Indifference curve analysis offers a stylised but powerful framework for thinking about the consequences of a price change. A rational consumer chooses the combination of Good X and Good Y where the budget line is tangent to the highest attainable indifference curve, so MRS_XY = Px/Py. A change in Px rotates the budget line, and the Hicksian decomposition splits the total effect into a substitution effect (switching toward or away from X along the original indifference curve) and an income effect (the move to a higher or lower curve as real income changes). The question is whether this framework earns its keep for policy makers.
For the claim that indifference curve analysis is useful. First, the decomposition answers a question policy makers actually ask: when a tax is introduced, how much of the fall in quantity demanded is a RELATIVE-PRICE response and how much is a REAL-INCOME response? For a sugar-tax or carbon-tax, the substitution effect indicates the targeted health or environmental benefit, while the income effect indicates the regressive burden on low-income households that spend a larger share on the taxed good. The compensating variation derived from indifference curves gives a money measure of the welfare loss, which can inform the size of a compensating cash transfer. Second, the model distinguishes NORMAL, INFERIOR and GIFFEN goods, warning policy makers that for some inferior staples (as in the Jensen-Miller 2008 Chinese rice evidence) a subsidy may REDUCE consumption — an important corrective to the 'cheaper = more consumed' intuition that would otherwise mislead food-security policy.
Further usefulness. The model also underpins welfare measures that are standard in cost-benefit analysis (compensating variation and equivalent variation), and it provides a visual, pedagogically tractable way of showing why distortionary taxes impose a deadweight loss — the consumer ends up on a lower indifference curve than they would under a lump-sum tax of equal revenue.
Against the claim. First, the model's behavioural assumptions are strong. Consumers are assumed rational, with stable, complete and transitive preferences, and full information. Behavioural economics evidence (loss aversion, framing, mental accounting, endowment effects) shows systematic departures. For many policy applications, especially 'nudge'-style interventions, these departures matter more than the price effect itself. Second, the model is restricted to two goods; real households allocate income across many goods, and cross-price elasticities often matter. Third, preferences and MRS cannot be observed directly, so the diagrams are illustrative rather than quantitative — in practice policy makers rely on econometrically estimated elasticities, not on tangency diagrams.
Further limitations. The standard model represents a REPRESENTATIVE consumer, but policy questions often turn on distribution across income deciles. The model is static, so it ignores habit formation, intertemporal substitution and uncertainty — all crucial for policies with long time horizons (e.g. carbon pricing). And the model is comparative-static: it compares two equilibria but says nothing about the ADJUSTMENT path, which may be long and costly.
Judgement. Indifference curve analysis is useful as a FRAMEWORK: it forces policy makers to ask the right questions — how large is the substitution response, how large is the income effect, who bears the welfare cost, and could the good be inferior enough to produce a counter-intuitive sign? For point predictions, however, it is a benchmark that must be supplemented with empirical elasticities, behavioural evidence and distributional analysis. The model is therefore a useful STARTING POINT but not a sufficient tool on its own. A policy maker who uses indifference curve analysis to structure the question and empirical evidence to answer it will arrive at better policy than one who relies on either alone.