Acknowledgments
Getting someone else to do something for you is not easy, as I describe at some length in this book. But doing things on one’s own is not easy either. Beyond building on the work of economists and political scientists in developing the arguments and illustrations in this book, I have had all sorts of help from coauthors, colleagues, and students. I began thinking about delegation issues in researching violations of human rights and sorting out the motives of the leaders and the led. Some of what is here and notably the opportunism of the principal in delegating or tolerating the wrongdoing of the agent was initially explored in Agents of Atrocity (2004) and Democracy’s Blameless Leaders (2012). In trying to isolate the leader’s and the agent’s contributions to violations, I worked on the issues of human rights, sexual violence and agentcentered violations first with James M. McCormick at Iowa State University, then with Chris Butler, Tali Gluch, Alok Bohara, and Mani Nepal at the University of New Mexico. To better understand the agent’s contributions to violations I became interested in the role of informal armed militias and vigilantes, teaming up with Sabine Carey, Mannheim University, to construct a database on these organizations. We wanted to find out what sorts of armed groups are out there, why governments outsource to these organizations, and with what consequences. That work continues. More recently, with Kristin Bakke and Dominic Perera (University College London) and Hannah Smidt (University of Zurich), I have been working on a project on civil society monitoring and how and why governments restrict civil society organizations to defeat fire-alarm monitoring. These collaborative projects also inform the argument and analysis presented in this book. At UCL I have been lucky enough to find not just good students but coauthors and generous colleagues. In afternoons spent in the Norfolk Arms, the Conflict and Change group
have given comments on drafts of various chapters. Thanks in particular to Rod Abouharb, Kristin Bakke, Zeynep Bulutgil, Kate Cronin-Furman, Niheer Disandi, Sam Erkiletian, Thomas Gift, Jennifer Hodge, Andreas Juon, Nils Metternich, Jon Monten, Dominic Perera, Kit Rickard, Tatjana Stankovic, Katerina Tertytchnaya, Manuel Vogt, and Sigrid Weber. Thanks also to Tom Dannenbaum, now of the Fletcher School, for his comments on my treatment of war crimes. Some of the examples used in the book have been tried out in the classroom with UCL students. A special thanks goes to Eric Fair, who graciously agreed to speak with me about his experience as an American interrogator in Iraq. For suggesting this project, for a newfound interest in the National Football League, and for such thoughtful advice on how to structure the book, thanks to David Pervin and to Macey Fairchild of Oxford University Press and to my current editor, James Cook, for seeking out such useful reviews and for his ongoing encouragement. Many thanks to the anonymous reviewers who challenged me to firm up the arguments and the contribution. Renée Danziger encouraged me to take on the project and has seen and commented on multiple drafts. I am lucky to have her insights and most of all her love. Finally, thanks to my children, Colin and Kate, and sisters, Alison and Laura, for their love and support.
Why Delegate?
Why Delegate?
