The economic decline that followed the COVID-19 epidemic was even “predicted” by it (although most economists attribute this to luck, and not the fact that it can predict natural disasters).
When does an inverted yield curve signal the start of a recession?
Every five years, the US economy experiences a recession. As a result, an inverted yield curve that forecasts recession three years in advance is similar to a stopped clock that is correct twice a day.
In other words, the median duration between the initial inversion of the yield curve and the commencement of a recession has been 18 months over the previous six decades, according to Brian Levitt, global market strategist at Invesco. Here are a few examples of why the curve isn’t very useful as a leading indicator:
- The next recession didn’t come until 1969, or 48 months after the yield curve inverted in 1965.
- In March 2001, the recession triggered by the burst of the IT boom began. The yield curve, on the other hand, had inverted 34 months earlier, in May 1998.
Do you think the yield curve can help you foresee a recession?
The yield curve usually slopes higher, reflecting the fact that longer-term debt holders have taken on more risk. One of the most accurate leading indications of a recession is an inverted Treasury yield curve.
Should we be concerned about a yield curve inversion?
On Tuesday, the yield on the 2-year Treasury note TMUBMUSD20Y,2.593 percent briefly jumped above the yield on the 10-year Treasury note, in what is known as a yield curve inversion, a rather rare occurrence.
Long-term bonds have higher interest rates, or yields, than short-term bonds, such as the 10-year Treasury TMUBMUSD10Y,2.385 percent.
While yield-curve inversions are significant economic indicators that might herald the likelihood of a prospective recession, market experts point out that they do not inform investors when an economic downturn is likely to occur.
An inverted yield curve is only a prediction of a recession if it remains inverted for more than a week, according to Barron’s. Furthermore, according to a research published by the Federal Reserve Bank of San Francisco in 2018, the duration between the inversion and the subsequent recession ranged from 6 to 24 months.
“When 2-year yields exceed 10-year yields, market observers may safely assume only one thing: Investors expect interest rates and/or inflation to be higher in 2 years than in 10 years,” according to Barron’s.
What does it mean for the economy if the yield curve inverts?
For the first time since 2019, 2-year Treasury yields have surpassed 10-year Treasury yields.
This is uncommon since investors usually expect a higher reward for taking on the risk that rising inflation will reduce the expected yield on longer-term bonds. As a result, a 10-year note usually pays out more than a 2-year note.
Inverted curves have historically predicted recessions and can serve as a warning indicator. The Federal Reserve of the United States has begun raising interest rates and is projected to do so strongly until 2022.
In certain ways, yes. When the curve is sloping, banks borrow short-term and lend long-term, making money on the difference in rates.
There is no spread to earn between borrowing for two years and collecting interest on 10-year Treasuries if the two-year and 10-year Treasury yields are inverted.
In practice, however, banks borrow and lend at diverse locations along the curve, with average loan and security maturities of fewer than five years.
At two years, they rarely borrow much and lend at ten years. They are more likely to borrow and lend near the front, or short-term, end of the steep yield curve. On Tuesday, the gap between the 3-month and 5-year Treasury notes, as depicted on the Treasury curve, was around 190 basis points US3MUS5Y=RR.
JPMorgan Chase & Co (JPM.N), for example, funds more than half of its balance sheet with low-cost deposits, which have a modest rate of increase. In the fourth quarter, the average rate for all of JPMorgan’s interest-bearing liabilities was merely 0.22 percent. That’s a far cry from the 2.4 percent yield on 2-year Treasuries US2YT=RR on Tuesday.
Many commercial and industrial loans are also floating-rate term loans, or revolving loan facilities with floating rates related to short-term benchmarks, which have risen dramatically this year in expectation of Fed rate hikes.
Banks have predicted that rate hikes will significantly improve their net interest income this year.
find out more
The greater threat to banks is the possibility of a recession, which would reduce consumer spending and make it more difficult for Americans to repay their debts.
What causes an inverted yield curve?
When the longer term yields fall substantially quicker than the short term yields, the yield curve inverts. This occurs when long-term government bonds (such as the 10-year US Treasury bond) are in higher demand than short-term bonds. The price of longer-term bonds rises in tandem with the demand for these products. Bond yields are inversely proportional to bond prices: as the price rises, the yield lowers. Short-term bond prices decrease and rates rise when investors shift their money to longer-term bonds by selling their shorter-term bond holdings. As a result, the yield curve is inverted.
When was the last time the yield curve inverted?
On Tuesday, a carefully watched gauge of the yield curve, which serves as one of the bond market’s most reliable recession indicators, inverted, highlighting concerns about the economy as the Federal Reserve considers raising interest rates quickly.
