Sections

Smart Beta: A smart beta taxonomy

Related Categories

Daniel Leveau and Des Morris categorise the alternative indexing universe and recommend that investors build diversified, smart-beta portfolios. By focusing on risk and return, they should achieve the desired results

Market capitalisation-weighted (MCAP) indexing still represents the majority of assets invested in passive equity strategies. The superiority of this indexing approach is increasingly being called into question by academics and market practitioners alike.
Criticism has been aimed at the theoretical underpinning, as well as its implications on the index construction process.

In response, a new breed of indexing methods has emerged, variously described as ‘smart beta', ‘strategy indices', ‘alternative beta' - all of which attempt to reweigh index constituents according to different criteria and rationale, with the objective of offering a more attractive index than MCAP.

Despite being different, there are several similarities between many of them, enabling us to classify them into five distinct beta sources. What, then, is the rationale of these beta sources and what characterises them? And even more importantly, what is in it for investors?

Beta 1 - Price-focused
In price-focused indexing, the higher the price (or market capitalisation), the higher the index weight. This group includes MCAP indexing, by far the most prominent indexing method, which rests on the theoretical assumption that the market is efficiently priced.
The implication for investors is that they should not bother to make active investment decisions but rather invest in a portion of the total market.

On a practical level, applying this index methodology will invariably result in pro-cyclical investment behaviour, as overvalued constituents per construction will be overweighted, and vice versa.

MCAP is ‘the market' and - by convention - style-neutral. However, its embedded pro-cyclicality and construction rules could easily render a MCAP index both growth and large-cap based, with an unintended tilt against value.

This group also includes price indexing, where the stock price per share of an index constituent - rather than the total market capitalisation - determines the index weight. Constituents with a larger per share price will get a higher weighting, and vice versa.
While simple in construction, the major point of critique for this index method is the flawed assumption that the price per share level and movement of the index constituent is representative of the general stockmarket.

Despite price indexing's poor representation of the market, the Dow Jones Industrial Average (first published in the 1880s) is one of the most widely-followed indices.
Finally, we have diversity indexing. The diversity of a stockmarket is a measure of the distribution of capital within a stockmarket index: the higher the concentration in a few index constituents, the lower the diversity measure.

From a historical perspective, the lowest market diversity is generally observed during periods of stockmarket bubbles such as the TMT bubble. Diversity indexing aims to improve the diversity of the index with respect to MCAP and can be viewed as a blend of MCAP and equal-weighted indexing, and it addresses problems associated with both to a certain degree.

For instance, smaller index constituents will be given a higher weighting at the expense of MCAP index heavyweights, resulting in reduced pro-cyclicality, but it also reduces some of the excess turnover associated with equal-weighted indexing.

How close it resembles either indexing method is a function of the investor's ability and willingness to take on a certain level of expected tracking error.

Beta 2 - Price-agnostic
Price-agnostic indexing fully eliminates the relationship between price and weight, and gives each index constituent the same weight. Equal-weighted indexing finds its origin in a simple arithmetic weighting scheme - the naïve 1/N portfolio.

It can be seen as a contrarian approach, as the weights to overvalued stocks will significantly be reduced, and vice versa. Its simplicity is compelling at first glance - but it is associated with some notable drawbacks.

The frequent rebalancing needed to stay equally-weighted results in higher transaction costs and the illiquidity of the smallest constituents may cause significant implementation problems and capacity constraints.

Furthermore, the performance of equal-weighted indexing is highly universe-dependent and can often result in a small-cap bias.

Beta 3 - Fundamentals-focused
Fundamental-focused indexing uses the absolute size of the index constituents' fundamentals (for example, book value) to determine the index weights: the larger the absolute value of the fundamentals, the higher the index weight.

Accounting-based indexing (often described as fundamental indexation) seeks to de-link weights from prices - thereby rendering pricing errors insignificant - by weighting each index constituent according to its so-called ‘economic footprint'. This is generally defined by each constituent's absolute level of fundamentals such as book value, sales and/or dividends, which are independent of market prices and therefore contain significantly less noise.

Compared with MCAP this invariably results in a value bias, and higher weights will generally be given to smaller index constituents. The chosen rebalancing date can also have a major impact on performance.

