One of the frequent rationales for efforts to strengthen governance is that this will lead to stronger economic growth (e.g. entrepreneurs are more certain about property rights, services, taxes and feel they are less likely to be exploited). Another debate is whether local governance or national governance matters most.
In a new paper, Neil McCulloch and Edmund Malesky use a unique data set from Indonesia to test these hypotheses. They link district level growth rates to firm perceptions of governance in those districts. Over 12,000 firms were surveyed in 243 of the country's 500 districts (representative of 15 of the country's 33 provinces). They mainly rely on some careful econometric analysis to explore the relationships between governance and growth.
The results are surprising:
* They do find a positive, statistically significant (and economically meaningful) relationship between district growth and local governance, but only after correcting for a whole host of data quality issues. This leads them to think that the relationship is not that strong and to explore other avenues of interpretation
* They find some evidence that structural features of districts (e.g. are they urban larger, more populous, conflict over natural resource endowments etc.) may be positively associated with growth but negatively with governance (e.g. corruption is easier). It is not clear to me from a quick read why their panel regressions cannot control for these relatively fixed features, but in any case they speculate that these are muddling up the relationship between growth and governance.
* Finally they speculate that governance at the local level may be less important than governance at the national level. The features that associate most strongly with growth are things like the presence of electricity and these have little to do with local governance, being determined nationally.
I would have liked to have seen more discussion on the influence of governance on different sources of growth (e.g. by sector, by labour intensity) and different qualities of growth (e.g. poverty reducing etc.), but perhaps this is for a later paper.
An interesting paper using an interesting data set which I am sure will generate many further important results.