Do you know how your fixed income portfolio is likely to perform if interest rates rise? In an environment in which rising rates are widely anticipated, investors are increasingly looking to the duration of their portfolio to understand the risks associated with it. However, providing a single duration measure is no longer an adequate response.
At its simplest level, duration is a measure of the sensitivity of the price of a fixed income investment to a change in interest rates, expressed as a number of years. Rising interest rates cause bond prices to fall, while declining interest rates lead to rising bond prices.
When fixed income managers report the duration of a portfolio, they typically reference effective duration*. While this measure of duration is generally accurate for government or corporate bonds, challenges arise when a portfolio has exposure to a variety of spread sectors and different types of risk premiums. Typically, rates and spreads have an inverse relationship (as rates rise, indicating an improving economy, spreads generally tighten), though as we will discuss later this relationship does not always hold true and can change over time.
Measuring the duration of unconstrained strategies
This challenge of accurately measuring a portfolio’s sensitivity to changes in yields is increasingly important with the proliferation of flexible or benchmark unconstrained strategies. In these types of portfolios, managers tend to have an absolute return or total return objective and do not invest in line with a traditional fixed income benchmark such as the Barclays Aggregate Index. Because they are not constrained to the sectors in this index, managers often look to generate return by indentifying opportunities across all of the different sectors within the fixed income universe.
Given all of the potential investments that could exist within these types of unconstrained fixed income portfolios—including spread sectors, high yield securities with embedded call dates, and mortgage structures that are sensitive to prepayment speeds—what is the best way to determine the portfolio’s true “duration,” or actual sensitivity to interest rate changes? In other words, how will the value of the portfolio actually change over time as interest rates rise and fall? In order to address this question, we use another measure of duration called empirical duration.
Empirical duration is the portfolio’s observed sensitivity, based on historical data, to changing interest rates. Despite the traditionally negative correlation between rates and spreads, Exhibit 1 depicts that the relationship is actually very dynamic. In fact, spread sectors such as emerging markets debt can even trade with a positive correlation to Treasury rates.
Exhibit 1: Fixed income sector correlations to 10-year U.S. Treasury rates
When calculating empirical duration, a purist might try to choose the interest rate most closely associated with an individual security and map the changes in that rate with the changes in the value of the security. While this approach is theoretically more accurate, the complexity of performing this exercise across an entire portfolio of several hundred securities is not efficient and has little incremental benefit. Instead, a simpler alternative is to map individual securities to related indices and then to determine the beta of the yield of each index to the yield of an on-the-run 10-year Treasury as a proxy for interest rates. Once the beta is determined, the empirical duration of the portfolio can be calculated as follows:
Empirical duration ≈ weighted effective duration x yield beta
However, because of the shifting relationship between rates and spreads (as depicted in Exhibit 1) as well as changing market volatility, it is important to monitor a portfolio’s empirical duration over different time periods and frequencies (such as 3 months, 6 months, 1 year and 3 years and daily or weekly data). For instance, if a significant market dislocation occurred a year ago, the 3-month or 6-month empirical duration number may be a more accurate representation of the portfolio’s current sensitivity to changes in interest rates than the 1-year number, which would include data from this unusual market environment. While the portfolio’s effective duration will remain consistent over all of the different data sets, its empirical duration will fluctuate depending on the time frame used to evaluate the relationship.
Not surprisingly, due to the changing relationship between rates and spreads, portfolios will more closely track different duration calculations at different times. By matching the changes in a portfolio’s valuation as rates change to the different types of duration measures, a manager can determine which duration is truly the best measure of the portfolio’s sensitivity to changing interest rates. As Exhibit 2 depicts, for an unconstrained fixed income strategy, the implied value of a portfolio based on the empirical duration measure more closely tracks the actual value of the portfolio than that implied by the effective duration. While this is not the ultimate solution to understanding how a fixed income portfolio will perform during different rate environments, it is another tool that could be used to more holistically evaluate and monitor the risk in unconstrained fixed income strategies.
Exhibit 2: Correlation of unconstrained portfolio’s value versus implied value based on effective and empirical duration measures
* Effective duration takes into account the way in which changes in yield will affect the expected cash flows. It takes into account both the discounting that occurs at different yield levels as well as changes in cash flows. This is a more appropriate measure for any bond with an option embedded in it.