Demand Theory, Price Response and Elasticity

Introduction

Demand forecasting and pricing strategy begin with a disciplined understanding of how buyers respond to prices and to the conditions surrounding a purchase. A demand forecast is not merely an extrapolation of past sales. It is an estimate of how many units customers will purchase under specified prices, income conditions, competitor actions, product attributes, expectations, and market institutions. Pricing decisions therefore require both economic theory and empirical judgment.

This chapter develops the foundations needed to interpret demand curves, utility, substitution, income and cross-price effects, elasticity, reference prices, and behavioral responses. It also addresses an important managerial danger: observing that sales changed after a price change does not establish that price caused the change. Competitor promotions, seasonality, stockouts, advertising, product redesign, or changes in customer composition may be responsible.

Learning objectives

By the end of this chapter, readers should be able to:

1. Explain the demand function and distinguish movement along a demand curve from a shift of the curve.
2. Use utility theory to interpret consumer choice, marginal utility, and the marginal rate of substitution.
3. Decompose price effects into substitution and income effects.
4. Calculate and interpret own-price, income, and cross-price elasticity.
5. Explain reference prices and behavioral departures from the standard rational-choice model.
6. Distinguish correlation from causal price effects and identify suitable evidence for pricing decisions.
7. Translate demand analysis into managerial pricing, segmentation, experimentation, and governance decisions.

1. Demand as a conditional relationship

Demand is the quantity of a product that consumers are willing and able to purchase at a particular price and under particular market conditions. The phrase “willing and able” matters: desire without purchasing power is not effective demand, and purchasing power without willingness does not create a transaction.

A general demand function can be written as:

Here, Q_d is quantity demanded per period; P is the product’s own price; Y is consumer income or purchasing power; P_s is the price of a substitute; P_c is the price of a complement; A represents advertising and marketing communication; T represents tastes and preferences; E represents expectations; and Z represents other relevant factors such as demographics, availability, regulation, weather, and channel access. The function is conditional: a stated quantity has meaning only with respect to a defined market, time period, product, customer population, and information set.

A simple linear demand specification is:

where a is the intercept and b is the absolute change in quantity demanded for a one-unit increase in price. For a downward-sloping demand curve, b > 0. If price is measured in dollars per unit and quantity in units per week, then b has units of units per week per dollar. The linear form is useful for illustration, but it is not a universal law. Many markets are better represented by constant-elasticity, log-linear, piecewise, or nonlinear models.

The law of demand states that, other things equal, quantity demanded generally decreases as own price increases. “Other things equal” is the crucial qualification. A price increase may coincide with a stronger advertising campaign, a shortage, or a change in product quality. The observed relationship may then differ from the underlying demand relationship.

Movement versus shift

A movement along a demand curve occurs when the product’s own price changes while other determinants remain fixed. A shift occurs when a non-price determinant changes. For example, a reduction in the price of a compatible accessory may shift demand for the focal product outward, while a change in the focal product’s own price produces movement along the existing curve.

Figure: A price change produces movement along a demand curve; a change in another determinant shifts the curve.

A manager should ask two separate questions: “What quantity is associated with this price on the current curve?” and “Has the curve itself moved?” Confusing these questions leads to poor forecasts and incorrect attribution of sales changes.

2. Consumer choice and utility theory

Utility theory represents preferences over bundles of goods. A consumer chooses a bundle that maximizes utility subject to a budget constraint. Utility is not necessarily a measurable psychological quantity; in modern microeconomics, it is commonly an ordinal representation of preference. If bundle A is preferred to bundle B, the utility representation assigns a higher value to A, but the numerical difference need not have psychological meaning.

For two goods, a utility function may be written:

where x and y are quantities of two goods. The budget constraint is:

where P_x and P_y are prices and M is income available for the two goods. The consumer’s feasible set consists of all bundles satisfying the inequality. If preferences are locally nonsatiated, the consumer normally spends the available budget, so the constraint binds:

At an interior optimum, the marginal rate of substitution equals the relative price:

MU_x and MU_y are marginal utilities, defined as the additional utility associated with a small increase in each good while holding the other constant. The marginal rate of substitution is the amount of y the consumer is willing to give up for an additional unit of x while maintaining the same utility. The condition applies to an interior solution with differentiable preferences and a tangency between an indifference curve and the budget line. Corner solutions require a different analysis.

