Among the most notable trends in the $47 trillion investment management sector has been a shift from actively managed to passively managed funds, whose investments are determined by a fixed procedure. Factor investing or smart beta has become the darling of investors and investment managers. These funds, which typically strive to mimic an index, now control around 50% of investable assets. Factor based funds – mutual funds, ETFs and closed-end funds -- have been among the fastest growing passively managed investment vehicles and now control over one trillion dollars in assets spread out among 1,000 vehicles. (primarily ETF funds.)
However, factor investing has several problems which has prevented it from living up to the promises of its advocates. First, there are an unwieldy proliferation of factors, and the lack of overarching framework to organize these factors. Second, factor investing has done a poor job of explaining and forecasting factor premiums.
A recent book, Popularity A Bridge between Classical and Behavioral Finance, by Roger Ibbotson, et.al., provides a framework that is a marked departure from the dominant CAPM risk-reward model for explaining and predicting factor-based asset pricing and managing investment portfolios. The Popularity framework brings together and organizes all the factors -- financial, behavioral and economic -- that influence the price of an asset. Popularity – investors’ rational and irrational attitude towards an asset - is the organizing principle which provides a usable model for explaining and predicting asset prices that incorporates all the characteristics of an asset from both classical risk-reward finance and behavioral finance.
The present article describes the basics of factor investing and the concept of factor premiums. It discusses the dominant CAPM model of factor pricing and variations of CAPM that have been developed over the years, primarily by increasing the number of factors included in the model. The following article Factor Analysis (2): describes a model for asset pricing based on the concept of popularity, which expands the classical CAPM model to incorporate an array of factors, resulting in improved explanatory and predictive.
Factors in this context are economic, financial and/or behavioral variables that attempt to explain and predict a fund’s performance by statistical analysis of historical data. The term is used in regression or factor analysis where independent variables (factors) are correlated to (and presumed to drive the performance of) the dependent variable (here the investment return). The general formula for the linear case is:
Y = a + b (factor 1) + c (factor 2) +d (factor 3) …
where Y is the return on a security, index or fund; a is the alpha; and b, c, and d, are the betas that show the degree to which each factor impacts the performance of Y. The larger the beta of a given factor the more it drives the funds’ performance and the larger the amount it is provided in asset allocation. The higher the t-statistic, the more statistically significant the factor is.
From Investment Alpha to Beta
One of the fallouts of the growth of factor investing is the decline in the use of investment alpha -- the return above a benchmark that is attributed to manager skill. Many factors that were introduced into factor investing models diminished the explanatory power of alpha, so that Alpha now measures the return of an asst after the influence of factors are removed, which often turns out to be small or even negative. With a shrinking alpha, hundreds of factor investing or smart beta funds have been launched with the goal of providing investors with increased portfolio diversification and/or improved risk adjusted returns.
(A word on names. Factor investing is the technical term that refers to the analysis of asset pricing using factors as explanatory variables. “Smart Beta,” on the other hand, is used to describe products that use the factor investing approach.)
Zoo of Factors
Over 400 factors have been put forth by academics and money managers, causing one researcher to describe a “Zoo of Factors.” Thousands of studies have attempted to tame the zoo by statistical analysis guided by economic theories based on economic and/or behavioral factors or using statistical techniques to identify correlated data which are then given a label (this approach is called “data mining.)
Several factors have become widely used by researchers and money managers in various combinations. Initially, the Capital Asset Pricing Model (CAPM) designated only one factor: market volatility or beta. The Fama-French three-factor model, the first popular approach to factor investing, included three factors: the major market (i.e., the S&P 500), value vs. growth, and size (market capitalization). They later added two additional factors (investment and profitability) to develop a 5-factor model; Andrew Lo introduced momentum as an important factor in explaining hedge fund and mutual fund returns. A study by Hansanhodzik & Lo, “Can Hedge Fund Returns be Replicated: The Linear Case,” identifies five factors that are correlated with hedge fund performance: equity markets; US Dollar; Credit spreads; bonds and Commodities. Cliff Assness, founder of AQR, identifies several factors as drivers of investment returns: value, momentum, profitability, low volatility, interest rate carry, and defensive company stock. S&P has developed factor indices for the following factors: Volatility, high beta, size, style, quality, momentum, enhanced value, dividends, capital expenditures, and multi-factor funds; and Invesco uses six factors in constructing investor portfolios: Value, Size, Momentum, Low Volatility, Quality and Dividend Yield.
What is “Smart Beta?”
The most common approach to asset allocation is based on market cap-weighted or equally weighted portfolios: an asset is included in a portfolio in proportion to its cap-weighted or equal weighted metric. "Smart Beta" (often called multi-factor or strategic beta) strategies use an alternative methodology for asset allocation based on the beta weights of the factors included in the statistical analysis. Factors with higher beta are allocated a proportionally higher share of a portfolio than factors with lower beta weights.
To understand asset pricing, it is necessary to understand the concept of factor premiums. Factors have different characteristics such as volatility, size, behavior during down markets, etc.
For any asset, these characteristics are popular or unpopular with investors. The less popular factors need to include a premium -- the expected excess return relative to an appropriate benchmark – to entice investors. Premium payoffs reward investors who are willing take on the risk or exposure to factors that other investors want to avoid or underweight. More generally, factors that are riskier or more unpopular than a benchmark provide investors with a risk or unpopularity premium in the form of a lower market price, which makes any gains a larger return in percentage terms than factors whose market price is high.
Most of the premiums analyzed to date have been “risk premiums,” which compensate investors for taking additional market or security risks. This is in line with the Capital Asset Pricing Model or CAPM theory for factor pricing, which states that a stock’s expected return should be proportional to how much it tends to swing with the ups and downs of the overall market. Those that swing more -- that have higher market risk, or “beta” -- should deliver larger returns over time to compensate investors for assuming this additional risk.
Financial economists have sought factor characteristics other than market or security risks that may provide a premium to investors. The payoff [in the form of a premium] is not just for risk but for anything investors find to be intrinsically unattractive, for example, less liquidity, high taxability, difficult to diversify, high search costs, bad management, distress, and so on. In fact, anything that investors eschew needs to provide a premium to entice investors.
This use of non-risk factors opens opportunities to expand the factors and premiums, modify the asset allocation process, and change portfolio management.
Two issues remain unresolved for asset pricing. The first is an absence of an overriding economic framework within which factors and factor premiums can nestle. Second, recognition that smart beta to date has not lived up to its promise.
My next article describes a comprehensive and innovative factor investing framework that holds the promise to help resolve or shed new light on many issues in asset pricing and portfolio management. The framework is presented in a new book by Roger Ibbotson, et.al, Popularity: A Bridge Between Classical and Behavioral Finance.