AI summary by TwinMind
Lecture 01: Monopoly Pricing and Durable Goods, Part 1
MIT OpenCourseWare
1h 22m 48s
Course Overview and Structure
- Course 14271 co-taught by instructor (21 lectures), Tobias (4 lectures), Shan terroll (2 lectures)
- Primary goal: teach students writing IO thesis what they need to know
- Also useful for students in other fields due to IO's theory-empirical integration and structural empirical techniques
- Course covers 12 topics, typically 2 lectures each
- Pattern: Wednesday theory, Monday empirical work
- IO inherently theoretical field requiring theoretical understanding for policy decisions
- Examples: Facebook-Instagram divestiture, Google-Apple $12 billion default search deal
- Cannot use natural experiments for one-time decisions
- Most successful IO students doing empirical work to test theoretical assumptions and measure effects
Basic Monopoly Pricing Model
- Standard setup: monopolist sells single product
- Inverse demand curve P(x), total cost function C(x)
- Maximize π = x·P(x) - C(x)
- First order condition: MR = MC
- Rearranged as Lerner index: (P - MC)/P = -1/ε
- ε = elasticity of demand
- Examples: ε = -2 → 50% markup; ε = -4 → 25% markup
- IO pricing based on demand elasticities, not demand levels
- Standard welfare analysis: consumer surplus, deadweight loss, producer surplus
Problems with Monopoly Power Beyond Deadweight Loss
- Productive inefficiency
- Monopolists may not produce at lowest cost
- Lack competitive benchmarks
- Employee laziness without detection
- Quality distortions
- Example: Microsoft Windows quality complaints
- May provide suboptimal quality without competition pressure
- Rent seeking
- Example: cell phone license "beauty contests"
- Story: Washington DC license to Katherine Graham (Washington Post editor)
- Socially wasteful expenditures to obtain monopoly rents
- Bribes and lobbying costs can dissipate monopoly profits
- Equity concerns
- Transfer from consumers to firms
- Important if consumers have higher marginal utility of income than shareholders
- Note: IO typically uses quasi-linear preferences (constant marginal utility of income)
Multi-Product Monopoly Pricing
- Most real firms sell multiple goods
- Two-product example: demand X₁(P₁,P₂), X₂(P₁,P₂)
- Constant marginal costs: C₁, C₂
- First order conditions include diversion ratio effects
- Diversion ratio: when price of good 1 increases, how does demand for good 2 change
- Modified Lerner index includes cross-price effects
- Critical insight: must think of pricing as system, not individual products
- Bad example: studying blue iPhone price elasticity holding other iPhone prices fixed
- Would find high elasticity due to substitution, but wrong pricing conclusion
Substitutes vs Complements Cases
- Substitutes: raising price of one good increases demand for other
- Markup exceeds single-product formula
- Larger diversion ratio → larger markup
- Complements: raising price of one good decreases demand for other
- Can price below single-product formula
- Examples: hamburgers/fries, iPhone/earbuds/chargers, printers/ink
- Can lead to below-cost or negative pricing
- Printer example: $99 printer, $59 ink cartridge
- Historical newspaper example: 50¢ price, $1.50+ production cost, subsidized by advertising
- Google search: would like negative price, pays Apple billions for default placement
Product Quality Choice
- Quality s as choice variable along with price P
- Marginal cost C(s) increasing in quality
- Consumer heterogeneity: types θ ~ Uniform[0,1]
- Utility: V(s,θ) - P if buy, 0 if don't buy
- Assume quality and type ordered: higher θ types have higher willingness to pay
Monopolist's Quality Choice
- Sells to consumers with θ ≥ 1-Q (highest types)
- Marginal consumer indifferent: V(s,1-Q) = P
- First order condition for quality: ∂V/∂s|₁₋Q = ∂C/∂s
- Monopolist optimizes quality for marginal consumer
Social Planner's Quality Choice
- First order condition: ∫[1-Q to 1] ∂V/∂s dθ = Q·∂C/∂s
- Dividing by Q: average ∂V/∂s across consumers = ∂C/∂s
- Social planner optimizes for average consumer
