• AIBC 2022 Research

The yield curve puzzle: Interpretations of recent fluctuations in US Treasury yields

By Michelle Mengxi Tian:


The US Treasury yield curve has long been scrutinised by market participants for its informational content. The curve is considered relevant in a twofold sense: first, it is valuable in forecasting business cycle and expectations of interest rates and inflation; and second, changes in the yield curve are useful in informing policymakers about the effectiveness of monetary policy transmission in the financial markets. Beyond these, the yield curve also plays an essential role in many economic decisions, functioning as a benchmark for lending and saving rates for banks for loans of different maturities. Following the success of Harvey Campbell’s 1986 dissertation documenting the empirical success of recession prediction models linked to the term structure of interest rates and the subsequent widespread interpretation of an inversion in the yield curve as a recessionary signal, it is unsurprising that a chorus of omniscient market commentary has recently emerged in response to inversions along parts of the yield curve.


Recent shifts: inversions (at some parts) and a negative risk premium


Under normal economic conditions, the yield curve tends to be upward sloping. Intuitively, this can be explained by the fact that investors tend to demand higher premiums [1] on longer-term bonds (commonly called a risk premium) to tie up their funds for more extended periods of time, in accordance with the Liquidity Preference Theory (O’Sullivan and Papavassiliou, 2020) and predicated upon the assumption that investors are risk averse [2]. Since April 2022, however, the US Treasury yield curve has deviated from this traditional upward slope, inverting at some points while maintaining a positive slope at other points, reflecting a mixed view of the markets towards recession and inflation.


Figure 1: US Treasury yield curve on various dates. Figure 2: 10Y2Y term spread and near-term forward spread from Oct-2021 to Apr-2022 [Source: Federal Reserve Bank of New York]

From Figure 1, we observe how the curve has evolved from April 2021 to August 2022. At the short end (for maturities less than two years), the yield spread has generally remained positive, with yields increasing in tandem with maturities. In contrast, in the latter half of the curve, medium-term rates have risen close to long-term rates, such that yields on 6M Treasuries are now above those for 10Y Treasuries. Looking at the famous 10Y2Y term spread, we see from Figure 2 that it has been on a downward trend since October 2021. Although not pictured here, the 10Y2Y spread has in fact been in negative territory since July 2022. Interestingly, we note two more trends against the backdrop of which this inversion has been occurring.


Firstly, the entire curve has shifted upward, likely due to revised expectations about inflation and monetary policy. For months after April 2021, the Federal Reserve (Fed)’s stance had been that the pandemic surge in inflation was largely ‘transitionary’ and would ebb in 2022. Even after two rate hikes earlier in the year, inflation has proved to be more powerful and persistent than expected – in June 2022, the US Consumer Price Index, a widely-tracked measure of average living costs, showed a 9.1% rise from a year earlier, close to the highest level in decades. From being a non-issue for decades, inflation became one of the dominant concerns in the market and the Fed recognised the need to drastically shift its monetary policy direction. In both June and July, the Fed lifted interest rates by 75bps [3], the third and fourth hikes in 2022 and the largest increases since 1994. Monthly bond purchases were also concluding earlier than expected, in March 2022. Overall, this represents the most aggressive Fed tightening cycle in more than 40 years [4].

Figure 3: YoY % Change in CPI [Source: NY Times].
Figure 4: Total Assets of the Federal Reserve [Source: Federal Reserve].

Second, the upward shift for bonds maturing in the medium-term (around 1-3 years) has increased noticeably more than those for the longest maturities. Just as the lockstep move of short-term rates with the US Federal funds rate can be intuitively conceived, the generally accepted view is that medium-term rates are primarily driven by expectations about the path of the policy rate and inflation in the future. With financial participants now expecting sharper rate hikes than they did in 2021 (evidenced by the Federal fund futures term structure in Figure 5), it is unsurprising that medium-term yields have been driven upwards further.


Figure 5: Federal Fund term futures as of June 2020 and June 2022 [Source: Dimensional] (See footnote [5] for an interpretation of this graph)

Formally, both shifts can be partially accounted for by the expectations theory, first developed by Thomas Sargent, which suggests that the return on holding a long-term bond to maturity is shaped by the expected future path of short-term interest rates [6]. Thus, in accordance with the expectations theory, a rise in the Fed funds rate raises the effective lower bound for short-term interest rates, thereby pushing up yields on the short end. Similarly, then, the rise in the level of middle-term yields reflects the market pricing in an extended period of restrictive monetary policy and market participants’ beliefs that the Fed will continue to raise the policy rate over the next few quarters.


At this point, it is reasonable to ask: against the backdrop of the above-mentioned trends, what are the underlying factors driving the rapid fall in spreads between mid- and long-term yields in recent months and the resultant inversion in the right half of the curve? We have, thus far, described three theories related to the structure and shape of the yield curve: the expectations theory, the liquidity preference theory, and the existence of the risk premium. We now shift our focus to whether any of these theories provide a reasonable explanation for the 2022 fall in the term spread, and how these could relate to a potential upcoming recession in the US.


