How well are workers doing?
The President’s year-end list of accomplishments noted the following:
The Economy Is Growing And Creating Jobs. Since May 2003, the economy has added nearly 4.5 million new jobs. The unemployment rate is down to 5 percent – lower than the average for the 1970s, 1980s, and 1990s. Last quarter, the economy grew at 4.1 percent and has been growing steadily for more than two years.
(From Fact Sheet: President Bush’s Accomplishments in 2005)
The jobs figure caught my eye, not because the figure is wrong (it isn’t), but rather because of what that single figure obscures. I thought it’d be useful to trace out the performance of employment growth in this recovery relative to the last recovery (which was itself acknowledged to evince slow employment growth).
First is a depiction of real GDP and payroll employment growth in the 1990′s episode.
Second is the graph of the same series for the 2001 episode.
In both cases, the series have been normalized to zero in at the end of the recession. The gray shading denotes NBER-defined recession periods. The GDP data are in Chained 2000 dollars, December 21, 2005 release. The total payroll employment figures are in thousands, from the December 2, 2005 release, as reported by the St. Louis Fed. The monthly figures are arithmetically averaged. Both variables are then logged.
It is obvious from this graph that the new jobs figure is maximized by the judicious selection of the base date (May 2003 being the trough in payroll employment). Comparing the November number against the peak employment number in February of 2001 yields only an increase of 1.7 million (or 1.8 million from the beginning of the first Bush administration).
Of course, employment numbers don’t tell the entire story. For instance, the widely acknowledged rapid productivity growth may have manifested in rising wages and/or compensation. But as noted in James Hamilton’s post on wages and compensation, matters don’t look altogether that good for workers on those other dimensions. Yet another way of looking at the issue is to examine labor’s share of income. Below, the share of paid compensation in gross domestic income is shown, along with the 1967-2005 average. (These data are from the December 21, 2005 BEA release, Table 1.10)
Brian Horrigan has made the point that normalizing by Gross Domestic Income may yield misleading inferences given the higher rates of capital depreciation during the last decade and a half. Hence, I’ve added the figure below, which normalizes by Gross Domestic Income subtracting out the capital consumption allowance (I’ve omitted 2005q3 because of some aberrant behavior in the CCA series).
The same patterns persist; the wage share is smaller in recent quarters, with the spike in 2004q2-q3 largely an artifact of the upward spike in the CCA series (which I believe, but do not know for a fact, has to do with the tax return data the CCA series are based upon). In other words, the wage share is below the 1967-05q2 average, and even below the 1995-05q2 average, although the deviation is somewhat smaller than that obtained using GDI.
Time will tell if these trends persist or reverse.