Difficulty: Intermediate | Prerequisites: Budget Constraints (Ch. 1) and Consumer Preferences (Ch. 2) study notes.
This is where everything comes together. You know the budget constraint (what the consumer can afford) and the indifference curves (what the consumer prefers). A utility function gives you a mathematical way to represent those preferences and compute the optimal bundle, the best affordable combination of goods. This section covers three main utility function types (Cobb-Douglas, perfect complements, perfect substitutes), how to derive the marginal rate of substitution, and the step-by-step method for finding the consumer's optimal choice. This is the core machinery you will use for the rest of the course.
A utility function assigns a number to each bundle so that preferred bundles get higher numbers. The marginal rate of substitution (MRS) is the slope of the indifference curve and equals the ratio of marginal utilities. At the optimal bundle, MRS = MRT (for Cobb-Douglas), the goods are consumed in fixed proportions (for perfect complements), or the consumer buys only the cheaper good (for perfect substitutes, unless the slopes match).
Utility function, U(X, Y)
A mathematical function that assigns a numerical value to each bundle of goods, representing the consumer's preferences. A ≻ B if and only if U(A) > U(B). Think of it as a scoring system for bundles.
Marginal utility of X (MUx)
The change in total utility from consuming one additional unit of Good X, holding Y constant. Formally, MUx = ΔU/ΔX.
Marginal utility of Y (MUy)
The change in total utility from consuming one additional unit of Good Y, holding X constant. Formally, MUy = ΔU/ΔY.
Marginal rate of substitution (MRS)
The rate at which a consumer is willing to trade one good for another while remaining equally satisfied. It is the slope of the indifference curve. Its value is -MUx/MUy. Think of it as your personal exchange rate, as opposed to the market's exchange rate (MRT).
Marginal rate of transformation (MRT)
The rate at which the market allows you to trade one good for another. It is the slope of the budget constraint, equal to -Px/Py.
Ordinal utility
Only the ranking of utility numbers matters (1st, 2nd, 3rd), not the size of the gaps between them. Consumer theory uses ordinal utility.
Cardinal utility
The actual numerical values of utility matter (e.g., U = 10 is "twice as good" as U = 5). Consumer theory does not require this.
Cobb-Douglas utility function
U(X, Y) = X^α · Y^β. Generates strictly convex, smooth indifference curves. The exponents α and β capture how much the consumer values each good relative to the other.
Perfect complements
U(X, Y) = MIN(αX, βY). Goods consumed in fixed proportions (like left and right shoes). Generates L-shaped indifference curves.
Perfect substitutes
U(X, Y) = αX + βY. Goods are interchangeable at a constant rate. Generates linear (straight-line) indifference curves.
Law of diminishing marginal utility
As you consume more and more of a good, at some point the extra utility from each additional unit begins to decrease.
Optimal bundle (X, Y)**
The bundle on the budget constraint that maximises utility. The consumer's best affordable choice.
A utility function U(X, Y) assigns a number to every bundle. The only thing that matters is the ranking: if U(A) > U(B), the consumer prefers A. The actual size of the numbers is irrelevant (ordinal, not cardinal). You can apply any increasing transformation to a utility function (e.g., doubling all values, taking the log) and it still represents the same preferences.
A utility function can be visualised as a 3D surface. Each indifference curve is a horizontal slice through that surface at a given utility level.
Along an indifference curve, utility is constant, so the total change in utility is zero:
ΔU = MUx·ΔX + MUy·ΔY = 0
Rearranging:
MUx·ΔX = -MUy·ΔY
MUx/MUy = -ΔY/ΔX
Therefore, the slope of the IC is:
ΔY/ΔX = -MUx/MUy = MRS
The MRS tells you how much Y you are willing to give up for one more unit of X. The MRT tells you how much you must give up.
Indifference curves: strictly convex, smooth, bowed inward. As you move along an IC and consume more X, Y becomes scarcer, and you become less willing to give it up. The slope flattens.
Marginal utilities:
MUx = α · X^(α-1) · Y^β
MUy = β · X^α · Y^(β-1)
MRS formula: MRS = -(α/β) · (Y/X)
This is the key result. Memorise it.
Finding the optimal bundle: two equations, two unknowns.
Equation 1 (MRS = MRT): -(α/β)·(Y/X) = -Px/Py, which simplifies to Y = (β/α)·(Px/Py)·X
Equation 2 (budget constraint): Px·X + Py·Y = M
Substitute Equation 1 into Equation 2 and solve.
General solution:
X* = (α/(α+β)) · M/Px, meaning the consumer spends the fraction α/(α+β) of income on Good X
Y* = (β/(α+β)) · M/Py, meaning the remaining fraction goes to Good Y
Worked example: U(X, Y) = X^(1/2) · Y^(1/2), M = $100, Px = $2, Py = $1.
Here α = β = 1/2, so α/(α+β) = 1/2.