Interpretations of the flattening and inverted yield curve


Since Harvey’s dissertation, a sizeable body of literature analysing the predictive power of the slope in forecasting recessions has emerged [7]. Historically, findings have pointed to the 10Y-2Y term spread as being a relatively reliable predictor of economic recessions – since 1955, an inversion has preceded each recession except one [8]. Similarly, a flattened yield curve, although less extreme, is also commonly associated with a sluggish economy and declining growth. But the statistical evidence of the inverted yield curve as a precedent to recessions has not equivalently shed light on the channels through which the yield curve acts as such a signal. Notably, two explanations seem to dominate existing market commentary that aims to explain why the yield curve flattens or inverts before a recession, with these generally falling into one of the following strands of thought:


Argument one: negative spreads reflect expectations of an upcoming downturn

One is that the flattening of mid to long-end of the slope reflects the market’s expectation of a likely cooling of the economy during that timeframe, and correspondingly, that there would be potential 1) rate cuts to spur economic activity and 2) lower inflation, features common during declining economic activity. Specifically, looking at the inversion of the 10Y2Y spread, one reasonable explanation is that inflation over the next two years is expected to be higher than during the next eight, naturally putting upward pressure on two-year nominal yields relative to ten-year nominal yields and compressing the spread. However, taking a step back, it is useful here to highlight that falling inflation expectation, in general, does not necessarily require a highlighter probability of recession, and neither do rate cuts or even a taper of rate rises beyond the horizon of around six quarters. To understand how much of the decline in the mid and long-term yield spread is driven by the changes in inflation expectations and short rates, we would require not only surveys on expectations of both, but also further analysis of the extent and manner through which these expectations are reflected by and affect treasury yields across the maturity spectrum.


Argument two: it reflects rising market stress and preference for safer and more liquid assets

Another common argument is that the fall in the term spread reflects investors’ preference for safer assets such as US treasuries during an economic downturn or period of market stress [9]. This is sometimes known as a “flight-to-quality” [10] and “flight-to-liquidity” [11] episode when treasuries become more attractive compared to equities and other investments for being able to provide a relatively safe return and liquidity in a risky market environment. While this argument has its merits, it appears to break down against the backdrop of today’s economic landscape. For instance, looking at Figure 6, yields for 20Y and 30Y bonds today are higher than that for 10Y bonds (and those for 20Y are higher than 30Ys) suggesting that there are other factors apart from a preference for longer-term safer assets that are putting downward pressure on 10Y yields and 30Y yields. These reasons could potentially be structural in nature. One such reason, for example, could be that most home mortgages in the US have a maturity of 30 years, so if a bank is trying to make money on the spread, they could reasonably try to match the duration risk of these mortgages.


Figure 6: US Treasury yield curve as of 14 August 2022

Where does this leave us? The overarching point here is that existing literature on the predictive power of the inverted yield curve seems to be largely empirical in nature – there is little consensus as to the precise mechanisms through which the yield curve functions as a recessionary signal. Furthermore, since the yield curve was first touted as a useful predictor of recessions in the 1980s, the fundamental explanations behind them that may have originally held true may no longer hold given the structural changes that have taken place in the underlying US Treasury market and global financial landscape in recent years.


Firstly, a collapse in the global supply of “safe assets” after the GFC has contributed to a shortage of instruments with qualities like long-term US Treasuries. Following the crisis, pseudo-safe assets such as the mortgage-backed debt of government-linked companies (like Freddie Mac and Fannie Mae), as well as debt by fiscally weak governments like Greece, were no longer perceived as safe. As such, the resultant rise in demand for substitutes like longer-term US Treasuries could have increased downward pressure on longer-term yields.


Secondly, there has been a secular decline in yields of long-term US Treasuries (Rudebusch and Bauer, 2013); from 1990 to 2012, 10Y US Treasury yields have fallen steadily throughout market cycles from just over 8% to around 2%. While identifying the source of the decline has proved challenging, two factors that seem to have played a role are a decline in long-term expectations of inflation and inflation-adjusted interest rates (See Figure 7). If we wanted to, we could further break down the causes of these two trends – potentially due to the trend of lower volatility and an accompanying downward shift in the level of US nominal GDP growth, adoption of forward guidance (introduced in 2013), and lower and less volatile inflation recently (See Figure 8). However, let us put aside our theories for the falling term premium for now, which would necessitate another discussion in itself. The point here is that the lower term premium itself, which some call the ‘term premium conundrum’, does not necessarily require an increased risk of an economic downturn, casting doubt on the reliability of inverted yield curves as a precedent for recessions.


Figure 7: Decline in long-term yield, inflation expectations and inflation uncertainty (1990 to 2000 and after 2010) [Source: Federal Reserve Bank of San Francisco]
Figure 8: Inflation rate (%) from 1960 to 2020. [Source: MacroTrends]

Peering beyond the yield curve: a recession on the horizon, anyway?


