Why do economists tend to look at employment and not earnings when both are important? Do they ever look at both? What does it matter if employment is high if everyone is earning only the minimum wage?

July 1, 2007

Both employment and earnings are important labor market indicators and economists look at both. As discussed below, employment numbers exhibit a stronger pattern over the business cycle than earnings. Additionally, average earnings are more difficult to report and the usefulness of the data is limited. A situation where employment is high while everyone is earning a minimum wage would not be a likely scenario – as discussed below, the share of minimum wage jobs in overall employment is rather low and has been declining over time.

Let’s look at some of the data economists use to analyze both shorter- and longer-term trends by reviewing the payroll employment and earnings data collected and published by the Bureau of Labor Statistics (BLS).

Payroll employment and earnings data
Nonfarm payroll employment (the number of jobs) and average hourly earnings data are collected monthly from the Current Employment Survey (CES) or the “Establishment survey” by the BLS. This national survey covers payroll jobs reported by a sample of about 160,000 business and government establishments with over 400,000 worksites.  In total, the CES sample is so large that it includes roughly one-third of all nonfarm payroll workers.

Timeliness is an important feature of the payroll employment and earnings data, and a key reason why the data are used frequently by economists for analyzing current conditions in the economy is that the BLS releases employment data for each month on the first Friday of the following month. The publication of these data with only a few days lag provides an important initial signal on labor market and economic developments. The payroll employment and earnings data also are important because they can provide clues to movements in other economic data series, like personal income and industrial production, and thus are used to help estimate those series.  

Because of its extremely large sample size, movements in the payroll employment data closely reflect overall labor market and economic conditions.1 Average hourly earnings data generated from the survey provide information on earnings trends and wage inflation in labor markets. However, average earnings data are more complicated to report for firms than is the number of employees—for example, calculations depend on both outstanding payrolls and the number of workers. Moreover, as the BLS notes, there are several characteristics of the data that limit their usefulness for analysis. Average earnings estimates may significantly underestimate average full-time worker earnings because the number of workers includes both full-time and part-time workers, as well as workers with multiple jobs. Other issues are that estimates of annual earnings may not match estimates from shorter periods because turnovers and layoffs reduce workers’ earnings part of a year, and that earnings only cover production workers for goods-producing industries and non-supervisory workers for service-providing jobs.2

Following the jobs data over the business cycle
For evaluating short-term economic developments, economists tend to look at payroll employment numbers because payrolls move so closely in line with the overall economy and with the business cycle. For example, the Conference Board classifies total nonfarm payroll employment as a coincident economic indicator and the National Bureau of Economic Research (NBER) uses total nonfarm payroll jobs to help date business cycles.

Chart 1:  Employment tracks the business cycle more closely than earnings

Chart 1: Employment tracks the business cycle more closely than earnings

Chart 1 shows three series to help explain why economists might focus on payroll employment rather than the monthly earnings series as an indicator of current labor market conditions. Over the 30-year period from 1977 to 2007, Chart 1 shows the number of total private-nonfarm payroll jobs (thick red line, left axis) and real (inflation adjusted) average hourly earnings in dollars for total private nonfarm payroll jobs (medium blue line, right axis), and nominal (not adjusted for inflation) average hourly earnings for total private nonfarm payroll jobs (thin green line, right axis). The private employment number is total nonfarm payroll jobs less government sector jobs.  

The employment number exhibits a strong pattern over the business cycle, which is shown here as periods of expansion (white) and periods of recession (gray bars). During expansions, the economy added jobs on a fairly steady upward trend. However, during each of the four recessions, the number of jobs declined.

In addition to the jobs and earnings data, economists do monitor many other labor market indicators, including, but not limited to, the unemployment rate, hours worked, new claims for unemployment insurance, layoffs, duration of unemployment, labor force participation, and even help wanted advertising. The Federal Reserve Bank of St. Louis publication, National Economic Trends, includes several charts on developments in a variety of U.S. labor market indicators.

Real and nominal earnings and the minimum wage
Let’s take a quick review of your question about minimum wage jobs, but first, notice that two estimates of average hourly earnings are shown in Chart 1. Real or inflation adjusted earnings, using 1982 as a base year, are shown by the medium blue line. The thin green line represents nominal earnings or earnings that are not adjusted for inflation. The nominal earnings series has risen much faster than real earnings over this period, indicating that most of the increase in wages over the 30-year period resulted from inflation, not an increase in real wages. Now that we have described both real and nominal earnings, we can move to your question about jobs and the minimum wage. 

To investigate into the share of minimum wage jobs over time, we can use BLS data that track the share of hourly wage jobs that are paid at or below the minimum wage. That figure has been trending down. In 1980, 15.1 percent of all hourly jobs were paid at or below the minimum wage. In 2006, only 2.2 percent of hourly workers—about 1.7 million out of 76 million—were paid either the minimum wage or below.    

We also can compare federal and state minimum wages to the nominal earnings data collected from the CES survey. The nominal earnings data for total private-sector jobs indicate that average wages are well above the minimum wage. As of July 2007, average nominal earnings per hour for total private sector jobs was $17.47 per hour, nearly three times the prevailing federal minimum wage of $5.85 an hour that took effect on July 24, 2007.3  Moreover, nominal average earnings per hour as of that same date were much higher than the minimum wage for both goods-producing and service-providing jobs, at $18.69 and $17.15 respectively.4 To see how nominal average hourly earnings vary by industry sector, you can review the BLS data on “Average hourly earnings of production and nonsupervisory workers (1) on private nonfarm payrolls by industry sector and selected industry detail, seasonally adjusted.” These data do show considerable variation in hourly earnings across sectors, with utilities showing the highest earnings, at $27.82 per hour for September 2007, while leisure and hospitality employment, at $10.46 per hour, recorded the lowest average hourly earnings. Even the latter, though relatively low, is well above the current federal minimum wage.


1. Payroll jobs data also show much less variation or noise on a month-over-month basis than does the household employment series, an alternative measure of employment at the national level.

2. For more information, see the BLS’s technical note on real earnings.

3. Many states have their own minimum wages laws; they can be viewed at the U.S. Department of Labor’s website.

4. Data reported as of March 20, 2008.