Discussion: Herd Behavior and Investment by Scharfstein and Stein (1)

In a serie of posts I will discuss the paper of Scharfstein and Stein called “Herd Behavior and Investment” from 1990. It can be found here. http://scholar.google.com indicates that to this day this paper is cited 1708 times,  quite a number. I will try to explain the paper as best as I can. The paper examines some of the forces that can lead to herd behaviour in investments. Although this is inefficient behaviour from a social standpoint, it can be rational from the perspective of the managers who are concerned about their reputation.

The model is a game theoretical model that falls under the category of dynamic games of imperfect information. In other words, the players in the game move sequentially (one after the other) and have imperfect information about the possible outcomes.

To understand this paper and post you will need to be familiar with Bayes rule. This post will mostly be an introduction. In subsequent post the model will be introduced and solved.

Introduction

The classical economic theory assumes that investment decisions are made by rational agents making rational decisions. The agent will use all available information to make a decision in an efficient manner. However, these decision may also be driven by group psychology, which will disrupt rational decision making.

One example is the stock market, the paper gives the example of the US stock market in 1987, leading to the 'famous' black monday on 19 October 1987. In that time the consensus among professionals was that price levels were too high. The market thus was expected to go down. However, few professionals were eager to sell their stocks. If the market did continue to go up, they were afraid to be one of the few missing the ride. On the other hand, if the market would go down, there would be comfort in numbers. How bad could they look if everybody else had suffered the same fate? In other words, they would share the blame. 

The model that the paper develops tries to give insight in to some forces that may lead to this herd behaviour.  To simplify the analysis the model assumes that the investments have perfect elastic supply, i.e. a horizontal supply curve. This means that changes in demand will not shift the price of making an investment, Later in the paper they comment on a situation when not making this assumption.  The authors give the following example that may reflect the model:

Suppose managers are in charge of capital investments at industrial firms. They each are considering a cost-saving technology. The value of this technology is realised in the future but may differ in different states of the future; in one state the investment is profitable in another state it is not. The managers receive a signal about this future state of the world, which may help them in their investment decision.

The question is: How well does the aggregate level of investment reflect all of the information. In other words, will the managers make full use of their private signals. Or do other factors, such as reputation, also play a role.

In the next post we will make a start with the model.