INTRODUCTION

Psychological studies demonstrate that most individuals are overconfident about their own abilities, compared with others, as well as unreasonably optimistic about their futures (e.g., Taylor and Brown 1988, Weinstein 1980). When assessing their position in a distribution of peers on almost any positive trait such as driving ability or income prospects, 90% of people say they are in the top half (see Svenson 1981). Few people say they are below average, although half must be.1

This paper explores one setting in which optimistic biases could plausibly and predictably influence economic behavior: entry into competitive games or markets. Many empirical studies show that most new businesses fail within a few years. For example, using plant level data from the U.S. Census of Manufacturers spanning 1963-1982, Dunne, Roberts, and Samuelson (1988) estimated that 61.5% of all entrants exited within 5 years and 79.6% exited within 10 years. Most of these exits are failures (see also Dunne, Roberts, and Samuelson 1988, 1989a, Shapiro and Khemani 1987).

There are many possible explanations for the high rate of business failure (reviewed below). In this paper we consider the hypothesis that business failure is a result of managers acting on the optimism about relative skill they exhibit in surveys.

The thesis of this essay is that decision makers are excessively prone to treat problems as unique, neglecting both the statistics of the past and the multiple opportunities of the future. In part as a result, they are susceptible to two biases, which we label isolation errors: their forecasts of future outcomes are often anchored on plans and scenarios of success rather than on past results and are therefore overly optimistic; their evaluations of single risky prospects neglect the possibilities of pooling risks and are therefore overly timid. We argue that the balance of the two isolation errors affects the risk-taking propensities of individuals and organizations.

The cognitive analysis of risk taking that we sketch differs from the standard rational model of economics and also from managers' views of their own activities. The rational model describes business decisions as choices among gambles with financial outcomes, and assumes that managers' judgments of the odds are Bayesian, and that their choices maximize expected utility.

Introduction

This is a rather conjectural report on the evolutionary role of decision biases— at both the level of individuals and of organizations—and, in particular, on their importance to the processes of corporate entry and the evolution of industrial structures. A growing and quite robust body of evidence highlights the pervasiveness of various types of biases in individual decision making, which accounts for systematic departures from predictions of the canonical model of rational choice (see, for example, Kahneman & Tversky, 1973, 1986, Shafir & Tversky, 1992). For our purposes, we will mainly concern ourselves with overconfidence or optimism, which frequently leads to bold forecasts of the consequences of one's own actions. Also, by way of example, we will examine risk seeking in the domain of losses, which often yields escalating commitments in the face of failures. Interestingly, these biases appear to carry over from the level of individuals to that of groups and organizations and, indeed, might even be amplified in the latter circumstances (see, for example, Kahneman & Lovallo, 1993, Lovallo, 1996a, and the literature discussed there). In this respect, a challenging domain of investigation – with vast ramifications into the analyses of the nature of entrepreneurship, technological change, and industrial dynamics – is that of corporate entry into an industry.