The ideal number of managers to be included in a £500m (e830m) UK equity multi-manager portfolio is probably between five and six, according to a study by London-based multi-manager firm Frank Russell.
The study, carried out by Steve Wiltshire, director of research at Frank Russell, tries to find the answer to the question that an increasing number of investors wishing to use a specialist multi-manager approach for their investments are faced with: how many managers should we include in our portfolio?
Investors who use too few managers within an asset class could end up with major unintended style biases in their portfolios. On the other hand, using too many managers in an uncoordinated manner could result in an expensive index fund.
The main factors affecting the ‘number of managers’ decision are the expected manager return and risk, the correlation of manager returns, manager fees and the volume of assets of the fund. “If you can understand how these factors interact then you can maximise the value of active management,” Wiltshire says. Russell has framed the problem in terms of a cost-benefit analysis. “Adding additional uncorrelated managers that are expected to outperform to a portfolio increases the benefits of diversification, but at progressively slower rate. At the same time, the aggregate cost will tend to rise at an increasing rate as more managers have smaller portfolios to manage, which attract proportionately higher fees,” says Wiltshire.
In the language of cost-benefit analysis, the optimum number of managers is reached when the marginal cost (effectively fees, measured in basis points) of adding an additional manager is exactly matched by the marginal benefit (risk-adjusted return, measured in basis points). This ideal number of managers will vary with the overall volume of assets invested in an asset class because of the non-linear relationships between costs, benefits and portfolio size.
Figure 1 is based on a UK equity portfolio of £500m, which is comparable to the current size of Russell’s UK equity fund. As can be seen, the theoretically optimal number of managers is between five and six. As the fund size increases, the number of managers also increases (see Figure 2).
But in practice things can be much more complex. Fee scales can vary dramatically between managers, and the scales tend to be stepped, X% for the first £10m, Y% for the next £15m and so on. “Also, some managers have minimum account sizes which are quite high, making it difficult for small pension funds to access specialists,” says Wiltshire. Also, investors with larger accounts are sometimes able to negotiate more favourable fees. All these factors make it clear that, in reality, not all investors face the same costs.
Wiltshire explains that estimating the benefit of diversification can also involve as much art as science since in practice this requires the manager-of- managers (whether Russell or another large investor) to assess how each manager’s performance will interact with that of other managers.
The portfolio of the Russell UK Equity fund is split between five managers – Schroders, Fidelity, GMO Woolley, BGI and Glasgow – which have mandates tailored to exploit their skills and to help maintain the balance of the overall portfolio structure. A further level of control on the expected level of risk and return from this combination of managers is achieved by weighting the managers according to their styles and performance characteristics. The number of managers in Russell’s fund increased from one to five over the past four years at the same time as the assets of the funds grew from a standing start to over £600m.
Russell does not share the perception that the concentration of assets in the largest UK managers necessarily means there is inadequate choice. “Our experience shows that careful research does generate effective strategies for multi-manager portfolios,” says Wiltshire. “At some point in the growth of this fund we may hit a manager supply constraint but that’s many years away and by then we will have figured out how to finesse the problem,” he says.
As manager performance characteristics vary, manager costs and style also vary widely and manager supply is finite, coming back to the question of how many managers should be included in a portfolio the answer seems to be “it depends”. “The marriage of investment theory and practical experience is the key to answering the ‘how many manager’ question,” says Wiltshire, “and getting it right is also key to delivering consistently attractive returns to our clients”.