The point of tit-for-tatis to deliver a limitedpunishment for defections from
cooperation. If the competitor cuts price one period, the firm cuts its price
next period. But if and when the competitor returns to a high price, the firm
returns to high prices too. As with the punitive strategies, the mutual play of tit-
for-tat supports a cooperative high-price equilibrium. With both using tit-for-
tat, the firms cooperate indefinitely. Neither can gain by a unilateral defection;
a one-period gain is not worth triggering an ongoing cycle of defections.
The mutual play of tit-for-tat, or of the punitive strategy, succeeds in sup-
porting a cooperative equilibrium. But these are only two of an endless num-
ber of possible contingent strategies. Not surprisingly, there has been
considerable research interest in strategies for playing the repeated prisoner’s
dilemma. An intriguing result of this research is how well tit-for-tat performs in
achieving cooperation. Tit-for-tat has four virtues. First, it is nice;it is never the
first to defect. Second, it is retaliatory; it immediately punishes an unwarranted
defection. Third, it is clear; a competitor can immediately see that it doesn’t
pay to mess with tit-for-tat. Fourth, it is forgiving; by mimicking the competitor’s
previous move, it always is ready to return to cooperation. This last feature is
the big difference between the punitive strategy (which satisfies the first three
features) and tit-for-tat.^10OTHER ASPECTS OF REPUTATION We have seen that a repeated game allows
a player to create and maintain a reputation for cooperation. Reputation can
play an analogous role in related contexts.
As a simple example, suppose a seller can produce medium-quality goods
or high-quality goods. A typical buyer is willing to pay a premium price for a
high-quality item, and the seller could make a greater profit from delivering
high quality. The trouble is that the two types of good are indistinguishable at
the time of purchase. Only after the buyer has purchased and used the good is
the difference in quality apparent. If only a single, one-time transaction is at
stake, we can argue (without needing a payoff table) that the only equilibrium
has the seller offering medium-quality goods at a low price. Why? Because a
seller’s claim for high quality would not be credible. Any buyer who believed
the claim and paid a premium price would be exploited by a self-interested
seller who delivered medium quality instead.
Of course, you might plausibly protest that honesty should be the best
policy for the seller. This is true—providedthe seller has an incentive to estab-
lish a reputation for delivering high-quality goods now and in the indefinite
future. Here is a simple way to establish a high-quality equilibrium. If a seller424 Chapter 10 Game Theory and Competitive Strategy(^10) The political scientist Robert Axelrod has been a pioneer in investigating the repeated prisoner’s
dilemma. In a famous series of experiments, he asked economists, management scientists, and
game theorists to devise strategies to be used in repeated prisoner’s dilemmas, such as price com-
petition. What strategy performed best on average when paired in turn against all other strate-
gies? Tit-for-tat! For more on repeated competition and the features of successful strategies, see R.
Axelrod, The Evolution of Cooperation(New York: Basic Books, 1984).
c10GameTheoryandCompetitiveStrategy.qxd 9/29/11 1:33 PM Page 424