Inverted Yield Curve (Final Part) – Is a Recession Coming in 2025?

Inverted Yield Curve - Is a Recession Coming in 2025?

Now, let's summarize the data from the inverted yield curve cases we've examined so far since 1960. The most important question is: “The yield curve is currently inverted. Does this mean that a recession will happen soon? If so, when and for how long?” I will attempt to answer this question to the best of my ability. What matters more than the conclusion itself are steps that lead to it so I encourage you to follow the discussion below closely to understand the logic leading to the final outcome.

Meantime, if you're curious about what an inverted yield curve is and why it is linked to recessions, as well as the earlier cases discussed, I recommend reviewing Part 1~3 first.

• Inverted Yield Curve (Part 1) - 2020, 2008, and 2001
• Inverted Yield Curve (Part 2) - 1990, 1981, and 1980
• Inverted Yield Curve (Part 3) - 1974, 1970, 1961 and exception!

Summary of Yield Curve Inversions since 1960

First, I’ve organized the 10 precedents we’ve looked at so far into a table:

• Year = The year when the recession occurred (e.g., the recession from Dec. 2007 to May 2009 is noted as 2008. However, 1961 is the year when the yield curve inversion occurred but no recession.)
• IP = Duration of the inverted yield curve (in months)
• R = Whether a recession occurred
• RP = Duration of the recession (in months)
• a = Time from the inversion to the onset of a recession (in months)
• b = Time from the yield curve normalization to the onset of a recession (in months)

In the four most recent cases, the recession occurred after the inversion was resolved. In the four cases before that, the recession occurred first, and then the yield curve returned to normal. Out of the 9 cases of yield curve inversions since 1960, only once (in 1966) did a recession not occur. In other words, when the yield curve inverted, the probability of avoiding a recession was only 11%. It seems fair to say that when the yield curve inverts, the likelihood of a recession is very high.

Out of the 9 recessions since 1960, there was only one case (1961) where a recession occurred without a yield curve inversion (although this case is somewhat debatable because the yield curve nearly reached zero). The probability of a recession without a yield curve inversion was also 11%. If we count the 1961 case, where the yield curve flattened to zero (but didn’t go negative), as half a case, this probability drops to 6%. Thus, it can be said that a recession without a prior yield curve inversion was a very rare occurrence.

Current 10-Year and 3-Month Treasury Bond Yield Spread

In Part 1 of the series, I started by confirming that as of end August 2024, the U.S. yield curve between 10-year and 3-month Treasury yields was inverted. But aren't you curious about how long this inversion has lasted? Let’s take a look at the graph below.

Source: https://fred.stlouisfed.org/series/T10Y3M

As of August 28, 2024, 10-year Treasury bond yield minus 3-month Treasury bond yield has been negative for 673 days (22.1 months) since the inversion on October 25, 2022. The negative spread is currently at -1.37%. The previous longest inversion lasted 18 months. Since the current inversion hasn’t normalized yet, it will easily surpass the previous record.

Is there Correlation between the Inversion Duration and the Length of Recession?

Aren't you curious if there is a correlation between periods of inverted yield curves and economic recessions? (I’m even more curious about the relationship between the depth of the inversion and its negative impact on the stock market, but that would require integrating the depth and width of the inversion, so we’ll skip that for now.) Even though the sample size is only 10, I ran a regression model, which may lack statistical significance. However, the R-squared value was only 0.17, meaning there wasn’t a significant correlation between the duration of the inversion and the length of the recession.

If the exceptional case of 1961 is excluded, the R-squared value increases to 0.28 (as a reference, I also tried using the time from the inversion to the onset of the recession and from the normalization to the recession, but neither produced an R-squared value above 0.15, so I discarded those models). While 0.28 is still low to be considered statistically significant, here are the results from the regression model:

(Length of the recession) = 2.3 months + 0.65 × (Duration of the yield curve inversion)

What Would It Mean for a Potential Upcoming Recession?

