Printer Friendly

The welfare effects of disclosure under cognitive dissonance.


There has recently been interest in the economic implications of the psychological theory known as cognitive dissonance. There are three basic premises to this theory: (1) for the same information set, people are able to manipulate their beliefs so as to make them compatible with preferences; (1) (2) people will seek out sources of information that confirm or augment desired beliefs; (3) once beliefs are formed on the basis of cognitive dissonance reactions, they persist over time. In essence, people who experience cognitive dissonance reactions engage in efficient rationalization and may even deliberately choose to believe something other than the truth, if believing the truth is an unpreferred state.

In their seminar paper, Akerlof and Dickens [1982] developed a model of a hazardous industry labor market where workers and firms know the true odds of accident, but workers are prone to experience cognitive dissonance reactions. Because of cognitive dissonance, knowing the odds imposes psychic costs of fear and under certain conditions, it will be efficient to suppress fear and choose an alternative set of beliefs more compatible with preferences, e.g., the odds of accident are lower. The main point of the model was to show that when people form such cognitive disonance reactions, government intervention such as safety legislation, where workers must purchase safety legislation, may be needed to restore Pareto optimality.

One question not addressed by Akerlof and Dickens (A&D) is what are the welfare effects of disclosing new (more accurate) information when the recipients of the new experience cognitive dissonance? In previous literature (2) on the welfare effects of disclosure, it is shown that disclosure has only positive social value in a production economy. This paper examines the welfare effects of disclosure in a production economy under cognitive dissonancee. The A&D model is extended to the case of asymmetric information, where firms have better information about accident probalities than workers. It is demonstrated that under certain conditions, disclosure with and without safety legislation will reduce social welfare.

Asymmetric Information and Disclosure in

the A&D Model

The A&D model is a symmetric information model. However, it is entirely possible that there may be asymmetric information in the hazardous industry labor market. For example, firms may have better information about the odds of accident than workers because firms are typically in the industry longer. The question then arises: what will be the welfare effects of firms (or government) informing workers of the true odds?

The results of the A&D model will be briefly recapitulated. There are two periods. In each period the demand for workers in the hazardous industry exactly equals the demand for the good, because it takes one worker to produce one unit of the good. In the first period, no safety equipment is available. In the second period, a worker in the hazardous industry has the option to buy a new safety device which eliminates the possibility of accident. Let q be the true probability of accident, [w.sub.s] as the safe industry wage, [c.sub.a] the costs of accident, [c.sub.s] the costs of safety equipment and [c.sub.f] the costs of fear. The worker will choose perceived probability of accident [q.sup.*]. The level of fear experienced f is assumed to be of the form f = [q.sup.*]/q.

A worker in the hazardous industry in period 1 will choose:

[q.sup.*] = 0 if ([qc.sub.a.-c.sub.s.)] < [c.sub.s.C.sub.f./] ([qc.sub.a.+c.sub.f.)]

[q.sup.*] = [qc.sub.s./]([qc.sub.a.+c.sub.f.)] > [c.sub.s.c.sub.f./] ([qc.sub.a.+c.sub.f.)].

If [W.sub.h.\q] is the first period's reservation wage in the hazardous industry when workers know q and there is no safety legislation, then

[W.sub.h.\q] = [w.sub.s] + [qc.sub.a] + min([qc.sub.a.-c.sub.s.,] [C.sub.s.Cf/(qc.sub.a.+C.sub.f.))].

Safety legislation, requiring the purchase of safety equipment in the second period, will induce workers to always choose [q.sup.*.=0] and lower the first period reservation wage to [W.sub.s.+qc.sub.a].

In Figure I, D([W.sub.h]) is the demand curve for labor (and product). The horizontal lines are labor supply curves. [S.sub.1] is the supply curve without legislation and [S.sub.1]' is the curve with legislation, both corresponding to the A&D case. In the symmetric information case of A&D, the welfare gain to legislation is the area A'B'C'D'.

Now consider the case of asymmetric information. Suppose that each worker, prior to his choice between work in the hazardous or safe industry starts off with an underestimate of the odds of accident. Worker's estimates of risk refer to an industry rather than a firm. Define q as an estimate of the true odds of accident. Then, q < q. We assume workers have homogenous beliefs. Then, the reservation wage is [W.sub.s] + [qc.sub.a] if workers know they must purchase safety equipment in the second period. If there is no safety legislation, the wage (3) is

[W.sub.h.\q] = [W.sub.s] + [qc.sub.a] + min ([qc.sub.a.-c.sub.s.,] [c.sub.s.c.sub.f.]/9[qc.sub.a + c.sub.f])).

