The cleverest data collection I’ve seen all year

How do poor people cope with injuries and other unexpected health shocks? One way to find out is to measure household responses to illnesses and injuries. The challenge, of course, is that illness and injury are irregularly bestowed. Poverty and risky behavior all raise the risk of something bad, clouding our ability to understand what life and behavior would have been like if they hadn’t taken ill.

Manoj Mohanan is a Harvard Health PhD student who took a term paper idea and has turned it into a dissertation chapter. He tracked down the victims of bus accidents in Karnataka, India, and compared their lives to people who rode the same bus routes on different and less dangerous days. Very clever, and even more intrepid.

I’ve been in more than one bus accident in India or Africa (including one prompted by copulating cows). While mine were never serious, I wouldn’t be surprised if road accidents are the chief source of injuries in poor countries. So this is no small question.

Even so, I asked Manoj what we learn from such a paper. I really admire the ingenuity. But, like any of the cute causal identification papers that are all the rage, can we generalize these results to injuries or even India more broadly?

Acknowledging that it’s hard to say whether the results are genral, Manoj makes a couple of good points:

what we do learn is that households rely largely on debt, most of which comes from money lenders, to meet health expenditures. This evidence differs from earlier studies that report consumption effects in response to health shocks.


In terms of the big picture, I think this evidence highlights potential welfare gains from providing insurance systems in developing countries.

Looking only at consumption smoothing as a measure of household welfare without examining the costly mechanisms adopted by households to smooth such consumption misses the big picture.

I recommend the paper.

3 thoughts on “The cleverest data collection I’ve seen all year

  1. Well bill, it might technically be a bull and a heifer, if the female had not produced a calf yet.