Such was the case with hurricane Mitch, which hit Honduras in October 1998. Mitch had a measurable and durable impact on the lives and livelihoods of Honduran citizens. Immediately upon impact, one third of the value of that year’s crop was destroyed, and homes, infrastructure, and farming structures were also destroyed. Farm land was eroded to the point that farms would be unproductive without significant investments in soil and landscape improvement. Farmers’ investments were wiped away, leaving them with lessproductive farms and with incomes that would be permanently lower without further investment. It is estimated that 5 percent of the population was immediately reduced to poverty, increasing the poverty rate to 75 percent.

In developing countries such as Honduras, where financial markets and formal insurance might be unavailable to most, the speed of recovery depends very heavily on the ability of informal mechanisms to function. Thus, the ability of the rural poor to care for one another, often without monetary or material reward, can help speed the recovery. Michael Carter and Marco Castillo set out to determine the extent to which altruism can help speed the recovery from a disaster. They collected experimental data from 389 farm households in Honduras a few years after hurricane Mitch.

Each household, in addition to answering several questions about the damage they suffered and their recovery from Mitch, also participated in a dictator game with other households. The dictator game asked them to divide money between themselves and another household. Any money passed to the other household would be tripled. On average, households passed 42 percent to others. More impressively, how much people shared with others was directly related to the speed of the recovery. Increasing the amount of altruistic giving by 10 percent was associated with a 1 percent increase in the rate of recovery (rate at which they returned to their previous level of assets). This suggests that in communities for which altruism is the norm, recovery from natural disasters may be much more effective for all. Perhaps such observations provide us some clues as to how altruism became so prevalent in modern societies.

History and Notes

Altruism is not only a topic of interest for economists or other social scientists. Evolutionary biologists have also taken an interest in altruistic behavior. Much like economics, the study of evolution often assumes that people maximize not their utility function but their own fitness, most often defined as one’s ability to pass on one’s genes to the next generation. At first blush, it would seem that risking your own life or reducing your own sustenance would reduce fitness. However, many examples of seemingly altruistic behavior can be found in nature. For example, Gary Becker notes that baboons often risk substantial harm to protect other baboons. Nonetheless, such acts of seeming altruism can help promote fitness for the group, increasing the probability that genes are passed on for any one individual. More prominently, sociobiologists argue that altruistic motives toward children may be directly motivated by fitness. Similarly, altruistic

discussed in Chapter 14. The notion of fairness has had many different definitions in the literature, and we will cover the most important of these.

One branch of the research on fairness concerns how people react not only to the distribution of outcomes but also to the motivations and perceptions of others. Such conceptions of psychological motivations in games can provide powerful explanations of the behavior commonly observed in laboratory games. These models of behavior also have implications for the real world. Much of work and business is conducted in teams or other groups in which the actions of each individual involved will affect the payouts of all. How teams perceive the diligence of each of their individual members can have a substantial impact on how any one individual will decide on how much effort to put forth. Similarly, firms can use the way they are perceived in the community (e.g., socially conscious vs. greedy) to market their products and services. In each case, how actions are perceived can have a big impact on the behavior of others and ultimately on the profits of firms and well-being of consumers and workers.

EXAMPLE 15.1 Giving Ultimatums

The notion that people desire outcomes in which the payoffs are evenly divided has been a common explanation for the behavior observed in the dictator game. However, as you recall, these results tended to erode once the dictators knew that their choices would be completely anonymous. Thus, the dictators seemed to be partially motivated by what others might think of their actions. One wonders how the second player in the dictator game felt about his portion of the money and the dictator who gave it to him. If the second player had the option of expensive retaliation, we would be allowed to see explicitly how others react to the dictator’s actions. This is the motivation behind the ultimatum game. In the ultimatum game, two players are to divide an amount of money, say $10. Player 1 proposes a split of the money between the two players. For example, Player 1 could propose that Player 1 will receive $7 and Player 2 will receive $3. Then, Player 2 can accept the proposal, and then both players receive the amounts Player 1 proposed, or Player 2 can reject the offer and both players receive nothing.

