Chapter 15

Strategic Asset Allocation

Aligning your portfolio with your investment objectives

‘Have a strategic asset allocation mix that assumes that you don’t know what the future is going to hold.’

Ray Dalio

Asset allocation is probably your most important investment decision. It determines somewhere between 50% and 90% of your portfolio’s return and risk. It is a critical decision, deserving attention and time.

For example, if your investment strategy is capital preservation, you might allocate 20% to equities, 60% to bonds and 20% to cash. This allocation should deliver modest returns with a relative low risk level.

However, if bond yields are low and expected to rise, bonds may suffer capital losses. In a low-yield environment, cash may deliver meagre returns, perhaps lagging inflation. This allocation might fail preserving capital in such settings. It must adapt to current market conditions.

If your strategy is income and growth, you might allocate 50% to equities, 40% to bonds and 10% to cash. Equities aim to deliver long-term growth, as well as income (dividends). Bonds aim to deliver income, as well as to diversify equity risk.

Whilst this allocation appears balanced, most of its risk comes from equities. In fact, well over 80% of its risk can come from the 50% equity allocation.

If your strategy is aggressive growth, you might allocate 80% to equities, 15% to bonds and 5% to cash. Here, equities should deliver high growth over the long term. The asset allocation is risky with high expected volatility. The role of bonds is mainly diversifying equity risk.

If you do not mind volatility and your time horizon is long, consider investing only in equities. However, you will miss out on an opportunity to reduce some risk through diversification without giving up the same amount of return. In other words, your strategy will not be optimal.

Asset allocation fulfils several roles:

• It allows for formulating an investment strategy that should deliver a level of return and risk in line with your objectives. Once you understand the characteristics of assets and project rational assumptions about their future expected returns, risks and correlations, you can blend them, aligning the mix with your objectives.
• It diversifies your portfolio across different assets to reduce risk. Correlations across different asset classes are normally lower than correlations amongst securities within an asset class. Therefore, multi-asset investing comes with diversification benefits.
• It expands the investment opportunity set, methodologically including a wide range of asset classes in your portfolio. It is a structured way to blend investments.

Asset allocation can be divided into a long-term, static allocation, often called Strategic Asset Allocation, and a short to medium term, dynamic asset allocation, often called Tactical Asset Allocation. Combine the two since what can spoil a long-term plan is short-term volatility.

Strategic Asset Allocation

Strategic Asset Allocation (SAA) commonly forms a portfolio’s long-term, anchor allocation. SAA does not take into account current market views, but rather allocates capital to different assets, based on their long-term characteristics.

SAA is often fixed. You choose your SAA, you allocate your portfolio accordingly and you do not change the allocation. However, forgetting about your allocation is a mistake. Markets change and with them assets’ characteristics.

For example, there is a big difference in the expected future behaviour of 10-year gilts when their yield is 2% or 5%. Yield of 2% means low carry and expected capital losses, whilst 5% yield means decent carry and potential capital gains. Whether the base rate is 0.5% or 4% has a huge impact on what cash will deliver.

Making the same SAA decisions regardless of current market conditions is naïve. Review SAA regularly, at least every year or two. But, unless circumstances have indeed changed, you do not need to modify it. Keep your SAA dynamic and relevant, not static and irrelevant.

Optimisation

Once we have formulated the expected returns, risks and correlations of asset classes (capital market assumptions – CMAs), the next step is blending them together into an asset allocation. Optimisation is an iterative quantitative process of generating an asset allocation with the highest expected return for a certain risk level or the lowest risk for a certain return level.

Maximising the ratio between return and risk (Sharpe ratio) means you get the most you can out of your portfolio, given your risk tolerance. An optimal or efficient allocation does just that.

Investment professionals use a computer program based on a mathematical algorithm to produce an optimised allocation. The inputs are expected returns, risks and correlations for asset classes in the portfolio’s universe.¹

The most common optimisation technique is mean-variance optimisation. The simplest computer program tries different allocations until it cannot find an allocation with a better return/risk ratio than the last one. Then it stops and the efficient allocation is identified.²

Commonly, optimisation tools plot a chart, called the efficient frontier, showing a curved line with an efficient allocation for each risk level. The allocations on the frontier offer the maximum expected return available for each risk level. You do not need to construct a frontier. The explanation here is just a jargon booster.

Since we are seeking simple, pragmatic solutions to help you manage your portfolio, you can do it yourself. Instead of sophisticated optimisation, try different mixes of a small number of up to six asset classes by calculating the expected return and risk of each mix. The asset allocation may not be optimal, but it should be close enough, matching your investment objectives.

