Has the U.S. labor market adjusted?

by on August 6, 2013 at 7:15 am in Data Source, Economics | Permalink

Via Ashok Rao, this is a scary chart:

labormarketadjusted

If read “crudely,” the chart suggests that in terms of percentage changes, we already have seen a historically normal level of adjustment.  There is an interesting discussion at the link.

By the way, if you’re not already, you all should be reading Ashok Rao.

Ashok Rao August 6, 2013 at 7:51 am

Thanks.

By the way, this is a comment I left on Scott Sumner’s blog which clarifies an important point since the study refers to job creation rather than unemployment reduction:

“This (http://research.stlouisfed.org/fredgraph.png?g=leZ) is perhaps a better graph, since the study refers to job creation, rather than U3. Though the signs are the same.

If the mechanism of min wage unemployment as suggested by the authors is the same as a Keynesian sticky wage mechanism, how do we know all further falls in unemployment have to do with sticky wages?

I’m not saying unemployment won’t fall further just curious about the extent to which wage adjustment is the key drawback.”

While unemployment rate can and will keep falling – if the mechanisms in this study are correct – it seems like sticky wages will not be the key driver. Again, that’s just a direct interpretation of what the study did to another context, whether that means something to you is only dependent on how well you think minimum wages act like sticky wages!

crs August 6, 2013 at 8:05 am

Ashok, I would not use a percent change of a *percent* to compare unemployment across cycles (first differences or deviation from pre-recession values) would be easier to interpret). The employment-to-population ratio (http://economix.blogs.nytimes.com/2013/08/02/an-employment-number-that-isnt-budging/?_r=0) still looks quite unusual ,,, it’s not that the unemployed are dramatically different from the non-participants (on the margin). Even if the labor market has settled into some moderate trend path, that does not mean a large positive aggregate demand shock would not re-shuffle the deck. Also do you know that the peak was usual relative to other cycles? One can’t assess the current state without a view on the recent past or the longer-run trends (demographics, natural rate, etc.). In any case, unused or underutilized labor is still there to be engaged, whether or not (or when) the natural forces of recovery will engage them is different question. … All that said, of course, I agree with the tip to follow your work.

Ashok Rao August 6, 2013 at 8:12 am

You’re right that using the absolute number is the better graph (I just think better in percents). You’ll see that the figure is effectively the same whichever you use, though.

The only thing I’m trying to point out is that if you find sticky wages and minimum wages as having the same theoretically-adverse effects on employment, you would expect sticky wages to halt job growth in the recovery as real wages are too high. Once the job growth is at its pre-recession level, however, why would you expect wage-mismatch to explain any bulk of labor market problems?

There are good reasons to believe the transmission mechanism is *not* the same, and the only empirical evidence I can offer is that points at which we’ve reached the pre-rec job creation peak also represent times that would seem to mark near full adjustments.

crs August 6, 2013 at 8:26 am

“Once the job growth is at its pre-recession level, however, why would you expect wage-mismatch to explain any bulk of labor market problems?” Well trends, like demographics, could have been altering the labor force … with maybe even an extra reset kick in the recession to unwind prior above-trend growth … so mismatch might still be there (or it might not or maybe never was) but simply pre and post comparisons are only suggestive.

Frederic Mari August 6, 2013 at 8:48 am

Maybe I am reading into your comparison between sticky wages and minimum wages too much but, for me, the key insight was that, IF minimum wages increases have little consequences on the level of unemployment in the short term AND you can more or less translate that effect for the sticky wages, THEN MMers are wrong that stoking inflation and the resulting lower real wages will do any good to unemployment rates.

Which, imho, was quite clear (I’ve linked my own relevant articles in a comment on Ashok Rao’s blog) but I had intuited it differently – that is to say, in one word, that it doesn’t matter how cheap Labour becomes, if a business doesn’t see Sales growth, it will not hire.

With the classic catch-22 that, if Labour keeps on getting cheaper, it means less and less disposable income, which means Sales are unlikely to grow.

I don’t pretend that those are original insights but I haven’t seen any supply-siders or MMers offering powerful counter-factuals.

Ashok Rao August 6, 2013 at 8:27 am

The point is this is scary in the sense the recovery may not accelerate to the extent the potential gap implies it will. As Scott Sumner notes, unemployment will continue to fall and jobs will be created. To the extent minimum wages and sticky wages are alike, our problems may not have to do with wage adjustments.

Frederic Mari August 6, 2013 at 8:49 am

To the extent minimum wages and sticky wages are alike, our problems may not have to do with wage adjustments.

+1. Amen to that. See above my own reply.

mulp August 6, 2013 at 12:35 pm

Why don’t economists understand that the wage price indicates the labor supply is too large and creative destruction of of labor is required.

