The yield curve has been indicating recessions for at least 100 years.
The 1-year vs. 10-year yield curve suggests the next recession will start December 2020, give or take a few months.
The worst 12-month period in the stock market should be December 2020 through November 2021, give or take a few months.
A Recession Warning
We’re going to cut to the chase and show the chart behind the timing call of the next recession and then show the several charts that build to this one.
The chart shows the correlation of GDP growth with the difference between the 10-year Treasury yield and 1-year Treasury yield when it is less than 0.4 (40 basis points). The limit of 0.4 is used because below 0.4 there is a strong correlation with growth. Above 0.4, the yield curve has a poor correlation with growth. The correlation implies the recession will start about December 2020 and reach the weakest growth around August 2021. The red scale for the yield curve is calibrated to the correlation between growth and the yield curve with the 0.4 limit.
Quarterly GDP growth over the last three decades has its strongest correlation with the yield curve using a 25-month lead time. This lead time is the basis for calling the recession’s start around December 2020.
A Hundred Year Record
While the constant maturity 1-year Treasury yield only goes back to April 1953 at the FRED database, I have data for the 3-month yield back to 1920. The yield curve using the difference between the 3-month and 10-year yields has a hundred-year record.
It called the four recessions starting in the 1920s. It missed the four recessions when the Fed held short-term rates below 2.1% from the mid 1930s to the mid-'50s, although it came close to warning of the 1953 recession. It warned of the nine recessions since then. It now warns of another.
The correlation of growth with the yield curve using 1-year vs. 10-year yields is stronger than either the 3-month vs. 10-year yields or the 2-year vs. 10-year yields.
As with the first chart, the yield curve axis is calibrated to the correlation. Since the part of the yield curve above 0.4 has a poor correlation, the change in the yield curve is shown as having a small influence on growth. When we plot the yield curve with an upper limit of 0.4, the correlation gets stronger and shows a bigger influence or stronger indicator of the declines.
This yield curve also correctly warned of the last nine recessions. It gave one false call in 1967 when the annual growth rate dropped from 7.7% to 2.7%, but did not fall into recession.
The influence of the yield curve on the last 3 recessions is significantly different from the others in both lead time and magnitude. In the recessions from 1960 through 1982, the fit of the yield curve and the depth of the recession match rather closely. The last three recessions were much deeper than the fit of the yield curve for the period from 1955 to 2018 would suggest in the chart above.
If I fit the period since the mid-1980s, the lead time with the strongest correlation jumps from 18 months to 32 months.
This chart is very similar to the first one in the article except it uses annual growth rates whereas the first one uses quarterly GDP growth. Using the correlation with quarterly growth probably allows a better estimate of when the recession will start.
The difference in the relationship between the yield curve and growth since the mid-1980s may be a function of a different monetary policy and reduced volatility in GDP growth.
In 1978, the Fed mandate changed from just maximizing employment to a dual mandate that included stable prices. Prior to 1978, the Fed mandate often resulted in an effort to keep interest rates low. Over time this backfired in higher rates of inflation. With the dual mandate, interest rates were raised to curb inflation. This likely contributed to the back to back recessions of the early 1980s. After peaking in January 1980, inflation trended down. Since inflation came under control in the 1980s, the dual mandate likely contributed to more stable economic growth.
The standard deviation in the quarterly GDP growth rate has fallen substantially in recent decades.
Prior to 1986, the average difference in the growth rate from quarter to quarter was 4.2%. Since 1986, the average difference has been 2.1%. The biggest difference prior to ’86 was 15 when Q4 1970 shrunk at 4.2% and Q1 1971 grew at 11.3%. Since 1986, the biggest range was 7 in mid-2000.
The standard deviation in the quarterly growth rate for 4-year periods dropped below 2 for the first time in 1986. It dropped below 1 for the first time in 2017. It hit its lowest level ever at the end of 2018 and has hardly budged in the last three quarters.
If reduced volatility of growth explains why the last three recessions were deeper than the yield curve fitted to the 1955-2019 period suggested they would be, the record low volatility of the last few years could mean the next recession will be deeper than the yield curve fitted to the 1986-2019 period suggests.
Stock Market Declines
The yield curve with a 0.4 limit also indicates stock market declines. The correlation is with the 12-month percentage change in the S&P 500.
The correlation suggests the stock market (NYSEARCA:SPY) for the 12-month period from April 2020 to April 2021 will finish about where it started; of course, there could be significant fluctuations within the 12 months. The curve also implies the weakest 12-month period will run from November 2020 to November 2021. The curve as far as I can analyze does not narrow down whether the stock market for this business cycle is topping now or will top sometime in the next 12 months.
If GDP growth continues at 2% or below, it’s quite possible the market is topping now and that there will be modest stock market declines in the next year preceding the worst declines that will come later. It seems more likely the high will be sometime in the next year, but the potential upside is likely dwarfed by the likely downside in the coming two years.
Disclosure: I/we have no positions in any stocks mentioned, but may initiate a short position in SPY over the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: There is no guarantee analysis of historical data their trends and correlations enable accurate forecasts. The data presented is from sources believed to be reliable, but its accuracy cannot be guaranteed. Past performance does not indicate future results. This is not a recommendation to buy or sell specific securities. This is not an offer to manage money.