The BLS (after reopening from the government shutdown) released the Employment Situation report for September, leaving the broad landscape of the economic recovery pretty much unchanged. The Great Recession is still casting a long shadow over the labor market as employment growth continues to be anemic and the employment/population and labor force participation remain at the lowest levels in more than thirty years. As we noted in an earlier blog, the weakness in the labor market continues to play against the Federal Reserve’s earlier attempts to provide forward guidance about asset purchases and interest rates based on thresholds for the unemployment rate.
Non-farm employment increased by 142,000, below most economist’s expectations. The private sector added only 126,000 jobs, of which 100,000 came from the service sector: the largest gains in Transportation and Warehousing (23.4k), Retail Trade (20.8K) and Temporary Help Services (20.2K). Given that the next report will come out on Friday November 8 (less than a month from this report) and that it will be influenced by the shutdown, it is unlikely to be very informative.
While a 142,o00 monthly employment gain is well below ‘normal recovery’ standards, the pace of improvement in the labor market has been consistent now for several years. Most labor market indicators have either returned to their pre-recession levels or are slowly nearing them. Consider the growth in non-farm employment:
While the level of employment has not yet reached its pre-Great-Recession peak in nearly six years, it has been climbing steadily. The monthly net employment gain has averaged 179,000 since January 2011.
The unemployment rate in September ticked down slightly from 7.3% to 7.2%. Initial claims, obviously affected by the recent shutdown spiked up, but the trend down is also indicative of an consistently improving labor market.
Interestingly, labor productivity measured by total output divided by the total number of labor hours looks very similar to all previous recoveries except for the 2001 cycle. Even as employment fell 5% below its peak level, productivity continued to rise.
As has been the case for some time now, the Taylor Rule indicates that the federal funds rate should be significantly higher than its current level. The rule prescribes an interest rate policy given the how far unemployment and inflation are from their long run stated targets of 6.5 and 2 percent, respectively. If the targets are hit exactly, the Taylor Rule gives a 4% fed funds rate. Currently, the rule suggests a rate of around 2%, well above the zero bound where the fed funds rate has been since 2010.
The question is what exactly the Fed is looking at that suggests that there continues to be a sufficient need to keep interest rates near zero and the rate of asset purchases high. One labor market indicator that might be particularly worrisome is the employment to population ratio. The employment to population ratio compared to all other recoveries looks much different, it fell nearly twice as much and has shown very little recovery.
Indeed, the ratio is now about what is was back in the early 1980’s. Unfortunately, we do not have any way of really knowing what this ratio “should” be! The reason is that female labor force participation, baby-boom retirees, schooling decisions and a host of other things determine the numerator, not just how many get employed given they choose to work. In other words, it is by no means clear that it is expected to rise back to its level in the year 2000 or even 2007. Note, though, it appears to have stabilized over the past few months.
Another possible troubling trend is the relationship between unemployment and job vacancies, referred to as the Beveridge Curve. Below we plot the vacancy-unemployment relationship since 2000. The variables moved down and to the right during the previous recession as job vacancies fell and unemployment rose. During the recovery it has shown a counter-clockwise loop as it moves back up to the left. This movement is not unexpected as noted by the Diamond-Mortensen-Pissarides workhorse model of search unemployment. The vacancy data in this graph come from JOLTS, however, these data only started in December of 2000.
The graph below uses a different data set (now defunct), the help wanted advertising index from the Conference Board, but we have only used the data from 1951-1998. This is the series that was formerly used by those looking at job vacancies before the advent of JOLTS. The colors indicate different decades. Ok, just as a quick quiz: Where are the 1950s on the graph and where are the 1990s?
The answer to the quiz can be found here. The point is that the outward shift in the Beveridge Curve might come about for different reasons. First, as noted in the standard labor search framework, the counterclockwise movement comes about since firms can quickly post vacancies, yet matching firms to workers takes time. However, outward shifts can occur due to skill mismatch, here is an example, or changing demographics. It’s not clear which factor is causing the trend in the Beveridge curve since 2010, as it is also not clear what the role of monetary policy is in correcting the outward movement.