Betting Against Duration

Calling the end of the duration trade

The duration of an asset is the sensitivity of its returns to interest rate shocks. Assets very sensitive to interest rate shocks are high duration; assets not very sensitive to the same are low duration.

The 2 year yield captures the expected path of the policy rate — when it goes up, we can infer that the market is pricing in additional rate hikes by the Fed. Changes in the 2 year yield are therefore a good proxy of interest rate or monetary policy shocks.

We proceed as follows. We obtain the prices of all S&P 500 stocks from Yahoo Finance. We compute duration betas by regressing weekly returns for each stock on weekly changes in the 2 year yield, controlling for the market return. We then construct a portfolio of low duration stocks by equiweighting the bottom quintile of stocks by duration. Similarly, we construct a portfolio of high duration stocks by equiweighting the top quintile of stocks by duration. So, we have about 100 stocks in each portfolio.

High duration stocks have been on a tear since the last quarter of 2020. Even though large cap stocks have done extraordinarily well since the covid panic of March 2020, high duration stocks have done even better. While the low duration portfolio has returned 29 percent per annum since March 2020, the high duration portfolio has returned 50 percent per annum.

But this recent outperformance is highly unusual and likely due to the Fed’s extraordinarily accommodative monetary policy. Let’s examine what happened to these portfolios during Yellen’s hiking cycle. The Yellen Fed began hiking in earnest in 2017. By the end of 2018, the brief hiking cycle was over. The two year yield crashed and the Fed soon began cutting rates. It had cut about 1 percentage point when the covid panic began.

During this cycle, the high duration portfolio gained an early lead but then moved sideways until Oct 2018, when it gave up its meager gains. During Yellen’s hiking cycle, between Oct 2017 and Feb 2020, the high duration portfolio returned just 4.5 percent per annum, even as the low duration portfolio gained 16.3 percent per annum.

Even in the previous period of monetary ease, when we should expect it to have done relatively well, the high duration portfolio didn’t do all that much better than the low duration portfolio. Although, to be sure, it did keep up.

More generally, over the full-sample, 2010-2022, the low duration portfolio has handily beaten the high duration portfolio. The premium is especially pronounced in risk-adjusted term: a full-sample Sharpe of 1.37 vs 0.77.

This result is puzzling. In theory, portfolio exposure to duration risk should be compensated — investors should be rewarded for bearing duration risk with an associated risk premium. We have shown that, like “betting against beta”, we should also be “betting against duration” — the risk premium on duration risk in US equities is negative. The reason is probably the same for both puzzles: leverage-constrained investors bid away the risk premium.

Finally, we show that two-way sorts with duration and market beta also demonstrate the wisdom of betting against beta and duration. The low beta-low duration portfolio has the highest return, the lowest volatility and max drawdown and the highest Sharpe ratio. The high beta-high duration portfolio has the lowest return and the worst Sharpe ratio. Moreover, duration is more important than beta: moving from high beta to low beta marginally increases risk-adjusted returns; moving from high duration to low duration assets has a much larger effect on Sharpe ratios.

So, despite their stellar performance since the covid panic, high duration stocks are likely to do poorly as additional hikes get priced in. In general, low duration portfolios are likely to strongly outperform high duration portfolios — especially if rates follow a much steeper path than current market pricing suggests. Time, then, to get rid of that duration risk.