Job growth is it enough? Too much?

Is job growth impressive enough to stop worrying?  

A look at unemployment and participation rates

The jobs picture: unemployment and participation rates

 

The household report, the source report for the unemployment rate calculation, registered 428,000 new jobs in February. It saw 847,000 jobs created in January. That’s nearly 1.3 million jobs in two months. And we still have doubters… And 2.6million jobs have been created according to the Household report in the last seven months.  That’s an average of 371K per month. Private nonfarm payrolls are up by just 1.3million over that seven month period for an average of 191K per month. That’s a lot less than the household figure but not a shabby showing.

Still the labor force participation rate was around 0.66 (66%) before the recession/financial crisis and it is only back to 0.639. So 2.1% - more - of the civilian population over the age of 16 has headed to the side lines helping to suppress the unemployment rate. Note that this is a proportion and in absolute terms the labor force is rising; at a 1.6% annual rate over six months. 

 

Population growth shows that ’women’ is the fastest growing class of labor followed by ‘teens’; ‘men’ is the slowest expanding class.

The labor force shows its fastest growth for women followed by men with teens the slowest

Employment is fastest growing for men followed by women and teens.  The number unemployed is shrinking fastest for men second fastest for teens but it is still rising for women.  As a result of these trends ‘men’ shows the slowest growth among the not in the labor force groups followed by teens with women having the largest not-in-the-labor-force (NITLF) growth rate.

Men are making the greatest gains since they are being employed faster than other groups have the slowest labor force growth, the fastest reduction in the ranks of unemployed and the slowest growth of being NITLF. Men’s labor force growth is slow even though their employment growth (part of the LF measure) is fast.

 

Cyclical Trends in Participation

 

The story of labor force participation rates is being re-written and not just in this cycle. Since 1990 there has been a gradual increase in the tendency for labor force participation rates to drop. In 1990 it was just the young and adult men along with all teens. The teen plus adult men’s rate and the adult men’s rate have steadily been shrinking. But the decline for adult men is accelerating in the business cycle. So is the decline for teen and adult men. The teen and adult women rate has been slowing and is now shrinking at an ever faster pace since 1973. Adult women continue to see their participation rate rise, but the pace of that increase has slowed, and slowed sharply. In an all-inclusive time-series format the adult women’s rate is down from its peak.

So there are several kinds of forces that are in play. There are a number of category-specific longer term trends that are in play. But it is also true that some of these effects are intensifying over the most recent cycles. Women’s participation rates once rising by 2.9 % points at the 32 month mark of recovery are up by just 0.10 pct points in this cycle. Historically, that was one of the important offsets to the long term structural drop in men’s participation rates. The men’s drop has had some vacillation but it has averaged a drop of 1% at the 32-month mark of the expansion cycle. In this cycle the men’s rate has dropped by 1.8 % points. The rate for teens has been dropping in a more accelerated fashion and the drop of five percentage points at the 32-month mark is identical to the drop at this point in the 2001 cycle.

In the 1990, 2001, and 2008 cycles participation rates have been more prone to drop than in earlier post war cycles. Interestingly 1960 is an exception.

People are screaming about the high rate of unemployment and the Fed has swerved its policies to try to reignite growth. Job growth is in gear again but participation rates channel much broader trends. The Fed needs to be careful about being pushed to achieve objectives that are beyond its grasp.

The metrics for participation by educational attainment do not have a long history. But in this recovery period the rate for those with less than a high school diploma is up by 0.4% points from the recession end (yes, that’s UP). But for all other educational; classes it is lower!  For participants with a high school diploma but not college the participation rate is lower by 3.8% points. For those with some college but no degree the participation rate is lower by 2.3% points. For those with a college degree their participation rate is lower by 1.4% points.   

