So much has been said about US labor market data in recent months. The deteriorating quality of survey data, the revisions of previous months’ data, etc. And all these issues are real issues that matter, especially for the Federal Reserve, which will meet again tomorrow to decide monetary policy. But one dimension, which has a significant impact not only on labor market data but also on US economic data in general, is getting short shrift.
Most economic data is calculated on an economy-wide basis. This includes the population. But population growth in the US is slowing, both because of demographic effects and now also because of the government’s crackdown on immigration. This means that, for example, the same level of GDP growth per capita will lead to lower GDP growth. Lower immigration figures in Europe are one of the main reasons why Europe has lower trend growth than, for example, the US.
Now put yourself in the shoes of the Fed, which must assess monetary policy with respect to inflation and unemployment. If population growth is slowing, it also means that the growth of the working-age population is also slowing. As a result, fewer jobs are needed each month to keep the economy at full employment.
Eat how much has calculated how many new jobs the US economy must create each month to keep unemployment stable. As long as monthly nonfarm wage wages are above this breakeven rate, the unemployment rate will remain stable or decline (assuming stable labor market participation). If nonfarm payrolls fall below that breakeven level, unemployment will begin to rise. The chart below shows that the U.S. needed to add about 120,000 to 160,000 jobs each month in 2023 and 2024 as a result of accelerated growth in the working-age population. But by June 2025, the required number of jobs had fallen to 86,000.
Break-even level for US non-farm payrolls to keep the unemployment rate stable
Source: Kolko (2025)
If we compare the actual number of jobs created per month in each of the previous calendar years, we can see that the US generated far more jobs than necessary between 2022 and 2024. But even with revised nonfarm payrolls in 2025, the pace of job creation is about what is needed to keep unemployment rates stable. What is of course worrying is the trend towards a slowdown in job creation, but we should be aware that the data currently do not point to a large increase in unemployment rates in the short term.
Actual nonfarm payrolls versus break-even rates needed
Source: Kolko (2025)
If the Fed does not take into account this slowdown in break-even rates, it will think the labor market is softening, when in reality it is not. The result would be a tendency to cut interest rates faster than necessary. Other macro data will reinforce this bias. As I said above, when overall population growth falls, real GDP growth falls, making the numbers look worse than they actually are compared to past averages.
In my post tomorrow I will discuss an unrecognized effect that has led the Fed to set interest rates that today are roughly one percentage point lower than they would have been had the Fed been truly unbiased in its decision-making. So we are already in a world where interest rates are “too low”. Now add to this the slowdown in equilibrium growth due to lower population growth, and the bias only increases. Not to mention the fact that the Fed is losing its independence and becoming a political tool for a president who thinks lower interest rates are the answer anytime, anywhere.
#macro #figures #structurally #worse