We can solve for the subgame perfect Nash equilibrium (SPNE) of this simple game by using backward induction. Given any split proposed by Player 1, a selfish Player 2 will prefer to receive any positive amount to $0. Thus, so long as Player 1 proposes that Player 2 receive at least $0.01, then Player 2 should accept. Given this strategy by Player 2, a selfish Player 1 should always choose to take all the money aside from $0.01. If Player 1 is altruistic, he might decide to allot more to Player 2. Whether Player 2 is selfish or altruistic, he should not reject an offer that provides him at least a penny. The selfish player will take the money because it makes him better off. The altruistic player will accept it because rejecting it would make both players worse off. Rejecting any positive offer of money would hurt Player 2 in order to hurt Player 1. It is an extreme and purely destructive move.

Werner Güth, Rold Schmittberger, and Bernd Schwarze engaged 42 participants in a laboratory experiment in which each pair played the ultimatum game. Each player

participated twice, once exactly one week after the first game. However, players were not aware that the second game would take place until after they had already completed the first game, thus eliminating the chance for strategic behavior. In the first game, two of the 21 players delivering the ultimatum chose a split in which they would receive all of the money, seven decided on an even split, and only five chose a split in which Player 1 would receive more than 67 percent of the money. On average, Player 1 offered a split giving Player 1 64.9 percent of the money. Clearly, this is out of line with the SPNE. On the other side of the table, nearly all of those playing the role of Player 2 chose to accept the split; two did not. One had been offered 20 percent of the money, and one had been offered none. Aside from the Player 2 who rejected 20 percent of the money, all of those playing the role of Player 2 behaved according to the rational selfish model: accepting a positive offer and either accepting or rejecting an offer of zero. These results are consistent with the results from other experiments involving the ultimatum game, where the majority of participants tend to offer a nearly even split, with Player 2 receiving 40 percent to 50 percent.

Given one week to consider the game, players returned and were paired with different participants for a second round. This time, 11 of those playing the role of Player 1 offered a split giving themselves more than 67 percent, and one participant offered the partner a single penny. Every player offered their partner at least a penny. Again, only one player behaved according to the SPNE, though the offers were much closer to the selfish and rational offer. On average, Player 1 offered a split resulting in Player 1 receiving 69.0 percent of the money. Those playing the role of Player 2 were incensed by the relatively unfavorable offers they received. Of the 21 participants playing the role of Player 2, six decided to reject the offer, even though all offers included a positive payout for Player 2. These players received offers in which they would receive 0.200 percent, 0.167 percent, 0.200 percent, 0.250 percent, 0.250 percent or 0.429 percent. Offers of more than 43 percent of the money were all accepted. In this case, about 25 percent were willing to pay money (by giving up the amount offered to them) to keep Player 1 from receiving any money. This behavior appears to be in retaliation for offers that gave Player 2 too little money. In similar experiments we find that 40 percent to 50 percent of offers in which Player 2 would receive less than 20 percent of the money are rejected.

This game provides an interesting illustration of the puzzle of other-regarding preferences. The offers by those in the role of Player 1 clearly deviate from the selfish rule of offering only one penny. In fact, they tend to gravitate toward offering about half (or almost half) to their opponent. This, and the behavior in the dictator game discussed in Chapter 14 suggest that people are willing to give up money in order to make others better off, which is consistent with altruism. Alternatively, many playing the role of Player 2 were willing to forgo money to keep their partner from receiving money; this behavior seems diametrically opposed to altruism. Moreover, the behavior seems to be contingent on the offer Player 2 received. Players wished to punish those who had given them an unfavorable split and were willing to pay in order to administer that punishment. If the game would be played only once, there was no plausible future monetary reward for punishing Player 1.