Consumer behaviour and utility

Economics (BA) · Microeconomics · free study notes for exams, placements and interviews

In short

Consumer behaviour explains how a household with a limited income chooses between goods, and utility is the measure of satisfaction economists use to model that choice. The cardinal approach (Marshall, Gossen) measures utility in units and gives the laws of diminishing and equi-marginal utility; the ordinal approach (Hicks–Allen) only ranks bundles and works with indifference curves and a budget line. From this one framework come the demand curve, the classification of normal, inferior and Giffen goods, consumer surplus and the welfare measures used to judge taxes and subsidies. It is core scoring matter in BA/MA microeconomics papers, UGC-NET Economics, ISI/DSE/IIT-M entrance tests and RBI/NABARD/IES exams, and it is applied daily by pricing analysts, FMCG brand teams, market-research firms (Kantar, Nielsen, Ipsos) and MoSPI's consumption-expenditure statistics.

Explained

Utility, wants and the cardinal approach

Utility: the want-satisfying power of a good — subjective, not the same as usefulness (cigarettes have utility, no usefulness) and not moral.
Cardinal utility: utility is measurable in notional units ("utils") and is additive across goods: TU = U₁(x₁) + U₂(x₂) + … Assumes utility of money is constant and goods are independent.
Total utility (TU): satisfaction from all units consumed. Marginal utility (MU): the addition to TU from one more unit, MU = ΔTU/ΔQ, or dTU/dQ.
Relation: MU falls as Q rises → TU rises at a decreasing rate; TU is maximum where MU = 0 (saturation); beyond it MU is negative and TU falls (disutility).
Law of diminishing marginal utility (DMU): as a consumer takes more of a good in a given period, MU of each successive unit falls. Reasons — wants are satiable, goods are imperfect substitutes for one another in different uses.
Assumptions of DMU: rationality, continuous consumption, homogeneous units, unchanged tastes/income/prices, "normal" consumer. Exceptions claimed: money, hobbies/collections, intoxicants at first doses, rare goods — mostly explained away as changed tastes or non-homogeneous units.
Law of equi-marginal utility (Gossen's second law / proportionality rule): a consumer maximises satisfaction by allocating income so that MU per rupee is equal across all goods: MUx/Px = MUy/Py = … = MU of money. If MUx/Px > MUy/Py, shift spending to X until equality.
Limitations of cardinalism: utility is not measurable, utility of money is not constant (it rises as income is spent), goods are not independent (substitutes/complements), it cannot handle indivisible goods well.

Ordinal approach: indifference curves and preferences

Ordinal utility: the consumer can *rank* bundles (A preferred to B, or indifferent) but cannot say by how much. Needs far weaker assumptions than cardinalism.
Axioms of preference: (i) completeness, (ii) transitivity (A≻B, B≻C ⇒ A≻C), (iii) non-satiation/monotonicity (more is better), (iv) continuity, (v) convexity (averages preferred to extremes).
Indifference curve (IC): locus of bundles giving the same satisfaction. Indifference map: a set of such curves, higher curve = higher satisfaction.
Properties of ICs: (1) negatively sloped; (2) convex to the origin (diminishing MRS); (3) higher IC means more utility; (4) two ICs never intersect (would violate transitivity); (5) an IC never touches either axis in the standard case; (6) they need not be parallel.
Marginal rate of substitution: MRSxy = units of Y given up for one more X with utility constant = −dy/dx = MUx/MUy. It diminishes along a convex IC because the more X one has, the less Y one will surrender for it.
Special shapes: perfect substitutes → straight-line ICs, constant MRS (U = ax + by); perfect complements → L-shaped ICs, no substitution (U = min(x/a, y/b)); one good a "bad" → upward-sloping ICs; neutral good → vertical/horizontal ICs; concave ICs → corner solution (spend all on one good).

Budget line and consumer equilibrium

Budget (price) line: Px·x + Py·y = M. Intercepts M/Px and M/Py; slope = −Px/Py (the market rate of exchange). The triangle below it is the budget set (what is affordable).
Shifts: income ↑ → parallel outward shift; Px ↓ → line pivots outward about the *Y*-intercept (the X-intercept moves out from M/Px to M/Px′); both prices and income changing in the same proportion → no change in the budget line (money illusion absent).
Equilibrium (interior solution): highest attainable IC is *tangent* to the budget line. Two conditions —
1. Necessary (first-order): MRSxy = Px/Py, i.e. MUx/Px = MUy/Py (the same proportionality rule as cardinalism, derived without measuring utility);
2. Sufficient (second-order): IC convex to the origin at the point of tangency.
Derivation outline (Lagrange, as examined): maximise U(x,y) subject to M = Px x + Py y. Form L = U(x,y) + λ(M − Px x − Py y); set ∂L/∂x = Ux − λPx = 0 and ∂L/∂y = Uy − λPy = 0 ⇒ Ux/Uy = Px/Py, with λ = marginal utility of income.
Corner solution: if MRS > Px/Py at every affordable bundle, the consumer buys only X (common with perfect substitutes or a strongly preferred good).

