Evidence describing the relationship between trade and food security one hundred years ago doesn’t translate directly into 21st-century policy recommendations, but this is neat stuff from Dave Donaldson and Robin Burgess:

Our district panel regression results suggest that the arrival of railroads in Indian districts dramatically constrained the ability of rainfall shocks to cause famines in colonial India. On average, before the arrival of railroads, local rainfall shortages led to a significant rise in our index of famine intensity. But after a district gained railroad access the effect of local rainfall shortages on famine intensity was significantly muted.

They’re still working on it:

While these preliminary findings are consistent with railroad expansion mitigating famine intensity by facilitating trade in food the same infrastructure investments may also have intervened in the weather-to- death relationship by facilitating movements of people, capital and famine relief. In future work we aim to gain a fuller understanding of how trade openness can mitigate or exacerbate the impact of weather shocks by building up a complete series of mortality statistics for the period and by analysing data on passenger flows, trade flows, trade imbalances, output, prices and famine relief.