The shape of the US Treasury yield curve generally contains useful information about future developments in the real economy. In particular, when the Treasury yield curve inverts – that is, when short-term rates exceed long-term yields – a recession usually follows in the next 12 months. Historically, the yield curve has been a very accurate forecasting tool: a curve inversion has preceded each of the last seven recessions in the US.
The Treasury yield curve flattened significantly in 2017, and last month, the spread between 10-year yields and two-year yields narrowed to about 50bps. It is therefore possible that the spread could turn negative at some point this year. As such, it is unsurprising that market observers have become nervous about the possibility of an economic slowdown. But given the current environment of loose monetary policy, to what extent should we trust the US yield curve as a harbinger of a recession?
Predictive powers: the relationship between the yield curve and economic activity
The slope of the Treasury yield curve is generally measured as the gap between 10-year yields and three-month T-bills, but due to liquidity issues it is now calculated as the difference between 10-year and two-year yields.
As aforementioned, the last seven US recessions have all been preceded by an inversion of the yield curve. In every instance, the slope of the Treasury yield curve turned negative between nine and 33 months before the official start of the recession (as defined by the National Bureau of Economic Research). After an inversion, the US yield curve quickly steepens and a recession ensues (see chart 1).
Chart 1: An inverted yield curve has correctly signalled an imminent recession in the past
Source: Reuters Datastream, Bureau of Economic Analysis, National Bureau of Economic Research (NBER) as at January 2018
However, the timeliness of the yield curve inversion as a forecasting tool has deteriorated over time. Recently, the lag between the curve inversion and the official start of a recession has lengthened. That probably mirrors the lengthening of the business cycle and the increased effectiveness of monetary policy, which has helped keep inflation at bay (see chart 2).
Chart 2: The lag between curve inversions and the start of recessions has widened, in line with the lengthening of the cycle
|Yield curve inversion start date||Yield curve inversion end date||Recession start date||Lag between curve inversion (start date) and recession (months)||Length of preceding expansion (months)|
|January 1969||August 1970||December 1969||11||106|
|February 1973||November 1975||November 1973||9||36|
|August 1978||May 1980||January 1980||17||58|
|September 1980||October 1987||July 1981||10||12|
|December 1988||April 1989||July 1990||19||92|
|June 1998||January 2001||March 2001||33||120|
|December 2005||June 2007||December 2007||24||73|
Source: Hermes, based on data from Reuters Datastream and NBER
Using the probit regression analysis, which is regularly updated by the New York Federal Reserve, it is possible to assess the likelihood of a recession1. This model uses the difference between 10-year and three-month Treasury rates to calculate the probability of a recession in the US in the next 12 months.
The New York Fed’s January update suggests that the likelihood of a recession occurring in the next 12 months is still quite low, running at 11-12% (see chart 3). That seems reasonable given that the term spread is still positive and the lead time is considerable (and it has lengthened over time). The model suggests that when the term spread touches zero, there is a 33% probability of experiencing a recession in the following 12 months.
Chart 3: The current yield spread implies a low recession probability in the next 12 months
Source: Federal Reserve Bank of New York, NBER as at January 2018
Drivers of yield-curve flatness and the outlook
In the last 12 months, yield curve flatness has been driven by both the front-end and back-end of the curve. Short-term rates have increased throughout 2017, up from 1.2% in early January to 1.9% by the end of December, reflecting the Fed’s ongoing hiking cycle. This upward move accelerated towards the end of the year, indicating the likely impact US tax reform would have on both the economy and monetary policy. More importantly, the primary driver of a flatter curve in the first half of the year was long-term yields. Despite three US interest rate hikes, 10-year yields declined in the first nine months of 2017. In the final quarter of 2017, long-term yields finally started to increase. Nevertheless, they continued to underperform short-term yields.
There are two possible explanations for the underperformance of long-term yields in 2017:
If the first explanation is valid, it is easy to envisage a scenario where the Treasury yield curve inverts at some point in the near future, perhaps in the next two years. As the Fed’s hiking cycle continues, markets will start to anticipate a slowdown (or recession) and the reversal of some interest rate hikes may be warranted. While financial markets can be wrong in predicting the evolution of the economy, developments in markets matter for the real economy as they determine financial conditions and, in turn, credit conditions. In other words, market expectations for the real economy can become self-fulfilling. In this context, expectations of an imminent recession in financial markets could cause a downward correction in asset prices and a sharp tightening of financial conditions, thereby weighing on the real economy.
That said, the scenario of a persistently low and flat yield curve would also pose risks in the medium term. Financial conditions that stay too easy for too long would be counterproductive as they can result in excessive risk taking, thereby increasing economic imbalances and laying the foundations for a crisis.
However, if the second explanation holds true, the informative value of the yield curve is poor because its shape mainly reflects distortions from monetary policies in financial markets, especially at the back-end, rather than expectations for the real economy. Additionally, other technical factors might be at play, such as expectations of increased supply of US Treasuries at the short end (in line with recent indications from the Treasury’s quarterly statement). From this standpoint, macroeconomic risks should be limited.
Furthermore, it would be reasonable to expect that the slope of the curve could stabilise over the next year as central banks worldwide become less accommodative, easing downward pressures on long-term yields by allowing for some normalisation in the term premium. Yet, the normalisation of monetary policy will likely be a very gradual process, and central banks’ balance sheets will remain large (the balance sheets of the Fed, ECB, BoJ and Bank of England (BoE) should stabilise at approximately $15.5tn towards the end of 2018), implying stock effects will continue to weigh on long-term yields in the near future.
