We are now 48 months since the start of the COVID downturn, which was the deepest economic shock since the Great Depression. We saw monthly unemployment hit 14.8 percent, with weekly estimates rising above 20 percent. By comparison, peak unemployment nationally during the Great Depression rose to just over 24 percent in the tough summer of 1933.

Following the widespread availability of the COVID-19 vaccine, the unemployment rate dropped precipitously. It has remained below 4 percent for two full years, the longest stretch in more than 50 years. Readers might be tempted to say, “So what? It is still bad that nearly one in 25 people are not working.”

That view seems sensible, but it turns out that a healthy economy has significant amount of underlying job churn. People change jobs to earn more money, to change a career or to dodge a bad boss. Businesses shed workers because they may be dissatisfied with a particular employee, or because they find a location or line of business unprofitable.

Most folks worry about the unemployment rate during recessions, but even in good times something like 4.5 to 5.5 percent of workers would be changing jobs each month. Because economists have a flair for sensitive language, we call this the ‘natural rate’ or ‘non accelerating rate’ of unemployment. We are today well below that level, which has several implications.

One result of very tight labor markets is wage growth. Very low unemployment rates imply a ‘sellers market’ for workers, who can more readily negotiate for wages or other job benefits. Real, inflation-adjusted wage growth can only happen if labor productivity is rising, but productivity rarely rises when the unemployment rate is very low. The reason for this is that a low unemployment rate typically means that the economy is performing so well that even the very least-skilled workers have a job. Good times also mean that low-productivity firms can stay in business.

Altogether, this means that during periods of low unemployment we don’t typically see strong productivity growth. This is particularly true as the recovery lengthens.

In 2020 and 2021, this business cycle looked just like previous downturns and recoveries. However, productivity growth has been improving for two years. Over the first three quarters of 2023, labor productivity grew at a shocking 4 percent rate. That recent productivity growth boosted both GDP and wages in late 2023. It should continue into 2024.

Labor productivity is dependent mostly on worker talent and education. But, there are other factors that affect productivity growth, primarily new technology. We economists aren’t yet sure why labor productivity is growing, but I can offer two educated guesses.

The first of these is demographics. For a variety of reasons, most people tend to hit their productivity peak about age 30 and remain there until their late 50s. On average this is when folks have finished their formal education or apprenticeship and embarked upon a career. Also, in the romantic language of economists, they have begun a period of ‘family formation’ that has an especially beneficial effect on the productivity of men.

The four biggest 5-year cohorts of American workers are aged 30 to 50. These are the Millennials and Gen-Xers. The movement of this age cohort into their most productive years of work is enough to explain some, but not all of the productivity growth. The effect is too abrupt to have a purely demographic explanation.

The most likely cause for increased productivity growth is the effect of accumulated investment in productivity-enhancing technologies. These technologies run the gamut of workplace applications from robotics to artificial intelligence to plain old digital interfaces. All of these replace tasks previously performed by workers.

The last two decades saw continued improvement in information technology, but business statistics didn’t reflect a great deal of investment in these technologies. One reason for that is simply that the price for computer technology has been declining for 40 years. So, data on spending cannot capture the value these items add to production.

However, the widespread adoption of workplace technologies continues to be apparent everywhere but the investment data. Today, most service-providing industries use customer interface technologies. These seem small, but paring one worker per shift could raise overall productivity by 15 percent.

Altogether, this makes me think we might be in one of those bursts of productivity growth that accompanies new technologies. How and where it plays out, I’m sure I don’t know. If we are in a period of sustained productivity growth, we are in the beginning of a strong economic environment that has eluded us for the past quarter century.