Many market participants are fearful that the narrowing gap between the yield on the 10-year and 2-year Treasury notes signals that the US economy may tip into recession as early as sometime in 2019. But a better guide to recession risk is the short-maturity yield curve and risk appetite in the corporate credit market. The message from short-term interest rate expectations and the credit market is that the risk of a recession over the next year is low. The US business and credit cycle is aged but is likely to live longer than many expect.
Inversion of the Treasury yield curve – yields on short-term Treasury notes above those on long-term Treasury securities –preceded virtually every recession since the 1950s. An inversion of the yield curve does not cause a recession, but rather reflects investors’ expectations of an economic downturn that will prompt the Fed to begin to cut interest rates at some point over the next several quarters.
The measure of the yield curve most commonly referenced by market participants is the difference between the yield on the 10-year and 2-year Treasury notes. The yield curve on this measure currently stands around 30bps (0.3%), the lowest level since the end of the last recession. With the Fed signalling further interest rate hikes, many commentators predict that the ‘2s-10s curve’ will soon invert signalling a pending recession that will bring the current expansion, the second longest in the post-WW2 era, to an end(1). But there are other, more relevant, measures of the slope of the yield curve that are not so close to inversion and imply that near-term recession risk remains low.
The New York Fed uses the difference between 10-year and 3-month Treasury rates to derive the probability of a recession twelve months ahead. Based on latest monthly data for July 2018, the probability implied by the ‘3mth10yr curve’ that the economy will be in recession in July 2019 is 13.6%, less than the unconditional recession probability of around 17% (in the post WW2 era, on average the US has been in recession one year in six).
A recent econometric analysis by Fed staff concluded that the ‘near-term forward spread’ – the difference between the 3-month Treasury current (or spot) rate and the market-implied rate six-quarters ahead is a much more accurate predictor of recession than the ‘2s-10s curve’(2). The near-term forward spread is much more sensitive to shifts in the near-term economic outlook as reflected in shifts in expectations for Fed monetary policy than the 2s-10s curve (see Fig. 3). Since late summer of 2017, the ‘near-term forward spread’ steepened in response to the boost to US economic growth from tax cuts and Fed forward guidance on rates even as the ‘2s-10s curve’ has continued to flatten (the 10-year Treasury yield also incorporates global investor demand for yield and ‘safe assets’ as well as the legacy of QE).
The forward-looking nature of the yield curve means that it is a better predictor of recession than current economic data (‘hard’ such industrial production or ‘soft’ such as business surveys). Credit spreads are also a forward-looking indicator that captures information about investor expectations for future economic activity. A Fed staff measure of investor sentiment or risk appetite in the corporate bond market is the ‘excess bond premium’ – the additional spread that is not directly attributable to firm-specific expected default risk(3). An increase in the excess bond premium may reflect investors’ raising their assessment of recession risk. A model of the probability of recession at some point over the next 12 months based on the excess bond premium is shown in Fig. 4 with a current probability of 12%. It is interesting to note that the credit market also gave false positives –predictions of recession that did not occur – in 2002 associated with the accounting scandals and bankruptcies of some large firms (notably Enron) and in 2016 on the back of a surge in commodity-related defaults and fears of a China ‘hard landing’.
The probability of a recession starting at some point over the next four quarters using a model developed by BlueBay that includes both the ‘near-term forward spread’ and the ‘excess bond premium’ is illustrated in Fig. 5. A positive feature of the model is that it provided a better signal of rising recession risk prior to the last recession that started at the beginning of 2008 than the other models (it has a better ‘fit’ with a pseudo R2 of 0.5). But compared to yield curve based models solely, it also sends more false warning signals reflecting episodes of risk aversion in the credit market. Nonetheless, current values of the near-term forward and excess bond spreads imply that recession risk is low and actually fell in the first half of 2018.
Using forward-looking financial variables – the yield curve and credit spreads – implicitly assumes that markets are ‘rational’ in the sense they appropriately capture all the available and relevant information. Yet history is replete with financial bubbles and subsequent crashes that reflect investor behaviour that proved to be far from rationale. Nonetheless, market expectations for short-term interest rates and current credit spreads suggest that the near-term (1-year) risk of recession is low despite the flattening of the ‘2s-10s curve’ that has garnered so much attention. The US expansion is one of the oldest, but also the least spectacular in terms of cumulative economic growth (even with the recent acceleration). Be wary of warnings that the yield curve is signalling a pending recession - It is too soon to ready portfolios for a US economic downturn.
1. However, market expectations of future Fed rate hikes are already reflected in the yield curve and thus a few more 25bps Fed rate hikes will not necessarily lead to inversion.
2. (Don’t Fear) The Yield Curve, FEDS Notes, US Federal Reserve, June 28, 2018
3. Recession Risk and the Excess Bond Premium, FEDS Notes, US Federal Reserve, April 8, 2016