Is lower US unemployment too good to be true?

The US unemployment rate has declined rapidly since October 2011, giving the impression that the labour market is improving much faster than was generally expected. If so, this will have significant implications for corporate profits, the federal budget deficit and US monetary policy. It could even have important implications for November’s presidential and congressional elections.

However, we believe the recent fall in the unemployment rate exaggerates the underlying improvement in the labour market. The main reason is the continuing decline in labour-force participation: a smaller percentage of the working-age population is participating in the workforce than a few years ago.

Indeed, not only is the percentage smaller, about 0.5 per cent fewer people were in the labour force at the end of 2011 in absolute numbers than at the end of 2008 – even though the population of working-age people increased 2.4 per cent over that period.

As participation declines, it pushes the unemployment rate lower for any given level of employment. In essence, the rate can decline, not because more workers are employed, but because some have left the labour force and are no longer reported as unemployed.

The participation rate measures the proportion of people aged over 16 who are working or looking for work. Participation rates vary with age and gender. They also vary over time within each age and gender classification. For example, the participation rate of young people has been falling fairly steadily since 2000, while the rate for people over 55 has been climbing since the mid-1990s.

The recent fall in the unemployment rate exaggerates the underlying improvement in the labour market.

These trends reflect changes in educational attainment and the types of jobs on offer. Younger people are spending more time in higher education so now stay out of the labour market longer. And more white-collar jobs mean older people can work longer and increase their lifetime earnings.

The shift in population away from the middle age brackets toward the younger and older extremes means the average labour-force participation rate has been on a steady downward trend since 2000. However, something happened after 2008, when the decline suddenly became much steeper than can be explained by demographic factors.

Labour-force participation is also affected by the ups and downs of the business cycle. Recessions cause people to drop out of the workforce while economic expansion means more people return to work and the participation rate rises. Following the 2001 recession, for example, the participation rate fell, then began to recover gradually, rising slightly above the trend rate by 2006.

In 2011, the unemployment rate averaged 9 per cent, or about 3.5 percentage points higher than the 5.5 per cent ‘natural rate’ estimated by the Congressional Budget Office. With an unemployment rate that high, the participation rate should have fallen from 65.5 per cent to about 65 per cent, but instead, it averaged 64.1 per cent in 2011, nearly a full percentage point lower than could be explained by the effects of the recession.

If the participation rate had stayed closer to 65 per cent, the labour force would have been larger by about 2.1 million people. And if all these potential workers were counted as unemployed, the unemployment rate would have been about 10.2 per cent, rather than the 9 per cent reported for the year.

Why have many additional people left the labour force over the past few years? The answer may lie in some of the structural changes going on in the economy. The participation rate for men has fallen much more than that for women. This is true for all age groups and all levels of educational attainment. Structural factors in the latest recession and the subsequent recovery may have affected job opportunities for men much more than for women, at least up to this point.

One obvious area of structural change is in construction, a mainly male industry. House building always declines in recessions but usually stages a strong rebound after they end. That has not been the case for the past two years. New housing starts in 2011, though up from 2010, were still less than half the level attained in 2007, the year before the recession started. Roughly three million fewer construction workers were employed in 2011 that represents a fall of 24 per cent since 2007. Unemployment among these workers rose from 6 per cent in 2007 to a high of 18 per cent by mid-2010 and has since fallen to 12 per cent – even though fewer people were employed in the construction industry in 2011 than in 2010.

The average labour-force participation rate has been on a steady downward trend since 2000.

The unemployment rate in construction has been falling, not because more workers are finding jobs, but because workers are leaving the industry. They could be leaving for jobs in other sectors, but for many workers, the chances of finding a job outside of construction are low. Most workers leaving construction are probably leaving the labour force entirely. So the decline in the unemployment rate in the sector does not mean the labour market has improved or that the market is tightening in any meaningful sense.

Since 2008, the male labour-force participation rate has dropped by nearly three percentage points. This rapid decline in participation has contributed to a decline in the measured unemployment rate among male workers. Note, however, that the male participation rate stabilised in 2011 at roughly 70.5 per cent.

If the participation rate holds steady or even begins to rise modestly in response to better labour-market conditions, then the downward pressure on the unemployment rate from lower participation should start to fade. This might occur if the ‘structural’ forces that have pushed the participation rate down begin to diminish. If so, the size of the labour force will start to grow in line with the growth of the working-age population. As it does, the unemployment rate will no longer shrink just because people are dropping out of the labour market.

And there may be structural changes in the economy forcing a sharper decline in labour-force participation among women. The recession in 2008-09 initially had less of an impact on female participation, compared with men but lately, the female rate has been falling faster, to about 58 per cent, even as the male rate stabilised around 70.5 per cent.

Female workers hold a disproportionately high proportion of jobs in education, child care, and local government administration, and large numbers of job cuts in these sectors began after the recession ended in 2009. With diminished prospects of finding suitable jobs, they may choose to retire early or even stop looking for employment altogether. So as they leave the labour force, the female participation rate, and hence the unemployment rate for women, will fall more than it might otherwise.

The unemployment rate, particularly given continuing structural shifts in different sectors of the economy, in our view is thus not the most accurate measure of labour market conditions.

This report must be read with the disclosures, analyst certifications and the disclaimer.