Almost Unbelievable Labor Market Strength
The strength of the U.S. labor market has not only defied expectations. It has given rise to the suspicions that the official figures are overstating job gains.
When the government releases the March jobs report on Friday, questions about the reliability of the figures are likely to resurface. Can we trust the numbers?
The Bureau of Labor Statistics (BLS) reports that payrolls expanded on a seasonally adjusted basis by 290,000 in December, 229,000 in January, and 275,000 in February. This puts the three-month moving average at 265,000, a very high level of employment growth by historical standards.
The median expectation in the Econoday survey is for 200,000 jobs, with responses ranging from 150,000 to 230,000. That is a relatively wide range of forecasts, reflecting the uncertainty about what is going on with the labor market. If the numbers come in as expected, the three-month average will decline to 234,000, high enough to indicate a still robust demand for workers.
The so-called “whisper number”—the figure traders are expecting—has moved slightly higher this week following the release of the Job Openings and Labor Turnover Survey (JOLTS) and the ADP private payroll report. By most accounts, the market is likely expecting around 215,000 jobs.
A Tale of Two Jobs Surveys
Skeptics of the recent jobs figures have focused on a large discrepancy that has emerged between the two distinct surveys that emerge out of the labyrinthine workings of the BLS to make up the monthly jobs report. According to one survey, we’re seeing the best of times. According to the other, the worst of times.
One survey relies on questionnaires sent out to approximately 119,000 businesses and government entities, known as the establishment survey. This behemoth of data collection serves as the bedrock for assessing the U.S.’s monthly job additions, normally considered a critical indicator of economic health.
Parallel to this, the BLS conducts a survey of about 60,000 households. This household survey is used to produce the unemployment rate and delivers an alternate estimate of employment, including segments of the workforce like the self-employed, who might escape the establishment survey’s gaze.
According to the establishment survey, there were approximately 2.7 million additional workers on payrolls compared to the previous year in February. The household survey shows job growth of only half as much. More strikingly, the establishment survey shows payrolls grew by 1.4 million in the six months through February while the household survey recorded a decline in employment of 532,000.
Those who see jobs as overstated can point to a lot of potential sources of error in the establishment survey. Response rates to the BLS surveys have declined since the pandemic, and we’ve had many large revisions as the government bean counters collect additional data to fill in the gaps. Some analysts think that the birth/death adjustments used by the BLS have been too generous, leading the official statistics astray. What’s more, the establishment data do not adjust for multiple jobholders, so that if someone takes on a third job this shows up as three jobs. The household survey avoids this because it totals jobs by just asking if someone is employed.
Choose Your Own Jobs Adventure
It should be noted that there’s a lot of motivated reasoning going on with the folks who think the household survey is more accurately reflecting reality. And unlike the partisan politics behind so many other divides about data, people on both ends of the political spectrum have incentives to believe the labor market is weaker than it might seem.
Progressives have been urging the Fed to cut interest rates, and they see a strong jobs market as an obstacle to that. Some conservatives are convinced that there’s no way jobs could be growing this fast while Biden’s policies weigh on the economy. And then there are the bitter Wall Street analysts who have decided that its not their forecasts that have been wrong but the data that is misleading.
Traders work on the floor of the New York Stock Exchange during afternoon trading on April 2, 2024, in New York City. (Michael M. Santiago/Getty Images)
The trouble is that if the household survey were telling the “real story” about the economy, we would already be in a recession. The collapse of employment it is describing is too large to fly under the radar. We would see it in declining consumer spending, soaring jobless benefit claims, and stagnating or falling personal income. JOLTS would not likely be indicating openings well-above pre-pandemic levels. Wage growth would have fallen off the edge of the table. GDP would be grinding to a standstill. Home prices would be stalled or falling.
None of that is happening. The Atlanta Fed’s GDPNow reads the economy as having grown at a 2.5 percent rate in the first three months of the year. Wage and salary income rose by 0.8 percent in February and 5.6 percent from a year earlier. Openings ticked up. Personal spending jumped 0.8 percent for the month and 4.9 percent for the year. Home prices are rising.
The rest of the economic data, in other words, supports the view of the economy from the establishment survey rather than the household survey. It’s very likely that the discrepancy is due to sampling errors in the latter that will be revised away over time—in the direction of more jobs.
If the discrepancy continues in Friday’s report, our approach will be to trust the establishment—which is not something we usually find ourselves saying around here.