The consumer’s demand for x is obtained by solving the utility-maximization problem subject to the budget constraint. Changes in P_x alter both the slope of the budget line and the purchasing power of income. This produces the substitution and income effects.

Substitution and income effects

When the price of a product changes, two mechanisms operate:

• Substitution effect: the product becomes relatively cheaper or more expensive compared with alternatives, encouraging substitution toward or away from it.
• Income effect: the price change alters the consumer’s real purchasing power. A price reduction effectively increases purchasing power; a price increase reduces it.

For a normal good, the substitution and income effects of a price reduction both increase quantity demanded. For an inferior good, the income effect works in the opposite direction, although the substitution effect normally remains positive for a price reduction. In the exceptional Giffen case, the adverse income effect is sufficiently strong to outweigh the substitution effect, so quantity demanded rises with price. Giffen behavior is theoretically possible but empirically uncommon and should not be assumed merely because a product is inexpensive.

3. Elasticity: measuring responsiveness

A slope measures an absolute change, whereas elasticity measures a proportional response. Elasticity is especially useful for comparing products with different units or price levels.

Own-price elasticity

Point own-price elasticity is:

Q is quantity demanded and P is own price. The partial derivative holds other demand determinants constant. For a conventional demand curve, ε_p is negative. Managers often discuss its absolute value, |ε_p|. Demand is elastic when |ε_p| > 1, unit elastic when |ε_p| = 1, and inelastic when |ε_p| < 1.

For finite changes between two observations, midpoint or arc elasticity reduces dependence on which observation is treated as the base:

This formula is appropriate for describing responsiveness over an interval, not as an exact local derivative. It assumes the observations are sufficiently comparable and that the measured changes are not confounded by major simultaneous changes in other determinants.

For a linear demand curve Q = a − bP, point elasticity is:

Thus, elasticity varies along a linear curve even though its slope is constant. Demand tends to be more elastic at high prices and low quantities and less elastic at low prices and high quantities.

Revenue implications

Total revenue is:

If demand is elastic, a price reduction tends to increase total revenue because the percentage increase in quantity exceeds the percentage decrease in price. If demand is inelastic, a price reduction tends to decrease total revenue. At unit elasticity, a small price change leaves revenue approximately unchanged. These are local comparative-static implications, not guarantees of profit maximization. Profit also depends on variable cost, capacity, customer acquisition cost, contribution margin, and strategic effects.

#### Worked example: linear demand and elasticity

Suppose weekly demand is Q = 1,000 − 20P, where price is measured in dollars per unit and quantity is units per week. At P = 30, quantity is:

Q = 1,000 − 20(30) = 400 units.

Point elasticity is:

ε_p = −20(30/400) = −1.5.

Demand is elastic at this point. Revenue is TR = 30 × 400 = $12,000 per week. If price is reduced slightly, the percentage increase in quantity is expected to exceed the percentage decrease in price, subject to the model’s assumptions. The result does not by itself establish that the lower price maximizes profit.

Income and cross-price elasticities

Income elasticity is:

A positive value is consistent with a normal good; a negative value is consistent with an inferior good. Goods with high positive income elasticity are often more sensitive to economic cycles.

Cross-price elasticity between product i and product j is:

A positive cross-price elasticity suggests substitution: an increase in the price of j raises demand for i. A negative value suggests complementarity: an increase in the price of j reduces demand for i. Near-zero values suggest weak measured association, though statistical insignificance may also reflect poor measurement or insufficient variation.

The signs are not mechanical proofs of economic relationships. Product reclassification, promotion targeting, distribution changes, and common shocks can distort estimated elasticities.

