Thursday, February 12, 2015

"It's just a recovery in low wage jobs." Not anymore

 - by New Deal democrat

Here is something I've been meaning to post for the last 6 months!  Over the last 5 years, from time to time the National Employment Law Project has posted periodic updates on the types of jobs created since the bottom for jobs in the beginning of 2010.  Each graph is cumulative, i.e., each graph includes the time frame included in the previous graph.  Note what happens over time, in particular with higher paying jobs.

The first update covered the period only through July 2010:

Note that only about 5% of all jobs created in the first few months after the bottom were in the top 2 quintiles.

The next update extended the period through January 2011, and divides into 3 rather than 5 wage brackets:

Note that now, 179,000 of 1,270,000 jobs, or 14% of all jobs created, are in the top 1/3.

The next update was in August 2012:

According to the NELP report accompanying the graph, high wage jobs accounted for 20% of all recovery growth, while mid-wage jobs accounted for 22%, and low wage jobs accounted for 58%.

The next update was February 2014:

By now, according to the NELP, 30% of all jobs created since the bottom in February 2010 were high wage jobs, 26% were mid-wage jobs, and 44% were low wage jobs.

Finally, in August 2014, the NELP published the following graph which covered only the period from July 2013 through July 2014:

According to their report, 33% of all jobs created in that period were high wage jobs, with only 26% mid-wage jobs, and 41% low wage jobs. In an accompanying footnote, they stated that 38% of all jobs created in the first 7 months of 2014 were high wage jobs.

In case it isn't already clear, here is the percentage of high wage jobs created over each time frame:

2/10 - 7/10* 5%
2/10 - 1/11 14%
Q1 2010 - Q1 2012 20%
2/10 - 2/14 30%
7/13 - 7/14 33%
1/14 - 7/14 38%

*(measured by top 2 quintiles)

Since all but the last two measures are cumulative, that means that the most recent job growth in each was even better than reported.  For example, the figures from February 2010 through January 2011 and through February 2014 make it easy to calculate the period from January 2011 to February 2014, which I've calculated for the list below:

1/11 - 2/14:
High wage jobs (2.424 million) 33%
Mid-wage jobs  (1.814 million) 24%
Low wage jobs  (3.211 million) 43%

While low wage jobs in this period predominated over mid-wage jobs, high wage jobs were created at a normal rate.

On the other hand, it is fair to say that since the NELP calculated that higher paying  jobs were disproportionately lost during the Great Recession - chiefly in manufacturing and government - the net change including both the recession and the recovery has still been from high paying towards low paying jobs.

One of my theories has been that this jobs recovery isn't so much different in kind from past job recoveries, but rather different in scale. The limited data I have been able to access suggests that lower wage jobs always come back first, with higher wage jobs increasing later in a recovery. The Great Recession was so deep that it took a long time for lower wage jobs to come back on for higher wage jobs to begin filling in.

On a similar note, here is Prof. Menzie Chinn at Econbrowser posting a graph from Torsten Slok of Deutsche Bank questioning whether it has been a low wage recovery at all:

Discusing Slok's work, Shane Ferro at Business Insider wrote:

What's the difference here? It's all in the methods. 
In the NELP analysis, the division of low, medium, and high jobs is made by lining up each industry's median wage, then dividing that into three equal groups. . . . . 
On the other hand, Slok's buckets aren't equal. He breaks down high, medium, and low wage jobs by occupational groups . . . . 
This delineation means that Slok puts far fewer workers in the low wage bucket than the other two, therefore the number of jobs created in that bucket is necessarily going to be smaller. It's not perfect, but breaking it down by occupation, rather than industry, gets a little more granular about how incomes correlate to each bucket.
I mentioned last week that median wage growth in the fourth quarter, as measured by the Employment Cost Index, outpaced average (mean) wage growth as measured in the monthly jobs reports.  That might just be an anomaly, or it might be the relative increase in higher wage jobs showing up in the official data.