As we saw earlier, the inversion has lasted 22.1 months as of the end of August 2024. Let’s assume the following: (i) the Federal Reserve starts cutting rates at each Federal Open Market Committee (FOMC) meeting from September onward, (ii) one of these rate cuts will be a 0.5% reduction, with the others being 0.25%, and (iii) the inversion gradually resolves in proportion to the extent of Fed rate cuts. With these assumptions, 4.5 rate cuts will be needed for the inverted yield curve to normalize. The fourth Fed meeting will be on 29 January 2025 and the fifth one is on 19 March 2025. The inversion could resolve around 22 February 2025 (in the middle of the 4th and 5th FOMC meetings), meaning the inversion could last 28.0 months. Plugging this into the regression model gives us an estimated recession duration of 20.4 months. The most severe recession since the Great Depression was the 2008~2009 global financial crisis, which lasted 18 months. The 20.4 months exceeds that.

Before accepting this conclusion, I want to reiterate three caveats: (1) a regression model with an R-squared value of 0.28 has low reliability; (2) the number of past recession cases is far too small for a statistically significant model; and (3) the model excludes the exceptional case of 1961. So, please take this analysis with a big grain of salt and consider it for a discussion purpose only.

What I haven’t addressed yet is whether a recession will occur soon and if so, when. While history doesn't repeat itself, I believe past examples are the best guide that we  have. Based on these examples, I would argue that a recession is more likely to happen than not. Of course, there’s a small chance (11%, based on past data) that we could avoid a recession, which would be a "soft landing." However, “this time it's different” is the four most dangerous words in investing, according to Sir John Templeton.

When Could the Next Recession Occur?

The average time from inversion to recession was 11.2 months, and from yield curve normalization to the onset of recession, it was 0.3 months. Given that the current yield curve inversion hasn’t been resolved yet, the first metric doesn’t apply. If we apply the second observation, and if the inversion normalizes around February 22, 2025, the recession could start around March 3, 2025. While this date is certainly not exact at all, it might point to a recession beginning around late Q1 2025. Adding the previously calculated recession duration of 20.4 months brings us to November 14, 2026 (again, please note that this date is purely for a discussion purpose and is not precise at all.)

In summary, we could imagine a scenario where the yield curve inversion resolves in Q1 2025, and the recession begins shortly after, lasting through Q4 2026. Of course, the actual length of the recession could be shortened if the Federal Reserve or U.S. government implements aggressive policies in response, although that might lead to a bubble before the economy fully recovers. Again, the regression model’s reliability is low due to various assumptions and small sample size. So, how one interprets this is completely up to each reader.

The Other Yield Curve Inversion: 10Y UST - 2Y UST

I have used the "U.S. 10-year Treasury bond yield - 3-month Treasury bond yield" so far while I mentioned in Part 1 that the "10-year / 2-year yield spread" is also frequently used for inverted yield curve discussion. Before wrapping up, let me show a chart of the 10-year / 2-year yield spread.

Source: https://fred.stlouisfed.org/series/T10Y2Y

As you can see, the yield curve inversion has already normalized for 10-year vs. 2-year. The first day the U.S. 10-year Treasury yield fell below the 2-year Treasury yield was July 6, 2022, and the inversion ended on August 27, 2024. The yield was 0.0% on August 27 and +0.01% on August 28. While the 3-month yield curve inversion still has a long way to go, the 2-year inversion has already ended (+0.06% as of Sep 6, 2024). In my view, the "10-year to 2-year" spread responds more quickly to change of interest rate environments, while the "10-year to 3-month" spread more accurately reflects the market impact. In any case, one yield curve inversion has just normalized, and the market is anxiously awaiting the resolution of the other.

Thanks for reading. Wish you grow rich slowly and surely!

This blog does not offer investment, financial, or advisory services. The information provided herein is for general informational purposes only and should not be interpreted as advice for making specific investment decisions. All investments involve risk, and past performance is not indicative of future results. It is advisable to consult with a qualified financial advisor to determine strategies or products that are appropriate for your individual circumstances.

The owner, writer, or operator of this blog accepts no responsibility for any direct or indirect losses that may arise from the use of or reliance on the content presented. The information provided on this blog is subject to change and may not be current. The content is based on personal opinions, as well as information from news sources and research, and may vary due to shifts in personal opinions, financial market conditions, or other influencing factors.

Most data referenced is derived from daily closing prices. Data finding, sorting, graphing, and analysis are performed primarily by myself, and while efforts are made to ensure accuracy, errors may be present. If you identify any inaccuracies, please inform me via comments or email, and I will endeavor to review and correct them as necessary. External content or images will be cited to the best of my ability.