operative behaviour. This wide range of possibilities creates a problem of multiplicity of
equilibria.14 In contrast to a situation were both players play simultaneously, in this game the
co-ordination problem might not be so serious. The reason is that the incumbent player has a
privileged position in the sense of being able to show a commitment to co-operate, which then
can be followed by the opposition party once in office. How can a particular level of reserves
from this feasible set be chosen as the co-operative stationary state? On what grounds an
incumbent decides to choose any particular level? A plausible answer can be given in terms of
the presence of a focal point (e.g., an amount of reserves with a salient feature). Also, by
considering the internal life of the incumbent party, the selection of a particular level can be
linked to the arrival of leadership sympathetic to co-operation.
In what follows I will look for those conditions required to make co-operation of the first type,
which is the more basic of the two. The second case will not be considered here. In the latter
situation, co-operation is based on the possibility of making sacrifices in the present in order to
lengthen the life of the game and requires far sighted parties. However, when myopia dominates,
co-operation on piling up reserves is very unlikely to emerge.15
3.1. Keeping constant the risk factor in the game
The study of the conditions for the emergence of co-operation is based on the standard
procedure that checks for the presence of incentives for defection. The focus is placed on the
Incentive Compatibility Condition (ICC), which consists of the difference between the pay-
14 The folk theorem of repeated games assures that, for a sufficiently small discount rate, any point of the
efficient frontier can be sustained as a subgame-perfect equilibrium. The good news is that a better result
than the inefficient equilibrium can be achieved; the bad news is that any point can be sustained. The
resulting multiplicity of equilibria creates a well-recognised co-ordination problem.
15 This second case is more complicated to treat formally. The analysis could be developed based on the
idea that for co-operation of this type to emerge there are two conditions that should be satisfied. In the first
place is the feasibility condition. It demands that the future benefit of co-operation of an electoral sacrifice
has to outweigh the opportunity cost in terms of re-election chances today, assuming that there is no
incentive to deviate once a greater level of reserves has been reached. In the second place is the credibility
constraint. It demands that once the level of reserves is increased, and given that there is any further
increase in reserves, the temptation for the next incumbent to deviate (whoever party it will be) cannot be
greater than the gains for continuing co-operation. The task, then, is to evaluate if there are values for the
parameters such that the build up of reserves is both feasible to be generated by the current incumbent, and
credible to be kept by the next one.