Factor-based models are often criticized for data mining. One way to address that charge is with “out-of-sample” testing over longer time frames. But that takes time. New research provides an alternative out-of-sample test – using emerging-market bonds.
I will explain that test, but first let’s review the context of how those factor-driven models have evolved.
Academic research into asset pricing now covers more than 600 factors that could add explanatory power to the cross-section of returns. From that “zoo” of factors, how can investors determine which exhibits are worthy of investment? In our book, “Your Complete Guide to Factor-Based Investing,” Andrew Berkin and I established the following criteria. For a factor to be considered, it must meet all of the following tests. To start, it must provide incremental explanatory power to portfolio returns and have delivered a premium (higher returns). Additionally, the factor must be:
- Persistent – It holds across long periods of time and different economic regimes.
- Pervasive – It holds across countries, regions, sectors and even asset classes.
- Robust – It holds for various definitions (for example, there is a value premium, whether it is measured by price-to-book, earnings, cash flow or sales).
- Investable – It holds up not just on paper but also after considering actual implementation issues, such as trading costs.
- Intuitive – There are logical risk-based or behavioral-based explanations for its premium and why it should continue to exist.
Among the small number that meet all the criteria are:
- Value: The tendency for relatively cheap assets to outperform relatively expensive assets.
- Momentum (cross-sectional): The tendency for an asset’s recent performance to continue, leading to outperformance of recent winners relative to recent losers.
- Carry: The tendency for higher-yielding assets to outperform lower-yielding assets. Carry is the expected return if nothing changes with the passage of time.
- Defensive (or quality): The tendency of safer, lower-risk assets to deliver higher risk-adjusted returns relative to their low-quality, higher-risk counterparts.
Jordan Brooks, Scott Richardson and Zhikai Xu provide us with a new out-of-sample test of the pervasiveness of those four factors by examining their performance in emerging market (EM) bonds. EM bonds are those issued by entities (corporate and sovereign) domiciled in emerging markets and can be issued in local and “hard” (typically USD) currencies. J.P. Morgan’s global EM indices, which capture the investible section of EM markets, have increased from around $350 billion in 2002 to nearly $2.5 trillion by the end of 2018.
EM bonds have two sources of risk. They inherit exposure to sovereign rates from the “hard” currency they are issued in. Second, they contain exposure to a relatively unique source of credit risk, and hence return. As of December 31, 2018, the J.P. Morgan EMBI Global Diversified Index contained 679 bonds from 66 sovereign and 76 quasi-sovereign issuers with an aggregate value of $873 billion – this is a large asset class.
Their March 2020 study, “(Systematic) Investing in Emerging Market Debt,” (which will be published in the fall 2020 issue of The Journal of Fixed Income) covers the period 2004-2018 and measures the performance of four systematic investment themes over 25 emerging sovereign reference entities. The authors defined the four systematic factors in emerging market bonds as follows:
- Value: A cheap EM bond (or credit default swap, CDS, contract) is one where the credit spread is wide relative to fundamental credit risk (measured by credit ratings and volatility of country equity returns). These two measures help capture the essential ingredients of distance to default (leverage and volatility).
- Momentum: An equal-weighted combination of three return-based metrics: (1) six-month trailing EM CDS returns, (2) six-month trailing foreign exchange (FX) returns, and (3) six-month trailing country equity returns.
- Carry: The spread of the five-year CDS contract at the start of each month, selling protection on CDS contracts with higher spread levels.
- Defensive: A measure of sovereign quality as indicated by its ability to achieve low and stable levels of inflation, selling protection on CDS contracts where there is a lower level of expected inflation. The authors also used a measure designed to capture the overall indebtedness of the sovereign entity (i.e., asset/debt ratio), which combines foreign reserves and the level of GDP grossed up by expectations of GDP growth for the next 12 months. The defensive theme is then an equal-weighted combination of forecasted inflation and indebtedness.
Following is a summary of their findings:
- The credit component is the most important driver of hard-currency EM bond returns, explaining 82% of the variation in returns versus just 18% for the duration component.
- Individually and in combination, the factors are significantly associated with future credit excess returns of EM hard currency bonds.
- The systematic factors provide economically meaningful excess of benchmark returns in the context of a systematic long-only, transaction cost-aware portfolio.
- The carry strategy retains some sensitivity to periods of market volatility (sharply lower returns around the financial crisis period and over the latter period of our sample). Given the negative correlation between carry and defensive, it is not surprising to see the defensive portfolio generating positive returns in those same periods.
- Sharpe ratios of long/short portfolios ranged from 0.3 for carry to 0.6 for momentum. Given the low correlation across the four themes (carry and defensive are strongly negatively correlated, and value and momentum are moderately negatively correlated), a portfolio allocating risk equally across the four themes had a Sharpe ratio of 1.1. There is an important benefit provided by diversifying across unique sources of risk.
- After accounting for transaction costs, a long-only portfolio generates an excess of benchmark return of around 2.5%. The authors noted: "This suggests that it may be possible to build a long-only EM bond portfolio seeking exposure to systematic investment themes.”
Brooks, Richardson and Xu then examined the returns of a broad set of 64 long-only, active EM managers with an explicit EM bond benchmark over the 2004-2018 time period.
Are active EM bond managers generating alpha?
The authors found that the average active EM manager had an excess of benchmark return of 0.8%. They then examined whether those excess of benchmark returns are alpha or a result of exposures to traditional risk premia. They found that most of the active returns can be explained by systematic (passive) exposure to traditional market risk premia – the “alphas” are about 75% lower than the excess of benchmark returns and are not significantly different from zero.
They also found that, “active EM bond managers both individually, and in aggregate, have non-trivial exposures to macroeconomic variables.” On the other hand, they noted that their, “long-only portfolio targeting market neutral exposure to carry, defensive, value and momentum has only minimal exposure to macroeconomic variables. This emphasizes a general benefit of a systematic investing approach as it allows you to control ex ante for exposures to traditional risk premia and macroeconomic variables.”
Summary
Their findings led Brooks, Richardson and Xu to conclude: “A combination of risk-based, cognitive error and market frictions gives rise to the positive risk-adjusted returns associated with each of these investment themes. The fact that these factors work in government bond, corporate bond, FX and now emerging bond markets is both a potential boon to fixed income investors and a wonderful ‘out-of-sample’ test of the original equity-centric results, enhancing our belief that the efficacy of these factors is the result of real forces and not random data mining.”
Larry Swedroe is the chief research officer for Buckingham Strategic Wealth and Buckingham Strategic Partners.
Important Disclosure: This information is for educational purposes only and should not be construed as specific investment, accounting, legal or tax advice. Certain information may be based on third party data which may become outdated or otherwise superseded. Third party data is deemed to be reliable, but its accuracy cannot be guaranteed. The opinions expressed by featured authors are their own and may not accurately reflect those of Buckingham Strategic Wealth® or Buckingham Strategic Partners®, collectively Buckingham Wealth Partners. LSR-21-21
Read more articles by Larry Swedroe
Factor-based models are often criticized for data mining. One way to address that charge is with “out-of-sample” testing over longer time frames. But that takes time. New research provides an alternative out-of-sample test – using emerging-market bonds.
