Weakening homebuilding cuts into construction worker employment numbers

Jan. 1, 2008
Although job losses in construction appear subdued in comparison to the sharp declines in some of the other housing indicators, homebuilding has continued

Although job losses in construction appear subdued in comparison to the sharp declines in some of the other housing indicators, homebuilding has continued to weaken, with starts, permits, and construction spending all falling, according to information from the Department of Labor.

From 2004 to 2006, construction was among the leading contributors to job growth, with employment expanding at over triple the rate of total nonfarm payrolls as measured by the Current Employment Statistics (CES), according to information from the Department of Labor.

Fueled by a boom in housing, the industry had added over a million jobs since its most recent trough in 2003. In 2006, the housing boom ended as homes sales stalled and the number of unsold homes rose. Construction employment peaked in September 2006 and then fell by 112,000 over the year. Residential construction has borne the brunt of the slowdown, shedding 192,000 jobs since March 2006.

In August 2007, the number of single family units under construction was 22.4% below the previous year's level, and residential construction spending was down 16.0% over the year. Throughout the recent turmoil in housing, the downturn in construction spending has been confined to the residential segment, which accounts for just under half of total construction spending.

Expenditures on nonresidential building and public-sector construction have continued to expand throughout the downturn, with annualized spending increasing by 14.7%, as of August 2007. Overall, construction spending has only declined 1.7% over the year ending in August.

CES employment has broadly mirrored construction spending, if with less pronounced movements. The residential construction components have seen the largest job losses, 192,000 since March 2006, with modest job growth across the nonresidential components tempering total construction job losses.

Shallow declines

The shallow declines have led to speculation that perhaps CES has not accurately captured construction employment. In fact, since the most recent benchmark in March 2006, construction employment in the CES series has closely mirrored the QCEW universe employment counts.

Preliminary benchmark figures for March 2007, released October 5, 2007, indicate a downward benchmark revision of 8,000 or 0.1% for construction. The Labor Department said it is possible that some firms classified as primarily engaging in residential construction have taken on more nonresidential construction projects, as activity in the two markets has diverged.

Additional statistics provide a basis for comparison of recent trends in construction. The Current Population Survey (CPS), a survey of households, produces monthly construction employment estimates.

Since January 2006, CPS private-sector construction employment has broadly emulated the CES construction series, with staffing at construction firms remaining steady over the year. Similarly, the unemployment rate among construction workers has remained essentially unchanged during this period — 5.3% in August 2007 compared to 5.9% in August 2006. Together these measures are in line with the relative flatness reported in CES construction employment.

Immigrant workers

“It is unclear as to what extent construction employment is being impacted by the industry's large number of immigrants, some of whom are assumed to be undocumented,” the report said. In 2006, 28% of all workers employed in construction and extraction occupations were foreign born.

It is likely that many of these workers are included in the payroll survey. However, the exact number is undetermined as CES does not collect demographic information.

BLS requests the number of persons employed by a given establishment, regardless of nationality or legal status. While BLS surveys are confidential and used only for statistical purposes, firms may be reluctant to report payroll data on undocumented workers to a government agency. Thus hiring and layoffs in construction may be larger than reported in the payroll survey due to exclusion of such workers.

Another dynamic affecting construction employment is the large number of temporary workers in the industry. Temporary workers allow firms to scale production up or down to meet demand, without incurring the costs associated with hiring or laying off permanent workers.

When choosing contract workers, construction firms have three options: hire temporary workers directly, subcontract from a third party, or subcontract with self-employed individuals.

Under CES methodology, directly hired temporary workers are indistinguishable from permanent employees. Workers from a third-party employment agency would be classified according to the industry in which the contract agency operates. A temporary help firm specializing in construction would be included in construction, while a diversified temporary help firm would be counted in the employment services industry.

According to data from the household survey, approximately one in six individuals working in construction is self-employed. That is over twice the average across all nonagricultural industries.

Much like temporary contract workers, self-employment is considered to be more sensitive to economic fluctuations. CPS data indicate that self-employed construction workers have been impacted, albeit modestly, by the slowdown in housing. In the first half of 2007, selfemployment in construction was down 0.5% from the prior year.

Through the March 2007 benchmark reference period, CES construction employment has trended in line with the universe employment counts derived from UI records. Over the past year ending in September 2007, CES has recorded a construction employment decline of over 110,000.