Thread by @rglenner

1. Massive investment in tubewells in #Bangladesh (incl by World Bank) is considered the poster child of development gone wrong (eg in West Wing) b/c many of the wells contained natural arsenic.

Our new @nberpubs WP shows how wrong this view is. Thread...
nber.org/papers/w25729.…

2. Diarrhea is 2nd largest killer of children <5. High population density and flooding make Bangladesh highly vulnerable. In 1970-90s, 8.6 million wells were dug in Bangladesh and 94% rural population moved to clean water sources. Massive achievement until arsenic was discovered.

3. Naturally occurring arsenic was found in many wells with 15% pop in great danger ie drinking water >50ppb arsenic. Big campaign in 1998/2000 to test & paint red contaminated wells. Very high compliance: noone wanted to drink arsenic. But shutting down wells came at a cost...

4. Within a village which wells had arsenic and which didn't was close to random. This allows us to test impact of shutting wells. Before 1998, trends in child mortality were similar in hh with high or low contamination wells. After, they diverge...

5. Why? When clean water is further away it is usually stored for longer and stored water gets contaminated. Its also possible ppl resorted to dirty surface water. Note the kids in these hh were drinking arsenic for yrs, its when they stop drinking arsenic the death rate jumps.

6. But arsenic is worse for the elder, at least they were better off right? We don't find much evidence of arsenic on the elderly. Death rates not higher prior to 2000 and if anything relative death rates rise when they switch out of arsenic wells (and have less clean water).

7. Note the difference between our findings and previous work on arsenic and elderly is we look at within village variation where arsenic is not correlated with socioeconomic status. Using across village variation in arsenic contamination is more problematic.

8. There is a positive end to the story. People responded by building more deep tubewells which are much less likely to be contaminated with arsenic (though much more expensive). Where deep tubwells have been built, mortality rates have fallen again.

9. Conclusion: that big investment in wells in 70-90s should be a poster child of development working. Arsenic is scary but diarrhea kills more ppl. *Convenience* of clean water matters.

Joint work with Nina Buchmann, Erica Field, Reshmaan Hussam
nber.org/papers/w25729.…