Moral Conditions of Economic Efficiency Book

Recommended by EJMR as I looked for good philosophy readings.

Sen in this essay criticizes the implicit assumptions grounding the First Welfare Theorem and other assumptions of welfare economics. Sen argues that ethics and economics have gotten too disconnected to the detriment of both. Economics had its start in two traditions: an Aristotelian stew in which concepts which we would today call ethics, politics, and economics were all mixed up and one in which economics is viewed as separate and instrumental. After the first generation of modern economists such as Adam Smith and John Stewart Mill, for whom morality was central, the engineering approach has been emphasized.

Sen examines the assumptions economics makes about human behavior. A core assumption of the vast majority of economics is that people behave rationally. Sen breaks rationality down into two sub-concepts: consistency (or rationalizability) and self-interest maximization. He later breaks self-interest maximization into sub-concepts as well.

A rationality is the system of rules by which an agent decides on an action given their preferences, beliefs, and options. It may also include rules restricting the types of preferences allowed and how beliefs are formed. Economic models assume many subtly different versions of rationality in different contexts. But the standard bare-bones version entails that an agent have complete and transitive preferences. Complete preferences means being able to make a choice between any two options that might conceivably be presented. Transitive preferences mean that if an agent prefers A to B, and B to C, he prefers A to C. While a person with intransitive preferences is conceivable, he might find it very difficult to go shopping because he'd cyclically replace stuff in his shopping cart with other preferred stuff.
A neat thing about rational preferences is that they can be represented as mathematical 'utility functions'. An individual doesn't need to be conscious of their utility function: to be 'consistently rational' they just need to act as though they were consciously maximizing it. Further, for a person to be consistent, their preferences don't necessarily have to have any connection to their good or pleasure or anything. If a suicide bomber or a drug addict acts sensibly in pursuit of their goals, they might be acting consistently, but you might not think they are acting in their self-interest.

The other definition of rationality economists work with is self-interest maximization. Why do economists want to assume that the utility function people maximize represents their well-being? First, factually, self-interest seems to be at least a powerful motivation. More fundamentally, if we want to be liberals and not automatically assume that the good for everyone is the same, then we need some other access to what individuals' goods' are. It makes sense to assume that individuals have better access to what their own flourishing entails than any outsider. Letting individuals tell us what their good is through their revealed preferences allows economists to dodge hard utilitarian problems of higher versus lower pleasures and the like. We don't have to decide what the good life entails: the consumer gets to. Sen argues that self-interest maximization, while certainly compatible with rationality, is not required by it.

Sen then moves from positive economics to normative economics. Sen argues welfare economics is grounded in a version of utilitarianism. Sen divides classic utilitarianism into three parts:
a) Welfarism - A state's goodness is completely described by its utility information
b) Sum Ranking - The best state is the one that has the highest total utility
c) Consequentialism - A choice should be judged based on its forseeable consequences

Sen thinks economics' strength is its embrace of consequentialism, as any moral theory must employ at least some consequential reasoning. For example, even if rights are intrinsically valuable, we still need to be consequentialists to deal with situations where different rights come into conflict.

As for sum-ranking, in PEL's utilitarianism episode, one of the huge problems that arose was inter-personal comparisons of utility. Utilitarianism sounds great until you are actually faced with practical problems of weighing the relative merits of saving one man's life versus the inconveniencing ten thousand people. Utility monsters also lead to paradoxes. And Arrow's Impossiblity theorem tells us that there is no voting system that can sensibly decide between a set of pareto efficient options. Engineering economics dodges these hard problems by dropping the sum-ranking criteria. Economists only feel on solid ground normatively when advocating for pareto efficiency. But Sen thinks the pareto criteria is too conservative. Extremely unequal societies can still be pareto efficiency, so long as the richest guy has some value from his hoard at the margin.

Sen also critiques welfarism. The move economists want to make is to say that people act to maximize their self-interest. By designing policies to get people more of their revealed preferences, economists hope to make them and society better off. Due to his analysis of rationality, Sen thinks extrapolating interests from actions is problematic. His analysis also suggests that there are at least two different things we should care about: getting people more of what they want and getting people more of what's good for them. If these two things don't line up perfectly, then a one dimensional measure of the good is inadequate. Sen is attracted to the idea of other intrinsic values as well. So Sen rejects moral monism. A ranking of social outcomes which squeezes heterogeneous considerations into a unidimensional scale may be desirable, but it is not necessary.

In the years since Sen's lecture, much progress has been made in economics, led by behavioral economists and decision theorists, in understanding a wider range of behavior. Many economists today shy away from using the loaded term 'rational' for any set of behavioral assumptions. However, versions of utilitarianism remain the dominant paradigm for understanding welfare.