4. What determines price elasticity?

Several structural factors influence responsiveness:

1. Availability of substitutes. More close substitutes generally make demand more elastic.
2. Share of budget. Products consuming a large share of income tend to receive greater price attention.
3. Time horizon. Consumers may be less responsive immediately but more responsive after they have time to switch brands, change habits, or replace durable goods.
4. Necessity and habit. Essential or habitual purchases may be relatively inelastic, although no product is universally inelastic.
5. Market definition. Demand for a narrowly defined brand is often more elastic than demand for a broad product category.
6. Search and switching costs. Information costs, contractual restrictions, inconvenience, and learning can reduce short-run responsiveness.
7. Capacity and inventory. A firm may observe muted sales responses when stockouts constrain quantity, even if underlying demand is highly responsive.
8. Customer segment. Elasticity can differ by income, use occasion, geography, loyalty, and channel.

Elasticity is therefore not a permanent attribute of a product. It is a property of a product-market-time-segment context.

5. Reference prices and behavioral responses

Standard demand theory assumes that consumers evaluate options consistently according to preferences and constraints. Behavioral research shows that judgments often depend on context, memory, framing, and perceived fairness.

A reference price is a benchmark against which a current price is evaluated. It may be internal, based on remembered prices or expectations, or external, such as a strikethrough price, competitor price, or “regular price.” A simple conceptual model is:

This is not a universal behavioral law and should not be treated as a calibrated demand equation. It indicates why an identical price can be perceived differently by different customers or at different times.

Important behavioral mechanisms include:

• Anchoring: an initial price influences subsequent evaluation.
• Reference dependence: gains and losses are evaluated relative to a reference point rather than only in absolute terms.
• Loss aversion: a price increase may feel more significant than an equivalent price decrease feels beneficial.
• Price-quality inference: consumers may interpret a higher price as a signal of quality when quality is difficult to assess.
• Fairness perceptions: unexplained or discriminatory increases may provoke dissatisfaction, switching, or negative word of mouth.
• Partitioned pricing: separating base price, delivery, fees, or add-ons can affect attention and perceived value.
• Decoy effects and assortment context: an inferior option may change preferences among other options.
• Mental accounting: customers may treat budgets for transport, entertainment, or subscriptions as separate accounts.

Behavioral responses can modify measured elasticity. A temporary promotion may generate stockpiling rather than sustained demand. A highly visible price increase may cause a stronger response than an equivalent increase hidden in fees. A discount can also train customers to delay purchases, reducing future willingness to pay.

Ethically, reference-price communication must be truthful and substantiated. Artificially inflated “was” prices, hidden mandatory fees, and misleading scarcity claims may violate consumer-protection rules and undermine trust even where they produce short-term conversion.

6. Correlation is not causal price response

Observed price and quantity are often jointly determined. A firm may lower price when it expects weak demand, or raise price when demand is unusually strong. In such cases, a simple correlation can be positive even though the underlying demand curve slopes downward.

A causal price effect asks: what would have happened to quantity for the same customers and circumstances if price had been different? This is a counterfactual question. Reliable estimation requires variation in price that is plausibly independent of unobserved demand shocks, or a credible design that controls for those shocks.

Common threats include:

• Confounding: advertising, distribution, product quality, or seasonality changes with price.
• Reverse causality: expected demand influences the price set by the firm.
• Selection: discounts are offered disproportionately to price-sensitive customers.
• Simultaneity: competitors change prices at the same time.
• Censoring and stockouts: observed sales reflect inventory limits rather than demand.
• Measurement error: transaction prices, list prices, coupons, and fees are recorded inconsistently.
• Interference: one customer’s treatment affects another through word of mouth or inventory depletion.

Useful approaches include randomized price experiments, geo-experiments, staggered tests, difference-in-differences designs, instrumental variables, and carefully specified panel-data models. Each depends on assumptions that must be examined rather than asserted.

A basic difference-in-differences estimator can be expressed as:

where Y is an outcome such as units per customer, T denotes the treated group, C the comparison group, and “pre” and “post” denote periods before and after the intervention. The key identifying assumption is that, absent treatment, treated and comparison groups would have followed parallel trends. This assumption should be assessed using pre-treatment data and substantive knowledge.

