«Equity Style Investing RONG, WU How to cite: Equity Style Investing, Durham theses, Durham University. RONG, WU (2013) Available at Durham E-Theses ...»
Style investing is based on asset classification, sensible categorisation of assets should be arguably based on characteristics that relate to the asset's cross-sectional expected returns. Under efficient market hypothesis that stock price contains all relevant information, style investing should not be more profitable than any portfolios containing randomly selected subset of stocks. Moreover, single style investing would not be mean-variance efficient as investors do not diversify across styles. Hence equity style investing might be fundamentally risky, and the findings of style premium would suggest that either the markets are inefficient or the traditional asset pricing models are misspecified. Rationalist like Fama and French (1992, 1996) argue that market values and book-to-market ratios (BM) are proxies for risk factors, thus the outperformance of small-cap and value investing is compensation for risk. Daniel and Titman (1997), however, disregard such risk-based interpretation. They argue that firm characteristics do not relate to the covariance structure of stock returns. On the other hand, behaviourists such as Lakonishok et al. (1994) propose that value premium is driven by irrational investors’ overreaction. Namely, investors mistakenly extrapolate past growth rate too far into future but subsequently experience disappointing financial results for the underlying stocks.
This PhD research is motived by several gaps identified in the existing literature. First, while academic study finds the relationship between stock returns, firm characteristics and the business cycle fluctuations, the relative importance of such driving sources is not extensively studied. The first part of this research fills the gap in the literature by explicitly examining how firm-specific characteristics and the business cycle conditions function separately to affect the stock performance based on the size and value-growth categorisations. Specifically, it aims to address a key question: what is the dominant driver that affects the relative style performance, the firm characteristics or the business cycle risk? To achieve that, a set of equity characteristics such as price to cash-flow (PC), dividend yield (DY), market-to-book values (MTBV) and market values (MV) are used to classify stocks into different size, value and growth categorisations and simple style investing strategies are tested. In response to the recent popularity of linking macroeconomic effects with the cross-sectional variations on average stock returns, following the framework of Chordia and Shivakumar (2002), Chapter 3 examines the relative importance of common risk factors and the firm-specific information in determining stock returns across styles by focusing on the role of the predicted risk premias and the pricing errors in the observed style premiums.
Second, this research is also motivated by the benefits of active portfolio management based on the relative style returns within equity style cycles. The divergence of style returns evolve all the time with cyclical nature. Over the time there are styles moving in and out of favour by investors according to their relative past performance driven by changes of investment opportunity set. There is no single style or a mix of styles that can dominate under all economy regimes. If equity style cycles do exist and are of long duration, the reward to take investment strategy by identifying the turning point of the leading styles and to opportunistically transition portfolio holding to next prevailing market segments should be massive. Motivated by that, Chapter 4 investigates a dynamic tactical trading strategy by applying a binomial approach to focus on the shifting between pairs of equity styles such as value versus growth or small versus large styles. Each time investors extrapolate the relative performance of different asset classes based on their past performance and bet 100% of investing on the ‘winner’ style financed by shorting the ‘loser’ style. Previous research documented the value of such price-driven strategies like the momentum of Jegadeesh and Titman (1993) and the contrarian of De Bondt and Thaler (1985). However, momentum strategies along the style level have not been well studied, in particular in the U.K. stock market. Chapter 4 contributes to the extant literature by providing valuable empirical evidence in the U.K. stock market to compare with other studies in different economic and institutional environments.
The research in this Chapter answers 2 key questions of whether investors can profit from the information of equity style cycles and whether the return dynamics of equity style momentum is distinct from price and industry momentum effects.
Chapter 5 of this PhD thesis is motivated by the apparent gap in the literature about the optimal multi-asset investing over the business cycles. Substantial evidence suggests that the distributions of stock returns contains time-varying predictable component in the business cycles. The benefit of considering business cycle predictors on asset allocations on the stock level is well studied. However, the portfolio choice implications of business cycle effect in prior studies often focus on the time-varying nature of return distributions driven by business cycle predictors, but the role such economic variables play in affecting optimal multi-style level allocation is less directly explored. Motivated by this gap, Chapter 5 implements an optimisation framework to test several equity style investing based on business cycle information and examine the ex-ante in-sample and ex post out-of-sample performance.
By answering questions such as which economic variable or a set of variables should be tracked when implementing optimal style and how to adjust the exposures to specific market segments given shocks to such underlying variables, Chapter 5 gives multi-style investors like ‘fund of hedge funds’ managers an intuitive advice to optimise their asset allocations when incorporating business cycle predictability.
This PhD research has yielded several meaningful conclusions. First, consistent with the literature, significant size and value premiums are found in the U.K. stock market over the period of 1980:01-2004:12, justifying the applicability of simple equity style investing strategies.
