Bond yields have declined to unprecedentedly low levels over the last three decades, resulting in stellar returns but also creating a more challenging outlook for the future. As a result, clients regularly ask whether our Robeco Lux-o-rente strategy will also be successful in a rising yield environment. An extended backtest that encompasses several decades of rising yields, notably in the 1970s, suggests this should be the case.
Robeco Lux-o-rente is a global bond fund that invests in government bonds tracking the JP Morgan GBI IG Index and which focuses solely on active duration management. It offers protection against rising yields by reducing its interest-rate sensitivity when yields are expected to rise. Conversely, when yields move lower, the fund can benefit from the increased interest-rate sensitivity by lengthening its duration.
The active duration positions are taken using US, German and Japanese bond futures. These positions directly follow signals given by Robeco’s proprietary quantitative duration model, which uses a set of variables ranging from macroeconomic to technical indicators.
This model has been proving its market-timing ability in practice since the 1990s, consistently and significantly outperforming the market over time. However, bond yields have generally declined since the model was developed and implemented.
Periods of rising yields have been shorter and occurred less frequently than periods of falling yields. To assess our duration model’s ability to weather prolonged adverse markets, we performed an extended backtest over a period that encompassed several decades of rising yields.
One period that is particularly important is the 1970s when the successive oil shocks of 1973 and 1979 unsettled bond markets in most developed countries. In the US, for example, inflation rose sharply to 12% in 1974 and 14% in 1980. In response, Fed Chairman Paul Volcker raised official interest rates to as high as 20% and long-dated bond yields rocketed up to 16% in 1981.
Sticking to this policy in spite of its negative economic consequences gradually restored confidence in the central bank and its ability and willingness to control inflation. This enabled interest rates to decline strongly again from 1981 onwards.
For our backtest, we used the historical data from the Ibbotson dataset for the US bond market to create a reasonable proxy for the model that goes back to 1951. We found the model would have performed well compared to the market in the period after 1981, when yields generally fell, but also in the decades before 1981, when yields generally rose. We also found that the model’s outperformance would have been especially strong during the months when yields moved significantly, regardless of the direction of these moves.