U.S. employment growth has been revised downward the most since 2009. | Wilnesh News

There is much debate over how much of a signal to read in the 818,000 U.S. payrolls revision, the largest since 2009.

Some facts worth considering:

• When the revised results for 2009 were released (824,000 jobs were inflated), the National Bureau of Economic Research had declared a recession six months earlier.
• Source: Unemployment claims surged to more than 650,000 during the same period, and the insured unemployment rate peaked at 5% that month.
• The gross domestic product announced at that time had been negative for four consecutive years. (Two of the quarters were subsequently revised upwards, with one revised upward to show growth rather than contraction. But the economic weakness was generally evident in the GDP numbers, the ISM and many other figures.)

The current revision covers the period from April to March 2023, so we don’t know if the current number is higher or lower. The model used by the Bureau of Labor Statistics likely overstates the strength of the economy at a time of economic weakness. While there are signs of weakness in the labor market and economy (which may well be further evidence), the same indicators in 2009 now look like this:

• A recession has not yet been declared.
• The four-week moving average of jobless claims is 235,000, unchanged from the same period last year. The insured unemployment rate has remained unchanged since March 2023 at 1.2%.
• Reported GDP has been positive for eight consecutive quarters. If there were no anomalies in the data in the first two quarters of 2022, this situation would be positive in the longer term.

As a sign of serious economic weakness, the current sharp revision is an outlier compared with data from the same period. As a sign, job growth averaged 68,000 per month during the correction, which is more or less accurate.

But this would only reduce average job growth from 242,000 to 174,000. How the Bureau of Labor Statistics addresses this weakness over a 12-month period will help determine whether revisions are more concentrated toward the end of that period, meaning they are more relevant to current conditions.

If that were the case, the Fed might not raise interest rates that high. If weakness persists beyond the correction period, the Fed’s policy may now become more accommodative. This is especially true if, as some economists expect, productivity numbers are higher because the same level of GDP appears to have been achieved with less work.

But inflation numbers are facts, and the Fed reacted to them more than employment numbers at the relevant time (and now).

Therefore, for the Federal Reserve, which is already leaning towards a rate cut in September, this revision may modestly increase the likelihood of a 50 basis point rate cut in September. From a risk management perspective, the data could fuel concerns that the labor market is weakening faster than previously thought. The Fed will monitor growth and employment data more closely as it cuts rates, just as it monitors inflation data more closely as it raises rates. But the Fed is likely to focus more on current jobless claims, business surveys and gross domestic product data rather than retrospective revisions. Remarkably, over the past 21 years, revisions were made in the same direction only 43% of the time. That is, 57% of the time, a negative correction is followed by a positive correction the year after, and vice versa.

Data organizations make mistakes, sometimes big ones. Even though the election was still three months away, they kept coming back to make corrections.

In fact, Goldman Sachs economists said late Wednesday they believed the Bureau of Labor Statistics may have overstated the revision by as much as 500,000 people. The Wall Street firm said illegal immigrants who are not currently in the unemployment system but were initially listed as employed accounted for some of the differences and that the overall trend in the initial revision was overstated.

Employment data can be affected by immigration hiring noise and can be volatile. But if the economy goes into a downturn like it did in 2009, a wealth of macroeconomic data will show signs of it. This is currently not the case.