Focus
April 01, 2022 | 13:14
The Curves of Tightening Past
Economics
Economics
Economics
Economics
Fixed Income Strategy
Economics
Economics
Economics
Economics
Economics
Economics
Fixed Income Strategy
Economics
Economics
The Curves of Tightening PastThe U.S. yield curve has inverted. During the past few weeks, benchmark 7-, 5- and 3-year Treasury yields have followed each other in moving above the 10-year node. This week, the spread between 2- and 10-year yields, the much-followed curve slope metric, crossed the negative line as well. Before the pandemic, the yield curve inverted in a persistent fashion before every recession since 1980 (Chart 1). With inversion at hand, should we be worried? The answer is both yes and no. |
|
‘Persistent’ means 2s10s averaging a negative value for at least one month. Such curve inversions have occurred before the first month of the past five pre-pandemic recessions. The signal ranged from 11 to 23 months ahead, for an average of around 17 months (Table 1). This means that if April were to average a negative value for 2s10s, history would suggest a recession starting around September 2023 and no later than March 2024. How robust is this recession signal? There has only been one false positive result, i.e., an inversion which is not followed by a recession within a year or two. In June 1998, there was the briefest and near-smallest possible inversion, just one month averaging -2 bps followed by zero in July (during these two months, the weekly average was a slight negative for six weeks). During the end of August 2019, the curve inverted for several days, enough to generate a negative weekly average but not enough to flip the monthly. If we employed a less stringent weekly standard as the recession signal, this would have been another false positive result. Although March 2020 marked the first month of a recession, it was solely due to the pandemic. While some economic indicators were weakening during the latter half of 2019, by the start of 2020, business conditions were strengthening significantly. The factory sector was expanding again after slipping into a recession (mostly owing to the trade war with China), and payrolls posted their strongest back-to-back gain in nearly a decade. So, if not for the pandemic, the economy would have likely kept on expanding, assisted by the Fed’s 75 bps of cumulative easing during the latter part of 2019. The 2s10s spread has a 70%-to-85% success rate in predicting recessions since 1980. Two-year notes started being issued only in 1976. For longer and/or corroborating studies on the yield curve’s effectiveness in signalling recessions or for simply tracking the slope, the 3-month T-bill rate (on a bond-equivalent basis) or the fed funds rate is typically paired with the 10-year note yield. For example, the NY Fed’s recession probability model includes the 3-month rate whereas the Conference Board’s Leading Index includes the fed funds rate. Keep in mind, whether it’s the 2-year, 3-month or overnight tenor that’s teamed with 10 years, the greater the degree of inversion, the higher the risk of recession. |
|
Using the spread between 3-month bill rates and 10-year note yields, an inverted curve has occurred before the first month of all seven pre-pandemic recessions since the 1969-70 downturn (Chart 1). The signal ranged from 6 to 17 months ahead, for an average of 12 months (Table 1). The spread currently sits around 180 bps which points to a potential negative value only after many more Fed rate hikes (which 2-year yields are pricing in). Again, how robust is this recession signal? It has had false negative results, i.e., no inversions despite eventual downturns, before the 1957-58 and 1960-61 recessions. This signal also has had false positive results. It inverted in 1966 with no follow-up recession. Unlike its 2-year cousin, it didn’t invert during 1998. It did for a few scattered days in September of that year, but not enough to turn even a weekly average negative. However, it did invert during 2019. As such, the 3-month/10-year spread has around a 65% success rate in predicting recessions since 1957 and around 85% since 1980 (in line with 2s10s). Note that apart from its recession signalling power, an inverted yield curve itself could contribute to slower economic growth by potentially affecting the credit creation process. Whereas a positively sloped curve provides an incentive for banks to borrow short and lend long, this incentive evaporates as the curve inverts. |
So, we should worry about yield curve inversions. But, this worry should be tempered by the fact that the yield curve has become inherently much flatter during the past 10 to 15 years. This makes it more prone to giving false positive recession signals, with 2019 being a prime example. Three key factors have caused an intrinsically flatter curve. First, the neutral policy rate has fallen sharply, mostly due to demographics (we’re getting older and retiring) and productivity (it’s gotten slower). When the FOMC first began publishing its longer-run projections of the neutral rate a decade ago, the median was 4.25%, which was in line with historic norms (Chart 2). The latest projection was 2.375%, down 187.5 bps since then. To the extent the neutral policy rate guides market expectations of the profile for short-term interest rates over the coming decade (particularly beyond the immediate years), this major component of 10-year yields will be commensurately lower. |
|
Second, the Fed’s adoption of a formal 2% inflation target in January 2012 served to better anchor the market’s longer-run inflation expectations and, importantly, reduce the perceived risks surrounding them. While the inflation risk premium embedded in longer-term yields may have lifted a bit in the wake of inflation’s recent surge to 40-year highs, let alone hoisting the front end of the expected inflation profile, we suspect that beyond the next couple years, ‘all-in’ inflation expectations (levels + risk premiums) remain noticeably below where they tracked more than a decade ago. Indeed, amid the current multi-decade-high inflation readings and angst, the TIPS’ 5-year forward measure of 5-year expected inflation has remained relatively stable (Chart 3). Note that the horizontal line is set at 2.3% (CPI inflation) which would be consistent with 2.0% PCE inflation given the pattern of the past decade. |
The Fed’s modification of its policy framework in August 2020 to target 2% inflation over time allowed it to aim for moderately above 2% for a while to offset past inflation underperformance. Part of the rationale for the modification was that this chronic underperformance was becoming embedded in longer-term inflation expectations. The new framework appears to have helped bring expectations back in line. Third, monetary policy has been purposely flattening the yield curve. In the wake of the Global Financial Crisis and Great Recession, the Fed conducted three rounds of quantitative easing (QE) as well as a Maturity Extension Program (‘Operation Twist’). By the end of 2014, it had amassed more than $2.3 trillion of Treasury notes and bonds with a skew to longer-term maturities, up from a starting point just over $400 billion (Chart 4). While maturities started being rolled off in October 2017, it only got down to $1.9 trillion before runoffs were stopped in July 2019 owing to an emerging scarcity of reserves. Then, in the wake of the pandemic, holdings surged to $4.9 trillion, with net purchases finally ending only in early March. The Fed intends to roll off maturities at a much faster pace this time, perhaps announced as early as next month. However, as a share of the market, Fed holdings of notes and bonds are going to remain elevated for a while longer, compared to where they were both before the pandemic and 2008. By owning an elevated share of the market, the Fed is still exerting downward pressure on yields. |
|
Bottom Line: The yield curve (2s10s) has inverted again. To the extent the inversion persists and cascades to the 3-month/10-year slope metric, a recession signal will be sounding for about late next year. While this is worrying, the chance of this being a ‘false alarm’ is relatively high given that the yield curve is now inherently much flatter than before. |