Last month, I wrote: "Next month will have to be an outlier for unemployment to come in above 5.2%." So, my perfect track record remains intact. It was apparently an outlier. Right again!
So, what's going on? I think it's a little bit outlier, a little bit persistence in higher unemployment numbers, and a little bit of a lull in employment growth.
Here is my graph comparing insured unemployment and total unemployment. For 18 months after the end of 2012, there was a steady trend in the decline of very long duration unemployed workers. Since July 2014, this decline has slowed, although it continues to have a downward trend. Since July 2014, the red and dark blue lines show the upper and lower bounds of expected total unemployment, based on a continuation of the previous trend at the lower bound and a complete halt to the decline of very long term unemployment at the upper bound. This month, that range was from 4.8% to 5.3%. Total unemployment came in at 5.5%. So, total unemployment has bumped above the range we would expect, given continued unemployment claims. This is not a pattern we expect to see. Here is the graph, smoothed, and going back to the 1970s. There is a counterclockwise pattern that we see through business cycles. The trigger for a change in trend at this point in the cycle should be a sharp increase in unemployment claims. But, before that happens, we would expect to see unemployment insurance claims level out while total unemployment continues to fall. So, the movement of the last few months is probably anomalous, and I would still expect to see a correction down in the unemployment rate of at least 0.2-0.3%.
This bump up in total unemployment is playing out in the shorter unemployment durations. One reason for this may be dynamics related to Part Time Employment for Economic Reasons, which is still slightly elevated from what we might consider recovery levels. (Note: The sharp downward shift in 1994 is a measurement change.) Not only is total unemployment elevated compared to insured unemployment, but the number of job losers is also elevated compared to insured unemployment. An elevated level of job losers, compared to insured unemployment, seems to have coincided with periods where there were high levels of part time workers for economic reasons. This might have to do with less universal unemployment insurance claims by part time workers. So, it may be continued claims that is the false signal right now. Or, more specifically, we may be seeing temporary frictions from the normalization of the labor market.
We tend to imagine workers having hours cut and then having full time status reinstated. But, I think we need to be careful about narrative explanations here. Both labor supply and labor demand have strong influences on the labor market. (Colloquially, and sometimes academically, we understate the influence of labor supply.) There is a lot of churn among workers as these markets normalize, and it seems reasonable that some of this exchange between part time and full time could create temporary unemployment from market frictions, even if the causal factor is employment growth that is pulling workers back into the full-time labor market. In the 1980's, this increased level of unemployment and job losers, relative to insured unemployment, appears to have persisted for several years. That might be the case again.
JOLTS data appear to continue to show generally positive trends. This data is a month behind, and a little noisy. A weighted moving average gives a little faster indication of changing trends, and here Hires and Quits are beginning to show weakness, while Job Openings continues to look very strong. But, all of these indicators were very strong in late 2014, so it is a little early to call this a problem.
A possible lull in hires might be related to some of the increase in unemployment. I had expected to see some of the unusually high very short duration unemployment dissipate this month. But there appears to be a hump of unusual unemployment moving through the durations in the seasonally adjusted data. On the bright side, very short durations moved back to normal levels this month. Maybe there has been a temporary hiring lull.
Looking at flows, there has been a decrease in net flows from Unemployed to Employed. But, flows directly from Not in the Labor Force to Employed continue to be very strong. In addition, the individual flows all continue to move in positive directions, even though they are at or near full recovery levels.
Any lack of flow out of unemployment into employment appears to be matched by very strong flows back into the labor force and directly into employment.
Wage growth was relatively strong in May. I only have real wage growth through April (because I use the PCE price index to adjust for inflation). Wage growth has recently crossed slightly below the long term trend given by the unemployment rate. But, that should recover slightly this month, once we have inflation numbers.
In total, I think these indicators still point to a healthy labor market and to an unemployment rate that should step down a bit from current levels. The forces that pushed it from my 5.2% call to the 5.5% print may reflect some persistent headwinds, but I think we should still expect a reversion to a trend in the unemployment rate that is lower than the last couple of months' numbers imply.