Friday, May 3, 2013

BLS Hiring and Firing Data

In discussing welfare with conservative friends, the discussion inevitably turns to all the people they know who are "able-bodied" but who simply refuse to work so they can collect welfare, in addition to all of the people they know who scam the welfare system. I'm not sure where these folks hang out that they personally know all of these people scamming the welfare system, apparently bleeding taxpayers dry, despite the fact that welfare is only 3% of the federal budget.

When I point out the BLS high rates of unemployment, especially for those under 30 and race minorties, conservatives typically critique the unemployment measure as "how many people aren't working because they would rather collect unemployment benefits." While federal data doesn't support this assumption, it is true that the "unemployment rate" isn't a masure of how many people are "unemployed who want work," which is important to my argument, since I claim the high rates of poverty and welfare use are because there are no jobs available, not because Americans are lazy and like to leech of our neighbors. This is the typical sociological frame--that poverty is due to structural forces that prevent people from access to skills, or access to jobs.

Based on a friend's question, "is there a 'jobs opportunity index,' an alternative to the unemployment index, but which shows actual availability of jobs?," I did some digging, but found little that adequately addressed that important question. In 2010, a pair of economists won the Nobel Prize for developing the Matching Theory Problem, that there is a severe structural problem with matching job seekers with job opportunities. There have been several nice studies that look at the lack of available jobs, and the subsequent rise in poverty; conversely, when job opportunities rise, then poverty and welfare use decreases. But all of these studies are micro-level studies of communities, based on researchers spending months canvassing from street-to-street to find employers, or utilizing online databases like Craigslist or There is no national database where all employers input available jobs, so therefore, there is no "available jobs index" to measure national, regional, or local employment opportunity rates.

The closest I could find was the BLS JOLTS database, the Job Openings and Labor Turnover Survey, which began in late 2000. The title sounds promising, but it actually doesn't measure "job openings." It takes a random sampling of 16,000 employers a month, primarily firms (larger, stable businesses), and asks how many people they have hired, and how many they have let go, or separations. The assumption is that separations minus hiring should represent jobs available. But later researchers found that when firms layoff workers and simply make businesses smaller, then the separations-hiring number is not job openings, but an actual contraction of the market.

In the table above, I track three measures: national unemployment rate (blue), hiring in the Midwest (gray), and layoffs in the Midwest (yellow). I purposefully left out "quits", to remove those people who leave the job market willfully, to emphasize the number of people who are taken out of the employment numbers due to structural factors. You can see the impact of the 2008/09 recession by the long-term spike in lay-offs, and the decreases in hiring that began in 2007, which is also when the unemployment rate began the slow incline. Compounding the lay-offs and decrease in hiring, you see the tremendous spike in unemployment rates.

Regardless, neither of these measure what my friend asked--whether there is a "jobs available index." But it seems to be about as close as we have for the time being, except for individual studies doing micro-level survey research.

No comments:

Post a Comment