And the vaults are stuffed with silver, / That the farmer sweated for.

January 22, 2011 9:33 PM

"People in these communities may not have experienced the Great Depression first-hand, but our research suggests that the cultural consequences of suspensions, especially as they relate to trust and demoralization, have been passed along for generations." Counties with higher bank suspension rates in 1930 experienced elevated suicide rates 70 years later.
posted by orthogonality (5 comments total)

Ah, but when governments start to fail from propping up bad business models on the backs of the general population, what will that do to suicide rates?
posted by Malor at 9:41 PM on January 22, 2011

It would have been cool if this had been posted by a user named "correlation vs causality", but "orthogonality" works too.
posted by intermod at 9:46 PM on January 22, 2011 [2 favorites]

^
posted by telstar at 11:13 PM on January 22, 2011

The interesting thing this research is saying is that it was the financial panic (bank suspensions) that caused long-term harm, rather than (just) the underlying economic distress. The researchers have tried to isolate the effects of the suspensions by "controlling for" measures of economic distress, like unemployment. I'm skeptical, though. The effect is very small. I would suspect bank suspensions track unobserved indicators of economic distress that their controls are not picking up. Prefer to see a clever instrumental variable model.
posted by grobstein at 11:26 PM on January 22, 2011 [2 favorites]

To isolate the link between the suspensions and suicides, Baller and his UI team controlled for other factors that may have influenced that relationship, including the value of agricultural goods, radio ownership, and unemployment in 1930, loss of male population during World War II, and contemporary measures of unemployment, social integration, inequality, poverty, divorce, religiosity, race and age.

I'm curious how they did this, wouldn't this only be possible if those confounding variables were uncorrelated with the independent variable in the underlying population?
posted by atrazine at 11:47 PM on January 22, 2011