Tag: Establishment survey

A Review of Recent Macro Data

Some interesting macro data were released during the past two weeks. On the key issues, the data indicate that inflation continues to run in the range of 3.0 percent to 3.5 percent, although depending on which series you focus on, you could conclude that inflation has dropped to a bit below 3 percent or that it is still in vicinity of 4 percent.  On balance, output and employment data seem to be indicating that the economy may be cooling in response to the contractionary monetary policy that the Federal Open Market Committee began implementing in March 2022.

We can summarize the key data releases.

Employment, Unemployment, and Wages

On Friday morning, the Bureau of Labor Statistics (BLS) released its Employment Situation report. (The full report can be found here.) Economists and policymakers—notably including the members of the Federal Reserve’s Federal Open Market Committee (FOMC)—typically focus on the change in total nonfarm payroll employment as recorded in the establishment, or payroll, survey. That number gives what is generally considered to be the best indicator of the current state of the labor market.

The previous month’s report included a surprisingly strong net increase of 336,000 jobs during September. Economists surveyed by the Wall Street Journal last week forecast that the net increase in jobs in October would decline to 170,000. The number came in at 150,000, slightly below that estimate. In addition, the BLS revised down the initial estimates of employment growth in August and September by a 101,000 jobs. The figure below shows the net gain in jobs for each  month of 2023.

Although there are substantial fluctuations, employment increases have slowed in the second half of the year. The average increase in employment from January to June was 256,667. From July to October the average increase declined to 212,000. In the household survey, the unemployment rate ticked up from 3.8 percent in September to 3.9 percent in October. The unemployment rate has now increased by 0.5 percentage points from its low of 3.4 percent in April of this year. 

Finally, data in the employment report provides some evidence of a slowing in wage growth. The following figure shows wage inflation as measured by the percentage increase in average hourly earnings (AHE) from the same month in the previous year. The increase in October was 4.1 percent, continuing a generally downward trend since March 2022, although still somewhat above wage inflation during the pre-2020 period.

As the following figure shows, September growth in average hourly earnings measured as a compound annual growth rate was 2.5 percent, which—if sustained—would be consistent with a rate of price inflation in the range of the Fed’s 2 percent target.  (The figure shows only the months since January 2021 to avoid obscuring the values for recent months by including the very large monthly increases and decreases during 2020.)

Job Openings and Labor Turnover Survey (JOLTS) 

On November 1, the BLS released its Job Openings and Labor Turnover Survey (JOLTS) report for September 2023. (The full report can be found here.) The report indicated that the number of unfilled job openings was 9.5 million, well below the peak of 11.8 million job openings in December 2021 but—as shown in the following figure—well above prepandemic levels.

The following figure shows the ratio of the number of job opening to the number of unemployed people. The figure shows that, after peaking at 2.0 job openings per unemployed person in in March 2022, the ratio has decline to 1.5 job opening per unemployed person in September 2022. While high, that ratio was much closer to the ratio of 1.2 that prevailed during the year before the pandemic. In other words, while the labor market still appears to be strong, it has weakened somewhat in recent months.

Employment Cost Index

As we note in this blog post, the employment cost index (ECI), published quarterly by the BLS, measures the cost to employers per employee hour worked and can be a better measure than AHE of the labor costs employers face. The BLS released its most recent report on October 31. (The report can be found here.) The first figure shows the percentage change in ECI from the same quarter in the previous year. The second figure shows the compound annual growth rate of the ECI. Both measures show a general downward trend in the growth of labor costs, although compound annual rate of change shows an uptick in the third quarter of 2023. (We look at wages and salaries rather than total compensation because non-wage and salary compensation can be subject to fluctuations unrelated to underlying trends in labor costs.)

The Federal Open Market Committee’s October 31-November 1 Meeting

As was widely expected from indications in recent statements by committee members, the Federal Open Market Committee voted at its most recent meeting to hold constant its targe range for the federal funds rate at 5.25 percent to 5.50 percent. (The FOMC’s statement can be found here.)

At a press conference following the meeting, Fed Chair Jerome Powell remarks made it seem unlikely that the FOMC would raise its target for the federal funds rate at its December 14-15 meeting—the last meeting of 2023. But Powell also noted that the committee was unlikely to reduce its target for the federal funds rate in the near future (as some economists and financial jounalists had speculated): “The fact is the Committee is not thinking about rate cuts right now at all. We’re not talking about rate cuts, we’re still very focused on the first question, which is: have we achieved a stance of monetary policy that’s sufficiently restrictive to bring inflation down to 2 percent over time, sustainably?” (The transcript of Powell’s press conference can be found here.)

