The politics behind business performance in Africa

…we use firm-level data for 89 countries to examine formal firm performance. Without controls, manufacturing African firms do not perform much worse than firms in other regions. But they do have structural problems, exhibiting much lower export intensity and investment rates.

Once we control for geography and the political and business environment, formal African firms robustly lead in sales growth, total factor productivity levels and productivity growth. Africa’s conditional advantage is higher in low-tech than in high-tech manufacturing, and exists in manufacturing but not in services.

While geography, infrastructure, and access to finance play an important role in explaining Africa’s disadvantage in firm performance, the key factor is party monopoly. The longer a single political party remains in power, the lower are firm productivity levels, growth rates, and sales growth for manufacturing.

In contrast, the business environment and firm characteristics (except for foreign investment) do not matter as much. We also find evidence that the effects of the political and business environment are heterogeneous across sectors and firms of various levels of technology.

A new NBER paper from Harrison, Lin and Xu. Sadly I do not see an ungated copy. Anyone who sees one please put in comments.

I’ve only been able to skim the paper, but the question, approach and the answer all strike me as incredibly important for lots of reasons. One is that it gets us started on addressing some of the puzzles of development raised by Banerjee and Duflo in their under-read Growth Through the Lens of Development Economics or this Pande and Udry piece on institutions and development (both of which I mentioned last week). In particular, it starts to get far more specific about government and institutional failure than the existing research (which is pretty crude when it comes to political failures).

I don’t think we should take this as evidence that party monopoly is actually what causes the problem (I think the authors would agree with this), but it does start to get us closer to what political failures need more investigation.