This methodology only allows for a limited scope for deviations from MCAP indexing, as the absolute fundamental value of constituents is closely related to its market capitalisation.

Beta 4 - Risk-focused
Risk-focused indexing solely focuses on risk when determining index weights, and normally applies an optimisation procedure.

This group includes risk parity indexing, where index constituents are equally weighted according to a particular risk measure. No universal measure of risk exists, but often volatility is used to determine the index weights.

In general, higher risk index constituents will get a relatively lower weight compared with constituents exhibiting a lower risk, resulting in a somewhat risk-averse index that is expected to lead to reduced drawdowns and a lower absolute level of risk.

On the flip side, according to traditional financial theory, one would also expect the accompanying return to be lower. This approach is sensitive to the choice of risk variable and is also often associated with high turnover.

This group also includes the maximum diversification indexing approach, which is very closely-related to MCAP investing.

The crucial difference lies in how return expectations are formed. Rather than using the capital asset pricing model (systematic risk) as the basis for return estimation, the maximum diversification approach forms expected (excess) returns on the basis of total risk, systematic and idiosyncratic risk.

In general, index constituents exhibiting either a relatively low level of volatility or a low correlation with other constituents will get a higher weight. This methodology is sensitive to the risk-input variables, is quite technical in its construction, and rather difficult to grasp intuitively.

Analogous to maximum diversification indexing, efficient indexing seeks to improve the risk/return efficiency compared to a MCAP index by applying a weighting scheme based on traditional portfolio optimisation, but with an alternative method of estimating the expected return inputs.

Whereas maximum diversification indexing assumes a linear linkage between total risk and expected return, efficient indexing assumes a direct link between expected return and risk as measured by downside volatility, which better captures the negative aspects of market movements.

Efficient indexing will, in general, favour constituents exhibiting either a relatively low level of downside volatility or a low correlation. It will also suffer from the same problems associated with the input variables as the maximum diversification approach.

Minimum variance indexing completely avoids the difficulties involved with forming any return expectations and focuses solely on risk, or more specifically variance. Return expectations of all index constituents are assumed to be identical. Avoiding return expectations can be viewed as an advantage, as expected returns are notoriously hard to estimate.

This approach will overweight low-volatility constituents and those exhibiting a low relative correlation. Many cautious equity investors have embraced this indexing approach due to its reduced absolute level of risk compared to a MCAP index.

However, the reduced risk is expected to be associated with a lower return long-term. As with all optimisation-based procedures, this methodology is sensitive to the estimated input variables (volatility and co-variance) and is associated with high turnover and transaction costs.

Beta 5 - Return-focused
Return-focused indexing focuses on optimising the long-term return by dynamically altering index weights according to return expectations.

This method assesses each index constituent's long-term attractiveness (its return expectation) and weights constituents such that the higher the expected return, the higher the index weight.

The weights can be a direct function of a single or combination of specific return-predicting factors such as value (a price/earnings ratio, for example), momentum or ‘quality'.

In the case of value, the lower the valuation multiple, the higher the index weight, and vice versa; for momentum, the stronger the momentum, the higher the weight. Studies have shown that return-focused indexing tends to yield the highest return over the long run, but it can come with higher levels of volatility than other indexing methods.

Key benefit
What is the key take-away for investors? There is no single indexing method that will be superior across all market conditions. Why?

At its most simplistic level, all indexing methods seek to take exposure to one or more characteristics or effects that drive return and risk. Any methodology that focuses on a single characteristic is only likely to perform strongly when the market is rewarding that specific characteristic. It will lag in other periods. We also know that predicting when the market will reward each effect is very difficult.

The most logical move would be to diversify across these indices, offering investors a diversified exposure to various characteristics and effects. As the various indexing methods exhibit markedly different risk/return characteristics, a diversified exposure is expected to result in a more attractive and even more broadly diversified index investment that is expected to be beneficial for investors over the long run.

How should investors proceed? Buying directly into the various single alternative index providers is an option, but is associated with various implementation problems.  Alternatively, directly accessing a ‘best-of' index that focuses on the two most important variables for investors - namely risk and return - could well deliver the desired results.

Daniel Leveau is head of portfolio management and Des Morris is director of institutional clients at 1741 Asset Management

 

Have your say

You must sign in to make a comment