged across 1976-2002. As for capital isolation, the average (log) distance from the capital is calculated for each county within a state. This measure is then normalized in a fashion that the county farthest from the capital receives a score of 1, the capital itself receives a score of 0. The distances are weighted by population to get a good composite measure of the capital’s isolation.
The correlation between these two variables is shown in the figure below.
In general, even with the addition of various control variables, we can see that isolation robustly and significantly affects corruption. Namely, a one standard deviation increase in isolation (equivalent to what Carson City, NV experienced between 1920 and 2000) can increase corruption by 0.75 standard deviation.
These results are robust to alternative measures of corruption and isolation. Placebo regressions also show that it is not the largest city’s isolation that’s driving the results, as that variable neither is significant nor does it decrease capital isolation’s significance. It is also shown that the isolation of the capital has no effect on non-corruption-related crime (such as drug cases); indicating that isolation has an effect on corruption per se, and not on crime in general.
Secondly, the authors address the issue of causality. For instance, there is a risk that states that were already corrupt in the past chose their capital cities to be in isolated places. If this were the case, then causality could (at least partially) run from corruption to isolation weakening the results presented above.
The instrumental variable used to address this issue is the isolation of the state’s centroid. This measure is correlated with the capital’s isolation because capitals tend to be close to the geographical center of the state, but it is obviously not correlated with corruption. The results of the IV analysis confirm that causality runs from isolation to corruption. The effects are mostly significant, and their magnitudes stay largely similar to what we saw in the previous analysis.
Finally, the authors look at how accountability, the role of money and public good provision are affected by isolation.
The find that newspapers with a higher reader concentration around the capital cover state politics significantly more. This means that if the population is not concentrated around the capital, newspapers in the state will be less likely to cover state politics. This can lead to a less informed and less engaged citizenry (in state politics). Indeed, individuals are found to know less and to a certain extent care less about state politics, the farther they live from the capital. On the other hand, general interest in politics at large is not affected by distance from the capital.
All these phenomena can lead to lower accountability of state officials. This can result in money having a larger influence in state politics. The authors indeed find that a one standard deviation increase in isolation can increase campaign contributions by 30%. Moreover, people/firms who are closer to the capital are more likely to contribute.
Low accountability and higher corruption can clearly lead to worse government performance. This is also evidenced in the data, as the share of public good expenditures (education, health, public welfare and hospitals) is significantly lower in states with isolated capitals. Other expenditures on the other hand are positively affected, suggesting that isolated capitals change the structure of state expenditures to the detriment of public good provision.
All in all, these are some quite interesting and novel findings. And I would not expect this relationship between corruption and isolation to change in the future, at least not without any specific intervention. While online media may, figuratively speaking, reduce distances, it will not necessarily give more coverage to state politics if readers are not interested. However, increasing citizens’ interest in state politics – a challenging but not impossible feat – could potentially be quite beneficial in some states.