Public Servants

In: Journal of Moral Philosophy
Mario I. Juarez-GarciaPhilosophy Department, University of San Diego, San Diego, ca, United States,

Search for other papers by Mario I. Juarez-Garcia in
Current site
Google Scholar
Alexander SchaeferSchool of Law, New York University, New York City, ny, United States,

Search for other papers by Alexander Schaefer in
Current site
Google Scholar
Full Access


Several political philosophers have recently pointed out that current electoral democracies fail to facilitate accurate and reliable feedback on the performance of public officials. Rather than rejecting democracy as a hopeless ideal, we defend an institutional reform called Service Responsibility, which introduces a superior incentive structure that better aligns the interests of citizens and public officials. Service Responsibility requires increasing or decreasing the income of public officials insofar as they succeed or fail to achieve democratically chosen goals. Later, we consider an alternative institutional scheme, recently proposed by Claudio López-Guerra: Piloting Responsibility. According to this alternative proposal, public officials must utilize a public provider whenever they seek out a basic service. We show that Piloting Responsibility fails to realize this ideal by generating a perverse incentive structure for public officials. We conclude that Service Responsibility outperforms both the status quo and Piloting Responsibility as an institutional scheme for ensuring competent governance and public justification.


Several political philosophers have recently pointed out that current electoral democracies fail to facilitate accurate and reliable feedback on the performance of public officials. Rather than rejecting democracy as a hopeless ideal, we defend an institutional reform called Service Responsibility, which introduces a superior incentive structure that better aligns the interests of citizens and public officials. Service Responsibility requires increasing or decreasing the income of public officials insofar as they succeed or fail to achieve democratically chosen goals. Later, we consider an alternative institutional scheme, recently proposed by Claudio López-Guerra: Piloting Responsibility. According to this alternative proposal, public officials must utilize a public provider whenever they seek out a basic service. We show that Piloting Responsibility fails to realize this ideal by generating a perverse incentive structure for public officials. We conclude that Service Responsibility outperforms both the status quo and Piloting Responsibility as an institutional scheme for ensuring competent governance and public justification.

I would have wished to be born in a country, where the sovereign and the people could not have but one and the same interest, so that all the movements of the machine always tended exclusively to common happiness …

jean-jacques rousseau, Discourse on the Origins and Foundations of Inequality among Men

1 Democracy and the Identity of Interests

A major strand in democratic thought emphasizes the alignment of interests between political leaders and citizens. Electoral democracies aim to achieve this alignment through the mechanism of popular elections. According to this view, citizens will scrutinize public officials and decide whether they have obeyed the general will. Those who truly represent the interests of the citizenry will be elected; those who do not will be fired. In this way, elections will incentivize rulers to further the common good. This alignment of interests thus realizes two values which are widely recognized by liberal democrats: competent governance and public justification.

Or so the story goes. Political theorists have become increasingly skeptical that electoral systems can realize these democratic goals. Critics of democracy assert that the feedback between political leaders and the electorate is too noisy and too delayed to discipline public officials who fail to serve the public interest.1 Does that mean that we must give up the democratic ideal of justifying political power on an identity of interests? Not necessarily. The fact that the electoral system is defective does not imply that we must abandon democracy, but that we must think harder about how to avoid its most serious pitfalls.

This paper defends an institutional scheme called Service Responsibility, which fixes the incomes of high-level public officials to some indicator of wellbeing that the governed choose, along with a metric used to measure performance on that indicator. Service Responsibility avoids the political rhetoric, voter irrationality, and disinformation of electoral democracy, and harnesses democracy to improve the alignment of elected representatives’ and people’s interests in a way that gives citizens good reason to believe that their rulers are properly using their delegated power.2

To show that Service Responsibility offers a way to improve upon status quo electoral democracies, this paper proceeds as follows. Section 2 outlines some desiderata for liberal democracies. In the third section, we present Service Responsibility, demonstrate its ability to improve the democratic identity of interests, and present its general features. Section 4 deals with some objections to Service Responsibility, specifically the problems that may arise when using an indicator to evaluate performance. Before concluding, in section 5, we contrast our proposal with Claudio López-Guerra’s competing institutional arrangement, Piloting Responsibility, in order to highlight three important virtues of Service Responsibility.

2 Desiderata: Incentive-Compatibility and Public Justification

For liberal democrats, competent governance requires that leaders effectively serve the interests of citizens.3 Moreover, nearly all liberal democrats hold that leaders must also provide public justification for their decisions and policies.4 The problem, as we understand it, is how to make institutions incentive-compatible and publicly justifiable. Hence, we lay out two key desiderata.


In order to ensure competent governance, that is, to reliably serve the interests of citizens, institutions must align the interests of rulers and citizens so that rulers are motivated to serve the public. If we are aiming to build a strong link between the interests of public officials and those of the people, we must examine the incentive compatibility of any arrangements that aim to do so. Hence, the desideratum …

Incentive-Compatibility (ic): Democratic institutions must be arranged so that they incentivize public officials to choose policies that will serve the interests of their constituency.

As one condition for satisfying ic, any institutional scheme that aims to ensure competent governance should pass the following test:

The Incentive Test: Is the proposed arrangement incentive-compatible, i.e., does it incentivize public officials to choose policies that will best serve the interests of their constituency?

An arrangement will perform better on the Incentive Test when we expect high-level public officials to choose policies that better match the interests of the citizens. By contrast, an arrangement will perform poorly on this test when (1) rulers have preferences that differ from those of a large part of the population, and (2) the satisfaction of the rulers’ preferences comes at a cost to those whom the ruler governs.

Public Justification

Those who focus on public justification, as most liberal democrats do, demand more from democratic institutions than the identity of interests.5 They hold that the justification of democracy rests heavily on its ability to yield publicly justified policies through, or in light of, a publicly justified procedure.6 The second desiderata is thus …

Public Justification (pj): Political leaders must provide reasons for citizens to believe that they are properly using their delegated political power.

There are different ways to justify political power. Reasons can be discursive or nondiscursive. The discursive justification approach understands public justification to occur through a process of public deliberation. For instance, deliberative democrats hold that institutions and laws are publicly justified through a fair and open discussion in a public forum.7 Similarly, public reason liberals often understand public justification to obtain when leaders or citizens provide arguments that appeal to the entire citizenry, despite their diverse moral commitments.8 Both groups emphasize the importance of providing an argument, verbal or written, to convince other citizens that their institutions are just, or at least acceptable.

While the discursive approach offers an important avenue for public justification, there exists a neglected but equally important kind of justification: nondiscursive justification. As its name suggests, this kind of justification encompasses all justifications that do not rely on speech. Rather than telling others that an institution or a policy is justified, we show them that it is. Nondiscursive justifications come in various forms, but tend to rely on an institutional structure to ensure that political leaders serve the interests of their citizenry, rather than merely seeking private gain.

Although nondiscursive justification has been neglected by most liberal democrats, there are some recent exceptions. Kevin Vallier, for example, defends a polycentric political order in which citizens enjoy exit power.9 Because they can switch jurisdictions at low cost, citizens are able to avoid laws and institutions that they find unjustifiable. Through a process of self-sorting, therefore, “exit can move the law into closer alignment with what is publicly justified.”10 Along these lines, Jeffrey Green argues for an “ocular democracy” that empowers the citizens by promoting “political institutions that seek to regularize candor as an everyday feature of democracy,”11 e.g., cross-examinations, interviews by unsympathetic journalists, questions from rival politicians, and so on. Finally, and most importantly for our project, López-Guerra has recently proposed an institutional arrangement that he calls Piloting Responsibility, which he believes would align the interests of public officials and citizens, thereby providing nondiscursive public justification.12

While discursive reasons aim to persuade and convince citizens that the use of power is justified, nondiscursive reasons show that public power is justified through actions, practices, or mechanisms that help substantiate and strengthen justificatory claims. Discursive justification may be important but it faces important limitations. With discursive justification, it is easy to distort the truth, provide misleading promises, or even flat-out lie. When rulers say “we are all in this together,” they often utter empty words to deceive the governed. Without mechanisms of accountability, cheap talk is ubiquitous in political discourse. Nondiscursive reason-giving avoids these issues by giving leaders skin in the game, thus constituting an indispensable supplement, if not a replacement, for discursive justification. When an institution provides nondiscursive reasons, citizens tend to be more confident that public officials will follow through on their word.

