In the early neoclassical era, the smooth adjustment of agents’ marginal utilities dampened the response of market prices to supply disturbances. But when agents do not make smooth trade-offs, they will tend consume goods in specific combinations, as in the model of safety bias. A negative shock to supplies can then send prices soaring, a common feature of energy markets for example. Postwar general equilibrium theory permits agents to be safety-biased but has dismissed these volatility scenarios on the grounds that they occur only at vanishingly unlikely output levels. The no-volatility conclusion fails to hold however once the production of goods is taken into account: the seemingly unlikely output levels arise systematically. Production on the other hand spells out countervailing forces that curb the most extreme cases of price volatility. Volatility is closely related to the indeterminacy of equilibria and the chapter critiques the claims of general equilibrium theorists that these phenomena are unlikely.

In the early years of neoclassical economics, agents were pleasure-seekers and their marginal utility for a good diminished continuously as consumption increased. This psychology ensured that agents could optimize and identify smooth trade-offs. Later generations abandoned this utilitarian inheritance in favor of a theory of rationality, but they had no replacement explanation for why agents can always order their options from best to worst. The same pattern has recurred in the theory of choice under uncertainty. In early theories, agents have preferences over lotteries where outcomes have known or objective probabilities. But once economists recognized the primacy of idiosyncratic risks, they turned to theories where agents form preferences based on their judgments of the likelihood of outcomes. These theories of subjective probability cannot explain why agents are capable of making these judgments. And if agents cannot pin down subjective probabilities they will not able to order their options. More broadly, agents face different classes of decisions, some where they are guided by an ordering principle that shows how to weigh their options and others where agents cannot judge trade-offs and form preferences.

Production rather than preferences should play the leading role in the theory of markets, as in the book’s analysis of volatility and policymaking. Production is not only the fount of social wealth, it is organized by firms whose decisions are guided by a clear ordering principle, their profitability. There is moreover an institution, the market, that can weed out the firms that fail to pursue profits effectively. But the purging of inefficiency by the market confronts policymakers with the dilemmas of Schumpeterian creative destruction. Policies that compensate firms and their owners for any declines in profitability will also dampen the threat of bankruptcy; the enforcement of productive efficiency can require unfettered competition. An application to international trade illustrates some of the challenges: For foreign competition to lead to productive efficiency, domestic prices must be aligned with world prices, which will push some firms into liquidation.

Individuals can rationally pursue their interests without the preferences and marginal utilities that have long taken center stage in economics. Economics without preferences lays out the microeconomics of individual behavior, markets, and welfare when agents cannot always come to judgment. Although economic theory has claimed that self-interest requires agents to form preferences, individuals can protect themselves from harm by refusing to trade options they cannot rank. Many of the anomalies uncovered by behavioral economics – from status quo bias to loss aversion – thus have a rationality design. The absence of preferences also resolves the puzzle that classical economic agents are almost never indifferent between options whereas real-world agents often are. When individuals cannot judge trade-offs, gaps appear between the marginal valuations of gains and losses. These gaps explain why market prices can be volatile and render orthodox efficiency criteria indecisive. Policymakers will no longer be able to pin down an optimal provision of public goods. Traditional schemes that try to harness preference information to compensate agents harmed by economic change will allow virtually any decision to qualify as efficient. Governments should instead spur productivity growth, the main benefit capitalism can deliver, while shielding agents from the price upheavals that result.

When agents, due to incomplete preferences, fail to have well-defined marginal valuations for goods, a great many government policies will maximize social welfare or achieve efficiency. Welfare economics then becomes useless as a practical guide to decision-making. For example, the values agents assign to increases in a public good will be discretely smaller than the values they assign to decreases. For society as a whole, a large valuation gap will form and a wide range of quantities of the public good will therefore qualify as optimal. Applied welfare economics and cost–benefit analysis bypass this obstacle by paying attention only to agents’ smallest valuations, thus slanting policymaking against public goods. The multiplicity of preferences that agents view as reasonable also neuters Pareto efficiency as a policy guide: virtually any policy change is likely to harm some of the preferences agents deem reasonable.

The social welfare function furnishes the primary tool of normative economics: It aggregates the utilities of different agents, by summing them for example. That technique is no longer available when preferences are incomplete since agents then cannot be modeled by utility functions. Agents can however have well-defined utility functions for groups of goods, though they will not know how to weigh the functions for different groups against one another. A policymaker can aggregate these utilities across agents and thus pin down a unique normatively optimal allocation for each group of goods. Government policies are usually debated in this fashion. Rather than solve a global welfare optimization problem, governments and advocates attack each domain of policymaking separately, whether it be education or health. The chapter’s approach illustrates Sen’s criticisms of welfarism.

