People Who Live in Glass Houses Shouldn’t Throw Stones

Stephen Herzenberg |

(NOTE: at a future date we plan to address the substance of the op-ed referred to below by Anthony Davies and a co-author in more detail.)

So, this is rich: in an op-ed critiquing our Pennsylvania Promise free tuition (not free college) proposal, Anthony Davies and a co-author snidely say “we use the term loosely” after describing the Keystone Research Center and Pennsylvania Budget and Policy Center as “think tanks.”

Why is this rich? Because Davies is the guy who wrote multiple papers for conservative “think tanks“ with flawed statistics claiming to show that states which have not privatized alcohol distribution (like Pennsylvania) have HIGHER traffic fatalities.

As Mark Price showed, Davies and a co-author manufactured this result by leaving two key variables out of their statistics.

One missing variable was average vehicle miles traveled. It turns out that the farther you drive on average the more likely you are to have an accident and a fatality. Duh. In the few states that still have state-controlled alcohol distribution—again like Pennsylvania—people tend to drive long distances. If you leave average vehicles miles traveled out of the statistics then the variable for “state control” erroneously picks up some of the fatalities associated with driving long distances.

The second variable Davies and his co-author left out was average per capita income. Turns out that states that control alcohol distribution tend to have lower per capita income. Having lower per capita income results in people having older, and less safe, cars on average. So once again what economists call “omitted variable bias” artificially drives up the estimate of fatalities in states that control alcohol distribution.

When you fix the statistics, Davies and his co-author’s original result reverses: privatized alcohol distribution states have more traffic deaths, properly controlling for other important factors. This makes intuitive sense. States with privatized distribution tend to have more places where people can buy alcohol and they also have a profit motive to sell to people who they shouldn’t (people who are under age or have had a few).

Given his track record producing shoddy research that no self-respecting think tank would publish, Davies might not want to throw stones.