A meditation on fairness

January 07, 2020

Let’s again consider the question I started with yesterday: “Is the economy fair to women?”

Before we can come up with an answer, we will need to define these things:

  • What do we mean by “the economy?”

  • What do we mean by “fair?”

  • What do we mean by “women?”

None of these are easy, but let’s start with fair. Why?

My blog. My choice.

Deciding fairness

There are two problems. The first: what is it we will measure as our proxy for fairness.

The second: what characteristic of our proxy measurement tells us that women (or men) are being treated unfairly?

The ultimatum game

Researchers have used the Ultimatum game to explore human ideas of fairness.

One player, the proposer, is endowed with a sum of money. The proposer is tasked with splitting it with another player, the responder. Once the proposer communicates their decision, the responder may accept it or reject it. If the responder accepts, the money is split per the proposal; if the responder rejects, both players receive nothing. Both players know in advance the consequences of the responder accepting or rejecting the offer.

In experiments in Western societies, proposers usually make equitable offers: the modal offering is 50%—which means that’s the offer that most people make. But some offer less. The mean ranges down to around 40%, and the minimum is almost always above 20% (and often rejected at that level.) But in other cultures the offers are different. In one study, people made offers above 50%; in another offers ranged down to 20%, and responders almost always accepted.

People can explain their reasoning, and their explanations are centered on fairness. People know what offer they would reject as unfair. They expect others to reject unfair offers. But they don’t know what others will judge unfair. So they tend to make offers that are likely to be deemed fair—or at least acceptable.

Researchers have studied the ultimatum game with chimps and bonobos, who behave in ways that are similar to the behavior of humans. At least in the ultimatum game, there’s a notion of what’s fair, even if people and chimps don’t know what others will think is fair.

What is fair?

One individual might consider that 40% is a fair offer (“after all, they could have offered 10%”). Another might consider anything less than 50% unfair. Individual people demonstrably have different ideas of fairness. So how might we develop a more objective definition?

We might collect some data: “At what percentage does an offer change from fair to unfair?” We’d get a range of responses. Should we define it to be the mean? The minimum? The mode? One standard deviation below the mean? Two? Three?

It seems a bit hopeless. But there is some sunshine.

Given our data, we can determine what everyone who has been polled will agree is fair, and what everyone who has been polled will agree is unfair.

And given that two people are trying to agree, they might both accept definitions of fairness and unfairness that are less demanding than the entire population.

So back to deciding whether the economy is fair to women?

Suppose we agreed that we’d use pay to determine fairness. Suppose we examined the pay of men and women doing the same job in the same environment. Suppose we found that men were paid twice what the women were paid (or vice versa.) We’d conclude there was probably something unfair about that.

Suppose the difference was 1%? Or 3%? Or 5%? Or 10%? When would a pay difference be strong evidence of unfairness? What pay difference might be acceptable?

And how could we come up with a more reasonable way to set the bounds for fair/unfair than “I think 3% is OK?” Versus “I think 3% sucks?”

​