Sahm ado about nothing?

Stay informed with free updates

You may have heard about the ‘Sahm Rule’ lately. It’s become the trendiest advanced indicator for economic recessions, though we’ll concede that competition in that category is a little light.

Coined by former Fed economist Claudia Sahm, it stipulates that a recession is coming when the three-month moving average of the US unemployment rate rises by half a percentage point off its 12-month low.

While people are still debating the usefulness of the yield curve (both if it works and if so why it works) the simpleness, intuitiveness and accuracy of the Sahm Rule in predicting recessions has meant it is the new hot thing.

Oh and here’s another reason it’s getting a lot of attention recently.

Line chart of US unemployment rate (%) showing Sahm Rule flickering amber

The US unemployment rate increased to 3.9 per cent last month — up from a low of 3.4 per cent — and the three-month average is now 0.33 percentage point above its April low. Cue lots of headlines about how “the jobs market is close to a major recession warning” and “the recession rules are close to being triggered”.

It should be noted that Sahm herself has repeatedly stressed that it’s just an indicator that could easily “break”, just like other economic rules of thumb. Here’s what she told former Alphavillain Colby Smith recently:

It is not a law of nature. Just because it worked in the past to signal early in a recession does not mean that it will necessarily work this time, because all kinds of empirical regularities have broken down in the post-pandemic recovery. 

So the Sahm Rule breaking would be if it hits half a percentage point, which would be consistent with unemployment running about 4 per cent for three months, but we don’t see a broad-based contraction. In fact, if you look at the forecasts that Federal Reserve officials have been writing down for quite some time now, essentially they have the Sahm Rule being triggered but no recession: the unemployment rate rises above 4 per cent and then it goes sideways.

You can tell a story right now as to what keeps it in bounds, but we’ve never seen it. The impossible is possible though and that’s been the theme of this year. The other empirical regularity had been the two quarterly declines in GDP growth, and that happened and we didn’t have a recession.

Analysts at Goldman “everything is awesome” Sachs are among those that are pretty unworried about the recent unemployment uptick.

While they expect underlying job creation to slow down from the current pace of ca 175,000 over the next year, it will remain above the 100,000 monthly pace to keep unemployment stable until the end of 2024.

“We don’t expect the recent uptick to presage an upward trend in the unemployment rate, for several reasons,” they argue. Here are their four main reasons.

First, the broader set of labor market data suggests that job growth is likely to remain strong. Job openings remain well above their 2019 levels in virtually every industry, and both the layoff rate and initial jobless claims remain low.

Second, the increase in the unemployment rate since April has come entirely from an increase in the size of the labor force, not a decline in employment that could set off the vicious circle between job loss and reduced spending that often leads to recession.

Third, the weakness in household employment has been concentrated in the most volatile and poorly measured parts of the labor force, such as self-employed workers and young workers. Indeed, higher unemployment rates for workers under 24 account for two thirds of the overall increase in the unemployment rate since April, even though these workers make up only 15% of the labor force.

Fourth, we estimate that residual seasonality depressed household survey employment growth by 100-150k in October and 40k on average over the last six months.

Goldman’s Manuel Abecasis concedes that technical factors like the birth-death model (how the creation and closure of new businesses are calculated) could be overstating the underlying pace of jobs growth by about 40k.

But Abecasis notes that jobs growth in the payrolls survey has recently been much stronger than in the household survey, and that household employment tends to converge to payroll employment over time. Which gives him confidence that things are still fine.

Not everyone is as sanguine, of course. BCA’s Peter Berezin — of “risk of Armageddon has risen dramatically. Stay bullish on stocks over a 12-month horizon” fame — sees “clouds on the horizon” (not nuclear mushroom clouds this time though).

Employment, as measured by the household survey, decreased by 40,000 over the past three months. While payroll growth has been robust — averaging 204,000 over the past three months — this can partly be attributed to an increase in the number of multiple job holders (which adds to the payroll survey, but not the household survey).

Moreover, payroll growth in October was heavily tilted towards a few industries, such as health care and government, that had become woefully understaffed. The breadth of job gains, as measured by the fraction of industries experiencing positive payroll growth, dropped to its lowest level since April 2020.

Worryingly, Berezin argues that this is part of a much wider trend of “slower employment growth, lower job openings, and in some cases, modestly higher unemployment” across most developed markets.

Although it’s still early days, that reinforces BCA’s view that the global economy will suffer a recession in 2024. So perhaps the Sahm Rule will notch up another successful prediction in the coming months.