August 27, 2014
It’s always nice when a prominent economist and the NYT pick up on a line of work that we started at CEPR. That is why we are all happy to see David Leonhardt’s piece on a new paper by Alan Krueger, the former head of President Obama’s Council of Economic Advisers.
The gist of the piece is that Krueger has discovered that many people do not respond to the Current Population Survey (CPS), the main survey used to measure the unemployment rate. Krueger discovered that the unemployment rates are higher for people the first month that they are in the survey than in later months. (People are in the survey for four months, then out for eight months and then back for four months.) The implication is that people who are not responding may be more likely to be unemployed than people who are responding.
This fits well with analysis done by John Schmitt and me nine years ago. That work noted a sharp gap between the employment rates reported in the 2000 Census and the employment rates reported in the CPS for the overlapping months, with the CPS rates being much higher. (The Census has a response rate close to 99 percent, whereas the coverage rate for the CPS is under 90 percent overall. It is under 70 percent for young black men.) The analysis focused on employment rates because employment is much more well-defined than unemployment.
The analysis also noted that the gap was largest for the groups with the lowest coverage rates. In particular the gap was largest for young black men, with the CPS showing an employment rate that was 8.0 percentage points higher than the Census data for the same month. Our conclusion was that the people who respond to the survey are more likely to be employed than the people who don’t respond. It’s good to see that Krueger appears to have concurred in this finding nine years later.
Note: Link and president corrected.