More Than Jobs Friday: A Long-Term Look At Employment

• Average hourly real wages remain 6.7% below its peak set in January 1973, meaning that the average American worker hasn't had a "real" pay raise in over 40 years.
• The gap between supervisory workers and nonsupervisory workers has expanded.
• The majority of the job growth since the 2000 peak and the Financial Crisis low has been concentrated in Professional and Business Services, followed by Education and Health Care, with Leisure and Hospitality rounding out third.
• Based on BLS projections to 2022, the economy would have to add about 100K jobs a month to keep the employment-to-population ratio constant at 60.1% or 53K a month to hold the unemployment rate at 4.5%.

Background

The Bureau of Labor Statistics' Employment Situation is comprised of two distinct surveys. The first is the Current Employment Statistics (CES) survey (also called the establishment survey), which surveys approximately 147,000 businesses, representing approximately 634,000 individual workers. The establishment survey is collected from the payroll records of a sample of nonagricultural businesses with statistical significance reached at 100,000 jobs. The second is the Current Population Survey (also called the household survey), which is more expansive in scope, sampling about 60,000 eligible households and is conducted by the U.S. Census Bureau. The survey is statistically significant at 500,000 jobs. The CPS can include self-employed, farm, unpaid family workers, and private household workers while the CES does not. In the CPS survey, individuals are only counted once as being employed even if they hold more than one job; however, in the CES, double counting is permissible, and an employee working for two firms will show up twice.

The CPS survey reports on population, labor force, employed (full time and part time), unemployed, not in labor force, etc. while the CES shows employment by industry, hours, and wages.

Employment growth since 2000

While many focus on the monthly jobs report I think garnering a longer perspective is in order. I will focus on the growth in employment from two dates. The first, February 2001, was approximately when the civilian labor force participation rate and employed-to-population ratio rate peaked at 67.3% and 64.7%, respectively. The second, February 2010, was the employment low set after the Financial Crisis.

Source: BLS

Since both dates, professional and business services, education and health, and leisure and hospitality have contributed to the majority of job growth. They have grown 24.9%, 19.4%, and 17.9% from February 2010, and 28.9%, 56.6%, and 29.0% since February 2001. Part-time employment, a subsector in professional and business services, has contributed 6.3% since 2010. These subsectors also constitute some of the lowest wage industries. Conversely, higher paying jobs in mining/logging and manufacturing have shown zero gains or outright contraction in the latter. The household survey data collaborates the aforementioned establishment data, showing that part time unemployment is 18% of the total workforce.

Source: BLS

Wages

Nominal hourly wages have shown growth since February 2010, particularly for supervisory employees, which we can estimate given that nonsupervisory employees comprise about 18% of the labor force.

Source: BLS

That said, average weekly earnings have only been growing at about 2.25% since 2012, with any wage gains offset by declines in average hours. Growth has been modest at best with no substantial acceleration to the upside except for the move off the bottom in 2010.

Source: BLS

Using the wages of nonsupervisory and production employees back to the 60s, and data from NBER prior to the 1960s, I was able to construct a multi-decade view of average hourly earnings. Not surprisingly, the estimate of supervisory average hourly earnings has grown disproportionately to its nonsupervisory counterpart.

Source: BLS, St. Louis Fed, NBER

Adjusting for inflation reveals an entirely different picture. Average hourly real wages peaked in January 1973 at $28.02. As of March 2017, average real hourly wages were $26.14. This means that the average American hasn't had a real pay raise in over 40 years. The high in real wages occurred right after America abandoned the gold standard, and corresponds with the exponential rise in debt.

Source: BLS, St. Louis Fed, NBER

The outlook

As of December 2015, the BLS forecasts that the civilian non-institutional population is expect to grow to 265.3mn people from the current 254.4mn while the civilian labor force is expected to grow to 163.5mn from 160.2mn. Assuming the employment-to-population ratio holds at 60.1% would require the addition of 93K jobs per month under the household survey, or about 102K under the establishment survey (assumes the current average of 9K in difference between the two). Assuming the unemployment rates holds at 4.5% would require the economy to add 45K jobs a month under the household survey. However, under that scenario the employment-to-population ratio would reverse course and fall to 58.9%. To return to the old highs in labor force participation rate and employed-to-population ratio would require the civilian labor force to grow by 11.4% over the next five years, not impossible but highly unlikely. More importantly, it would require the addition of 270K jobs a month in the household survey or 279K jobs in the establishment, both of which seem extremely improbable.

Given the current dichotomy between supervisory and nonsupervisory wages, the lack of real wage growth, and the seemingly impossible climb to regain 2000 peak employment, I don't think the labor market as tight as some sources claim.