What Is It to Be a Leader? And Why Do We Need One in the First Place?
The questions that constitute the title of this post are the topic of a very recent and stimulating paper by the economist Kaushik Basu, titled “Hume and Hobbes, with a Dash of Nash. Why Have Leaders at All?”. Over the last couple of decades, Basu has established himself as one of the most insightful and original contributors to institutional analysis, making innovative use of game theoretic arguments to explore the nature and functioning of institutions. His last book, The Republic of Beliefs, does so in the case of law, arguing that the effectiveness of law is ultimately located in its ability to create focal points based on which individuals’ beliefs converge toward some equilibrium outcome.
The paper provides an account of leadership in games that follows similar ideas as in this book. Basu’s main point is that what makes one a leader is the fact of being a “focal person”, i.e., a player whose strategy choice in a game will serve as a focal point to coordinate the beliefs of each other player about everyone’s choice. A “leadership game” is a game where one of the players plays the role of the leader with respect to a subset of the population of players. The strategy choice of the leader (“the order”) is known and followed by the members of this subset, provided that it is a best response when every other member of the subset follows the leader. A Nash equilibrium in a leadership game is “deviation proof” if there is no order such that it is a best response for any member of the subset to follow the order if everyone else in the subset does. Finally, a “Leviathan equilibrium” is a deviation-proof Nash equilibrium such that there is no other deviation-proof Nash equilibrium that gives the leader a higher payoff.
Basu proceeds to show through simple examples that having a leader is not necessarily beneficial. This is especially the case when not all the players in the game are members of the subset exposed to the leadership. This may not strike as being very surprising. Maybe more counterintuitively, suboptimality may also arise when all the players are exposed to the leadership. Basu also explores an interesting case where several leaders compete in the same game. Again, we see that the presence of leaders may make the situation worse than without them.
The most insightful aspect of this game-theoretic analysis is to suggest that leadership coordinates people’s expectations in the same way as social norms or other institutional devices. Indeed, we may not need leaders at all provided that we already have the right set of social rules at hand. This does not show that the authority and power of leaders are an illusion. But that suggests that these are endogenous and fragile phenomena that can easily unravel as soon as some members of the subset start to doubt the fact that others will follow the order. This also shows that leadership is not a substitute for institutions: leadership is at the bottom an institutional phenomenon that is embedded into social practices which already have de facto a normative force from the perspective of the individuals taking part in them.
Pursuing in this direction, there are at least two further possible developments to Basu’s model of leadership. A first one is related to the idea that leadership not only works through people’s beliefs but also preferences. I have in mind here Timur Kuran’s now relatively old but still relevant book Private Truths, Public Lies. Kuran develops a model of preference falsification to account for the fact that authoritarian leaders are able to maintain their authority over a population, but also for its fragility that emerges as soon as some persons realize that everyone’s private preferences are turned against the regime. My intuition is that at least in some cases, the authority of the leader is grounded on a conditional preference for conforming to what is perceived as a social norm of legitimacy, complemented as it is by the fear of formal or informal sanctions in case one is alone revealing her true “dissident” preferences.[1] More generally, I think that part of the phenomenon of leadership is also based on the leader’s ability, within a particular institutional framework, to change people’s preferences in such a way as reinforcing his authority by making it more robust to “exogenous shocks” affecting people’s beliefs.
The second possible development is partially related to the first. In Basu’s model, the existence of the leader is exogenous and postulated by the definition of the leadership game. Following my remark above about the social embeddedness of leadership, we may want to investigate what we could call the institutional construction of leadership games. This is important to understand the roots of authority and political legitimacy in all kinds of regimes. This is for example especially relevant in liberal democracies where the identification of political leaders primarily relies – to use Max Weber’s famous ideal types – on a “rational-legal” form of legitimacy. But we see also that this source of legitimacy is more and more contested. An important and wide research question is then to determine what are the conditions for institutions to foster the establishment of leaders as focal persons. This is not only a matter of efficiency because Basu’s analysis suggests that having leaders is not automatically beneficial. Other considerations must be at play.
[1] Basu’s point in the Republic of Beliefs is that in the case sanctions are law-based, or at least partially formal, individuals must have reason to believe that others, starting with public officials, will be incentivized to enforce the sanctions. The authority of the leader is therefore the product of a complex network of beliefs about the actions of different classes of players (citizens, officials).