The philosophy of microeconomics encompasses various theories and approaches that seek to understand the principles, assumptions, and implications of individual decision-making within the context of markets and economic systems. Some key theories in the philosophy of microeconomics include:

Rational Choice Theory: Rational choice theory posits that individuals make decisions by maximizing utility or satisfaction given their preferences, constraints, and available information. It assumes that individuals act in their self-interest and make choices that maximize their well-being.

Marginalism: Marginalism examines how individuals make decisions at the margin, weighing the benefits and costs of small changes or incremental units of goods and services. It emphasizes the importance of marginal analysis in determining optimal decision-making and resource allocation.

Utility Theory: Utility theory explores the concept of utility as a measure of satisfaction or happiness derived from consuming goods and services. It investigates how individuals allocate their limited resources to maximize utility, subject to budget constraints and preferences.

Consumer Choice Theory: Consumer choice theory analyzes how consumers make decisions about what goods and services to purchase based on their preferences, budget constraints, and the prices of goods in the market. It explores consumer behavior, demand curves, and the determinants of consumer choice.

Production Theory: Production theory examines the behavior of firms and producers in allocating resources to produce goods and services. It analyzes the relationship between inputs (such as labor and capital) and outputs, the concept of production functions, and the factors influencing production decisions.

Market Equilibrium: Market equilibrium theory explores the interaction of supply and demand in determining prices and quantities exchanged in markets. It examines how markets reach equilibrium through the adjustment of prices and quantities to balance supply and demand.

Game Theory: Game theory studies strategic interactions between rational decision-makers, such as individuals, firms, or governments, in competitive or cooperative settings. It analyzes the outcomes of strategic interactions, including the Nash equilibrium, cooperation, and competition.

Information Economics: Information economics investigates the role of information and uncertainty in economic decision-making. It examines how individuals gather, process, and act on information in markets, the impact of asymmetric information on market outcomes, and the role of signaling and screening mechanisms.

Behavioral Economics: Behavioral economics integrates insights from psychology and economics to study how cognitive biases, heuristics, and social factors influence economic behavior. It challenges the assumptions of rationality and explores deviations from standard economic models.

Welfare Economics: Welfare economics evaluates the efficiency and equity of resource allocation in economic systems. It assesses the welfare implications of market outcomes, including market failures, externalities, income distribution, and the role of government intervention.

Henry Bolingbroke, Duke of Hereford

One problem I often come across is the presumptuousness with which certain observed behaviours are called irrational, "irrational" here used in the economic sense of utility-nonmaximising. For instance, if someone gives you two options, (A) $80 for certain or (B) a 90% chance of $100, many people would probably choose A. It's common for people to believe that anyone who chooses A is acting irrationally. Such behaviour is chalked up to a "bias" in human reasoning. Only through education in statistics can someone overcome this irrationality.

However, anyone who's familiar with utility theory in the slightest knows this is nonsense. It's quite consistent, for someone with diminishing marginal utility of income, to prefer A to B, since a fall in income by $1 causes a bigger fall in utility than a gain in income by $1 causes a gain in utility. Therefore lower variance is preferred to higher variance.

Another example is time preference. It's often assumed without any real argument that a rational person has a consistent time preference ratio thoughout time. But I think this is intuitively wrong. Knight's discussion in chapter four of Risk, Uncertainty and Profit I thought was quite lucid:

[I]t is reasonable and common for human beings to prefer future goods to present, within limits, as compared with a uniform distribution in time. Most civilized persons, in fact, plan for a rising standard of living through life rather than a constant, much less a falling one. But when infinite time comes under consideration the case is different.

Any finite amount of consumption or enjoyment distributed uniformly through infinite time becomes a zero rate of real income. Hence there must be an apparent discount on the future in the demand for perpetual income goods.

What motivated this post was seeing how the Allais Paradox was cited as proof of people's "biases" and "irrationality." The "paradox" is as follows. In two scenarios people can choose between option A or option B.

Scenario 1:

1A.  $24,000, with certainty. 1B.  33/34 chance of winning $27,000, and 1/34 chance of winning nothing.

Scenario 2:

2A. 34% chance of winning $24,000, and 66% chance of winning nothing. 2B. 33% chance of winning $27,000, and 67% chance of winning nothing.

Expected utility theory is the theory that agents maximise their expected utility, "expected" used in the statistical sense (the mean). According to a certain interpretation of expected utility theory, if you prefer 1A to 1B then U($24,000) > (33/34)*U($27,000). And if you prefer 2A to 2B, (34/100)*U($24,000) > (33/100)*U($27,000). It should be clear that preferring 1A to 1B is equivalent to preferring 2A to 2B in this framework, since the two inequalities can be rearranged to produce one another.

The "paradox" is that most people in experiments choose A in scenario 1 but B in scenario 2.

The Allais Paradox is no doubt inconsistent with that certain interpretation of expected utility theory, but I don't think such behaviour can be called "irrational" at all. Notice a key assumption: the utility function's only direct dependence is on the income. All aversion to higher variance is explained by the diminishing marginal utility of income.

However, if a direct dependence on variance is introduced (i.e. E(U(Y)) is replaced by U(E(Y),σ^2)) then the algebraic inconsistency goes away.

I think this is not just a fudge factor thrown in to save an unworkable theory. After all, if you really grill people on why they make their preferences it's clear that the difference in variance between 1A and 1B is very large whereas 2A and 2B have similar variance, and this fact is what informs their preference. The ratio of the variance of 1B to 1A is infinity whereas the ratio of the variance of 2B to 2A is about 1.25.

What people should be drawing from the Allais Paradox is that diminishing marginal utility cannot wholly explain the preference for lower variance over higher variance. However, as in the previous examples, the Allais Paradox is treated as a proof of biases in intuitive human reasoning. Even Eliezer Yudkowsky takes this position. I think this is a mistake.

While we may gasp at the ridiculousness of preferences, internally consistent preferences are by definition never irrational, only the means to satisfy those preferences can be irrational.

EDIT #1: Isn't what I'm saying in some sense a departure from expected utility theory? Yes, though not a radical departure. Instead of agents maximising expected utility, agents maximise plain-old utility. In my opinion the assumption that expected utility should be maximised is somewhat ad hoc and can't be derived from utility maximisation (why not maximise median utility?).