Over-diversification: The quickest way to overpay for active management

There is one mathematical feature of blending diversified portfolios that is more mis-understood than any other in investment management; the cost per unit of active risk.

Conventional wisdom would have it that the more active fund managers are added to a portfolio the less systematic risk is in the portfolio – allowing the bottom-up geniuses to thrive and leaving only the stock-specific risk at work.

Unfortunately, this is untrue. Adding too many active equity fund managers is the simplest way to remove the benefits of active fund management.

A significant paper from the CFA explains the problem neatly.

In ‘What Free Lunch? The Costs of Overdiversification’ authors Shawn McKay, Robert Shapiro and Ric Thomas analyse 88 of the major US defined benefit pension plans and find that on average these funds allocate to 17 different equity managers.

The authors use active share and active risk to formulate two new measures ‘fees per unit of active risk and ‘fees per unit of active share’.

The authors then analyse randomised portfolios covering nearly 700 active funds in order to examine the impact that diversification has on the level of active risk in a portfolio.

They conclude that invariably more diversified portfolios express far less active risk the more managers are added. If this point seems obvious we should note that many providers of multi-asset funds and model portfolios argue against this principle.

This conclusion leads to a corresponding need for a higher information ratio (the value a fund manager adds for the active decisions they take) from each of the underlying fund managers if they are to justify their fees.

The authors quantify it as follows. On average, when the level of diversification in a portfolio rises from one active equity manager to ten, the ex-ante active risk, on average, falls from 3.3% to 1.2%. This leads to a rise in the ‘fee per unit of active risk’ from 19.4 basis points to 53.2 basis points.

In other words, however good an active equity manager is, if you overdiversify them with others then they are forced to work incredibly hard to justify active fees. The optical cost of the active management may look the same on the factsheet, but in reality it is a multiple higher than a more focused portfolio.

To put it another way, a group of five ‘good’ active fund managers achieving, say, an information ratio of 0.25 could significantly outperform an ‘exceptional’ group of active fund managers delivering an information ratio of 0.5.

So how do investors respond to this? Firstly, by challenging the providers of model portfolios.

Those who do not add value through asset allocation and instead claim that they have found dozens of exceptional stock-pickers are likely producing index returns at active cost.

Secondly, by gaining a key understanding of how much diversification is actually needed.

The seminal work of 1968 by John L. Evans and Stephen H. Archer famously demonstrated that stock portfolios of 20 stocks eliminate almost all the systemic risk of concentration. We can ask then why a portfolio of perhaps a 1,000 underlying stocks is necessary across equity portfolios within a model?

At Albemarle Street Partners we are clear on how we address this problem.

Firstly, we place emphasis and weight on the need to outperform through our asset allocation, allowing ourselves enough active risk to express a meaningful view within the risk profile of clients.

Secondly, we are clear on why we own every active equity fund manager – requiring them to express a distinct factor that contributes to the overall balance of a portfolio which is aligned to the long-term evidence for what drives equity markets.

Thirdly, we interrogate each fund manager in enough detail that we are willing to express a position in the fund with enough weight to allow the fund manager to add value.

These are the keys to gaining appropriate diversification without creating closet index trackers at a high fee for investors.