IMPLEMENTATION

Building portfolios based on views about a small number of individual securities is relatively straightforward. After ranking the universe of securities and deciding upon which ones to own, the portfolio manager simply buys shares of those securities. Part of the challenge associated with dynamic factor approaches to portfolio management rests in the implementation. Because the returns associated with individual securities contain an element of idiosyncratic risk, imperfect exposure to various factors is generally unavoidable. As such, the implementation of portfolio management strategies based on dynamic factor allocations is fraught with error. However, the problem of implementation error has been alleviated in recent years by several developments facilitating the implementation of factor-based strategies.

Physical Securities

The implementation of dynamic factor strategies with individual stocks requires a portfolio with a large number of holdings in order to achieve suitable factor exposure and avoid an excessive degree of security-specific risk. A portfolio of securities can be constructed in order to provide exposure to a specific set of factors. With a large enough number of holdings, the idiosyncratic return is minimized and the overall portfolio will reflect its factor content rather than the constituent securities. The securities are included in the portfolio in order to capture the returns associated with the characteristics they exhibit. Much like an insurance company manages risk by relying on a law of large numbers, a factor-based equity portfolio built with a large enough number of individual securities will perform in a manner driven the factor content rather than the individual stocks in the portfolio.

The decline in the cost of computing power over the past two decades has not only improved the modeling side of portfolio management, but it has also improved the ability to implement macro strategies through the use of portfolios of individual securities. First, access to inexpensive computing power is central to the modeling and construction of portfolios of large numbers of individual securities. Equally as important, technology has dramatically reduced trading costs and introduced efficiencies into the portfolio management process that allow for the implementation of strategies that were impractical in the past.

Trading in Aggregates

Some dynamic factor strategies may be implemented at the macro level through direct (or semidirect) exposure to a particular set of factors. Many index providers have created benchmarks designed to proxy one or more of the various factors discussed earlier in this chapter. For example, S&P, Russell, Wilshire, MSCI, and Morningstar all have developed indexes to proxy the performance of value stocks and growth stocks, as well as various segments of the capitalization spectrum. While these benchmarks are far from flawless (e.g., the definition of value in some indexes results in excessive turnover that has little to do with changes in equity style), they are useful for categorizing segments of the equity market, and the number of indexes appears to be mushrooming. Recently, in response to Arnott, Hsu, and Moore⁵³ and Siegel,⁵⁴ a host of new active or fundamental indexes have been developed, opening the door for an endless array of factor-based indexes that can be used to model, benchmark and implement dynamic factor approaches to portfolio management.

Investable versions of these indexes are increasingly available to implement style tilts and gain exposure to other factors. Futures contracts are available for various equity size and style exposures. However, the use of futures in equity style management has been plagued by a lack of liquidity in the underlying futures contracts.⁵⁵ Futures contracts are often used to implement global tactical asset allocation strategies, which trade country-level exposure to various equity and fixed income markets and currencies.

Swap contracts can also be used for various factor exposures. Indeed, because of the ability for customization, swap contracts are one of the most flexible tools available for implementing factor modeling strategies. Swaps are especially attractive for taxable investors, since gains on futures are generally realized frequently and are taxed unfavorably, while swaps allow deferment of capital gains through the term of the contract, as well as loss harvesting by terminating a contract prior to its intended settlement.

Exchange-traded funds (ETFs) represent a basket of securities that are traded on an exchange, and they are typically designed to replicate an index or benchmark. Along with the proliferation of indexes based on various factors and segments of the equity market, an increasing number of ETFs designed to replicate these indexes has emerged.

ETFs have provided portfolio managers the ability to gain exposure to a particular style, industry, sector, country, market capitalization segment or other factor for an indefinite period at extremely low cost in a single trade. As a result of their advantages over traditional mutual funds (e.g., low-costs, intraday trading and high tax efficiency) and their lack of the regulatory burdens associated with trading futures contracts, these funds have experienced explosive growth in number and total assets since their introduction. As the ETF market grows, it appears that an increasing number of factor-like exposures will become easily tradeable on equity exchanges. A comprehensive treatment of ETFs can be found in Gastineau.⁵⁶