Olivier Laplénie (pictured), head of quantitative fixed income management, BNP Paribas Asset Management Charles Cresteil, quantitative investment specialist, BNP Paribas Asset Management
Following the spectacular development of factor investing in equity markets, fixed income investors are increasingly turning to factor-based strategies too, and there are good reasons to think that factor-based investments should be an integral part of their strategic allocation.
While the traditional approach to bond investing essentially focuses on actively managing duration, credit risk and/or currency exposure – three parameters that can be characterised as directional risks or ‘beta’ – factor investing, on the other hand, aims to delve into all the other sources of risk — and returns — most likely to drive markets.
The objective of factor-based strategies is thus to improve risk-adjusted returns by actively targeting these non-directional sources of performance, also called factor premiums. In other words, factor-based strategies are designed to generate pure alpha, i.e. returns that are not linked to any active view on the direction of interest rates or credit markets.
The factors that drive bond markets The four following groups of factors are considered critical drivers of performance in fixed-income markets:
• Value/carry factors: focused on price-related data, mostly with a value for money – or carry for risk – approach
• Fundamental quality factors: focused on non-price-related data, such as company-specific data for corporate bonds and country-specific data for government bonds (macroeconomic data, for example)
• Momentum-type factors: based on market dynamics and sentiment, they are a way to capture market moves that may not be fully reflected by fundamentals
• Low-risk factors: built on the empirical observation that the bonds with the lowest risk tend to offer higher risk-adjusted returns over the long term.
These factors have been documented in academic research for their capacity to explain long-term returns, and have been the subject of extensive historical testing. They are based on relatively simple concepts which have been used by active portfolio managers for decades – only in a less formalised and disciplined way.
Building a systematic exposure to factors While the concepts that underpin factors are straightforward and well-recognised, building an efficient factor exposure requires an elaborate investment process. Indeed, selecting factors is only half of the story, as methodological choices made to implement factor strategies are equally important, if not more, to ensure a factor-based strategy can achieve long-term out-performance.
In particular, one area which requires sophistication is the removal of potential directional biases in a factor portfolio. For instance, consider a naïve carry/value factor, which would overweight bonds with high carry and underweight those with low carry. If this factor is not adjusted for beta, it will be biased toward high-beta bonds/ countries/sectors, and much of the final performance of the factor will be the result of this beta bias.
Combining factors Another key methodological choice is how factors should be combined in portfolios. Factors have been shown to outperform over the long-term, however extended periods of underperformance for a given factor are not uncommon. This is why, multi-factor strategies are generally preferred by investors, as diversifying across multiple factors not only improves total risk-adjusted returns, it also results in shorter drawdown periods.
For this to be true however, the underlying correlations between factors need to be low, not only on average, but also on tail events – i.e. during drawdown periods. Using factors from different investment styles is a first step in that direction – for example value and quality styles tend to be complementary as the former essentially focuses on finding good yield, while the latter favours companies with the most robust business model. However, the construction methodology used to implement factors is also crucial in that respect, because if the factors do not properly account for potential biases, they risk ending up being highly correlated during some periods.
Static or dynamic allocation: we favour a balanced approach In view of the differences in the short-term performances of factors, some investors might be tempted to deviate from a balanced factor allocation to introduce a ‘timing’ mechanism.
The aim would be to anticipate which factor should outperform based on its recent performance or on where we stand in the economic cycle. One such strategy can be to look at the valuations of factors with a mean-reverting approach, i.e. under-weighting factors that have outperformed compared to their long-term rhythm and over-weighting factors that have been out of success recently.
However, the analyses of our Quant Research Group at BNP Paribas Asset Management suggests that the added value of such strategies tends to be quite low historically, and is often negative when turnover-related costs are factored in. Another option can be to look at the sensitivities of each factor to the economic cycle, but our research shows that the correlations between factor returns and the cycle are usually the result of directional biases in the factor construction. Hence, such correlations are much less apparent when ‘purified factors’ are used. As a result, maintaining a balanced allocation to the different factors seems to be the most efficient and straightforward approach for most long-term investors.
Using factors: a tailored approach for institutional investors Factor investing can be of interest to different institutional investors for different reasons. For pension funds it can be valuable to analyse the factor exposure of their existing portfolio. For example, they may find out that they have a significant allocation to value-style factors from their active managers. In that case, it would make sense for them to consider a tailored factor solution, in order to balance the factor exposure for their overall portfolio, by favouring the factors that are under-represented in their portfolio.
Insurers have been under pressure in recent years as they try to preserve their investment returns while controlling risk. With limits to how much they can invest into risky assets, it is particularly important for insurers to extract greater return from less risky portfolios of assets. Factor-based products can serve as powerful tools to capture uncorrelated return opportunities and improve portfolio diversification.
Conclusion With the development of factor investing, a new breed of strategies has emerged with the objective of improving risk-adjusted returns by focusing on the underlying drivers of fixed-income markets, such as carry, fundamentals, risk and momentum. By building purified factors and combining them efficiently, it is possible to build investment solutions that can help address many core needs of institutional investors, enhancing long term risk-adjusted returns and improving portfolio diversification.
