Friday, December 17, 2004

Where do inefficiencies exist? Where imperfections exist!

Efficiency does not mean perfection!

This came up in class the other day, so I figured even though the Ofek, Richardson, and Whitelaw paper itself has been around for a while (it has been accepted but has yet to be published at the JFE), I would do a quick recap of the discussion and add the links to the actual paper in question.

We were speaking about market efficiency, and suggesting that if markets were not efficient, the inefficiencies could be from two main sources: 1. True irrationality or 2. market imperfections such as a lack of liquidity, short sales constraints, poor information, etc.

While acknowledging group irrationality can develop, its existence is (IMO) quite rare and a function of a lack of differing views. Hence the more participants in a market, the less likely group irrationality will develop.

Market imperfections can be broadly classified as transactions costs (lack of information, a lack of liquidity, high trading costs, short sale restrictions etc). These imperfections allow any temporary mispricing to continue by raising the cost of arbitraging the errors away. Additionally, to the degree that the imperfections shrink the size of the market, transaction costs also help to allow the group “irrationality” mentioned above to both develop and to persist.

It should be noted that the existence of these conditions does not guarantee inefficiencies. And even if the conditions and inefficiencies do exist, it very well may be that the market is still better than any other allocation mechanism. (So relax, I am still a believer in largely efficient, albeit not perfect, markets.)

Rather than consider this question from the emotionally charged debate on market efficiency, consider the question to be: “if I had to look for inefficiencies, this is where I would look.”

In a forthcoming JFE article, Ofek, Richardson and Whitelaw investigate the idea that imperfections may limit arbitrage in conjunction with the put-call parity relationship. (Put call parity states that the put + stock = call + PV (strike) (remember: all positive at Park and Shop ;) ).

This put-call relationship should always hold in equilibrium. The authors find that it does not hold in the presence of short sale restrictions.

In their words:
…consistent with the theory of limited arbitrage, we find that the violations of the put-call parity no-arbitrage restriction are asymmetric in the direction of short sales restrictions. These violations persist even after incorporating shorting costs and/or extreme assumptions about transactions costs (i.e., all options transactions take place at ask and bid prices).

Need more “evidence”? I will ignore the preponderance of existing literature on the topic (ask if you want some more) and give you some “real time” research. I am currently working on a paper with Jonathan Godbey and Rodney Paul that looks at this idea along different lines. We look at the predictive ability of implied volatilities of equity options. While the paper is still a ways off (hopefully it will be done in January), preliminary results suggest fairly convincingly that the efficiency of the option market (as measured by the ability to forecast future volatility) is strongly tied to option liquidity. That is, for active options, implied volatility works well, for illiquid options, it does not.

So what this all mean? I will give the same conclusion that ended our class discussion:

“From a societal point of view, we should strive to reduce market imperfections and lower transactions costs: Reduce barriers to entry, end short sale restrictions, increase competition. Why? Because where there are significant imperfections, the market prices we see are less likely to be “correct” and consequently allocational efficiency is suspect.

...Are markets perfect? No. Are they pretty good? Yes. Are they better in some incidences than others? Yes. It is this last area where we may be able to do something. As an investor (and not just in financial securities), look for areas where you have a sustainable advantage. It is unlikely, although not impossible, that this is in the area of financial markets. In the financial markets, you can earn a very good return, but it is unlikely that actually beat the market on a risk adjusted basis. "

The Ofek, Richardson and Whitelaw paper is available at:
Forthcoming JFE version remember it will be taken down when the paper goes to print
FEN-SSRN (working copy version)


Anonymous said...

Little slow on responding to this; like you, I'm just now catching my breath from grading finals, etc. Glad to have it finally over!

I was a little confused regarding your separation of ineffieciencies as coming from the two possible sources of irrationality and market imperfection. Are you suggesting that either is sufficient for inefficiency, but they're not both necessary? Back when I took Nick Barberis' behav fin class at Chicago, Lecture #1 was limits to arbitrage... basically, that we wouldn't have too much to say in the behavioral literature if irrationality could be arbitraged away. It's precisely because there are so many limits to arbitrage (and I take Ofek et al as additional evidence in that literature) that anomalies are allowed to persist. I think your point was consistent with this perspective, but I wasn't sure.

As Barberis and Thaler state in their excellent review chapter, limits to arbitrage is one essential building block of behavioral finance. Investor psychology is the other building block, which brings me to my second comment:

I wasn't sure what you had in mind when you talk about "group irrationality". Are you specifically thinking of a mob-type mentality in which everyone gets swept into believing the same wrong assumptions (as many believe happened with the 90's internet bubble)? Although I do think this happens, I agree with you that it doesn't happen all that often. The types of psychological biases that drive many of the behav fin anomalies, however, have more to do with individual-level errors than a group mentality. What causes the problem is that the bias often works in the same direction for everyone, so those individual-level errors fail to cancel each other out. Examples might include overreaction (individuals expect trends to continue) and disposition effect (individuals hate taking losses). If we are all making these same individual-level mistakes, than it doesn't really matter how many of us are in the market.

