From smart beta indexing to factor investing

Post with image

Over the last few years, interest among investors in indexation has driven asset managers to develop new forms of indexation also known as smart beta. Raul Leote de Carvalho, co-head of Financial Engineering at BNP Paribas Asset Management, explains that smart beta strategies are, however, driven by exposures to factors known to generate a positive premium.

These factor premiums have been extensively discussed in academic literature. They include value, low volatility and small capitalisation. Raul believes that a growing number of investors are now becoming aware of the compelling arguments in favour of this approach and that their attention is already shifting away from the more empirically defined smart beta indices available today towards more efficiently crafted factor-centric strategies.

What is smart beta?

Smart beta indices, based on relatively transparent quantitative approaches, represent a move away from traditional market-cap weighted indices toward alternative weightings purporting to offer either better diversification or greater efficiency than the standard market cap-weighted approach. Historical back-tests of these indices show better risk-adjusted returns than market capitalisation indices, often with not only higher returns but also lower risk.

However, smart beta indices should not be regarded as a form of passive investing. Passive management can be used to replicate smart beta indices but the indices themselves are a form of active investing in that subjective assumptions and choices are required as to where the emphasis is to be placed: they establish the rules to decide on how to deviate systematically from the market capitalisation portfolio, how frequently and by how much. Only the capitalisation-weighted portfolio can be considered as a genuinely passive strategy – it is the only buy-and-hold portfolio that theoretically could be held in equilibrium by every investor.

What are factors?

Factors can be thought of as characteristics of stocks that are important to explain their risk and performance. According to the Capital Asset Pricing Model (CAPM) of the 1960s, the performance of a stock should be determined by one single factor, i.e. the stock exposure to the market portfolio which is measured by a stock characteristic known as beta. Beta is calculated from estimating by how much a stock price co-moves with the price of the market portfolio, usually proxied by a market capitalisation index.

However, as demonstrated in a large number of academic papers published since, academics found that the CAPM is not verified empirically when tested with historical stock prices and more than one factor is needed to explain the performance and risk of stocks. Such papers highlighted that other factors also play an important role. Today it is widely accepted that factors such as value, volatility, quality, momentum and capitalisation may also play a role in explaining stock returns.

Value, volatility, quality, momentum and capitalisation factors

Value characteristics such as the price-to-book or the price-to-earnings of a company measure how cheap a stock is relative to others. It has been found that on average, over time, cheaper stocks tend to out-perform other stocks, in particular expensive stocks.

Volatility is a risk measure indicating by how much the price of a stock fluctuates. Evidence that less volatile stocks, with less price fluctuations, generate at least comparable returns to riskier stocks renders low volatility stocks much more attractive for investors: same returns in the medium to long-term with less uncertainty.

It is also known that higher quality stocks, e.g. the most profitable companies, tend to generate higher returns than other stocks, in particular when compared to the least profitable companies. Different measures of profitable can be used, e.g. return-on-equity.

Stocks with the strongest price trends, e.g. stocks with the strongest out-performance relative to other stocks as measured over the previous 12 months also tend to continue to outperform. This is known as momentum.

Finally, smaller capitalisation stocks tend to generate higher returns than larger capitalisation stocks over the medium to long-term.

What are factor premiums?

Factor premiums are the stock returns explained by their exposure to factors. The premium of value stocks, low-volatility stocks, quality stocks, strongest-trending stocks and smaller capitalisation stocks is positive on average over time. Investors have an interest in tilting their portfolios in favour of such stocks to earn their factor premiums. Conversely the premium of expensive stocks, risky stocks, poor quality stocks, stocks with the weakest price trends and largest capitalisation stocks is negative on average. Investors do better to stay away from such stocks and avoid their negative premium which would reduce returns.

Many papers discussing the source of factor premiums have been published by academics over the past decades. Factor premiums are believed to result from the fact that neither investors behave as financial theorists assumed in the 1960s nor are markets as perfectly functioning as they believed. The assumptions behind the CAPM simply do not hold, which explains why more than one factor is needed. Guideline constraints on portfolio allocation, regulations constraining what investors can do, investors’ reluctance, for example, to employ leverage, investor greed, investor over-confidence, transaction costs, and even the motivation for investing which is not always to maximise absolute returns, all play an important role and contribute to explaining why factor premiums exist.

Why do smart beta indices outperform?

Over the last years, we have undertaken extensive research into these alternative weighting approaches. We were among the first to highlight that risk and returns to smart beta strategies can be almost entirely explained by the fact that these strategies tilt portfolios in favour of the cheapest stocks, the least volatile stocks and sometimes also the smallest capitalisation stocks. Smart beta portfolios are exposed to the value, low volatility and small capitalisation factors. As we demonstrate in a paper published in the Journal of Portfolio Management – Spring 2012, risk and returns of a class of smart beta strategies known as risk-based strategies, which include exotic names like minimum variance, maximum diversification and risk parity, are derived almost entirely from such factor exposures. This evidence has been more recently confirmed by a number of other researchers.