As we have hinted, the two desiderata (ic and pj) are interrelated in a crucial way. The duty to satisfy pj is a second-order duty that depends upon the first-order duty of properly exercising political power. In essence, pj is about publicizing the successful discharge of this first-order duty. As a matter of logic, therefore, one cannot satisfy pj if doing so undermines the likelihood, as appraised by citizens, that one has, in fact, properly exercised political power. It follows, then, that a second-order duty of public justification may not be discharged by any action that undermines the likelihood of fulfilling the associated first-order duty, as reasonably appraised by the moral patient to whom the duty is owed.13 To clarify, suppose that Avital owes a first-order duty to Ben, as well as a second-order duty to show Ben that the first-order duty was adequately discharged. If an action Φ reasonably leads Ben to doubt that Avital has discharged her first-order duty, then Φ-ing cannot provide a means of discharging the second-order duty.

The upshot of this relationship is that one cannot satisfy pj if doing so undermines ic. A political arrangement with any hope of satisfying liberal hopes for democracy must satisfy the conjunction of ic and pj within democratic institutions. Although electoral democracy aims at fulfilling these desiderata by linking the performance of the elected politicians to the popular vote, many democratic theorists fear that it performs poorly on both fronts because the feedback supported by popular elections is both noisy and delayed. We now turn to the task of reforming the democratic process so as to satisfy the two desiderata.

3 Service Responsibility

Realizing the desiderata of liberal democrats requires setting up incentives for the rulers to actually act, and show that they act, for the interests of the citizens. This is the goal of the institutional scheme...

Service Responsibility: As a condition for the occupancy of high public office, the income (including fringe benefits and salary) of public officials for the duration of their term—and perhaps some time after—will be determined by their performance according to a particular indicator (and metric) of wellbeing that the governed choose, either directly or indirectly.

Through a democratic process, citizens articulate their priorities by choosing indicators, e.g., gdp, and determine how to measure performance on this indicator by choosing a metric, e.g., U.S. dollars. Service Responsibility provides an incentive structure that punishes high-level public officials for harming the public interest and rewards those who serve it. The details of its application, e.g., the duration of validity of the indicator, clauses that could modify it before it expires, the length of the obligation for former leaders, and the exact procedure of selection (majority vote, supermajority, etc.), are to be sorted out by each society. Our goal is to provide the framework for our proposal, to describe its general features, while allowing the specifics to be determined by each democratic society.14

Three General Features: Elected Officials, Income, Indicators

First, Service Responsibility imposes obligations for high-level elected officials of the legislative and executive branches of power, i.e., legislators, members of the congress, governors, presidents, prime ministers, and so on. Since the goal is to amend the gap between the interests of high-level public officials and those of the citizens, it makes sense that elected public officials are the main subject of our institutional scheme. That said, some unelected occupants of key positions of power would fall under the obligations of Service Responsibility, as well. For example, members of the cabinet should be submitted to it, as they exert extraordinary power over citizens’ lives. This is an important extension, since elected officials often delegate complex tasks to cabinet members. Other members of the administrative state, however, need not be subject to this limitation inasmuch as their performance is typically assessed by their hierarchical superiors according to whether they follow orders or not. In this, our proposal differs from New Public Management proposals, which focus on setting incentives of performance for the members of the administrative state, but seldom for elected officials.15 Service Responsibility is neutral regarding the duties of mid-level and street-level bureaucrats as it mainly concerns decision makers who determine the direction of national policy.

Secondly, Service Responsibility stipulates that the incomes of political leaders (i.e., their fringe benefit and salary) be indexed to particular welfare indicators. It is important to differentiate fringe benefits from essential goods and services required for good performance. Personal security, transportation to official events, basic healthcare, and a comfortable working space are services that any employee with important responsibilities requires for good performance. It is in the interest of the governed that their public officials (their employees) enjoy good physical and mental health so that they can concentrate on achieving optimal results. However, a place of residence, clothes, education for their children, personal social media management, a summer house, and a private lawyer are all examples of fringe benefits. Basically, any goods or services used in the public officials’ free time or that are not directly related to the public office must be paid for by the rulers. Alternatively, they can choose to utilize a public provider at no cost. Like any other citizen, they have the option of acquiring goods or services from a private or a public provider as long as they can afford them.

Service Responsibility also opens the door to setting incentives that apply after a public official leaves office. While this is not essential to the institutional arrangement, this feature might foster rulers’ attention to the long-term effects of their decisions. To fix ideas, we present some examples of how this could be done: high-level public officials can be allocated a pension after leaving office which could increase or decrease according to the fluctuation of a previously chosen indicator of performance. Alternatively, former decision makers could be sanctioned if their policies yield damaging effects, e.g., former Presidents could receive a fine if the decision to start a war led to an indicator to show low economic performance due to this decision. That said, it might very well be the case that societies decide that they do not want them to have any obligation (and hence no reward) after rulers leave office. The particular arrangement must be chosen by each society.16

Third, the key condition to prevent high-level public officials from manipulating the indicator is that they have no power over it. If the standard of evaluation is set by those who are under evaluation, it would undermine the incentive compatibility of the arrangement; rulers could simply select indicators or metrics on an ad hoc basis. In electoral democracy, public officials will often cherry-pick the most flattering indicator or metric, while ignoring or dismissing any that reveal their failures. Service Responsibility therefore aims to disentangle politicians from welfare and pushes voters to differentiate their choice of political candidates from their preferences for welfare standards.

While Service Responsibility requires that the selection of the indicator and the election of political representatives be separated, it remains the task of each democratic society to determine the particulars of the choice process. Consider the following concrete proposals:

  1. The indicator can be chosen directly, either by simple majority vote, or by a supermajority.
  2. The election can be indirect: congress, or a special body of political representatives, can select indicators (by majority, supermajority, or consensus) for the other branches of power (as said, they cannot choose the standards that apply to themselves).
  3. A weak version of Service Responsibility could involve choosing politicians who are required to propose their own indicators and metrics before taking office, so as to avoid the issue of ad hoc cherry-picking. In this way, when voters choose politicians, they are also choosing indicators and metrics. Although the choice of leader and indicator occurs simultaneously, these are still distinct choices insofar as citizens are not delegating the choice of indicator to the political leader, but, rather, are approving it or rejecting it through direct democracy.
  4. Voters can choose goals—not indicators or metrics—and allow a group of experts to translate those goals into indicators, and define metrics. In this way, voters could articulate vague desires, such as lowering unemployment, while the specialists could focus on devising the best indicator and metric to incentivize political leaders to strive for this goal in a meaningful way.
  5. There are also more complex variations and combinations of these proposals. For example, citizens could vote directly for welfare goals and, instead of a winner-takes-all election, some of the best-ranked goals could be taken into account by the assembly of experts to shape a composite indicator that would respect the heterogeneity of citizens’ preferences.

There exists a wide variety of ways to implement Service Responsibility. The choice between these alternatives will depend on the nature of particular democratic systems. In all cases, however, Service Responsibility separates preferences for candidates and preferences for welfare indicators, thus offering a way to steer public officials towards policies that satisfy the interests of the governed.

Satisfying Liberal Desiderata

How would Service Responsibility fulfill the desiderata of liberal democracies? Satisfying pj requires transparency: citizens must know what goals their rulers should be pursuing. Satisfying ic requires motivation: citizens must know how their rulers will be incentivized to pursue those goals. Service Responsibility transforms public officials into transparent and efficient public servants by indexing their income to some indicator of social wellbeing to create a fluctuating income that depends directly on performance.17 This indicator may be a composite and encompass any number of simple indicators: gdp per capita, the Gini index, hedonic measurements, human capabilities, the human development index, the index of economic freedom, or others.18

As things currently stand, the way in which we allocate income to public officials is quite unique. In a typical firm, performance partially determines the incomes of both employees and employers. A typical employee earns a fixed salary somewhat insulated from variations of the market, but subject to the discretion of his employer. In general, employees who perform well receive raises and promotions, while those who perform poorly are demoted or fired. Employers, on the other hand, decide their own incomes, but the amount they can afford to withdraw from their revenues is constrained by the firm’s performance in the face of market conditions. In both cases, therefore, the quality of one’s work is a partial determinant of one’s income. This is not so for high public officials. Like employers, they decide their own incomes, and, like employees, their incomes do not fluctuate according to market conditions. Thus, the incomes of public officials are not directly tied to any performance indicator. For the governed, this is the worst-case scenario, because high public officials’ incomes are independent of whether they yield good results or not.

