The US is heading for recession. That’s the message some investors are taking from the fact that yields on 10-year US Treasury bonds fell below those on three-month bills last week – something that hasn’t happened since 2007.
On the two previous occasions when we saw this inversion of the yield curve, a recession followed: in 2001 and in 2007-09. For this reason, the yield curve has also predicted equity returns in the past. Since 1987, an upward-sloping curve (in the sense of 10-year yields being above three-month bill yields) has led to the S&P 500 rising by 32.6 per cent on average in the following three years. But an inverted curve has led to the index falling by an average of 22.8 per cent in the next three years*.
There’s a simple reason for this. To see it, ask why anybody would want to hold a 10-year bond when it yields less than you could get on cash. One possibility is that they expect interest rates on cash to fall so that they would rather lock in a 10-year return than have to reinvest cash at a paltry rate. Another is that they expect prices of 10-year bonds to fall so they get a capital gain. Whichever it is, the point is that an inverted yield curve means that investors expect falling yields. And why would they expect this? It’s because they expect the economy to do badly. History tells us that this expectation is often correct. There’s wisdom in crowds. An upward-sloping yield curve predicts economic growth; a downward-sloping one predicts recession. And it's a better predictor of recession than economic forecasters.
So, should investors worry? Only a little. Two facts should mitigate our concerns. One is that while an inverted yield curve does predict recession, it does not predict an imminent one. For example, the 10-year/three-month yield gap first turned negative in August 2006, but the recession did not begin until December 2007. And it turned negative in August 2000 but the recession didn’t start until seven months later. And if we look not at T-bill rates but interbank rates, the lag between the first inversion and start of recession is even longer.
Because of this, the yield curve is not a great predictor of shorter-term equity returns. The correlation between the 10-year/three-month bill yield gap and subsequent annual changes in the S&P 500 has been statistically insignificant since 1987. And while the shape of the curve (whether it is upward or downward sloping) has some predictive power, it explains only 3 per cent of the subsequent variation in annual returns on the S&P since 1987. There are stronger leading indicators of annual returns – not least of which is foreign buying of US equities – and these are sending a bullish message, at least for now. History therefore tells us that there is plenty of potential for a relief rally if investors believe that a recession is not imminent.
A second comfort is that people sometimes learn from the past. The Fed knows the yield curve is a great economic forecaster: one of the economists who has pointed this fact out, John Williams, is now vice-chair of the Federal Open Market Committee. It is therefore keen that the curve does not invert. That is why its latest statement hinted that it won’t raise the Fed funds rate for some time – a promise that should help hold down three-month rates.
It’s not just the Fed that learns from the past. In theory, so too should investors. By now everybody should know that the yield curve predicts recessions. The bad news that the curve has inverted should therefore be already discounted by the stock market. Such inversions should therefore no longer lead to share prices falling months later.
Whether this is really the case or not is, however, questionable. The fact that shares fell so much last Friday is consistent with this possibility. So too is the fact that the market fell a lot last December when 10-year yields fell below three-month interbank rates.
On the other hand, though, we’ve reason to suspect that a recession is not fully discounted. People are terrible at forecasting changes in their future tastes. Harvard University’s Matthew Rabin calls this projection bias: it is why people pay more for houses with swimming pools in the summer than they do in the winter. It’s possible therefore that even if people do see a recession coming they fail to anticipate the fact that this will reduce their appetite for risk. If so, a recession would cut share prices when it actually happens even if people do expect it.
As long as the US avoids recession, however, we have reason for optimism about equities. We must respect the message of the yield curve, but not fear it.
*This is based on end-month data.
