Preferences and utility

Compare ordinal utility representations and distinguish preference rankings from interpersonal welfare claims.

Paper packet. Every task here also exists on screen, where it is checked automatically; answers written on paper are not assessed by Nydus. When you are back at a device, enter your answers there.

1. What you will learn

You will compare ordinal utility representations and distinguish preference rankings from interpersonal welfare claims, showing the calculation and stating the assumptions that make the conclusion valid.

2. Starting point

Use the supplied definitions and units. Separate an accounting identity, a behavioral assumption, and a normative criterion before drawing conclusions.

3. Terms to use precisely

4. A preference relation comes before its numerical representation

A consumer-choice model begins with a way of comparing feasible bundles. A bundle might specify quantities x and y of two goods in the same period. Saying that bundle A is weakly preferred to B means A is at least as good as B according to the modeled decision maker's preferences. Strict preference and indifference are distinct possibilities. The relation concerns that decision maker and the specified objects, not an automatic social ranking.

Completeness means that the model permits comparison of every pair in its domain. Transitivity means that if A is at least as good as B and B at least as good as C, then A is at least as good as C. These assumptions help make choice coherent and representable. They are analytical conditions, not an instruction to describe every observed person as perfectly consistent or to dismiss context-dependent behavior without investigation.

A utility function assigns numbers to bundles so that higher numbers represent more preferred bundles. It is a representation of an ordering. The numerical labels do not automatically measure units of happiness, moral worth, or a quantity that can be compared between different people. The distinction between a useful representation and a literal psychological measurement is central to interpreting the model responsibly.

For a bounded example, suppose U(x,y)=xy on strictly positive quantities. Bundle (2,6) receives twelve, while (3,5) receives fifteen. The function ranks the second bundle above the first. That conclusion follows from this particular representation. It does not show that every person prefers the second bundle, nor that the improvement is three objectively measurable units of wellbeing.

Term	What it means
Weak preference	An ordering in which one bundle is at least as good as another for the modeled person.
Ordinal utility	Numbers representing a ranking without an automatic cardinal or interpersonal interpretation.
Strictly increasing transformation	A relabeling that preserves every strict ordering and indifference on the relevant range.
Indifference curve	Bundles assigned the same preference rank.
Marginal rate of substitution	A local trade-off between goods holding represented utility constant under suitable smoothness conditions.
Pareto comparison	A comparison based on each person's ordering rather than an arbitrary sum of ordinal utility levels.

Another way: Ordinal transformations preserve the ordering

If a strictly increasing function g is applied to a utility representation U, the transformed representation V=g(U) preserves the same preference ordering. Higher U values still become higher V values, and equal U values remain equal. For instance, V=5+2U ranks every pair exactly as U does. The level and scale of the printed numbers change without changing the underlying choices implied by the ranking.

This result explains why utility differences and ratios generally lack an ordinal interpretation. If one bundle has U twelve and another fifteen, their numerical gap is three. Under V=5+2U, the gap is six. Under a different strictly increasing transformation, it can change again. The preference ranking is preserved, so the original gap cannot by itself be a uniquely measured amount of satisfaction gained.

The transformation must be strictly increasing over the relevant utility range. Squaring preserves order when the relevant U values are nonnegative, but it does not preserve order over the entire real line. Negative three squared exceeds negative two squared, reversing their original numerical order. A transformation is not valid merely because it looks mathematically simple. Check its monotonicity on the domain actually used.

A decreasing transformation reverses the ordering unless the interpretation is correspondingly changed. Multiplying U by a negative number and continuing to maximize would generally produce different choices. Likewise, a transformation that is flat over part of the range can create indifferences not present before. The precise ordinal-invariance statement uses strictly increasing transformations, not arbitrary relabeling of utility numbers.

Another way: Indifference curves describe equal-ranked bundles

An indifference curve collects bundles with the same represented utility. For U=xy, the bundles (2,6), (3,4), and (4,3) all have utility twelve and lie on the same indifference curve in the positive domain. The curve is a set of bundles, not a claim that the consumer chooses all of them simultaneously. A budget or another feasible-set restriction is needed to identify which of them can actually be selected.

Where preferences are smooth and both marginal utilities are positive, the marginal rate of substitution describes a local willingness to exchange one good for another while remaining indifferent. With x horizontal and y vertical, its positive magnitude is MUx divided by MUy. For U=xy, MUx=y and MUy=x, giving y/x units of y per additional x locally. The signed slope of the indifference curve is the negative of this magnitude.

A strictly increasing differentiable transformation with positive derivative scales both marginal utilities by the same factor, leaving their ratio unchanged. This is another way to see why the economically relevant local trade-off can remain meaningful even when utility levels do not. The derivation requires the differentiability and nonzero conditions stated; not every preference representation is smooth at every point.

Perfect substitutes, perfect complements, and satiation can produce different shapes or corners. The familiar smooth convex curve is not the definition of a preference relation. For complements, a kink can make a single derivative-based rate inappropriate at the corner. Later choice analysis must accommodate the actual representation rather than forcing every consumer into one convenient graphical pattern.

