Dueling Job Data, Revisited

Last month I wrote about some mixed signals in the data from two different job surveys.  While the rate of job loss at San Diego companies was improving, the rate of loss among San Diego’s residents — regardless of where they are employed — was hitting new highs.  The graph accompanying this article shows that this gap widened further in October.

I asked local economist Kelly Cunningham what he thought of the disparities between the two job surveys since the recession began.  Kelly, who is with the National University System Institute for Policy Research, has talked me through a couple of economics topics before and had some thoughts to share.

He believes that the occasional differences between the survey results stems primarily from the fact that the labor force (or “household”) survey includes self-employed people, while the payroll (“establishment”) survey only counts people who work for businesses.

As he puts it, “[P]art of the discrepancy occurs as workers are laid off or lose their payroll job and attempt to find temporary work or work independently for themselves. This would show up as a loss in payroll job, but the household survey data would not indicate a decrease.” Labor force employment (and self-employment in specific) could even be increasing as this happens due to population growth and the other usual factors.

As the recession wears on, says Kelly, some people who have been giving self-employment a go find themselves unable to keep it up and become unemployed.  This could result in a catch-up effect as apparent labor force employment drops according to the household survey.

This would help to explain both why the household survey painted such a rosier picture than the establishment survey in 2008, and why it is making things look worse now.  It also supports Kelly’s suggestion that the two surveys are usually in pretty good alignment but tend to get out of phase during the beginnings and ends of recessions.

I have my own little additional theory about the 2008 increase in household employment even as the establishment survey showed a decline in payrolls.  In addition to the self-employment lag effect described by Kelly, I suspect that the collapse of the Inland Empire housing market may have helped to boost the number of employed people living in San Diego.

Here’s my thinking.  We know that foreclosures were rampant in Temecula and other Inland Empire towns.  We also know that many of the foreclosures weren’t due to job loss, but to the fact that that the severe home price declines had put many buyers so far underwater that they felt no reason to keep paying their mortgages.  Many residents of the Inland Empire commute to jobs in San Diego, so it’s plausible to believe that many people who walked away from their Inland Empire homes but still had San Diego jobs might have simply moved to San Diego.

If my theory is correct, this would have resulted in an increase in the number of employed people living in San Diego even as local companies were cutting payrolls.  This effect would have further reinforced Kelly’s self-employment-related lag.

Now the tables have turned and household employment is looking worse than payroll employment on a year-over-year basis.  But due to the effects described above, the household survey has to compare to a much higher year-ago number than the establishment survey.  It’s the catch-up effect in action.

It seems that there are some reasonable explanations for the disparate behavior between these two surveys.  I’ll keep updating both measures of employment, but given the apparent lags in the household survey, I suspect that the establishment survey — the one that has started to look a little better — is providing a more forward-looking read on what’s happening with local employment.

— RICH TOSCANO