2. The typical S&P 500 index fund owns all 500 stocks in the index but doesn’t invest an equal amount in each. Traditional index fund weights are determined by market capitalization. However Smart-beta indexes tilt toward value stocks that perform well over time.
3. Smart-beta ETFs are rules-based. The rules are established in advance. Using the PowerShares S&P 500 ETF (SPLV) as an example, the fund takes the 100 least volatile stocks in the broader S&P 500 index, and then every three months it rebalances by selling what no longer fits its rules-based criteria and buys what does fit.
EQUITY SMART BETA
1. The equity smart beta universe can be split into two high level categories: thematic and systematic.
2. Thematic Approach
(i) Thematic smart beta aims to take advantage of secular or temporal mispricing issues (for example, emerging wealth exposure, demographic impacts). The underlying thesis is essentially that many investors are excessively focused on the short term and, therefore, mispricing exists in respect of opportunities that are longer term in nature.
(ii) These strategies require beliefs to be held about what the future has in store, the appropriateness of such strategies is particularly specific to an investor’s existing portfolio, beliefs and preferences.
3. Systematic Approach
(i) Opportunities for systematic smart beta approaches arise from a range of factors; for example, investor heterogeneity and systematic mispricing (such as the value weighted index), and some structural issues associated with market capitalisation approaches (for example, constituent change dates, arbitrary rules for inclusion or exclusion).
(ii) The systematic categories can then be split into four specific definitions: equal weighted, economically weighted, risk weighted and factor tilts.
4. These approaches aim to tilt portfolios toward attributes that have been shown in academic literature to add value over time. Typically these are factors that are associated with active management ‘styles’.
(i) Value — ‘Cheap’ stocks (where the current market price is perceived to be less than the market average as measured by metrics such as Price/ Earnings and Price/Book) outperform ‘expensive’ stocks (those where the market price exceeds the market average) as the market readjusts towards fundamental valuations.
(ii) Momentum — Stocks that have performed well recently, say over the last 12 months, continue to perform well (and vice-versa).
(iii) Size — Stocks of smaller companies have higher returns.
EQUALLY WEIGHTED APPROACH
1. This is the most simplistic approach to removing the link between prices and stock weights. No assumptions are used, or needed, for stock returns, volatility or correlations. Implicit in this approach is that stock returns, volatility and correlations are unknown (or are at least expected to be the same on an ex-ante basis)
2. Although simple, equally weighted approaches have historically been hard to beat both within and across markets, at least before transaction costs. Certainly they have outperformed market capitalisation approaches.
3. The approach in its purest sense is, however, not practical, particularly for broad market portfolios. The approach gives a large allocation to small and illiquid stocks. This means that a portfolio cannot be implemented with any significant size, so the strategy has low capacity.
ECONOMICALLY WEIGHTED APPROACH
1. This smart beta approach weights stocks by fundamental business metrics such as sales, earnings, book value and so on, rather than market capitalisation. The main aim of this approach is to remove the link between the price and the weighting of stocks in a portfolio, and replace it with weightings based on economic size.
2. Relative to a market capitalisation portfolio, economically weighted portfolios tend to have a value ‘tilt’, although this varies over time. However, relative to typical value oriented indices, economically weighted strategies include all stocks in a given universe, rather than just the ‘value half’.
3. There are a number of ways of combining different economic measures to form an economically weighted strategy. In most cases, several measures are used, and the weighting from each is averaged.
4. The underlying philosophies of the various approaches do differ, and this can lead to some differences in the construction of the indices. However, we believe the effects they capture are broadly similar.
RISK WEIGHTED APPROACHES
1. These approaches aim to improve portfolio efficiency (return per unit risk) by making assumptions about future volatility, return and/or correlations, generally based on historical observations.
2. There are three main risk-weighted approaches:
(i) Naïve volatility-based weighting — The weight of the stock in the portfolio is related to the inverse of its variance. No assumption is made about return or correlation and, in a similar manner to equal weighting, the implication is these are unknown.
(ii) Diversification strategies — These strategies assume that correlation is mispriced and that by constructing a portfolio to have low correlations (a high level of diversification), superior risk-adjusted returns can be generated.
(iii) Volatility minimisation strategies — These strategies combine volatility and correlation through an optimisation framework to produce a portfolio with the lowest volatility or beta, depending on the strategy.
POPULAR SMART BETA STRATEGIES
1. Minimum variance: To construct the minimum variance strategy we use the method of Clarke, de Silva and Thorley (2006).
2 Maximum diversification: Portfolio optimised to maximise expected diversification ratio, which is defined as the ratio of weighted average risk to the expected portfolio risk. For details see Choueifaty and Coignard (2008).
3 Risk-efficient (λ=2): Mean-variance optimised portfolio assuming that expected excess returns are proportional to the stocks’ downside semi-deviation, and with stringent constraint to limit portfolio concentration. For details see Amenc et al (2010).
4 Risk cluster equal weight: Applying statistical methods to identify major market risk factors, assumed to be driven by industries and geographies, and then equally weight these uncorrelated risk clusters.
5 Diversity weighting: Weighted based on the market capitalisation weight raised to the power of a constant that is between zero and one to tilt the portfolio towards small cap stocks while limiting tracking error. We used the value of 0.76 in our simulation.
6 Fundamentals weighted: Weighted based on the five year averages of cash flows, dividends, sales and the most recent book value of equity. We introduce a two year delay to avoid forward-looking bias. Following the original method, we select top stocks with the largest fundamental weight. For details see Arnott, Hsu and Moore (2005).
7 Cap-weighted: Weighted based on market capitalisation. The market capitalisation is computed using December values at the close of the year prior to index construction.