Overall, I agree with you that 1) we should be searching out these limits to arbitrage if we want to make markets more efficient, and 2) trying to beat the market is a tough way to make a living. But I'm not ready to admit that markets are not heavily influenced by psychological biases. As Barberis & Thaler point out, just because there's no free lunch doesn't mean that prices are right.


FinanceProfessor said...

Great comments!

I agree that I may have been too strong in merely assigning two forces for inefficiencies but am am a bit less willing to admit the degree of importance to place on psychological biases. I have no doubt that they exist and even that they may play key roles at times, but I am not willing to say that they "heavily influence" prices. Again as you say, in the absence of limits to arbitrage.

I whole-heartedly agree with your classification of this paper into this "limits to arbitrage" school of literature.

Your point that individual biases play a role through their accumulation is true, but I believe (maybe hope is a better word :) ) that if we remove the barriers to arbitrage even many of these will fall by the wayside.

For instance, an example. We will go with the idea that people do not like to admit misatkes and hence will not sell their losers.

Suppose that the stock price has fallen and a large majority of the shareholders in the stock own it at a price significnatly higher than it is now. It then follows that these shareholders are less willing to sell and at some level the stock price is now overvalued.

At this point, IF short selling is allowed, outside investors should short the shares. Obviously (and I will never admit to anything near market perfection), there may be limits on shorting (including some that may be psycholically based themselves--See Wisdom of Crowds for a nice recap of this).

Thus, I would agree with your view that you need the limits of arbitrage to keep the prices incorrect.

As for definition of group mentality, yes mob mentality, but it does not even need to be that severe. Rather just a lack of diversity on an issue. If everyone in the world owned the stock and they were all reluctant to sell because it had gone down, then this lack of thought diversity could lead to mispricing. (gee I really was influenced by the Wisdom of Crowds...if you have not read it, I highly recommend it--here is what I wrote on another of my blogs about it:

As you say:

>with individual-level errors than a group mentality. What causes the problem >is that the bias often works in the same direction for everyone, so those >individual-level errors fail to cancel each other out. Examples might include

this would be my lack of diverse thought...and if you have that, then you will get mispricing.

>overreaction (individuals expect trends to continue) and disposition effect
>(individuals hate taking losses). If we are all making these same individual->level mistakes, than it doesn't really matter how many of us are in the >market.

tue...again see above...

BTW I love the line:

>As Barberis & Thaler point out, just because there's no free lunch doesn't mean that prices are right.



Anonymous said...

Jim -

Thanks so much for responding to my comments. I've been enjoying your blog for several months now and it's nice to get the chance to indulge in some dialog!

I do know the Wisdom of Crowds book - haven't read it all the way through, but some good friends (Soll and Larrick) published some of the original work on how crowds can be "more right" than individuals, which Surowiecki borrows from. I'm trained in behavioral science at UChicago (Thaler was my advisor) and so I'm a bit sensitive to the differences between group behavior (as in Surowiecki's book) and biases in individual judgments. I certainly think both types show up in behavioral finance anomolies, but I find the individual biases more interesting and possibly harder to arbitrage away. For example, it wasn't all that hard during the 90's internet bubble to step back and see it as a bubble, based on some type of mob mentality about putting your money into .com stocks. Limits to arbitrage is the best explanation for why that bubble was able to persist. But when the bias is at the individual level it seems harder to spot. Continuing with the disposition effect example, the arbitrageurs may be willing to short the stock if they know that a large proportion of the investors are stubbornly holding onto losing positions and overvaluing it, but that's a difficult thing to measure without knowing exactly when each person bought in relative to a long history of past prices.

In a sense, this difficulty in spotting exactly where individual-level biases are causing mispricing is, to me, part of why there is no free lunch. It's also a problem for the research itself. Although lots of researchers believed in the disposition effect, it wasn't all that possible to measure it in the aggregated market level data; it wasn't until Terry Odean got ahold of the individual level data that he was able to prove it existed. There's plenty of other work done at the individual level (on both investors and analysts) that is hard to roll up to the market level. I'd even venture to suggest that myopic loss aversion falls into this category - I have no doubt that investors fall prey to the behavior, but proving that it's the main cause of the equity premium is not such an easy argument.

Thanks again for the comments, and for such a great website!