The general idea of efficiency in financial markets is that information is being processed almost instantaneously because active investors arbitrage away any arising price discrepancies.
On the other hand, more and more money is being allocated to passive investment vehicles (mainly ETFs) because of the empirical fact that active managers are not worth their fees and/or you cannot discern upfront those who will be worth their fees.
This obviously constitutes a paradox because if fewer and fewer active investors arbitrage away price discrepancies the case for passive investing begins to crumble...
I did a little googling and found lots of "opinions" on this, so the idea of this paradox as such is not new. What I did not find are any serious academic papers on the issue or some kind of consensus.
My question
To the best of your knowledge, is there some consensus on this matter and/or could you point me to good academic papers which thoroughly investigate this paradox (theoretically, by simulation or empirically)?
Answer
On more than a few occasions, I have attempted to extrapolate the current trend towards passive allocation to its logical conclusion: more passive allocation means more inefficiency.
I am not aware of any research which directly measures the correlation between market efficiency and active versus passive allocation. In general, the level of market efficiency should be hard to measure (see related topic: What are the empirical limitations to testing market efficiency?). However, I believe that the converse is indirectly measurable: pricing "anomalies" and/or risk-premia tend to dissipate following their discovery and/or the publication of a more efficient (i.e, "better") model.
If one concludes that the dispersion of knowledge regarding pricing inefficiencies has lead the market to become more efficient as market participants act on this knowledge, then one may tentatively accept the converse: the flight of capital away from active management may lead to greater inefficiency.
Examples of "alpha" diffusion and then decay abound. Notably:
- Fama-French factors have decreased in efficacy over time. In the 2004 paper, "On the Cross-Section of Expected Stock Returns: Fama-French Ten Years Later", the authors found that "using the Fama-Mac Beth two-pass regression, that the size effect becomes insignificant during the post-1981 period, and the Book/Market effect becomes insignificant during the post-1990 period."
- Following several papers on the momentum anomaly, notably with Carhart's 1994 paper, momentum's ability to predict future returns has notably diminished.
- On a recent Master's in Business podcast with Barry Ritholz, guest Ed Thorp discusses the efficacy of the Black-Scholes model following its publication in 1973. Thorp had been using equations similar to the Black-Scholes for approximately five years before the paper was published. He noted in the interview that traders were slow to latch on to more rigorous pricing mechanisms, but eventually did so.
This concept of active versus management and inefficiency might be exemplified through a thought experiment in which all market participants woke up tomorrow fully believing in the absolute truth of EMH. These converts would then be motivated by the belief that active management is futile; allocation decisions would be purely driven by the mechanistics of MPT and mean-variance optimization. Throw out risk-premia, fundamental analysis, technical analysis, everything -- the market prices already reflect all expectations of future risk and reward. Given that it is impossible for the average money manager who charges fees to out-perform the "average", the shift towards passive management is justifiable. Yet, some price discovery through active allocation is necessary for the EMH to work.
The problems of herd behavior in the context of MPT are more eloquently outlined in the 2014 paper, "Modern portfolio theory and risk management: assumptions and unintended consequences", in which the authors define some problems underlying the assumption of MPT, namely: "information asymmetry and bounded rationality; the joint hypothesis problem and random walk hypotheses."
The question then may become, "how much active capital is required to provide an efficient price discovery mechanism?". To my knowledge, this remains an open question but I believe that the threshold whereby flight from active management leads to inefficiency may still be a ways off. Currently, I believe about half of the equity money managers are active versus passive. However, much of the assets under active management are actually passively managed (vis-a-vis, "closet indexing") as measured by a fund's "active share". While this seems like a lot of passive money, the dismal performance of hedge funds over the past decade is evidence that active management is still saturated. Moreover, the rise of algorithmic finance through vast computational power, data analytics, and instantaneous communication has probably left less room for active management to exploit anomalies and inefficiencies.
No comments:
Post a Comment