SMART BETA VS FIXED INTEREST
1. Traditional fixed interest indices, which weight issuances according to outstanding debt, have a fundamental flaw in our opinion — the most indebted issuers constitute the greatest portion of the index. Government bond benchmarks, in particular, have high risk concentrations in the US and Japan.
2. Any approach which breaks the reliance on issuer size in a sensible manner represents a significant improvement compared to the traditional indices.
3. An example of this style of approach (for global sovereign debt) is to establish two simple caps on exposures which significantly alter the characteristics of the index — such as a cap on individual countries, and a collective cap on Eurozone countries.
4. In the current environment of record-low sovereign bond yields globally, this style of smart beta strategy allows investors to maintain global sovereign bond exposure without assuming the same level of exposure to the US, Japan and heavily indebted Eurozone countries.
5. Other alternative methods of weighting fixed interest securities in the government bond space could be to weight according to issuer GDP or fiscal sustainability (based on economic fundamentals)
SMART BETA FOR COMMODITY FUTURES
1. Commodities provide diversification across the business cycle, with prices being highest at the end of the cycle when capacity is lowest, whereas equity returns tend to peak earlier in the cycle. Additionally, commodities provide reasonably strong inflationlinkages, being part of the consumer basket in most cases.
2. Passive investment in commodities has a number of vulnerabilities. We believe that many of these weaknesses can be improved by systematic tweaks to the investment process. For example, the common commodity futures indices (such as S&P GSCI) are based solely upon the front future contract, which doesn’t seem to fully represent the entire commodity sector anymore and is open to being front-run by active investors.
3. Investing in longer dated contracts or implementing continuous/daily rolling should theoretically avoid the crowding of passive investors in the shorter dated future contracts, and should also make better use of market liquidity.
SMART BETA FOR INFRASTRUCTURE
1. Unlisted, illiquid assets which provide the following desirable characteristics:
(i) Long duration assets
(ii) Strong cash flows that are often regulated, contracted or otherwise protected
(iii) Inflation linkages
(iv) Low correlation with traditional equity markets.
2. Due to the illiquid nature of unlisted assets, investors often attempt to obtain exposure to these characteristics through listed vehicles. Historically, these vehicles have had an unacceptably high correlation to listed equities. Another problem with investment in listed infrastructure markets specifically is the typically high exposure to energy generation assets (which have relatively high volatility of earnings).
3. An investment strategy which specifically targets stable, ‘core’, developed assets with predictable earnings throughout the economic cycle and relatively low levels of gearing, mitigates many of these issues while providing access to the desirable characteristics of infrastructure and property assets due to their lower correlation with equity markets than their traditional counterparts, given their lower exposure to assets with high leverage and development risk.
SMART BETA FOR FX
1. Currency carry strategies invest in high yielding currencies using capital from low yielding currencies. The attraction of this strategy is that the real interest rate differential rewards the willing investor over time.
2. One way to mitigate the risks of adverse exchange rate movements is to diversify across countries. For example, with 10 liquid developed market currencies, there are 45 currency pairs an investor could potentially take positions in.
3. A simple smart beta approach would weight each of these pairs equally, with the higher interest rate currency being bought against the currency with lower interest rates using forward contracts. The rationale for such an approach is that higher interest rate currencies tend to outperform lower interest rate currencies over the long term.
SMART BETA FOR EMERGING MARKET CURRENCIES
1. Emerging market currency strategies invest in cash and short-dated bonds denominated in emerging market currencies. This provides the investor with an unhedged exposure to emerging market currencies. Emerging market currencies will outperform developed market currencies over the longer term and offer protection against low growth in the developed world due to:
(i) Higher productivity growth in emerging markets should lead to currency appreciation over the long term.
(ii) Many emerging market central banks have historically suppressed the value of their countries’ exchange rates, however trade imbalances are now stressing this process.
(iii) There is a heightened risk of devaluation or nonreturn of assets in emerging market countries, for which a premium typically has to be offered to investors.
2. A semi-passive approach be taken to this asset class, with a largely buy-andhold approach taken to emerging market currencies subject to the manager’s view on country-specific risks and which countries might relax their currency suppression sooner.
SMART BETA FOR REINSURANCE
1. Catastrophe bonds, which usually have a term of three years, have the same pay-off as a traditional bond, except in the event of the insured risk occurring, when some or the entire principal is used to cover the reinsurance claim. This style of investment gives direct access to a number of different catastrophe-related risks.
2. It is important to adopt a reasonable time horizon for reinsurance investments. Like any type of insurance, a premium is collected to reward the insurer over the longer term for protecting the insured against certain events. Therefore, although a loss will assuredly be incurred at some point, over the longer term (three to five years) a reinsurer with a well-diversified portfolio is expected to have made a profit in aggregate.
THOUGHTS
1. Advantages of smart beta as an alternative to active management are:
(i) Capacity is higher than for most actively managed strategies (lower turnover and less concentrated).
(ii) Higher levels of portfolio redundancy in traditional active multi-manager configurations can lead to a convergence towards capitalisation weights with some residual style tilts that can be easily and cost effectively replicated using smart beta strategies.
2. Smart beta strategies have high tracking error relative to a market capitalisation portfolio. This means that the strategies will perform very differently to market capitalisation weighted portfolios and can underperform market capitalisation substantially, potentially over a long period of time — typically during strong market rallies and ‘bubbles’. Investors should therefore have a longterm approach to the investment — we would suggest a timeframe of at least five to ten years.
3. Beta Strategies should be benchmarked against a suitable combination of a market capitalisation index and cash (which would be another way of lowering the beta of the equity portfolio) over the medium term.
4. Risk and return characteristics of a range of smart beta strategies from February 2002 to December 2012
(Source: ThinkAdvisor, Towers-Watson)
