As I said Friday, the trend will probably start to level out in unemployment, making this indicator less informative for macro-forecasting. The unemployment rate came in a little high this month, so there will probably be some decline off this level in the near term. But, I think we are likely to see a kink here to more level trends until the next downturn. Let's hope that is a way off. There is no sign of an imminent downturn in the productive economy.
Durations are probably settled into long term ranges. I expect durations under 26 weeks to move sideways from here. It is looking more and more like there will be some persistence in the various groups of long term unemployed, so December will be closer to 5.7% than to 5.5%, and the trend looks like it might flatten. Early in the year, I had hoped to see the quarterly exit rate of long term unemployed workers (this graph uses workers over 15 weeks) go over 40% and stay there. It will get there, but the rate has not increased as strongly as I'd hoped.
I am also starting to see a sustained leveling of the very long term unemployed, suggesting that this portion of measured unemployment may have some persistence. This is corroborated by the average duration data. By using the average duration data published by the BLS and estimating the average duration of workers unemployed for less than 26 weeks, we can also estimate the average duration of workers unemployed for more than 26 weeks. This is currently divided about half and half between regular unemployment and the unusual group of very long term unemployed. We can presume that the regular group of unemployed workers with durations over 26 weeks has an average duration of about 60 weeks (the well established range before 2009). We can further infer that the remaining unusual group has an average duration of more than 100 weeks. The average duration of workers unemployed for more than 26 weeks has remained around 80 weeks throughout the recovery, and jumped to 87 weeks this month, suggesting that the reductions in long term unemployment are coming from the regular group instead of the very long term group, and/or the average duration of the unusually long term group is continuing to age as time passes.
As I have mentioned, insured unemployment is a big reason why I still see some room for a drop of a couple tenths in the near term. Insured unemployment continues to fall sharply, and there is no historical precedent for a stagnating unemployment rate when insured unemployment continues to fall. We should expect to see a leveling of insured unemployment, after which, total unemployment will continue to fall slowly as the recovery ages. If insured unemployment continues to fall, then the trend in unemployment won't level off as quickly as I expect. Demographics could put the nadir of insured unemployment at a lower level than it has been in the past (baby boomers, messing with every statistic these days), but it probably is at a place where further declines will come more slowly. In any case, since total unemployment this month went against the trend suggested by insured unemployment, I expect to see some snap-back next month. 5.8% seems high.
Another reason to expect a snap back is flows. They all continue to trend toward recovery levels. The flow from Employed to Unemployed (EtoU) popped up to 1.3% this month. At this point, the EtoU flow should be in the 1.1%-1.2% range, and it has been for a while. If it had been in the expected range, this month would have come in at 7.6%-7.7%. The move up will certainly be reversed in coming months. There is typically a net flow between E and U of about 0.2%. This month there was very little net flow from unemployment into employment, which clearly is not reflective of the employment market. These are very noisy series, and the EtoU flow has been especially noisy while these flows have been elevated.
Wages continue to grow, albeit somewhat slowly. I attribute this mostly to low inflation. Wage growth has been unusually high during the recession. Now, even if, as I suspect, labor markets are more reflective of 5% than 6% unemployment, real wage growth of just under 1% is not out of the ordinary. Back in 2009, when real wages were growing at 2% and unemployment was over 8% - that was out of the ordinary, and it was a problem.
I had been positioning for rising long term interest rates over the past 1-2 years because of QE3 and the end of EUI, which I expected to push up inflation and pull down unemployment, relative to expectations, pulling back the date of the first rate hike. Forward rates did rise a bit, more from a steepening of the yield curve than from a movement of the first rate rise. The inflation never really materialized. So, the date of the first hike remains about where it was 2 years ago. (This, itself, is a huge win compared to pre-QE3 Fed policy, which had seen the expected rate hike continually moving forward in time, like a carrot on a stick.)
As QE3 tapered, mortgage credit failed to break out, and so over the course of 2014, I believe that limits to credit growth have overtaken improvements in the labor market as the critical element for forward rate increases. As we move into 2015, I believe that mortgage credit will be a more important signal for interest rates than employment, and much of the change in forward rates, if credit market expansion grows, will come from a steepening yield curve.
While all the moving parts in the Federal Reserve balance sheet make this difficult to forecast, it seems likely to me that the yield curve will either flatten in a worst case scenario, or rates will increase at a rate of 2-4% per year, as they have in past episodes. Right now, the yield curve suggests a rise of about 1% per year. So, I expect there to be a speculative opportunity here.
I was beginning to worry that the expected date of the rate hike was starting to move forward again as we move away from QE3. But, after the November employment report, the large increase in rates was almost entirely the result of a move back in time of the expected first rate hike of about 1 1/2 months (to approx. the end of 2015 2Q), where it had been before October and right where it had been at the beginning of QE3. This last graph shows the movement of forward Eurodollar rates since the beginning of QE3. The short end of the curve is almost exactly where it was a year ago. At the long end, terminal rates have dropped by about 1.5% over the past year, back to where they were at the start of QE3.
My "expected" rate trajectory accepts the market's expected first rate hike, but follows a rate of 1.5% of hikes per year. Even this is a much slower rise than past episodes of rate increases.
No comments:
Post a Comment