X* = (1/2)·100/2 = 25
Y* = (1/2)·100/1 = 50
The consumer spends $50 on each good.
Another worked example: U(X, Y) = X^(1/4) · Y^(3/4).
α/(α+β) = (1/4)/1 = 1/4. The consumer spends 1/4 of income on X and 3/4 on Y.
X* = M/(4Px)
Y* = 3M/(4Py)
Indifference curves: L-shaped. Extra units of one good without the other provide no additional utility. The kink of the L is where αX = βY.
Finding the optimal bundle:
Equation 1 (kink condition): αX = βY, i.e., the consumer uses goods in the exact required proportion
Equation 2 (budget constraint): Px·X + Py·Y = M
Substitute and solve.
Worked example: U(X, Y) = MIN(3X, Y), M = $25, Px = $2, Py = $1.
At the kink: 3X = Y.
Substitute into BC: 2X + 1·(3X) = 25, so 5X = 25, so X* = 5.
Y* = 3·5 = 15.
Another worked example: U(X, Y) = MIN(4X, Y), given bundle X = 2, Y = 6.
U(2, 6) = MIN(8, 6) = 6. At the kink: 4X = Y. Setting MIN(4X, Y) = 6 gives two cases: either 4X = 6 (so X = 3/2 and Y ≥ 6) or Y = 6 (so X ≥ 3/2). The IC is L-shaped with the corner at (3/2, 6).
Indifference curves: straight lines with slope -α/β. The MRS is constant everywhere.
Marginal utilities: MUx = α, MUy = β. Both are constant.
Finding the optimal bundle: compare |MRS| with |MRT|.
If |MRS| > |MRT| (i.e., α/β > Px/Py): the consumer values X more highly relative to Y than the market does. Buy only X. X = M/Px, Y = 0** (corner solution on the X-axis).
If |MRS| < |MRT| (i.e., α/β < Px/Py): the consumer values Y more highly relative to X than the market does. Buy only Y. X = 0, Y = M/Py** (corner solution on the Y-axis).
If |MRS| = |MRT|: any bundle on the budget line is optimal.
Worked example: U(X, Y) = 4X + Y, M = $20, Px = $5, Py = $1.
MRS = -4/1 = -4. MRT = -5/1 = -5.
|MRS| = 4 < |MRT| = 5. The IC is flatter than the budget constraint.
The consumer is willing to give up 4Y per unit of X, but the market demands 5Y per unit of X. X is too expensive relative to what it is worth to this consumer.
Optimal bundle: X* = 0, Y* = 20.
For strictly convex ICs (Cobb-Douglas), MRS = MRT is both necessary and sufficient for an interior optimum.
The proof uses the contrapositive. Suppose MRS ≠ MRT. Then either |MRS| > |MRT| or |MRS| < |MRT|.
If |MRS| > |MRT|, then MUx/Px > MUy/Py. An extra pound spent on X yields more utility than an extra pound on Y. The consumer should buy more X and less Y. They are not at the optimum.
If |MRS| < |MRT|, then MUx/Px < MUy/Py. The consumer should buy more Y and less X.
In either case, MRS ≠ MRT means the consumer is not maximising utility. Therefore MRS = MRT is necessary.
The price-consumption curve traces out how the optimal bundle changes when the price of one good changes, holding the other price and income constant.
For U(X, Y) = X^(1/2)·Y^(1/2) with M = $100 and Py = $1, as Px changes from $1 to $2 to $5:
Px | X* | Y* |
|---|---|---|
$1 | 50 | 50 |
$2 | 25 | 50 |
$5 | 10 | 50 |
With this utility function (equal exponents), spending on Y does not change when Px changes. All the adjustment falls on X.
\text{MRS} = -\frac{MU_x}{MU_y} \quad \text{(slope of the indifference curve)}\text{MRT} = -\frac{P_x}{P_y} \quad \text{(slope of the budget constraint)}\text{Cobb-Douglas: } U(X,Y) = X^\alpha Y^\beta \implies \text{MRS} = -\frac{\alpha}{\beta} \cdot \frac{Y}{X}\text{Cobb-Douglas optimal: } X^* = \frac{\alpha}{\alpha+\beta} \cdot \frac{M}{P_x}, \quad Y^* = \frac{\beta}{\alpha+\beta} \cdot \frac{M}{P_y}\text{Perfect complements: } U = \min(\alpha X, \beta Y) \implies \text{kink at } \alpha X = \beta Y\text{Perfect substitutes: } U = \alpha X + \beta Y \implies \text{MRS} = -\frac{\alpha}{\beta} \text{ (constant)}\Delta U = MU_x \cdot \Delta X + MU_y \cdot \Delta Y = 0 \quad \text{(along an IC)}Cobb-Douglas captures the idea that people spread their spending across categories rather than going all-in on one. This is why households buy both food and entertainment rather than spending everything on food. Perfect complements describe goods that must be used together: a printer and ink cartridges, a phone plan and a phone. Perfect substitutes describe goods you would swap freely at some rate: generic paracetamol vs. branded paracetamol, or different bus routes to the same destination.
