Bernardo Guimaraes and Kevin D Sheedy look at Power sharing and institutional stability. Clearly if you are creating a smaller pie then your institutions are not right. This in turn raises the question, Why inefficient institutions persist,
Institutions determine the allocation of resources and power among the individuals in the model. However, institutional choices have to survive popular uprisings, coups d'états, and all types of insurrections against them. Hence the elite in power chooses the rules that maximise its members’ payoffs subject to avoiding challenges that would dislodge them from power. For instance, excessively large taxes on some groups would prompt the disaffected groups to fight to take power and change the institutions.
In many cases, rules necessary for economic efficiency require restrictions on the untrammelled exercise of the elite's power. Including the possibility of investment in the model immediately raises this issue: rules against expropriation are needed, because protection of private property is optimal for the elite ex-ante, but not ex-post. There needs to be a means of preventing the elite from using its power to sweep aside existing rules and reshaping institutions to conform to their ex-post interests. The problem is essentially the classic question of ‘who will guard the guardians?’.
The model provides an answer: there must be many guardians if institutions are to be safe. A larger group holding power emerges as a commitment mechanism that allows them to act as a government bound to a set of policies that would otherwise be time inconsistent. This large group in power could perhaps be interpreted as including armed forces, police, parliaments, judges, etc. This resonates with Montesquieu's doctrine of the separation of powers, now accepted and followed in well-functioning systems of government.
How does it work? Once sunk investment costs have been incurred, many groups will want to destroy the current institutions and replace them with new ones. That is the time inconsistency problem: a group taking power would have incentives to expropriate the fruits of investments previously made. But importantly, this group would also have some optimal degree of power sharing in mind. For instance, a large legislative body might be thought of as unnecessary when institutions are rebuilt. Now suppose that the current institutions share power among a wider group. If a coup d’état occurs, some of those in power will lose their privileged status. Anticipating that, they will be willing to ‘fight’ to defend the current institutions.
The key idea is that power has to be shared so that power is less concentrated than it would be if institutions were destroyed and the group that took power were to rebuild institutions from scratch. Interestingly, the mechanism through which this operates is not bringing in new types of individuals who represent or care about other citizens. All members of the elite care only about their own rents under the status quo, and that is what gives them incentives to fight against insurrections from both insiders and outsiders. But alas, that is also what precludes efficient institutions from arising.
Although it is possible to sustain protection against expropriation by sharing power among a sufficiently large group, in equilibrium there is too little power sharing and thus too little commitment to rules such as protection of property rights. The reason is that sharing power requires sharing rents – because were powerful individuals not to receive rents commensurate with their status, they would have incentives to join rebellions against the current institutions, and no incentive to defend them. Hence those in power might want to choose a small pie (inefficient institutions) in order to guarantee their slice will be larger (little power sharing).
In many parts of the world, the threat of conflict has played and continues to play an important role in shaping institutions. Understanding that connection might teach us why some parts of the world are still so underdeveloped, and perhaps help us to devise better ways to address the problem.