Recognizing the Risk of Flow-Driven Performance

In the “assets” module of the Advanced Due Diligence and Manager Selection course, there is a section on the relationship between performance and flows.  As discussed there, it is widely accepted that flows are responsive to performance.

Good performance leads to inflows and bad performance leads to outflows.  It makes sense given human nature — and study after study has shown it to be true across asset classes, public and private.

A simple visual accompanies the narration:

After some notable examples of that principle, the arrow on the slide starts to rotate, resulting in this:

The effect of flows on performance isn’t talked about much at all, even though the impact can be dramatic in certain situations.  There is little research on the topic; one good example was published in the Financial Analysts Journal in 2014, “Flows, Price Pressure, and Hedge Fund Returns.”  The authors, Katja Ahoniemi and Petri Jylhä, concluded that “one-third of estimated hedge fund alphas are due to flows.”

A new paper

Enter “Ponzi Funds,” a paper from Philippe van der Beck, Jean-Philippe Bouchaud, and Dario Villamaina.  As for the use of the word “Ponzi”:

We emphasize that the title of this paper does not suggest that concentrated investment funds are literal Ponzi schemes as defined by the SEC. . . . Instead, the term ‘Ponzi funds’ merely conveys the notion of self-inflated returns.

It is difficult to estimate the degree of performance that stems from price pressure related to flows, especially for open-ended mutual funds, hedge funds, and separate accounts, each of which presents data challenges.  In contrast:

ETFs are ideally suited because their portfolio holdings and flows are observable at a daily frequency, and because the vast majority of ETFs perfectly reinvest flows in their existing positions on the same day.

Funds (of any kind) that on average have ownership stakes that are large in comparison to the daily volume of their holdings are the most likely to have returns that are affected by flows:

When concentrated funds become large, their fund illiquidity spikes giving rise to self-inflated returns.

The importance of self-inflated returns in explaining funds’ overall returns increases monotonically in both the size and the concentration of their portfolios.

In regards to one anonymized thematic ETF, when it “received a 1% inflow on a certain day and proportionally rescaled its positions, it bought 20% of the daily volume in the underlying stocks.”  That being the case, it is not hard to imagine flows driving performance.  (Of course, the process works in reverse when there are outflows.)

From their research, the authors conclude that “investors are unable to differentiate between self-inflated returns (price impact) and fundamental returns (stock-picking skill).”  And there is a compounding effect:

Because investors place a higher weight on the most recent return, even short-lived price pressure can have an impact on the distribution of fund flows.  This can cause an endogenous feedback loop:  Flows cause a price impact, which is a realized return that causes further inflows and amplifies the initial price impact.

The paper includes charts that show a significant difference in return patterns between “run-up ETFs” — ones that have had substantial outperformance — from “bubble ETFs,” those which have had cumulative Ponzi flows at the upper end of all of the run-up funds.

For run-up funds, “excessive outperformance is on average not followed by a subsequent crash,” with cumulative returns reverting to market returns.  But the bubble-ETFs crumble well below those levels.

Examples and strategies

The first sentence of the paper provides an egregious example from outside the world of ETFs:

The collapse of Archegos Capital Management prominently showed that when investment funds trade concentrated positions, portfolio returns can be driven by the funds’ own price impact.

A more typical recent example is ARK, which very much fits the pattern of Janus in the the dot-com era:  performance begat flows which begat performance as the cash was plowed into a limited number of relatively illiquid stocks — until the music stopped and the machine reversed.

The ebb and flow of performance comes with the territory, but capital allocators often overlook the flow effects buried in the performance record and select a fund at the worst possible time.

While it is impractical to calculate the “Ponzi flow” before choosing a manager, it is easy enough to have that be a factor that must be addressed before making a commitment.  Estimating the general risk is possible even if the specifics are lacking.

As the authors note, “timing the crash is difficult.”  Strong moves in relative outperformance can last a while, and some investors are willing to chase a bubble while it is still inflating, although that can be a dicey strategy if the vehicle used is not easy to exit quickly.

Everyone wants to be in on the virtuous part of the cycle, but not its evil twin.  Managers adept at self-inflating returns tend to be particularly alluring just before the gears shift.

 

In addition to the course mentioned, The Investment Ecosystem offers training and facilitation for organizations regarding due diligence, innovation, and communication.