WASHINGTON — The White House said Tuesday that changes to the U.S. Census will make it “easier” to see the before-and-after effects of the Affordable Care Act.
The New York Times reported earlier in the day that the Census will make changes to the annual Current Population Survey, designed to measure the number of Americans who go without health insurance in a given year. Questions have been tweaked in the survey to correct what administration officials say were long-standing flaws in the survey. The net result of the changed questionnaire could reduce the official number of uninsured Americans listed by the Census.
Administration officials said Tuesday the changes will be implemented for the year of 2013, before the exchanges launched.
“The main point here is that these changes will actually make it easier to measure the impact of the ACA, since we will have an improved baseline for 2013, which we can use to see changes from 2013 to 2014,” an administration official said Tuesday.
Many analysts said Tuesday that the changes are so fundamental to the survey that comparing data before the change to after is difficult.
Vox’s Sarah Kliff reported a top administration official told the site, “What’s being missed here is that the Obama administration will use the new survey questions to collect data for 2013, the year prior to Obamacare’s health insurance expansion.”
The Times reported that when both questionnaires were used last year, the number of Americans without insurance came in around 2% lower under the new questions.
Observers have long described flaws in the survey’s methodology. Administration officials pointed to a 2003 Congressional Budget Office report criticized the questionnaire even after it was tweaked in 2000 to improve data.
“First, it has a relatively long reference period for measuring insurance status. That long reference period may partly explain how the CPS estimate overstates the number of people who are uninsured all year,” read the CBO report. “Second, although the CPS estimate appears to closely approximate the number of people who are uninsured at a point in time, the survey provides no information on what fraction of the year people do have coverage.”
The new changes to the questionnaire were announced in September 2013 and submitted for public comment.
“Health insurance questions have measurement error due to both the reference period and timing of data collection,” the Census Bureau wrote in its announcement of the changes on the Federal Register. “Qualitative research has shown that some respondents do not focus on the calendar year reference period, but rather report on their current insurance status.”
The new questions, the Census wrote, “integrate questions on both current and past calendar year status” when it comes to insurance, ask questions designed to track coverage in exchanges created by the Affordable Care Act and ask recipients whether or not their employer provides health insurance.
The Census and the Department of Commerce, which oversees the Census, did not respond to a request for comment on why the changes were being made to the insurance survey now, rather than years ago when concerns were first raised.
On Twitter, Kaiser Family Foundation vice president Larry Leavitt praised the new survey and said it won’t stand in the way of determining whether Obamacare is working when it comes to the uninsured.
“The new census questions on the uninsured will be able to measure the change from 2013 to 2014,” he tweeted. “We just have to be patient.”
The administration says the survey changes have been in the works for years before Obamacare was even under discussion.
“Census’s annual report on insurance coverage has had language about the shortcomings of the CPS for measuring insurance coverage (shortcomings that these revisions aim to address) every year back to President George W. Bush’s first term,” an administration official said in an email.