Hedge fund size determines performance

Research shows that performance persistence is reduced significantly when hedge fund size and share restrictions, such as notice, redemption and lock-up periods, are incorporated into rebalancing rules

Institutional investors chase hedge fund performance

Large funds no longer deliver best performance. To get a good return investors need to put money into small funds. 

This aphorism has become the standard mantra for many, particularly those who are trying to get institutional investors to put more money into funds that have $500 million or less in assets under management (AUM).

In The Effect of Investment Constraints on Hedge Fund Investor Returns, Robert Kosowski examines the effects of investor restrictions – what academics call ‘frictions’– and real-world investment constraints on the returns that investors can earn from hedge funds. 

The research shows that performance persistence is reduced significantly when fund size and share restrictions such as notice, redemption and lock-up periods are incorporated into rebalancing rules. The ‘frictions’ also include minimum diversification requirements that are commonly used by institutional investors, as well as liquidity constraints.

Using a large consolidated database gathered from Hedge Fund Research, Tass, Eurekahedge, BarclayHedge and Morningstar, Kosowski and his co-authors Juha Joeväärä and Pekka Tolonen, both of the University of Oulu, show that the size/performance relationship is positive/negative when past/future performance is viewed.

Evidence of performance persistence is reduced significantly when fund size and share restrictions are incorporated into rebalancing rules.

The research, first presented by Kosowski at the annual hedge fund research conference sponsored by Lyxor and Euronext held in Paris in January, notes that it is crucial to distinguish between the forward-looking and the backward-looking size/performance relationship. Size matters, concludes the research.

There is an x-shaped relationship between past performance and size today, which is upwardly sloping. “If I were in charge of marketing for a fund, I’d be happy to point to this, saying ‘we are big now and bigger funds have done better in the past.’ But if I were a hedge fund investor, I want to know whether, given you are big now, if you will perform well in the future. And that is the mirror image. The relationship between size in the past and performance in the future is inverse,” says Kosowski.

Larger funds tend to have generated higher returns than smaller funds in the past, but larger funds tend to perform worse than smaller funds in the future.

The research also finds that size is key for the persistence of hedge fund performance. The investor’s choice of the fund size limit is crucial. Looking at hedge funds in the ‘billion dollar club’, performance is very different to that of smaller funds of less than $500 million. 

The research also concluded that incorporating realistic investment restrictions actually reverses widely accepted conclusions about performance persistence among large hedge funds that are the focus of institutional investors.

Looking at the size/performance relationship for 1994–2012, a portfolio of the largest funds – corresponding to AUM of at least $1 billion in 2012 – generates a Sharpe ratio of 0.72, about half the size of the Sharpe ratio (1.34) generated by a portfolio of the smallest funds with AUM of less than $10 million in 2012.

This relationship is monotonic and holds even when the data is controlled for risk-adjusted returns or what the paper calls ‘alpha’. Out-of-sample tests, using returns adjusted for backfill bias, show that the largest group of funds generates a statistically insignificant alpha of 1.7% monthly (1.4% annually). This is in sharp contrast to the corresponding returns exhibited by a portfolio of the smallest funds. These generate an alpha of 3.9% annually.

“These findings do not bode well for the largest hedge funds and management companies since the reported values suggest that larger funds will be unable to continue generating superior risk-adjusted returns. Our results are supported by the remarks of experienced hedge fund investors,” notes the research paper.

Rick Sopher, chairman of LCH Investments, the fund of hedge funds run by the Edmond de Rothschild group, has consistently observed that a critical component of longer-term success for the top 20 hedge funds is to avoid becoming too big. Sopher produces a list looking at return on capital invested.

The research uses this list, too. When looking at the one thing virtually all the managers on the list have in common, they have all, at some point or consistently, restricted how much money they will accept.

Mega hedge fund management firms may point to evidence that past performance is positively related to current fund size. However, that claim is not relevant to investors, who are concerned with a fund’s future performance.

Kosowski and his co-authors’ data confirms the positive relationship between past performance and fund size. A portfolio consisting of the 700 largest funds with December 2012 AUM of at least $1 billion generates a Sharpe ratio of 1.59 when calculations are based on their historical track record, whereas a portfolio of the smallest funds (10,763) generates a Sharpe ratio of 0.66 when calculated in the same way. 

Nevertheless, investors are more interested in whether larger funds will continue to outperform smaller funds. 

For the 20 years of data analysed by the paper, the forward-looking, out-of-sample tests show that this is not the case. When plotting the forward-looking and the backward-looking size/performance relationship in the same graph, the result is an ‘X’-shape. 

The research also explores whether the diversification requirements and fund size restrictions that typically apply to investors, as well as share restrictions more broadly, affect an investor’s ability to exploit performance persistence. Exploiting such persistence may be difficult in practice because hedge funds normally restrict capital withdrawals.