Fundamental indexing, another type of smart beta strategy, has also been shown to derive its excess returns from exposures to factors, in particular the value factor.

Are smart beta indices efficient?

Our research lead us to believe that while smart beta investing represents a good start to capture factor premiums, investors can do much better. We demonstrated that the much proclaimed simplicity of smart beta indices comes at the cost of an inefficient harvesting of the factor premiums they are exposed to. In addition, obtaining exposures to both the quality and the momentum factors is much more difficult in smart beta indices. This is because smart beta indices have not been designed to optimally capture factor premiums. When smart beta strategies were introduced, little was known about the sources of their risk and performance, and explanations of why they should achieve higher risk-adjusted returns than market-capitalisation indices, were vague and unconvincing. Only now is it clear that factor premiums are the key drivers behind their higher risk-adjusted returns.

From smart beta indexing to factor investing

If a portfolio of multiple smart beta indices earns excess returns from multiple factor exposures why not just build a robust portfolio with the desired factor exposures? The fact that there are only a handful of known factors believed to pay a premium constitutes a major advantage in facilitating such an approach – there’s no need to analyse and choose among the hundreds of smart beta indices available today. Another advantage is that investors can decide which factors to include in their portfolios and the risk exposures to those factors they should aim at.

Efficient factor investing

Investors can build robust portfolios with the desired factor exposures. In a paper published in the Journal of Asset Management – March 2014, we propose a framework that is efficient at controlling factor exposures and at capturing factor premiums even after applying a set of typical portfolio constraints, e.g. prohibiting the use of leverage.

The risk associated with factors such as value and momentum can be controlled rather successfully, and the volatility of factor returns can be forecast thanks to a property know as clustering, i.e. the fact that the volatility tends to stay high when it is high and tends to stay low when it is low. Furthermore, controlling the risk exposure to factors can significantly increase the risk-adjusted returns they generate. This is explained by the fact that factor premiums tend to be much higher when factor volatility is low and they tend to be much smaller, sometimes even negative, when factor volatility is high. We discussed this phenomenon in a recent paper published in the Journal of Investment Strategies – December 2014.

Exhibit 1: Decomposition of the excess returns of a simulated risk-controlled global multi-factor strategy exposed to value, momentum, quality and low-risk factors and benchmarked against the MSCI World Index (unhedged net returns).

excess returns versus msci world

Source: BNP Paribas Asset Management, THEAM, MSCI and Exshare on 31 July 2014.

Hypothetical performance results have many inherent limitations and have been obtained with the benefit of hindsight. Past performance is not indicative of future results.  [divider] [/divider]

In exhibit 1 we show the cumulative outperformance relative to the MSCI World index (net return) derived from the application of this framework to build an efficient multi-factor strategy for global stocks. The excess returns are derived from exposures to value, quality, momentum and low volatility factors. The portfolio is rebalanced every month. In exhibit 2 we show the cumulated performance of the same strategy compared to the cumulated performance of the MSCI World index, the benchmark index. The results in both exhibits are based on monthly total returns in USD and gross of fees. The tracking error risk over the MSCI World index is 4.3% annualised. The back-tested average annual excess return over the period is 7.7% annualised. Market impact and transaction costs were not included.

Exhibit 2: Simulated net asset values for a risk-controlled global multi-factor strategy exposed to the value, momentum, quality and low risk factors and benchmarked against the MSCI World Index (unhedged net returns).

simulated asset values

Source: BNP Paribas Asset Management, THEAM, MSCI and Exshare on 31 July 2014.

Hypothetical performance results have many inherent limitations and have been obtained with the benefit of hindsight. Past performance is not indicative of future results.  [divider] [/divider]

The future of smart beta

We believe that the focus on smart beta will continue to shift away from the more simplistic approaches available today. They often lack a strong theoretical background or are based on erroneous assumptions, e.g. it can be demonstrated that the maximum diversification strategy is efficient only if the Sharpe ratio of all stocks is equal, something neither observed empirically nor expected by financial theory.

We expect that factor-centric strategies will increasingly take over. We also expect investors to grow aware of the fact that these strategies, even if they may offer large capacity, will never be as liquid as the market-capitalisation indices and that market impact should be taken into account. Ultimately, investors of different sizes should realise they need to contemplate different factor strategies with different levels of turnover and potentially even different risk budget allocation to factors and to the active risk they generate to most efficiently capture factor premiums.

Raul Leote de Carvalho

Deputy Head of Quant Research Group

Leave a reply

Your email adress will not be published. Required fields are marked*