Service Responsibility uses income as an incentive for improving public officials’ performance in much the same way as income is used to incentivize the typical employee or employer. If a policy increases the wellbeing of the citizens, then citizens reward the rulers with a raise, while a policy that decreases social wellbeing translates quickly into a pecuniary penalty. When Service Responsibility has teeth, rulers become servants of the public, and the governed are their employers.

To illustrate how this institutional scheme would work, consider an instantiation of Service Responsibility in which there are no fringe benefits and the salary is determined by a particular indicator of wellbeing. For simplicity, imagine that the citizens are concerned about the healthcare system, so they choose a related indicator of performance and a composite metric that combines the number of hospital deaths, successful surgeries, preventive care consultations, and other relevant factors for the population. Avital, the newly elected President of Kazanistan, reforms the healthcare system. After some years, the indicator reveals the deterioration of social healthcare. Because Service Responsibility is in place, Avital’s income takes a hit due to her unsuccessful policies, which affects her ability to afford important things – the private education of her child, for example. Citizens know that Avital is sharing the cost of her bad policies, so they know that it is in her personal interest to quickly correct the course.19 Avital eliminates her reforms so as to regain the previous performance of the healthcare system and her previous levels of wellbeing for her and her family. Citizens are not satisfied with Avital’s performance, but they know that she tried to fulfill her first-order duties and corrected them as soon as it was clear that her policies were failing. The citizens do not need to wait for the next election to correct unsuccessful policies.

Service Responsibility, therefore, passes the incentive test, inasmuch as its incentive structure leads high-level public officials to choose the policies that, they believe, best serve the interests of the citizens (satisfying ic). Because it passes this test in a transparent way, it also satisfies pj: under Service Responsibility, the governed have nondiscursive reasons to believe that the rulers strive to advance the interests of citizens because personal income is linked to performance. Service Responsibility thus emerges as a good candidate for satisfying ic and pj.

4 The Danger of Indicators

Linking the income of political leaders to indicators of performance raises several concerns: (i) selecting the indicator might be problematic; (ii) voters’ might choose an indicator reflecting their preference but not their interests; (iii) high-level public officials can game the chosen indicator or metric by ignoring aspects of wellbeing not captured in its evaluation; (iv) Service Responsibility might undermine rulers’ intrinsic motivations and become detrimental to their performance; (v) random shocks might create unfair outcomes for public officials. Alas, the perfect incentive structure does not exist. We hold, however, that Service Responsibility will yield substantial benefits, even if implemented imperfectly. In what follows, we deal with these concerns about the realistic working of Service Responsibility.

Choosing Indicators

We have already admitted, along with critics of electoral democracy, that voters are not particularly good at retrieving information about how to vote based solely on outcomes. Service Responsibility seems to place an immense cognitive burden on subjects as voters: selecting, through some aggregation mechanism, the components and weightings of different statistical measurements that make up the (composite) standard of performance evaluation, deciding how to attach punishments and rewards to this indicator, evaluating whether or not the metric and incentives are producing the desired outcomes, and so on. How does Service Responsibility fare in the light of these challenges given that it requires that voters choose not only the representative but also the indicator?

We distinguish two features that render this cognitive load so burdensome. First, it requires a large amount of fact-gathering. The governed face difficult questions such as: What are the most promising indicators and metrics experts use to assess social states? What do these abstract metrics represent and what do they leave out? How accurately and economically can the requisite data be accessed for each of these indicators? For example, unemployment may seem like a rather straightforward indicator, but this is an illusion. Unemployment metrics involve myriad complications and assumptions, and before endorsing unemployment as a valuable metric for performance evaluation, the governed would need to grasp these many nuances. Second, there is the problem of noisy feedback. Because of general political ignorance, irrationality, partisan bias, and other shortcomings of human cognition, citizens may not be capable of discerning whether or not a given metric has led to improvement or decline. Even absent these cognitive distortions, assessing counterfactuals is epistemically questionable. If it is impossible to know what would have happened under an alternative indicator, metric, or incentive scheme, how can voters appraise the success of one scheme relative to another?

It is worth noting that both of these problems (fact-gathering and noisy feedback) arise in democratic decision-making whether or not Service Responsibility is satisfied. In every democracy, voters exhibit political ignorance and performance is difficult to assess. One way of coping with this issue – surely not a silver bullet – is to elect representatives, rather than to vote on policy directly. Representatives, it is hoped, can spend their working hours learning the intricacies of legislation and leveraging this knowledge to best serve their constituency.20 Meanwhile, ordinary voters can focus on their own, non-political work. Similarly, under Service Responsibility, voters might choose an assembly of specialists trained in the uses and hazards of social measurement. The assembly of experts presents significant improvements over the democratic status quo: it mitigates the issue of fact-gathering by leaving the task in the hands of experts who can disentangle complex information and strive to produce indicators that could progressively articulate the voters’ preference more accurately.

With respect to noisy feedback, the democratic choice of the standard of performance evaluation has a decisive advantage over normal democratic electoral politics. While politicians typically serve short terms, the time-horizon for a standard of performance evaluation is much longer. It can be improved piecemeal and incrementally over time. For example, citizens might first choose unemployment as their indicator and define a metric of employment that simply divides the number of unemployed individuals by the labor force. Now, imagine that an efficacious but simple-minded ruler succeeds in lowering this metric of unemployment by creating unproductive jobs, which leads to inflation and unnecessary government spending. Arguably, in this case, voters were not deceived; they set an inadequate metric for their indicator and learned that they do not care about unemployment as measured by that metric. They may also have learned that their indicator was too narrow: they care about more than mere unemployment. This is feedback about the choice of the indicator. Next time that voters have to select it, they have reasons to choose a better metric for unemployment and a more complex indicator of performance, perhaps including inflation to yield a composite indicator. Our preferred arrangement, an assembly of experts, might contribute to elucidating the adequate metrics for the preferences of the citizenry after several iterations of trial and error. Therefore, the capacity for cumulative learning is much greater once Service Responsibility is in place, and, for this reason, even a noisy stream of information can lead to meaningful progress over time – provided that the learning process is not hopelessly weak. Service Responsibility offers voters a way to leverage learning processes that would otherwise be too slow and too imperfect to effectively implement.

Service Responsibility improves the noisy feedback problem in yet another way. Although determining the right indicator and metric remains difficult and public accountability will always be imperfect, it nevertheless marks a major improvement over a system in which rulers act on a host of unknown incentives to pursue a variety of unknown aims. Satisfying pj requires that citizens know, to some degree, the goals that their leaders should be pursuing and how they will be incentivized to pursue such goals. This achievement is more attainable under Service Responsibility than under the status quo given that the appraisal of high-level public officials becomes relatively straightforward. By specifying an indicator beforehand and reducing the number and variety of possible means of evaluation down to a democratically-chosen metric, the issue of determining whether a leader has been successful is greatly simplified. Under Service Responsibility, therefore, it is more likely – alas, not guaranteed – that voters learn the right lesson from their choices.

The choice of indicators is not a one-shot choice. Instead, it unfolds and must be reaffirmed over time, which allows for reflection and learning. Because of its iterative nature, this choice involves a series of built-in “cooling off” periods, which are thought to reduce the effect of cognitive biases, leading to more rational choices.21 In this way, the democratic choice of a metric and indicator presents a particularly low-risk context for cognitive biases and failures of rationality.

Preferences and Interests

Another issue with selecting the metric and indicator is also worth mentioning. While voters will generally choose to satisfy their preferences, preferences and interests (or welfare) are not equivalent.22 We often have preferences for things that do not affect our interests. One may prefer, for example, that the human species continues to exist for the next several millennia. Moreover, choice and interests tend to come apart when cognitive biases are present that distort our choice behavior in seemingly irrational ways.23 How can we have faith in voters’ ability to vote for their own interests given the widely recognized tendency to behave in irrational ways?

