Smart beta

Smart beta is a marketing term used to promote investment products such as exchange-traded funds (ETFs) employing a mild type of active management. More specifically, smart beta funds use a rules-based system for selecting investments. Another way to state this is that smart beta is “active investing with an algorithm”.

This differs from traditional, passively managed ETFs that follow market-cap weighted indices, which are implied to be “dumb beta”. Smart beta funds can invest in equities or fixed income, depending on their mandates.

Depending on the definitions, smart beta or ‘strategic-beta’ strategies can include fundamentally weighted indices, equal weighted indexes, value, growth and dividend-oriented funds, and low-volatility strategies. At the end of 2019, after several years of strong inflows, 'factor' ETFs had a 12% market share among equity ETFs in Canada, while "dividend/income" ETFs had a 13% share, compared with 59% for cap-weighted ETFs. Globally, assets in smart beta products were more than $1trn in 2017.

The original beta
The term beta comes from the Capital Asset Pricing Model (CAPM) developed by Sharpe and others in the 1960s. Beta can be thought of as related to market (systematic) risk and portfolio volatility : the stock market as a whole has a beta value of 1.0. A portfolio with a 50% greater market risk will have a beta of 1.5: CAPM predicts that it will be more volatile, but achieve higher returns.

The CAPM equation has another important term, alpha, that represents excess return not explained by beta. This ‘alpha’ can be interpreted as manager “skill”, or manager luck, and should not exist persistently if the market is efficient. By investing in a cap-weighted index fund, you are accepting the market returns (minus the small MER) and are accepting that alpha will be zero: you will not, by definition, beat the index. In contrast, traditional active management (stock picking) aims to generate positive alpha (beat the market, hopefully at equal risk) but typically fails to do so after costs, especially over decadal timeframes, due to the "arithmetic of active management".

New betas
But that is not the end of story. Since the 1970s and 1980s, academic researchers have begun to identify other factors that can contribute to explain investment returns: CAPM was incomplete. By the early 1990s, that research had shown that in theory, “high, risk-adjusted returns could be generated by simple rules-based investing in: low volatility stocks; small cap stocks; stocks with low PEs (price to earning ratios); stocks with high book-to-market value (i.e., low P/B); stocks with high price momentum; and high dividend yield stocks”.

A milestone was the Fama & French three factor model of 1992. This included market beta, a value factor, and a size factor: there were now three betas instead of one. The three factor model became the standard for portfolio analysis. In this type of analysis, true 'alpha' is the over-performance that can't be explained by exposure to the three betas (market, value, size). The three factor model was upgraded to a five factor model in 2014, with the addition of ‘profitability’ and ‘investment level’.

The factor zoo
In recent years, according to Rob Arnott, "several hundred factors have been published". Arnott calls this the factor zoo and notes that "no one will bother to publish a factor or a strategy that fails to add value historically; this encourages data mining and selection bias".

All of these new factors are often referred to “alternative betas” in the academic community, but ‘smart’ beta sells better, and is more media-friendly. The term "smart beta" was coined by consulting firm Towers Watson. William Sharpe says that smart beta makes him “definitionally sick”.

Dividend investing
Dividend-focused strategies predate the smart beta label. However dividends can be seen as a factor among others, and dividend-focused ETFs can be included in the smart beta label, for example by BMO and Ishares Canada.

Factor investing
Factor investing also predates the smart beta label. It is an investment approach that involves targeting quantifiable firm characteristics or “factors” that can explain differences in stock returns. A factor-based investment strategy involves tilting equity portfolios towards and away from specific factors in an attempt to generate long-term investment returns in excess of benchmarks.

Well known factors include company size (small cap versus large cap), value (value versus growth), momentum, low volatility, quality, etc.

Factor investing can target a single factor, or several factors simultaneously, in which case it is known as “multi-factor”. A classic multi-factor approach combines value and size and is described in Multifactor investing - a comprehensive tutorial. Other combinations are becoming popular.

Most "smart beta" funds target a single factor for a single asset class, for example low-volatility Canadian stocks, or small US stocks, and use long-only approaches, without leverage. The factor exposure can be obtained explicitly by tilting the portfolio towards, or restricting it to, stocks that display the desired factor, or it can be obtained indirectly through alternative indexing methodologies such as equal weight or other non-price measures.

Why these funds exist
Some of the smart beta funds may be useful for some investors. It is possible that some of these strategies will outperform after fees, or achieve the same returns with lower risk. But the smart beta marketing can be cynically explained as follows: investors, and the investment industry, are always looking for a better mousetrap.

The original ETFs were passive, based on cap-weighted indices, like the original index funds. However, economies of scale and competition between providers has lowered the management fees on passive ETFs to near zero. There is no space in the ETF market for many new funds based on the TSX Composite index or the S&P500: the market is already saturated.

New entrants in the ETF space must therefore innovate to carve out a market share, and existing large providers are also looking to launch new funds, preferably with higher management fees. There must be a promise of higher returns, or lower risk, to entice investors to pay these higher fees.

Potential issues
Potential issues with smart beta strategies include:
 * the factors are back-tested, but some factors may not deliver in the future
 * constructing factor portfolios in academic research can be different from how smart beta ETFs work or perform
 * additional costs (higher MERs, higher turnover, etc.) may overwhelm any actual out-performance relative to a broad market cap-weighted index
 * performance chasing: "investors often choose these strategies, as they previously chose their active managers, based on recent performance"
 * factors may under-perform over long periods, which could lead to behavioral issues (not being patient enough to actually capture the premiums)

A study of a sample of US-listed smart beta ETFs representing 80% of assets under management in the category showed that "the claimed stellar performance of smart beta indexes only exists in backtests before ETFs listing, and smart beta index performance deteriorates significantly after ETF listing", yielding negative alphas for investors, even before management costs are deduced. The authors explore a number for hypotheses but conclude that this is due to data mining, i.e. "data overexploitation in backtests".