Bonds have been a mess since the election as the market correctly fears huge budget deficits (via tax cuts) and tariffs. The equity market, on the other hand, loves the idea of cutting the corporate tax rate. As stocks have risen relentlessly, bonds have fallen in their standard negatively correlated manner. We have written about how bonds (with a lag) affect stock prices. The only unusual thing about this environment is the degree to which bonds have fallen.
In this study, we shall look to see if the large scale of the recent interest rate spike is particularly onerous for the future of equities. Let’s first establish the base case. We know that the S&P is a mean reverting time series. If you wait for a pullback, you are rewarded far more than if you chase a breakout:
If we use a ten-day RSI and buy when it is below 50, we get a vastly better return outcome than if we buy when the RSI is above 50. The latter strategy lost money over the 12 year horizon while the market gained a total of 900 points.
The next chart shows you the RSI of the S&P and the bond market at the same time:
You can see how the two are highly correlated. That’s because of their high daily negative correlation. Let’s add them together:
We can see here that the sum hit a level a week ago that was extremely high. Let’s compress the chart so we can see more history:
Last week’s reading was the highest since May 2007 (a spectacular time to get short). If we use this indicator as a filter we will see whether it succeeds in giving us better/good entries:
The filter (<100) captures 1200 points out of a 900 point total move. The alternative lost money. We know that bond moves affect stock market performance with a lag so we will now include a lag and rerun the filter:
This captured 1400 points. If we made our filter more selective, we would get fewer trades but we would also see few losses. Yes, this is a very unusual set of conditions for the stock market. It may well be very happy with tax cuts and tariffs but the bond market has historically acted like a bucket of cold water on all this happiness.
These conditions have been put into a trading system to test it more severely. The optimal holding period for a 10-day RSI look-back is 7-8 days. Pyramiding and trailing stops would be needed as would a test for the non-reversion problem.