While the exact relationship between choice, preference, and interests can be specified in a number of ways, there are two extreme views that can be ruled out. First, philosophers of economics generally agree that preference satisfaction is not equivalent to interest satisfaction.24 The claim that satisfying preferences logically entails increasing wellbeing does not withstand critical scrutiny. Second, the claim that preferences rarely or never reflect an agent’s interests is similarly implausible. While a typical agent will sometimes have preferences for outcomes that do not pertain to any specifiable interest of that agent, it is even more common that an agent’s preferences correspond to some interest(s) of that agent. In rejecting these two extreme views, we follow Daniel Hausman in holding that “self-interest is common, and people are often better judges of what serves their interests than are others.”25

While the claim that voters will sometimes fail to serve their own interests is certainly correct, it does not follow that there exists a superior mechanism for translating information about voters’ interests into policies. Allowing government agents to decide how to instantiate voters’ interests by selecting indicators of wellbeing raises obvious incentive issues. Due to the principal-agent problem, the preferences and choices of government agents are less likely to reflect voters’ interests than are the preferences and choices of the voters themselves. While one can cite various cognitive biases to argue that voters would be bad at instantiating their interests in a chosen indicator, government agents are also prone to cognitive biases.26 Thus, while Service Responsibility will not be a perfect mechanism for translating voters’ interests into policy choices, it will outperform mechanisms that rely solely upon the indicators preferred by the rulers.

Gaming the Indicator

At this point, a further worry appears. No indicator covers all aspects of wellbeing or all dimensions that we would like to measure. Every metric highlights some aspects by casting a shadow on others: exam scores reward particular performance but not innovative thought or intellectual curiosity; hourly wages reward time inputs, but not product outputs. Socio-economic indicators are no exception to this general rule. A monomaniacal focus on the selected indicator or metric can easily prove deleterious to the common good. Racking up public debt and spurring inflation can increase gdp, destroying valuable business enterprises can increase equality, and so on. This produces the problem of strategy: elected officials might game the system by boosting the metric at the expense of other important priorities. As noted above, using expert representatives who specialize in designing incentive-compatible indicators and metrics can help here, but two considerations are important in the implementation of this approach.

First, those choosing the indicator and defining the metrics should seek guidance from qualified experts with training in game theory and mechanism design. Such experts will not foresee all possible abuses, but they will be able to prevent the most obvious and detrimental by designing metrics that are difficult to game. A key principle here will be selecting multipart, composite indicators and metrics that correspond to the desired indicator. If the indicator is simple, such as gdp, then it can be pursued at the expense of various other concerns. A more complex indicator – such as a composite of gdp, health outcomes, equality measures, poverty reduction, and public debt – will be far more difficult to game.

Second, if we rely on representatives to choose indicators and specify metrics, then they must be insulated from partisan politics. Perhaps the best model for this in the United States is the Federal Reserve, which aims to be nonpartisan, often employing economists who retain their jobs despite new administrations entering the White House. If the representatives were partisan, then politicians would have an undue influence over the selection of metrics, undermining the realization of Service Responsibility. In practice, the Federal Reserve is, of course, subject to various political pressures, but it is remarkably insulated from these pressures given its key role in the American economy. If the agency tasked with selecting the metric is equally independent, then it could greatly improve the quality of our governance.

The representative system will not appeal to all societies; some populations will prefer a more direct form of democracy, which means that they would prefer to have a direct choice over the indicator. As noted above, this is compatible with Service Responsibility. The only caveat is that the problems of fact-gathering and strategy will likely be worse in the absence of experts to select the metric. For some societies, however, this may seem a trade-off worth making.

Crowding Out

The economist Bruno Frey claims that linking performance to monetary rewards undermines personal, intrinsic motivation to work; he called this the Crowding-Out Effect. Frey concludes, “external interventions crowd out intrinsic motivation if the individuals affected perceive them to be controlling. In that case, self-determination, self-esteem and the possibility for expression suffer, and the individuals react by reducing their intrinsic motivation in the activity controlled.”27 Some might worry, then, that Service Responsibility produces a Crowding-Out Effect in high-level public officials, hinders their performance, and, consequently, becomes a counterproductive institutional arrangement. We argue that there are reasons to be skeptical about this concern.

There is a vast literature on crowding-out effects. However, a recent review of 100 tests based on the canonical experiment challenges the view that crowding-out effects are strong and omnipresent. The economists Constança Esteves-Sorenson and Robert Broce hold that the existing literature has failed to test a full range of effort indicators.28 Thus, in testing whether or not monetary incentives boost performance, past experiments may have produced false negatives. Moreover, in their own field experiment, they conclude that “with one exception, our results across output, productivity, and quits are consistent with a standard economics model,” i.e. money payouts incentivize greater effort.29 These empirical results show that we must be skeptical about crowding-out effects. The evidence is too mixed to be conclusive.

Even if crowding-out effects are real and powerful, they may not be omnipresent. Roland Bénabou and Jean Tirole hold that these effects only appear in particular cases of the application of the principal-agent model, particularly they “requir[e] that the agent be less knowledgeable in some dimension than the principal.”30 The relation between the ruler (the agent) and the citizens (the principal) does not satisfy this requirement inasmuch as political leaders have more information about their actions, the circumstances, and the challenges to achieve a certain outcome, than the citizens. If Bénabou and Tirole are right, then, a crowding-out effect might exist as a phenomenon, but it would not represent a worry regarding Service Responsibility.

For the sake of argument, however, let us suppose that monetary incentives will crowd out intrinsic motivation in the case of political office. The question then becomes: which incentives will do a better job of motivating good governance, the intrinsic ones or those introduced by Service Responsibility? This depends on the nature of the “intrinsic” motivations that are displaced by pecuniary incentives. If one holds that politicians are moved by a genuine desire to promote the values of citizens, then Service Responsibility may present an undesirable reform. If, however, politicians are more often moved by less pristine motives, then using monetary incentives as a tool for directing their behavior might present a massive improvement over the status quo. In the absence of monetary incentives, for example, politicians might be driven by a desire for power and prestige. Or, arguably worse, they may have a genuine, ideologically-tainted desire to promote a particular conception of the good. Both of these scenarios are far worse than a profit-seeking politician attempting to maximize an intelligently chosen metric for some indicator of societal success.

In the end, therefore, the purity of motives matters less than the quality of outcomes. Even if politicians were saints, they may underperform competent, profit-seeking leaders who aim to maximize an incentive-compatible indicator and metric. One reason why this might be the case is that, for all their altruism, leaders may not know exactly what citizens value or the priority rules to apply in addressing their concerns. Prices (in the form of income) may impart such information to political leaders insofar as the indicator and metric are democratically chosen. In other words, even if a desire for profit drives the behavior of political actors, so long as profits are acquired and maintained by producing excellent results, both ic and pj will be satisfied. We do not need our leaders to be saints; we need them to be competent executives.

Random Shocks

Some might worry that public officials are unfairly rewarded or punished when unforeseen and uncontrollable events affect the indicator. A clause within Service Responsibility that prevents the harm that uncontrollable events cause to public officials may alleviate those worries. However, such a special clause is undesirable. If unexpected shocks occur, an institutional arrangement should not protect public officials from damages. Doing so would grant public officials an unequal and unjustified immunity. We hold that politicians ought to be subject to the vicissitudes of chance, just like the governed. More than a matter of equality, tying the fate of governors and governed, even in unexpected circumstances, is important first and foremost as a matter of satisfying ic and pj.