When preferences are incomplete, an agent or policymaker cannot order options from best to worst. Decisions and policymaking are then slanted in favor of the status quo. Individuals and institutions are governed by customary decisions, until a new option appears that allows for an unambiguous improvement. The reshaping of preferences provides a rarely explored escape hatch to this conservatism and is illustrated by how the flexibility of preferences can cure Baumol’s cost disease (low productivity growth in services).

An agent unable to decide how to trade-off rival goods can turn to a family of welfare functions and choose the option that guarantees the greatest welfare level. The agent will then display a bias in favor of the ‘safe’ options to which all of the welfare functions impute the same level of welfare. When the rival goods are ordinary items of consumption, the safety-biased preferences that result will exhibit complementarities. When the rival goods are commodities delivered at different states of nature, safety-biased preferences will generate the maxmin model of choice under uncertainty. In any setting, the agent’s sets of preferred options will display kinks at the safe options but not at other options. Such agents will therefore tend to consume at just the points where slight changes in consumption have a large effect on agents’ marginal valuations of goods and hence on prices.

Agents are classically considered rational in economics if their preferences satisfy the completeness and transitivity axioms. But no matter how preferences are defined, an agent can always violate these axioms without making decisions that leave the agent worse off. If preferences are defined by an agent’s welfare judgments, those judgments can be incomplete, while if preferences are defined by an agent’s choices then sequences of those choices can be intransitive. In both cases, agents can shield themselves from harm simply by maintaining the status quo unless offered an unambiguously superior option. Agents will then display several of the anomalies discovered by behavioral economists, including the endowment effect, loss aversion, and the willingness-to-accept/willingness-to-pay disparity. These behaviors are therefore consequences rather than violations of rationality. Moreover, a single set of preference judgments can explain all of the choices an agent makes through time; the theory therefore wields predictive power. Behavioral economics in contrast courts unfalsifiability by positing a separate preference for an agent at every decision-making juncture.

Preferences are incomplete when agents cannot rank some of their options. Incompleteness can arise when individuals believe that multiple utility functions or preference relations are reasonable ways to make decisions. When these potential preferences conflict, an agent will fail to come to a bottom-line judgment. Faced with such conflicts, agents can still specify the options preferred to some reference point but that set will be kinked at that point. Agents will therefore fail to have well-defined marginal valuations for goods: they will sacrifice a unit of a good only when offered a large amount of another good in exchange. The chapter also presents a more abstract model of incomplete preferences that begins with kinks in preferred sets of options rather than deriving them from a multiplicity of potential preferences. Indecisiveness and kinks also appear in decision theory when agents are unable to identify unique subjective probabilities.

This chapter presents an operational test of indifference versus incompleteness. Indifference obtains when two alternatives have the same sets of more preferred and the same sets of less preferred options. This definition turns out to be equivalent to declaring indifference when a trade of alternatives cannot convert a harmless sequence of trades into a sequence that takes an agent from a better to a worse option. Incompleteness obtains when a trade of unranked alternatives makes a harmful sequence of trades possible. The distinction resolves the puzzle that agents frequently cannot strictly rank alternatives even though classical economic theory claims that indifference is rare. When an agent’s preferences are allowed to be incomplete, pairs of alternatives where no preference obtains will abound. Indifference on the other hand nearly disappears.

The problems that afflict Pareto efficiency can be overcome if the criterion is rebuilt on preference-free foundations. A policy change passes the ‘availability test’ if it allows agents to afford whatever they purchased originally: Agents might not then be better off but no one can legitimately object to the change. One way to pass the availability test is to give agents the right to repeat their original transactions; a reform of rent control serves as an example. A second strategy stabilizes prices for consumers while letting the prices that firms face promote efficiency in production. A deregulation of a public utility, for example, can preserve consumer prices while giving firms an incentive to innovate. These policy alternatives show how to resolve the Schumpeterian dilemma of creative destruction: They harness the progressive feature of capitalism, that it fosters technological change, while protecting the individuals who can be harmed by the same forces. Conventional laissez-faire policies are in contrast difficult to justify even from within the orbit of traditional economic theory and can generate bitter social conflict. An application to opening an economy to free trade shows how to combine the advantages of technological change while satisfying the availability test.