Price and income changes: PCC, ICC, Engel curve and the demand curve

Price consumption curve (PCC): locus of equilibria as one price changes, income and the other price fixed. Reading quantity of X against its own price off the PCC gives the individual demand curve — this is how ordinal theory derives demand without cardinal utility.
Income consumption curve (ICC): locus of equilibria as money income changes with prices fixed. Upward-sloping for normal goods; bends back towards the axis for an inferior good.
Engel curve: quantity (or expenditure) on a good against income. Engel's law: as income rises, the *share* of income spent on food falls (though the amount may rise).
Classification by income elasticity: normal (ηM > 0) — necessities 0 < ηM < 1, luxuries ηM > 1; inferior (ηM < 0, e.g. coarse cereals, bidis in some data).
Giffen good: a strongly inferior good whose negative income effect outweighs the substitution effect, so demand *rises* when its price rises — an exception to the law of demand. Classic Indian illustration: a staple coarse grain for very poor households (empirically contested).
Veblen/snob goods: demand depends on the high price itself (status), so the demand curve can slope upward; conspicuous consumption, not a utility-theory exception in the Giffen sense.

Decomposing the price effect: substitution and income effects

A fall in Px does two things: it makes X relatively cheaper (substitution effect, always towards the cheaper good, i.e. always negative price-wise) and it raises real income (income effect, positive for a normal good, negative for an inferior good).
Price effect = substitution effect + income effect. Normal good: both reinforce, demand rises. Inferior non-Giffen: income effect opposes but is smaller. Giffen: income effect dominates and reverses the sign.
Hicksian decomposition (compensating variation): adjust income so the consumer returns to the *original indifference curve* at new prices; movement along the same IC is the substitution effect.
Slutsky decomposition (cost-difference): adjust income so the consumer can just buy the *original bundle* at new prices; the substitution effect is thus slightly larger than Hicks's and is observable from market data.
Slutsky equation: ∂x/∂Px = (∂x/∂Px)|U − x·(∂x/∂M) — total effect = substitution + income, with the income term scaled by how much X is bought. It also yields the Marshall–Hicks law of demand: demand must fall when price rises unless the good is inferior and accounts for a large budget share.
Hicksian (compensated) vs Marshallian (uncompensated) demand: Hicksian holds utility constant and is always downward-sloping; Marshallian holds money income constant and may not be.

Consumer surplus, welfare measures and revealed preference

Consumer surplus (CS): what a consumer would have been willing to pay minus what is actually paid = area under the demand curve above price. Marshall measured it in money assuming constant marginal utility of money — the weak point Hicks replaced with compensating variation (income to be taken away after a price fall to restore old utility) and equivalent variation (income the consumer would accept instead of the price change). CV ≤ Marshallian CS change ≤ EV ordering depends on whether the good is normal.
Uses: measuring the welfare loss of an indirect tax, the gain from a subsidy or a tariff cut, and (with producer surplus) deadweight loss — the standard tool for appraising GST slab changes or a price cap.
Cash versus kind: indifference-curve analysis shows an unconditional cash transfer is never worse than an in-kind transfer of equal cost and is strictly better when the in-kind quantity exceeds what the household would freely choose — the analytical case behind the PDS-versus-DBT debate in India.
Revealed preference theory (Samuelson, 1938): behaviour reveals preference — if the consumer chooses A when B was affordable, A is revealed preferred to B. WARP (weak axiom): B is then never chosen when A is affordable. It derives demand (including the substitution effect, via the Slutsky compensation) with no utility function and no indifference curves at all; SARP adds transitivity so preferences can be reconstructed.
Utility under risk (briefly): expected utility EU = Σ pᵢ U(xᵢ); concave U means risk aversion (diminishing MU of wealth), which explains insurance purchase and the St Petersburg paradox. Departures from this model belong to behavioural economics.
Other applications of the apparatus: labour–leisure choice (wage = price of leisure, backward-bending supply), intertemporal choice (interest rate = price of present consumption), index numbers and the cost-of-living comparison behind CPI.

