Multi-factor credit portfolios: more stable alphas
We have found strong empirical evidence for the existence of Size, Low-Risk, Value and Momentum factor premiums in the corporate bond market. All factors have substantially higher returns and Sharpe ratios than the market. The tracking errors, however, are relatively large, highlighting the risk of underperforming the market over shorter investment horizons. By investing in a multi-factor portfolio, which diversifies across the four factors, the tracking error and drawdowns versus the market are reduced while the high returns and Sharpe ratios are preserved.
Corporate bond factor allocation pushes up the Sharpe ratio of a multi-asset portfolio
As most investors are also invested in other asset classes, such as equities or government bonds, we have researched the added value of corporate bonds factors in a multi-asset portfolio as well. We have analyzed a hypothetical multi-asset portfolio containing 20% government bonds, 40% equities, 20% investment grade corporate bonds and 20% high yield corporate bonds. In a traditional portfolio, all allocations are to the market indices. Next, we have tested three alternative allocations, where we:
allocate only the equity portfolio to a multi-factor portfolio
allocate only the corporate bond portfolios to the multi-factor portfolios and
allocate both the equity and corporate bond portfolios to multi-factor portfolios
Figure 1: Sharpe ratios market and factor portfolios over the risk-free rate for government bonds, investment grade (IG), high yield (HY) and equities

Sources: Robeco, Barclays, Bloomberg, Data Library Kenneth French. Period: 1994-2013.
Figure 2: Sharpe ratios multi-asset portfolios over the risk-free rate

Sources: Robeco, Barclays, Bloomberg, Data Library Kenneth French. Period: 1994-2013.
Restez informé sur l'investissement quantitatif
Receive our Robeco newsletter and be the first one to get the latest insights, or build the greenest portfolio.
Figure 1 shows the Sharpe ratio of the market and the multi-factor portfolio per asset class. The improvements of the factor portfolios versus their respective market indices are substantial, with Sharpe ratios increasing by 0.16 (IG), 0.28 (HY) and 0.26 (equities). Figure 2 shows the Sharpe ratio of the multi-asset portfolio and the three alternative portfolios. Investing in the corporate bond multi-factor portfolios boosts the Sharpe ratio from 0.70 to 0.81. Investors that already allocate to the equity factors and decide to invest in the corporate bond multi-factor portfolio too, see their Sharpe ratio grow from 0.89 to 0.97. In both cases, the corporate bond factor allocation contributes almost 1% to the improved return, while the volatility is virtually unchanged.
Robeco’s enhanced approach: smarter definitions and portfolio construction
The results we have analyzed the benefits of factor investing following academic conventions. However, Robeco research has shown that it is possible to improve upon these results in two ways, i.e. by using:
smarter factor definitions and
smarter portfolio construction rules
Smarter factor definitions
To smarten the factor definitions, it is important to understand the latent risks in each factor and to mitigate these risks if they are not properly rewarded with higher returns. Moreover, since risk itself is unobservable and multi-dimensional it is advisable to diversify across risk measures. We found that is especially beneficial to expand the scope of the risk measures beyond bond market characteristics, and use accounting and equity data as well. For instance for Low-Risk, we do not only control for risk via rating and maturity, but also for the amount of leverage the company is taking on, and how much risk its equity shows. To enhance Value, we do not only look at rating and maturity to calculate the ‘fair’ credit spread, but also at company characteristics. Moreover, we use equity market information in our Momentum definition.
Smarter portfolio construction rules
To smarten the factor definitions, it is important to understand the latent risks in each factor and to mitigate these risks if they are not properly rewarded with higher returns. Moreover, since risk itself is unobservable and multi-dimensional it is advisable to diversify across risk measures. We found that is especially beneficial to expand the scope of the risk measures beyond bond market characteristics, and use accounting and equity data as well. For instance for Low-Risk, we do not only control for risk via rating and maturity, but also for the amount of leverage the company is taking on, and how much risk its equity shows. To enhance Value, we do not only look at rating and maturity to calculate the ‘fair’ credit spread, but also at company characteristics. Moreover, we use equity market information in our Momentum definition.
Conclusion
There is strong empirical evidence for the existence of Size, Low-Risk, Value and Momentum factor premiums in the corporate bond market. By investing in a multi-factor portfolio, the tracking error and drawdowns versus the market are reduced in comparison with single-factor portfolios, while the high returns and Sharpe ratios are preserved.
In a multi-asset context, by allocating to corporate bond factors investors can improve the Sharpe ratio by 0.1 and their return by about 1%, regardless of whether they already allocate to factors in their equity portfolio. Although these results are already strong, there is still much to be gained by enhancing the investment process. This can be done by using smarter factor definitions and by improving portfolio construction rules.