When random shocks – such as global economic crises, pandemics, or external wars – severely damage social welfare indicators, it is desirable that rulers personally feel such damage so as to minimize the harm caused by such events. If public officials see their living standards decline due to an external shock, they will have additional reasons to strive for effective and ingenious ways to overcome the crisis. Their personal quality of life is anchored to social welfare, and the need for solutions to these problems will help spur their entrepreneurial spirit. Rather than the cliché that political leaders are like captains of a ship, rulers should be thought of as similar to business owners. As restaurant owners are motivated to innovate and adapt to unfavorable circumstances (increasing their home delivery capacity in the case of a pandemic, for example), under Service Responsibility public officials have personal motivations to increase their alertness to solutions and innovative policies. They also have incentives to revise and correct policies that prove unsuccessful. In status quo electoral democracies, leaders have less incentive to seek out new solutions if their emergency strategies fail. Instead of working to respond creatively to the new challenges, public officials can devote their efforts to crafting effective political rhetoric to mask their failure. For example, politicians often devise ways to blame adverse outcomes on their political opponents. As a last resort, a political leader can, with low cost, simply throw in the towel and wait to be fired in the next election. Elections impose a low cost on underachievement or running out of ideas. Conversely, Service Responsibility imposes a high cost on such failures, thus incentivizing political leaders to explore new strategies until they find one that triggers recovery.

Consider the opposite case: new technological innovations, the discovery of new resources, or even favorable weather are examples of shocks that can improve the welfare of citizens and, in all likelihood, the indicator of performance of the leader. Should leaders benefit from these fortunate events that they did not provoke? Maybe society simply got lucky. Why should public officials be rewarded for that? One reply is that fortunate social conditions are not, in general, only a matter of luck. New innovations tend to appear when inclusive economic institutions are in place.31 The state shapes the framework within which such innovations are possible; therefore, high-level officials are – at least, partly – responsible for constructing the basic institutions that create the conditions for technological innovation. At the very least, they are responsible for not destroying these institutions and stifling innovation. It is easy to overestimate the importance of fortuitous events and to underestimate the difficulty of constructing and maintaining political institutions that allow society to benefit from such events. Both Norway and Venezuela possess rich oil reserves, but their different institutional structures have produced quite different outcomes.32

Furthermore, as a second reply, rewarding leaders for presiding over good outcomes – even if they did not cause these outcomes – sends a signal to future leaders about the value of serving the public interest as well as the importance of maintaining institutions that permit innovation and social progress. Thus, although external shocks can create unexpected advantages, there are major benefits to rewarding leaders for these fortuitous events. Service Responsibility incentivizes appropriate action in the case of external shocks.

5 Piloting Responsibility and the Virtues of Service Responsibility

Although we have shown that Service Responsibility will improve upon status quo electoral democracy regarding ic and pj, it is worthwhile to compare it to an alternative institutional arrangement that also aims to create a strong identity of interests between rulers and governed: Claudio López-Guerra’s Piloting Responsibility. We will present this alternative and contrast it with Service Responsibility to make salient some major virtues of our institutional proposal.

Piloting Responsibility

The core of Piloting Responsibility can be expressed by the long-standing analogy between a ruler and the captain of a ship. All citizens are like passengers of a ship in the quest for a better future. However, the vast majority are merely passengers while a few are also officials. The latter are in charge of the ship, while the former must obey orders. Even if all are equal qua passengers, officials, qua officials, have more powers and duties. Since they all share the same fate and burdens, passengers have nondiscursive reasons to think that officials steer the ship as best as they can. To recreate this strong identity of interests in a democracy, López-Guerra proposes …

Piloting Responsibility: As a condition for the occupancy of high public office, public officials and their dependents, for the duration of their term and some time after, will be required to reside within the polity, and if they seek out a basic service, they would be randomly appointed to a public provider, say, a public hospital, a public school, a public defender, etcetera.33

Piloting Responsibility imposes three conditions on high-level public officials: living in the country that they govern, being subject to the rule of law, and acquiring goods and services exclusively from public providers, when available. The aim of this institutional scheme is to publicly align the incentives of citizens and public officials (satisfying ic) by providing nondiscursive reasons for citizens to believe that officials are properly using their delegated power (satisfying pj).

In López-Guerra’s view, Piloting Responsibility eliminates the rulers’ exit option, confining them to face the full consequences of their policy decisions. When public officials exit the polity, they abandon the ship, leaving the governed to their own fate. And when such an option exists, the governed lack reasons to believe that their representatives are doing the best that they can, since, if their performance is poor, public officials can always “[escape] in lifeboats unavailable to the public” when the ship sinks.34 The governed know that, because high-level public officials can abandon defective public services, they lack a strong incentive to improve these services. On the contrary, when public officials cannot exit, the governed have nondiscursive reasons to believe that public officials will strive to provide the best public service possible for the governed because they will strive to provide the best service for themselves. Removing the exit option for rulers both creates an identity of interests and makes it plausible to believe that there is an identity of interests.

Despite its initial appeal, we believe that Piloting Responsibility would give rise to a serious incentive incompatibility, which threatens to undermine public justification. We refer to this incentive incompatibility as the Alignment Problem.

The Alignment Problem

To understand the nature of the Alignment Problem, and how Piloting Responsibility gives rise to it, first notice the divergence of interests and values between citizens and rulers. The divergence of interests and values arises partially, though not entirely, from the special position of high-level public officials in society, which is both political and, most likely, socio-economic. Even absent these distinguishing features, however, a divergence of interests would likely remain. An irrepressible feature of modern society is the fact of pluralism, i.e., the fact that under free institutions there will exist “a plurality of reasonable doctrines, both religious and nonreligious, liberal and nonliberal” and that these doctrines will often clash on fundamental matters of value and belief.35 An important way of coping with such diversity is to define spheres of autonomy within which citizens can make private decisions that are not subject to public control. For instance, in many situations, citizens pursue their idiosyncratic preferences by choosing personalized consumption bundles in a market setting. But, under Piloting Responsibility, if “public officials and their dependents [...] seek out a basic service, they would be randomly appointed to a public provider [...].”36 Consequently, they are forbidden from seeking out individualized consumption bundles using their own funds; they must now consume the same goods and services as citizens who choose to use the public option, and they must do so at public expense. The conflict of interests, endemic to any diverse society, is thus exacerbated by López-Guerra’s proposed institutional scheme.37 A diversity of interests transforms into a conflict of interests insofar as Piloting Responsibility is in place. This conflict of interests is what we call the Alignment Problem.

An example may illuminate how Piloting Responsibility creates tension between the interests of the political leader and the general populace. Suppose that Avital, the president of Kazanistan, must allocate resources for the purpose of promoting public health. As is often the case, the most effective use of these funds – that is, the most lives saved or improved per dollar – lies in taking basic measures to prevent avoidable maladies such as malaria and dysentery. However, the political leader faces very different health priorities. Well into her 50s, she likely lives in an urban area, in comfortable housing, free of malaria-causing mosquitoes or dysentery-causing amoebas.38 She is far more concerned about treating heart disease, cancer, and other so-called diseases of affluence. Since under Equal Subjection she cannot use her own private funds to secure special treatments – advanced diagnoses, precision surgeries, high-tech mri s, etc. – she has a strong incentive to direct social resources to provide these services at public expense. More basic and effective health measures may go unfunded because the maladies that they would prevent on a mass scale are not a pressing issue for the political leader. Therefore, under Piloting Responsibility, satisfying public officials’ personal preferences comes at a cost to the governed. In other words, in a diverse society Piloting Responsibility fails The Incentive Test.

Since there are no real instances of Piloting Responsibility, we rely mostly on theoretical considerations, rather than empirical ones. However, there are many examples of how political leaders shape the public sector according to their own preferences. Probably, the clearest cases are found in Soviet republics, where something similar to Piloting Responsibility was the default institutional scheme, as leaders had to acquire goods from public suppliers. The Soviet leaders created a free public healthcare system; however, this did not lead to the same quality of medical care for every comrade; in reality, “the Soviet government created two systems of medical care; one for regular citizens, and one for highly privileged, high-ranking government officials, essentially creating an upper class with regards to access high-quality health care.”39 The members of the nomenklatura hired foreign physicians from the best schools in the West, offered them jobs in selected public hospitals, and employed them to train elite Soviet physicians. The rest of society had to deal with public, low-quality medical services from physicians trained in their country. They were equal subjects, but the quality of the service was not equal. This is not a unique feature of the extinct Soviet republics. A more recent example of the same phenomenon in a welfare state is the public spending in Mexico, where the President, Andrés López, a diehard baseball fan, spent more than 20 million dollars to remodel baseball stadiums in 2021, in the midst of the covid-19 pandemic and even when soccer is a far more popular sport nationwide.