Investors in the bond market reacted to Powell’s press conference by pushing down the interest rate on the 10-year Treasury note, as shown in the following figure. (Note that the figure gives daily values with the gaps representing days on which the bond market was closed) The interest rate on the Treasury note reflects investors expectations of future short-term interest rates (as well as other factors). Investors interpreted Powell’s remarks as indicating that short-term rates may be somewhat lower than they had previously expected.

Real GDP and the Atlanta Fed’s Real GDPNow Estimate for the Fourth Quarter

On October 26, the Bureau of Economic Analysis (BEA) released its advance estimate of real GDP for the third quarter of 2023. (The full report can be found here.) We discussed the report in this recent blog post. Although, as we note in that post, the estimated increase in real GDP of 4.9 percent is quite strong, there are indications that real GDP may be growing significantly more slowly during the current (fourth) quarter.

The Federal Reserve Bank of Atlanta compiles a forecast of real GDP called GDPNow. The GDPNow forecast uses data that are released monthly on 13 components of GDP. This method allows economists at the Atlanta Fed to issue forecasts of real GDP well in advance of the BEA’s estimates. On November 1, the GDPNow forecast was that real GDP in the fourth quarter of 2023 would increase at a slow rate of 1.2 percent. If this preliminary estimate proves to be accurate, the growth rate of the U.S. economy will have sharply declined from the third to the fourth quarter.

Fed Chair Powell has indicated that economic growth will likely need to slow if the inflation rate is to fall back to the target rate of 2 percent. The hope, of course, is that contractionary monetary policy doesn’t cause aggregate demand growth to slow to the point that the economy slips into a recession.

Unraveling the Mysteries of the May 2023 Employment Situation Report

(Photo from the Associated Press via the Wall Street Journal.)

During most periods, the “Employment Situation” report that the Bureau of Labor Statistics issues on the first Friday of each month includes the most closely watched macroeconomic data. Since the spring of 2021, high inflation rates have made the BLS’s “Consumer Price Index Summary” at least a close second in interest to the employment report. The data in the CPI report is usually more readily comprehensible than the data in the employment report. So, we think it’s worth class time to go into some of the details of the employment report, as we do in Macroeconomics, Chapter 9, Section 9.1, Economics, Chapter 19, Section 19.1, and Essentials of Economics, Chapter 13, Section 13.1.

When the BLS released the May employment report, the Wall Street Journal noted that: “Employers added 339,000 jobs last month; unemployment rate rose to 3.7%.” Employment increased … but the unemployment rate also rose? How is that possible? One key to understanding media accounts of the report is to note that the report contains data from two separate surveys: 1) the household survey and 2) the employment or establishment survey. As in the statement just quoted from the Wall Street Journal, media accounts often mix data from the two surveys.  

The data showing an increase of 339,000 jobs in May are from the payroll survey, while the data showing that the unemployment rate rose are from the household survey. Below we reproduce part of the relevant table from the report showing some of the data from the household survey. Note that total employment in the household survey falls by 310,000, so there appears to be no contradiction to explain—the unemployment rate increased because the number of people employed fell and the number of people unemployed rose. But why, then, did employment rise in the payroll survey?

Employment can rise in one survey and fall in the other because: 1) the types of employment measured in the two series differ, 2) the periods during which the data are collected differ, and 3) because of measurement error. The household survey uses a broader measure of employment that includes several categories of workers who are not included in the payroll survey: agricultural workers, self-employed workers, unpaid workers in family businesses, workers employed in private households rather than in businsses, and workers on unpaid leave from their jobs. In addition, the payroll employment numbers are revised—sometimes substantially—as additional data are collected from firms, while the household employment numbers are subject to much smaller revisions because data in the household survey are collected during a single week. A detailed discussion of the differences between the employment measures in the two series can be found here.

Usefully, the BLS publishes a series labeled “Adjusted employment” that estimates what the value for household employment would be if the household survey was measuring the same categories of employment as the payroll survey. In this case, the adjusted employment series shows an increase in employment in May of 394,000—close to the payroll survey’s increase of 339,000.