The paper finds that performance persistence when evident is much weaker for value-weighted than for equal-weighted portfolios. 

Investors using historical fund performance as a guide for capital allocation should adopt the latter portfolio-weighting approach. 

However, equal weighting may not be feasible in practice if the investor accounts for too high a proportion of a given fund’s AUM than allowed by its own investment guidelines. 

When the data was controlled for the effect of share restrictions on the rebalancing of investors’ portfolios, it showed investors may not be able to exploit performance persistence at annual horizons. For large (AUM in excess of $500 million) and mega funds (AUM of $1 billion or more), the research found no evidence of performance persistence. 

This, concludes the paper, confirms that fund size is an important determinant of performance persistence. The investor’s choice of fund size limit is crucial.

Although the volatility of the larger funds is lower, the performance risk relationship is less favourable than for the smaller funds, says Kosowski.

Out-of-sample tests show that an investment policy focusing on the 20 historically best-performing funds could deliver superior performance for investors. These results shed light on discussions about decreasing hedge fund industry risk-adjusted performance as a whole. 

Over the period covered, the best-performing hedge funds generate a significant risk-adjusted performance, although the average hedge fund has not been able to deliver significant alpha.

Theoretical and empirical research has addressed the relationship between fund size and performance. One equilibrium model, developed by JB Berk and RC Green in 2004, relating to mutual fund flows and performance in rational markets, shows that mutual funds with positive alphas incur costs that are an increasing convex function of fund size.

In this model a fund with positive alpha receives inflows until its size reaches the point where expected alpha, net of costs, is zero. In equilibrium, all active funds have positive expected alpha before costs, but zero expected alpha net of costs.

Kosowski and his co-authors show that larger funds tend to perform worse in the future than do smaller funds.

However, using economically motivated size categories that are relevant to real-world investors leads to different conclusions about the performance of mega funds, which account for more than 60% of industry AUM. The research found no statistically significant evidence of performance.

Interestingly, the research also finds that there is a group of funds with AUM in excess of $500 million that generate a statistically significant alpha of 8% by chasing performance in less liquid areas.

For funds with shorter than six-month notice periods, the spread between the top and bottom alphas is 4.07%. There is a similar pattern for mega funds, but it is not as strong, concludes the research. The key ingredient is strategy distinctiveness. Based on existing published research, choosing funds based on this criteria is often a better guide to future performance than past performance, according to Kosowski.

However, “we found that the inverse size/performance relationship, forward-looking, holds by style”, he adds.

Another area covered by the research was headline or operational risk. Larger funds tended to have lower drawdowns, which was good news for investors. However, there does appear to be a trade-off between performance and drawdowns. 

Size–performance relationship for forward-looking and backward-looking alphas (PDF)

“Our study supports the trade-off, which is four dimensional – performance, volatility, liquidity and something that I would call operational risk or headline risk/maximum drawdown – with the persistence test. Less liquid funds have a higher persistence,” notes Kosowski.

“We caution against investing in the largest funds but if the investors care about maximum drawdown, then yes, maximum drawdown is lower [in the larger funds]. But this is not the same thing as Sharpe ratio. Volatility may be lower but performance falls more than is commensurate with lower maximum drawdowns. With the persistence analysis for large funds, if we look at large and mega funds, there is some performance persistence especially for illiquid funds.”

Another area the research team looked at was investor, rather than fund, size. For example, large investors, such as sovereign wealth funds or big pension plans, could have a sizeable pure hedge fund allocation of $1 billion. These investors are likely to have limits on investment such as not investing more than 10% of the AUM of a fund. A high-net-worth individual or smaller family office may have only $100–500 million to invest and may be limited because the larger funds need a much more sizeable allocation than a smaller investor is able to make.

Anecdotal evidence, says Kosowski, also suggests investment in 20 hedge funds is manageable, given the amount of monitoring and due diligence required. The research does find that if investors allocated across all funds, independent of size, there is a statistically significant likelihood of positive performance and that the investment will be more profitable. However, the trend still appears to be for allocators to choose larger funds rather than widening their universe to include small and medium-sized funds (in the range of $10 million up to $500 million).

The research findings confirm what many have suggested: fund size is an important determinant of hedge fund performance persistence. The results are strongly affected by the fund’s choice of size limit. The results for $1 billion and above funds differ significantly from those under this threshold. Although the $1 billion-plus funds account for more than 60% of industry AUM, on average they do not generate statistically significant alphas and do not exhibit performance persistence. 

While this should be good news for smaller funds hoping to capture more money from investors, the problem of producing performance with more money is a problem both hedge fund managers and investors will need to consider in future.