One of the most difficult things to statistically explain, let alone forecast, is productivity. And, as the chart below shows, past productivity performance is not a very good predictor of productivity in the future. This is another way of underlining the Fed's (and all economists') inability to forecast.
The Fed's recent policy tilt to have a projected 'best case forecast' of super low interest rates belies its own ability to know if such a forecast is correct. Bernanke has even admitted as much in his 'on-the-Hill' testimonies. Real GDP is essentially productivity growth times total hours worked. The productivity part is nearly unforecastable. Still, the Fed has this long-dated forecast in play saying that the economy will remain weak and inflation low; increasingly we are given some reason for doubt.
The recent unit labor costs (ULC) and productivity report shows one set of nagging concerns. It depicts a pattern since 2010Q1 of productivity on the wane and of unit labor costs on the rise. This is an uncomfortable pair of trends for a Fed that 'presumes' inflation will remain so low that it will not be raising rates until 2014.
Of course, the Fed statement is not a promise. It is what economists call, 'a conditional statement.' But to put meat on its bones the Fed also says it is a policy tilt based on its 'best-case' forecast. But now very early in the process of low-rate encouragement, we begin to see the flies in the ointment.
The Fed's Theory: The Fed is of the opinion that there is a large GDP gap. That refers to the notion that current GDP is far below its potential. We can represent that by either a low capacity utilization rate or a high rate of unemployment. Both of those traits would be expected if the GDP gap were high.
The Fed builds from this observation of a GAP, a measure of economic slack, to the opinion that, with slack, factor prices (like wages) are unlikely to gather much momentum. The availability of idle resources should mean that, if business expands, new employees will offer their services at current wages or that idle existing capital can be used with no pressure on inflation 'for some time.' But the rise in unit labor costs with unemployment still so high raises questions about that view.
(Click chart to enlarge)
Historic data on unemployment rates raise their own questions…for this set of inquiries we turn to data on 'reason for unemployment.' In the table above we have recast these data as contributions to the overall rate at this, the 32nd month of the economic recovery cycle. What we see is that the level of unemployment at this stage of an economic recovery is more or less normal except for the category 'Job Losers' and one of its main components, 'not temporary job losers.' 'Not temporary job losers' look to be the big issue in this business cycle. Those unemployed because of 'temporary' job losses raise the unemployment rate by 0.7 pct points, close to the historic norm of 0.8pct points. Job leavers add 0.7pct points to the rate of unemployment where 0.8% would be the expected norm. Re-entrants to the labor market add 2.2pct points to the rate of unemployment, higher than the norm of 1.8% but within one standard deviation of 'normal.' Still, this metric is the high-one among all post war past cycles at this stage of recovery. New entrants to the labor market are raising the rate of unemployment by 0.9pct points, close to the 0.8% that is average.
So we are left with 'not temporary jobs losers' as the big problem. These are workers who lose jobs and are not expected to go back to them because a plant was closed or a firm failed. While some want to focus on a different breakdown of the unemployed and feature the 'long term unemployed' as 'THE' problem I would argue that that statistic is an artifact of this one. The 'not temporary job losers' are hard to put back to work, they tend to be unemployed for longer; they make up 'the bulk of the unemployed. 'Not temporary job losers' or, if you will, the 'permanently unemployed' account for 47% of overall unemployment in this cycle at this time. This proportion usually averages 35% of total unemployment, but it has been elevated in each of the recent economic cycles. 'Permanent' job losses have added 3.9% points to the overall rate of unemployment (which stands at 8.3%) compared with a contribution of 2.5% points normally at this point. Thus the unemployment rate would be lower by 1.4% points (it could be 6.9%) if permanent job losses were normal.
Progress in reducing the rate of unemployment in an expansion usually comes from reducing temporary unemployment. Temporary unemployment is typically 20% of total employment as the recession ends. But in this cycle it was only 12%. Temporary unemployment tends to fall by 48% (median drop) by the 34th month of expansion. In this expansion it has fallen by 35%, which is the average not the median result; moreover, because the category was small, it had a lesser impact on the overall rate of unemployment.
This recovery also had 'not temp' or 'permanent' unemployment as 53% of its unemployment at the recession end when the norm is only 35%; permanent unemployment usually falls by only 5% by this point in the cycle. So the larger presence of permanent unemployment has made the overall rate sticky.
And, in fact, the other categories on unemployment: job leavers, new entrants and labor market re-entrants, normally tend to remain elevated above their end recession levels at the 32nd recovery month.
In fact in this cycle permanent unemployment has dropped faster than we would have expected; it had fallen by 22% from its end recession level compared with the average drop of 8% and the median drop, 5%.
So in some ways we are doing 'well' but only compared with past trends within categories. The word of caution to the Fed is that the unemployment bulge is not the right focus because it likely stems from permanent job losses and structural imbalances. If the Fed, in pursuing its dual mandate, tries to reduce the unemployment rate it will be fighting a real structural imbalance, something monetary policy is ill-equipped to handle. The risk is higher inflation.
Over the past seven months the payroll report has averaged job gains of 184K per month. The household report has averaged a stunning 374K per month. Job growth has been good/solid/strong. In time the two reports do tend to converge. The Fed should be cautious about a too-dogmatic adherence to its belief in the GDP Gap model. When the GDP gap hits the reality gap it is time for policy to opt for realty.