Building and measuring outcome-oriented investment strategies
For too long there has been little connection between real-world investment problems and the way portfolios built to solve them are measured. We suggest an alternative to capital-weighted benchmarks which could provide a reality check on whether objectives are achievable.
We are all faced with real-world investment problems. These might be the need to achieve adequate savings for retirement, to generate income to cover living costs, to meet the pension promises made to employees, or simply to maintain purchasing power over time. For many years, however, there has been a fundamental disconnection between these problems and the benchmark related portfolios which have been built to solve them. The use of market indices as de facto proxies for the intended risk exposures has allowed segmentation into convenient building blocks and easier portfolio monitoring, but has also led to differences between asset owners and their investment managers in how success is defined.
By using real-world outcomes as the explicit target results for portfolios, investors can dispense with these market related proxies and better align the success measures of their appointed managers with their own. Portfolios are judged on the delivery of the outcome, rather than compared with an abstract market measure (Figure 1).
Figure 1: What’s the problem you’re trying to solve?
Source: Schroders, for illustration only.
These five target outcomes can be distilled into more specific objectives. This article outlines a method to ensure these objectives are realistic and achievable. But such targets do not exist in isolation. As well as primary and secondary objectives, investors also have tolerance levels and investment restrictions to consider – it is the combination and relative importance of these elements that define the portfolio’s shape.
The likelihood of the successful achievement of target outcomes depends on:
– the reasonableness and consistency of primary and secondary objectives
– the extent of constraints imposed on the manager(s)
– the market environment experienced over the investor’s time horizon, including startingvaluations, and
– the manager’s ability to control outcomes through active management decisions.
Creating an outcome-oriented framework
As the basis for our analysis, we created an unbiased range of permutations of five asset classes – US equities, global equities, US Treasury bonds, commodities and cash. Rather than being random, these asset allocations attempt to replicate the full universe of possible portfolios. By populating them with actual market data, we can then define ranges of achievable outcomes, such as returns, volatility, drawdowns, income-generation or capital preservation. This demonstrates how outcomes change with asset class opportunities and over time, allowing us to identify realistic parameters for outcome-oriented portfolios.
Such a framework can be used to demonstrate how investors’ objectives and theirconstraints are inherently linked by showing, for example, the returns achievable from portfolios within certain risk limits or with restricted capital allocations. This framework can also be used as a monitoring tool for outcome-oriented portfolios. By combining information about the medium-term market environment with the actual long-term objectives, an intelligent assessment can be made of the portfolio’s performance.
The remainder of this article is a worked example of a risk-controlled growth portfolio, demonstrating the setting of realistic parameters and the use of our framework to monitor the outcomes. The main growth objective here is a typical specified portfolio return, often in absolute terms or in excess of a cash or an inflation measure. The secondary objective is the “risk-controlled” part. Here again, definitions can vary, with risk being seen variously in terms of loss tolerance, return variability/smoothness or in measures like value at risk. Figure 2 plots the range of return outcomes from our simulated portfolios over time. We also show a few real return target levels.
Figure 2: Real returns over five-year periods are often impossible to achieve
1001 simulated portfolios for each five-year period. Source: Bloomberg, Thomson Reuters Datastream and Schroders. Indices
used are the S&P 500 TR, US 10 year Treasuries, MSCI World, US 1 week LIBOR and DJ AIG Commodity.