As Benoit Coeure, a member of the ECB’s Executive Directive, suggested in a recent speech , spill-overs on long-term yields from monetary policies abroad can be significant, which explains the high correlation across long-term yields in the main developed economies. In recent years, the term premium in the US has been depressed by international capital inflows triggered by overseas QE programmes. In particular, the ECB’s asset purchase programme has resulted in substantial cross-border capital flows as bond investors have rebalanced their portfolios moving out of Europe and into US Treasuries. According to an ECB study, spill-over effects from the euro-area currently account for 30-40% of the variance in US Treasuries3.
Chart 4: The US term premium has been compressed in recent years, probably reflecting significant monetary policy stimulus domestically and abroad
Source: Federal Reserve Bank of New York, Federal Reserve, European Central Bank, Bank of Japan, Bank of England as at January 2018. Data for the term premium is provided by the New York Fed as at December 2017. It is updated on a monthly basis.
These two explanations do not necessarily rule each other out. While long-term yields are probably distorted by central banks’ unconventional policies, the yield curve also incorporates financial markets' expectations of future economic developments. The Treasury yield curve has been a reliable predictor of recessions in the past, even when observers found reasons to dismiss it. Therefore, it is a good idea to monitor its trajectory and take any signals into consideration.
The implications of Treasury curve flattening for the Fed
The flattening of the curve has been a subject of discussion for Fed officials recently. Policymakers are also facing the same interpretative challenges that we have outlined above.
In December, at her final Federal Open Market Committee (FOMC) press conference as Fed chair, Janet Yellen dismissed concerns about the yield curve – echoing her predecessors Bernanke and Greenspan, who discounted its efficacy before the 2007-2009 US recession began. However, the minutes from the December Federal Open Market Committee meeting showed a more nuanced discussion. They noted:
“Some [FOMC participants] expressed concern that a possible future inversion of the yield curve, with short-term yields rising above those on longer-term Treasury securities, could portend an economic slowdown, noting that inversions have preceded recessions over the past several decades, or that a protracted yield curve inversion could adversely affect the financial condition of banks and other financial institutions and pose risks to financial stability4.
In other words, the Fed is not currently worried by the flatness of the yield curve. However, a yield curve inversion would likely change its view, but that depends on general conditions in the economy and financial markets.
A yield curve inversion could imply that the market knows something about future economic developments which the Fed is disregarding. However, other market indicators do not imply particular concerns about the recent flattening of the term structure yet – a fact Fed officials drew comfort from, as referenced in the minutes of the December FOMC meeting.
In general, a curve inversion is unlikely to stand out as an issue for the Fed in isolation, but it would become one in conjunction with persistently low inflation. Should a yield curve inversion occur when inflation is still undershooting the official 2% target, the Fed would have a strong case to slow down or even pause its hiking cycle. Moreover, the uncertainty surrounding the estimate of the neutral interest rate would also stoke caution during the hiking cycle, as the Fed is wary not to overshoot the neutral rate while inflation remains low.
Curve flattening elsewhere
Shifting our focus outside the US helps put the recent US Treasury yield curve flattening into perspective. A couple of observations stand out.
First, the government yield curve is steeper in the eurozone, the UK and Japan, compared to the US (see chart 5). That probably reflects the fact that these economies are at an earlier stage of the business cycle, which calls for looser monetary policies. Accordingly, differences across term structures have been mainly driven by short-term rates, while, as previously discussed, long-term yields have tended to be highly correlated across developed markets since the crisis (see chart 6).
Chart 5: The yield curve has not displayed the same predictive power outside the US
Source: Reuters Datastream as at January 2018
In addition, government yield curves in jurisdictions outside the US have not displayed the same predictive power for recessions historically. In Germany, the relationship between the government yield curve and the occurrence of recessions seems quite loose. Recent recessions have not been preceded by a curve inversion and the lead time can be considerable (in one instance, almost 3 years).
Similarly, in the UK, the relationship between the government yield curve and the occurrence of recessions is not as strong as in the US. The curve inverted in 1985 and in 1997 but recessions did not take place in the following 12 months.
Chart 6: Long-term yields have been highly correlated across developed markets, while short-term rates have displayed a low correlation, reflecting policy divergence
Source: Reuters Datastream as at January 2018
The most interesting case is Japan, where the curve has not inverted since the early 1990s, but recessions occurred in 1997-98 (Asian crisis), 2001-02, and 2008-09 (global financial crisis). The absence of curve inversions since the 1990s possibly reflects the BoJ’s policies which have almost constantly depressed short-term yields. Many major central banks followed Japan in unconventional policy territory in the last 10 years. For that reason, there might be a case to believe the government yield curve has lost some of its informative value as it has been increasingly distorted by monetary policy.
For now, trust the yield curve
In summary, there are a few reasons to doubt the validity of the US yield curve's predictive power in the current circumstances. In particular, distortions from extremely loose monetary policy could jeopardise the validity of the signals from yields. Conversely, a recession could occur even in the absence of a curve inversion, as evidenced from Japan’s recent economic history. Indeed, Japan’s experience suggests that permanently loose monetary policies might prevent curve inversions from occurring, depriving the yield spread of its signalling properties. This possible issue underscores the need to include other macroeconomic and financial indicators in an effective analysis of the business cycle.
Today, it seems too early to worry about an imminent US recession. Although the Treasury yield curve has flattened over the last year, the spread between long-term and short-term yields is still in positive territory. Furthermore, macroeconomic data has shown positive momentum at the turn of the year, suggesting the current positive mini-cycle has probably still some way to go.
Nonetheless, any foreboding by the yield curve cannot be ignored given its reliable track record. And when policymakers and commentators dismissed the significance of the yield curve as a forecasting tool in the past, it has proven its worth. Therefore, a curve inversion, should it occur, deserves market – and policymaker – attention.