Pace López-Guerra, consuming the same services need not provide assurance that both parties are well served. Put differently, equivalent consumption does not entail equal welfare. Subjective differences in needs, interests, and preferences mean that one person’s good will often be another person’s frivolous expense, or even her bad. Indeed, one way of encapsulating the main point is that López-Guerra ignores diversity in his claim that, by utilizing the same public services, the leader provides assurances that those services are serving the interests of the citizenry at large. To the contrary, in areas where the political leader has preferences that diverge from those of the populace, the exact opposite is true: constraining a politician to utilize public services may provide reasons to believe that those services are not provided at the optimal quality and quantity. Or rather, these services are more likely to be provided at the optimal quality and quantity according to the standards of the ruler, not those of the populace. The average citizen typically faces very different trade-offs and holds very different values than the political leader. The same bundle of public services will not serve them equally well. This fact alone serves to undermine López-Guerra’s claim that “[a]s full and equal users of the state’s goods, officials stand to lose as much as the citizens if they have not in fact ruled in the public’s interest.”40

One might defend Piloting Responsibility by claiming that democratic elections would prevent public officials from abusing their power. However, López-Guerra joins us in dismissing the feedback process as being too noisy and overwhelmed by countervailing considerations. As he puts it, “public officials have ample opportunities to misrepresent their performance in office.”41 His claims are borne out by many democracies, where leaders regularly win reelection despite perverse levels of corrupt and self-serving behavior. And even if democracy effectively facilitates feedback, that is, even if self-serving behavior does impose some cost in terms of reelection, optimizing politicians will seek to strike the best trade-off between the benefits of reelection and the benefits of self-serving behavior.42 By removing outside options, Piloting Responsibility will change the relative payoff structure, pushing rulers to increase their self-serving behavior on the margin even if it comes at the cost of somewhat reducing their probability of reelection.43 If we accept López-Guerra’s claim that this reduction in probability will be slight or imperceptible, then, according to basic price theory, this shift in costs and benefits will result in significant behavioral change, that is, in much higher levels of self-serving behavior.

Solving the Alignment Problem

Piloting Responsibility generates The Alignment Problem. Service Responsibility, on the other hand, alleviates this problem in two crucial ways. First, it permits political leaders to utilize markets, thereby allowing them to serve the interests of the governed without extreme personal sacrifice.44 Second, by indexing the rulers’ salary to performance according to some indicator of wellbeing, Service Responsibility softens the conflict between rulers’ interests and the common good. Under Service Responsibility, as long as public officials want to maximize their utility − which is linked to their purchasing power − they want to efficiently serve the interests of the governed. The common good and the interests of decision makers are intertwined, and any trade-offs between satisfying one or the other tend to be relatively small. Consequently, under Service Responsibility, political leaders can effectively satisfy pj, since doing so will not entail undermining the satisfaction of the first-order duty to serve the interests of citizens.

The crucial drawback of Piloting Responsibility is that leaders may provide goods and services that they find highly valuable at the expense of goods and services that the governed much prefer. Service Responsibility is a superior framework for fulfilling ic and pj because it promotes policies enacted for the sake of social welfare by installing a mechanism that links the rulers’ personal, private, pecuniary interests to what the citizens consider to be their own interests. The measure of social welfare emanates, not from the leader’s private conception of the good, but from the citizens themselves through a democratic aggregating mechanism that sets the indicator to serve as the standard of performance. Service Responsibility thereby provides an institutional scheme that offers nondiscursive reasons to believe that public officials do the best that they can according to the standards of the governed.

Incentives, Standards, & Steering

How does Service Responsibility deal with public officials’ deviant interests? As a public servant, the leader has an obligation to improve people’s lives. Of course, in most cases, monetary income does not exhaust a leader’s interests. People, including political leaders, have deviant interests, that is, they often pursue aesthetic, moral, ideological, or otherwise idiosyncratic objectives, even if it harms their pecuniary interests. This is true whether a political system operates under Piloting or Service Responsibility. But how do the two institutional schemes compare when dealing with deviant interests?

To answer this question, let us consider a hypothetical scenario. Imagine that Avital, President of Kazanistan, faces an important decision: whether or not to invade the distant nation of Mazanistan, toppling its oppressive government and spreading the values of equality and democracy. Avital truly believes that all citizens deserve to taste the fruits of liberty and equality which grow only under conditions of liberal democracy. She also knows that many of her friends and wealthy campaign donors would benefit from access to Mazanistan’s rich oil reserves. How would Avital’s incentives differ under Piloting Responsibility as opposed to Service Responsibility?

Under Piloting Responsibility, Avital faces little cost, from her perspective, for this policy. First, the standard that is used to motivate her is rigidly fixed to her own evaluations of the quality of state services. She benefits or suffers to the extent that these services, which she must utilize, are improved or diminished according to her own standards. Since Avital believes that invading Mazanistan is a worthwhile endeavor, she anticipates that doing so will improve the quality of the national defense service. Second, the cost imposed on Avital for policy errors is rigidly fixed to the effect that the policy has on the publicly provided goods and services that she consumes. Here, because of her ideological bent and her private interest in helping her friends and donors, the cost of invasion will be minimal (or negative) for Avital, even if most citizens view it as unnecessary, burdensome, and destructive.

This example reveals that Piloting Responsibility is constrained by two kinds of rigidities …

Standard Rigidity: The indicator and metric by which political leaders assess success or failure is difficult or impossible to adjust.

Incentive Rigidity: The harshness of punishment (or the generosity of reward) for failing (or succeeding) according to the institutionalized standards is difficult or impossible to adjust.

In physical terms, standard rigidity relates to the direction of institutional incentives, while incentive rigidity refers to the force of these incentives. Together, these two rigidities entail that it is difficult to control the velocity of institutional incentives, that is, the goals toward which rulers steer society and the vigor with which they do so.

Service Responsibility is flexible where Piloting Responsibility is rigid. Citizens can choose their own standards (which may include life expectancy, economic growth, or education outcomes) and they can choose the extent to which leaders are punished for failing to perform well according to these standards. This power of selection does not guarantee that citizens will choose the best indicators or metrics, but it allows for correction if the wrong ones are chosen. For example, Avital’s invasion of Mazanistan will probably raise the gdp of Avital’s country and lower its unemployment rate. If her salary is determined by these indicators, Avital’s monetary incentives will point towards invasion. However, in all likelihood, the invasion will also lead to a host of economic problems: a decrease in consumption and investment, higher public debt levels, and currency inflation. Citizens might choose to include these measures in a composite standard of performance evaluation. More simply, they may, as peace-loving folk, include a salary penalty for any non-defensive war. Avital may think twice if she knows that the cost of her policy error would result in not only the deaths of strangers and the squandering of valuable resources, but also a large decrease in her personal income due to the adverse effects of war. By avoiding standard rigidity, Service Responsibility allows citizens to steer the decisions of their public officials with much greater precision.

Yet, the penalty may not suffice to deter Avital from making a reckless decision. Unlike Piloting Responsibility, Service Responsibility provides a straightforward remedy: increase the penalty. By adjusting the magnitude of punishments and rewards, citizens can ameliorate the issue of insufficient incentives. One can gauge the index so that public officials feel public losses and gains more acutely. For instance, upswings in the public debt could be indexed in a one-on-one relation: a 1% increase in national debt corresponds to a 1% decrease in presidential salary. If this is inadequate, a one-to-ten relation may suffice.

Even with this kind of flexibility, one might worry that Service Responsibility would be impotent for extremely wealthy leaders. Those with higher incomes can afford severe reductions in their salaries that result from implementing ineffective policies, perhaps for ideological reasons or to secure non-pecuniary rewards.45 Considering only reductions in income, however, understates the flexibility of Service Responsibility. In addition to decreasing a leader’s income, citizens may also opt for a scheme that permits leaders to earn a negative income. Even a billionaire leader would then have much to lose by implementing or failing to reform detrimental policies. Another solution might involve placing restrictions on the rulers’ access to personal wealth during their time in office as a punishment for poor performance. This would further enhance the power of incentives by ensuring that reductions in income be felt more directly. Third, as discussed above, extending the period of time in which one’s income is affected by indicator performance could also enhance the power of incentives. Finally, the incentives could be defined in relative terms instead of absolute terms, so that the penalties and rewards for independently wealthy rulers are greater than the penalties and rewards for less wealthy rulers.