The widely studied differential between 2-year TMUBMUSD02Y,2.464 percent and 10-year Treasury yields TMUBMUSD10Y,2.385 percent has gone below zero on many occasions and is down from around 160 basis points a year ago. The last time the spread flipped was on August 30, 2019, according to Dow Jones Market Data data at 3 p.m.
Traders are reacting to the likelihood that, in order to combat inflation, the Fed would need to deliver a larger-than-normal half-point rate hike, and potentially more, shortly. Chairman Jerome Powell of the Federal Reserve opened the door to boosting benchmark interest rates by more than a quarter percentage point at a time earlier this month, a view shared by other officials. Some argue that signs of progress in the Russia-Ukraine peace talks allow the Fed to tighten as needed.
Despite a rebounding stock market, “bond markets continue to reflect mounting pessimism regarding the outlook for economic growth,” according to Mark Haefele, chief investment officer at UBS Global Wealth Management.
“The likelihood of an abrupt slowdown or recession has increased,” he said in a note Tuesday, “together with the prospect of a faster sequence of Federal Reserve rate hikes and disruptions related to the crisis in Ukraine.”
Because of its predictive power, investors pay close attention to the Treasury yield curve, or the slope of market-based yields across maturities. According to Principal Global Investors, every recession since the 1950s has been preceded by an inversion of the 2s/10s. That was true of the early 1980s recession, which followed former Fed Chairman Paul Volcker’s inflation-fighting efforts, and the early 2000s downturn, which was marked by the dot-com bubble bursting, the 9/11 terrorist attacks, and various corporate-accounting scandals as well as the 2007-2009 Great Recession, which was triggered by a global financial crisis, and the brief 2020 contraction fueled by the pandemic.
Economic downturns, on the other hand, tend to lag behind 2s/10s inversions. Anshul Pradhan and Samuel Earl of Barclays BARC,-0.85 percent said in a note Tuesday that the lag “has been roughly 20 months, and in several instances, it has been greater than two years.”
Inversions have already occurred elsewhere on the US Treasury curve. The spread between the 5- and 7-year Treasury yields and the 10-year Treasury yield, as well as the difference between the 20- and 30-year yields, have all remained below zero.
According to Ben Emons, managing director of global macro strategy at Medley Global Advisors in New York, the 2s/10s spread has been flattening at a quicker rate than it has at any time since the 1980s, and has moved closer to zero than at similar stages throughout previous Fed rate-hike campaigns.
Normally, the curve does not approach zero until after rate hikes have been implemented. Even with only a single quarter-point rate boost under the Fed’s belt, it ended up there.
The graph below, created in February, compares how long it took the 2s/10s to invert ahead of previous recessions to the current pace. This time, the 2s/10s spread has gone toward zero in a matter of months, rather than the years it took during its previous two forays into negative territory.
Is the yield curve currently inverted?
The yield curve in the United States is not inverted now, but it has become much less steep in recent months. Today, the 10-year and 2-year US Treasury bond yields are separated by 42 basis points. The spread was treble that in March 2021. As a result, we could be on the verge of seeing an inverted yield curve, especially given how quickly the yield curve has flattened in 2021.
What if the central bank used Operation Twist to reverse the inverted yield curve, which is a sign of impending recession?
Central banks can sell long-term bonds and buy short-term bonds, increasing long-term bond yields while decreasing short-term bond yields. In this approach, the inverted yield curve can be transformed into a normal-looking ascending slope, masking the true recession indicator.
Why does a yield curve that slopes downward usually signal a recession?
The slope of the Treasury yield curve has been shown in numerous studies to be a reliable predictor of recessions. 2 The empirical data in figure 1 displays the term-structure slope, which is assessed by the spread between the yields on ten-year and two-year US Treasury securities, as well as coloring that signifies US recessions (dated by the National Bureau of Economic Research). Since the 1970s, the yield-curve slope has grown negative before every economic downturn. 3 In other words, a “inversion” of the yield curve, in which short-maturity interest rates exceed long-maturity interest rates, is usually associated with a near-term recession.
: Ten-to-two year yield-curve spread
Previous research has used statistical models like probit specifications to estimate recession probability based on this empirical regularity. Figure 2 depicts a probit analysis that demonstrates the fitted likelihood of a recession over the next year when the explanatory variable is the ten-to-two-year yield-curve spread. With the exception of a false positive in the mid-1960s, the fitted likelihood peaks before the start of each recession.