Randomized experiments provide stronger identification when implementation is controlled. However, experiments may have limited duration, contamination between groups, novelty effects, or weak external validity. Observational models can cover wider settings but require stronger assumptions about omitted variables and selection.

7. Managerial use of demand and elasticity

Managers should translate elasticity into decisions rather than treating it as a single answer. A practical workflow is:

1. Define the product, customer segment, channel, geographic market, and time horizon.
2. Specify the decision variable: list price, net transaction price, subscription fee, bundle price, or promotion depth.
3. Map substitutes, complements, costs, constraints, and likely behavioral reactions.
4. Estimate baseline demand and uncertainty using appropriate data.
5. Test whether price variation is causal or confounded.
6. Estimate heterogeneous effects across segments and contexts.
7. Simulate revenue, contribution margin, retention, acquisition, and long-run strategic effects.
8. Implement guardrails, monitor outcomes, and update the model.

A profit function may be represented as:

where π is profit per period, c is unit variable cost, Q(P) is quantity sold at price P, and F is fixed cost. The formula assumes a single price, constant unit variable cost, and no additional strategic effects. Real pricing often requires multiple segments, capacity constraints, inventory costs, acquisition costs, returns, taxes, and dynamic customer lifetime value.

For a differentiable single-product problem with unconstrained capacity and constant marginal cost, an interior profit maximum satisfies:

Rearranging gives the Lerner condition:

where ε_p is own-price elasticity at the optimum and is negative. This condition applies only under restrictive assumptions: a firm with market power, a well-defined demand curve, constant marginal cost, no capacity constraint, and an interior optimum. It is not a license to raise prices whenever estimated elasticity is low. Regulation, fairness, competition, customer lifetime value, and reputational effects remain relevant.

Applied case: illustrative omnichannel coffee subscription

Background

Consider BrewPath, a fictional coffee company selling beans through supermarkets, its website, and a monthly subscription. The company increased the online subscription price from $24 to $27 while simultaneously expanding digital advertising and introducing improved packaging. Online units declined by 3 percent. Executives concluded that customers were highly price sensitive and proposed reversing the increase.

Problem

The observed decline cannot be interpreted as a causal price elasticity without considering the counterfactual. The advertising campaign may have offset part of the decline; packaging may have increased willingness to pay; competitors may have launched promotions; and customers may have shifted from subscriptions to supermarket purchases.

Analysis

The analytics team first defines the relevant outcome as active subscribers, units per active subscriber, contribution margin, and churn over a three-month horizon. It separates list price from net price by including discounts, shipping, and taxes consistently. It also examines stock availability and customer segment.

The team finds that the price increase was applied to existing subscribers in some regions but not others because of a phased billing-system rollout. Regions not yet exposed provide a comparison group. Pre-period trends are similar. The team estimates a difference-in-differences model and conducts a placebo test using an earlier period when no price change occurred.

Results suggest that existing subscribers reduced quantity modestly, but churn increased substantially among low-frequency buyers. High-frequency subscribers showed little response and valued the packaging improvement. Some customers migrated to supermarket purchases, so online unit decline overstated total brand demand loss.

Approach and solution

BrewPath adopts segment-specific retention offers rather than a universal price reversal. It maintains the new price for high-frequency subscribers, offers a temporary transition credit to vulnerable low-frequency subscribers, and redesigns communication to explain the packaging and service improvements. It also tests a two-tier subscription with different delivery flexibility. The company reports the reference price and total mandatory charge clearly and avoids presenting an unsubstantiated “regular price.”

Outcome

After two billing cycles, contribution margin improves because the firm retains the higher price among relatively inelastic customers while reducing avoidable churn in the sensitive segment. The company continues testing because the initial estimate applies only to the observed horizon and customer groups.

