On the first Friday of every month, the markets get ready to process the latest Employment Situation Report. The report is released by the Bureau of Labor Statistics and details the state of the labour market in the United States (US). The first sentence of the report usually reads something like this:
"The unemployment rate rose to 4.3 percent in July, and nonfarm payroll employment edged up by 114,000, the U.S. Bureau of Labor Statistics reported today.”
How Employment Figures Influence Economic Policies and Markets
Employment indicators are some of the timeliest measures of the state of the US economy. In the days, hours and minutes following their release, markets can react decisively if they miss expectations.
This is because the Federal Reserve (the Fed) uses the latest employment data to check whether its monetary policies are achieving their goals and what direction the economy is heading. For example, over the past 18 months, the Fed has raised rates and held them at their highest level for some time to combat inflation. Higher interest rates tend to slow down the economy, and a key sign of this slowdown is higher unemployment. Consequently, rising unemployment rates suggest to the Fed that its strategy to control inflation might be working, which in turn influences market reactions.
Where Unemployment Stats Fall Short
However, the unemployment rate is often revised after it is released because, like most economic data, it is subject to errors. While it gives us some details about the state of employment in the country, it doesn’t tell us everything. This often results in the market forming distorted views about the state of the US labour market and the effectiveness of the Fed's monetary policy.
We’ll break down what the unemployment rate is, its limitations, and why it’s often revised after the fact.
1. Scope: The Mechanics of Unemployment Rates and Their Different Types
The Bureau of Labor Statistics (BLS) has been calculating unemployment statistics in largely the same way since 1940. The BLS follows the international guidelines for reporting labour statistics, which are set by the International Labour Organisation (ILO).
To be classified as unemployed, a person must be (1) without work, (2) available for work, and (3) have actively searched for work. Although that classification is relatively simple, the BLS recognises that employment dynamics are nuanced and no single measure holistically reflects the state of employment in the economy. To that end, the BLS publishes six different unemployment rates every month. They are colloquially known as U-1 to U-6, with U-3 being the official unemployment rate.
The first two measures, U-1 and U-2, define unemployment more narrowly than the official rate (U-3), meaning they include fewer people as unemployed. On the other hand, measures U-4 to U-6 have broader definitions of unemployment and usually show higher rates than the official U-3 rate.
The table below defines U-1 to U-6 and shows reported rates over the past year.
The BLS refers to each of these statistics as “alternative measures of labour underutilisation”. For simplicity, they are best thought of as stricter and more lenient definitions of the official unemployment rate, or U-3.
This table shows significant differences between unemployment measures U-1 to U-6. Understanding these differences requires knowing some key terms.
Unemployment Terms Explained
Civilian Labour Force – People aged 16 years or older and classified as employed or unemployed.
Marginally Attached to the Workforce – Those persons who are not in the labour force, who want and are available for work, and who have looked for a job in the past 12 months but are not counted as employed because they haven’t looked for a job in the past four weeks preceding the employment survey.
Discouraged Worker – A subset of the marginally attached but stopped looking for work because they believed no work was available for them, therefore not included as unemployed.
The Formula Behind the Unemployment Rate
If we think of the basic unemployment rate calculation, across all six measures, the numerator is an estimate of the number of unemployed people that varies depending on how unemployment is defined. Likewise, the denominator is an estimate of the labour force, which is the number of people in a country who are available to work.
Formulaically, U-3, which is the official unemployment rate, is defined as follows: U-1 to 2 and U-4 to 6 either expand or reduce the size of the numerator and/or the denominator.
The market focuses almost exclusively on U-3, the official unemployment rate. However, the mere existence of the above classification tells us that U-3 only provides us with a narrow view of the labour market.
The fact that discouraged workers and marginally employed people are excluded from this statistic has a meaningful impact on the reported numbers. This can be seen by the change in the numbers from U-3 to U-5 above.
2. Context – Uncovering the Limitations in U-3 Unemployment Data
In addition to the varied definitions of the labour force and unemployment, U-3 has several limitations.
