IT’S A measure of how unhinged the debate over the Affordable Care Act has become that a minor change to a single question on a Census Bureau survey became a roiling political controversy.
Sure, the Census Bureau did not do everything right. But, no, that does not mean that the Obama administration is cooking the numbers on the ACA, and, no, it’s not worth a congressional investigation.
On Wednesday, the New York Times reported that the Census Bureau was changing a survey in a way that will reduce the number of people reporting that they are uninsured. Speculation flew about how the change would conveniently make it look like the ACA dropped the uninsurance rate more than it really did. “Another crazy conservative conspiracy theory is proven to not be so crazy after all,” railed Mediaite Editor Noah Rothman. On Thursday the House oversight committee demanded documents from the Census Bureau.
The Census runs its annual Current Population Survey to measure U.S. families’ well-being. Among other things, the bureau tracks how many have health coverage. Previously, the bureau asked respondents if they had insurance at any time over the past year. That formulation of the question, however, seems to have encouraged underreporting of coverage respondents may have had months before. The bureau’s solution was to begin asking more detailed questions about their past insurance status. That way, the bureau would be able to get a more accurate picture of who in the country has had health insurance, when and for how long. The change made sense.
What of the suspicious timing? Bureau Director John H. Thompson insists there was nothing untoward. The bureau had been planning for years to change questions. It tested its new questions in 2010. Besides, the bureau is phasing them in now, as it gathers data on last year, so that it will be able to fairly compare numbers from 2013, before ACA insurance plans came into effect, with the post-ACA years.
We don’t see much smoke here, let alone fire. When it knew the ACA phase-in timeline, the Census Bureau should have moved faster to change its survey so that there would be at least a couple years of pre-ACA numbers to compare against the post-ACA figures. Data analysts like multi-year trends. But there are other measures of the health-insurance rate, including from the bureau, that can give a sense of what may have been missed. Some health-care analysts would like the bureau to find a way to ask both sets of questions for a time, so they can get a clearer sense of how much the change in questioning is altering survey results. The Census’s statisticians should take another look at doing that — and if Republicans are really worried about cooked books, they should consider giving the bureau the extra resources it would surely need.