Four models of economic agents (Pettit, 1995)

In one of this week’s readings, Pettit begins by identifying four different theses of egocentricity found within economic models. These make different assumptions regarding the content of what human beings desire.

1. Self-centredness.
2. Weak non-tuism.
3. Strong non-tuism.
4. Self-regardingness.

Weak non-tuism

Four economic theses

The explanatory value of these different economic models ultimately depends on how well they capture real human behaviour.

Question: Do they?

Answer: Sometimes, but not always.

The ultimatum game

Definition

Two players, the Proposer and the Receiver, have to decide how to share a certain amount of money $M$ as follows:

1. The Proposer is given $M$ and can offer any amount $0, 1, \dots, M$ to the Receiver. Call this offer $\Omega$.
2. The Receiver can Accept or Reject the offer.
   • If the offer is accepted, the Receiver gets $\Omega$ and the Proposer gets $M-\Omega$.
   • If the offer is rejected, both players get $0$.

The ultimatum game

The standard solution

What should rational players do?

Assuming that something is better than nothing, the Proposer should offer the Receiver the smallest positive amount.

The Receiver should then accept this offer.

The ultimatum game

What people actually do

The ultimatum game

What people actually do

The same subjects played the game a second time after having a week’s break.

The offers were less generous, but still inconsistent with the game-theoretic analysis:

• In the second experiment, the mean offer was 37% and only two people offered an even split.
• There was only one offer of less than 1DM, and it was rejected.
• Three offers of 1DM were rejected, as was one offer of 3DM.

Reactions to the original findings

One common reaction was that there wasn’t enough money at stake. If there was “real money” on the table, it was claimed, people would behave differently.

To test this, Hoffman et al. (1996) ran an ultimatum game with monetary stakes of $M=\$100.$ Two different treatments were tested:

1. Random assignment of people to Proposer/Receiver roles.
2. A contest/exchange treatment, where subjects first answered 10 questions about current events. The subject ranked #1 (Proposer) was paired with the subject ranked #7 (Receiver); the subject ranked #2 (Proposer) was paired with the subject ranked #8 (Receiver), etc.

Hoffman et al. (1996) experiment results

Random assignment

Contest/exchange treatment

Raising the stakes even further

One hundred dollars is a decent amount of money (or it was back in 1996), but was it enough to “really matter”?

Slonim and Roth (1998) ran an experiment in the Slovak Republic, with variation in stake size ranging from 60 Slovak Crowns (SK), to 300SK, to 1,500SK.

With an average hourly wage of 24SK, the amount on the table varied between 2.5 and 62.5 hours of wages.

Slonim and Roth (1998) results

Plot of offers and rejections; a rectangle of no height indicates no offers of that type in that period were rejected.

Numerical data for the figure to the left. Subjects were given 1,000 tokens, which were convertible into cash at the end of the experiment, with possible offers in 5 token increments.

Raising the stakes really high

Just to be sure that subject behaviour was really robust in high stakes games, Cameron (1999) reports an experiment conducted in Indonesia in 1994, where:

[T]he per capita gross domestic product (US$670) was less than 3% of that in the United States. Conducting the experiments in Indonesia made it possible to increase the implied monetary stakes to a level much greater than that of previous experiments. The largest stakes used were approximately three times the average monthly expenditure of participants. Cameron, pg. 47, 1999

How widespread are the findings?

What the experiments revealed

The underlying philosophical problem

The self-interested nature assumed by economics does not fit with how we understand our motivations and reasoning in ordinary life.

We generally accept that we, and others, act in ways that “transcend and occasionally confound the calls of self-interest” (Pettit, pg. 315, 1995).

Here, Pettit means that we feel our motivation to act often derives from higher motives and abstract concerns, sometimes perhaps even the fact that other people have interests, and not that we have an interest in helping other people achieve their interests.

The underlying philosophical problem

Pettit articulates the point quite forcefully, as follows:

The economic mind is that of a relatively self-regarding creature. But the mind that people display towards one another in most social settings, the mind that is articulated in common conceptions of how ordinary folk are moved, is saturated with concerns that dramatically transcend the boundaries of the self. So how, if at all, can the economic mind be reconciled with the common-or-garden mind? Pettit, pg. 316, 1995

A diagnosis of what might be going on

One suggestion is that, although most reasoning takes place as it seems to us, the economic mind is still implicitly present.

The challenge is to make sense of what this would mean. Pettit considers two possible interpretations:

1. A focal/peripheral interpretation.
2. A virtual/actual interpretation.

Let’s consider each of these in turn.

The focal/peripheral interpretation

Painting by Simon Stålenhag.

The focal/peripheral interpretation

Advantages of the interpretation

The model certainly gives us a picture of what it might mean to say that implicitly people are economically minded. It would mean that even as people pay attention to the sorts of concerns engaged in ordinary exchanges with others, even as they keep their eyes on the needs of a friend, the job that has to be done, the requirements of fairness, they in variably conduct some peripheral scanning of what their own advantage dictates that they should do. Pettit, pg 317, 1995

The focal/peripheral interpretation

What it gets wrong:

We all admit […] that weakness of will and self deception are pretty commonplace phenomena. But what the focal-peripheral model would suggest is that the whole of human life is shot through with this sort of failure: that what we take to be a more or less occasional, more or less localised, sort of pathology actually represents the normal, healthy state of the human organism. That is a fairly outrageous claim. Pettit, pg. 318, 1995

The virtual/actual interpretation (allegedly) provides a better answer.

The virtual/actual interpretation

Consider the following question:

Europe has more than 10,873,274 people.

Do you believe that? Yes, of course. 😁 👍 🥳 🎉

Did you ever consider that specific proposition before in your life?

No.

So, could we then say that you didn't really believe that proposition before I asked you right now? 😳 😬 😩

That’s what the virtual/actual distinction is supposed to address.

The virtual/actual interpretation

A virtual belief exists when a person would readily assent to the proposition, if they entertained it.

If a person would not readily assent, if they entertained it, but would assent upon further reflection, then we are more likely to characterise this as a process of belief formation, rather than actualising a virtual belief.

The virtual/actual interpretation

So how does this help?

Homo economicus, according to Pettit, is virtually real in the sense that we are virtually self-regarding agents.

Most of the time, these two models will be behaviourally equivalent.

But when our self-interest becomes threatened, we flip into homo economicus mode and become actual self-regarding agents.

Conclusion

And so we have the following solution:

The common-sense viewpoint is valid to the extent that ordinary folk manage their affairs most of the time without adverting to their own interests; they are guided in their decisions by what is required of them under the cultural framing of the situations in which they find themselves. The economic viewpoint is valid to the extent that even when this is so, even when people are not explicitly self-regarding in their deliberations, still self-regard has a virtual presence; it is there, ready to affect what people do, in the event that any of the alarm bells of self interest ring. Pettit, pg. 323, 1995