Learning points

The case illustrates five principles. First, a sales decline after a price increase is not automatically a price elasticity estimate. Second, segmentation can reveal heterogeneous responses hidden in an average. Third, channel substitution must be included in the market definition. Fourth, revenue effects and profit effects differ. Fifth, ethical price communication and transparent experimentation support both compliance and long-run trust.

Challenges, limitations, and contemporary developments

Demand models face structural uncertainty. Preferences may change after product innovation, economic shocks, regulation, or competitor entry. Historical elasticities may not transfer to a new price range. Machine-learning models can improve prediction but may still fail to identify causal effects. A highly accurate forecast of what happened under old prices is not necessarily a reliable estimate of what will happen under a new price.

AI-enabled pricing systems can integrate browsing behavior, inventory, competitor information, and experimentation. However, governance is essential. Firms should document objectives, data sources, model versions, features, constraints, and monitoring procedures. They should test for disparate effects across protected or vulnerable groups, prevent use of sensitive personal data without a lawful basis, and preserve human accountability for consequential decisions. Dynamic pricing can be efficient, but frequent individualized changes may create perceptions of unfairness or enable exploitative pricing during emergencies.

Best practice is to combine economic structure, statistical identification, behavioral insight, and managerial judgment. Models should report uncertainty intervals, not only point estimates. Experiments should have pre-specified success metrics and stopping rules where feasible. Managers should distinguish short-run conversion from long-run retention, brand equity, and customer trust.

Chapter Summary

Demand is a conditional relationship between quantity and its determinants. Own price usually produces movement along a demand curve, whereas income, preferences, advertising, competitor prices, expectations, and availability can shift the curve. Utility theory explains consumer choice through preferences and a budget constraint; at an interior optimum, the marginal rate of substitution equals the relative price.

Price changes generate substitution and income effects. Own-price elasticity measures proportional quantity response to proportional price change, while income and cross-price elasticities describe responses to purchasing power and other product prices. Elasticity varies by market definition, time horizon, segment, substitutes, switching costs, and purchasing context. Revenue implications do not automatically determine profit implications.

Reference prices, anchoring, fairness, loss aversion, and mental accounting help explain behavioral responses that standard models may not capture. Most importantly, observed correlation between price and sales is not necessarily causal. Experiments and credible quasi-experimental designs are needed to estimate counterfactual price effects. Effective pricing combines theory, causal evidence, segmentation, financial analysis, ethical governance, and continuous learning.

Key Terms

Demand function: A relationship between quantity demanded and its economic and market determinants.

Quantity demanded: The amount consumers are willing and able to buy at a specified price and context.

Demand curve: A graphical or mathematical representation of quantity demanded as price varies, holding other determinants constant.

Movement along a demand curve: A quantity change caused by a change in the product’s own price.

Demand shift: A change in the entire demand relationship caused by a non-price determinant.

Utility: An ordinal representation of consumer preferences over bundles.

Budget constraint: The combinations of goods affordable given prices and available income.

Marginal utility: The change in utility associated with an additional unit of a good, holding other goods constant.

Marginal rate of substitution: The quantity of one good a consumer will give up for another while maintaining utility.

Substitution effect: The response caused by a change in relative prices.

Income effect: The response caused by a change in real purchasing power.

Own-price elasticity: Proportional quantity response to a proportional change in own price.

Income elasticity: Proportional quantity response to a proportional change in income.

Cross-price elasticity: Proportional response in one product’s demand to a proportional change in another product’s price.

Reference price: A benchmark used to evaluate a current price.

Price anchoring: The influence of an initial or salient price on later evaluation.

Causal effect: The change in an outcome attributable to an intervention relative to a counterfactual.

Confounding: A situation in which another factor affects both the apparent treatment and the outcome.

Difference-in-differences: A quasi-experimental estimator comparing changes over time between treated and comparison groups.

Lerner condition: A markup relationship connecting price, marginal cost, and own-price elasticity under restrictive profit-maximization assumptions.

Review Questions

Short Answer Questions

1. Distinguish a movement along a demand curve from a shift in demand.
2. What do the substitution effect and income effect describe?
3. Define own-price elasticity and explain the meaning of an elastic demand response.
4. Why can a reference price affect willingness to pay?
5. Why does correlation between price and sales not establish a causal price effect?