Anyway, many optimisations are flawed since they are based on historic relationships that may not hold in the future and on numerous assumptions that may turn out incorrect – garbage in garbage out (GIGO).

Asset allocation models

Table 15.1 includes five asset allocation models for different investment objectives. Use these models as guidance for a starting point.

Table 15.2 shows the assumptions used to calculate the expected returns and risks for the allocation models. Update the assumptions using the methodology that we have covered earlier to calculate the current expected returns and risks of the allocation models.

Expected returns do not include any potential alpha from active management or an allowance for fees and costs. By adding alpha, as well as diversifying across additional asset classes, you can achieve potential higher returns with lower risk. However, being conservative about assuming alpha is prudent, in particular if most of your portfolio is invested in index trackers.

Table 15.1 Asset allocation models

Source: Bloomberg. FTSE 100 Index, iBoxx £ Gilts Index, UK Cash LIBOR TR 1 Month Index, UK IPD TR All Property Index. January 1998 to December 2015. Based on monthly total returns, measured in £

Table 15.2 Assumptions

Source: Bloomberg. FTSE 100 Index, iBoxx £ Gilts Index, UK Cash LIBOR TR 1 Month Index, UK IPD TR All Property Index. January 1998 to December 2015. Based on monthly total returns, measured in £

Current expected returns are low because of the low-yield environment. When base rate is higher as well as gilts’ yields, the expected returns of the asset allocation models would be higher.

For example, assuming a 1.5% base rate and 3% 10-year gilt yield, the expected return of capital preservation increases by 0.7% from 2.9% to 3.6% (allocation to bonds and cash is large) and that of aggressive growth increases modestly from 4.9% to 5.0% (allocation to bonds and cash is small).

Achieving a 6% total return per year is challenging in current market conditions of not cheap equity markets and low yields. An all equity portfolio, including emerging market equity and assuming alpha, can achieve it. However, risk level would be high. A 6% return is more achievable when equity markets are cheaper and bonds’ yields are higher.

Figure 15.1 plots the allocation models on the risk/return plane using their forward-looking returns and risks. The risk/return space is commonly used to illustrate how different assets and allocations compare with each other.

Figure 15.2 shows the backtested cumulative performance of capital preservation and aggressive growth, against an annual 5% investment objective.⁹

This figure illustrates how the two models would have comfortably surpassed the objective. It also shows that during the particular backtesting period, an aggressive strategy only marginally outperformed a conservative one, but with a much higher risk. During the majority of the back-testing period, the aggressive strategy lagged the conservative one.

Be vigilant extrapolating the past into the future. During the specific backtesting period, fixed income investments enjoyed a strong lift, due to falling rates. Do not assume such a decreasing bond yield environment going forwards. Yields are expected to rise over time (eventually).

Figure 15.1 Asset allocation models on risk/return plane

Source: Bloomberg. FTSE 100 Index, iBoxx £ Gilts Index, UK Cash LIBOR TR 1 Month Index, UK IPD TR All Property Index. January 1998 to December 2015. Based on monthly total returns, measured in £

Figure 15.2 Backtesting of asset allocation models

Source: Bloomberg. MSCI World Index, MSCI Emerging Markets Index, iBoxx £ Gilts Index, iBoxx £ Non-Gilts Overall Index, Barclays Capital UK Govt Inflation Linked Index, BofA Merrill Lynch Global High Yield Index (measured in $), JPM Emerging Markets Bond Index Plus EMBI+ Composite, UK Cash LIBOR TR 1 Month Index, UK IPD TR All Property Index. January 1998 to December 2015. Based on monthly total returns, measured in £ except when indicated otherwise

Dynamic SAA

SAA is designed for the long term. However, since our CMAs can be updated regularly and they are linked to current market conditions, you can update your SAA periodically.

For example, when equity markets are expensive, dividend yield is low and P/E is high.¹⁰ Equities’ expected return is therefore low. Similarly, when bonds are cheap, their yield is high and so are their expected returns. This is in line with the principle: higher price today means lower returns in the future and vice versa.

With such expected returns, the allocation to equities should be lower and that to bonds should be higher, promoting a healthy discipline of selling high, buying low.

For example, you implement the growth asset allocation model of 70% equity, 15% bonds, 5% cash and 10% alternatives. Given the updated expected returns, you underweight equities to 60% and overweight bonds to 25%.

