Gallup has some new health insurance poll results out today. Perhaps they're just in time for the President to use for his presser tonight. One item from Gallup's report really surprised me. Apparently, the percentage of uninsured has not increased significantly since January 2008. Unemployment has. What gives?

Here's that chart from Gallup:

uninsured timeline.gif


As you can see, in January 2008, 14.8% of those polled did not have health insurance. As of June 2009, that percentage increased to 16.0% -- an increase of a mere 1.2%.

Over that same period of time, unemployment has increased as follows:

unemployment insurance.PNG


Here, in January 2008, unemployment was 4.9%. As of June 2009, that percentage increased to 9.5% -- an increase of a whopping 4.6%. Keep in mind, that increase would be even greater if discouraged workers were included.

Clearly there's very little correlation here -- and you'd think there would be. After all, many people have employer-based health insurance -- approximately 59% from the Census Bureau's 2007 estimate. So a little multiplication shows us that, of that 4.6% of the population who lost their jobs, assuming those laid off had employee-based health insurance in numbers representative of the rest of the population, 2.7% should have lost their health insurance. In realty, it's less than half that. What are some potential explanations?

One possibility is the stimulus bill's COBRA provision. That allowed anyone laid off to only pay one-third of their employer's health insurance costs to continue on their old plan -- the stimulus covered the other two-thirds. This provision went into effect for anyone laid off after mid-February 2009. So instead, you can do this analysis in February, before the COBRA change took effect. You still have a disconnect. Then, the uninsured was 16.1%, while unemployment was 8.1%. Doing the math for February instead, you would still expect 1.8% more to be uninsured -- nearly 40% more than the actual increase of 1.3%.

One obvious possibility is that some people who lost their employer-based insurance are just paying for private insurance. This effect might not be so widespread, however, since anyone unemployed has even less money to be spending on health insurance, and might feel it's a cost that can be put off until employed again. Still, this might explain some of the effect.

Another possibility is overlapping coverage. For example, if a husband was laid-off, maybe he just got on his wife's employer-based plan. This likely explains it for some cases, but not everyone is married.

Can those two effects really make up for that 40%? I can't think of any other possibilities, so I suppose they must. My only other thought is that those workers who did not have employer-based health insurance disproportionately suffered lay-offs. I'm not convinced, however, given employers with health care plans would save even more money by laying people off than those without.

If you can think of other reasons, feel free to make some suggestions as comments. But overall, this outcome is pretty surprising. As unemployment increases, the number of uninsured does not increase by nearly as much.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.