Where are the workers?

WHEN America's job market isn't disappointing, it's perplexing. Take the month of January, when a net figure of just 36,000 non-farm jobs was created. Even by the miserly standards of the current recovery, that was low.

But in the same month the unemployment rate tumbled to a 21-month low of 9% from 9.4% in December. Its drop since November, when it was 9.8%, is the largest over two months since the prosperous days of 1958. Which of these bits of news gives the truer picture?

One reason that America's jobs figures often send mixed signals is that they are drawn from two separate surveys. In this case, the survey of employer payrolls was almost certainly affected by snowstorms that ravaged much of the east of the country and which economists estimate may have kept 60,000 to 150,000 people away from work.

The survey of households, from which the unemployment rate is derived, tells a far more optimistic story. It finds that employment has surged by 882,000, or 0.6%, in the past two months, after adjusting for new estimates of America's population. This is more in keeping with other data, such as car sales and GDP, which suggest that the recovery is picking up steam.

Yet for several reasons the fall in unemployment should not be taken as evidence of a job market on a roll. First, it was inexplicably high to start with. According to Okun's law, an economic relationship that predicts how much unemployment will rise according to how GDP is doing, unemployment should be only about 8.5%, reckons Alan Krueger, a labour economist at Princeton University. So some of the latest drop may simply be an overdue statistical correction.

More troubling, unemployment in the past two years would have been even higher but for a disturbing lack of growth in the labour force: that is, the number of people either working or searching for work. Its growth has been slowing for some time because of the ageing population. After expanding by 1.3% a year in the 1990s, it grew by just 1% from 2000 to 2010, and the Congressional Budget Office predicts that over the next decade it will grow by only 0.7%.

Since the start of 2008 the labour force has not grown at all. One reason appears to be reduced immigration, both legal and illegal, because of diminished job opportunities and more effective border enforcement. In each of the past four years the Census Bureau has had to revise down the estimated Hispanic population, leading to smaller estimates of the overall labour force.

Not only is the population growing more slowly, the share of it in the labour force (that is, either working or looking for work), known as the participation rate, has also fallen. That rate commonly falls during recessions when some of the unemployed give up the search, go back to universities or training colleges or retire early. The decline usually reverses during recovery. Not this time: since the recession ended in mid-2009, the participation rate has kept on sliding (see chart). It has fallen most among the young, many of whom have stayed in education, and least among those over 55.

Most striking has been the drop for men aged 25-54, who have long had the highest participation rates. Some of these men will re-enter the labour market when the economy and job opportunities revive, but many will not. The participation rate of men has been declining for years, apparently because many who lost their high-paid, low-skilled jobs in manufacturing, transport and construction have retired or registered as disabled rather than retraining. Julia Coronado, an economist at BNP Paribas, reckons a wave of early retirements by state- and local-government staff and manufacturing workers may explain the latest downturn. If these men never rejoin the job hunt, it would, paradoxically, help to bring the unemployment rate down faster. That might look like good news; but it is not.