Thursday, September 3, 2015


Bruce B. de Mesquita, TLPS

The ease with which graft can be hidden in construction projects makes this industrial activity unusually attractive as a way for leaders to indulge in cronyism and nepotism. Every society requires construction, of course, and not all of it involves the misappropriation of funds. But there is sufficient anecdotal evidence to suggest that construction involves an unusual degree of corrupt practices.

We test the relationship between our indicator Construction and the effects of W and W/S using our standard models. The selectorate model performs well when it comes to assessing Construction, though the substantive impact is more modest than for corruption. W by itself is significantly associated with a dampening of construction projects and remains significant after we control for other factors. WS:Income, in contrast, turns out to contribute significantly to increasing construction.


Right. I was able to replicate this finding as well; see the updated notebook attached to this post. Once the general richness of society is accounted for, there is a negative correlation between construction and the winning coalition size, W, meaning better governance means less construction that involve theft. Rich societies naturally build more, or they build in proportion to their income, and if they are governed well no more than necessary. The "more" of this extra construction can involve shady dealings and that part is caught by the regression.

We also control for what it is traditionally known as democracy mirroring BBM and his book. In a separate regression we take out effects of W,S from inside this traditional "democracy" variable, so what remains (the residuals) are taken to be all other parts (benefits?) of democracy that is not related to W or S. This residual has no effect on the regression, meaning traditional concept of democracy is irrelevant.

This finding is interesting at so many levels..  As we all know, TR is not a "large W" environment - after the latest election, neither the 60% of the votes cast for the opponents of AK, or AK itself in conjunction with either of the 3 parties were able to form a coalition. At the same time there have been allegations around AK over inappropriate dealings related to the construction sector. Coincidence? 

Q&A - 21/5

Question How do you empirically prove interest rates do not cause increase or decrease in GDP growth? There is a test for that Data ,...