Estimated recession probabilities, long-spread model
While the literature has discovered that this variable has predictive value, it has been less successful in determining why such an empirical link exists. Financial experts appear to believe that the slope of the yield curve contains information regarding present and predicted future monetary policy actionsi.e., the Federal Reserve’s raising or lowering of the federal funds ratewhich are linked to predictions of future business cycle outcomes. However, because the yield curve is impacted by factors other than monetary policy expectations, this is an imperfect picture. The yield curve, in instance, indicates market perceptions of various risks, which are influenced by economic outcomes. If the discrepancies between long-run and short-run values for policy expectations or risk variables represent market judgments of the likelihood of a future recession, then they will be useful in forecasting recessions to the extent that these assessments are right or self-fulfilling.
Furthermore, both the predicted interest rate path and the risk premium have numerous components. Market participants’ expectations for the future evolution of inflation and real (inflation-adjusted) interest rates impact expected rates. Similarly, bondholders’ risk reflects uncertainty about future inflation as well as the path of future real interest rates. The slopes of these multiple components of expectations and risk premia could reveal diverse information about future economic scenarios, which could help projections.
We investigate the differing effects of these channels on the estimated chance of a recession in this Chicago Fed Letter and find that they actually have diverse influences. We also provide some economic interpretations for these yield curve-to-broader macroeconomic correlations.
The yield-curve slope and recession risk
Many different metrics of the yield-curve slope, or term spread, have been studied in the literature. The difference between the yield on the ten-year Treasury note, which reflects bond investors’ long-term views, and the three-month Treasury bill rate, which is a close substitute for the federal funds rate targeted by the Federal Reserve’s policymaking body, the Federal Open Market Committee, has frequently been used in academic studies (FOMC). The difference between the ten- and two-year yields, on the other hand, is frequently used by practitioners and financial analysts. 4 Both criteria have merit, and when used in statistical models for recession predictions, they generate qualitatively identical results.
Why do these yield-curve slopes appear to be useful in predicting recessions? Remember that the interest rate on a long-term bond is influenced by the predicted path of short-term interest rates over the bond’s tenure. Views on the business cycle and monetary policy, in turn, influence the projected course. If market investors predict a downturn, they will almost certainly expect the FOMC to decrease its policy rate in the near future to provide monetary policy relief. Longer-term rates fall as the expectation of lower future rates falls, perhaps resulting in an inverted yield curve. A further explanation is that market investors may believe that the FOMC’s aggressive monetary policy tightening, which would raise current rates compared to future rates, would increase the odds of a future economic downturn. Such movements in the yield-curve slope will be associated with a higher probability of a future recession to the extent that the market’s expectation of a downturn is right.
Engstrom and Sharpe (2018)5 investigate an alternate yield-curve slope metric, the difference between the six-quarter-ahead forward rate on US Treasury securities and the present three-month Treasury bill rate, to better isolate this monetary policy channel “forward spread in the near term.”
6 They suggest that the six-quarter-ahead advance rate is more accurate than spreads based on the ten-year yield in identifying important market expectations of future monetary policy measures. This is because the long-term yield is a weighted average of forward rates over ten years, diluting the signal in shorter-term rate changes associated with business cycle variations. Engstrom and Sharpe (2018) find that the near-term forward spread crowds out other slope measures in probit models that predict U.S. recessions, and they conclude that a negative near-term spread may only predict recessions because it reflects the market’s expectation that a contracting economy will cause the Federal Reserve to lower its policy rate.
However, monetary policy expectations may not be the only link between the yield curve slope and future economic activity. Changes in the slope could be influenced by shifts in market participants’ risk attitudes, and these shifts could help predict economic downturns. We can use a dynamic term-structure model (DTSM) to break down the nominal yield y on a Treasury security of any maturity into components associated with expected inflation, expected real interest rates, and the risk premia that investors require to invest in a security that exposes them to inflation and real interest rate risks in order to study this second channel:
where $$ is the average expected path (EP) of consumer-price inflation throughout the bond’s life and $$ is the inflation-adjusted real rate’s average expected path over the same horizon. The expected path of the nominal interest rate, $$, is the total of these components. The bond’s yield additionally includes an inflation risk premium $$ and a real rate risk premium $$, which reflect compensation for the risk of future inflation and real interest rate movements, respectively. The term premium, $$, is the overall compensation that investors want in exchange for bearing interest rate risk. Finally, an error term $$ may cause the observed Treasury yield to diverge from the DTSM-implied value; because $$ is modest, we ignore it for the rest of the analysis.
