Reading the vital signs in the jobs report
By Jared Bernstein Lawrence Mishel 04-03-08

April 3, 2008 | EPI Issue Brief #243

Reading the vital signs in the jobs report

by Jared Bernstein and Lawrence Mishel

The Bureau of Labor Statistics release of its jobs report on Friday, April 4, provides an important set of labor market vital signs, which are key indicators of our overall economic health and the living standards of the vast majority of working families who rely primarily on wage income. This issue brief outlines some of the most important indicators and benchmarks from the monthly jobs report, what they mean for families' economic well-being in the current downturn, and the longer-term trends that give them context. As always, EPI will carefully monitor and report on these indicators every month on the morning the data is released.

Income recovery incomplete
First, it is important to recognize that this is the first recession and recovery cycle on record in which median family income has not regained the ground lost since the last downturn. According to our analysis of Census Bureau data, the income of the typical (median) working family was lower in real terms in 2007 than at the end of the last business cycle, in 2000.1 Given that family incomes will almost certainly decline in 2008, the typical family's income—which failed to recover over this 2000 cycle—will now begin a new descent, as unemployment rises throughout 2008 and into 2009 (as is predicted by virtually every forecast).

Weakening labor market causes broad-based income declines
Based on past experiences with how weakening job markets effect income growth, the rise in unemployment in 2008 will cause middle-class family incomes to fall by about $750, on average. The reasons for the decline include:

• Higher unemployment;
• Fewer hours worked per week;
• Fewer people employed per family;
• Declining rate of wage growth as unemployment rises;
• More people working part-time involuntarily;
• Higher underemployment (a monthly measure based on broader, more inclusive data on labor utilization).

Understanding the indicators
The BLS releases data on various indicators that are important to take into consideration when gauging the health of the labor market. They include:

1. Unemployment and net job gains/losses
2. Jobs in the private sector
3. Employment rate and work hours
4. Weekly work hours
5. Wage momentum
6. Long-term unemployment

1. Unemployment and net job gains/losses
The most highly watched numbers from the jobs report are payroll employment growth and the unemployment rate. A related indicator that gets a lot of attention is whether the report meets forecasters' consensus prediction of payroll growth. In March, for example, the consensus is that payroll jobs contracted by 50,000.

Whether job growth meets analysts' expectations is important, but even more important, especially in a weak period like the present, is whether enough jobs are created to absorb new workers entering the labor force while keeping the unemployment rate from rising. Our calculations find this "steady-state" number is 127,000: that is the number of payroll jobs that need to be created each month to forestall unemployment from growing higher (see the appendix for description of how we arrived at this number). In other words, at least 127,000 new jobs per month are needed just to maintain the equilibrium in unemployment.

It is possible for unemployment to decline in any given month, even if payrolls shrink. In fact, this occurred in February, when payrolls declined by over 60,000 and unemployment fell from 4.9% to 4.8%. But the reason the jobless rate went down was because hundreds of thousands of workers left the job market, were no longer actively seeking employment, and thus were not officially counted as unemployed. The point is that these monthly unemployment rate changes are volatile and can be driven by job loss and labor market exit and entry. Our benchmark of 127,000 should be viewed as the number of payroll jobs needed simply to keep unemployment from rising, assuming the labor force is not shrinking.

2. Jobs in the private sector
Private-sector jobs changes, another important indicator, tell us how quickly the economy is slowing down. In the last three months, private-sector job growth declined.

3. Employment rate and work hours
A key factor in raising middle class incomes over the last few decades has been that more family members work, and they work longer during the year (i.e., more weeks per year and/or greater number of weekly hours). As the labor market weakens, this process reverses itself. One can trace this by tracking changes in the employment rate (the ratio of employment to population). This ratio is important to watch because other measures, such as the unemployment rate or the labor force participation rate, depend on how people are classified as in or out of the labor force.

In February, the employment rate and the unemployment rate both declined. In this situation the decline of the employment rate provided a better indication of labor market weakness than the unemployment rate. In fact, it is well known that the employment rate has fallen since 2000, and that has led to an understatement of the growth of unemployment.

4. Weekly work hours
Another important measure to follow is weekly work hours. This number falls in a downturn as less overtime work is available, more people are employed part time, and employers cut back on hours as the demand for their goods and services falters.

5. Wage momentum
The hourly wage data provided in the monthly BLS report are the only monthly measure of wages that can be used to track shifts in the wage trajectory, specifically for the roughly bottom 80% of the workforce captured by the BLS measure of production/nonsupervisory workers' wages. Recently, the rate of nominal (i.e., not inflation-adjusted) wage growth, measured year-over-year, peaked in March 2007 at 4.2%. By February 2008, that rate had slowed to 3.7%, a 0.5% deceleration. Note that inflation has grown more quickly over this period, due largely to rising energy and food costs. Real hourly wages have thus been flat or negative since October 2007.

6. Long-term unemployment
An issue that arises in each downturn is whether, when, and how to provide extended unemployment insurance (UI) to the long-term unemployed, that is, those who have been jobless for 27 weeks or more. The logic is that in a downturn those who do not find jobs are facing unusual problems because of the pervasive weakness in the job market, and thus need further income support. The federal government has provided extended unemployment insurance benefits in every recession since the 1970s, but Congress usually waits too long, letting unemployment, and especially long-term unemployment, rise substantially before it acts. The 2001 recession actually ended before Congress managed to extend benefits.2 As of this February, there were already about the same number of long-term unemployed—1.3 million—as in March 2002 when Congress finally extended unemployment compensation.

Appendix: Calculating the jobs needed to prevent a rising unemployment rate

To come up with the "steady-state" number of payroll jobs needed to keep the unemployment rate stable, we used as our benchmarks the labor force participation rates and unemployment rate in the fourth quarter of 2007 (66% and 4.8%, respectively). The question is then: given population growth, how many jobs are needed to keep the unemployment rate at 4.8%?

We took the projected growth of the working age population—1.1% in 2008—from the most recent Social Security annual estimates. (We used the "intermediate" estimate; note that the Census 10-year projections yield the same rate.)

We then make one further, important adjustment. The labor force statistics used thus far in this exercise come from the household survey, but our benchmark should apply to the establishment survey. To adjust for this difference, we take advantage of the BLS publication [PDF] of household survey employment using the payroll survey definitions. For 2007, the ratio of payroll/household jobs was 95%, and we applied this adjustment factor to the household data to derive a payroll jobs number of 127,000 per month.

Specifically, a 1.1% growth of the working age population adds 2.554 million people. If the employment population rate remains at 62.8%, this implies an employment growth of 1.604 million (or 133,600 each month). If payroll growth were to be 95% of this household employment growth, it would need to be 127,000 per month.

Notes
1. Census data are available only through 2006, at which point the real median family income was about 2%, or about $1,000 below its 2000 level. Our forecast for 2007 finds that the difference, 2000-06, reduces to -1%.

2. Even though the official recession was over, it was good that Congress did extend benefits because unemployment continued to rise for many months. In fact, the period became commonly referred to as the "jobless recovery" due to the historically long time it took for the labor market to bounce back after the recession.