Indian data, codes and practice

MoSPI's Household Consumption Expenditure Survey (HCES) replaced the older quinquennial NSSO consumer-expenditure rounds; HCES 2022-23 and 2023-24 report monthly per-capita consumption expenditure (MPCE) and item-wise budget shares. Food's share of MPCE is now roughly 46-47% rural and about 39-40% urban — the clearest Indian evidence for Engel's law. (Figures are revised with each round; quote the latest MoSPI release.)
CPI weights are built from HCES budget shares; the series is being rebased using the newer HCES rounds, so weights quoted from the 2011-12 base are dated.
GST has 5%, 18% and 40% slabs since 22 September 2025 (plus nil-rated items and the special rates for bullion), and the slab assigned to a good changes its relative price and hence the equilibrium bundle — merit goods are kept at nil/5% precisely to exploit the substitution effect, "sin" and luxury goods at 40%.
Consumer Protection Act, 2019 and the Central Consumer Protection Authority (CCPA) cover misleading advertisements, unfair trade practices and e-commerce rules — the legal counterpart to "consumer sovereignty"; the Legal Metrology (Packaged Commodities) Rules, 2011 mandate MRP and unit-price declaration, which is what makes MU-per-rupee comparison possible for a real shopper.
Poverty lines (Tendulkar, Rangarajan) were anchored in consumption baskets from these surveys; NITI Aayog now also tracks the multidimensional poverty index.

Key points

Utility – want-satisfying power of a good; subjective, ethically neutral, and only *ranked* in modern theory.
TU and MU – MU = ΔTU/ΔQ; TU is maximum where MU = 0; MU negative ⇒ TU falling.
Law of DMU – successive units of the same good in a given period yield falling MU; the basis of a downward-sloping demand curve in cardinal theory.
Equi-marginal principle – maximise utility where MUx/Px = MUy/Py = MU of money; the universal allocation rule (also used for budgets and inputs).
Ordinal utility – needs only completeness, transitivity, monotonicity, continuity and convexity.
Indifference curve – equal-satisfaction locus; downward-sloping, convex, non-intersecting, higher = better.
MRS – MRSxy = MUx/MUy = −dy/dx; diminishes along a convex IC.
Budget line – Px x + Py y = M, slope −Px/Py; income shifts it parallel, a price change pivots it.
Equilibrium – tangency: MRS = Px/Py plus convexity; λ in the Lagrangian is the marginal utility of income.
Perfect substitutes/complements – straight-line ICs (corner solution) and L-shaped ICs (fixed proportions) respectively.
PCC → demand curve; ICC → Engel curve – the ordinal route to both relationships.
Price effect = substitution + income effect; the substitution effect is always opposite in sign to the price change.
Hicks vs Slutsky – compensation to the original *utility* vs to the original *bundle*; Slutsky's version is observable and slightly larger.
Giffen good – inferior good with a dominating negative income effect; demand rises with price.
Consumer surplus – area under demand above price; refined by Hicks into compensating and equivalent variation.
Cash ≥ kind – equal-cost cash transfer weakly dominates in-kind transfer; the DBT-versus-PDS argument.
Revealed preference – WARP/SARP derive demand from observed choices without any utility measurement.
Engel's law – food's budget share falls as income rises; visible in HCES rural/urban MPCE shares.

Formulas and examples

MU = ΔTU/ΔQ = dTU/dQ ; TU = Σ MU. Units: utils (notional); Q in physical units; prices in ₹ per unit.
Cardinal equilibrium: MUx/Px = MUy/Py = … = MUm, subject to Px·x + Py·y = M. (MUm = marginal utility of one rupee.)
MRSxy = −Δy/Δx |U constant = MUx/MUy (units of Y per unit of X).
Budget line: Px·x + Py·y = M ; slope = −Px/Py ; intercepts M/Px (X-axis), M/Py (Y-axis).
Ordinal equilibrium: MUx/MUy = Px/Py ; Lagrangian L = U(x,y) + λ(M − Px x − Py y), λ = ∂U/∂M.
Cobb-Douglas U = x^a·y^b ⇒ x* = aM/[(a+b)Px], y* = bM/[(a+b)Py] (budget shares a/(a+b), b/(a+b)).
Perfect substitutes U = ax + by ⇒ buy only X if a/Px > b/Py, only Y if a/Px < b/Py.
Slutsky equation: ∂x/∂Px = (∂x/∂Px)|U − x·(∂x/∂M) [total = substitution + income].
Income elasticity ηM = (ΔQ/Q)/(ΔM/M); normal if > 0, luxury if > 1, inferior if < 0.
Consumer surplus (linear demand P = a − bQ): CS = ½ × (a − P) × Q = ½ b Q².

