Does the Recent Trend in Labor Demand Presage Recession?
The number of job openings or vacancies posted by employers constitutes a good measure of unmet labor demand. Assuming employers spend some time and resources to recruit workers, this measure could give us a nice clue about their expectations of future labor market conditions.
The longest time series of vacancies that we have is the Help-Wanted Advertising Index (HWAI) provided by the Conference Board. This index is monthly and tracks help-wanted ads in more than 50 major metro area newspapers. HWAI is normalized to 100 for 1987. A higher index value indicates higher numbers of help-wanted ads are appearing in newspapers.
The HWAI experienced sharp declines in every postwar recession. More interestingly, every decline in the index has been accompanied by a recession, with the exception of the mid-1960s. After hovering in the 40s for most of the 2003–2005 period, the index started to fall gradually at the beginning of 2006. In January 2008, the index hit 21, its all-time low.
Ironically, the index by itself may not be very informative about the difficulty employers have in filling positions, because that difficulty depends not just on how many vacancies there are, but also on the number of workers who are looking for jobs. For instance, the index could be low (indicating few vacancies), but employers could expect to fill vacant positions relatively easily if many unemployed people are searching for work.
In order to assess employers’ difficulty in finding workers, we need measure of market tightness, which we have in the ratio of help-wanted newspapers ads to the number of unemployed workers. Movements in this ratio closely follow those of the HWAI. During expansions, both market tightness and the HWAI rise, and during recessions, they both decline. Recent labor market conditions, according to this measure, have been exceptionally slack. Currently, the ratio stands at 0.205, the lowest it has ever been.
However, the declining trend in these measures might be related to factors independent of labor market conditions. In particular, a shift toward posting vacancies online rather than in newspapers could be responsible for it. The Conference Board started to gather and report data on online help-wanted ads in May 2005. Although this series is not long enough to cover a full business cycle, we still see that vacancies, as measured by online ads, have grown from about 3.1 million to more than 4.3 million in two years (May 2005–May 2007). These numbers suggest that the HWAI might be understating the true availability of jobs in the labor market. In addition to tracking the number of help-wanted ads posted online, the Conference Board also tracks how many of those postings are new. It would be fair to assume that movements in total help-wanted ads are driven by the new postings every month. However, the raw data captures a lot of seasonal movements. When we look at year-over-year changes in online ads to remove this seasonality, we see an increasing trend in job postings until mid-2007, after which postings decline. In February 2008, total new ads increased by only 103,000 relative to February 2007, the smallest year-over-year increase since May 2006.
One other major source of data on job availability, and one that is more comprehensive than the HWAI, is the Job Openings and Labor Turnover Survey (JOLTS) published by Bureau of Labor Statistics. It samples from the same universe as the Current Employment Survey, and each establishment in the Survey provides data on job openings in a given month.
The picture painted by JOLTS data confirms the view that the HWAI might be understating the actual availability of jobs. According to JOLTS, employers were creating more and more vacancies every month up until mid-2007. Since then, the trend seems to have reversed. According to the most recent data, there were about 3,925,000 job openings in January 2008.
Four industries accounted for almost two-thirds of the total monthly job openings on average—education and health services; professional and business services; trade, transportation, and utilities; and leisure and hospitality. All sectors roughly follow a similar pattern over time, although three sectors experienced larger declines in response to the last economic downturn: professional and business services; manufacturing; and trade, transportation, and utilities. Even though job openings have leveled off recently in these sectors, we have not observed a decline similar enough the one observed at the onset of the last recession to indicate a significant slowdown in the labor market.
Overall, different measures of job availability all suggest that the number of new job vacancies advertised might be falling. Total job openings are still far from their pre-recession peak of 4,580,000 (in December 2000), which is consistent with the last recovery’s designation as a “jobless” one.