In March 2010 in the Bronx, Sharif Stinson got a ticket for trespassing and obscene language. It triggered a very expensive class action lawsuit. His lawyer said there was no crime. Foul language, he noted, was not unfamiliar to New Yorkers. Stinson’s aunt had given her permission for him to be in the building, and the paperwork left the obscenity unspecified. The court found the summons lacked legal justification. Sharif Stinson was one of many New Yorkers to receive tickets for disorderly conduct, drinking and urinating in public from police officers who, on their part, neglected the legal niceties. It cost the City tens of millions of dollars to settle the class action lawsuit. According to the lawsuit, the officers’ reward structure was outcomebased. Those in charge geared officers’ performance to the number of tickets they issued. The nature of the task was communicated by watch commanders in roll-call meetings, as a Bronx patrolman secretly recorded: “Until you decide to quit this job and become a Pizza Hut delivery man, this is what you’re going to be doing. . . . If you think 1 [arrest] and 20 [summonses] is breaking your balls, guess what you are going to be doing? You’re going to be doing a lot more, a lot more than what you think.” In Brooklyn, commanders were similarly pithy: “I see 8 fucking summonses for a 20-day period . . . I told you last month, they’re looking at the numbers. Ain’t about losing your job, [but] they can make your job real uncomfortable. . . . Go through the motions and get your numbers anyway. . . . If they’re on a corner, make ’em move. If they don’t want to move, lock ’em up. Done deal. You can always articulate [a charge] later.”1 While the City denied using quotas, the settlement required it to notify all officers that ticket quotas were not police department policy.2 It was an expensive lesson in the difficulties of delegating. As those in
charge cannot observe the effort of the officer on the beat, the number of summonses indicated productivity. Of course, the community wants officers to work rather than shirk. But they do not want motion-going officers and just “numbers.” The community did not benefit from this productivity. As we go up the chain of delegation from the officer on the beat to the mayor, who in turn answered to the community, nobody really benefited. Taxpayers were out of pocket to no purpose. Even the police officers who met their “numbers” were deprived of meaningful work, which is thought to be an important part of what makes us happy
It is not easy to keep others on task, whether it is policing the Bronx, guarding the safety of the nation, or just fixing your car. As mayor of New York, as a political leader facing a security threat, or as a car owner, delegation brings with it an underlying anxiety. Who is really in control? Is the police officer going through the motions? Is your security official selling you out? Is the mechanic installing an unnecessary part?
The world turns on the delegation relationship and on one party authorizing another to do a task. Social, economic, and political life is inconceivable without it. It has a profound impact on us all, from mundane domestic interactions, to the running of a country or the international community, and from our recreational enjoyments to our spiritual life. Why Delegate? investigates the diverse and sometimes questionable mix of incentives we have for delegating in the first place, and then the difficult choices of why we select one agent over another, why we delegate one task rather than another, or why we sometimes continue to delegate to untrustworthy and opportunistic individuals. Large organizations cannot exist without delegation. Parceling out tasks to contractors, employees, or elected representatives is essential for efficient production, democracy in large communities, and just to get us through the day. There seem to be few tasks too precious for us as individuals or as decision-makers in large organizations to leave to others.
There is the story told by the journalist and speechwriter James Fallows that President Jimmy Carter took it upon himself to manage the scheduling of the White House tennis court. He was a busy man, with Middle East peace to negotiate and hostage crises to settle.
There were presumably others with the time and skill to sort out the order of play, but the president wanted to control the court. Tennis aside, even the most intimate matters may be delegated. Lovers may be matched and mothers surrogated. Ethical as well as intimacy boundaries to delegation are porous. We expect students to take their own tests and write their own papers, which does not prevent others from offering to perform these services for them. Delegation is part of the human condition, as are the difficult choices it brings with it. Its upside can be efficiency, shared responsibility, and even happiness; its downside can be conflict, corruption, and the attenuation of morals.
1
There is an underlying structure to delegation relationships. Economists offer the insights of principal-agent theory, where the principal contracts with an agent to accomplish a task. The principal and the agent are assumed to have conflicting goals. The agent may be in a position to act in a self-interested and surreptitious way at the expense of the principal. The car owner principal wants a roadworthy car or, perhaps, with a prospective buyer in mind, a paper trail showing that the owner is telling the truth and the car has been well looked after. The mechanic, on the other hand, may be tempted to pad the bill. For successful delegation, and we look at this relationship from the principal’s perspective, the challenge is to keep control of the agent and to achieve an outcome in line with her interests. The amount of control exercised by the principal depends on the amount of task-related and agent-related information the principal possesses. Conventionally, the person in charge is at a disadvantage and lacks information about the agent’s qualities and about what he is actually doing.