If your expectations are correct, deviating from SAA can add value. Remember, however, that these are long-term expectations. You can lose for years before potentially benefiting from deviating from SAA. You might not benefit at all since expectations are expectations, they are not certain. You need perseverance, patience and an ability to handle your regret risk.

Regret risk is regretting unprofitable investment decisions. When managing investments, expect to make regretful, wrong decisions. It is part of the game. Learn how to not regret too much when your decisions seem to be losing.

Good things come to those who wait. Sometimes.

One advantage of dynamic SAA is that you do not formulate views about how financial markets are likely to behave over the short term. Dynamic SAA is based on CMAs, which form objective estimates of markets’ future long-term returns, reflecting current valuations. CMAs reiterate what financial markets are implying, not our own subjective market views.

Unless you have a special insight about markets, leave predictions to professionals. It is challenging enough for them. They do it, but only occasionally get it right. Nobody owns a crystal ball foretelling the future. Sadly, successful active investing firmly relies on the skill of correctly forecasting the future.

However, adjusting your allocation once every six or twelve months, based on dynamic SAA, can help systematically buying undervalued assets and selling overvalued ones (taking profits). It follows the investing principle of Warren Buffett of long-term value investing.

Identify attractively priced assets, buy them and hold them for the long term. The long time horizon allows low prices to revert back to fair valuations. However, try avoiding value traps. Sometimes an asset is cheap since it should be cheap. Rubbish is cheap because it is rubbish, not because it is a bargain.

Setting an investment strategy

The plan is in place. You have articulated your investment objectives and designed an investment strategy with the highest likelihood of achieving them. The next step is implementing a solution – turning your plan into reality.

Implementing a solution includes three main activities: actively changing asset allocation to position your portfolio to current market conditions; choosing investments under each asset class; and constructing and managing a portfolio reflecting the asset allocation and investment choice.

Summary

• SAA aligns your portfolio with your investment objectives. This is probably your most important investment decision.
• Optimisation is a process of generating an asset allocation with the highest expected return for a given risk level. Choose a broadly optimised asset allocation aligned with your objectives.
• Use the asset allocation models as a starting point to formulate your SAA.
• Dynamic SAA updates the asset allocation periodically (every 6–12 months) based on current market conditions with a medium-term view (1–5 years).

Notes

¹ The portfolio’s universe usually includes only a subset of all available asset classes. The universe may be defined by investment constraints, such as excluding illiquid assets, such as property, due to liquidity constraints.

² Solver in Microsoft Excel is sufficient to perform simple optimisations with a small number of asset classes. You need to constrain each asset weight between 0% and 100% and the sum of weights to 100%.

³ EMD soft currency should not be hedged to British pound. The hedging costs are too high.

⁴ Backtested return and risk are based on calculating hypothetical returns assuming the static asset allocation (rebalanced monthly) and using monthly total returns of indices representing the asset classes of each asset allocation model. Backtesting demonstrates how the models would have performed in the past.

⁵ The value in each cell in the covariance matrix is Cov_{i, j} = ρ_{i, j}σ_iσ_j. Since in the diagonal of the matrix i = j the values in its cells are the assets’ variances.

⁶ 5.9% = 75% × 5.3% + 25% × 7.6%. For property, the assumed expected return is 7.6%, considering a 3.5% net rental yield.

⁷ 10.9% = square root of (75%² × 14.2%² + 25%² × 4%² + 2 × 75% × 25% × 14.2% × 4% × 0.19); where 14.2% is the volatility of UK equities, 4% is the volatility of UK commercial property and 0.19 is the correlation between them.

⁸ 0.80 = 75% × 1.00 + 25% × 0.19. This is the weighted average of correlation of equities with themselves (1.00) and correlation of equities with property (0.19).

⁹ Capital preservation 18% global developed equity, 2% EME, 42% gilts, 12% IG corporate bonds, 3% global high yield, 3% EMD, 10% cash, 5% commercial property, 2.5% REITs and 2.5% commodities. Aggressive growth 76.5% global developed equity, 8.5% EME, 3.5% gilts, 1.0% IG corporate bonds, 0.3% global high yield, 0.3% EMD, 5% cash, 2.5% commercial property, 1.3% REITs and 1.3% commodities.

¹⁰ Dividends at the numerator are more stable than rising prices at the denominator, pushing dividend yield downward. In P/E the price at the numerator rises more rapidly than earnings at the denominator, pushing P/E upward. Prices reflect the expected future rises in earnings.