The issue of incentive strength also bears on the relationship between Piloting and Service Responsibility. While Piloting Responsibility and Service Responsibility are not logically incompatible, the foregoing discussion does help to elucidate a tension between these two institutional arrangements. As a hybrid form of Service and Piloting Responsibility, consider, for example, a formal institutional application of Service Responsibility, constrained by an informal application of Piloting Responsibility as a recognized moral duty. In this case, high-level public officials would exclusively acquire their goods from public suppliers – even if they are not institutionally constrained to do so – while their income would be determined by their performance. The tension becomes clear when one considers that Service Responsibility’s power to steer and incentivize is enabled by the fact that reducing a political leader’s income reduces her access to market options. If political leaders are required to consume public goods and services, regardless of their performance, then the incentive effects of increased or decreased income become much weaker. Therefore, prima facie, this hybrid form might hinder ic inasmuch as rulers would only need to care about the quality of public providers while neglecting welfare indicators in general – as they do not need income to acquire goods.

Other variations also exist. For instance, there might be an institutionalized hybrid form in which rulers can draw upon better – or worse – public providers depending on their performance. This institutional arrangement would require a difference in quality in public services so the access to better services depends on the rulers’ performance. We will not work out the details of such hybrid forms, but merely acknowledge their existence. Nonetheless, our points about incentives and steering suggest reason to believe that – in terms of satisfying pj and ic – Service Responsibility will outperform hybrid schemes that weaken the strength of its incentive effects.

Risk & Experimentation

Finally, some might worry that our institutional scheme favors the status quo because it rewards conservative policies that do not greatly alter the indicator of performance, so as to maintain the initial salary. In other words, Service Responsibility might disincentivize risk-taking behaviors, thus losing the possibility of beneficial innovation in public policy. Alternatively, one might worry that politicians, in their quest for monetary gain, might take on excessive risks to the detriment of the public. Service Responsibility has straightforward solutions to both of these potential issues.

Under Service Responsibility, citizens are able choose the level of risk-taking behavior that they want their ruler to exhibit. The punishment and reward structure need not prescribe punishments and rewards that run parallel.46 If policy mistakes are more disastrous for citizens than policy successes are good, if things are going particularly well, or if citizens are simply risk averse, then citizens may wish to induce less risky behavior in their leaders by punishing errors more harshly and rewarding successes less extravagantly. For example, suppose that citizens choose gdp as part of their composite standard of performance evaluation. If the gdp increases 1% in a year, perhaps public officials obtain a 1% raise; but, if the gdp decreases by 1% in a year, high-level public officials lose 10% of their salary. In this way, the indexation can be calibrated for high or low risk according to citizens’ levels of risk tolerance. Piloting Responsibility, by contrast, has no way of precisely striking this balance. Consequently, citizens would not be reassured that their risk preferences are adequately accounted for by the leader’s use of power; this is yet another reason to predict that pj would not be satisfied under Piloting Responsibility.

Our claim about optimal risk-taking connects to an important issue that has recently received attention in the philosophical literature: under conditions of complexity, politics must perform a search function, experimenting and learning in order to find new and better solutions to political problems.47 This means that political leaders must be given some latitude to take calculated risks in search of political improvements. At the same time, we do not want politicians to be constantly undertaking costly experiments when adequate policy responses already exist. In other words, rulers face the classic “explore/exploit trade-off.”48 Service Responsibility allows citizens to articulate their risk preferences by structuring the incentives that leaders face when determining whether to explore new solutions or to exploit tried-and-true policy responses. Due to incentive rigidity, Piloting Responsibility can perform no such feat. Citizens’ risk preferences remain largely irrelevant to the decision calculus of politicians. Adequate incentives are crucial to assure citizens that rulers strive to implement effective innovations where feasible without subjecting citizens to reckless and unnecessary schemes of social engineering. Without such assurance, pj is not satisfied. Service Responsibility, in contrast to Piloting Responsibility, provides a straightforward mechanism for incentivizing cautious experimentation, tailored to the risk preferences of the population.

6 Conclusion

At this point humility is due. There may exist institutional arrangements that neither we nor López-Guerra have considered, and some of these institutional arrangements may outperform any arrangement that involves Service Responsibility. While we cannot rule this out, we do maintain that Service Responsibility is well-designed to align the interests of public officials and citizens, and it thus presents a plausible way to improve currently existing democracies.

We have argued that well-designed democratic institutions must tend toward an “identity of interest” between the citizens and high-level public officials. Service Responsibility offers a way to improve current-day democratic practices. It promotes an actual identity of interests between governors and governed, which, compared to both the status quo and Piloting Responsibility, better satisfies both desiderata: Incentive-Compatibility (ic) and Public Justification (pj).

Service Responsibility takes into account that passengers care about reaching their destination, not whether or not the captain is on board when the ship sinks. The captain fulfills her duty of public justification as long as people have reasons to believe that they are going safely and efficiently to wherever they want to go. Piloting Responsibility, at best, gives reasons to trust that the captain would go down with the ship, whereas Service Responsibility provides reasons to trust that the captain is doing the best that she can to keep the ship afloat and sailing it in the direction of the common good.


We are thankful to the editors (especially, Thom Brooks and Chiara Cordelli) and to the anonymous reviewers whose comments made our argument considerably stronger and more sophisticated. We also thank Jacob Barrett, Rhys Borchert, Tom Christiano, Charles Delmotte, Matthew Jeffers, Santiago Sanchez-Borboa, and David Schmidtz for their valuable help and support. The inception of this project took place in a Social Choice seminar taught by the late Gerald Gaus; we would like to thank Jerry for all that we learned from him.

Biographical Note

Mario I. Juarez-Garcia is a Visiting Professor of Philosophy at the University of San Diego. His work has been published in Res Publica, Social Philosophy and Policy, and Synthese.

Alexander Schaefer is a postdoctoral researcher at the New York University. His work has been published in American Journal of Political Sciences, Philosophy and Economics, and the Social Theory and Practice.


Bryan Caplan, The Myth of the Rational Voter: Why Democracy Choose Bad Policies (Princeton, NJ: Princeton University Press, 2007); Jason Brennan, Against Democracy (Princeton, NJ: Princeton University Press, 2016).


For the sake of conciseness, we will often use the word “rulers” in place of “high-level public officials.” Unless otherwise specified, the term “rulers” refers to elected representatives and other high-level public decision makers.


To best satisfy the “interests of the people” is a fraught issue due to social choice problems. There are different methods of generating a collective welfare function from a set of individual utility functions, none of which succeed in satisfying all important desiderata. There is no perfect collective welfare function, and philosophers have varying views about which is the most just. For this reason, when we refer to the “interests of the citizens,” this is actually short for “the ranking according to the best possible welfare function,” where “best” is left open. Our argument will hold for a wide range of social welfare functions, though not for those in which “interests of the citizens” is equivalent to satisfying the rulers’ preferences (Kenneth J. Arrow, “A Difficulty in the Concept of Social Welfare,” Journal of Political Economy 58, no. 4 (1950): 328–346). Philosophically speaking, interests and preferences are not equivalent (Daniel M. Hausman, Preference, Value, Choice, and Welfare [Cambridge: Cambridge University Press, 2011]). For the purposes of our argument, however, what matters is that rulers have a choice function that differs from the “best” possible welfare function of their constituency. The relationship between interests, preferences, and choice is discussed below in subsection Preferences and interests.


For example, John Rawls, A Theory of Justice: Revised Edition. (Cambridge, MA: Harvard University Press, 1999), 297; Thomas Christiano, The Constitution of Equality: Democratic Authority and its Limits (Oxford: Oxford University Press), 47.


Hélène Landemore Democratic Reasons: Politics, collective intelligence, and the rule of the many (Princeton: Princeton University Press, 2013).


Joshua Cohen. “Procedure and Substance in Deliberative Democracy,” in Philosophy and Democracy, ed. T. Christiano (Oxford: Oxford University Press, 2002).


Thomas Christiano, The Rule of the Many: Fundamental Issues in Democratic Theory (Boulder, CO: Westview Press, 1996); Jurgen Habermas, Between Facts and Norms (Cambridge, Mass: mit Press, 1998).