To summarize, the May employment report indicates that payroll employment increased, while the non-payroll categories of household employment declined, and the unemployment rate rose. Note also in the table above that the number of people counted as not being in labor force rose slightly and the employment-population ratio fell slightly. Average weekly hours (not shown in the table above) decreased slightly from 34.4 hours per week to 34.3.

A reasonable conclusion from the report is that the labor market remains strong, although it may have weakened slightly. Prior to release of the report, there was much speculation in the business press about how the report might affect the deliberations of the Federal Reserve’s Federal Open Market Committe (FOMC) at its next meeting to be held on June 13th and 14th. The report showed stronger employment growth than economists surveyed by Dow Jones had expected, indicating that the FOMC was likely to remain concerned that a tight labor market might continue to put upward pressure on wages, which firms could pass through to higher prices. Members of the FOMC had been signalling that they were likely to keep their target for the federal funds rate unchanged in June. The reported employment increase was likely not large enough to cause the FOMC to change course.

Solved Problem: How Can Total Employment and the Unemployment Rate Both Increase?

Photo from the New York Times.

Supports: Macroeconomics, Chapter 9, Section 9.1, Economics Chapter 19, Section 19.1, and Essentials of Economics, Chapter 13, Section 13.1.

As it does on the first Friday of each month, on September 2, 2022, the U.S. Bureau of Labor Statistics (BLS) released its “Employment Situation” report for August 2022. According to the household survey data in the report, total employment in the U.S. economy increased in August by 442,000 compared with July. The unemployment rate rose from 3.5 percent in July to 3.7 percent in August. According to the establishment survey, the total number of workers on payrolls increased in August by 315,000 compared with July.

  1. How are the data in the household survey collected? How are the data in the establishment survey collected?
  2. Why are the estimated increases in employment from July to August 2022 in the two surveys different? 
  3. Briefly explain how it is possible for the household survey to report in a given month that both total employment and the unemployment rate increased.

Solving the Problem

Step 1: Review the chapter material. This problem is about how the BLS reports data on employment and unemployment, so you may want to review Chapter 9, Section 9.1, “Measuring the Unemployment Rate, the Labor Force Participation Rate, and the Employment–Population Ratio.” 

Step 2: Answer part a. by explaining how the data from the two surveys are collected. As discussed in Section 9.1, the data in the household survey is from interviews with a sample of 60,000 households, chosen to represent the U.S. population. The data in the establishment survey—sometimes called the payroll survey in media stories—is from a sample of 300,000 establishments (factories, stores, and offices).  

Step 3: Answer part b. by explaining why the estimated increase in employment is different in the two surveys.  First note that the BLS intends the surveys to estimate two different measures of employment. The household survey includes people working at jobs of all types, including people who are self-employed or who are unpaid family workers, whereas the establishment survey includes only people who appear on a non-agricultural firm’s payroll, so the self-employed, farm workers, and unpaid family workers aren’t counted. Second, the data are collected from surveys and so—like all estimates that rely on surveys—will have some measurement error.  That is, the actual increase in employment—either total employment in the household survey or payroll employment in the establishment survey—is likely to be larger or smaller than the reported estimates. The estimates in the establishment survey are revised in later months as the BLS receives additional data on payroll employment. In contrast, the estimates in household survey are ordinarily not revised because they are based only on a survey conducted once per month.  

Step 4: Answer part c. by explaining how in a given month the household survey may report an increase in both employment and the unemployment rate.  The BLS’s estimate of the unemployment is calculated from responses to the household survey. (The establishment survey doesn’t report an estimate of the unemployment rate.) The unemployment rate equals the total number of people unemployed divided by the labor force, multiplied by 100. The labor force equals the sum of the employed and the unemployed. If the number of people employed increases—thereby increasing the denominator in the unemployment rate equation—while the number of people unemployed remains the same or falls, as a matter of arithmetic the unemployment rate will have to fall. 

The BLS reported that the unemployment rate in August 2022 rose even though total employment increased. That outcome is possible only if the number of people who are unemployed also increased, resulting in a proportionally larger increase in the numerator in the unemployment equation relative to the denominator. In fact, the BLS estimated that the number of people unemployed increased by 344,000 from July to August 2022. Employment and unemployment both increasing during a month happens fairly often during an economic expansion as some people who had been out of the labor force—and, therefore, not counted by the BLS as being unemployed—begin to search for work during the month but don’t find jobs.

Source: U.S. Bureau of Labor Statistics, “The Employment Situation—August 2022,” bls.gov, September 2, 2022.