By continuing on this site you have agreed to cookies being placed and accessed by this website. More information and adjusting cookie settings.
Most academic studies on factor investing are about equities. Patrick Houweling and Jeroen van Zundert show that factor investing also works for bonds. How has their research paper been used to create a fund?
On 15 June 2015, the Robeco Global Multi-Factor Credits Fund was launched. This fund, which invests worldwide in investment-grade bonds, enables institutional investors to take advantage of four factors: Low-Risk, Value, Momentum and Size. Low-Risk selects low-risk bonds of low-risk companies, Value selects bonds that are cheap relative to their risk and Momentum selects recent winners. In addition to these three factors which Robeco also uses in its equity factor strategies, Robeco Global Multi-Factor Credits takes the Size factor into account. Amongst others, this enables the strategy to take advantage of the liquidity premium, which plays a more significant role in less liquid markets such as the credit market than it does for equities.
Patrick Houweling manages the fund and since 2012 has also been portfolio manager of the Conservative Credits strategy, which implements the Low Risk factor for corporate bonds. In 2014, together with quantitative researcher Jeroen van Zundert, Houweling wrote a research paper titled 'Factor investing in the corporate bond market'. In this study they show that factor investing can also be successfully applied to credits. Their research demonstrates that multi-factor credit portfolios achieved attractive Sharpe ratios and Information ratios in the research period from 1994 to 2013.
Where does the interest for this fund come from?
“Mostly from institutional investors who do not want to use factor investing just for equities – they want to apply factor investing to their entire portfolio and are interested in the prospect of a higher Sharpe ratio and higher returns."
‘An alternative to passive investing’
“Realistic historical simulations that take into account liquidity and transaction costs show that the Global Multi-Factor Credits fund has a significantly higher Sharpe ratio than the Barclays Global Investment Grade Index. One of the advantages of a multi-factor portfolio compared to a single factor one is the effective diversification of relative risk, in other words the likelihood of lagging the index. Weak years for one factor are often canceled out by strong years for others. This results in a more stable outperformance relative to the market than the individual factors produce.”
“Apart from institutional investors who use factor investing as an alternative to their actively managed portfolio, there are also investors who see it as a replacement for passive investing and are considering using it to replace their index portfolios."
“The fact that there is interest in both camps doesn't surprise us because factor investing is an intermediate strategy that falls between passive and active investing. Characteristics of the active form are the high tracking error, focus on outperformance and the more concentrated portfolios. The similarity with the passive style is the more mechanical and systematic way of investing."
What led you to write the research paper on factors in the bond markets?
"This academic research was driven by questions from clients. Robeco already has factor strategies for equities. Many institutional investors apply factor investing to stocks or have decided to do so. They asked us how factor investing could be used for corporate bonds. This is why we documented our work on this subject in an academic research paper.”
‘This academic research was driven by questions from clients’
“It is no surprise that investors first start with equities, because the market there for factor investing is more developed. There are more asset managers offering this style of investment and more has been published about it than about bonds.”
How much exposure is there to factors in clients' existing bond portfolios?
“Active bond portfolios often have a fundamental management style and no quantitative approach, which is why they cannot be termed real factor portfolios. We always offer to analyze clients' portfolios in terms of their factor exposures. For example, if a manager expresses a preference for Low Risk, how strong is that exposure and what is the exposure to the other factors? It is important that clients are aware of what their portfolio actually contains."
What were the focus points in your research to find workable strategies?
“It is important to make the research into factor strategies as realistic as possible: transaction costs and liquidity are essential focus points in this. Certain bonds are more expensive to trade in than others, depending on their liquidity. For example, recent issues are more liquid than older ones. Limited liquidity also means that you cannot buy all the bonds you would like to. This is why we have developed a model that calculates how likely it is that you will be able to buy or sell a bond, given its characteristics.”
"Another element is portfolio construction. Academic research usually assumes an investment horizon of twelve months. But this evaluation is too rigid – sometimes too long and sometimes too short. In the actual strategy, we constantly evaluate a bond's attractiveness. As long as it remains attractive you don't need to sell it. Our strategy's turnover is dynamic and differs from one bond to another. This helps keep down transaction costs. The result is an average investment horizon of two years; longer than in the academic methodology and so this also saves on transaction costs.”
“The integration of the various factors in a portfolio is another major focus point. How do we achieve balance in the model so that, on average, you achieve equal factor exposure? We want to maintain a well-diversified exposure to all four factors.”
Which factor was the most difficult to implement?
“Momentum, because this changes so fast. Bond rankings change quickly over time, which means that the turnover rate is higher than it is for the other factors. This makes it difficult to set up a pure momentum strategy for bonds at the current time. The challenge is to maintain a good exposure to this factor without trading too much."
“But because we combine the Momentum factor with other factors we can keep turnover down: for example, if two bonds score the same on Value, it is better to buy the one that scores best on Momentum. It is important to incorporate this in a multi-factor strategy. The Momentum factor combined with the others improves diversification and generates a higher Sharpe ratio than a multi-factor portfolio without Momentum.”
Does the Size factor cause liquidity risks?
“If we were to select directly on the basis of the Size factor, then liquidity would be a risk as the portfolio would be almost entirely made up of bonds issued by smaller companies. But our indirect approach helps limit the possible risks."
“We position ourselves for the Size factor indirectly, by choosing a broad universe which also includes smaller names. Furthermore, we construct a well-diversified portfolio with a large number of issues. We strive to achieve equal weighting in the portfolio: so relative to the benchmark we underweight large companies and overweight the smaller ones. We do not need to take any large positions, because we hold a wide range of issues in the portfolio. This approach ensures that the factor exposure of Size is almost the same as those of Value, Momentum and Low Risk.”
Have your experiences with Conservative Credits helped you establish this strategy?
“We have used many parts of the Conservative Credits investment policy. We follow the same five steps as used in the Low Risk strategy. We start with a broad universe, restrict that universe, rank the remaining bonds, carry out analysts’ checks and finally construct the portfolio."
“But there are important differences too. We use much stronger restrictions on risk characteristics such as maturity for our Conservative Credits strategy, because this is predominantly a Low Risk strategy. We also have other factor weights in the multi-factor model, because the factors are equally weighted, while in Conservative Credits the emphasis is of course on the Low Risk factor.”
Why have you opted for a global universe?
“For a quantitative strategy, we want to have the largest possible universe in order to make a good selection and construct a portfolio. You need effective diversification and factor exposure to do this."
“The fund's universe comprises more than 14,000 investment-grade corporate bonds from more than 3,000 companies worldwide. The portfolio contains 150-200 issuers, more than a typical active corporate bond portfolio. Many of our clients also want to have a global investment universe and with Robeco Global Multi-Factor Credits we can fulfil this demand.”