Friday, September 21, 2012

Measuring the Entitled: Thoughts on that 47 Percent Figure

One of the things that I teach in my econometrics class is that the way in which economic variables are measured is important to the conclusions you are able to draw from data.  A recent example is Mitt Romney's assertion that 47 percent of American households do not pay federal income tax in the past year, and therefore, the 47-percent will vote for Obama -- the entitlement candidate -- no matter what (my paraphrase of Romney's recent gaffe).

What's wrong with this assertion?  To my knowledge, 47 percent is accurate, but it doesn't mean what Romney wanted it to mean.  That's because the 47 percent of people who pay no federal income tax includes students and elderly, who at other points in the lives may have paid significant federal income taxes.  There are also people who had a bad year this year, but will do just fine next year and did just fine in the past.  So, if your goal is to measure "fraction of entitled/dependent on government," 47 percent likely overstates that true fraction ... by a lot.  This is a point that the media, and commentators have made quite forcefully (Krugman, Thoma, Jon Stewart, Planet Money).

Is all lost with this measure of "entitlement"?  Is it useless?  Not exactly.  If a statistic that you know overstates what you're hoping to measure is consistently reported over time, we can examine how the statistic evolves over time, and learn something about how entitlements have changed.  This is informative as long as the measurement problem was as approximately as bad in the past as it is in the present.  That's why I found this passage  from a recent Market Watch article interesting:
In 1979 only the lowest fifth of households paid no federal individual income tax, according to data published in July by the nonpartisan Congressional Budget Office. By 2009, the lowest 40% not only made no contribution to income tax revenues, they actually received individual income tax transfers. They were net recipients of federal payments because of programs such as the earned income tax credit, the child tax credit, and Making Work Pay credit. 
To be fair, there are still problems with using these changes over time by themselves.  Immediately, two come to mind:
  1. There are composition changes in the population from 1979 to 2009 (more baby boomer elderly) that could help inflate even this "changes" figure.  
  2. So called "Entitlement Spending" is supposed to address problems with the economy, and there are currently more problems with our economy (one figure: in 1979, unemployment was ~6% while in July 2012 it was 8.3%).
Other evidence, presented elsewhere, is needed to address these concerns, but it certainly seems to be the case that the safety net has gotten more generous.  Rather than flame out over whether a candidate is using an inflated measure to make his point, maybe a more appropriate response is to seriously debate the issue to get down to what it really means.  Of the linked posts above, I think Planet Money's post is the best at objectively breaking down the 47%, but it still doesn't answer the tough question.  Is the safety net (as measured this way) a problem, a response to a problem, or both?


Pre-post edit: On discussing this post, my friend Ryan suggested this article on the larger issues at play.  I've only taken a quick read, but it strikes me as relevant and (potentially) the right way to start thinking about the issue of entitlements.

Another Edit (9/23/12):  Steve Landsburg has some interesting thoughts on the issue.

No comments:

Post a Comment

Please feel free to share your ideas about this post in the open forum. Be mindful that comments in this blog are moderated. Please keep your comments respectful and on point.