Example 1 — Equi-marginal utility (cardinal). Income ₹60; Px = ₹10, Py = ₹5. MU schedules:

Unit	MUx (utils)	MUx/Px	MUy (utils)	MUy/Py
1	100	10	50	10
2	90	9	45	9
3	80	8	40	8
4	70	7	30	6
5	50	5	20	4

Buy in order of MU per rupee: both goods give 10, then 9, then 8. Take X₁,Y₁,X₂,Y₂,X₃,Y₃.
Cost = 3×₹10 + 3×₹5 = ₹30 + ₹15 = ₹45; remaining ₹15. Next best ratios: X₄ gives 7 per rupee, Y₄ gives 6 → buy X₄ (₹10), leaving ₹5 → buy Y₄ (₹5).
Final bundle x = 4, y = 4, spending ₹40 + ₹20 = ₹60. Check: MUx/Px = 7, MUy/Py = 6 — the closest attainable equality with indivisible units.
TU = (100+90+80+70) + (50+45+40+30) = 340 + 165 = 505 utils.

Example 2 — Cobb-Douglas equilibrium. U = x^0.5·y^0.5, M = ₹900, Px = ₹30, Py = ₹15.

Budget shares are equal (a = b = 0.5): spend ₹450 on each.
x* = 450/30 = 15, y* = 450/15 = 30. U = √(15×30) = √450 ≈ 21.21.
Verification: MUx = 0.5(y/x)^0.5, MUy = 0.5(x/y)^0.5 ⇒ MRS = y/x = 30/15 = 2 = Px/Py = 30/15 = 2. ✔

Example 3 — Slutsky and Hicks decomposition. U = x·y, M = ₹1,200, Py = ₹10; Px falls from ₹20 to ₹10.

Demands: x = M/2Px, y = M/2Py. Before: x = 1200/40 = 30, y = 1200/20 = 60. After: x = 1200/20 = 60. Total price effect = +30 units.
*Slutsky:* income needed to buy the old bundle at new prices = 10×30 + 10×60 = ₹900. At M = 900, Px = 10 ⇒ x = 900/20 = 45. Substitution effect = 45 − 30 = +15; income effect = 60 − 45 = +15.
*Hicks:* restore the old utility U = 30×60 = 1,800. With U = xy, utility at the optimum = M²/(4PxPy) ⇒ M² = 1800×4×10×10 = 720,000 ⇒ M ≈ ₹848.53. Then x = 848.53/20 ≈ 42.43. Substitution effect ≈ +12.43; income effect ≈ +17.57. (Slutsky's substitution effect exceeds Hicks's, as theory predicts.)

Example 4 — Consumer surplus and a tax. Demand P = 100 − 2Q (P in ₹, Q in units).

At P = ₹40: Q = (100 − 40)/2 = 30; CS = ½ × (100 − 40) × 30 = ₹900.
A tax raises the price to ₹52: Q = (100 − 52)/2 = 24; CS = ½ × 48 × 24 = ₹576.
Loss of consumer surplus = ₹324, of which tax revenue = ₹12 × 24 = ₹288 and deadweight loss = ½ × 12 × 6 = ₹36.

Example 5 — Engel's law from budget shares. A rural household's MPCE rises from ₹3,000 to ₹4,500 (+50%); food spending rises from ₹1,500 to ₹1,950 (+30%).

Food share falls from 1500/3000 = 50% to 1950/4500 = 43.3% — Engel's law holds even though food spending rose.
Income elasticity of food = 30/50 = 0.6 ⇒ a necessity (0 < ηM < 1). Non-food spending rose from ₹1,500 to ₹2,550 (+70%), ηM = 70/50 = 1.4 ⇒ luxury on average.

