JF Sept. 16, 1996 A.A.S.

How can Skinner's categories of reinforcers be put into economic terms?

Consider a boss concerned with labor slacking (substandard work).

1. Reinforcers are feedback from the environment that increase the probability of the desired behavior.

2. Punishments are feedback that leads to suppression.

P- Remove a previous positive reinforcer.

eg. A $ bonus is withdrawn, former baseline pay is reduced for slacking.

Skinner's research results generally support the idea that R+ positive reinforcement works best in the sense that it is more reliable and produces fewer undesirable counter behaviors. The R- increases the probability of the desired behavior, but less reliably. In neoclassical econ. a $ increase in baseline wages is the same as the removal of a $ fine since both increase net wages by the same amount. But Skinner suggests that we inquire into the probability of the conditioned behavior and expects that the effect is not the same.

Some economists believe that the threat of job or income loss is sufficient to prevent labor slacking. They then are not much interested in labor relations of the sort that emphasizes praise or labor involvement in work design ala Herbert Gintis or Robert Lane. But Skinner suggests that R- is not as strong as R+.

Skinner finds that punishment is widely practiced because it often gets an immediate improvement in the desired behavior. But over time it leads to deviant behavior because it suppresses the previous behavior but does not replace (shall we say internalize? or volunteered) it. Punishment for slacking often leads the subject to search for ways to avoid detection rather than to reduce slacking. It may lead to defiance, sabotage, or withdrawal.

Again it is possible to view P- and P+ as equal in effect if equal in $ amount. For example a $ bonus withdrawn (perhaps expected promotion withheld) has the same effect on net income as the addition of a $ fine. But Skinner would ask us to inquire into the effect of this on behavior rather than assume additivity.

We are often interested in stopping some behavior. The boss (or teacher) says stop slacking or else you will be fired (receive a bad grade). This threat is a negative reinforcer (or punishment). Skinner calls this aversive control and generally finds that it is less effective than positive and immediate reinforcement of the desired behavior (eg. praise for the non-slacking). Note that more pay for non-slacking is not a good option as it would seem to reward slacking for awhile which would demoralize the non-slacker co-workers. Some non-monetary reinforcement seems the only way to go.


Some additional applications of Skinner's ideas:

Price does not change the demand curve in some theories, but it does in Skinner & behavioral econ. The phenomenon can be observed in behavior that Elster calls "sour grapes".

The attempt to recover sunk costs is often reinforced in spite of economists recommendation to ignore them. Perhaps sunk costs leaves chemical record (brain searches for rationalization) which can't be erased simply by the logic of maximization by the planner brain.

Some economic policy is based on R- (or is it Punishment P+ ?).

If you want to decrease a Behavior, raise its cost. eg. drug law fines and prison. The variable identified by standard theory is price. Does this assume people are calculating future payoffs when they act? What is the empirical evidence? Are our rising expenditures for prisons paying off?

How conceptualize R+ like advertising which appears to change preferences? Does advertising work by helping you reason that conception of the good will bring future satisfaction or by helping you see that the product fits?

Savings behavior:

In some theories, the only variable is the interest rate (prices)

Skinner and Akerlof --Suggest how saving enhances self control.

Are public pension plans additive to personal savings decisions? The Chicago School regards public plans as meddlesome.

Economists say U.S. saves too little to compete with Japanese. Do we have any more advice than exhortation (or subsidies)? Some advocate tax break IRA's and capital gains tax decrease. The emipirical evidence is mixed.

It is partly P.D. Doesn't help individual if others don't also increase savings. Again we have Arrow's point about the fact that is is hard to be rational in an irrational world made by others actions.

\skinner.htm