The unemployment rate in the United States is currently 4.4 percent and is at a ten-year low. Naturally, the Trump administration has been trying to interpret low unemployment numbers as a sign of their successful job-creating policy. This stands in strong contrast to comments that Trump made during his election campaign, alleging that the “real unemployment rate” was as high as 42 percent. But pre-inauguration Trump is not the only one who is concerned that the official unemployment rate does not fully capture the gap between the US economy’s current state and its full potential, often refered to as “slack”. Despite low unemployment numbers and a still very low interest rate, inflation has been timid, giving rise to theories that the slack in the US economy is higher than the unemployment rate might suggest. I take a brief look at falling labour force participation rates and underemployment to investigate the amount of truth in these theories and discuss implications and some potential remedies.
During the Great Recession, the official unemployment rate for the US peaked at 10.0 percent in October 2009 and it has been on a steady decline ever since. It has now reached its pre-crisis level and is approaching a stage last seen during the boom years that lead up to the turn of the century. The data is based on the monthly Current Population Survey (CPS) and everybody who did any work at all for pay or profit during the survey reference week counts as employed. Unemployment is defined as not having a job, having actively looked for work in the prior four weeks, and being currently available for work. That this can be a restrictive definition is indicated by the case of Germany, where 6.0 percent of the labour force have been registered as unemployed in March but only 4.0 percent counted as unemployed according to the definition of the International Labour Organisation (ILO), which is close to the US definition. Anybody who is unemployed but is not currently looking for a job, for example out of desperation, does not show up as unemployed in official US statistics. This gives rise to the theory that after the crisis, a lot of workers got left behind, stopped looking for work altogether and dropped out of the official unemployment statistics.
The unemployment rate is calculated as the ratio of the number of unemployed workers to the number of workers in the labour force, defined as all employed plus all officially unemployed workers. Proponents of the theory that the official unemployment rate is biased downwards often point to the fact that the labour force participation rate, defined as the share of the population (16 years or older) that is currently counted as employed or unemployed, has been declining since 2000 and especially in the years following the start of the Great Recession.
A part of this can obviously be explained by demographics and educational attainment: The share of old, retired workers is increasing and more students stay in tertiary education longer. However, this is not the whole story. A look at the participation rate of the so-called prime age demographic, workers aged between 25 and 54, shows that also a group which should neither be influenced by a higher educational attainment nor by an aging society exhibits a declining labour market participation. This drop in prime-age labour force participation started in 2000 when female participation growth ceased to offset the decline in male participation that started already in the 1950s.
Black et al. (2016) conclude that the drop in prime-age male labour force participation is primarily due to “reduced labour market opportunities for lower-skilled workers, a factor that is also consistent with the decline in relative wages of lower-skilled workers”. Moreover, they find that rising reliance on government assistance cannot be an important factor in this development, as coverage has declined. The fact that in the past two years, labour force participation has slightly rebounded further suggests that a part of decreasing participation rates can be explained by workers who would like to work dropping out of the labour force because of dire labour market conditions. Aaronson et al. (2012) estimate that half of the drop observed in the labour force participation rate is due to economic fluctuations. This cyclical drop in labour force participation seems not to be evident in recessions that occurred before 2000. This supports the notion that in recent decades, the official unemployment rate in the US has lost some of its validity as a measure for economic prosperity, especially for the low-skilled. If we assume that the drop in prime-age labour force participation is solely caused by workers who would like to work and is representative across the labour force, the unemployment rate would still have to be corrected upwards by about 1.5 percentage points.
On a side note, since the US has by far the highest incarceration rate of any developed country (0.7 percent – most other OECD countries are close to 0.1 percent), it might be interesting to note that while the direct effect of increased incarceration is to increase the participation rate because the official statistics omit prisoners, it indirectly lowers the participation rate since former inmates are often barred from jobs due to hiring restrictions and give up looking for a legal occupation altogether.)
A second way in which the unemployment rate might shroud slack in the economy is by not distinguishing between sufficiently employed workers and workers that would like to work more but are working part-time or in less-qualified jobs than they would like to.
There are still more people working part-time for economic reasons than before the crisis and their share of the labour force is more than one percentage point higher than during the boom years around the year 2000. Unfortunately, I was not able to find data on full-time workers that are overqualified for their current job.
It is worth noting that the BLS also publishes the measure U-6 which captures total unemployed, plus all marginally attached workers, plus total employed part time for economic reasons, as a percent of the civilian labor force plus all marginally attached workers. At the moment, this share is 8.9 percent, one percentage point higher than before the crisis and two percentage points higher than in 2000. However, this measure does unfortunately not capture workers who have been discouraged for more than 12 months or full-time workers who are overqualified for their job. But especially the former group is crucial since they are the ones hit hardest by economic crises.
Considering falling prime-age labour force participation rates and elevated levels of underemployment, there is reason to believe that the official unemployment rate does underestimate the remaining slack in the US economy. This undervaluation can have both economic and political consequences.
If the unemployment rate is used as a measure for the output gap, a deterioration of its meaningfulness can cloud the decision making of the central bank and other economic institutions. However, the case of European countries, where labour force participation rates do not exhibit the same drop as in the US and alternative unemployment measures are used, suggests that the issue of falling labour force participation has not played a vital role in the recent puzzle of low inflation (as I have argued before when discussing US monetary policy). Nevertheless, rising underemployment might indicate an increasing polarisation of the US labour market (see e.g. Autor & Dorn, 2013) that is part of a secular development that hampers inflation.
The political implications might currently be more severe. If official statistics cease to reflect the reality of many workers that have either given up looking for work or are trapped in unsteady and unsatisfying employment, trust in these statistics might be lost and will be hard to restore. If government agencies fail to openly communicate issues with published statistics, these issues can act as the grain of truth in hyperbolic populist claims.
A first step to make US unemployment statistics more transparent could be to include workers who have been unemployed but gave up looking for work for more than 12 months. At the moment, official US statistics offer no information on these workers who can be considered truly left behind. Moreover, it should be considered worthwhile to offer more information not only on the quantity but also on the quality of jobs. For example, an interesting approach to measuring the US labour market situation is Gallup’s Good Job Indicator, derived from asking 30,000 people above the age of 18 each month whether they work 30+ hours per week for an employer who provides a regular paycheck. At the moment, 44.9 percent answer this question with “yes”.
It is likely that the same forces that have led to higher inequality (technological change, globalisation, changes to institutions such as taxes and unions) have also pushed workers into more unsteady jobs or completely out of the officially registered labour force. This development does not show up in the official statistics on unemployment, complicating economic decision making and undermining trust in official statistics in general. It is desirable that in the future more information on the US labour market situation is provided and emphasised.
Aaronson, D., Davis, J., & Hu, L. (2012). Explaining the decline in the US labor force participation rate. Chicago Fed Letter, (Mar).
Autor, D. H., & Dorn, D. (2013). The growth of low-skill service jobs and the polarization of the US labor market. The American Economic Review, 103(5), 1553-1597.
Black, S., Furman, J., Rackstraw, E., Rao, N. (2016) The long-term decline in US prime-age male labour force participation. Voxeu.org
written by Jonas