This construal of public reason liberalism applies most readily to consensus theorists, such as Rawls, who believe that all citizens must endorse the same justification for institutions and laws. Convergence theorists such as Gaus and Vallier, by contrast, believe that citizens provide their own justificatory arguments from within their comprehensive doctrines. Both groups tend to focus on arguments, and therefore emphasize discursive justification by our definition. Gerald Gaus, The Order of Public Reason (New York: Cambridge University Press, 2011); Kevin Vallier, Liberal Politics and Public Faith: Beyond Separation (New York: Routledge, 2014).


“Exit, Voice, and Public Reason,” American Political Science Review 112(4) (2018): 1120–1124; Also, Brian Kogelmann “Justice, Diversity, and the Well-Ordered Society,” The Philosophical Quarterly 67(269) (2017): 663–684; and Alexander Schaefer, “Reasonable but Non-liberal: Another Route to Polycentrism,” The Philosophical Quarterly (2021).


Vallier, “Exit, Voice, and Public Reason,” 1121.


Jeffrey E. Green, The Eyes of the People: Democracy in Age of Spectatorship (New York, NY: Oxford University Press, 2009), 14.


Claudio López-Guerra, “Equal Subjects,” Philosophy & Public Affairs 45 (2017), 321–55. We discuss this project in Section 5.


The duty of Public Justification may seem, to some, to be an obligation, rather than a duty, since it is undertaken by voluntary choice of occupying public office. More plausibly, it may be considered a positional duty, since it is owed in virtue of a certain position that rulers occupy in relation to citizens. A. John Simmons, Moral Principles and Political Obligations (Princeton, NJ: Princeton University Press, 1979), 12–4.


In this, we are influenced by Sen’s Capability Approach, which allows flexibility for societies to determine the specific capacities that are most salient for them, given their history and culture. Amartya Sen, The Idea of Justice (Cambridge, Massachusetts: Belknap Press, 2009), 231–235.


Peter Aucoin, “Administrative Reform in Public Management,” Governance 3 (1990): 115–37; Owen E. Hughes, “New Public Management,” in Public Management and Administration (Palgrave: London, 1998): 52–80. Christopher Hood, “The ‘New Public Management’ in the 1980s: Variations on a Theme,” Accounting, Organizations and Society, 20. 2–3 (1995): 93–109.


See subsection Incentives, Standards, & Steering, for more on the importance of rewards and sanctions.


This is an application of the principal-agent model in a political setting. George Downs and David Rocke, “Conflict, Agency, and Gambling for Resurrection: The Principal-Agent Problem Goes to War,” American Journal of Political Science 38(2), 1994: 362–380.


We believe that citizens would certainly prefer a composite of these indicators. See subsection Choosing indicators for deeper treatment of this question.


We consider only realistic public officials, those who are neither egoists nor saints.


Representatives or party leaders can function as information-conveying mechanisms. Thomas Christiano, “Voter Ignorance is Not Necessarily a Problem,” Critical Review: A Journal of Politics and Society, 27 (2015): 253–269. Also, Alexandra Oprea, “Adam Smith on Political Judgment: Revisiting the Political Theory of the Wealth of Nations,” The Journal of Politics, 84.1 (2022).


Colin Camerer et al., “Regulation for Conservatives: Behavioral Economics and the Case for Asymmetric Paternalism,” University of Pennsylvania Law Review 151.3 (2003): 1211–1254.


We thank an anonymous referee for pressing us to clarify the relationship between preference, interest, and choice.


The literature on cognitive biases is now massive, but an accessible discussion is available in Richard Thaler and Cass R. Sunstein, Nudge: Improving Decisions about Health, Wealth, and Happiness (New York, NY: Penguin, 2021).


See, for example, Amartya Sen, “Rational Fools: A Critique of the Behavioural Foundations of Economic Theory,” Philosophy & Public Affairs 6 (1997): 317–44, and Daniel M. Hausman, Preference, Value, Choice, and Welfare, Ch. 6.


Hausman. Preference, Value, Choice, and Welfare, 91.


As Viscusi and Gayer explain, while private individuals exhibit a host of irrational behaviors, government actors “are also subject to similar behavioral inadequacies across a broad range of policies.” Kip Viscusi and Ted Gayer, “Behavioral public choice: The behavioral paradox of government policy,” Harvard Journal of Law & Public Policy 38 (2015): 973.


Bruno Frey, Not Just for The Money: An Economic Theory of Personal Motivation (Cheltenham, UK: Edward Elgar, 1997), 18.


Constança Esteves-Sorenson and Robert Broce, “Do Monetary Incentives Undermine Performance on Intrinsically Enjoyable Tasks? A Field Test,” Review of Economics and Statistics (2020), 33.


Esteves-Sorenson and Broce, “Do Monetary Incentives … ,” 34.


Bénabou and Tirole, “Intrinsic and Extrinsic Motivation,” Review of Economic Studies 70 (2003), 492.


Daron Acemoglu and James A. Robinson, Why Nations Fail: the origins of power, prosperity, and poverty (New York: Crown Business, 2012).


Leif Wenar, Blood Oil: Tyrants, Violence, and the Rules that Run the World (New York: Oxford University Press, 2015).


López-Guerra, “Equal Subjects,” 327.


López-Guerra, “Equal Subjects,” 343.


Rawls, Political Liberalism, xvi.


López-Guerra, “Equal Subjects,” 327.


López-Guerra, “Equal Subjects,” 348–50.


An important feature of this example is that the socio-economic and demographic background of the leader is very different from that of the average citizen. This entails, we suggest, a systematic divergence between the interests of leaders and those of the citizens. As an example of how political leaders fail to represent a broad range of citizens’ interests, consider the United States: only one of their 46 presidents has been a person of color (about 2%) and none of their presidents has been female. In addition, the median age to become the U.S. President is 55 years, and their average income is far, far above the U.S. median. In general, high-ranking public officials are drawn from a shockingly narrow swath of age, class, gender, racial, and religious backgrounds, given the diversity of the United States as a nation. Even with the best intentions, it will be difficult for leaders to reliably serve the interests of communities of which few or none of them will ever be a member.


Boris Petrikovsky, Y. Nemova, E. Petrikovsky, “Russian Leaders and their Foreign Doctors: A Historical Review,” Lupine Online Journal of Medical Sciences 4 (2019), 338.


López-Guerra, “Equal Subjects,” 340.


López-Guerra, “Equal Subjects,” 341.


This trade-off is explored rigorously in Susan Rose-Ackermann, Corruption: A Study in Political Economy. Academic press, 2013.


On this model, politicians with the largest margin of victory (e.g., a democrat in California) will engage in more self-serving behavior than politicians in “battleground” states, ceteris paribus.


López-Guerra rejects the right of political leaders to seek out goods in markets, because, for him, this represents an “exit” from the polity; this claim is, however, false. Well-functioning markets are in large part provided or enabled by the state. This is the predominant view in contemporary social science. See Acemoglu and Robinson, Why Nations Fail; Joseph Stiglitz, The Price of Inequality (New York: Norton, 2013).


This is an implication of the so-called “wealth effect.” Michael Darby, “Wealth Effect,” The New Palgrave: A Dictionary of Economics, v. 4, (1987): 883–4.


Economists increasingly recognize that losses and gains are not felt equally, which affect the optimal punishment-reward structure. Daniel Kahneman & Amos Tversky, “Prospect theory: An analysis of decision under risk,” Econometrica, 47, (1979): 263–291.


Elizabeth Anderson, “The Epistemology of Democracy,” Episteme 3 (2006): 8–22; Jacob Barrett, “Social Reform in a Complex World,” Journal of Ethics and Social Philosophy, 17 (2019): 103–132; Gerald Gaus, The Tyranny of the Ideal (Princeton, NJ: Princeton University Press, 2016).


J.G. March, “Exploration and Exploitation in Organizational Learning,” Organization Science 2, (1991); Scott Page, Diversity and Complexity (Princeton, NJ: Princeton UP, 2011), 122–124.

Content Metrics

All Time Past Year Past 30 Days
Abstract Views 239 239 31
Full Text Views 16 16 9
PDF Views & Downloads 36 36 20