Carlo Beretta

Making use of and following substantive rationality is a costly and not always feasible option. In games, this option generates two levels of reasoning. The player has to decide how to play the overt game, but this requires him to decide how to play an underlying game in which the alternatives are to follow either substantive rationality or a different rule, here that of reasonableness, in playing the overt game. A strategy is reasonable if coupled with its best response it results in the player that adopts it achieving the highest payoff among those dominating the Nash equilibria of the overt game. The underlying game can have more than one equilibrium. Substantive rationality in the underlying game can justify behaviour which is at odds with substantively rational behaviour in the overt game. The case of the centipede is used as an example. Key words: Common knowledge, Rationality, Backward induction, Reasonableness, Incomplete information, Centipede game. JEL Classification: C72.

Each country produces just one good, corn, by means of corn, labour and land. Capitalists allocate their funds between corn production and the purchase of land so as to equalise the rate of profit on these activities under equilibrium expectations. Autarky in natural equilibrium is briefly discussed. Opening the economy allows capitalists to invest either domestically or abroad. In the small country case, equilibrium expectations make the economy very unstable. Under most circumstances, the economy will register an inflow of capital, which can be employed productively only in the short run, if at all, and thereafter will translate in speculation in land, followed by a balance of payments crisis and collapse on this market.

Social sciences aim to give a rational explanation of the regularities which characterise the interactions observed. Most of these interactions can be described as games and in game theory rationality comes down to substantive rationality. The paper stresses that in some cases substantive rationality is not enough to explain the particular regularity one is interested in. In others, it requires a behaviour which is apparently at odds with individual rationality. One has to distinguish the rationality of having a rule from the rationality of the rule itself. Moreover, in judging the rationality of a rule, one must take into account the constraints deriving from the fact that the rule must be supported by social sanctions.