Limited Information on Job Quality: The unemployment rate tells us if people have a job but nothing about the quality of that job. For example, it doesn’t distinguish between someone working a single-hour week and someone working 60 hours a week. According to the ILO, a person is classified as employed if they worked at least one hour during the reference week in paid employment, regardless of job type or hours worked. Employment statistics also don’t tell us if people are working in jobs that they are overqualified for, known as underemployment.
Limited Information on the Unemployed: Information like the education level of the unemployed as well as whether or not people are recently discouraged workers is important to understand. For example, if most of the unemployed are skilled workers, that may lead to different conclusions about the state of the economy than if most of the unemployed are unskilled workers. Likewise, to be classified as unemployed you need to have actively looked for work in the past four weeks. So technically, if a person last looked for a job five weeks ago they are classified as discouraged, but they may not in fact be a discouraged worker. Therefore, the official unemployment rate may actually be higher as a result.
Ultimately, these factors all have a meaningful effect on the makeup of the labour market and having sufficient information on each may help us to better judge the state of the economy.
3: The Lag Effect of Unemployment Data
Unemployment statistics are classified as a lagging economic indicator because they reflect past economic activity. That is, employment activity reacts to the state of the economy.
In addition, employment statistics take some time to gather. In the US, it typically takes the BLS around two weeks to conduct employment surveys and then another two weeks to report labour statistics. That means that the most recent employment statistics report on the state of the labour market four weeks earlier.
Short-term abnormalities in the labour market like strikes, natural disasters and temporary lay-offs may also distort labour market dynamics, especially if they coincide with the weeks over which the BLS surveys the economy.
4: How Data Errors Impact Employment Statistics
Unemployment statistics are typically prone to four kinds of data errors:
Measurement Errors – These tend to occur when survey data is misinterpreted due to incorrect reporting by survey respondents.
Sample Size Errors – When the size of the sample which is surveyed is too small, this leads to a distorted view of the labour market.
Classification Errors – When data reported through the surveys is misinterpreted and not classified correctly.
Lack of Data – Ultimately, employment data is only useful if survey respondents respond and the size of the sample is sufficient enough to make broad-based conclusions about the state of the labour market.
While these may seem like relatively trivial errors, they can have a substantial impact and ultimately lead to data revisions.
A Practical Example of a Measurement Error – Nonfarm Payroll Revisions
One of the most reported statistics in the Employment Situation Report is the Nonfarm Payroll (NFP) number. It measures the total amount of employed people (on a company’s payroll) working in goods, construction and manufacturing. It represents about 80% of the US workforce and monthly changes in the NFP number are closely followed by the market.
The number usually reported is the difference between the latest and the previous month’s NFP number, which indicates the total jobs in those sectors that the economy added or lost over the last month. However, NFP numbers are often revised because of some of the data errors we highlighted above. Most recently, the BLS preliminarily revised the March 2024 NFP figure down by 818 000 jobs.
That means that in March, the BLS over-reported the total number of jobs in the goods, construction and manufacturing sectors by 818 000 (i.e. there were 818 000 fewer jobs than originally thought). That is a material overestimation and has led many analysts to believe that the unemployment situation in the US is worse than originally estimated.
Conclusion: Navigating the Complexities of Employment Data
Employment statistics provide the market with important information about the state of the US labour market, the effectiveness of the Fed’s monetary policy and ultimately the direction of the US economy.
However, they also provide us with less information than we need to properly characterise the state of employment in the US.
Crucial data like NFP numbers, which the market uses to form expectations, is often revised. Understanding these shortcomings may help us form a more balanced economic view and ultimately lead to better investment decisions.
Moreover, although the Fed’s rate-hiking regime has been effective in bringing down inflation, the latest revised employment data suggests it might have been too effective. It remains to be seen whether or not the Fed can still keep the US economy out of recession like the market expects.
Opmerkingen