Paying politicians theory and evidence

Pdf File 296.09 KByte, 37 Pages

Paying Politicians: Theory and Evidence

Timothy Besley London School of Economics

Abstract This paper looks at the theory behind the idea that paying politicians better will improve their performance. The paper lays out a political agency model with adverse selection and moral hazard where politicians are subject to two-period term limits. This model provides a number of predictions about how the pay of politicians affects agency problems. We also consider what happens when the pool of politicians is endogenous. The main ideas in the model are confronted with data on U.S. Governors.

This paper was first given as the Joseph Schumpeter Lecture at the 18th Congress of the European Economics Association in Stockholm. I am grateful to Robin Burgess, Rohini Pande, Ray Fisman, Sanjay Jain, Rocco Macchiavello, Torsten Persson, Michael Smart, and Daniel Sturm for useful discussion and comments on a earlier draft. Daniel Sturm and Ray Fisman were also kind in offering me access to their data. Marit Rehavi and Silvia Pezzini provided excellent research assisitance.

e-mail address:


1 Introduction

While a great deal of attention has been paid to the remuneration of chief executives in the private sector, comparatively little has been written about political chief executives such as Prime Ministers and Presidents. But pay setting for politicians is an important issue and the pros and cons of wage increases are often hotly debated. This lecture focuses on two questions -- (i) What is the effect (theoretically speaking) of changing the remuneration paid to politicians? (ii) Is there any evidence that such remuneration affects the behavior of politicians or who chooses to run for office?

To set the scene, it is interesting to look at a small selective sample of political chief executives. To this end, Table 1 gives some figures for four countries -- the U.K., U.S.A., France and Sweden. The U.S. President is the best paid at around $400,000 with the British Prime minister coming second at $270,000. The worst paid is the French President at $70,000 -- worse than a starting assistant professor in Economics in a U.S. University! Relative to income per capita, the ranking is also preserved. The U.K. and U.S.A. both pay around ten times national income per capita, while France pays only three times. However, these jobs carry very different responsibilities. The third row standardizes the salaries of these chief executives relative to the size of the government budget. On this basis, the U.S. President is actually the worst paid with Sweden now topping the league table followed by the U.K.. The fourth row standardizes this by population size. The Swedish prime minister is now comfortable the best paid, with the U.K. second. France and the U.S. look fairly similar on this basis. These numbers are only illustrative, but they serve to focus our mind on the issues at hand.

Officials may also have very different domains of responsibility, terms


and methods of accountability. Any serious comparison of wages would have standardize across payments from different sources and job description. It is largely for this reason that the empirical analysis in this lecture focuses on a very specific group -- Governors in the U.S..

A data set for U.S. Governors provides a more comparable source since they perform largely similar tasks in a common institutional setting. The Governor of California is paid $165,000 to preside over a budget of $150 billion (including Federal funds) with a population of roughly 34 million, while the Governor of Montana is paid around $84,000 to preside over a budget of $3 billion and a population of just 900,000. The average pay of a Governor in 2000 was $105,000. To put this in perspective, the median CEO compensation package for the Mercer/Wall Street Journal Survey of the 350 biggest publicly held companies is around $6.1 million.1

The main justification for the high level of chief executive compensation levels is that it generates good incentives and enhances corporate performance. This can be thought of in a standard principal agent model where it is difficult to monitor the actions of managers -- monetary incentives are needed to diminish any conflict of interest between managers and shareholders. The efficacy of monetary incentives in political settings is far from clear. Discussions have frequently adopted a view based on the notion that political service is a calling and that money is a distraction. This is nicely illustrated by the following quote from an article published in The Idaho Statesman of 1945 complaining about the behavior of the incumbent Senator:

"The best of our public officials don't seek their positions for



the money. There are other and infinitely greater rewards. For a senator who can win it, there is the respect of his ablest colleagues, and that is something no money-chaser has ever had. There is his own respect to win and keep, and his conscience to live with, while devoting himself faithfully to the welfare of his country. And after his long service is over there is the monument to himself in the memory and the gratitude of the people he served. If such deep and abiding gratifications don't make your salary seem rather incidental, then you don't belong there, and the people of Idaho will be happy to relieve you of the job in 1950."2

This view belongs to a respectable tradition that views the political process as a means of getting the right people to make policy decisions. In this vein, the great American political scientist, V.O. Key argued that:

"(t)he nature of the workings of government depends ultimately on the men who run it. The men we elect to office and the circumstances we create that affect their work determine the nature of popular government. Let there be emphasis on those we elect to office. Legislators, governors and other such elective functionaries ultimately fix the basic tone and character of state government." V.O. Key (1956 page 10).

This sentiment should also make sense to economists. Thus, Lazear and Rosen (1980) note that:

2From "Salaries of Politicians by Vardis Fisher, The Idaho Statesman, January 16, 1945 available at .


"Public service has always been viewed as a social obligation, somehow different from other careers, and the responsibilities, duties, and personal conduct of public officials has been regarded somewhat differently from those of people holding positions in the private economy." Lazear and Rosen (1980 page 101).

But how exactly do such sentiments factor into political agency and the role of elections in improving the quality of policy? How should they affect wage setting policies? These are the issues that will concern me here. While not seeking to re-instate the benevolent government paradigm entirely, the approach that I will take will put greater weight on the importance of selection. Following Key and others, I take it as a possibility that some individuals are more publicly spirited than others and that one role of good governance is to get the right people (persons of character to use a phrase favoured by James Maddison) making social decisions. This is not to argue that standard self-interest concerns are not important -- problems such as bribery and corruption which infect many aspects of public life are grounded in the self-interest motive in politics.

The possibility of benevolence is somewhat at odds with the dominant traditions in political economy emanating from Chicago and Virginia. They are founded on applying standard economic reasoning to political life -- this includes the presumption that actors are narrowly selfinterested. Thus, in his characterization of the Public Choice approach, Buchanan (1989) takes it as given that "Individuals must be modeled as seeking to further their own narrow-self interest, narrowly defined, in terms of measured net wealth position, as predicted or expected." (Buchanan (1989, page 20)). He also suggests that "(t)o improve pol-


itics, it is necessary to improve or reform rules, the framework within which the game of politics is played. There is no suggestion that improvement lies in the selection of morally superior agents who will use their powers in some "public interest" (Buchanan (1989, page 18)).

In political life, perhaps the most basic incentive comes from the need to be re-elected -- politicians who do not do what voters like are removed from office. Barro (1973) suggested an approach to incentives in politics where elections are used to discipline incumbents.3 This has spawned a significant interest in political agency problems. Since wages affect the value of holding office, then using the standard logic of efficiency wages, it should be easier to inculcate policy in the voters' interest when wages are high.4

The approach that I take here allows for unobserved heterogeneity in the types of politicians (adverse selection) as well as unobserved actions (moral hazard). We are then able to consider the role of wages in sorting politicians as well as their role in achieving discipline in the standard way. I construct an agency model with two main features. First, time is infinite and there are successive generations of politicians each of whom can serve at most two terms. Second, the politicians can choose to enter politics and their decision to do so depends upon their outside option and their payoffs in political office. The main innovation proposed here is to embed the agency model in a framework where the pool of politicians is endogenous.

This lecture will look primarily at theory. However, it also considers evidence from a very particular source -- U.S. Governors. The standard political agency model with a single agent directly accountable

3See also Ferejohn (1986). 4This argument was put forward for public sector employees in general in Becker and Stigler (1974).


Download Pdf File