A tale of two graphs
In an interview published Sunday, the US Treasury’s Chief Economist Alan Krueger told the Financial Times that: “The jobs situation – as difficult as it is – has actually started to improve earlier than in the last two recoveries.”
Today, Ezra Klein picked up on this same point, running with this graph from a post by Invictus at Ritholtz.com:
Source: Invictus, Ritholtz.com
This first graph shows the change in total private-sector employment from the trough of the last three recessions (assuming that the Great Recession ended in the summer of 2009). By this measure, the last year or so doesn’t look so bad. Private employment turned around sharply about the beginning of this year, well ahead of the schedule for the previous two recessions.
But there are two problems here. The first is that the Great Recession was a lot longer than the other two. That is why the gray area marking the recession is so much thicker in the late 2000s. In effect, the graph above doesn’t give the early 1990s or the early 2000s recessions any credit for being short –because all the employment data in the chart start after the recession is over.
The second problem is closely related: the chart above sets the post-recession private employment level equal to 100 and compares the recovery only to the absolute depths of the recession.
The chart below (also from Ritholtz.com, but not reproduced in Klein’s post) illustrates just how much of a buzz kill factoring in the length and depth of the Great Recession can be.
Source: Invictus, Ritholtz.com
The top three lines are exactly the same as the ones in the first chart. The bottom three lines, however: (1) start tracking private employment from the beginning (not the end) of the recession and (2) use the employment level immediately before –not after– each recession as the benchmark (employment at the benchmark is set equal to 100).
There is just nothing to feel good about in that second graph.
You clearly have a point here, but a quibble that might (or might not) be significant occurs to me regarding the benchmark: it seems to me that bubbles nearly always have some inflationary effect on employment, and a housing bubble is likely to have a highly inflationary effect on employment. If I am right about this, shifting the benchmark merely temporally seems unlikely to get closer to the economic reality. Deflating the size of the bubble we have just gone through took longer, and had deeper effects that the previous two graphed — but there is still some (may be not much, but some) good news in the slope of the upturn after the low point.
If I am right about this, then there is also some (again, may be not much, but some) good news politically for the Obama administration: having had no responsibility for the bubble itself, it had no choice about riding the deflationary downturn, but it is hard to imagine producing a sharper upturn after bottoming out; and easy to imagine making other policy choices that would have deepened and/or prolonged the trough, and decreased the slope of the recovery.