The Dreaded Omitted Variable Bias Red Card

Posted in:

Most people would expect that states like Pennsylvania that more tightly control alcohol sales would have fewer of the social problems associated with excessive drinking, including alcohol-related motor vehicle fatalities.

So you can imagine our surprise when analysis by economists John Pulito and Antony Davies reached the very opposite conclusion. Their work was published by self-avowed “free market” think tanks, including Pennsylvania’s Commonwealth Foundation and the Mercatus Center at George Mason University.

The Pulito-Davies findings are at odds with those of a panel of public health experts who recently studied the effects of privatization of retail alcohol distribution. Based on a systematic review of the available research, the panel found that privatization contributes to increases in alcohol consumption, creating a greater risk of alcohol abuse and its associated social costs.

Who made these findings, you ask? The Task Force on Community Preventive Services, a panel of 21 national public health experts appointed by the director of the Centers for Disease Control and Prevention (CDC). They published their findings in the April edition of the peer-reviewed American Journal of Preventive Medicine.

So to help us make sense of Pulito and Davies' puzzling findings, we asked a rising senior at Dickinson College named Jue Wang to replicate their work last summer. Through Jue's impressive work, we discovered that Pulito and Davies’ findings were plagued by a case of omitted variable bias.

Jue demonstrated that Pulito and Davies’ results were reversed once you account for two variables excluded from their original analysis but key to understanding the differences in alcohol-related motor vehicle fatality rates among the states — average vehicle miles traveled and average per capita income.

Include these variables, and states that more heavily control the sale and distribution of alcohol do have lower alcohol-related fatality rates among adults than either states that do not regulate or only lightly regulate alcohol sales.

Including vehicle miles traveled and per capita income is important, since control states tend to be ones in which people drive further and have lower incomes. Driving more increases the likelihood of fatal traffic accidents, while lower incomes mean people are less likely to be driving in newer cars with strong safety features.

We laid out all the findings in a Keystone Research Center policy brief this week. Check it out to get the full story.


0 comments posted

Post new comment

Comment Policy:

Thank you for joining the conversation. Comments are limited to 1,500 characters and are subject to approval and moderation. We reserve the right to remove comments that:

  • are injurious, defamatory, profane, off-topic or inappropriate;
  • contain personal attacks or racist, sexist, homophobic, or other slurs;
  • solicit and/or advertise for personal blogs and websites or to sell products or services;
  • may infringe the copyright or intellectual property rights of others or other applicable laws or regulations; or
  • are otherwise inconsistent with the goals of this blog.

Posted comments do not necessarily represent the views of the Keystone Research Center or Pennsylvania Budget and Policy Center and do not constitute official endorsement by either organization. Please note that comments will be approved during the Keystone Research Center's business hours.

If you have questions, please contact [email protected]

The content of this field is kept private and will not be shown publicly.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <p> <img>
  • Lines and paragraphs break automatically.