Measuring investment performance to highlight if a manager has added real value through skill is a key issue for wealth managers charging for performance.
According to a study released in October 2008 by research firm NOVEO Conseil for BNP Paribas Private Banking, of the 3,000 independent wealth management companies in Switzerland at the time, 49 per cent were found to have less than 100 clients, with only 8 percent registering more than 500 customers.
Meanwhile, 83 per cent managed less than SFr1.0 billion (around $955,000) in assets, with 44 per cent having between SFr100 million and SFr500 million. In addition, the study found that independent managers expect to work with a growing number of family offices going forwards.
Given this profile presented by the study it would appear the focus for Switzerland’s independent asset management community tends to be on serving a relatively small but demanding client base, predominantly of high net worth private clients, and providing them with investment solutions – and increasingly services such as tax planning – to meet their wealth preservation and accumulation goals.
Naturally, the asset manager’s ultimate measure of success will be the investment performance it is able to achieve. But more than that, for the firm to show it is adding value to its clients it has to demonstrate that its performance was achieved by skill and design, rather than by luck or the whims of the marketplace. And it is here that performance attribution becomes so critical.
As such, attribution has a dual role. For one, it provides the manager with a breakdown of where their returns have come from and their role in achieving those – information they can communicate to their clients and so (hopefully) demonstrate their worth.
Alongside that it provides the firm’s portfolio and risk managers with valuable analysis regarding which investment decisions are providing the desired results and which are not, feedback the firm can use internally to guide them in the investment choices they make going forwards and thus be better equipped to fulfil their duties.
As with all forms of analysis, what you measure and the way you do it will impact the outcome. And since attribution can take various forms and be calculated in different ways there is an element of analyst’s discretion that will affect the results obtained (a more extensive discussion of this topic can be found in Advent’s white paper Performance Attribution: A Powerful Tool for Identifying the Sources of Investment Performance to view click here).
One common tool though is absolute attribution, which examines the total return of the portfolio and what contributed to it. This is often used by managers to highlight their most and least successful investments.
Another common measure is relative attribution. This focuses on the excess return i.e. assessing the difference between the actual portfolio return achieved and the benchmark or index return against which the manager’s performance is being compared. The goal is to achieve an understanding of the contribution made by actively managing the portfolio.
The impact of active management can then be considered according to the role of two elements: asset allocation, meaning the manner in which the manager spreads a portfolio’s assets between equities, bonds, real estate, cash, etc; and stock selection, where the analysis will reveal, for example, the return earned by the manager’s choice of equities compared to the overall equity benchmark. And depending on the attribution model used, the analysis may also produce a third component, called interaction, which accounts for that portion of the return that cannot be attributed purely to the asset allocation or stock selection.
But more than merely assessing the total or excess return achieved, meaningful performance analytics need to take account of the investments’ risk profile, which means being able to account for, and separate, both the alpha and beta.
Beta measures the sensitivity – and thus the risk – of an asset or portfolio in terms of its correlated volatility relative to the overall market, for instance the S&P 500. A security with a beta of 1.5 therefore would be expected to move 1.5 times the market’s excess return. Alpha represents the excess return achieved over what can be expected given the beta. As such, it measures the risk-adjusted performance.
But while a high alpha obviously is good, if it also comes with a high beta then the risk involved may be greater than clients are willing to tolerate. Therefore the two must be considered together for the performance of a portfolio to be appropriately controlled.
According to the NOVEO Conseil study, 87 per cent of independent asset managers use their depositary bank for performance calculations. What is of paramount importance though is that managers ensure they are benefiting from sophisticated analytical tools, so as to obtain the requisite degree of insight into portfolio performance that will allow them to make better investment decisions.
For example, accurate performance attribution means having a system that is capable of identifying the sources of portfolio performance relative to a benchmark or model portfolio, evaluating the impact of weighting and security selection decisions, and having the ability to drill down from sector to individual security.
Meanwhile, calculating performance contribution involves pinpointing the drivers of performance by quantifying what sectors, industries, or securities had the greatest or least impact, and ranking them accordingly. A performance analytics engine should also be able to analyse a portfolio’s risk, volatility and risk-adjusted return based on historical performance with ex-post risk statistics.
In addition, where the performance analytics engine is built into or seamlessly integrated with the portfolio management platform then it eliminates reconciliation problems with third-party providers, ensuring the firm’s attribution numbers match its performance numbers.
It is worth pointing out too that accurate attribution and contribution reporting depend on obtaining good constituent data from the different benchmarks being used as the points of comparison. And to do that means having a provider that can supply the necessary data.