Principals and agents might be individuals like the car owner and the mechanic, collections of individuals like shareholders who appoint a manager to run a company, and voters who elect a president to run a country, or they might member states of the United Nations who task that organization to coordinate responses to
pandemics, implement a peacekeeping operation, and address pollution, climate change, or some other policy task that crosses national borders. The theory encompasses hierarchical relationships in private and public life. It describes a principal giving an agent a task and holding the agent accountable for the delivery of the task. If the agent fails to deliver, the principal punishes the agent or terminates the contract. Shareholders, for example, punish unsatisfactory performance by cutting a manager’s pay, or replacing him. The theory simplifies the outcomes of the relationship to the interests of the parties and the information they hold. It describes the tension in the relationship produced by the different parties having conflicting interests and the principal’s likely lack of relevant knowledge, or information asymmetry.
Viewing the relationship from the perspective of the principal, interesting questions stem from the assumption that the agent is selfinterested, rather than motivated by the goals of the principal. Her challenge is to shape the agent’s incentives and to limit his opportunities to exploit his position. Getting and maintaining control over the agent is central to principal-agent theory and drives much of the research. The agent’s pursuit of his self-interest at the expense of the principal is the principal-agent problem. Borrowing from the insurance industry, the agent’s opportunism is conceived as adverse selection and moral hazard. Adverse selection is the challenge facing the principal before entering the relationship; moral hazard is the concern after agreeing to delegate. Both result from the principal’s information disadvantage.
Before entering the relationship, the principal lacks information about the qualities of the agent. The principal runs the risk of a poor choice of agent. The agent may be untrustworthy, unsuitable, and understandably unforthcoming about these characteristics. Like the seller of an unreliable car, or a sick person seeking health insurance, a lazy, dishonest, or incompetent agent is expected to behave rationally and to prefer to hide off-putting preexisting conditions from the principal.
After making an agreement or a contract with the agent and as the agent undertakes the task, the principal then worries about how the agent will perform and his capacity for hidden action. She worries
that the agent will be tempted to act in a hidden, dishonest way at her expense. An agent may shirk rather than work, or sell a secret to the enemy. The principal cannot constantly watch the agent. Even if she can, she may not have the knowledge to know whether the agent is acting in her interest or not, for example, regarding the part he is installing in her car.
We do know from the car-maker Henry Ford that the division of labor has benefits. Yet in seeking benefits from someone else’s knowledge and efforts, the principal exposes herself to the opportunism of the agent. Trust in others to carry out a task is not always repaid. There are inherent agency costs in delegating, having to do with the time and effort necessary to find the agent and to agree to the specifics of the task. They also mount up as the more far-sighted, but possibly over-anxious principal, takes costly countermeasures in order to minimize opportunism by the agent. Agency losses accrue for the principal as the self-interested agent acts against the principal’s interest. Much of the principal-agent literature investigates the efficacy of the responses available to a principal to manage and deter opportunism. By exercising due diligence and not selecting high-risk candidates, by writing lengthy and elaborate contracts, by providing incentives to make opportunism less attractive, by inculcating honor codes and professional ethics, by monitoring, and by punishment and dismissal, the principal strives to deter shirking and to align the agent’s interests and actions with her own.
This simple theory about the give and take and to and fro between principal and agent has great reach. Regarding a passage from Exodus about sending an angel to guide the Israelites out of Egypt, one treatment of the topic observes that “even God delegates.”3 In a passage from Paradise Lost, one discovers that God, too, has principal-agent problems; there are worse as well as better angels and even “Rebel Angels,” which the celestial principal has to contend with and to punish (at great length). From such epic struggles to ordinary daily interactions, our understanding of a vast array of issues benefits from bringing them under the simple structure of the delegation relationship, more or less as economists have described it. It alerts us to the possibility of conflicting interests. It identifies how
opportunities might be seized, helps to clarify who is in charge, and suggests how to respond to control problems. While principal-agent theory’s economic applications receive the most systematic attention, thinking about control and accountability and the best response to a loss of control is a widely relevant exercise. It makes us attentive to the standards, norms, and targets (“get your numbers,” as the watch commander urged) expected to guide the delivery of a task, the attribution of responsibility to the relevant actors in the delivery of a task, and, when necessary, what can be done to improve matters. What is more it suggests puzzles presented by actors not acting as the theory expects.
