- We evaluate the risk from extending the duration of a hypothetical multi-asset-class portfolio as a hedge against a possible recession.
- Surprisingly, extending the bond duration of a 60%-equity/40%-bond portfolio in the current environment does not significantly impact portfolio risk, according to the MSCI Multi-Asset Class (MAC) Factor Model.
- Our analysis also finds that the longer end of the Treasury curve has been a powerful friend to investors during past recessionary periods when coupled with aggressive Fed easing.
As more market participants worry about recession risk,1 asset allocators may consider changing the duration of their bond portfolios. By extending duration,2 institutions may believe they are better positioned to take advantage of a potential decline in rates triggered by a slowing economy and aggressive Federal Reserve (Fed) easing.3 But couldn’t duration extension push portfolio risk above a hypothetical institution’s target thresholds? Maybe not, our MAC model suggests.4
A brief history of downturns and Treasury returns
In the exhibit below, we highlight three periods of substantial monetary easing by the Fed during the past 30 years. Each of these periods started off with a slowdown in economic growth and a rise in the unemployment rate.
Three periods of economic slowdown and aggressive monetary easing
The 2007-2013 easing cycle also included the Fed’s quantitative-easing program. On May 22, 2013, then-Fed Chairman Ben Bernanke suggested that the Fed might soon begin tapering its purchases of bonds. We use this date as the end of the latest easing cycle. (Harding, R., Politi, J., and MacKenzie, M. “Ben Bernanke says bond buying could slow.” Financial Times, May 22, 2013.)
As shown in the exhibit below, in each of these periods, returns on longer-maturity Treasurys (10+ years) outperformed the intermediate part of the Treasury curve (as represented by the seven-year Treasury). History does not necessarily predict the future, but the longer end of the Treasury curve was a powerful friend to investors during these past slow-growth periods.
Annualized return outperformance relative to the seven-year Treasury
We computed Treasury returns using yields from the MSCI USD Govt OTR II curve for dates after 2006 and constant-maturity Treasurys from the Federal Reserve Bank of St. Louis’s FRED database for any previous date. N/A means the yields are not available for the full period.
Historical return may not be the only criterion to consider when contemplating duration extension. It may be important to look at the impact on the risk budget. For bond-only portfolios, extending duration will generally result in higher forecasted risk.5 But what about multi-asset-class portfolios?
Extending duration may not always increase risk …
Our hypothetical institution has a multi-asset-class portfolio with a 60% weight in equities (as represented by the MSCI ACWI Index) and 40% weight in bonds (as represented by a U.S. Treasury ETF with duration of approximately eight years).6 This institution also has an assumed total risk budget of 8%. Would extending duration breach the risk budget?7 What would be the impact of changes in correlations and volatilities?
Surprisingly, in the current environment, extending the portfolio’s Treasury duration by five years does not materially impact the risk of the 60/40 portfolio, according to the MSCI MAC model. The drivers of this result are the current negative correlation between bond and equity returns and relatively low bond-market volatility.
… but things could change with increases in correlations and volatilities
Our analysis shows that risk in a hypothetical 60/40 portfolio would rise significantly from duration extension in the event of a major regime change in which the bond-equity correlation turns positive and bond-market volatility experiences large relative increases. In our example, portfolio risk (as measured by volatility of returns) would rise to 11% from 7%.8
We invite you to use the interactive exhibit below to test different hypothetical correlation and volatility assumptions and measure the impact of duration extension on the risk of a multi-asset-class portfolio.
Extending duration: An interactive portfolio-level look
What could cause correlation and volatility to change at the same time?
One possibility could be a dramatic departure by the Fed from its current inflation-targeting regime, with the result that inflation uncertainty increases substantially. This scenario characterized the late 1970s and early 1980s, when bond-market volatility was relatively high and the bond-equity return correlation was positive. Nevertheless, in the current market environment with benign inflation, the new MSCI macroeconomic model suggests that the bond-equity correlation will remain negative in more than three-quarters of scenarios for the full decade ahead.
1See, for example: “Duke University/CFO Global Business Outlook.” Duke University’s Fuqua School of Business and CFO, Sept. 18, 2019.
2Duration refers to the sensitivity of bond prices to a movement in interest rates.
3Our analysis assumes that leverage cannot be used to position the portfolio for a potential decline in rates.
4We use the MSCI Multi-Asset Class (MAC) Factor Model Tier 1 USA factors in the analysis. They reflect the high-level drivers of risk and returns across asset classes and markets.
5For example, the iShares 7-10 Year Treasury Bond ETF has a duration of approximately 7.5 years and our risk-factor model projects 5% annual return volatility. The iShares 20+ Year Treasury Bond ETF has a duration of approximately 18 years and a projected annual volatility of 11.8%.
6We use the iShares 7-10 Year Treasury Bond ETF to represent the fixed-income portfolio allocation.
7To achieve the desired increase in duration, the fixed-income portfolio allocation is changed to a 20% weight in the iShares 7-10 Year Treasury Bond ETF and a 20% weight in the iShares 20+ Year Treasury Bond ETF.
8A similar scenario of positive return correlation between Treasurys and equities, along with higher Treasury volatility, was observed at the start of the 1989-1992 period.