We have selected for retrospective review a few all-time “best selling” research papers of the past few years from the General Financial Markets category of the Social Science Research Network (SSRN). Here we summarize the August 2003 paper entitled “A Theory of Large Fluctuations in Stock Market Activity” (download count nearly 2,700) by Xavier Gabaix, Parameswaran Gopikrishnan, Vasiliki Plerou and Eugene Stanley. Why do stock prices vary more than company fundamentals? Why do stock markets crash? This paper proposes a theory of large stock market movements based upon a linkage between market activity and the size distribution of large financial institutions. Motivated by empirical findings that stock returns, trading volumes, number of trades, price impacts of trades and sizes of large investors all have power law distributions, the authors propose that:
- Large volumes and large returns (“fat tails” compared to those of a normal distribution) result from the trades of large investors as they attempt to exploit stock mispricings while minimizing the effects of their own trades on prices. Optimal trading under these conditions results in power law distributions for returns, volumes and number of trades.
- The fatter the tails of the distribution of large investors, the fatter the tails of the distributions of trading volumes and returns.
- The price impact of trading varies with the square root of the volume because large players trade more slowly than do small players. Since volume moves prices even without new information (liquidity trading), prices may deviate from fundamentals, but will revert in the long term.
In summary, the power law distribution of sizes of large investors, along with the optimal trading behavior of those investors, explains the excess volatility observed in asset markets.
This paper is part of the econophysics movement, the use of tools from physics to study economic issues.
An October 2005 retitled version of this paper is available as “Institutional Investors and Stock Market Volatilty.”
