Indifference Curves

How indifference curves work

Imagine you're on a desert island stocked with coconuts and bananas. Some days you eat 8 coconuts and 2 bananas; other days, 2 coconuts and 8 bananas; a third day, 4 coconuts and 4 bananas. Suppose you find all three combinations equally satisfying — you're indifferent between them. The set of all such bundles forms a curve in the (coconuts, bananas) plane. That's an indifference curve.

Stack many such curves and you get an indifference map: a contour map of utility, like elevation lines on a topographical map. Higher curves represent strictly higher utility — every bundle on a higher curve is strictly preferred to every bundle on a lower one. The map encodes everything needed to predict consumer choice.

Four properties define a well-behaved indifference map:

1. Monotonic. More is better — bundles with strictly more of every good lie on higher curves.
2. Convex. Curves bow toward the origin. Mixing two indifferent bundles is at least as good as either alone.
3. Negatively sloped. Giving up some of one good must be compensated by more of the other to stay equally happy.
4. Non-crossing. Two indifference curves can never cross; if they did, transitivity of preferences would fail.

The marginal rate of substitution

The slope of an indifference curve, in absolute value, is the marginal rate of substitution (MRS). At any bundle, MRS answers: how much of good Y am I willing to give up for one more unit of good X?

MRS_xy = − dY/dX  (along an indifference curve)
        = MU_x / MU_y

The second equality follows from the chain rule: along a constant-utility curve, MU_x · dX + MU_y · dY = 0, so dY/dX = −MU_x/MU_y. The MRS is the ratio of marginal utilities at the point.

Convexity translates directly into diminishing MRS: as the consumer moves rightward along the curve (more X, less Y), each extra X is worth less in terms of Y. Walk along the curve from coconut-rich to banana-rich and the consumer trades fewer and fewer coconuts for each banana — until at the right end, even one more banana barely tempts them.

The consumer optimum: tangency with the budget line

The consumer doesn't get to pick any bundle; they pick the one that gives the highest utility subject to a budget. Plot the budget line on the same axes:

Y ↑
   |\
   | \
   |  \  ← budget line: Px·X + Py·Y = M
   |   \
   |    \    ← higher indifference curves
   |     \   _________
   |      \ /         \
   |       *           ← optimum: tangency
   |      / \_________/
   |     /
   |    /  ← lower indifference curve
   +────────→ X

The consumer wants to reach the highest curve that still touches the budget line. Inside the budget set the curves stack vertically; outside, they're unaffordable. The unique solution (for a smooth, strictly convex indifference map) is a tangency: the highest curve that just kisses the budget line without crossing it.

At the tangency, the slope of the indifference curve equals the slope of the budget line:

MRS  =  Px / Py     (tangency condition)

Equivalently:
   MU_x / MU_y  =  Px / Py
   MU_x / Px    =  MU_y / Py     (equimarginal principle)

The last form is the bedrock of all consumer theory: in equilibrium, the marginal utility per dollar is equalized across goods. If MU_x/Px exceeded MU_y/Py, the consumer would be wasting money by holding any Y — every dollar shifted to X would buy more utility.

Worked example: Cobb-Douglas utility

Suppose Aleksei has utility U(X, Y) = X^0.4 · Y^0.6, prices Px = $4, Py = $6, and income M = $120. Find his optimal bundle.

Marginal utilities:

MU_x = 0.4 · X^(−0.6) · Y^(0.6)
MU_y = 0.6 · X^(0.4)  · Y^(−0.4)

MRS = MU_x / MU_y = (0.4 / 0.6) · (Y / X) = (2/3)(Y/X)

Set MRS equal to the price ratio:

(2/3)(Y/X)  =  4/6  =  2/3
Y / X        =  1
Y            =  X

Substitute into the budget constraint Px·X + Py·Y = M:

4·X + 6·X = 120
10·X       = 120
X*         = 12,    Y* = 12

Aleksei buys 12 of each. Utility at the optimum: U = 12^0.4 · 12^0.6 = 12. The Cobb-Douglas form has a famous shortcut: with exponents α and β summing to 1, expenditure shares equal the exponents. Here α = 0.4 means Aleksei spends 40% of $120 = $48 on X (12 units at $4) and 60% = $72 on Y (12 units at $6). The math checks.

Special shapes: substitutes and complements

Preference type	Shape	MRS	Example
Standard convex	Smoothly bowed	Diminishing	Coconuts and bananas
Perfect substitutes	Linear (parallel lines)	Constant	$5 bill vs five $1 bills
Perfect complements	L-shaped	Undefined at the kink	Left shoe + right shoe
Quasilinear	Vertical translates	Depends only on X	Money-and-burgers utility
Cobb-Douglas	Bowed; never touches axes	Proportional to Y/X	Constant expenditure shares
Concave (anti-convex)	Bowed away from origin	Increasing	Bingeing on one good

Concave preferences violate the convexity axiom and produce corner solutions: the consumer spends everything on whichever good has the lowest price-per-utility ratio. Most real preferences are convex, but corner solutions reappear in finance (specialization in one asset) and labor economics (full-time vs no-work labor supply).

Hicks vs Slutsky decomposition

When a price changes, the consumer slides along the existing indifference curve (the substitution effect) and jumps to a new curve because real income changed (the income effect). Two ways to decompose:

	Hicks decomposition	Slutsky decomposition
Compensation rule	Income adjusted to keep utility constant	Income adjusted to keep the original bundle just affordable
Geometric meaning	Same indifference curve	New budget line through old bundle
Theoretical purity	Cleaner — preserves welfare	Easier to compute from observed prices and quantities
Empirical use	Welfare analysis (CV, EV)	Empirical demand systems
Substitution effect always non-positive?	Yes (convexity)	Yes (revealed preference)

The Hicks substitution effect is the rotation of the price line tangent to the original indifference curve; the new tangency point is the Hicksian-substitution bundle. The Slutsky version uses a different rotation — the line through the original bundle at the new prices — and lands on a slightly different intermediate point. For small price changes, the two coincide; for large ones, they differ but both produce non-positive substitution effects.

The income effect can be positive (normal goods) or negative (inferior goods); when it's strongly negative, it can flip the sign of the total response, producing a Giffen good — a curiosity but a real one in studies of staple grains in extreme poverty.

Common pitfalls

• Confusing MRS with the price ratio off-equilibrium. MRS = Px/Py only at the optimum. Away from it, the two are different and tell you which way to reallocate.
• Treating utility numbers as cardinal. Utility on indifference-curve diagrams is ordinal — only rankings matter. Doubling utility numbers is meaningless; a curve labelled U=10 and one labelled U=20 just say one is preferred to the other, not "twice as good".
• Drawing crossing curves. If two indifference curves cross, transitivity is violated: bundle A is preferred to B, B to C, but C to A. Curves can never cross.
• Forgetting that convexity is an assumption. Most textbooks assume strictly convex preferences, but it's not a logical requirement. Concave preferences (bingeing) and quasilinear preferences (no income effect on the non-money good) are perfectly coherent and produce different demand behavior.
• Mixing the Marshallian and Hicksian demands. Marshallian (uncompensated) demand holds income constant; Hicksian (compensated) holds utility constant. They give different elasticities and different welfare measures. Mixing them silently leads to errors in cost-benefit calculations.
• Reading the slope as marginal utility. The slope of an indifference curve is the ratio of marginal utilities, not either marginal utility alone. MU_x and MU_y are individually unobservable; only the ratio shows up on the diagram.