Another way: Additional preference assumptions have separate roles

Monotonicity means that, in the relevant sense, more of a good is not worse when other goods are held fixed. Strict monotonicity supports choosing on a budget boundary when more can be obtained and no other constraint intervenes. It is an assumption about the modeled goods and domain. A bad, a satiation point, or an unwanted excess can violate the simple 'more is better' description.

Convex preferences favor mixtures relative to extremes in a particular ordering sense. In common smooth models this can support an interior balance between goods, but it does not guarantee that every constrained optimum is interior. Prices, income, and the exact shape of preferences can produce a corner. Do not convert a general taste for variety into a claim that every purchase includes positive quantities of every good.

Local nonsatiation means that arbitrarily nearby improvements are available around a bundle. It is weaker than requiring every additional unit of every good to be desirable. This distinction matters in welfare and consumer-theory results, where a theorem may rely on a specific assumption rather than every intuitive property associated with ordinary goods. Naming the assumption accurately prevents a proof from claiming more than it establishes.

Completeness, transitivity, monotonicity, convexity, continuity, and differentiability should not be treated as interchangeable labels for rationality. Each does a different job. A numerical exercise may stipulate a utility function that already implies some of them on its domain. A conceptual interpretation should explain which conditions are relevant to the inference, instead of assuming that the presence of a formula proves all desirable properties.

Another way: Represented choice is not an interpersonal welfare sum

A utility representation for one person cannot generally be added to another person's arbitrary ordinal utility numbers to obtain a uniquely meaningful social total. One person's representation could be multiplied by a positive constant without changing any of that person's preferences, yet the numerical sum and its maximizing allocation could change. An interpersonal welfare calculation therefore needs additional normative and measurement assumptions.

This does not make all welfare analysis impossible. Pareto comparisons can ask whether each person weakly prefers one allocation and at least one strictly prefers it, without adding arbitrary utility levels. Other social objectives can be defined explicitly, but they must state how individual outcomes are compared or weighted. The distinction is between an acknowledged additional assumption and an unnoticed use of an ordinal scale as a common measuring rod.

Observed choices can provide evidence about preferences only alongside information about feasible options, prices, constraints, and context. A person who buys a small quantity may face a tight budget rather than dislike the good. A choice from one menu does not reveal how every unavailable alternative would be ranked. Economic interpretation should separate the model's latent preference ordering from the observations used to infer it.

For the exercises, evaluate the supplied function at each labeled bundle, identify the highest-ranked bundle, and apply the stated increasing transformation. Check that the maximizing label remains unchanged. Then explain the limit: the transformed numbers can differ while the modeled ordering remains identical. This combination of calculation and interpretation is the lesson's goal, rather than treating a utility table as a literal report of happiness.

Another way: A preference representation is not an observation by itself

Writing down U=xy is a model specification. To evaluate whether it describes observed choices, a researcher must examine the available menus, constraints and repeated decisions. A good fit on three selected bundles need not establish the same ranking over every possible bundle. Distinguish mathematical implications of the representation from empirical evidence supporting its use for a particular population and setting.

5. Two analysts encode the same ordering

A fictional research exercise gives one consumer the utility representation U(x,y)=xy over three positive bundles. A is (2,6), B is (3,5), and C is (4,3). Their represented utilities are twelve, fifteen, and twelve. B is strictly preferred to the other two, while A and C are indifferent under this model. The bundle labels are names, not rankings.

A second analyst uses V=7+3U. The transformed values are forty-three, fifty-two, and forty-three. The numbers and gaps are different, but the ranking remains B above A and C. Both analysts represent the same preferences. A report claiming that the second analyst found the consumer three times happier would mistake a change of numerical representation for a change in wellbeing.

Suppose a third analyst proposes squaring all utility values. On the present positive range this also preserves the ordering, but that conclusion depends on the range. If a different model used negative utility labels, squaring could reverse their order. The analyst should state why the proposed transformation is increasing on the relevant domain rather than assume every formula is an admissible ordinal transformation.

Finally, the exercise does not say which bundle is affordable. A consumer-choice prediction requires a feasible set as well as a preference ordering. If B is unaffordable, the fact that it has the largest value in this list does not make it a feasible choice. The next lesson joins the ordering with a budget, preserving the distinction between what is preferred, what can be obtained, and what the numerical utility scale actually means.

6. Check the tempting inference

Utility levels and gaps are not automatically measurable happiness or interpersonal welfare. A transformation must be strictly increasing on the relevant range; squaring is not order-preserving over all real values. A preferred bundle can be infeasible, and a smooth utility representation does not justify every additional assumption about preferences.

7. Evaluate and rank three bundles

8. Transform the scale without changing choice

9. Check a transformation's domain

10. Finish an ordinal transformation

11. Guided practice

Use U=xy on A(2,6), B(3,5), C(4,3). Transform with V=5+2U. Give the highest-ranked label, its U and its V.

12. Guided practice

U=xy evaluates A(2,5), B(3,4). V=3+2U. Complete the ranking and transformed value for the preferred bundle.

13. Guided practice

Use U=xy on A(4,4), B(2,7), C(3,6); V=1+3U. Give the highest-ranked label and its two represented values.

	Your result
Highest-ranked bundle label
Its U value
Its transformed V value