Saturday, April 24, 2010

Trading Strategies and Market Efficiency

Here is David Merkel's tenth rule:
The more entities manage for total return, the more unstable the financial system becomes.

The shorter the performance horizon, the more volatile the market becomes, and the more index-like managers become. This is not a contradiction, because volatile markets initially force out those would bring stability, until things are dramatically out of whack...


Momentum persists in the short run, so play it if you must, remembering that the market gets more volatile when many play momentum.
This is an important point with some very interesting implications for market efficiency and volatility clustering.

In a well functioning market, asset prices are supposed to reflect the best available information about anticipated earnings flows and the risk-sensitive rates at which these should be capitalized. But the only way in which such information can come to be reflected in prices is through the trading activity of those who are alert to changes in information. A market dominated by such "information traders" will tend to be stable in the sense that prices will reliably track changes in information about underlying asset values.

Such adjustments may be rapid but they are never instantaneous. This means that in a market that is functioning well, price movements (and other market data) can reveal some information about changes in underlying values to those who have not incurred the resource costs of acquiring the information directly. This, in turn, can make technical analysis profitable.

On the other hand, in a market dominated by technical analysis, changes in prices and other market data will be less reliable indicators of changes in information regarding underlying asset values. The possibility then arises of market instability, as individuals respond to price changes as if they were informative when in fact they arise from mutually amplifying responses to noise.

But if market stability depends on the composition of trading strategies, what determines this composition itself? A key determinant is past performance: strategies that have recently given rise to strong returns will come to represent a greater share of total trading volume both because wealth has been transferred to those executing such strategies, and because they will attract new funds. In stable markets with informative price movements, the incidence of technical analysis will grow. If it grows too much, the market will be destabilized and an asset price bubble could form. Information based strategies will initially suffer from this, but those with the confidence and liquidity to persist in using them will prosper eventually once the inevitable correction arrives.

In other words, one ought not to expect markets to be efficient or inefficient, but rather to experience periods of relative efficiency that are interrupted from time to time by severe disruptions. This is a phenomenon I described in a 1996 paper as endogenous regime switching:
Behavior conducive to stability... may be most profitable when the market is unstable, and behavior conducive to instability... may be rather lucrative in a stable market when it is costly to collect information about fundamentals. There is a sense in which the two groups of speculators enjoy a symbiotic relationship with each other: each group benefits from an increase in the numbers of the other.... As a result, one would expect heterogeneity of practices to persist in the long run, with the dominance of one set of practices giving way to that of the other. The implication for speculative markets is that periods of tranquility are likely to be punctuated from time to time by periods of excessive volatility.
This paper builds on work by Beja and Goldman and Carl Chiarella, who had earlier examined the price implications of heterogeneous trading strategies without allowing for endogenous changes in population composition. The model itself is simple and stylized, but I believe that the basic idea underlying it is sound. Destabilizing strategies will prosper and spread when they are sufficiently rare, giving rise at some point to market instability, and the eventual return of stabilizing strategies. Neither perfect efficiency nor drastic inefficiency can last indefinitely, as each regime gives rise to changes in behavior that serve to undermine its own existence.

3 comments:

  1. On the other hand, in a market dominated by technical analysis, changes in prices and other market data will be less reliable indicators of changes in information regarding underlying asset values.

    This begs the question: how much information is transmitted by traders making non-technical decisions? Also, in a market dominated by technical traders, fundamental analysis is more profitable since spreads are lower and liquidity is higher. If XYZ Co. trades at $20 and you think it's worth $25, the returns from this trade will rise significantly if the bid/ask spread drops from $1 to $.05.

    Your model is interesting, but I wonder what happens when you add macro traders into the mix. Their relationship to fundamental traders is analogous to the relationship between fundamental traders and technicians: when fundamental traders trade, they make the market incrementally more efficient, but they are likely to miss the huge economic changes that invalidate their analysis. From that perspective, 2007 was the peak of a 'value bubble', in which people were too confident in discounted cash flows and asset values, and insufficiently careful about global capital flows and natural resource prices.

    ReplyDelete
  2. Hmmm... That's interesting. Where did you first hear about this? Do you have other blog posts I can take a look at?

    Justin Davis
    MK Partners
    Workplace Efficiency Experts

    ReplyDelete
  3. Byrne, you're right about the liquidity effect but it works both ways - orders based on fundamental analysis also lower bid/ask spreads faced by other speculators.

    Justin, if you type "efficient markets" (with quotation marks) in the sidebar search box you'll see a few other related posts and links.

    ReplyDelete