On March 24, 2015 Associated Builders and Contractors (ABC) launched its state-by-state economic analysis with the release of economist Bernard Markstein's analysis of construction's contribution to each state's gross domestic product. Markstein also produces monthly analysis of the construction job market and unemployment rate in each state, background on state-level construction job markets is below. The state level analysis is produced monthly in addition to ABC's existing national economic data and analysis.

By Bernard M. Markstein

Measuring Employment, Unemployment, and Unemployment Rates
Measuring employment and unemployment would seem to be straightforward. However, it turns out to be a little more complicated than it first appears. The Bureau of Labor Statistics (BLS) produces the employment statistics for the United States based on a monthly survey conducted by the Census Bureau known as the Current Population Survey (CPS).

Each month, the Census Bureau surveys roughly 60,000 households (about 110,000 individuals). To be included in the survey panel, a person must be 15 years of age or older and not in the Armed Forces or in an institution, such as a prison, long-term care hospital, or nursing home (however, employment statistics are only reported for individuals 16 or older). The survey is conducted during the week that contains the 19th of the month. This is designated the survey week. Survey questions refer to activities that occurred during the prior week (i.e., the week that included the 12th of the month). This is designated the reference week. The BLS then uses the CPS survey results to produce its estimate of U.S. employment and unemployment.

A person is considered employed if they worked during the reference week. No distinction is made between full-time and part-time work—although this observation is tracked and used elsewhere in the employment data. A person is considered unemployed if they did not have a job during the reference week but actively looked for work during the four weeks prior to the survey week. The labor force is equal to the number of employed people plus the number of unemployed people. People who were not employed nor actively looking for a job are considered not in the labor force. The unemployment rate is defined as the number of people unemployed divided by number of people in the labor force.

In addition to this official unemployment rate, the BLS produces five other unemployment rates, such as a rate that includes discouraged workers. Discouraged workers are people who did not actively search for a job during the four weeks prior to the survey but did look for a job sometime during the prior 12 months and would re-enter the labor market if job market conditions improved.

Construction Unemployment
Determining employment by industry is relatively easy. Starting with the definition of whether a person is employed, it is simply the industry in which a person is employed. There is a complication if the person works two or more jobs in different industries—e.g., a person who works at a restaurant and has a second job as a taxi driver. The BLS solves this problem by placing the person in the industry in which they worked the most hours.

People who are unemployed are considered unemployed in the industry in which they were last employed. New entrants (i.e., people who have not been previously employed) are not assigned an industry. The unemployment rate for an industry is the number unemployed in that industry divided by the labor force for that industry. New entrants seeking employment are not included in the industry unemployment figures regardless of where they are focusing their job search activity. However, they are included in the general unemployment statistics.

Based on the CPS, the BLS produces a monthly unemployment rate for major industries, including the construction industry. These industry unemployment rates are available at the national level on a monthly basis, but not at the state level. (State unemployment rates for major industries are available as annual averages with a considerable lag.)

Using a model that incorporates state employment data, we produce monthly estimates of state construction unemployment rates. Each month the estimates for the state construction unemployment rates are updated as state employment data are revised and new data become available. As a result of revisions in the state employment data, the state construction unemployment rates for the months in which the state data are revised can change.

National and Construction Industry Unemployment Rates
Both the national unemployment rate and the national construction unemployment rate are an estimate based on the results of the CPS. Thus, each estimate of the unemployment rate is subject to sampling error. Given the smaller sample for the construction industry (a subset of the national CPS data), the construction unemployment rate is subject to greater variation and error.

The variations in the unemployment rates from month to month may not indicate changes in the true rates (i.e., the rates that would result from accurately surveying the entire U.S. population) but may be due to sampling error. However, over time, sustained movement either up or down would reflect labor market conditions for the entire United States population. Economists use unemployment rates, with other labor force and economic data, to draw conclusions about the health of the labor market, the economy, and the construction industry.

The national construction unemployment rate is reported on a not seasonally adjusted (NSA) basis. Therefore, the construction unemployment rate generally increases in the fall and winter, and decreases in late spring and summer. Consequently, month-to-month comparisons should be approached with great care. Year-over-year comparisons (comparing the same month a year apart or several years apart) are a much better approach. To better compare the construction unemployment rate with the overall unemployment rate, the tables in the articles reporting on the monthly state construction unemployment rates use only NSA data.