How factor investing also works for corporate bonds

How factor investing also works for corporate bonds

11-11-2014 | インサイト

Two Robeco researchers have become the first to analyze the effect that factor premiums can have on corporate bond investing.

  • Jeroen  van Zundert
    van Zundert
    Researcher Quantitative Credits
  • Patrick  Houweling
    Head of Quant Credits

Speed read

  • Factors used in equity investing also work in corporate bonds
  • Multi-factor portfolios proved to outperform the benchmark
  • Research was back-tested over 20 years using USD bonds
  • Strategies target low-risk, value, size and momentum factors

The work by Patrick Houweling and Jeroen van Zundert aims to show that the factors used successfully in equity market investing can also work in the corporate bond market.

Investors have long used the four main factors low-risk, value, size and momentum to pick stocks that on average have a higher-risk adjusted performance than others due to their characteristics. However, relating this quantitative investment style to corporate bonds is still in its infancy.

Their research into the effects of value, for example, is so new that the industry does not yet even have a commonly agreed definition as to what ‘value’ means when applied to corporate bonds.


Starting a dialogue

“The paper is meant to start a dialogue with investors after we received more and more questions about this,” says Patrick Houweling, quantitative researcher and portfolio manager for the Robeco Conservative Credits strategy. “Some had already implemented factor investing on the equity side and wanted to apply it to fixed income too, and in particular, to corporate bonds.”

“There are a couple of papers out there on low-risk and momentum credits, and we know of one paper on value, which is quite technical. But we don’t know of any papers on size, and ours is the first one to apply the whole concept of factor investing to a single portfolio of corporate bonds.”

“We have assessed the benefits of combining multiple factors in a single portfolio to see what kind of returns you could get,” he explained. “We also looked at the added value this generated in bonds versus those for the equity market factors. As far as we are aware, this is all brand new research.”

Houweling and his fellow quant researcher Van Zundert spent months writing a ground-breaking white paper, Factor Investing in the Corporate Bond Market, posted on 31 October 2014 on the Social Science Research Network (SSRN) academic website.

“The research on the equity side was already there, so we built on the same philosophy,” says Van Zundert. “This does not mean that we could simply copy and paste the equity factor definitions. For factors such as momentum, the idea is similar by looking at past returns, but for the value factor we think that a fixed income investor will typically look at spreads instead of book-to-market ratios. So we only looked at bond market characteristics to capture the essence of the factors: how a bond investor would use them, and not an equity investor.”

“The hardest part was that we didn’t have a base paper saying ‘this is the definition of the factor when applied to the bond market’. We had to define it ourselves,” he says. “And there’s no single way to do it, so we tested a number of different measures one could come up with, such as by looking at price instead of spread for value, or by looking at past 3-month returns instead of 6-month returns for momentum. Our findings are robust to such alternative definitions.”

‘We only looked at bond market characteristics to capture the essence of the factors’

Back-tested over 20 years

As there are currently no funds with long track records in this field, their work was back-tested from a data set from 1994 to 2013 using US dollar corporate bonds. “We ran the tests over a big pile of historical data and tracked the factor portfolios through time to see if the effects are really there, to make sure that the higher returns were not just a lucky shot,” says Houweling. “The low-risk and momentum results were confirmed by other studies which had similar results on these factors.”

Applying the four factors to corporate bond investing is going into mostly uncharted territory. Robeco’s Conservative Credits strategy is focused on the low-risk factor, also taking momentum, value and size into account. “We don’t have a pure value strategy yet for corporate bonds, but we have done all the research so that we can deliver one,” says Houweling.

“It’s the same for multiple factors – we want to be able to offer a multi-factor solution if clients ask us to do so. That’s the stage where the market is at – people need time to get up to speed with factor investing in the corporate bond market. We expect a lot of activity next year on this topic.”

The research showed that investing in such a multi-factor portfolio doubles the Sharpe ratio that could be achieved versus investing in the market index. The annualized alpha of the multi-factor portfolio is 1.00% for investment grade and 3.21% for high yield, which is sizable given their market premiums of 0.59% and 2.46% respectively.

‘People need time to get up to speed with factor investing in the corporate bond market’

How the factors work

So how do the factors themselves work? “The themes that lie behind the factors used in equities are usually related to human behavior, or how markets function, or how regulators look at markets, and these explanations apply equally well to corporate bonds. Herding behavior, for example, which causes the momentum effect does not just apply to equities,” says Houweling.

Low-risk investing exploits an anomaly which has been proven to show that apparently ‘dull’ stocks can generate returns that are just as good as the ‘exciting’ ones, but without the higher risk. The study shows that the same is true for low-risk corporate bonds.

Momentum works on the principle that past returns indicate future returns: a stock doing well in the past is likely to continue to do well in the future. “What we find in corporate bonds is that this premise works in high yield but not for investment grade bonds,” says Houweling.

The size factor targets companies that have been overlooked by the market. “Smaller companies tend to be ignored by the vast majority of investors, so we can get a neglect premium,” says Van Zundert.

Value looks for companies whose market price does not reflect its fundamentals. “We look at the expected return that you would get after taking into account what the rating agencies say and what the spread curve indicates” says Van Zundert. “If the expected return of the bond more than compensates for the risk being taken, it is worth owning. You can say it is cheap then.”

For their application in actual investment portfolios, the study recommends combining multiple factors instead of relying on a single factor. It is important to make sure that one factor doesn’t counteract another. “For example, we don’t want to buy the most expensive low-risk bonds or cheap high-risk bonds” says Houweling. “Therefore we also take value and momentum into account in a low-risk strategy. Also, we construct equal-weighted portfolios to benefit from the size premium.”

‘We don’t want to buy the most expensive low-risk bonds or cheap high-risk bonds’

Trading costs

But isn’t it more expensive to trade in corporate bonds, which are sold ‘over-the counter’ by financial institutions rather than over an exchange like equities? “Corporate bonds are less liquid than equities, and so transaction costs are higher,” says Houweling. “So we devoted a section of our research to finding out whether the returns were still higher than the transaction costs. If you have to replace a holding, for example, to keep exposure to the value factor, then how costly is that for the investor? We found that all these factors still delivered substantially higher returns after costs.”

“It’s important to recognize that some bonds are more expensive to trade than others, but even when we doubled the transaction costs to be on the conservative side in our analyses, we still found that the returns were still there. So this is not just theoretical research – it would actually give you higher returns in your portfolio.”


当資料は情報提供を目的として、Robeco Institutional Asset Management B.V.が作成した英文資料、もしくはその英文資料をロベコ・ジャパン株式会社が翻訳したものです。資料中の個別の金融商品の売買の勧誘や推奨等を目的とするものではありません。記載された情報は十分信頼できるものであると考えておりますが、その正確性、完全性を保証するものではありません。意見や見通しはあくまで作成日における弊社の判断に基づくものであり、今後予告なしに変更されることがあります。運用状況、市場動向、意見等は、過去の一時点あるいは過去の一定期間についてのものであり、過去の実績は将来の運用成果を保証または示唆するものではありません。また、記載された投資方針・戦略等は全ての投資家の皆様に適合するとは限りません。当資料は法律、税務、会計面での助言の提供を意図するものではありません。




商号等: ロベコ・ジャパン株式会社  金融商品取引業者 関東財務局長(金商)第2780号

加入協会: 一般社団法人 日本投資顧問業協会