For one, what’s interesting is that although academic literature classifies theories on the term structure of interest rates into four main stands of thought: the preferred-habitat hypothesis, expectations hypothesis, liquidity preference theory and existence of the risk premium, to translate these theoretical models to a real-world interpretation of the yield curve that considers the idiosyncrasies in today’s financial markets is no mean feat. We need to understand the dimensions along and conditions under which each theory succeeds in describing the term structure, and the conditions under which each one fails. To apply this, we need in turn to have a clear assessment of current conditions and a time-tested understanding of the cause-and-effect linkages, which will then allow us to clearly assess what might come next, and how we can respond.


Furthermore, taking a step back and looking at today’s markets, we see that although the yield curve’s link with recessions may have become increasingly spurious, a mild recession may nevertheless be before us. A range of indicators including consumer spending, housing and manufacturing activity and purchasing manager index (PMI) all seem to indicate that an imminent and significant weakening in real growth and persistently high level of inflation is on the horizon. To benefit from today’s market environment, using the yield curve as a proxy for what is already discounted in the markets could serve as a useful benchmark against which we play our cards, testing our understanding of what is likely to unfold.


 

[1] In academic literature, this premium is known as the ‘risk premium’ – the compensation for uncertainty and risk around future central bank policies, inflation, and growth.

[2] That is, given the choice between two financial investments with the same average expected return, they will prefer the one with less anticipated variability in the return.

[3] Basis Point (bp): One basis point equals 0.01%

[4] In comparison, coming out of the Great Recession of 2007 to 2009, the Fed took two and a half years to raise rates at the same pace.

[5] The term structure for the expected Federal funds rate is constructed using prices of Federal funds future contracts for consecutive months. Looking at the term structure as of June 2020 and 2022, the flat term structure in June 2020 indicates low chances of a rate hike (potentially due to pandemic-related uncertainty), while the upward slope in 2022 indicates expectations of a higher Fed funds rate.

[6] In the original form, the theory suggests that long rates are forced into equality with the expected future path of short-term interest rates. For this reason, it has been rejected in many studies, but for interpretation purposes, the theory serves as a useful starting point for understanding the determinants of the term structure. For example, the expected return from investing in an n-period bond should equal the expected return from investing in a one-period bond over n successive periods. It also suggests that if policymakers wish to alter long-term rates through their influence on short-term rates, they must succeed in altering the market’s expectations of future interest rates (Robert G. King 2002).

[7] A more comprehensive list includes Fama, 1986; Harvey, 1988 and 1989; Estrella and Hardouvelis, 1991; Estrella and Mishkin, 1998; Rudebusch and Williams, 2009.

[8] It inverted before the early 1990s recession, before the Great Recession in the early 2000s, before the 2008 recession, and before the pandemic recession in 2020.

[9] Furthermore, what is considered ‘safe’ depends largely on the underlying market environment. During the COVID-19 crisis, for example, prices of US Treasuries became extremely volatile and yields of similar-maturity Treasuries started to deviate significantly from each other amid heavy investor demand to liquidate Treasury securities for large amounts of cash. In other words, US Treasuries did not serve their traditional safe-haven role during the COVID-19 fragility. For more details about the mutual fund selling crisis and subsequent dysfunction in the US Treasury market during the initial declaration of COVID-19 as a pandemic in 2020, see this.

[10] When investors in aggregate begin to shift their allocation away from riskier investments into safer ones

[11] When investors shift their allocation away from illiquid investments into more liquid ones

 
References
  1. D. Bauer, M., & Mertens, T. M. (2022, May 9). Current recession risk according to the yield curve. San Francisco Fed. Retrieved August 14, 2022, from https://www.frbsf.org/economic-research/publications/economic-letter/2022/may/current-recession-risk-according-to-yield-curve/

  2. Engstrom, E. C., & Sharpe, S. A. (2022, March 25). (Don't fear) the yield curve, reprise. The Fed - (Don't Fear) The Yield Curve, Reprise. Retrieved August 14, 2022, from https://www.federalreserve.gov/econres/notes/feds-notes/dont-fear-the-yield-curve-reprise-20220325.htm

  3. Tang, D. Q., Li, Y., & Tandon, A. (2019, March). The term premium conundrum. Retrieved August 14, 2022, from https://www.nb.com/documents/public/global/u0064_0319_wp_the_term_premium_conundrum.pdf

  4. Bauer, M. D., & Rudebusch, G. D. (2013, July 8). What caused the decline in long-term yields? San Francisco Fed. Retrieved August 16, 2022, from https://www.frbsf.org/economic-research/publications/economic-letter/2013/july/cause-decline-long-term-us-government-bond-yields/

  5. The Group of Thirty. (2021, July). U.S. Treasury Markets - Steps Toward Increased Resilience. Retrieved August 14, 2022, from https://group30.org/images/uploads/publications/G30_U.S_._Treasury_Markets-_Steps_Toward_Increased_Resilience__1.pdf

 
Michelle Mengxi Tian
Michelle is currently in her second year of undergraduate study at the London School of Economics.