White Papers

How flexible do we need to be?


Greg Cooper

Greg Cooper

CEO Schroders Australia / Global Head of Institutional

Simon Doyle

Simon Doyle

Head of Fixed Income & Multi-Asset

Chris Durack

Chris Durack

Country Head of Hong Kong and Head of Institutional Asia Pacific

Simon Stevenson

Simon Stevenson

Deputy Head of Multi-Asset


In the paper “Why strategic asset allocation is flawed” we analysed a ‘stylised’ traditional balanced fund using asset class data commencing in 1900 and showed that historically, while in the long run, average return objectives have been met, the time horizons required are significant.  Fixed strategic asset allocations have generated significant medium term volatility of outcomes, making them unsuitable for consistently achieving objectives.  This has been due to the fact that equity markets have delivered real returns in long term cycles.  It begs the question as to what level of flexibility is required in the asset allocation process in order to consistently achieve investment objectives. What degree of asset allocation flexibility would have been required historically to achieve a typical investors’ investment objectives (in this case taken as CPI+4.5% p.a. over rolling 5-10 year periods)?

What is the optimal portfolio?

To understand the degree to which the constraints of a traditional asset allocation strategy would need to be unwound to achieve a more consistent pattern of returns, we analysed the historical dataset assuming perfect foresight on the returns, risks, and correlations between the asset classes. That is, we determined the optimal portfolio for a real return of 4.5%p.a. based on the minimisation of downside risk and a minimum level of diversification by utilising a ‘diversity factor’. The diversity factor is based on two elements: a measure of concentration, in this case the Herfindahl Index1 , which is basically the sum of the squares of the portfolio weights (i.e. extremes are magnified); and a multiplier that determines how much influence the concentration measure has on the optimisation output.

Chart 1: Optimal portfolios with diversity factor - decade by decade

Chart 1 displays the results decade by decade and shows that achieving a real return objective of 4.5%p.a. over shorter (ie 10 year) timeframes was possible but required substantial flexibility in asset allocation. Importantly, the portfolios met the rolling time constraints set by funds with much higher frequency than a fixed SAA portfolio, with one exception, the 1970’s.

What is clear is need to have very wide asset allocation ranges in order to consistently deliver on objectives. Chart 2 shows the difference in asset allocation flexibility required on a decade by decade basis relative to a traditional 60/40 balanced fund.

Chart 2: Difference in asset allocation required to meet investment objective

Comparing the performance of the optimal portfolio with the traditional 60/40 balanced portfolio decade by decade shows that flexible asset allocation portfolio captured much of the upside in good decades, but also did not experience the poor performance of the traditional fixed strategic asset allocation (SAA) portfolio in challenging decades (refer Chart 3). Average real returns over the period were 5.3% p.a. for the fixed SAA portfolio and 5.4% p.a. for the unconstrained portfolio.

Chart 3: Optimal portfolios performance - decade by decade

Forecasting returns

While this analysis shows it would generally have been possible to meet rolling real return targets it assumes some level of foresight with respect to future asset class returns. The next question is then to what degree are future market returns predictable on a systematic basis?

Fixed asset allocation funds typically rely on long run equilibrium returns as the base assumption. We require a more accurate forecast of future market returns, particularly equity market returns and have utilised a systematic process that requires no qualitative judgement.

Research by Campbell and Shiller2 shows that simple analysis based on cyclical adjusted PEs can provide relatively accurate forecasts over 7 years or longer time horizons. The quality of this simple relationship is shown powerfully in charts 4 and 5, which plots the actual 10 year returns versus a predicted 10 year return based purely on Shiller PE (real price divided by real 10 year rolling earnings) for the US equity and Australian equity markets. Campbell and Shiller used this relationship in their writings of the late 1990s and early 2000s to argue that the outlook for the US equity market in the 2000s was very poor, which with the benefit of hindsight, was correct.

Chart 4: US Equity Market – Actual and Forecast Returns Based on Shiller PE

Chart 5: Australian Equity Market – Actual and Forecast Returns Based on Shiller PE

Given such a straightforward forecast methodology can produce successful outcomes we contend that investors should be making use of this information to develop asset allocations that are more likely to meet investment objectives. Such portfolios are likely to result in a much wider divergence of asset allocations than traditional fixed strategic asset allocations. In the same way that most investors would never expect to hold a fixed allocation to a particular stock in an equity portfolio irrespective of price, we see no reason why investors should be comfortable doing the same with their asset class exposure.

Developing asset allocation ranges

What asset allocation is required to meet investment objectives on a forward looking basis? In Schroders’ multi-asset team’s forward looking base case we assume the aftermath of the global financial crisis and the deleveraging process continues for a total of 7 years, consistent with historical experience. The base case analysis provides returns for equity markets that at first look very high. However, while the assumptions underlying the forecasts are relatively conservative, they are consistent with analysis based on the Shiller PE, and they are not that different from the long term performance of these markets (since 1900 13%p.a. for Australian equities and 11%p.a. for unhedged global equities).