Students often confuse MRS and MRT. MRS is the slope of the IC (based on preferences, what you are willing to trade). MRT is the slope of the budget constraint (based on prices, what you must trade). At the optimum for Cobb-Douglas, they are equal, but they measure different things.
Students sometimes set MRS = MRT for perfect complements or perfect substitutes. For complements, use the kink condition (αX = βY) instead. For substitutes, compare the slopes and pick a corner solution.
Students forget that utility is ordinal. Saying U(A) = 10 and U(B) = 5 does not mean A is "twice as good" as B. It only means A is preferred to B.
Students mix up the Cobb-Douglas formula for X*. Remember: X* = (α/(α+β))·(M/Px). The fraction α/(α+β) tells you the share of income spent on X, and dividing by Px converts from spending to quantity.
For perfect substitutes, students sometimes assume there is always an interior solution. There is not. Unless the IC slope exactly equals the budget constraint slope, the optimal bundle is at a corner.
⚠️ The Cobb-Douglas MRS = -(α/β)·(Y/X) is one of the most-tested formulas in the course. Know how to derive it from the marginal utilities.
⚠️ For Cobb-Douglas, the share of income spent on each good is constant (α/(α+β) on X, β/(α+β) on Y). Exams test this property.
⚠️ For perfect complements, the exam will test whether you can identify the kink condition from the utility function and solve correctly.
⚠️ For perfect substitutes, you must compare |MRS| with |MRT| before solving. Do not blindly set MRS = MRT.
⚠️ The proof that MRS = MRT is necessary for an optimum (using the contrapositive) is a common exam question.
True or false: A utility function of U = 2X + 3Y represents perfect substitutes. (True)
Fill in the blank: For Cobb-Douglas U = X^α · Y^β, the MRS equals _______. (Answer: -(α/β)·(Y/X))
True or false: For perfect complements U = MIN(2X, 5Y), the kink of the IC occurs where 2X = 5Y. (True)
Fill in the blank: At the optimal bundle for Cobb-Douglas preferences, the consumer spends the fraction _______ of income on Good X. (Answer: α/(α+β))
True or false: If |MRS| > |MRT| for a perfect substitutes utility function, the consumer buys only Good Y. (False: they buy only Good X.)
Q: Given U(X, Y) = X^(1/3) · Y^(2/3), M = $90, Px = $3, Py = $6. Find the optimal bundle.
A: α = 1/3, β = 2/3, so α/(α+β) = 1/3. X* = (1/3)·90/3 = 10. Y* = (2/3)·90/6 = 10. The consumer buys 10 units of each good, spending $30 on X and $60 on Y.
Q: Given U(X, Y) = MIN(2X, 4Y), M = $60, Px = $4, Py = $2. Find the optimal bundle.
A: Kink condition: 2X = 4Y, so X = 2Y. Substitute into BC: 4(2Y) + 2Y = 60, so 10Y = 60, so Y* = 6 and X* = 12.
Q: Given U(X, Y) = 3X + 2Y, M = $30, Px = $5, Py = $3. Find the optimal bundle.
A: MRS = -3/2. MRT = -5/3. |MRS| = 1.5, |MRT| ≈ 1.67. Since |MRS| < |MRT|, the consumer values X less than the market charges for it. Corner solution: X* = 0, Y* = 30/3 = 10.
Q: Explain in words why the MRS equals the ratio of marginal utilities.
A: Along an indifference curve, total utility is constant. If you gain some X and lose some Y, the utility gained from extra X (MUx·ΔX) must exactly offset the utility lost from less Y (MUy·ΔY). Setting the total change to zero and rearranging gives MUx/MUy = -ΔY/ΔX, which is the slope of the IC.
Q: For a Cobb-Douglas consumer, what happens to the optimal quantity of Y when Px increases? Explain.
A: Nothing. The share of income spent on Y is β/(α+β), which does not depend on Px. So Y* = (β/(α+β))·M/Py is unchanged when Px changes. All adjustment falls on X.
The optimal bundle framework here is the starting point for income and substitution effects (Slutsky decomposition), which decompose the impact of a price change into two components. The demand curves you will derive later come directly from the optimal bundle solutions, X*(Px, Py, M) and Y*(Px, Py, M). The price-consumption curve introduced at the end of these notes leads directly into the derivation of the individual demand curve.
utility function, marginal utility, MUx, MUy, marginal rate of substitution, MRS, MRT, Cobb-Douglas, perfect complements, perfect substitutes, optimal bundle, utility maximisation, ordinal utility, cardinal utility, diminishing marginal utility, indifference curve slope, corner solution, interior solution, kink condition, price-consumption curve, ECON 302, intermediate microeconomics, Penn State