11 - An Experimental Comparison of Alternative Rules for Competitive Market Exchange

Summary

In short, the argument that we cannot experiment in the behavioral sciences because the problems are too complex is no more than a blanket rationalization of our ignorance as to what experiments to perform, and how to go about performing them

(Abraham Kaplan, 1964).

INTRODUCTION

Double oral auction trading on the New York Stock Exchange, and on most organized stock and commodity exchanges throughout the world, has evolved over more than two centuries. This evolution has consisted, in part, of a gradual formulation of rules governing the mechanics of “floor” trading. Our scientific curiosity as to why these institutional rules exist has motivated an interest in studying the effect of contracting rules on the convergence and efficiency properties of competitive markets. Does the “specialist book,” and the New York Stock Exchange trading post rule requiring admissible bids and offers to narrow the bid-ask range, have identifiable affects on market performance? Our working hypothesis is that the survival value of a rule is manifest in measures of improved market performance.

The 21 experiments reported in this chapter use computerized transformations of the oral double auction that have been programmed by Williams using the PLATO computer system's TUTOR language. Our interest in real-time electronic trading institutions is motivated in part by the gradual evolution of both private and government securities markets into systems based on electronic quotations and in part by the scientific value of the computer in providing closer control over the' procedures, and recording of data, in group exchange experiments.