Diversification is one of the central tenets of investment management and receives endorsement across the global financial planning industry.
Its validity was set in stone by Harry Markowitz in his PhD dissertation and 1952 Journal of Finance article “Portfolio Selection”, which demonstrated that combining uncorrelated assets could improve a portfolio’s return per unit risk.
The now-famous work showed diversification to be the closest thing to a Holy Grail of investing and possibly the only ‘free lunch’ (rebalancing may be the free dessert), as it was possible to improve the return expectation of a portfolio without necessarily increasing risk (or to maintain the expected return whilst decreasing risk).
Today, the concept of diversification may seem second nature to all of us in the investment industry but some of its fundamentals are often misused and misrepresented. Diversification is probably the most commonly used justification for investment recommendations and the word conveys a sense of lower risk, which is always appealing. Unfortunately, its meanings appear to have shifted in potentially misleading ways.
The purpose of this article is to revisit the foundations of diversification and, it’s hoped, address potential misconceptions or misunderstandings.
A simple example
It is not uncommon for investment advisers to recommend a portfolio as an improvement on an investor’s existing portfolio on the basis of greater diversification. A simple example may be the investor who has much of their wealth tied up in a single stock, maybe because of an employee share scheme or inheritance.
The adviser would be concerned about the concentration risk of this single stock and would recommend a sell-down or reduced exposure to the stock, to spread its risk across a portfolio of managed funds or perhaps a larger stock portfolio. The justification is that doing so doesn’t necessarily compromise return potential.
This may be quite a valid recommendation with a valid justification. Spreading the risk from one stock to many others is a simple example of diversification but there is a little more to this than meets the eye.
How to measure diversification
Modern Portfolio Theory (MPT) defines the completely diversified portfolio as the market portfolio. Without going into too much explanatory detail, because the market portfolio contains all assets, it cannot be diversified away, except by other markets. So increasing diversification is an exercise in shifting a portfolio to make it more market-like. In the simple example above, this was a shift from the specific risk of one stock to many more stocks.
This all means that to determine the level of diversification of a portfolio consistent with MPT, one has to measure it in the context of the market or market risk. Measures commonly used include active share, tracking error or systematic risk, as defined by the R-Squared of the Capital Asset Pricing Model (CAPM).
Using the CAPM R-Squared measure, an index portfolio is close to 100 per cent market risk, and active strategies will have variable market risk depending on how active and how diversified or concentrated they intend to be. An active strategy’s obvious goal is to ensure that non-market risk produces excess risk-adjusted returns (alpha), but it will always have less diversification than the market.
As examples, the following three charts show the market risk (blue) and non-market risk (green) through time (using Markowitz defined risk) for:
- an Australian equities index fund
- a popular actively managed Australian equities fund
- a popular actively managed small-cap Australian equities fund.
Source: Delta Research & Advisory
The index fund, as expected, is all blue and therefore all market risk; the active strategy is dominated by market risk but with a substantial proportion of non-market risk (sometimes called active or idiosyncratic risk), and the restricted small-cap strategy, expectedly, has an even higher proportion of green (or non-market risk), as it excludes the large-cap shares from the fund.
So when recommending changes based on diversification, it is possible to explicitly measure and demonstrate the changes in diversification using past performance risk measures.
Diversification misrepresented
Many may argue that investment recommendations are sometimes looking to diversify away specific risks, as opposed to non-market risks, which may not result in the portfolio becoming more like the market. A popular example is when an adviser recommends a small-cap Australian equities strategy to diversify away the large-cap bias of the Australian equities market, which is dominated by large banks and materials companies. On the surface, this justification appears reasonable, but there are some issues of which advisers need to be aware.
Firstly, this is not diversification, it is the opposite.
Considering the first step of portfolio construction is asset allocation, which is designed based on market expectations of asset classes (i.e., beta), a small-companies strategy restricts the portfolio and, therefore, increases concentration risks to small caps and away from the market (or beta) recommendation. This shift away from the asset allocation decision potentially increases risks of market-relative performance failure. Don’t forget, complete diversification contains all assets, which the restricted small-cap strategy cannot.
The decision to move away from the market, dominated by large caps, is an active decision that is probably based on the belief that small caps are likely to outperform large caps. So it is a decision designed to outperform the market and generate alpha risk (similar to tracking error) and, therefore, is not based on diversification. Diversification is, in fact, alpha risk minimisation. A portfolio that contains a single security is the simplest example of a massive alpha bet, whilst an index portfolio contains no alpha bet whatsoever.
Alternatives aren’t automatic diversification
For the last 10 to 15 years, alternatives have made appearances in more and more investment portfolios, often for reasons of diversification. Sometimes this is effective and sometimes not.
Alternatives can bring diversification benefits by accessing markets that do not exist elsewhere within a portfolio. This may be true of soft and hard commodities, private equity, and maybe unlisted infrastructure. This is because these asset classes are not represented in the traditional asset allocations of bonds and equities.
A market cannot be diversified away, except by a different market.
Alternatives do not necessarily increase diversification, though. They can also increase concentration and non-market risk. This happens within the various equity and bond strategies many hedge funds execute. This includes long-short, variable beta, and potentially other arbitrage or concentrated strategies. These strategies do not increase diversification because it is always possible they will have the same market exposure as an index fund. What this approach does is increase the concentration risks linked to the success or failure of the specific strategy bets. Like the small-cap example above, the inclusion of equity or bond alternatives is based on hopes of capturing manager skill (alpha) and ability to outperform a market, not on improving diversification and minimising non-market performance risk.
Over-diversification may not exist
Over-diversification is often mentioned within investment circles; however, in a cost-free world, it isn’t possible. Over-diversification occurs when the costs of adding securities or investments to a portfolio detract from performance potential. When costs are nil or very low, over-diversification is difficult or impossible.
As an example, a portfolio of many index funds for the same asset class will be detrimental to portfolio returns compared to one index fund only if there are flat fees charged per investment. There is no over-diversification because there is no non-market risk to diversify, and the return, irrespective of the number of funds, will be the market minus the average of the low management fees. There is rarely any value in having multiple index funds of the same market.
Over-diversification most frequently occurs when combining active managers of the same asset class. This is because the more active managers in a portfolio, the more they diversify away the portfolio’s non-market risk (because you can’t diversify away market risk), potentially leaving a portfolio that resembles an index fund but at active manager costs. Measuring this is possible using historical data, as already discussed and shown with Chart 1-3, but predicting the optimal number of strategies is difficult and the number can vary depending on the degree to which each strategy is active and correlated.
Final thoughts
It seems diversification is used to justify more than it should. It is a relative concept and is about reducing non-market risks and not increasing market outperformance potential.
There is no right or wrong level of diversification, as there are many schools of thought and examples with reasonable evidence as to what works in markets and what doesn’t. Warren Buffett is quoted as saying that “diversification is ignorance”, yet he says most should invest in index funds. Lower levels of diversification may be safest when investment skill exists, but finding true skill is difficult, sometimes expensive, and persistent skill is rare.
Investment recommendations are designed to reflect one’s philosophy. If you believe markets are efficient, you have defined the appropriate level of diversification, and will recommend market portfolios. If not, the portfolio construction question to be answered is, how much diversification is enough?
