Modern Portfolio Theory has been a cornerstone of our industry ever since Harry Markowitz published his Nobel Prize winning work in the 1950s. However, its core principle of “portfolio diversification” is still poorly understood.
The idea that a collection of investments that display high levels of risk individually can be low risk overall is intuitively a difficult notion to understand.
While most people begrudgingly accept this concept at the highest level when considering asset allocation, they revert to an individualist view when choosing funds. This is depriving investors of the full benefits of diversification that were identified more than 60 years ago.
Diversification harnesses the differences between investments and offsets their individual fluctuations against each other to smooth out the effect felt by the investor. For this approach to produce a noticeable impact, it is necessary to combine investments with as low a correlation to each other as possible.
Consider two possible investments; one in the UK stockmarket, represented by the FTSE All Share index, and one in UK government bonds, represented by the FTSE Actuaries UK Government All Stocks index.
Individually one is considered high-risk, high-reward and one is considered low-risk, low-reward. By combining these two investments it is possible to receive the returns from both but for much less risk.
This can be seen in the chart; although there has been little benefit of investing in equities over the past 15 years, the time period selected demonstrates the concept quite accurately.
The approach is understood at the asset allocation level. Asset classes such as equities, bonds, property and cash are chosen for the low correlation to each other.
For a desired level of risk, different mixes of these asset classes are selected to offer the best possible expected returns. This is where diversification usually ends.
Many advisers implement their strategic asset allocation by picking funds for each asset class. When assessing suitable funds for their portfolio, they usually look at them in isolation.
Factors such as performance, volatility, the quality of the team, the track record of the manager are the mainstays.
Even professional analysts who strive to delve deeper into the viability of a strategy, or fund ratings are still treating funds as separate entities and ignoring their interaction with each other.
This is fine for identifying good funds, which is obviously an important part of portfolio construction, but does not guarantee the best portfolio. Even once you have narrowed down your choices to the best funds in each asset class, you are still faced with a choice of at least 100 funds and, for a 10-fund portfolio, that amounts to 17.3 trillion possible portfolio combinations.
Instead of looking at each fund individually, investors would reap vast benefits if they further leveraged the principle of diversification by studying how funds interact with each other
This number is greatly reduced once asset-class restrictions are applied, but it demonstrates the unseen problem in portfolio construction.
For every asset class, it is possible to identify three or four outstanding fund choices; typically at this point the most minor of differences becomes the deciding factor, with personal preference playing no small part in the final decision.
On the face of it, this looks like a pretty good outcome. A portfolio containing a diverse mix of asset classes, populated with the best funds available sounds like a win in anyone’s book; but by ignoring the diversification effect between funds, huge improvements in the risk-reward profile have been forgone.
Instead of looking at each fund individually, investors would reap vast benefits if they further leveraged the principle of diversification by studying how funds interact with each other.
Active funds do not mirror their underlying asset classes exactly. Differences in strategy, style, economic outlook, and even the moods of the fund managers, all combine to ensure that no two funds behave alike, and this can be used to our advantage.
By analysing the relationship between funds, identifying where these differences offer a significant advantage, we are able to further diversify and lower the overall risk of the portfolio.
Imagine a moderately cautious investor who had identified the need for a 20% exposure to UK equities. A likely outcome would be a cautiously minded, defensive fund, possibly in the UK Equity Income sector that offered access to the asset class but fitted the risk profile of the investor. But a better solution exists.
Within the spectrum of UK equity funds available that meet the required standard for excellence, there will be a superior solution.
Assume there are two alternatives: a fund with an aggressively minded fund manager, with a bullish outlook on the economy and who believes in the growth prospects of the technology sector; and a fund that relies on pure stock picking to identify undervalued companies and has a bias towards only the most stable and financially secure firms.
These two funds invest in the same asset class but will behave differently, the conditions for one to excel are different to the optimal conditions for the other. The same principles of diversification apply; the combined returns are available for much less risk than average.
The tables show just how much diversification it is possible to add through fund selection.
The first table shows the correlation between four main sectors, UK All Companies, Global, UK Gilt and Property. These represent the core asset classes most investors hold in their portfolios, and are considered to be diverse. The correlation statistics say otherwise however, with only property showing any significant behavioural differences.
The second table shows how dramatically the situation can be improved by selecting active funds and considering the relationships between them.
An equally weighted portfolio consisting of the four sector averages would have had a volatility of 9.26% over the past three years, whereas the comparable fund portfolio would have had a volatility of 8.44% over the same period, despite the individual constituents being much more volatile individually. The fund portfolio would have outperformed the other by 19.5% as well.
If we expand this principle across all asset classes, it is possible to use the active element of funds to add further diversification benefits and obtain the same asset class exposure for a much lower level of risk than expected. Alternatively, the same level of risk can be achieved while having greater exposure to risky, higher returning assets.
An example of how this works in the real world is seen above: a portfolio following the same asset allocation strategy using active funds is able to have a much lower risk profile than the same strategy implemented with index tracking funds, over a six month period.
This seems quite obvious, and most advisers will no doubt try to get a broad range of styles and characteristics when picking funds for a portfolio.
The problem, though, is how to objectively assess which funds offer the best mix and interaction. While it is possible to understand how two funds react quite easily, and even three funds, the multitude of interrelationships within a larger portfolio are far too many and complex.
The only way to truly maximise the diversification benefits within funds is to try to analyse every possible combination, then select the mix that offers the greatest reduction in portfolio volatility relative to the volatility of the individual funds. This requires some considerable number crunching.
Accessing the risk profiles of 17 trillion possible portfolios is a task that is beyond even the most avid Excel user, and is the main reason diversification has traditionally stopped at the asset class level.
Selecting the optimal portfolio from the near infinite range available is like looking for a needle in a haystack
To complete the task successfully investors would need a database of fund information and past performance; the expertise to calculate and interpret volatility; the ability to create, store and analyse all the portfolios and the computing power to prevent this from taking several lifetimes.
Selecting the optimal portfolio from the near infinite range available is like looking for a needle in a haystack. However, by modelling the characteristics of each portfolio, it is possible to create an efficient frontier of portfolios that contain the optimal combination of funds at every level of risk.
Developments by leading academics and fund research house FE have made great strides in this area. Having the required technological infrastructure and quantity of fund data at their disposal, they have managed to perfect this process.
Diversification has been the driving force enhancing investors’ returns for more than a generation and now, thanks to recent breakthroughs, is poised to offer more benefit than ever before.
For too long funds have been viewed in isolation, merely as vehicles providing access to an asset class. The importance of viewing a portfolio as a whole, something that has been understood for more than 60 years, is finally possible for the majority of investors – not just a few privileged institutions.