Long Answer Questions

1. Explain how utility maximization, the budget constraint, and the marginal rate of substitution describe consumer choice. State the assumptions behind the interior tangency condition.
2. Compare own-price, income, and cross-price elasticity, including their formulas, signs, managerial uses, and limitations.
3. Discuss the factors that make demand more or less price elastic over short and long horizons.
4. Evaluate how anchoring, fairness, loss aversion, and mental accounting complicate conventional demand theory.
5. Explain why the Lerner condition cannot be applied mechanically to every pricing decision.

Analytical / Application Questions

1. A firm estimates demand as Q = 1,200 − 30P. Calculate quantity and point own-price elasticity at P = 20. Interpret the result for a small price reduction, and identify two reasons why the conclusion may not predict profit.
2. A retailer raises price while also increasing advertising and experiencing a competitor stockout. Design a credible analysis to estimate the causal effect of the retailer’s price change.
3. A subscription business observes that high-use customers remain after a price increase while low-use customers churn. Explain how segmentation, reference prices, channel substitution, and customer lifetime value should influence the pricing response.

Multiple Choice Questions with Answers

Question 1

Which change normally causes movement along an existing demand curve rather than a shift?

A. A change in the product’s own price

B. A change in consumer income

C. A change in the price of a substitute

D. A change in consumer preferences

Correct Answer: A. A change in the product’s own price

Question 2

At an interior consumer optimum, which condition generally holds under differentiable preferences?

A. Marginal utility of every good is zero

B. The marginal rate of substitution equals the relative price

C. All goods have identical prices

D. Income is greater than expenditure

Correct Answer: B. The marginal rate of substitution equals the relative price

Question 3

If own-price elasticity is −1.5, demand is best described as:

A. Perfectly inelastic

B. Inelastic

C. Unit elastic

D. Elastic

Correct Answer: D. Elastic

Question 4

A positive cross-price elasticity between products A and B generally suggests that the products are:

A. Complements

B. Substitutes

C. Inferior goods

D. Unrelated by definition

Correct Answer: B. Substitutes

Question 5

For the demand function Q = 1,000 − 20P, what is quantity demanded when P = 30?

A. 300 units

B. 400 units

C. 600 units

D. 1,600 units

Correct Answer: B. 400 units

Question 6

Which statement about a Giffen good is correct?

A. Its demand necessarily has zero elasticity

B. Its quantity demanded can rise when its price rises because a strong adverse income effect outweighs substitution

C. It is always a luxury good

D. It has a positive cross-price elasticity with every product

Correct Answer: B. Its quantity demanded can rise when its price rises because a strong adverse income effect outweighs substitution

Question 7

Which issue most directly threatens a causal interpretation of a sales decline after a price increase?

A. The product has a brand name

B. The firm changed advertising at the same time

C. The demand curve is displayed graphically

D. The price is stated in currency units

Correct Answer: B. The firm changed advertising at the same time

Question 8

In a difference-in-differences design, the parallel-trends assumption means that:

A. Both groups must have identical sales levels at all times

B. The groups must receive identical prices

C. Absent treatment, the groups would have followed similar outcome trends

D. The treatment group must be larger than the comparison group

Correct Answer: C. Absent treatment, the groups would have followed similar outcome trends

Question 9

If demand is inelastic at a particular price, a small price reduction will generally:

A. Increase total revenue

B. Decrease total revenue

C. Leave quantity unchanged

D. Make marginal cost equal to zero

Correct Answer: B. Decrease total revenue

Question 10

Which is an appropriate governance practice for an AI-enabled pricing system?

A. Hide the model’s objectives from decision makers

B. Use sensitive personal data without legal or ethical review

C. Monitor disparate effects and document data, assumptions, and model versions

D. Assume predictive accuracy proves causal validity

Correct Answer: C. Monitor disparate effects and document data, assumptions, and model versions