When the US government tried to analyze data distorted by the pandemic, it greatly underestimated the number of jobs created this year, which has brought new challenges to policymakers in a highly turbulent economic environment.
During 2021, the government agency that releases monthly employment statistics in the United States has raised its initial estimate of employment growth by 976,000 jobs, which is the highest level of such adjustments in a year.
Revising past estimates based on new data is a routine task: the initial estimates calculated by the Bureau of Labor Statistics (BLS) are updated twice, the second estimate is published in the next month’s employment report, and the third estimate is below Released in months.
However, it is during the major economic turning points that these numbers are not only the most difficult to measure, but they are also the most closely monitored. This means that data revisions that are usually not released with much fanfare suddenly have the potential to significantly change the concept of labor market health and thus how fine-tuned policies should be.
“The fundamentals of the economy have been changing at an unprecedented rate. In my lifetime and in the lives of most people alive we see today…. The economic recovery has been as rapid as it has been since the spring of 2020.” Said David Wilcox, head of the research and statistics department. “The challenge of economic measurement in a pandemic environment is huge.”
It turns out that for economists studying the American labor market, this epidemic is a uniquely thorny issue because it not only causes a lag in data reporting, but also distorts the seasonal rhythms that are usually included in estimates. The sheer scale of wage growth during the economic recovery has also brought challenges.
Ernie Tedsky, a senior economic adviser to the Biden administration and a member of the Economic Advisory Council, said: “You are taking on an already arduous task, making it absolutely arduous.”
This difficulty reappears in the latest employment report Publish On December 3, it shows that only 210,000 new positions were created in November. The unemployment rate dropped further to 4.2%, a far cry from the nearly 15% reached last year.
These numbers confuse economists, especially because of the huge gap between the two surveys that make up the monthly report. The company’s “establishment” survey showed that the number of recruits has slowed sharply, and from this, overall employment data that may be revised. The individual “household” survey used to calculate the unemployment rate showed an increase of 1.1 million.
To further complicate the situation, the Bureau of Labor Statistics has extensively updated the September and October data, increasing the number of employed persons to 379,000 in September and 546,000 in October. In total, approximately 6 million jobs have been restored in 2021.
“Everything is much bigger than it used to be,” said Stephen Crestol, an economist who has worked at BLS for more than 30 years. “Slumping and rising, we are not used to it.”
The task of estimating the number of payrolls during a pandemic is complicated by two main factors. First, more and more companies are late in submitting survey responses.
Krestall said that during the pandemic, it is difficult for the agency to let companies choose to participate in corporate surveys. Except for the three states, all other states participate voluntarily, Halved Since last February.
Even if companies agree to fill out the survey, their responses may appear after the initial estimated deadline in the employment report, which is released on the first Friday of the month.
For the November report, 65.3% of companies Responded In time, this is the lowest interest rate reported in November in more than a decade.
To generate the first number, BLS calculated the expected payroll numbers for companies that have not yet responded. Late responses were incorporated into subsequent revisions.
“In a normal month, your bottom line won’t change much,” said Erica Groshen, a former BLS commissioner at Cornell University. But she said that in a year like 2021, marked by a rapid economic recovery, the slowest responding companies may hire the fastest, leading to a greater upward adjustment.
Seasonal adjustment factors also make measuring employment growth a difficult task.
Economists pay close attention to the “seasonally adjusted” data, because once we remove the frequent fluctuations in the data related to events such as the beginning of the school year, these data can provide the most direct understanding of current trends.
For example, compared with October, retail companies reported an increase of 331,600 employees in November.But the U.S. Bureau of Labor Statistics adjust This number has fallen because hiring in the retail industry tends to pick up before the holidays begin. The agency reported that after taking into account seasonal adjustments, the industry “lost” 20,400 jobs.
Overall, the original data showed that the number of employees in November increased by 778,000, and the US Bureau of Labor Statistics lowered it by 568,000, a record high.
“Seasonal patterns are not certain laws, they are just patterns,” said Bessie Stevenson, a professor of economics at the University of Michigan and a member of the CEA during the Obama administration. “Covid has done more to disrupt our model than I thought.”
The seasonal factor is based on the data of the past 10 years, and the weight is greater in recent years. In 2020, many sports will be Considered outliers And it is not included in the adjustment model.
As new data enters, the model will continue to adjust, leading to more revisions. “The more data you get, it will affect your understanding of past seasonal trends,” said Nick Bunker of the employment website Indeed.
How the new coronavirus variant unfolds will determine whether the recent wave of revisions will peak or become larger from here.
This has a huge impact on the U.S. Central Bank, which pays close attention to the employment situation in the United States. Green light Tighten monetary policy next year. Governor Christopher Waller stated that the Fed has supplemented its model with high-frequency data and other sources, including weekly ADP employment reports.
“The more we see the future recovering from the recession, the better the effect of statistical techniques,” Groschen said. “The more it is driven by new variants and different policy responses… The more likely it is that models will fail because they are based on the past.”