The declines in the unemployment rates have favored the less educated. For the least educated group (no high school diploma) the drop in the unemployment rate since the recession ended is 2.6% points; for high school grads it is -1.5% points for those with some college it is -0.8% and for college grads the drop is 0.6% pts. Now this also reflects the fact that the level of the unemployment rate is highest for the least educated. But they still have had the greater proportionate drop. 

                                                                  Caution Fed caution!!

The table above breaks down unemployment rate progress by different categories of unemployment: BY REASON. Job losers probably most come to mind when think of this category. They make up 55% of it and included workers on temporary layoffs and on permanent layoff (factories closed, out of business etc). Job leavers made up another 8%, of the unemployed with re-entrants to the labor force accounting for one quarter of the unemployed. New entrants are currently about 10% of the pool of unemployed. 

Compared to historic norms the group ‘not temporarily unemployed’ is quite large. It makes up 46% of the unemployed pool right now compared with a normal proportion of 37% and a previous high of 40% (1990 and 2001). The proportion unemployed temporarily is 8.7% compare to an average of 12.1% and a previous low of 10% (1970).

One implication here is that a large bulk of the unemployed reflects people that are not simply going back to their old jobs when the economy picks up.

The Fed may take its dual mandate seriously. But these data suggest that a great big chuck of the main unemployment problem is not simply cyclical. ‘Job leavers’ is at a low proportion historically making up only 7.9% of total unemployment compared to 12% on average at this point in the recovery cycle . But ‘re-entrants’ and ‘new entrants’ are at a near-normal proportion of total unemployment. Job leavers may be few in numbers because, with such high unemployment, not many are willing to leave jobs. So fewer wind up being unemployed. This low reading is not a mark of success. The low quit rate reinforces this view.

There is another way to assess unemployment by reason and that is to look at the contribution to the rate of unemployment. Looked at through this lens the categories for unemployment are all within one standard deviation of the previous mean for unemployment by category at this point of the cycle except for ‘not temporary job losers.’ Of course this means that the overall category that includes this measure, Job Losers, also is distorted up. But the other component of job losers, ‘temporary job losers’ is actually normal. What stands out it that the unemployment rate is elevated by about 1.4 percentage points (it could be 6.9%) except for the excessive strain on the category ‘not temporary job losers’. The reading for this category is three standard deviations above the mean., a reading that makes it impressively significant. And this category is a proper target for fiscal policy not for monetary policy. 

The Fed, as much as it wants to boost growth, will have to be aware that this large crop of ‘permanent’ unemployed may be harder for the labor market to absorb. Having said that, the table above show that the current level of permanently unemployed as a percentage of the recession end peak is only 75% which is the second lowest among all these cycles. Thus progress in reducing ‘permanent unemployment’ has been quite good. But this has also been a very large pool of unemployed by historic standards. Temporary unemployment is at 65% of its peak and that is about normal at this point of the expansion.  

Job leavers are 26% above their peak level but this class averages a higher number; at his stage the expansion the average rise should be more like 11% not 26%.

Re-entrants are ‘doing well’ by historic comparison as their ratio to their end recession level is low. It is up 1% from its recession-end level and the norm would be to rise by 3% to 4%.

New entrants unemployed are up 39% from their recession-end level while the norm is somewhere between 16% (average) and 25% (median).  New entrants are struggling to get back into the labor market.

On balance a number of factors are in play in understanding participation rates and unemployment rates. Many of them are out of the Fed’s control. Some are structural and others are demographic. The danger is if the Fed pushes too hard to try and get the unemployment rate down it could backfire. This past week we did see unit labor costs elevate and they are up for two quarters running and are up to 3% Yr/Yr. This may not be a sign that the Fed has reached its limits but job growth is getting stronger has become more reliable with nearly all job gauges giving off consistently firm readings. Still, the Fed is obsessed with its multi-mandate. We can only hope that it will recognize, not just that it is pushing on a string, but that it is pouring stuff into a punch bowl that is already overflowing.

If might do a better job of policy recognition error if is not as guilty as I am in mixing its metaphors.