2
The idea behind Why Delegate? is to explore and to develop the logic of delegation, to show its wide application in our private and public lives, and to do so in an informal and accessible way. Of the pioneering treatments of delegation,4 few are written for a general audience as well. In the process, of course, I do not want to leave behind academic colleagues.
To contribute to understanding how we negotiate hierarchical relations in our organizational and everyday lives with real-world applications in mind, I adopt the useful terms of principal and agent, but not the mathematical language of a principal-agent theorist. In the collision of the simple, elegant theory of economists and rational choice scholars with the actual policy and practice of delegation in a variety of situations, there are, I argue, further contributions to be had in working toward a broad, more descriptively useful logic of delegation. Important insights are to be found where the principal behaves in ways that are unexpected and perhaps puzzling to a rational choice eye.5 First, I argue and show that opportunism lies on both sides of the relationship, and nowhere as notably as when it comes to the distribution of blame.6 Here, it is the principal who acts opportunistically at the expense of the agent. Often she does so in a way not specified in the terms on which the agent entered into the relationship. While it may be difficult to predict which tasks a
principal will be willing to delegate, that is not the case with blame. Carrying off the blame is one task that decision-makers almost invariably decide to pass to an agent (or, in a passage from Leviticus, to a goat). The issue of which tasks to pass to an agent and which not is an interesting question in itself. It turns out to be more complicated than the dictates of efficiency.
Second, in showing how the principal behaves in ways unexpected by the theory, the book uncovers her peculiar passivity under some conditions. Even when confronted by an agent’s truly dreadful performance, an unexplored puzzle is why principals in some organizational contexts choose not to punish the opportunistic agent, as the conventional account of the relationship would lead us to expect.
On blame, Why Delegate? describes how shifting—or when the whistle blows, shoving—this unwelcome burden onto an agent is one of a range of incentives shaping delegation relationships. In principal-agent theory, opportunism belongs to the agent. Yet with the incentive to avoid damage to one’s power and position, neither side in the delegation relationship can be trusted. A modified logic of delegation recognizes that while the division and specialization of labor and the allocation of the production and delivery of goods and services to all sorts of trades and professions are surely the engine of economic growth, the boundaries of delegation are not drawn by efficiency calculations alone. We delegate unwanted tasks, complex tasks, and sensitive, high-risk tasks. The incentives to delegate extend from efficiency calculations and saving time and effort to political calculations and saving reputations and managing the blame. It follows that although we equate the division of labor with economic and social progress, not all delegation contributes to the public good. As the motive for the division of labor and the decision to delegate shifts from the economic to the political, from saving time and effort and increasing efficiency to that of preserving power and saving reputations, delegation becomes more dubious. Sometimes the apparently rogue behavior of someone we give a task to is not a case of can’t control, as the principal-agent theorists fear, but of won’t control. Those in charge of the large corporation look away as contractors down the supply chain ignore child labor laws. Or when a
scandal is exposed, those at the top shift the blame down to the lowest plausible level in the chain of delegation, to the “fall guy.”
The success of the “fall guy” strategy rests on an intuitive acceptance by the relevant audience of the tenets of principal-agent theory. These tenets include conflicting goals, the difficulty that those in charge have in knowing what those who work for them are up to and the presence of a control problem at the heart of the delegation relationship. An understanding of the mechanics of delegation as presented here should reveal tell-tale indicators of a principal who won’t control rather than a principal who can’t control. It should allow us to be more discerning judges of when individuals have genuine control problems and when individuals use delegation as a blame management device to deny responsibility and evade accountability.
The principal’s claim that there was a control problem may be barely credible, but it still may offer cover for those in charge when external pressure grows for accountability. It seems enough to get a Saudi prince off the hook for the murder and dismemberment of a former advisor by his hit squad of fifteen security officials, flown in from Riyadh and fully equipped to do the deed in an Istanbul consulate in 2018.
On October 2, 2018, a trap was sprung. It was set for a fifty-nineyear-old ex-employee of the Saudi royal family, Jamal Khashoggi, at the time a columnist for the Washington Post. In his columns, Khashoggi wrote of the Crown Prince Mohammed bin Salman’s use of repression. He compared the prince to Vladimir Putin. His final column described Saudi Arabia’s loss of dignity in the Yemen war. Khashoggi was invited to the consulate in Istanbul to sort out the documents he required to wed his Turkish fiancée. She waited outside. He never came out. He was tortured, killed, and disposed of. A look-alike Khashoggi did leave the consulate.
The Riyadh arrivals that morning included members of the Crown Prince’s security service. The Turkish staff in the consulate had the afternoon off. As the Guardian journalist Martin Chulov puts it: “The loyalties of those remaining in the building could not be questioned. The assembled hit squad was drawn from the most elite units of the Saudi security forces, whose fidelity had been repeatedly tested.”7 But, once the ghastly details of Khashoggi’s fate became public, their
loyalty was questioned by their political master and by those who benefit from economic and strategic relationships with the Crown Prince.
The Saudis’ management of the blame followed the usual sequence of denial, delay, and delegation.8 They first denied Turkish accusations: Khashoggi had left the building and after that they knew nothing. That strategy failed. Turkish intelligence had recorded what happened. After delay, and with international attention not moving on from the Turkish revelations, the Saudis shifted to delegation. Instead of leaving the building, they claimed Khashoggi had gotten into a fight in the consulate. The squad, supposedly sent to persuade the journalist to go back to Riyadh, had gone rogue and killed him. The fidelity of this squad had been repeatedly tested, the reporting states, yet the Saudi foreign minister described a “rogue operation . where individuals ended up exceeding the authorities and responsibilities they had.” The Crown Prince claimed not to know what his security people were doing and denounced the “heinous crime.”9 Arrests followed, and eleven faced trial in Saudi Arabia for the murder. The purpose of punishing these agents is not to correct noncompliance, as punishment is normally depicted in principalagent theory (unless one buys the Saudi, not the Turkish version). Rather, punishment is to shift responsibility onto the agent and is an integral part of the strategy when delegation is used for blame management. Delegation, the claim of a breakdown of control, and then the punishment of individuals who were said to have implemented the killing, was an effort to buffer the Crown Prince from the blame. Others who saw the economic and strategic value of the Kingdom were now not at a total loss for words. The White House, with a stated priority of protecting its ally and its economic interests, could claim some doubt about where the buck stopped. Four months after the killing, it was reported that at the time of Khashoggi’s first column for the Post in September 2017, the Crown Prince had a conversation with an aide that was intercepted by the US National Security Agency. The Crown Prince had said that if the journalist could not be lured back to Saudi Arabia, he would use a bullet on him.10 Publicly, they stuck to the “rogue operation and those responsible will be punished” story. Delegation is used, more
or less successfully, for political and face-saving advantages in addition to gains in efficiency, and in ways the agent may not have bargained for.
Beyond blame and principals disingenuously claiming a lack of control, I argue that principals under some conditions can be peculiarly unresponsive. This part of the argument looks at agents organized in professions. This discussion of professionalism differs from the usual treatment in examining its costs as well as its benefits. Among agents, professionals are perhaps the one group to escape principal-agent theory with their characters tolerably unscathed. From Nobel economist Kenneth Arrow’s work in the 1960s onward, a sense of professional responsibility or “idealism” is recognized as an alternative mechanism to financial incentives and to correct for opportunism. This sense of responsibility allows the principal to overcome the trust gap and the principal-agent problem. Codes of ethics and a sense of mission increase effort beyond the monetary incentive on offer, and they reduce opportunism. Yet when opportunism does occur among these agents, it repays analysis. It turns out that principals have particular difficulty addressing the rogue behavior of these agents. The paradox of professionalism is that in the process of developing the highly specialized skills that agents need for the successful accomplishment of the task, professionalism may raise singularly tenacious control issues. In some real-world contexts, the use of multiple agents, as opposed to the single agent found in the usual principal-agent models, may raise the cost of punishment and encourage passivity on the part of the principal. Where the principal-agent theorist expects a response to correct agent opportunism, an informed but passive principal is happy to leave the hidden action hidden.
3
In short, I argue that there are modifications to be made to the standard account of the principal-agent relationship to better fit what happens in the world around us. The use of ordinary language sets the argument apart from its insurance industry origins and the
formal, rational choice and game-theoretic workhorses of the economists and political scientists that give us principal-agent theory. These scholars may not see this logic of delegation as belonging in their stable, but the aim is to build on the achievements of their approach. The argument draws on the central features of principalagent theory in seeing the interaction between those who give the orders and those who carry them out as driven by the preferences of actors, who are engaged in a struggle for information and control, anxious about betrayal, and with a capacity for opportunism. It departs from standard principal-agent theory in claiming that we should be alert to opportunism on both sides of the relationship, that the preferences of principals and agents may be substantively very different but not necessarily in conflict, that there are situations of won’t control as well as the canonical can’t control, and that agents may have a less rational, cooperative side, not just a competitive regard for each other. As one scales up from the individual in a relationship with a mechanic all the way to relationships between organizations on the international stage, this dynamic plays out. It is my view that an understanding of the logic of delegation can help to make sense of a marvelously disparate set of questions:
How do we outsource tasks in our ordinary lives for ourselves or our family members?
How do shareholders keep managers on task? Or how do those running car firms like Volkswagen ride out pollution scandals? What happens with more than one actor in charge? What particular difficulties do members of a “collective principal,” like shareholders or owners of National Football League (NFL) teams, encounter? The team owners’ agent, Commissioner Roger Goodell, hands out punishments to players and owners and attracts harsh criticism in the media. Why do not the owners rein him in?
Was the corruption surrounding the World Cup the responsibility of the high-handed and brazen Sepp Blatter, the FIFA president? If so, why did the national football associations repeatedly reelect him to conduct FIFA business?
Why is an organization with such a firm view of what is right and wrong as the Catholic Church apparently indifferent to moral hazard? Like an army that ignores the war crimes of its soldiers, the church has covered up for those who use their positions in the church for private gratification. These opportunists are protected by those in charge from the negative consequences of their actions. The Catholic child abuse scandal is the biggest scandal of all time, according to the March 2019 Atlantic Monthly poll of Twitter followers (followed by the “WMD hunt” in Iraq as a distant second, then Iran-Contra and the 1919 World Series in fourth place). Does the logic of delegation help make sense of it?
How does delegation allow principals to get away with wrongdoing? Why do states continue to delegate violence to vigilantes, motorcycle gangs, and irregular armed groups, as Russia does in Ukraine? They have regular armed forces at their disposal.
Alternatively, how does delegation allow principals to do, or to appear to do, moral or worthy things in the face of the temptation to do otherwise? To elevate the discussion to the international level, why do member states of the United Nations delegate
authority to a human rights regime, tying the hands of their own security forces?
An understanding of delegation simplifies a myriad of important and seemingly separate problems in private and public life. Attention to delegation theory helps dispel myths about, for example, the military doctrine of command responsibility and democratic accountability for abuses of human rights and provides insights on a wide variety of issues, be it corruption in FIFA or the church’s child abuse scandal. 4
The approach in this book is to use diverse examples to show the variety of incentives at work and the general application of a wider logic of application. Such a strategy might be described as “cherrypicking.”11 It is certainly not random. It is also true that some cases are invitingly ripe for analysis. And there always might be other unpicked cases, out of view and awkward for the argument. The important task is not to forego illustration but to describe the argument clearly and draw out its implications in a way that opens it to empirical challenge and to further empirical research that would add confidence to the claims.
My selection of examples is driven by three concerns. First, I want to show that the theory has application to a wide range of social and political, as well as economic, real-world issues, from day care, to corporate corruption, to the Catholic Church. Much of the work in principal-agent theory is abstract. Rational choice scholars point out: “too little attention is paid to a central empirical issue: What realworld institution or process is being modeled by a particular formulation? . . . many modelers (present company included!) have been rather casual about the prima facie plausibility of their models.”12 I want to establish the prima facie plausibility of my arguments. Second, I want to show real-world applications that are of current concern and so I pick from the headlines. Third, I select examples that illustrate the theoretical processes at work, that are
diagnostic, and that reveal puzzles unanticipated in the models. With regard to war crimes in the military, for example, I choose the least likely case for agents not to be held accountable by principals: the civilian-controlled armies of long-term, rule of law, Western democracies. While I cannot examine all democratic armies engaged in conflict in searching for cases awkward for the argument, I point out what sort of evidence would disconfirm my argument. In these ways, the selection of real-world examples helps develop our understanding of the logic of delegation and illustrate its usefully wide application. 5
The principal’s incentives to delegate inform the structure of this work. While there is no definitive list of incentives, and others develop further subcategories, the effort is to contribute to a broader logic of delegation by examining incentives to delegate beyond efficiency gains. These incentives match different types of tasks, as illustrated in Table 1.1
Table 1.1 The Principal’s Incentives and the Agent’s Tasks
Menial Tasks
Complex Tasks
Judicial Tasks Chapter 4
Affirming Tasks Chapter 5
Controversial Tasks Chapter 6
The most familiar incentives to delegate are to save time and effort and to obtain expertise. We outsource menial and disliked tasks in order to take the opportunity to spend our time more pleasantly or
lucratively Chapter 2 examines the incentive to save time and effort in individual and organizational contexts. For parents, it might involve finding others to look after the baby, or with surrogacy, even to make the baby. For shareholders, it involves finding managers to take proper care of their investment. For citizens, it involves electing representatives to run the government.
Some tasks are complex. We seek an agent with expertise to take out a tooth or a tumor. The challenges of delegating to professionals with specific and difficult-to-replace knowledge and training are analyzed in Chapter 3 The chapter moves from the use individuals make of professionals of one sort or another, to the use religious institutions and governments make of those with specialized training. It shows the advantages these agents extract from their positions, illustrated by control problems in the Catholic Church and in the army. Within the principal-agent literature, professionals are in good standing. Yet when these agents go bad, they may be very difficult to control. This chapter modifies the existing account with the theoretical argument that group loyalties, asset specificity, and what I refer to as the agent confidence factor protect these noncompliant agents from the normal consequences of their actions.
We delegate in order to resolve important disagreements. In a “friendly” tennis match, we live with the bias of our opponent, but when trophies are involved, we give the task of making the call to an umpire. When marriages get into difficulties, the couple task a therapist to get them back together, or they appeal to a judge to part. Whether disciplining players in the NFL or settling disputes between governments in the World Trade Organization, we turn to an agent to sort it out. Chapter 4 discusses the use of delegation to solve disagreements. Some depart from principal-agent theory to replace the concept of agent with that of trustee to describe the autonomy in this type of delegation relationship.13 We can manage without this complication.
Rather than the biases of others, as Chapter 5 discusses, at times it is our own biases that we do not trust and that persuade us to delegate. There are circumstances where the principal doubts her self-control. She wishes to protect her long-term interests from ephemeral whim and desire. Or she may not lack willpower, but she
expects others may doubt her determination to exercise appropriate self-control. With such doubts about, she turns to someone else, a respected agent, to affirm her trustworthiness to a valued audience. In this chapter, the incentive to delegate is to deliver on a commitment. The task for the agent is to affirm the principal’s trustworthiness, despite other priorities and temptations. For a healthier future, and to keep us from short-term indulgence, we give a coach or a trainer authority over our diet and exercise regime. Or a football player finds an agent to look after the money he makes for the protracted period when no longer playing. The Argentina and Barcelona soccer star Lionel Messi used his father for this task. The pair ended up on tax fraud charges in a Spanish court. Or governments hand over the control of interest rates to central bankers in order to commit to a sound economic policy for the long term. They protect themselves from the temptation of a monetary stimulus for short-term electoral gains. More surprising at the international level, governments commit to human rights treaties and agencies, ostensibly giving up repression when dealing with opponents, dissenters, and critical journalists. While third-party or “fire-alarm” monitoring of these commitments by civil society organizations such as Amnesty International provides some added robustness to this commitment device, as the real-world application to human rights makes clear, fire alarms can be tampered with in a way unanticipated in principal-agent models.
Finally, there is the desire to evade blame for a controversial action. In Chapter 6, passing the buck is the incentive to delegate. The eighteenth-century Scottish economist Adam Smith did not underestimate the “horror of blameworthiness,” which allows “neither quiet nor repose . . . from which no assurance of secrecy can protect them, from which no principle of irreligion can entirely deliver them, and from which nothing can free them but the vilest and most abject of all states, a complete insensibility to honor and infamy.”14 To avoid this horror, we delegate to a “fall guy” lower down in the organization. This task may well be hidden from the agent and not spelled out in the contract. Principals may have multiple tasks for an agent and various incentives to delegate, and these may change in the course of the delegation relationship. Blame is a task that the principal
opportunistically allocates the agent when the unanticipated contingency of a whistle-blower appears on the scene and wrongdoing becomes visible. Volkswagen software engineers, hired to write code, were tasked with the blame when the scandal over diesel emissions erupted.
The question Why Delegate? provides an easy way into a relationship of great practical and theoretical consequence. There is more to delegation than meets the eye. We cannot take for granted that we know who is doing what for whom.
1 United States District Court, Southern District of New York: Sharif Stinson et al., against City of New York. https://www.clearinghouse.net/chDocs/public/PN-NY0012-0002.pdf.
2 “New York City to Pay Up to $75 Million over Dismissed Summonses,” New York Times, January 23, 2017. https://www.nytimes.com/2017/01/23/nyregion/new-york-city-agrees-to-settlementover-summonses-that-were-dismissed.html.
3 Jon Bendor, Amihai Glazer, and Thomas H. Hammond, “Theories of Delegation,” Annual Review of Political Science 4 (2001): 235
4 See Kathleen M Eisenhardt, “Agency Theory: An Assessment and Review,” The Academy of Management Review 14, no 1 (1989): 57–74 for a review of the early literature See also Jean-Jacques Laffont and David Martimort, The Theory of Economic Incentives: The Principal Agent Model (Princeton, NJ: Princeton University Press, 2002); Bengt Holmström, “Pay for Performance and Beyond,” American Economic Review 107, no. 7 (2017): 1753–1777. There are excellent applications to particular policy areas, e.g., John Brehm and Scott Gates, Working, Shirking and Sabotage (Ann Arbor: Michigan University Press 1997), and to the bureaucracy, e.g., Gary J. Miller and Andrew B. Whitford, Above Politics: Bureaucratic Discretion and Credible Commitment (Cambridge: Cambridge University Press, 2015). Others have applied principal-agent theory to parliamentary government, e.g., Kaare Strom, Wolfgang C. Müller, and Torbjörn Bergman, Cabinets and Coalition Bargaining: The Democratic Life Cycle in Western Europe (Oxford: Oxford University Press, 2008), and to international organizations, e g , Darren Hawkins, David Lake, Daniel Nielson, and Michael J Tierney, Delegation and Agency in International Organizations (Cambridge: Cambridge University Press 2006)
5 I thank an anonymous reviewer for challenging me about the “puzzle for the rational choice eye.”
