One of the things you can always bet on with surety is that the conservatives will always try to convince the public that a cyclical event is, in fact, a ‘structural’ event. This has two, linked purposes. First, they can downplay any hint that aggregate fiscal policy interventions, which work at the macroeconomic level are necessary no matter how bad the problem is. Second, they can then wheel out their favourite ‘structural’ remedies, all of which just happen to result in national income being distributed to profits or high income earners, less capacity for low-wage workers to enjoy real wage rises or reasonably share in national productivity growth, and lower government income support payments to the disadvantaged. A double-whammy strategy. Here is an example of that sort of lie. The US Employment-Population ratio has fallen dramatically since the onset of the crisis and remains stuck at low levels. The reason is clear – there was a huge collapse in employment in 2008 and 2009 and, in the recovery, the rate of job creation has not been sufficiently strong to reverse that decline. Total employment growth has been around or just above the underlying growth in the civilian population (above 16 years), which is why the ratio is stuck. There needs to be much faster employment growth for the US to make back the ground that was lost in the downturn. While the US civilian population is growing older and that is having an impact on the calculated Employment-Population ratio, the impact is small and doesn’t alter the fact that a huge negative cyclical event occurred in the US and the fiscal intervention was not large enough to fix the problem.
There was a good technical discussion of the Employment-Population Ratio in the February 1981 edition of the US Bureau of Labor Statistics (BLS) journal the Monthly Labor Review.
[Reference: Leon, C.B. (1981) 'The Employment-Population Ratio: its value in labor force analysis', Monthly Labor Review , February 1981. http://www.bls.gov/opub/mlr/1981/02/art4full.pdf]
We learn that the unemployment rate is the “best known” summary measure of labour market conditions but is sensitive to shifts in the participation rate. The unemployment rate is the ratio of total measured unemployment as a percentage of the labour force. The labour force can fall in a recession because of the discouraged worker effect and so official unemployment is replaced by ‘hidden unemployment’.
Using the population as the basis for the ratio is better in this sense because the population is invariant to participation rate swings that are of a cyclical nature. The hidden unemployment leave the measured labour force but remain within the population.
The Monthly Review paper provides additional reasons for preferring the Employment-Population Ratio as a more accurate indicator of the state of the labour market. You can follow up if you are interested.
The BLS defines the Employment-Population ratio is the proportion of the civilian noninstitutional population aged 16 years and over that is employed.
The following graph shows the Employment-Population ratio for the US from the first-quarter 2004 to the first-quarter 2014. The drop off the cliff is symbolic of the extent of the labour market collapse in the US.
The Employment-population ratio also peaked (in terms of the period leading up to the crisis) in the month of December 2006 – 63.4 per cent compared to the dismal 58.9 per cent in the month May 2014.
The next graph decomposes the employment shifts into age groups. The blue bars are the absolute change in employment (EMP is total) between the first-quarter 2004 to the first-quarter 2014 and the red diamond is the same change in percentage terms (noting the zero point on the left- and right-axes do not line up.
The older workers (GT55 = above 54 years) have increased their employment by 20.2 per cent over the six-year period while total employment has fell. The big losers in percentage terms have been America’s teenagers, which is a world-wide trend.
In an environment where there are not enough jobs, the least-skilled and more easily vilified (ethnicity, disability, etc) will be shuffled to the end of the jobless queue.
The depth of the current malaise in the US is also seen when we examine the shift in participation rates. The maximum participation rate in the US was 67.3 in the first three months of 2000. The most recent (pre-crisis) peak has been 66.4 in December 2006-January 2007. The current participation rate (May 2014) is 62.8. So participation is now 3.6 percentage points below that pre-crisis peak indicating a massive withdrawal of the workforce.
One might be tempted to say that was due to a surge in retirements associated with the ageing population, which is the argument used by those who wanted to downplay the cyclical nature of the collapse.
But the following graph suggests otherwise. It shows the change in labour force participation rates between the first-quarter 2008 to the first-quarter 2014 in percentage points.
Consistent with the growing employment for the over 55-years cohort in the previous graph, their participation rate has risen. All other age groups have experienced a fall in their participation rate, none greater in percentage point terms than the teenagers.
So not only are the 16-19 year olds losing out in employment significantly, they are also dropping out. Not a good sign for America’s future. That is also a common trend in the job-starved advanced economies and a principal casualty of the austerity obsession. These nations will live to regret this myopia.
The last thing to do is see whether the ageing population is the reason the Employment-Population ratio has plunged so far and is not improving. The argument is if more people with a lower ratio are now in the population then you can see a decline in the ratio because of that compositional shift alone with no intrinsic changes in each cohorts fortunes. That is easy to test.
First, there is ageing in the US population. The following graph shows the weights in the civilian population above 16 years for the first-quarter 2008 (blue bars) and the first-quarter 2014 (grey bars).
The following graph shows the shifts in the Employment-Population ratio between the first-quarter 2008 and the first-quarter 2014 for the different age-groups in the civilian population above 16 years. It is quite clear that the fortunes of each demographic group have been quite different.
Finally, we can see what the Employment-Population ratio would be if there had been no change in these weights by computing a fixed-weight ratio (using the weights as at the first-quarter 2008).
The following graph shows the current-weight ratio (using the actual proportions of each demographic group) and the fixed-weight ratio (holding the first-quarter weights constant). The two time-series are indexed to 100 as at the first-quarter 2008.
There is some impact of the changing demographics apparent. The decline in the fixed-weight Employment-Population ratio from the first-quarter 2008 to its trough is of the order of 5.8 percentage points whereas using current weights the decline was 6.6 percentage points (as at the last-quarter 2009).
The current difference between the two series is around 1.5 percentage points or about 25 per cent of the total change in the Employment-Population ratio from the first-quarter 2008 to the first-quarter 2014.
The predominant reason the Employment-Population ratio has fallen so much and is struggling to recover is the shortage of jobs rather than the shifting age-composition of the civilian population.
Whenever you see a rapid surge or decline in an economic time-series that is cyclically variable you can be almost certain that the cause is cyclical in nature. Structural shifts work over longer periods of time.
The point is that structural change is on-going but should never divert the policy debate from dealing with the cyclical issues.
The neo-liberals try to muddy the waters by introducing these structural red-herrings. The big return on policy would be for the government to directly stimulate employment to push the ratio up instead of wondering about the ageing civilian population.