The outperformance of investing small-cap and value stocks are more pronounced during recessionary periods. It is again found that the underlying driving forces determining the dynamics of relative style performance are indeed much controversial. Overall, the divergent returns of small-cap versus large-cap stocks and the value versus growth stocks as characterised by PC and MTBV are mainly driven by the cross-sectional pricing errors in the context of a multifactor business cycle model. This would suggest that the outperformance of small stocks and the better returns of investing in value stocks with low PC or MTBV (i.e. high BM) may be caused by investors’ irrational trading behaviour to such stock groups that result from cognitive biases like underreaction to firm-specific news. In contrast, the outperformance of value stocks with high dividend yield (DY) is likely to be attributed to cross-sectional difference in conditionally expected returns predicted by business cycle model. Therefore it represents the compensation for bearing business cycle risk. It is also concluded that although on the individual stock level the relative returns of value stocks based on PC and MTBV sorting are not likely driven by the business cycle risks, on the portfolio level the business cycle model could still partly capture the time-series expected value premiums.
Hence equity valuation multipliers such as PC, DY and MTBV contain time-varying predictable component in the expected returns, which is consistent with findings of empirical studies focusing on time-series relations among expected returns, risk and equity characteristics (e.g.
Fama and French (1993, 1996), Kothari and Shanken (1997), and Chan et al. (1998), among others).
The profit of style momentum strategy would suggest the existence of U.K. equity style cycles. Since styles perform differently during various stages of a market cycle, investing strategies to buy stocks in current in-favour styles could continue to outperform those in current out-offavour styles for a period up to 12 months or possibly longer. Such payoffs generally increase with longer ranking periods and decrease with longer test periods. Consistent with the literature, it is found that style momentum effect has strong independent explanatory power for the future individual stock’s expected returns, and style momentum is distinct from price momentum of Jegadeesh and Titman (1993) and industry momentum of Moskowitz and Grinblatt (1999) documented in the literature.
The empirical findings in Chapter 5 concludes that on a strategic perspective investors tend to significantly hold value stocks or smallcap stocks, and short sell growth stocks or large-cap stocks in their optimal style allocation process. It is much different for style investing incorporating or disregarding business cycle effects. Disregarding the business cycle predictability would usually introduce a strategy that is relatively conservative regarding the overall net equity exposures as compared to those that incorporate strong prior beliefs about the business cycle conditions. Style investing incorporating business cycle predictability generally result in more extreme weights to some styles at both long and short sides, possibly because investors believe that because of predictability such extreme exposures can be eventually reduced at bad times when the investment opportunity set changes. It is also suggested in Chapter 5 that business cycle predictors affect the conditional equity style returns and the optimal style investing in a different mechanism. Indeed, economic pervasive variables such as yld, term, div and def exert a strong influence on the shape or location of the optimal style investing frontier. Style investing capitalising on the conditional business cycle information consistently beat that disregarding such business cycle influence, both in-sample and outof-sample.
6.3 The practical implementations
The empirical findings in Chapter 3, 4 and 5 would have practical implementations in the investment practice. First, the findings in chapter 3 provide practical guidance for active portfolio management.
Portfolio managers who pursue style investing by allocating their funds to characteristic-sorted asset groups must first understand the different risk-related mechanism behind the observed divergent style returns. For example, if the return differentials are driven by bearing macroeconomic risks, active style management should aim to incorporate the business cycle effect. Conversely, if risks outside the business cycles drive the mispricing are the major driving forces of the relative style returns, style timing should focus on identifying the underlying stock groups related to investors’ trading behaviour.
The profitability of style momentum documented in Chapter 4 would suggest how investors could manage their portfolio’s style exposure efficiently. Namely, style exposures can be bought and sold and the investing portfolios can be constructed with desired style exposures, both positive and negative according to the style performance relates to market cycles. This technique can be easily implemented to help passive investors enhance the returns. Passive investors normally invest on an index fund. Index fund is a mutual fund or exchange traded fund (ETF) with a clearly predefined set of constituents that are constant regardless of market conditions. Passive investors do not expect to beat the overall market but rather pursue average market returns. Such strategy may be supported by efficient market hypothesis but clearly lacks efficiency. The divergent style returns under different market regimes indicates that equity style exposures can be used to hedge the inefficiency of an index fund by eliminating its least attractive portion. Extant literature regarding index hedging focuses primarily on the application of derivatives such as options and futures. The results in Chapter 4 provide a plausible method of adaptively constructing long short market neutral style portfolios to hedge the deficiency of an index fund under different state of the economy.
The research findings in Chapter 5 offer a simple yet intuitive way for mean-variance investors to optimise their style allocations. First, investors like ‘fund of funds’ managers are advised to incorporate business cycle information when implementing active style investing.
Using macro information to assist in style selection has always been a hot topic in the quant circles in the investment community. There is certain evidence to suggest that different style factors are more or less relevant during different states of the macroeconomy conditions.
Hence conditional multi-style investing strategies following business cycle information generally outperform the unconditional strategies.
Second, mean-variance multi-style investors could follow simplified optimisation approaches such as Brandt and Santa-Clara (2006) to parameterising directly on business cycle predictors when applying optimal style allocation. Namely, investors could follow a dynamic approach that is ‘macro driven’ to timing their style investing. By doing this a set of business cycle related economic variables should be tracked which forms the ‘tradable environment states’. With the popularity of Exchange Traded Funds (ETF) and its flexibility and low trading expenses and high liquidity in leading financial markets, a combination of such optimal hybrid strategy should arguably help investors to squeeze more juice from the investing returns.
6.4 Recommendations for areas of future research Despite enormous effort has been devoted to this PhD research, due to data availability and the time constraints, the author has identified several directions where further research is needed. The areas of
recommended further research include the following: