What Choices People Would Make in Ignorance of Their Own Personal Interests

Both Rawls's A Theory of Justice, 1971, and my own theory of moral value judgments (see, e.g., Harsanyi, 1953, 1977, chapter 4) can be interpreted as theories that try to answer the question of what social institutions people would choose if their choices were wholly unaffected by their own personal interests.

In Rawls's theory, this question takes the form of asking what social institutions people would choose in the original position where a “veil of ignorance” would prevent them from knowing what their own social positions and even what their own personal characteristics were and therefore from knowing their own personal interests.

In my own theory, this question takes the form of asking what social institutions people would choose for their society if they had to make their choiceson the assumption that each of them would have the same probability n of ending up in any one of the n possible social positions.

Yet, even though the basic questions Rawls and I ask are rather similar, our theories by which we try to answer them are very different. One important reason for this is that Rawls assumes that people in the original position would use the maximin principle as their decision rule, whereas I assume that people making moral value judgments would base their choices on expected-utility maximization in accordance with the Bayesian concept of rationality.

The Maximin Principle

Rawls's use of the maximin principle as a decision rule is rather surprising because it has been known since the early 1950s that it is an irrational decision rule, with very paradoxical implications (see Radner and Marschak, 1954; see also Harsanyi, 1974).

Introduction

The Aggregation Theorem is one of the main arguments used by Harsanyi in support of utilitarian ethics. It was first presented in his 1955 article and further developed in chapter 4 of his book in 1977. Since then, several authors have constructed alternative proofs of this theorem in more general settings. It is generally presented as relating a “single profile” of individual utility functions {Ui}, to the utility function W of a moral observer by means of the Pareto Indifference rule. In this context, the theorem states that if all utility functions (including the moral observer's) are von Neumann–Morgenstern (VNM), then the moral observer's utility is an affine transformation of the individual utilities, that is, W= ΣβiUi + γ.

The relevance of this result in giving proper foundations to utilitarianism has been questioned on several grounds. First, the weights {βi} are not necessarily positive, and hence the welfare of some individuals might not affect, or worse, might negatively affect total welfare. This first problem can be solved quite naturally by strengthening Pareto Indifference into the Strong Pareto condition; the latter implies that all weights are positive. A second problem is that the weights might not be uniquely defined, creating an indeterminacy. This further problem can be solved by introducing an additional condition, called Independent Prospects, which says that for every individual there exists a pair of lotteries for which that individual alone is not indifferent.

RISK IS UNDESIRABLE to most people, especially risk involving household income. However, for the rural poor in developing countries, fluctuations in income pose a particularly grave danger. Suppose the income of a typical North American family fell abruptly by $1,000. Although certainly unwelcome, it would be of little real consequence, perhaps vacation plans changed from a hotel stay to a camping trip. But an income loss of $1,000 for a rural Central American family would be devastating. It would probably necessitate the sale of assets, such as land, that are crucial to the household's long-term livelihood. It could force the migration of a family member. It might even force a change to a cheaper and less nutritious diet: more tortillas, fewer meats and vegetables.

The irony is that while the consequences of risk are more disastrous in developing countries, risk is also more prevalent. To continue the example, despite a far lower base of wealth, an abrupt $1,000 loss in income is arguably more likely for the rural Central American household than for the household in North America. This may be true for several reasons. First, risk inherently follows the uncertainties that surround rural households in developing countries.

JOSEPH KONY IS not an ordinary guerrilla leader. A former priest and witch doctor, he heads the Lord's Resistance Army (LRA), an insurgent movement that roams the jungle areas of northern Uganda and southern Sudan. Regarded as one of the world's most brutal revolutionary movements, the LRA's stated purpose is to overthrow the Ugandan government, replacing it with one that rules the country by the Ten Commandments. Ironically, it is difficult to think of an insurgency that violates the Ten Commandments more thoroughly than the LRA. The modus operandi of the LRA has been to conduct nightly raids on the houses of rural homesteaders, pillaging food and supplies, raping women and young girls, and abducting young boys to flesh out its ranks. The number of its victims is staggering. According to the United Nations, in the two decades since its foundation in the mid-1980s, members of the LRA have abducted more than 20,000 boys and driven more than 2 million people from their homes. When abducted, the boys are regularly forced to kill or maim members of their own families in the hope that they will be ashamed to return home. The LRA then assimilates the abducted boys into its company as guerilla soldiers by brainwashing them with its quasi-religious doctrine.

THIS APPENDIX is here to give you a little more background on the basic solution concepts and techniques used in game theory. A warning: This is only a brief overview, and it is somewhat terse. To delve into these concepts at a more satisfying level, I recommend several books that can serve as excellent introductions to game theory at the end of this section.

A game consists of two or more players, and often we index the players by a number or letter (e.g., 1, 2, 3, …, n), where n represents the number of players in a game. Each player in a game has a set of strategies. For example, in a game of peasant farmers we might represent the set of strategies available to Player 1 and Player 2 as S1 = S2 = {Beans; Coffee} in a two-player, two-strategy game where the players have the same strategies. Any combination of strategies, one by each player in the game is called a “strategy profile.” Here, each player i chooses one of the strategies in her strategy set, or chooses one particular strategy si that is part of Si. Thus a strategy profile for n players is a combination of strategies, one by each player, (s1; s2; s3; … sn). The strategy profiles in our two-player, two-strategy game would be (Beans; Beans), (Beans; Coffee), (Coffee; Beans) or (Coffee; Coffee).

THE DOHA ROUND of world trade negotiations began in November 2001 in an atmosphere of heightened anticipation. The 148 members of the World Trade Organization began the round of talks (named after the capital of Qatar where the kickoff meeting was hosted) with an objective that was both lofty and noble: The Doha round was to incorporate the poorest countries of the world into a free and fair global trade system. The talks sought to rectify a long-identified bias against the poor countries in international trade – the forbidding shield of protection erected by the rich countries to defend domestic farmers against foreign agriculture. The sheer size of agricultural protection in the OECD countries is staggering. Recently calculated at $279 billion, it equals 30 percent of total agricultural receipts, and six times the amount spent on foreign aid to the developing countries.

Serious reductions in agricultural protection in the Doha round would have momentous implications for economic growth in the developing world. Indeed, as the Doha round began, the World Bank had estimated that better poor-country access to rich-country markets would increase world income by $520 billion, and would lift 144 million people out of poverty by 2015.

Benefiting most directly by a Doha trade agreement would be the rural poor in the less-developed countries (LDCs).

WHAT DO TEENAGE ice-cream scoopers, taxi drivers, door-to-door salesmen, and peasant day laborers all have in common? The answer is, at least at work, they all can be difficult to monitor and motivate. Economics considers a fundamental problem in which a principal, who needs a task carried out, hires an agent to carry out the task. The problem is that once the agent is hired, the agent's interests may not match those of the principal. As any supervisor of human resources can attest, a hired worker left unmonitored and unmotivated is a worker tempted by the twin evils of sloth and self-indulgence. The labor supervision problem is a definitive example of moral hazard, the incentive for an agent to act in his own interest rather than the interest of the principal when the agent's actions are hidden. As a result of moral hazard in labor markets, the principal may never offer a labor contract in the first place unless he can design a contract that sufficiently lines up the agent's incentives with his own.

In this chapter, we will examine three different types of contracts between principals and agents: fixed-wage, fixed-rent, and share contracts. We will explore how these different types of contracts address the issues of moral hazard and risk sharing in the context of some industrialized country labor market examples, and then analyze how they shape the agrarian institutions that order rural economic life in developing countries.

IN CHEATING MONKEYS and Citizen Bees, biologist Lee Dugatkin describes the grooming behavior of the impala. (For Americans who think of an impala as gas-guzzling Chevy, the impala, closely related to the gazelle, inhabits the savannah of Kenya and other parts of southern and eastern Africa.) Impala face a problem unfamiliar to humans: They are unable to clean important parts of their body, especially their backs, since they have no arms, and their legs are facing the wrong way. This is a problem particularly with ticks, which are itchy and carry nasty impala diseases. Whereas the rhinoceros solves this problem though his symbiotic friendship with the oxpecker (tick bird), who stands on his back feasting on his ticks, impala are too jumpy for a piggybacking mate. As a result, they usually ask another impala for help. But there are costs in grooming: lapses in vigilance for lions and wild dogs, hairballs, loss of saliva, and so forth. The substantial health benefits of grooming would justify these costs if another impala would reciprocate. Yet what is to keep a “groomee” from bounding off on his own business after the groomer does his work? Here, impala play a Trust game, with a guarantee of reciprocity hard to secure.

WHAT ACCOUNTS FOR the wide spectrum of poverty and prosperity in the world today? This is arguably the most important question in the social sciences, but it has also proven to be one of the most difficult. Many books written about economic development contain a plethora of macroeconomic statistics that document the widening span of the economic chasm between rich and poor. This is not one of those books. There are relatively few statistics in it. You will not find many references here to GDP, macroeconomic growth rates, inequality coefficients, or statistics about hyperinflation. This book addresses this question not by reexamining the statistics on world poverty or looking at the successes, failures, or potential of grand development schemes. Instead, it examines how patterns of human interaction form the basis for poverty and prosperity.

Game theory is a formal structure used to understand human interaction. Because human interaction is both frequent and desirable for most of us, game theory covers a lot of ground: Games occur in social relationships, during market exchange, in the fulfillment of contracts, in the use of environmental resources, in educational and technology choices, in politics, and myriad other aspects of everyday life. By analyzing human interaction in a formal structure, game theory can make predictions about how people will behave and the consequences of their behavior. This makes game theory a powerful tool. It can also give us insight into difficult questions, such as why some countries have become rich and others remain poor.

ECONOMIC DEVELOPMENT, AS we know it, is a relative newcomer to human society. For most of history, the world languished in a kind of economic limbo. Centuries after the early Roman era, world per capita income hardly changed, lingering around $400 per year in both the rich and poor areas of the world. (Ironically, for centuries it remained slightly higher in what is now considered the “developing world.”) The world economy was dormant, technological change was slow, and advances in human welfare were virtually imperceptible, bringing to mind the ageless and changeless millennia of J. R. R. Tolkien's Middle Earth. By the time of the Renaissance, however, per capita income in Europe slowly began to grow – to around $700 by 1500, creeping to $1,100 by the eve of the Industrial Revolution, the beginning of a spectacular economic takeoff.

What finally got the economic ball rolling? The answer, according to many, was the Big Push.

EASTER ISLAND IS one of the most isolated spots on earth, lying 2,100 miles to the west of Chile and 4,000 miles to the southeast of Hawaii. The island is renowned for its massive stone faces that stoically gaze over the Pacific, the moai statues, carved by the native Rapanui half a millennium ago during the zenith of an ancient and mysterious civilization. This same thriving island civilization of perhaps 10,000–15,000 inhabitants once survived on cultivation of sweet potatoes, yams, bananas, domesticated chickens, and, of course, fishing. But by the arrival of Dutch explorers in 1722, the Rapanui had dwindled to a few thousand; by 1877, the entire population of Easter Island had plummeted to just 111 half-starved natives.

In his book Collapse (2005), UCLA geographer Jared Diamond explains the principal blunder of the Rapanui civilization: deforestation. Trees and timber had been vital to the ancient Rapanui. Trees prevented crop erosion and provided a native habitat for birds and animals important for supplementing the local diet. Wood provided raw materials for hand tools, logs used in the erection of the moai statues, fuel for warmth during cool and rainy nights, and most importantly, for constructing fishing canoes. Since Easter Island receives only 50 inches of rainfall per year (scanty by tropical standards), trees grow slowly, leaving the island's inhabitants more vulnerable to the “tragedy of the commons”.

ADAM SMITH'S CONCEPTION of the “invisible hand” postulated that a market consisting of self-interested individuals would yield an outcome that was best for society simply through the natural course of free exchange. One of the most striking contributions of game theory has been to demonstrate that the benevolence of the “invisible hand” is merely a special case, rather than a general truth about the fruits of economic self-interest. We will observe a number of cases, in fact, where the invisible hand can become an angry, malevolent hand, punishing players for their selfish behavior. What we will see is that self-interest is sometimes good for society, but often it is not. Indeed, self-interest may yield very poor economic outcomes outside the discipline of well-functioning institutions.

Game Theory

The origins of game theory are fascinating, and they help us to understand how such important insights came about. The work of French economist Augustin Cournot revealed the earliest glimpses of a formalization of strategic behavior in the 1830s. Cournot developed a famous model of strategic competition between two firms that foreshadowed some of the later insights of game theory. But Cournot was never able to generalize his concept of an equilibrium solution to other contexts, and for many decades, his results were regarded as insightful as applied to only a restricted type of competition between two firms.

ON SEPTEMBER 5, 2006, an electoral court proclaimed Felipe Calderón to be the next president of Mexico. The court ruled that Calderón had won the election, which had taken place two months before, over his rival, Andrés Manuel López Obrador. Calderón's presidential victory was by the slimmest margin in Mexican history, igniting street protests organized by López Obrador's followers that would last for months.

Both before and after the election, the political battle between Calderón and López Obrador grew increasingly tense over the issue of corruption. As political tensions mounted, each tried to paint the other candidate as a contributor to the problem, while simultaneously presenting himself as the best solution to it. López Obrador's accusations were particularly painful for Calderón, who had campaigned under the nickname El Sr. de Manos Limpias (Mr. Clean Hands), pledging to eliminate the scourge of corruption in Mexico.

Presidential pledges to combat corruption, however, were not revolutionary. Vicente Fox had been elected as an anticorruption warrior six years earlier, as had been Ernesto Zedillo six years before him. The tenacity of corruption in Mexico has not kept politicians from promising to eradicate it. Because citizens correctly identify corruption at the root of Mexico's development problems, promising to break its stranglehold on society garners the votes of many. But were Calderón fully aware of the difficulties before him, he might have considered easier alternatives, like changing the national language to Swahili.

IN THE 1960s, Walter Mischel at Stanford University carried out an experiment to test the effects of delayed gratification in children. The subjects of his study were a random sample of 4-year-old children. The children were led into a plain room, one-by-one, where Mischel presented each with a marshmallow on a plate. Children were told that they were free to eat the marshmallow, but any child who waited to eat the marshmallow until Mischel returned from an errand, would receive two marshmallows.

Some of the children immediately crammed the marshmallow into their mouth with Augustus Gloop–like voracity as soon as the researcher left the room. Others were able to wait for a few moments, but then succumbed to the overpowering temptation of the marshmallow. Another group of children engaged in a variety of self-distraction exercises: covering their eyes so they could not see the marshmallow, walking over to sit in a corner, singing, and playing clapping games with themselves. When Mischel returned, he rewarded these children with their second marshmallow. Then he waited for the children to grow up.

What he found fourteen years later was astonishing. The children who had waited for the second marshmallow scored an average of 210 points higher on the SAT than those who couldn't wait.

Chapter Two

1. Consider a game in which there are two peasant farmers. One has land on which it is slightly better to grow corn, and the other has a land on which it is slightly better to grow wheat. They want to buy a simple harvesting machine together that can be used only for corn or for wheat. What kind of coordination game is this? Is it a Stag Hunt, Battle of the Sexes, or a Pure Coordination game? Explain.

2. Suppose that one player abides by one convention over the use of a resource, while a second player abides by another. Show in the context of a Hawk-Dove game how this can produce conflict.

3. Consider an application of the Prisoners' Dilemma to cooperatives. Suppose that a development institution establishes an agricultural cooperative. There are n people in each cooperative, who equally share all of the output. Each person in the cooperative can produce 100 pesos/day, and the effort for each person to produce this amount is given by e. What is the threshold number of people in the cooperative below which people will still choose to Work rather than Shirk?

4. Is it possible to turn a Prisoners' Dilemma game into a coordination game by changing only one player's payoffs? If so, which payoff? If not, what kind of game results when only one payoff is changed?

ONE OF THE fundamental human dilemmas is that individuals potentially stand to gain from competition as well as cooperation with one another. Costly mistakes have been made from failing to recognize the two horns of this dilemma and that the tension between these conflicting incentives underlies nearly all social behavior. For example, this failure has led some to overemphasize the competitive nature of market-oriented societies, where social benefits are misperceived to accrue chiefly from the competitive aspects of markets. Economists have more recently come to understand that the cooperative aspects of market-oriented societies are just as important as their competitive aspects. The institutional constraints that check self-interested behavior are instrumental to the freedoms that allow for the pursuit of self-interested gain.

Institutions dictate the rules of the game in any society. They serve as guidelines for human interaction, and they set the limits for human freedoms. In his Institutions, Institutional Change, and Economic Performance (1990), Nobel laureate Douglas North defines institutions as “the framework within which human interaction takes place.” He draws an analogy between the rules that govern competitive economic activity, and rules that govern competitive sports. In each, the rules of the game create a stable, predictable structure to behavior. Even in a sport as seemingly violent as American football, an elaborate (some would argue too elaborate) set of rules carefully governs play.