Both wages are lower than the wages that would prevail in the A&D model with symmetric information. These wages are shown in Figure I: the labor supply curve is [S.sub.1] in the no-legislation case and [S.sub.1]' in the legislation case.

What will be the welfare effects of disclosure? Without safety legislation, disclosure will reduce social welfare by the shaded area BB'CC' in Figure I. This loss occurs because with disclosure, cognitive dissonance reactions will be stronger than without disclosure; when workers are told the truth, they have more to fear than before and there will be greater returns to suppressing fear. Reservation wages will be higher with disclosure, hence fewer workers will be hired and less product will be demanded. What if disclosure is coupled with safety legislation? While disclosure raises the wage to [W.sub.h./q], legislation will reduce it to [W.sub.s] + [qc.sub.a]. There will be a net gain in welfare of A'BCD'.

The socially efficient arrangement is safety legislation without disclosure. With legislation, the reservation wage falls to [W.sub.s] + [qc.sub.s]. Welfare rises by ABCD, a net gain of AA'DD' compared to the case of disclosure. Thus, regardless of whether safety legislation is introduced or not, the social value of disclosure is negative.

The model can be extended to consider two additional cases. First, the social value of disclosure will be positive if all workers systematically overestimate the risk of accident (q > q). In that case, wages will be lower with disclosure. Second, if beliefs are heterogenous but unbiased (the rational expectations case), disclosure will have no social value. While this paper has focussed on the case of workers systematically underestimating the true odds of accident, beliefs of the population as a whole could be unbiased: given the existence of a safe industry and a risky industry, equilibrium sorting of workers with initially unbiased beliefs yields the case considered in the paper.

Concluding Remarks

This paper has demonstrated using the A&D model, that in a world of cognitive dissonance it is possible that disclosure of more accurate information to the public could reduce social welfare. One implication of this finding is that it will be socially efficient for firms to withhold the truth from workers and for markets to maintain (perhaps even enforce) informational asymmetry. To eliminate the asymmetry would be to reinforce cognitive dissonance reactions. If cognitive dissonance exists, then a reevaluation of the social efficiencies of disclosure laws may be warranted.

A problem with testing the implications of this model is that it may be difficult to empirically distinguish between an explanation of wages and disclosure based on cognitive dissonance from an explanation based on a conventional theory of information processing. For this reason, a case study of one firm or industry or an experimental approach, where for example reservation wages of subjects are elicited before and after disclosure, might be appropriate.

In the absence of case studies or experiments, a possible test would be to focus on the relationship between the degree to which wages respond to disclosure and the length of time worked in the firm/industry. Workers with greater tenure in the industry will be more familiar with working conditions and have more accurate estimates of the odds of accident than workers with lower tenure. It follows that wages of experienced workers will change less than inexperienced workers at the time of disclosure. Also, in those industries where information is available at lower cost, it is expected that the disclosure effect on wages will be less than in industries where information is costly.


George A. Akerlof and William T. Dickens, "The Economic Consequences of Cognitive Dissonance," American Economic Review, 72, 3, June 1982, pp. 307-19.

Jack Hirshleifer, "The Private and Social Value of Information and the Reward to Inventive Activity," American Economic Review, 61,4, September 1975, pp. 562-74.

John Marshall, "Private Incentives and Public Information," American Economic Review, 64, 3, June 1974, pp. 373-90.

David S. Ng, "Information Accuracy and Social Welfare Under Homogenous Beliefs," Journal of Financial Economics, 2, 1, March 1975, pp. 53-70.

(1) One interpretation of cognitive dissonance is that not only does utility depend on consumption states, but also on beliefs (and information sets used to form beliefs) about states.

(2) See Hirshleifer [1971], Marshall [1975] and Ng [1975].

(3) In the assymetric information case, the worker in period 1 will choose:

[q.sup.*] = 0 if ([qc.sub.a.-c.sub.s.)] < [c.sub.s..C.sub.f/] ([qc.sub.a.+c.sub.f.)]

[q.sup.*] = [qc.sub.s/] ([qc.sub.a.+c.sub.f.)] if ([qc.sub.a.-c.sub.s.)] > [c.sub.s.c.sub.f/] ([qc.sub.a.+c.subf.)]
COPYRIGHT 1991 Atlantic Economic Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Bodvarsson, Orn B.
Publication:Atlantic Economic Journal
Date:Mar 1, 1991
Previous Article:Blaug on Kuhn versus Lakatos and the marginalist revolution.
Next Article:Forward market intervention and exchange rate variability: a simulation approach.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters