Lots has been written about the yield curve over the last couple of weeks, after the two-year US Treasury yield rose fleetingly above the 10-year yield. As it was only an intraday move so you won’t see it in a chart using close of business data. 
As such, we’ve answered some of the most frequently asked questions by our clients.
Does it matter which part of the curve we look at?
Historically, pretty much all parts of the curve have co-moved, give or take a few months. The 3y-5y curve inverted last December, the 3m-10y curve inverted in May; the 1y-10y moved in and out of inversion in March, May and August. So it should come as no surprise whatsoever that the 2y-10y - probably the most commonly cited version of the “curve” - has started to flirt with inversion too.
Source: Datastream, NBER, Rathbones
What’s the yield curve’s track record as a recession indicator?
As you can see from the first chart above, an inverted yield curve - when the yield on longer-dated bonds falls below that of shorter-dated bonds - has been a great indicator of recession. It has predicted the last eight recessions, with no missed signals and only two false signals - one big false signal in the mid-1960s and one fleeting false signal in 1998. But it tells us relatively little about timing, as the table below details. Consultancy Capital Economics recently described the yield curve as “blind in one eye”. That’s a great analogy: it can tell you what’s in front of it, but not how close it is. The window between inversion and recession tends to be long, 14-15 months on average, and has been getting longer with time. This makes it a difficult tool for investors to use.
Why is the yield curve a good indicator of recession? Is there causation or just association?
The balance of expert opinion suggests that the yield curve’s relationship with the economy is not causative. Banks could have a role in a causative association, and there is substantial proof that net interest income falls with the yield curve. Furthermore, a survey by the St Louis Fed found banks saying that they would tighten lending standards when the curve inverted. But net interest income isn’t the same as operating profits. Banks hedge and banks trade. So studies that look at operating profit have actually tended to observe higher profits with a lower curve.
Most people do not think the relationship is causative. It is just indicative. There are four reasons why it might indicate a forthcoming recession:
1) An inverted curve could mean that investors think the central bank is making a policy mistake, raising rates too far, choking off growth and inflation, and will have to start cutting in the future.
2) It could also mean that inflation is too high and that will entail a short period of higher rates, but that will ultimately choke off growth and lead to lower rates and lower inflation in the future.
3) It could have nothing to do with monetary policy. It could simply be a reflection of sentiment: The perceived safest asset in the world is the US 10-year Treasury bond. When people believe there is increased risk of a crisis or a recession, they shift their assets into this bond. This is sometimes known as the ‘flight to quality’. Buying pressure on the 10-year drives prices up, thereby lowering yields. This is my favoured theory.
4) The yield curve could be a proxy for the return on capital relative to the cost of capital.
Is the yield curve still a useful indicator?
There are two reasons why the yield curve may not be quite so useful as it once was, or why it might be prone to more errors.
1) It could be distorted by quantitative easing (QE). To understand how, we need to remember that bond yields can be broken down into three components: the expected path of real base rates, expected inflation, and the term premium (which reflects compensation required for uncertainty, and also changes in the balance of supply and demand for bonds). QE predominantly affects the term premium (it introduces the central bank as a new buyer, changing the balance of supply and demand and pushing down the term premium). If QE - both in the US and globally (there are international spillovers) - mainly affects the term premiums of two-to-10 year bonds, a yield curve that makes use of these yields could be compromised. There are a couple of ways to correct for this, however.
It is interesting to note that the biggest false signal in the yield curve’s history occurred in the 1960s when the term premium, like today, was unusually negative.
We could look at a shorter-term measure of the curve unaffected by term premia - the San Francisco Fed propose the 3m-18m forward rate curve (see left-hand chart below). Or we could subtract the term premium from the yields we use. Unfortunately, each approach currently sends a different signal (right-hand chart below). The 3m-18m spread is sending an even stronger recessionary signal than more conventional measures; the term premium-adjusted approach suggests no risk of recession (note the threshold signal is no longer 0 but -0.5). A further problem is that if you favour third theory above, that the yield curve works because it indicates demand for 'safe haven' assets, then you need a decent long-term measure of the term premium in your chosen version of the yield curve.
Source: Datastream, NBER, FRBNY, Rathbones
2) It could also be prone to more false signals if the yield curve is permanently flatter because longer-term rates are lower, due to the profound decrease in productivity growth and/or evidence of an increased proclivity for saving cash relative to investing in productive projects. A flatter curve is arguably more prone to (brief) false signals.
Does the fact that the Fed is now cutting rates invalidate the signal?
Sadly not. The Fed has cut rates before the curve inverted in some (not all) previous episodes. But what is unusual today – and definitely worth noting well – is that the Fed is starting to cut when there is no clear macroeconomic indication that it has raised rates too far – above the “neutral” rate of interest – in the first place.
Could the yield curve be a self-fulfilling prophecy?
Many people are asking if investors are now more aware of the foreboding signal sent by the yield curve than ever before. And that the curve could itself cause a deterioration in confidence and a retrenchment of discretionary spending. This is possible. It’s my 13th year in asset allocation, and I learned of the predictive powers of the curve from my bosses pretty early on. But that’s just one man’s experience. Conversely, the top-ranked Wall Street strategist, Francois Trahan, wrote a report just two years ago confessing that he had only just fully appreciated the yield curve’s track record.
Looking at the number of stories on Bloomberg containing the phrase “yield curve” does not suggest that the market is more aware of the curve today than it was when it last inverted in December 2005.
Why do we only look at the US yield curve?
Other countries’ yield curves are more prone to missed or false signals. Again, this is why I favour the third theory from the above list of why the yield curve works as a recession predictor. It tells us about demand for safe-haven assets, and the US Treasury market is the world’s safe haven.
How should investors use the yield curve?
Since the mid-1950s the yield curve has inverted on average 14 months before a recession (15 if you’re watching the 5y-3y). But equity markets only lead recessions by three-to-six months. Selling equities as soon as the curve inverts could cause you to miss out on rising equity markets for rather a long time. For risk-aware investors, missing out on the final few months of returns is often prudent. But missing out on the final few years may be less forgivable. Note that last time around it took two years for the recession to start.
The second column in the left-hand table below shows the performance of US equities over the 12 months following a yield curve inversion. They often still post decent returns, especially since the late 1980s. The right-hand table suggests that markets may get a little more jittery as they digest the news over the first three months after an inversion, however.
Source: Datastream, Rathbones
Looking at the other columns in the left-hand table we can see that more defensive parts of the market are more likely to outperform. The US, and to a lesser extent the UK, tend to outperform global equities. The earnings underlying the US and the UK have a low beta to global economic activity compared to other markets. Cyclical sectors frequently underperform.
So the bottom line is that history does not suggest that investors should underweight equities as soon as the yield curve inverts. But, investors are likely to be rewarded for positioning in the more defensive parts of the market.
Ed Smith is head of asset allocation research at Rathbones. The views expressed above are his own and should not be taken as investment advice.