The labor market is creatively destructing as seen in the rapid fall in the relative pool of labor supply. This is the same joyous result of Wards, Sears, K-Mart, and dozens of other department store chains going bankrupt in the 90s. Wards, Sears, K-Mart could not cut prices enough to drive demand because its costs were too high. Wal-Mart has focused on selling to the people on welfare, hiring people on welfare to sell to the people on welfare. Wards, Sears, K-Mart built their business on shoppers who had good jobs making stuff sold in the US and globally.

The jobs being destroyed at those that provide living wages, and the minimum wage merely limits the amount of welfare a worker requires.

Economists should analyze the optimal minimum wage against the total welfare cost. If the minimum wage is lower, then welfare to workers will need to increase, so will lowering the minimum wage increase the total welfare to workers more or less than the reduced welfare from the reduction in the supply of labor.

Social Security benefits are not reduced much if a beneficiary enters the labor pool at takes a minimum wage job. As they have a sizeable welfare benefit, they can afford car and are thus reliable workers. The 20 year can not afford a car, so he is a faster sharper but unreliable worker. If the minimum wage were higher, then the 20 year old might be able to buy a reliable car and thus be a better employee than the 70 year old. But an employer can’t afford to pay the higher wage to employ the 20 year old if all his competitors hire the 70 year olds and rely on the government to make them acceptable workers by paying for their cars.

Of course, welfare levels are set very low, below the levels required to drive economic growth, so that means businesses face a stagnant demand and thus with technology have lower labor demand, which shifts more of labor to welfare….

The pool of labor must fall because the price of labor is below the cost of labor.

Z August 6, 2013 at 8:43 am

Just looking at the graphic in isolation, it confirms a trend observed going back to the Clinton years. That is, since the 80′s, recessions have been followed by increasingly tepid recoveries. The Clinton recovery was not as robust as the Reagan recovery. The Bush recovery was not as strong as the Clinton recovery. So far, this recovery as been a credit fueled stagnation. Given what we see in earnings, the prospects for a boom over the next 18-months are remote, even if you are optimistic.

This may be as good as it gets.

mulp August 6, 2013 at 12:36 pm

Tax cuts kill jobs.

Brian Donohue August 6, 2013 at 12:52 pm

Signed,

An Accountant

B.B. August 6, 2013 at 11:13 am

What the graph of percentage change in unemployment misses is the huge drop in the labor force participation rate in this “recovery”. When people drop out of the labor force, they are not employed, but they are not “unemployed” either. So we have seen a big drop in unemployment without much of a rebound in employment. Unless you think all of the labor force dropping out is “structural,” there is plenty of room for further job growth.

mulp August 6, 2013 at 12:45 pm

It is supply and demand.

Employers want fewer workers, as indicated by the low prices offered, so with prices lower than costs, the supply of labor must fall.

If you are not able to live with your parents, or don’t have Social Security, how can you mean the demands of employers to be a clean and on-time worker when you are living under a bridge, maybe in your car that no longer runs, and get to work by hitchhiking?

McD’s hired financial advisors to lay out a budget for its workers to live well working for McD’s with the result being the assumption of a second full-time job, and what must be subsidized housing and transportation.

Ashok Rao August 6, 2013 at 10:35 pm

You’re right, but using simple “employment” data shows the same dynamics.

Rich Berger August 6, 2013 at 11:42 am

I think this is a very misleading chart. Since the chart shows the change in unemployment year over year, the height of the spike is very important, rather than its shape. In the Obama recession, unemployment continued to increase at a high rate, so the hole was deeper. This is a chart of the first derivative of the rate of unemployment and read “crudely:, I find it very confusing.

Charles D August 6, 2013 at 12:17 pm

It also seems misleading to say that “the adjustment in unemployment is over,” since the graph doesn’t measure unemployment directly. The y-axis plots a first-difference in the number of unemployed over the previous year. So long as that number remains below zero, the number of unemployed will continue to “adjust” (that is, drop, relative to conditions a year ago) The “length” of our stint below zero will help determine *how much* unemployment will be reduced in total by the recovery, and the lengths of these stint seems to vary greatly, given the historical data in the chart.

Ashok Rao August 6, 2013 at 5:58 pm

Yep, it would be wrong to infer that unemployment has adjusted to its current state. Rather, that the labor market has equilibrated at the current wage rate, which is why job growth is back to the pre-recession level. This isn’t necessarily scary, but I worry because to the extent sticky wages are a big component of the output gap, we won’t see an accelerating pace of recovery.

Comments on this entry are closed.

Previous post:

Next post: