Is Passive Investing Better than Active Investing?A Critical Review.
1. Introduction
In an influential piece, Sharpe (1991)Sharpe, W.F. 1991. The arithmetic of active management. Financial Analysts Journal, 47(1), pp.7–9.9 put forward the proposition that active investing must be a losing pursuit in the aggregate, as it amounts to a zero-sum game in gross terms and hence must be a negative-sum game after costs. I take a critical look at the underlying concepts and assumptions behind Sharpe’s proposition and link it to the issue of whether it is worthwhile for investors to consider using active fund managers. I highlight some of the related academic research along the way. This paper has been written for investment industry practitioners and is mainly a synthesis of existing ideas.
The broad thrust of this paper is that Sharpe’s proposition is not water-tight upon closer examination, and certainly should not be received as gospel. In particular, it should not be taken as sufficient to support the general conclusion that indexed funds should necessarily be favored over active managers. Rather, the evidence indicates that whether active or passive management is likely to deliver a better outcome will vary with the circumstances of the investor and the asset category being considered. It is wrong to make any sweeping generalizations. ‘It depends’ is the main message.
Before continuing, let me put some views on the table. Personally, I am neither for nor against either active or passive management, but rather see both options as worth considering. Fortunately, in some quarters the choice between active and passive no longer seems viewed as a ubiquitous either/or decision, which I wholeheartedly endorse. In other quarters, the mantra that ‘active management adds no value’ seems to have been embraced and is helping to fuel a switch to passive which may not be in the best interests of all investors. Further, the active versus passive debate is still often presented in adversarial terms. It would be much better framed around identifying situations where one approach might be preferred, or both should be used in tandem. To this effect, section 9 presents a framework that may assist to identify situations where active managers may form part of the mix. Hopefully, this paper may help spur more fruitful discussion and some deeper thought.
A PDF version of this is available here.
2. Sharpe’s Proposition
Sharpe’s logic is seductive. He calls his paper “the arithmetic of active management”, and presents a case that active investing must be a negative-sum game as an inconvertible mathematical truism. His argument is that, as all investors hold the market in aggregate, then active investing must fail on average as a matter of an adding-up constraint. Passive investors who hold the market will earn the gross market return pre-costs. The remainder are active investors who each deviate from market weights in some way, but in aggregate must also hold the market for the holdings to add up. Thus active investors also get the gross market return in the aggregate, with anyone active investor’s gain being another active investor’s loss. This also means active investors must, therefore, do worse than passive investors in net returns as they are incurring greater costs in terms of fees and trading. Active investing is thus a zero-sum game in gross terms and a negative-sum game in net terms. QED.
French (2008)French, K.R., 2008. Presidential address: The cost of active investing. Journal of Finance, 63(4), pp.1537–1573 backed up Sharpe in his Presidential Address to the American Finance Association. He presented an estimate of 67 basis points as the aggregate cost that investors incurred in pursuing active returns in US equities over the period 1980–2006. His estimate incorporates all fees and expenses in mutual funds, hedge fund fees, the investment management costs of institutional investors, and an estimate of trading costs; and deducts the estimated cost of investing passively. The implication is that this is a big cost to pay in the hope of outperforming the market.
French acknowledges that active investment may provide a social benefit by making the market more efficient. Even so, active investors are paying the price for this social benefit without reaping any private gain. French estimates that the total cost incurred of providing this ‘service’ amounts to about 10% of the total market capitalization. While greater market efficiency may raise returns for all investors, he argues that active management remains a negative-sum game nevertheless. Essentially passive investors are free-riders on any positive externalities generated by active managers through their efforts to outperform.
The implications of Sharpe’s proposition are worth pointing out as background for interpreting the evidence from academic research. A key test of Sharpe’s proposition that active investing amounts to a zero-sum game at the gross return level and a negative-sum game at the net return level is whether the aggregate (i.e. asset-weighted) return earned by active investors relative to the market is zero before costs. If this were found to be the case, then it naturally follows that the net active return will be negative. As we will see, there is a fair amount of dissonance between this condition and the way that most academic research is conducted. In particular, it is difficult to observe the returns earned by all active investors. Further, Sharpe’s proposition does not imply that active managers cannot outperform, but rather that other active investors must underperform for them to do so. This nuance leaves open the door for active managers to outperform where they possess some competitive advantage over other active investors: a matter that will be addressed later.
3. Grossman and Stiglitz offer a different perspective
Sharpe’s proposition grates against an alternative perspective that was put forward earlier by Grossman and Stiglitz (1980) Grossman, S.J. and Stiglitz, J.E., 1980. On the impossibility of informationally efficient markets. The American Economic Review, 70(3), pp.393–408. Grossman and Stiglitz (GS) describe an equilibrium where those that invest in ‘information’ receive higher gross returns on average, but these returns are just enough to cover the costs of seeking out the information. Meanwhile, those who do not invest in information get lower gross returns. However, the net expected return is the same for both groups on average. The implication is that a GS equilibrium would see active investors generating higher gross returns than passive investors, but only just enough to offset the difference in costs. In these research pieces, we have two Nobel Prize laureates pitched against each other
GS’s perspective makes sense as a plausible equilibrium. It presents a rational explanation for why active management can cohabitate with passive investing. Indeed, the GS view effectively implies that Sharpe’s proposition does not accord with a rational equilibrium, as one would need to assume that active investors are irrational to explain the existence of active management. That is, active investors are either fools, or their decision-making is impaired by behavioral effects. This is possible, and perhaps the current strong growth in passive investing is a sign of people waking up to the error. However, relying on pervasive and persistent irrationality to explain the broad and extended use of active managers over many years seems a longbow. Mass delusion on this scale seems unlikely. Could people really be THAT stupid?
Another argument for why equilibrium might look like that proposed by GS arises by considering the plausibility of the two extremes of either 100% active or 100% passive investing. Either extreme is likely to be unstable. A 100% active approach would amount to a costly competitive game over active returns. The losers of the game — those with limited ‘skill’ — would have the incentive to give up paying to play and go passive. On the other hand, 100% passive investing would likely produce significant mispricings, given that no one would be focusing on whether prices are right. This should encourage some investors to go active, attracted by the existence of opportunities to exploit. The equilibrium mix between active and passive should sit in between these two extremes. It will probably be similar to the equilibrium described by GS, with active investors receiving some compensation for their cost and effort. The markets may currently be working towards such an equilibrium, having started from a 100% active approach some decades ago.
For the above reasons, I am personally attracted to the GS perspective, whereas my reaction to Sharpe’s proposition is one of suspicion. Let’s look at some of the evidence.
4. Berk and van Binsbergen show how to do the sums, and Leippold and Rueegg follow up
While total returns for ALL active investors are hard to observe, evidence exists on aggregated fund returns. I am going to start with the work of Berk and van Binsbergen (2015, 2016)Berk, J.B. and Van Binsbergen, J.H., 2015. Measuring skill in the mutual fund industry. Journal of Financial Economics, 118(1), pp.1–20. Berk, J.B. and van Binsbergen, J.H., 2016. Active managers are skilled: On average, they add more than $3 million per year. Journal of Portfolio Management, 42(2), p.131–139., rather than the more traditional investigations into active returns earned by the ‘typical’ manager. Berk and van Binsbergen (BvB) separate out the question of whether active managers add value from the question of what net active returns are received by investors. They argue that the net returns earned by investors reflect how the market clears for fund management services, i.e. who extracts the value-add associated with any skill (or, from the GS perspective, the return from investing in information). Recall that under Sharpe’s proposition, active investors as a group should generate zero gross active returns but deliver negative net returns. Under GS’s equilibrium, one would expect to see positive gross returns and net returns of around zero. What BvB find for US equity mutual funds appears much closer to a GS equilibrium than Sharpe’s proposition.
BvB estimates that the gross value-add versus the index over the period 1977 to 2011 is positive in dollar terms, averaging about $3.2 million per year per fund. This seems inconsistent with Sharpe’s proposition, which tags active management as a zero-sum game in gross terms. However, it does not necessarily imply that the zero-sum game notion is wrong. It could be that active US mutual funds are generating returns at the expense of other investors not included in BvB’s calculations. Given that the US equity market is dominated by institutional investors, it is not immediately clear who is the losing group of active investors in this case. Other institutions such as hedge funds, or asset owners that invest directly rather than outsourcing, are typically well-resourced professional investors. The evidence on hedge funds suggests they do even better than mutual funds (see Agarwal, Mullally and Naik, 2015Agarwal, V. Mullally, K.A. and Naik, N.Y. 2015. The economics and finance of hedge funds: A review of the academic literature. Foundations and Trends in Finance, 10(1), pp.1–111.). Fund managers could be extracting substantial value-add from private investors, given that there is evidence that private investors tend to underperform (e.g. Barber and Odean, 2013)Barber, B.M. and Odean, T. 2013. The behavior of individual investors. Handbook of the Economics of Finance, 2, pp.1533–1570).. Perhaps, but this seems unlikely given that private investors are a limited pool.
When examining net returns after fees, BvB find outperformance based on a simple average, but underperformance weighted by assets. The negative asset-weighted result is also a relevant result in the current context, and it is not fully consistent with a GS equilibrium which predicts it should be zero. But then again, the net returns are not statistically significant. BvB (2016, p138)Berk, J.B. and van Binsbergen, J.H., 2016. Active managers are skilled: On average, they add more than $3 million per year. Journal of Portfolio Management, 42(2), p.131–139. comment on their net return results as follows:
“A positive net alpha implies that capital markets are not competitive. A negative net alpha implies that some investors are irrational in that they are committing too much money to active management. We find that the average net alpha to investors is statistically indistinguishable from zero. Thus we cannot reject the hypothesis that investors are rational and capital markets are fully competitive. As a result, managers are able to capture all economic rents their skills provide.”
The broad message arising from BvB is that active managers possess ‘skill’ and that those with greater skill accrue greater AUM and earn more fees in aggregate dollar terms. In equilibrium, the value that managers create is captured for themselves via the fees. Meanwhile, the average investor in US equity mutual funds is not significantly worse off for investing actively relative to investing passively. Overall, what BvB describe approaches a GS equilibrium where the fees paid for manager insight is tantamount to the cost of seeking information, and managers are capturing the value arising from generating that information.
BvB only examines US equity mutual funds. In a recent paper, Leippold and Rueegg (2019)Leippold, M. and Rueegg, R., 2019. How rational and competitive is the market for mutual funds? Review of Finance, 1–35 (in press). examine aggregated performance for all equity and fixed income categories within the global Morningstar mutual fund database. Leippold and Rueegg (LR)Leippold, M. and Rueegg, R., 2019. How rational and competitive is the market for mutual funds? Review of Finance, 1–35 (in press). compare the performance of active funds versus passive funds within each category, including returns both equally-weighted and asset-weighted, before-fees (gross) and after fees (net), and for institutional versus retail funds. While the results vary across categories, the typical finding is that active funds outperform passive funds in aggregate in gross terms, but are closer to line ball in net terms. They explicitly comment that only three categories (out of 63 categories, as listed in their appendix) reveal results that are inconsistent with GS and hence consistent with Sharpe’s proposition, including US Equity Large Cap Blend, Canada Fixed Income and Euro Fixed Income. A key summary chart from their paper is copied below, although bear in mind this aggregates across fund categories. In this chart, only the aggregate results for retail fixed-income accord with what might be expected under Sharpe’s proposition. The other series plotted in accord with a GS equilibrium where active funds generate positive gross returns. LR also has a lot to say about how active returns vary across categories, which I will draw on further below.
Source: Leippold and Reuugg (2019, p13)Leippold, M. and Rueegg, R., 2019. How rational and competitive is the market for mutual funds? Review of Finance, 1–35 (in press).
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index
5. What about all the evidence that ‘active management underperforms’?
There is a general perception that there exists a strong body of evidence that active managers do not create value for investors, and in fact, largely destroy value on average. I argue here that this view is too harsh, and in any case, does not in itself provide proof of Sharpe’s proposition. As there are a number of dimensions to the story, I will take you on a journey. I start by outlining some aspects of the academic literature for context, then drill down into some of the key issues, before circling back to provide an overall summary.
Aspects of the academic research on active management
A consistent conclusion drawn from the research is that the average US equity mutual fund has not delivered positive active returns to investors. For example, Jones and Wermers (2011)Jones, R.C. and Wermers, R., 2011. Active management in mostly efficient markets. Financial Analysts Journal, 67(6), pp.29–45. state in their review article:
“Following Jensen’s seminal study (1968), numerous studies have reached virtually the same conclusion: The average actively managed mutual fund does not capture alpha, net of fees and expenses.” Jones, R.C. and Wermers, R., 2011. Active management in mostly efficient markets. Financial Analysts Journal, 67(6), pp.29–45.Jones and Wermers (2011)
This result is sometimes taken as clear evidence that active management should be avoided, often in conjunction with an appeal to Sharpe’s proposition. This is overreach. A number of aspects of this body of research mean that it is wrong to draw the conclusion that it implies active management should always be avoided. Nor should this research be taken as proof of Sharpe’s proposition. The finding that the average manager underperforms after fees arises in the following context:
- Returns are typically measured after deducting fees as appearing in the database, the average of which can be heavily influenced by the higher ‘rack rate’ fees paid by retail investors. This skirts two issues:
- What gross returns were delivered: the main test of Sharpe’s proposition
- If there are positive gross returns, whether the net return could be positive for investors who pay lower fees than those who pay higher fees
- The bulk of findings relate to US equity mutual funds, which comprise a majority of retail funds along with some institutional funds offered as pooled mandates to larger investors. Any finding based on this data need not generalize to other areas, such as:
- Institutional funds that are managed under segregated mandates (i.e. separately managed accounts)
- Other equity markets
- Other asset classes
- The research often tends to focus on the average or median returns across a manager sample.
