Improving Smart Beta strategies

By Roger McIntosh

Smart beta investing, or rules-driven active investing, has gained widespread popularity and usage, particularly amongst investors looking for alternatives to high cost managers.  Almost 50% of institutions surveyed by Russell Indexes with assets over USD 1Billion are looking to increase their usage of smart beta strategies (1).  There is now over US400billion in smart beta exchange traded products offered in the United States (2).

There have been significant benefits for investors with the development of smart beta products.  Typically they provide low cost access to an active strategy the investor expects will produce superior portfolio performance over time.  They are also a useful yardstick to measure and analyse the outcomes of investment manager returns that identify with a certain style factor exposure.  Why pay high fees for an active strategy that claims outperformance from the broader market when a smart beta index built upon one or more different style factors that can be described by a set of rules can do a similar job at lower cost?

There are two concepts to consider in this discussion: the underlying style factor and the smart beta index set of rules that aims to describe it (3).  Finance practitioners and academics have long been able to identify and describe a number of style factors that outperform the broad equity market over time, such as value, momentum, quality, and low volatility.  Many of these factors have found their way into smart beta products.

As the style factor return can’t be directly measured it has to be estimated through standard quantitative techniques.  Smart beta strategies take a simplification short cut and define a set of rules to describe the style factor for ease of explanation in marketing material or a sales pitch.  There is rarely though one set of rules that can best describe a style factor.  For instance, several index vendors have multiple definitions of a value index, each with a different approach to capture this style factor.  The other challenge with describing a style factor portfolio by a set of rules is that it is difficult to ensure that is independent of other factor effects, or that at least these other factor interactions can be controlled.

The main problem with smart beta indexes is that they often don’t do a great job of capturing the preferred style factor premia.  At best a rules based approach can provide a portfolio that may appear to have high exposure to a preferred characteristic (say high book/price), but if it is applied with a simplistic weighting scheme it ends up only really being a tilt towards the underlying style factor.

What is not considered, and often disregarded, in the clamour to build the latest smart beta strategy is the evaluation of risk and investor risk preferences.  That is, how to best select securities and determine their portfolio weight in the context of trading off each security’s expected return (which should be related to the exposure to style factor: higher exposure, higher preference) with expected risk and also the investors overall risk preferences.

For instance if two securities in the same industry or country have a similar exposure to a desired factor, but one is more volatile than the other, how should you allocate weight to each them?  The simplistic approach adopted by many smart beta products doesn’t consider this and allocates either on the basis of equal weight or market capitalisation.

These simple weighting approaches cannot consider that the riskier security effectively has greater uncertainty associated with its style factor exposure and hence should have a higher expected return as compensation for choosing to own it as part of the portfolio.  The optimal portfolio maximises the risk-adjusted exposure to the style factor and controls for other risks.

This principle is deeply embedded in modern portfolio theory and also the concept described by Grinold and Kahn as the ‘fundamental law of active management’ (4).  Failing to control risk introduces unintended bets, produces inefficient portfolios and can result in portfolio performance being driven by exposures to market, industry, country and other risk index factors.

Index vendors typically include large numbers of securities in many smart beta strategies, presumably to provide greater diversification.  Holding many securities may be helpful to determine in the broadest sense the opportunity set of the style factor, but often ends up capturing only a fraction of the style factor premia.

Smart beta index construction rules can also impose other limitations.  By rebalancing on a calendar cycle there can be lags in the information embedded in the index construction and outdated information compared to the market, that a continuously evolving investment process can better appraise and respond to in a timely manner.

The important point to remember when evaluating and deciding upon any of these smart beta strategies is that the investment manager is the fiduciary charged with the management of your assets, not an index vendor.

The best of both active portfolio management and smart beta rules combines sound stock selection and sophisticated, risk-controlled portfolio construction techniques to determine security holdings and weight.  This approach more efficiently provides the style factor premia without needlessly holding securities that have a partial or tenuous association with that type of style factor return.

There are significant benefits of this risk-controlled approach compared to implementing a smart beta index set of rules:

  • Better portfolio construction: security selection and weight allocation focuses on maximising risk-adjusted returns from the desired style factor exposures in the most efficient manner.
  • Risk control: greater control of risk preferences and factor exposures avoids performance discrepancy arising from unintended bets.
  • Cost control: better turnover and portfolio cost control is achieved from holding fewer securities and managing trading activity.

 

Factor based investing provides some advantages in the pursuit of active investment performance.  As an investor there is better control around the desired type of style factor risk premium, which market sectors to target, and greater accountability of sources of portfolio return for the manager.  We believe smart beta provides a better way to measure active manager performance, but really is the starting point to better portfolio performance.

If the philosophy of the manager can be captured by the use of smart beta then any active fee should only be paid for improving on the risk return profile of the smart beta, not the cap-weighted index. We completely embrace the use of smart beta and have been involved with the development of style factor strategies for many years.

Building and implementing a strategy based around capturing style factor exposures shouldn’t mean outsourcing the investment process through replicating the rules of a smart beta strategy from an index vendor.  In fact an active approach to factor investing is the best way for investors to successfully take advantage of this market evolution.

A disciplined investment approach evaluates the ability of all securities within an investment universe to deliver the style factor exposure in the context of their contribution to portfolio risk.  Adopting quantitative processes and qualitative judgement to determine weights in a risk aware context provides an efficient portfolio that appropriately reflects investor risk preferences.  Combining this portfolio construction approach with a thoughtful trading process gives a powerful way to efficiently obtain exposures to preferred style factors with superior long term risk and return outcomes compared to an equivalent smart beta index.

References:

(1) Russell Indexes Global Smart Beta Survey April 2014

(2) www.investmentnews.com 1 Feb 2015 ‘Smart beta booming despite a name everyone hates’

(3) ”Smart Portfolios”, Jason MacQueen, Northfield Information Systems, November 2014

(4) ”Active Portfolio Management”, Grinold & Kahn, 2nd edition, McGraw-Hill Education, 1999

Related Posts