Unemployment Creep and Recession Statistics

Click here to read the original Cautious Optimism Facebook post with comments

As of June 2024 the official U.S. unemployment rate is 4.1%, up 0.7 points from the April 2023 post-Covid low of 3.4%.

Why is this interesting?

First of all unemployment virtually never jumps through the roof right before a recession. That always happens during the recession itself (see attached St. Louis Federal Reserve chart).

https://fred.stlouisfed.org/graph/?g=1q0rV

Preview:

However one might notice that unemployment does creep up slightly right before recessions begin. When the recession starts, other economic indicators also suggest a slump may be starting but since unemployment remains relatively low—albeit not an absolute low—debate usually rages as to whether a recession is coming or not. The debate is usually settled six months later when it's obvious to everyone.

But here’s an interesting factoid about the recent 0.7% increase in the unemployment rate:

Going back to the 1980’s, the unemployment rate has never risen 0.7% from a cyclical low without entering a recession.

Before the 2007-09 recession, unemployment bottomed out at 4.4% in May of 2007. By the time it had risen 0.7 points to 5.1% it was March 2008 and the economy had just entered recession—although the NBER didn’t make its official recession announcement until eight months later.

Before the 2001 recession, unemployment bottomed out at 3.8% in April of 2000. By the time it rose 0.7 points to 4.5% it was June 2001 and the economy had also just entered recession. Once again the NBER didn’t make its official recession announcement until six months later.

Before the 1990-91 recession, unemployment bottomed out at 5.0% in March 1989. By the time it rose 0.7 points to 5.7% it was August 1990 and the economy had just entered recession. Once again the NBER announcement came much later.

The 2020 Covid recession was unique in that it was manmade with states shutting down their economies due to Covid. However the pattern remains: by the time cyclically low unemployment of 3.5% (February 2020) had risen 0.7 points to 4.2% (4.5% one month later) the economy was in recession, although given the sudden nature of the pandemic there was no “creep up” period prior.

Unemployment bottoming out and rising 0.7 points adheres to the same predictive pattern in the 1981-82 Volcker recession, the 1980 Jimmy Carter recession, the 1973-75 OPEC recession, the 1969-70 recession, and the 1960-61 recession—basically all nine recessions going back to 1960.

The first and only exception occurs in June 1959, 65 years ago, when unemployment rose from the June low of 5.0% to 5.8% by November without an official recession, although it was close: Q3 and Q4 GDP growth in 1959 was just +0.07% and +0.28% respectively.

The pattern then resumes without exception during the recessions of 1948-49, 1953-54, and 1957-58.

Additional Fed tables indicate the pattern continues without exception during the recession of 1945, and economic historians’ unemployment estimates continue to match the 0.7% pattern during the Great Depression: in the Depression of 1937-38 and Herbert Hoover's Great Depression contraction of 1929-1933.

One last “exception” (which really isn’t an exception) was a rapid 5% increase in unemployment during 1934 after Franklin Roosevelt signed the National Industrial Recovery Act. The slump was swift, rapid, and painful, but because contraction only lasted a few months it didn’t meet the “two consecutive quarters” definition and wasn’t officially recorded as recession. i.e. contraction began the last month of Q1 but Q1 as a whole was left barely positive, there was a massive contraction in Q2 which was dizzyingly negative, and contraction the first month of Q3 was followed by two months of growth leaving Q3 also barely positive—the net result being a 2007-09 sized Great Recession compressed into a five-month period that didn’t produce two consecutive negative quarters.

But it was a recession, and a big one at that.

So the pattern holds up for 15 of the last 15 recessions going back for a century, with a single false signal in 1959. And once unemployment was 0.7 percentage points above the cyclical low, the economy was already in recession in 15 of the last century's 15 recessions, not "about the enter one."

Disclaimer: Although the Economics Correspondent thinks the U.S. economy has entered a time window where the beginning of recession is highly likely—April 2024 to December 2024—he doesn’t claim this unemployment statistic guarantees recession has begun, only that, like so many indicators we’ve seen the last year or two, it’s yet again consistent with imminent recessions of the past.

Past performance is no guarantee of future results, but if the economy doesn't enter recession in the next 12 months this particular case will be a statistical unicorn.