We Are Looking at the Unemployment Rate Incorrectly Part II: What is a job?
**This is the second part of a multi-part series examining issues with the unemployment rate. Part I can be found here. Interactive versions of these charts and the code to make them can be found here**
The United States added 178,000 jobs to the labor market in November of 2016. On the surface, this seems to be pretty positive. But what exactly does 178,000 new jobs mean? If we added 178,000 jobs, how many did we lose? Are these jobs open or filled positions? Are they full- or part-time? Does that mean 178,000 more Americans are employed this month?
What is a job?
The Bureau of Labor Statistics calculates new jobs as:
To be on a payroll, a job must have been filled. When job creation numbers are reported, they indicate a net increase or decrease in jobs.
A new job does not necessarily mean a new person is employed, as an employed individual could possess multiple jobs. As discussed in Part I of this series, any person who earned income in the previous four weeks is regarded as employed. The number of hours worked or amount income they made from that job does not impact a persons employment classification. So what does this number say about the change in employment? Less than you might expect:
While researching this topic, I was shocked to see how often employment levels dropped while job creation was positive. Monthly job creation levels have been positive each month since 2011; however, there have been 17 months in which the employment level has decreased. Since 1980, these two metrics moved together only roughly 3/4 of the time.
The top left quadrant of this graphic is particularly concerning; as it indicates that jobs were created, but employment still decreased. A change in jobs with no corresponding change in employment can occur for many reasons, such as individuals taking multiple part-time jobs, or full-time employees are picking up a second job. While these scenarios are not categorically negative, neither scenario is a particularly inspiring economic indicator.
The distribution of monthly changes in job and employment growth yields some interesting observations. Decreases in jobs occur much less frequently than decreases in employment, and the distribution of potential job growth numbers is much more narrow than that of employment growth.
Full-Time vs. Part-Time Jobs and Employment
If 178,000 jobs were created in November of 2016, how many of them were full-time jobs? The simple answer is we don’t know. The BLS measures people as full- or part-time employed, not jobs. This makes job creation numbers particularly difficult to interpret.
The BLS reports the changes in full- and part-time employment, but you will need to scroll down to the 20th page of the jobs report to find these numbers. If a person usually works more than 34 hours in a week, they are regarded as full-time employed (even if they reach this number through multiple part-time jobs). A quick look at monthly full- and part-time employment level changes shows indicates why these numbers are not highlighted:
The data for full- and part-time employment possesses a lot of noise. Not only do the monthly changes vary dramatically, the sub-components of total employment do not add up to the total employment level. In November of 2016, the jobs report stated the total employment level rose 160,000. The individual gains for full- and part-time employment were 9,000 and 118,000 respectively.
The BLS acknowledges the discrepancies between individual data components and their totals. These discrepancies exist because the BLS seasonally adjusts individual components separately from totals. The BLS offers the following guidance for analyzing seasonally adjusted data:
“Separate adjustment of components and summing of them to the total usually provides series that are easier to analyze; it is also generally preferable in cases where the relative weights among components with greatly different seasonal factors may shift radically. For other series, however, it may be better to adjust the total directly if high irregularity among the components makes a good adjustment of all components difficult.”
For this analysis, I will generally ignore the reported total employment and use the sum of full- and part-time components.
Given how noisy changes in employment components are, how exactly should full- and part-time employment growth be analyzed? The Wall Street Journal uses the index approach (a visualization approach I glowingly reviewed in this post):
While the index approach provides fantastic perspective on the full-time employment situation, the short-term changes are barely visible.
At this point, it is probably worth questioning whether or not monthly employment changes should even be analyzed. An alternative to looking at absolute monthly changes would be to aggregate employment data into quarterly averages. This method provides a removes the noise found in monthly changes, while still providing an understanding of short-term changes. It also provides the ability to measure annual percent change, a method which should not be used in highly volatile data such as monthly employment levels.
Is Job Creation Important?
Job creation is in fact an important economic indicator, and it is important driver of economic growth. That said, job creation numbers are not especially relevant to unemployment levels. Frequently these two metrics move together, but they differ enough that job creation should not be discussed as if it were an input to unemployment rate.