We can deconstruct the slope of the yield curve into its expectations and risk-premia components using equation 1:
Fluctuations in each of these parameters may be linked to various growth potential. We investigate this issue by constructing these distinct components using an estimated DTSM and then allowing these factors to have varied affects in a probit model that estimates the likelihood of the US economy entering a recession within the next year. Benzoni and Chyruk provide more information on the analysis (2018). 7
The exact breakdown described in equations 1 and 2 is not used. We use slope $$ terms that concentrate on near-term spreads because the expected monetary policy explanations focus on a shorter time frame than ten years. We utilize the difference between the model’s six-quarter-ahead inflation forecast and its projection of average inflation over the next three months for slope $$, and the difference between the model’s six-quarter-ahead and current three-month forecast for real interest rates for slope $$. Changes in risk attitudes become more obvious in the pricing of longer-term bonds for a variety of reasons. 8 We utilize the difference between the ten-year and two-year $$ and $$ terms from the DTSM model for slope $$ and slope $$. Finally, we adjust for current monetary policy by inserting the DTSM-implied estimate of the current real rate, following Harvey (1988) and Estrella and Hardouvelis (1991). 9
We discover that monetary policy loosening, whether present or predicted, is linked to a higher risk of future recession. A decrease in the spot real rate or a decrease in the near-term spread in the real projected rate path, all else being equal, increases the model’s chance of recessions. Statistics show that the impacts are statistically significant.
The predicted inflation component’s spread has little effect on model fit. As a result, it is not included in the baseline probit definition. This data supports the hypothesis that inflation expectations were relatively steady across the post-1985 sample period, even while recession risk was high. The recession signal is contained in the anticipation of future monetary policy, as assessed by changes in the slope of the predicted real-rate path.
Long-term risk premia changes provide predictive information as well.
10 A decrease in the $$ slope raises the chances of a recession. Why is this the case? During periods of low inflation, the fixed nominal cash flows from a nominal bond become more appealing, causing bond prices to rise and interest rates to fall. The risks of surprisingly low inflation have increased in recent recessions, compared to the risks of unexpectedly high inflation. 11 As a result, if investors believe the chances of a recession are increasing, the long-term inflation risk premium in Treasury bonds will decrease. Increases in the long-term $$ spread, on the other hand, are linked to an increased risk of recession. One view is that if investors believe there is a larger danger of recession, they would place a higher value on short-term assets that can be rapidly liquidated to fund consumer spending. As a result, they will want higher compensation, i.e., more money, in order to continue holding long-term securities. It’s worth noting that the direction of the $$ spread defies common wisdom: a decrease in the yield-curve slope as a result of a lower $$ spread indicates lower, not higher, recession probabilities.
Figure 3 shows the fitted probability that the US economy would undergo a recession in the coming year for the sample period 1985:Q1 to 2018:Q1. The red line depicts the recession probability calculated by Benzoni and Chyruk (2018), while the blue line depicts the fit of a probit model that adds the ten-to-two-year yield-curve spread as a single predictor of downturns (the blue line) “model (also known as a “long-spread” model). Before a recession, the projected likelihood increases in both models. The red line, on the other hand, has more prominent peaks before each of the three recessions in the sample. In addition, the red line, when compared to the blue line, downplays the likelihood of recession in the 1990s and recent years. This suggests that the model proposed by Benzoni and Chyruk (2018) performs better than the more traditional probit specification with a lone yield-curve slope predictor; in particular, a statistical test of equal predictive accuracy between the long-only model and the BenzoniChyruk model is rejected with a 1 percent $$.12 error. In either event, the two models estimate the probability of a recession to be roughly 17 percent as of the first quarter of 2018.
Conclusion
What makes an inverted yield curve such a good forecast of future recessions? The yield-curve slope is a single indicator that aggregates several diverse elements that impact the state and evolution of the economy. We present our findings in Benzoni and Chyruk (2018), which show that decomposing the yield-curve slope into its expectations and risk-premia components helps separate the mechanisms that connect Treasury rate swings with the economy’s future state.
We find that monetary policy loosening, as measured by a lower current real-interest rate or a narrowing of the predicted real-rate differential, is related with a higher likelihood of a recession in the coming year. In contrast, depending on the source of the downturn, a decrease in the slope of risk premia is related with either a larger or reduced recession probability. In recent years, a drop in the inflation risk-premium slope has been associated with a higher risk of recession, but a lower real-rate risk-premium slope has been associated with a reduced risk of recession. This indicates that not all yield-curve slope decreases are bad news for the economy, and not all steepenings are good news either. 13
Of fact, the evidence presented in this article does not prove that a yield-curve inversion causes a recession. Rather, it’s possible that the slope changes to reflect shifting economic expectations, and that these expectations are useful predictors of economic downturns.
In addition, our findings are dependent on a variety of modeling assumptions. To begin, we use a specialized DTSM to interpret the term-structure components’ expectations and risk premia. Second, probit models of recession probability can be specified in a variety of ways. Our findings are resilient to a number of alternative modeling options that we investigated, but not all of them. The findings presented here are based on a specific set of assumptions and are part of a larger project aimed at better understanding the relationship between the yield curve and the macroeconomy.