Mike Mauboussin's latest write up Reflections on the Ten Attributes of Great Investors is a must read for investors (while his recent podcast with Barry Ritholtz is a must listen). And while I broadly agree with his ten main points, I thought it was worth discussing details / providing an alternative take to his main point under the section 'What's Next?', which touches on the impact passive / ETF asset growth will have on the ability for active investors to outperform.
His specific thesis (#'s and bold mine):
Thirty years ago,
index funds were less than one percent of assets under management, and today
they (along with other passive vehicles such as exchange-traded funds) are
about one-third. Think of it this way: (1) For you to have positive alpha, the
industry’s term for risk-adjusted excess return, someone has to have negative
alpha of the same amount. By definition, alpha for the market must equal zero
(before fees).
So you want to compete against less-skilled investors
because they are your source of alpha. (2) It is disadvantageous for you if the
weak players flee the market (selling their stocks and buying index funds), or
if the least capable professional investors lose assets to passive funds,
because it means that only the smartest investors remain in the active game.
The truth is that weak players, whom the strong players require to generate
excess returns, are fleeing at a record pace.
To paraphrase, (1) the market is a zero sum game (relative winners only exist if there are relative losers) and (2) less skilled investors are going passive, resulting in the amount of negative alpha available for remaining investors shrinking.
Is the Market a Zero Sum Game? Yes and No.
I agree with the view that the market is zero sum when looking at dollar-weighted performance (i.e. if everyone is fully invested and there are no flows, then if some investors are up relative to the market, then others must be down, resulting in net returns equal to the market less fees). BUT
investors are known to be horrific market timers, regularly underperforming the return of the funds in which they invest.
As a result, the negative performance gap between the performance of funds / ETFs and the performance of investors can be captured by fund managers and/or investors, meaning that it is theoretically possible that all funds outperform their benchmark, while all investors underperform the funds they allocate to (the zero sum nature should be viewed in time-weighted performance across all asset categories, rather than dollar-weighted performance within one category, as flows exist in the real world). This was the essence of my favorite investment read of 2016
Active vs. Passive Investing and the “Suckers at the Poker Table” Fallacy.
Are There Less Weak Players in the Market When They Go Passive? Sure Doesn't Seem So.
Weak players may simply be making poor investment decisions elsewhere. The chart below shows the performance of mutual funds and ETFs across a variety of U.S. styles (growth, blend, value), as well as market caps over the last five years per Morningstar methodology as of June 30th, 2016. Note that I think the total return and investor return calculation used by Morningstar is flawed* (see footnote below), but it is the only source I am aware and is used (fairly or not) throughout the industry (so I'll use it here, but if readers have a better source let me know).
Note that ten year data was unavailable for 4 of the 6 ETF categories, though reported investor underperformance was similar in scale for large blend and large value.
Using this calculation, the average total return of ETFs was higher than mutual funds across categories over the last five years, BUT ETF investors
massively underperformed both the return of the ETF and the investor return of similar style mutual funds over the same time frame.
What Might This All Mean?
It means ETFs may actually make weak players weaker. ETFs have certainly enabled irresponsible investors to more readily negatively impact their performance through the following capabilities previously unavailable / less prevalent within mutual fund structures:
- Trade asset classes intra-day and after market
- Build negative carry portfolios through shorting
- Invest easily with leverage in horrific structures
- Performance chase in highly esoteric / niche areas of the market
- Become asset allocating tourists (vs manager selection tourists)
Investor returns in the chart above would indicate ETFs have seemingly ADDED to the alpha pool within these plain-vanilla categories, which does not include the more esoteric asset classes / structures that have already
removed billions of dollars of investor capital from their wallets. As a result, the poor timing by ETF investors may make more alpha available for those institutions and investors that can act as a counter to this poor market timing.
But given most investors have underperformed within mutual funds as well, where has all of this excess performance gone? I need to think more about what I am going to say next, but what if it is has gone to the companies themselves in the form of stock appreciation and senior level stock compensation? If investors have been poor timers this cycle, then perhaps companies that have consistently bought back their own stock have been good market timers?
Footnote: Morningstar Total Return and Investor Return Calculation
I believe there are a lot of issues with Morningstar category total return calculations, as well as Morningstar category and fund investor return calculations.
Category total return calculation: see
here for their methodology, but in a nutshell the calculation averages the total return for each fund within each style universe irrespective of asset size and benchmark; a tiny fund with poor performance holds as much weight as a huge fund with outperformance. As a result, it is theoretically possible for the average dollar weighted investment in funds to outperform the index even if the category underperformed.
As is the case of the Morningstar Intermediate Bond category… for the five years ending 6/30/16, the average fund underperformed the Barclays Aggregate Bond Index. BUT, the dollar weighted fund return outperformed the index and funds benchmarked against the applicable Barclays Aggregate Bond Index outperformed the index further. Narrowing the focus to institutional share classes added another 20+ bps to excess performance. Note in this analysis I used the most recent share class AUM, which may overstate the relative performance given investor performance chasing (i.e. AUM is likely overstated for the best performing funds).
Investor return calculation: I've written about the issue of comparing time weighted and asset weighted returns
previously. There is an additional issue, similar to the one within the total return calculation outlined above, which is the calculation seems to average the investor return for all funds in each category. As a result, a tiny fund with tiny poorly timed flows may have an oversized impact to the category average. For example... if a $10 million fund received a $5 million outflow, which was then allocated to a much larger fund, the investor return of the small fund could be severely impacted by the outflow, while having an immaterial impact to the larger fund. Averaging the two at a 50/50 weight would mean the flow could have a huge impact to the category average, while being a wash in reality. Finally... investor returns for ETFs also seem to include flows related to short investors / hedgers within these ETFs, but my view is this is less of an issue as shorts that underperform still add alpha to the overall "alpha pool" in a zero sum world (i.e. shorts / hedgers that lose money have negative alpha... a boon to other investors).
Smell Test: In terms of whether the direction of the investor returns looks right, I did take a look at flows and returns within the Large Blend ETF category going back five and ten years. Over the last ten years, the slope of the regression between 12-month flows and the 12-month historical return was positive 10x (i.e. if returns were -10% in the previous 12-months, then the regression assumes outflows of -$10 billion over that same window). On the other hand, the slope of the regression between 12-month flows to the Large Blend category and 12-month forward returns was negative at -10x (i.e. if flows were -$10 billion, then the slope of the regression assumes future returns were positive 10%). In other words, investors zigged after markets zigged, then markets zagged, which would lead to poor investor returns.
