[roc]

If the correlation between income disparity and crime rates remains despite a net drop in crime, how is that a refutation of inequality of disparity as a component to crime?

Wouldn't a larger drop in crime in more-equal countries than less-equal countries only re-affirm that though there are clearly other factors involved, inequality is still a component?

I don't actually know the numbers but I'd think it would be pretty easy to look and see whether crime rates are dropping uniformly, or if less-equal countries are seeing a lower drop than more-equal ones.

[ajkessler]

Sure. But since income disparity has soared in the last 40 years and crime rates have plummeted, even if they have plummeted everywhere, this indicates that inequality, if it has any effect at all, is likely pretty minor.

I think this would be near impossible to test for though. So many factors affect crime rates across difference societies, and different societies experience wildly different types of income disparity. Intuitively, I imagine places with smaller net income disparity but larger real income disparity (ie a city where everyone is relatively poor but some are destitute) have much more crime than places with larger net income disparity but smaller real income disparity (ie a city where everyone is relatively wealthy, but a few are unbelievably rich).

[roc]

> "since income disparity has soared in the last 40 years and crime rates have plummeted, even if they have plummeted everywhere, this indicates that inequality, if it has any effect at all, is likely pretty minor."

The big problem with a simple comparison of the plot of inequality and the plot of crime, is that inequality and crime aren't evenly spread across demographics. The young are disproportionately on the losing end of inequality and the perpetrating end of crime. (Regardless of how poor they are and how rich their neighbors, the elderly simply can't flip a car anymore.)

If you had, say, a bulge in the poor demographic around 1970 and expected crime to increase with a rise in inequality, you would expect to see the spike in crime from 1970 to 1990 and you would expect to see it subside as that bulge grew older and transitioned out of the age groups that perpetrate crime.

So, again, we cannot trivially rule out inequality as minor just because a simple plot of crime rates and a simple plot of inequality do not suggest a straightforward correlation.

[scott_w]

It's not impossible. Levitt discusses issues like this in Freakonomics.

Just because all crime falls over a period of time doesn't discount the effect. It doesn't even mean the effect of income disparity is minor. If the rate of crime falling in countries with a lowering income disparity is vastly higher than the rate of crime falling in countries with high income disparity, then that suggests a link that can be followed up on statistically.