Quality Distortion Analysis
- Two opposing effects:
- Monopolist focuses on marginal vs average consumer (may underprovide quality)
- Monopolist serves higher θ types than social planner would (may overprovide quality)
- Effects can go in opposite directions
- Corollary: homogeneous population model gives optimal quality theorem
- Marginal = average consumer
- All-or-none sales decision
- No scope for quality distortion
Durable Goods Monopoly
- Goods provide services over multiple periods
- Intuition: walk down demand curve over time, price discriminate by timing
- High-value consumers buy early at high prices, low-value consumers wait for price drops
Two-Period Model with Commitment Power
- Consumers buy at t=1 or t=2, then world ends
- Consumer utility: 2θ - P₁ if buy at t=1, θ - P₂ if buy at t=2
- Types θ distributed F(θ), marginal cost c
Optimal Policy with Commitment
- Counterintuitive result: optimal to sell at monopoly price in period 1, make no sales in period 2
- Constraint: θ₁ ≥ θ₂ (early buyers must have higher values)
- If constraint non-binding, would get θ₂ > θ₁ (contradiction)
- Reason: goods sold in period 1 have zero incremental cost (would produce for period 2 anyway)
- Example: Apple iPhone pricing - constant price until new model
- P₁ = θ₁ + θ₂, P₂ = θ₂ (early buyers pay premium for extra period of use)
Why Do We Observe Price Declines?
- Different preferences (status from being first)
- Changing costs over time (learning, capacity constraints, technology improvements)
- Commitment problems
Commitment Problems
- Without commitment, firm tempted to cut price in period 2
- Forward-looking consumers anticipate this, won't buy in period 1
- Numerical example: uniform demand, zero cost
- With commitment: P = 1/2 both periods
- Without commitment: P₁ = 9/10, P₂ = 3/10
- Lower profits, decreasing price sequence
Coase Conjecture
- Extreme commitment problem: continuous time, frequent price changes
- Consumer utility: θe^(-rt) if purchase at time t
- As Δt → 0, monopolist profits → 0
- Monopolist essentially forced to marginal cost pricing
- Reasoning: any finite price decline creates first-order gain from accelerating sales, second-order loss from price reduction
- In limit, price must fall at infinite rate initially
Why Coase Conjecture Doesn't Hold
- Firms don't change prices continuously (menu costs, consumer adjustment time)
- Reputation effects
- Inflow of new high-value customers
- Non-rational/non-forward-looking consumers
- Strategic responses: renting vs selling, most-favored-customer clauses, capacity constraints
Information Design in Monopoly Pricing
- Recent active area in IO theory
- Observation: some goods provide lots of information (laptops), others very little (mattresses, used cars)
- Wholesale used car auctions: minimal information provided intentionally
- Platform examples: Yelp, Amazon providing information about third-party products
Basic Framework
- New product (Clover fried cauliflower sandwich)
- Consumer values V ~ Uniform[0,18], marginal cost = $6
- Three scenarios: full information, no information, partial information
Full Information Case
- Standard monopoly pricing
- Price = $12, 1/3 of consumers buy, profit = $2 per unit sold
No Information Case
- Consumers know distribution but not own value
- Expected value = $9 for all consumers
- Price = $9, everyone buys, profit = $3 per unit sold
- Higher profits by hiding product information
- Creates inefficiency (low-value consumers buy) but eliminates monopoly deadweight loss
- Transfers more surplus from consumers to monopolist
Partial Information Case (Platform Design)
- Yelp provides partial information: "low value" (0-$6), "medium value" ($6-$14), "high value" ($14-$18)
- Creates point masses in demand: 1/3 at $3, 4/9 at $10, 2/9 at $16
- Monopolist prices at $10 (better than $16 due to demand structure)
- Results in lower price than full information case ($10 vs $12)
- Both consumer surplus and profits higher than monopoly case
- Demonstrates how information design can benefit both parties