Basis	Cardinal (Marshall)	Ordinal (Hicks–Allen)	Revealed preference (Samuelson)
Utility	Measured in utils, additive	Only ranked	Not used at all
Tool	MU schedules, equi-marginal rule	IC map + budget line	Observed choices, WARP
Equilibrium	MUx/Px = MUy/Py	MRS = Px/Py	Chosen bundle revealed best
Price effect	Not split	Split into substitution + income	Substitution effect observable
Main weakness	Utility not measurable; MU of money assumed constant	ICs not directly observable; two-good diagrams	Says nothing about why preferences are what they are

Where it is used

FMCG and retail pricing: pack-size and sachet strategy (HUL, ITC, Dabur), price-pack architecture, bundling and multi-buy offers all rest on diminishing MU, budget constraints and consumer surplus extraction.
Market research and analytics: conjoint analysis and discrete-choice models used by Kantar, Nielsen, Ipsos and in-house teams estimate utility weights for product attributes; roles — consumer insights analyst, pricing analyst, revenue-management analyst.
Telecom, OTT and fintech: Jio/Airtel data-plus-voice bundles, OTT tiering and credit-card reward design are versioning and price-discrimination problems built on willingness to pay.
Government and policy: MoSPI's HCES and the CPI basket, the PDS-versus-DBT cash-transfer debate, GST slab design (5/18/40 since 22 September 2025) and electricity/LPG subsidy design.
Banking and insurance: risk aversion from concave utility underpins insurance pricing, product suitability and portfolio-risk profiling at RBI/SEBI-regulated firms.
Teaching and research roles: UGC-NET/JRF, economics faculty posts, research associate roles at NCAER, NIPFP, IEG and bank research departments — all examine this theory directly.

Interview questions

Q: What is the difference between cardinal and ordinal utility? A: Cardinal utility assumes satisfaction is measurable in units and additive; ordinal utility assumes only that bundles can be ranked. Ordinal theory needs weaker assumptions and still yields the same equilibrium condition.
Q: State the law of equi-marginal utility and why it matters. A: A consumer maximises utility when marginal utility per rupee is equal across all goods. It matters because the same logic applies to any allocation of a scarce budget — government spending, input use, time.
Q: Why are indifference curves convex to the origin? A: Because the marginal rate of substitution diminishes — the more of a good you already have, the less of the other you will give up for another unit.
Q: Why can two indifference curves never intersect? A: The intersection point would be indifferent to two bundles on different utility levels, violating transitivity of preferences.
Q: Decompose the effect of a price fall. A: Into a substitution effect (the good is now relatively cheaper, so consumption rises) and an income effect (real income has risen) — positive for a normal good, negative for an inferior one.
Q: How do the Hicks and Slutsky methods differ? A: Hicks compensates income to keep the consumer on the original indifference curve; Slutsky compensates just enough to buy the original bundle. Slutsky's measure is observable from market data and gives a slightly larger substitution effect.
Q: What is a Giffen good and is it the same as an inferior good? A: A Giffen good is an inferior good whose negative income effect outweighs the substitution effect, so demand rises with price. All Giffen goods are inferior, but most inferior goods are not Giffen.
Q: Why does economic theory prefer cash transfers to transfers in kind? A: A cash transfer of equal cost lets the household reach any bundle the in-kind transfer allows plus others, so utility is at least as high — strictly higher when the in-kind quantity exceeds free choice.
Q: What does the Lagrange multiplier mean in the utility-maximisation problem? A: It is the marginal utility of income — the extra utility from one more rupee of budget at the optimum.
Q: Why did Hicks replace Marshallian consumer surplus with compensating variation? A: Marshall's measure assumes a constant marginal utility of money; compensating and equivalent variation measure welfare change in money without that assumption.
Q: What does revealed preference add? A: It derives demand theory from observed choices alone, with no utility measurement and no indifference curves, making the theory testable (WARP/SARP).
Q: If a consumer spends her whole income on two goods and one price doubles while her income doubles, is she better off? A: Weakly better off, yes — with the other price unchanged, doubling income more than offsets one price doubling: the new budget set contains the whole old one, so the old bundle is still affordable, and she is strictly better off if she was buying any of the other good. (Had both prices *and* income doubled, the budget line would be unchanged and she would be exactly as well off as before.)

Common mistakes

Treating TU as falling as soon as MU falls — TU keeps rising while MU is positive, and peaks only at MU = 0.
Writing equilibrium as MUx = MUy instead of MUx/Px = MUy/Py; the prices must be in the denominator.
Confusing the slope of the indifference curve (MRS, the consumer's valuation) with the slope of the budget line (price ratio, the market's rate) — equilibrium is where they are equal.
Calling every inferior good a Giffen good, or claiming a Giffen good disproves the substitution effect (it does not; the income effect merely dominates).
Drawing intersecting or concave indifference curves and then still asserting an interior tangency solution.
Mixing up Hicks and Slutsky compensation, or assuming the substitution effect can be positive for a price rise — it never is.
Interpreting Engel's law as "the poor spend more on food in rupees" instead of "the food *share* falls as income rises", and quoting old NSSO rounds rather than the latest HCES figures.