Are We in a Recession? Okun’s Law Says No
(9 min read)
(9 min read)
A recession is a prolonged period of generalized economic contraction. There are technical and academic definitions of recession. Technically, a recession occurs when quarterly GDP growth is negative for two quarters in a row.
Academically, a US recession occurs when the Business Cycle Dating Committee decides there is one. The Committee contains eight highly respected economics professors (some of whom have been members since 1978!).
For the Committee, the definition of a recession is ‘a significant decline in economic activity that is spread across the economy and lasts more than a few months’. The Committee has no fixed formula to determine if the economy is in recession but rather considers many economic indicators. That said, ‘in recent decades, the two measures we (i.e., the Committee) have put the most weight on are real personal income less transfers and nonfarm payroll employment.’
The Committee identifies dates of peaks and troughs in activity, and they define recession as the period between a peak and a trough. The committee waits until enough data is available to declare a peak or trough with confidence, therefore avoiding having to revise the business cycle calendar. In the past, the Committee has waited 4-21 months after a trough to call a recession.
Since WWII, two consecutive quarters of negative growth have always ended up being called a recession by the Committee. (And non-consecutive quarters of negative growth have sometimes as well; Chart 1.) However, I do not think the Committee will call a recession in 2022, despite the first quarter and second quarter showing negative growth. The reason is largely because of the strength of the labour market.
It makes sense that members of the recession Committee pay so much attention to employment in their dating process. Employment reflects business strength and drives labour income and therefore consumption, which represents 70% of the US economy.
A recession has never started when employment growth was as strong as it is currently, about 4.3% year-over-year (Chart 2).
Furthermore, labour demand (i.e., employment) growth is currently twice as high as labour supply (i.e., the labour force, the sum of employed and unemployed workers) growth. This implies stable or falling unemployment.
Unemployment typically rises before the start of a recession. The economist Claudia Sahm has found that since the 1970s, all recessions except one were preceded by an increase in unemployment of at least 50 basis points (0.5%) relative to the low of the previous 12 months (Chart 3).
That said, current labour market strength needs to be put in perspective. The economy is rebounding from an unprecedented increase in unemployment. The strength of current labour demand reflects the depth of the pandemic decline and self-sustaining economic dynamism. In that sense, we cannot take current labour demand strength at face value and a higher standard of evidence is needed to gauge the economy actual strength.
This standard is Okun’s law.
Okun’s law is an empirical relationship between unemployment and growth. It says that, in the US, a 1% deviation of unemployment from trend requires a 2% deviation of GDP from trend. The relationship was first highlighted by Arthur Okun, a member of President Kennedy’s Council of Economic Advisors in the 1960s.
The intuition behind Okun’s law is that businesses tend to adjust their workforce with a lag, relative to their turnover. For instance, during a recession, it makes sense not to immediately fire redundant workers as firms need to ascertain the strength of the downturn. In addition, hiring workers is costly, so firms tend to initially keep more workers than they need during recessions. If they did not, they could struggle to rebuild their workforce once the recession is over.
Okun’s law has proven surprisingly robust. It appeared to break down during the GFC, only to be proven right after GDP revisions.
On Chart 4, I have plotted a very simple Okun law, using the CBO estimates of trend unemployment (U*) and growth. My very rough and unsophisticated estimate shows an Okun coefficient of about -1.8.
This suggests GDP growth in the first half of 2022, which has led many to say the US is in a recession, has been underestimated. First, the distance between the first- and second-quarter data points (red dots) relative to the average relationship before the pandemic (the ‘trend’ line) is large.
Second, the CBO estimates that U* has been falling since the pandemic, which is very likely wrong (Chart 5). U*, the natural rate of unemployment, is the rate consistent with stable and low inflation. A key driver of U* is frictional unemployment, i.e., unemployment caused by workers moving between jobs. Frictional unemployment is likely to be higher when the economy is going through restructuring and workers need to move from contracting to expanding sectors.
The US economy is currently restructuring as it adjusts to the post-pandemic world. We can see this, for instance, in the unusually high vacancies relative to unemployment. We can also see it in the rising initial unemployment insurance claims that are not matched by an increase in continuing claims. That is, workers are getting laid off but quickly finding new jobs. The falling ratio of total to initial claims demonstrates this (Chart 6).
Therefore, the difference between unemployment and U* is likely wider than suggested by the CBO’s current estimate of 4.4% for U*. For instance, if U* in Q2 2022 had increased by 0.25ppt relative to Q2 2011, the GDP growth implied by Okun’s law for Q2 2022 would be about 4.5% YoY, against Q2’s 1.6%.
These numbers are only illustrative, and we must not take them at face value. But they highlight the disconnect between labour market and GDP data. This suggests an upward revision in the GDP is coming – though it could take several years. (GDP updates are carried out through three estimates, an annual update of the five most recent years, and a comprehensive update carried out every five years.)
We might therefore find out later that the US was not in a recession in 2022. But for now, two quarters of negative GDP growth means we are in a technical recession.
At Macro Hive, we think there is more chance of a proper recession in 2023 than in 2022. Although the US is technically in a recession now, a more severe economic downturn is likely next year due to the considerable amount of interest rate hiking the Fed must still undertake to combat inflation.
At Macro Hive, we are currently underweight stocks in general, although we see more opportunity in the S&P 500 than the FTSE 100 or Euro Stoxx 600. Traditionally, the best recession stocks tend to be in sectors that people cannot cut back on during hard times, such as healthcare. Consumer staples were also traditionally though of as good recession stocks, but we are currently underweight as they seem to be struggling to pass on rising costs to consumers. This could change as inflationary pressures ease, however. As of mid-2022, energy stocks look like they will outperform. For all the details on our best stock market picks right now, see here.
A recession is an extended period of broad-based economic contraction. It is commonly defined as two quarters of negative growth.
Technically, yes, the US is in a recession this year because GDP growth was negative in the first and second quarter. But labour market strength suggests this reflects measurement errors or data quirks and that growth is likely to turn positive again in H2 2022.
Okun’s law is a robust empirical relationship between unemployment and growth. It says that a 1% change in unemployment relative to trend is associated with a 2% change in GDP relative to trend.
Although the US is technically in a recession for 2022, we think a more severe economic downturn is likely in 2023. This is because the Fed will need to keep hiking interest rates to curb inflation, which will probably cause a hard landing. But we are not there yet.