The key return assumptions are as follows:

Prospective 7-10 year return forecasts

Portfolio modelling scenario

The return distribution assumptions were optimised in Schroder Multi-Asset Risk Technology (SMART) relative to inflation, to provide a portfolio consistent with a real 4.5%p.a. real return objective. The optimised portfolio is outlined below. We have compared this portfolio relative to an “industry average” portfolio.

There are a number of key observations about the portfolios to note.

  1. The base case gives a very different asset allocation to the typical industry portfolio. This is largely a function of the base case portfolio emphasising objectives over return maximisation. Given the wide distribution of possible outcomes embedded in the forecasts, a portfolio with greater certainty is favoured. This can be achieved with a much lower exposure to riskier assets.
  2. The base case portfolio has a very high weighting to cash. This is driven by the strong risk-return nature of cash, where in Australia we expect a relatively high real return while the volatility of cash is very low.
  3. The base case portfolio has a bias towards index-linked bonds over nominal bonds. This is not surprising given the portfolio was optimised against inflation and reflects the better relationship between index-linked bonds and inflation relative to nominal bonds.
  4. The base case portfolio has a relatively low exposure to equities and a relatively high exposure to high yielding credit. Given expected returns from credit securities are relatively high based on the elevated level of credit spreads, this provides the ability to access the corporate risk premium but do so in a risk controlled manner, by being higher up the capital structure.
  5. While the industry average portfolio has a higher expected return, this comes with considerable downside risk. The VaR and Conditional VaR results show the potential for significant negative return outcomes (not inconsistent with what was observed in 2008).


Historical evidence shows that it is not possible for fixed asset allocation portfolios to generate reliably real returns within required timeframes that are consistent with many individuals’ objectives. Instead fixed asset allocation portfolios require a very long term time horizon, given that equity markets in particular have delivered real returns in long term cycles or ‘regimes’. Fixed strategic asset allocations generate significant medium term volatility of outcomes, making them unsuitable for consistently achieving objectives.

In particular, we observe that there are three key portfolio management capabilities required for a plan to achieve a real return objective of circa 4-5% p.a. over a defined 5-10 year time frame:

  1. The breadth of asset allocation ranges needs to be wide, with our analysis suggesting that unconstrained ranges are most likely to be required. This is the only way to be reasonably assured of the potential to achieve the real return objective over the time frame stipulated. Narrower asset allocations ranges will require the time frame over which to achieve the real return objective to be lengthened substantially.
  2. Some capability around the forecasting of asset class distributions over the objective timeframe will be required. While academic literature suggests this is possible, a robust process is necessary to successfully use the wider asset allocation ranges.
  3. Ability to change asset allocation when required. The frequency of asset allocation changes will not be high in managing to the 5 to 10 year part of the time frame, given the long run regime nature of financial markets. However, managing to a shorter time period would involve managing the cyclical nature of financial markets.

How Flexible do we need to be? (2012)

Revisited: How Flexible do we need to be? (2014)

1The Herfindahl index (also known as Herfindahl–Hirschman Index, or HHI) is commonly used as a measure of the size of firms in relation to the industry and an indicator of the amount of competition among them. In this case we use the measure to reduce the dominance of any one asset class in a portfolio.
2 Campbell, John Y., and Robert J. Shiller, “The Dividend–Price Ratio and Expectations of Future Dividends and Discount Factors,” Review of Financial Studies, 1:195–228, Fall 1988(a). Campbell, John Y., and Robert J. Shiller, “Stock Prices, Earnings, and Expected Dividends,” Journal of Finance, 43(3): 661–676, July 1988(b).

Important Information:
Opinions, estimates and projections in this article constitute the current judgement of the author as of the date of this article. They do not necessarily reflect the opinions of Schroder Investment Management Australia Limited, ABN 22 000 443 274, AFS Licence 226473 ("Schroders") or any member of the Schroders Group and are subject to change without notice. In preparing this document, we have relied upon and assumed, without independent verification, the accuracy and completeness of all information available from public sources or which was otherwise reviewed by us. Schroders does not give any warranty as to the accuracy, reliability or completeness of information which is contained in this article. Except insofar as liability under any statute cannot be excluded, Schroders and its directors, employees, consultants or any company in the Schroders Group do not accept any liability (whether arising in contract, in tort or negligence or otherwise) for any error or omission in this article or for any resulting loss or damage (whether direct, indirect, consequential or otherwise) suffered by the recipient of this article or any other person. This document does not contain, and should not be relied on as containing any investment, accounting, legal or tax advice. Schroders may record and monitor telephone calls for security, training and compliance purposes.