Practice Midterm 2 - Module 3 Cases

How to Use This Practice Exam

This is a second pass over the same four cases as Practice Midterm 1, but the questions attack them from different angles - so doing both gives broad coverage. As the syllabus states, the real exam rewards reasoning, framing, evidence evaluation, and clear argument, not recall.

Designed for about 75 minutes. Total: 120 points.
Each question hides a model answer in a callout - in Obsidian, click the bar to reveal it. Always attempt your own answer first.
Strong answers name the concept, apply it to the case, and admit the trade-off.
The italic line under each question names the exam skill being trained.

Part	Case	Points
1	Pricing the EpiPen	25
2	The Elevator Saga	27
3	Liconsa and Social Assistance for Milk	26
4	The Effects of Rice Subsidies in Thailand	22

Part 1 - Pricing the EpiPen

Q1 (13 pts). Why does widespread health insurance tend to raise the equilibrium price of a drug like the EpiPen? Identify the underlying principal-agent or moral-hazard mechanism, and say who ultimately bears the cost.

Tests: explaining a market distortion and tracing who pays.

[!success]- Model answer In a normal market the consumer is also the payer, so a seller who charges too much loses sales - price discipline works. Health insurance breaks that link.

The patient consumes the drug, but the insurer pays the bill, and an employer or the government funds the insurer.

The decision-maker therefore does not feel the price, so demand becomes insensitive to price - a moral hazard in insurance and a principal-agent problem (the agent choosing the drug is not the party paying).

With demand made inelastic, sellers can raise list prices with little loss of volume.

Who pays: the cost is spread to all insured people through higher premiums, and - because over 40% of U.S. children are on Medicaid or CHIP - onto taxpayers. The uninsured, with no third party, feel the full list price directly. The high price does not vanish; it is shifted onto premium-payers and the public budget.

Q2 (12 pts). Martin Shkreli defended Mylan by noting it earned only an 8.9% net profit margin, so it was "not gouging." Critique this defense. Why is the figure misleading?

Tests: critically evaluating a quantitative claim.

[!success]- Model answer The defense is misleading because it uses the wrong unit of analysis.

8.9% is the company-wide net margin, blending Mylan's many low-margin generic products with the EpiPen.

The relevant figure is the product-level economics of the EpiPen: a price around USD 300 per pen against a marginal cost of about USD 1-2 - a product gross margin near 99%.

A company-wide average hides an extremely profitable single product. A firm can show a modest overall margin while one product is priced enormously above its cost.

Also, net margin reflects choices like heavy marketing spend (Mylan's selling and administrative expense far exceeds its R&D) and interest on debt - costs the firm chose, which do not show that the EpiPen price tracks the EpiPen's cost. The correct test is price versus marginal cost for that product, and on that test the markup is extreme.

Part 2 - The Elevator Saga

Q3 (13 pts). Using the second-price (Vickrey) auction analogy, explain why pledging your true value is a dominant strategy in the pivot mechanism. Why is "dominant strategy" a stronger property than "best response"?

Tests: explaining a mechanism-design result precisely.

[!success]- Model answer In a second-price auction the highest bidder wins but pays the second-highest bid. Because your bid decides whether you win but not how much you pay, shading your bid cannot lower your price - it can only make you lose something you wanted or win something you did not. So bidding your true value is optimal.

The pivot mechanism copies this split:

Your pledge decides whether the elevator is built, but you pay a fixed fair share, not your pledge - so your pledge cannot change your own bill.

If your pledge is pivotal, the penalty makes you pay exactly the cost you imposed on others.

Overstating risks tipping an inefficient build (you pay for something worth less to you); understating risks tipping an inefficient rejection (you pay a penalty). Pledging b = v avoids both.

Dominant strategy means truth-telling is best no matter what everyone else does - you need no belief about others' pledges. That is far stronger than a best response, which is only optimal given a specific guess about others. Dominance is what makes the mechanism robust.

Q4 (14 pts). A second-floor homeowner dislikes the elevator because of noise: his value is v = -EUR 500. His fair share is f = EUR 2,050, and everyone else pledges a total of B_other = EUR 72,300. Should the mechanism allow him to pledge a negative number? If he pledges honestly, what happens, and is honesty still his best move? Show your work.

Tests: handling an edge case and verifying incentive compatibility.

[!success]- Model answer Yes - negative pledges should be allowed. A homeowner who is harmed by the elevator has a genuine negative value, and the mechanism stays efficient only if it can hear "I want this not built."

Honest pledge: b = v = -EUR 500.

Total B = B_other + b = 72,300 - 500 = EUR 71,800. Since 71,800 is below 72,000, the elevator is not built.

Pivotal test: B - b + f = 71,800 - (-500) + 2,050 = EUR 74,350, which is above 72,000 while B is below it - opposite sides, so he is pivotal in killing the elevator.

Penalty: F_other = 72,000 - 2,050 = 69,950. Penalty = B_other - F_other = 72,300 - 69,950 = EUR 2,350.

Payoff = 0 (not built) - 2,350 = -EUR 2,350.

Is honesty still best? Check the alternatives:

If he pledged his fair share (EUR 2,050): B = 74,350, the elevator is built, he is not pivotal, he pays f = 2,050, payoff = v - f = -500 - 2,050 = -EUR 2,550.

If he pledged 0: B = 72,300, built, not pivotal, pays 2,050, payoff = -EUR 2,550.

Honesty (-2,350) beats both alternatives (-2,550). So even for someone who is harmed, truthful pledging remains the dominant strategy.

Q5 (13 pts). Decompose the effect of the milk price subsidy on a family into an income effect and a substitution effect. Why is the substitution effect called the "distortion"? Under what condition is that distortion small?

Tests: applying consumer theory to a real policy.

[!success]- Model answer A price subsidy lowers the price of milk. Its effect splits into two parts:

Income effect: the lower price makes the family effectively richer, so it buys more of all normal goods, including milk.

Substitution effect: milk is now cheaper relative to other goods, so the family substitutes toward milk.

A cash transfer produces only the income effect; a price subsidy produces both. The substitution effect is the distortion because it pushes the family to consume milk that it values at less than milk's true (unsubsidized) resource cost - a deadweight loss. An equal-cost cash transfer avoids it.

The distortion is small when the family is inframarginal - that is, when it would have bought at least the subsidized allotment even at the full price. For those units the subsidy is just a pure income transfer, with no behavior change. The 4-liter-per-week cap helps by keeping much of the benefit on inframarginal units.

Q6 (13 pts). Why are the Progresa (Oportunidades) cash grants made conditional on school attendance and clinic visits rather than unconditional? Why did the government's decision to assign communities to the program at random matter for evaluating it?

Tests: justifying program design and judging evidence quality.

[!success]- Model answer Why conditional: a pure cash transfer fixes current poverty but does not guarantee that families invest in their children's human capital. Poor families may under-invest in schooling and health because of credit constraints, short time horizons, or because the person controlling the cash is not the one who benefits. The conditions keep the flexibility of cash but steer behavior toward the specific under-provided investments - education and health. The grants are even scaled (larger for higher grades and for girls) to offset the opportunity cost of staying in school.

Why random assignment matters: by randomly assigning communities to Progresa, the government created a randomized controlled trial. Random assignment makes the treated and untreated groups comparable in every other respect, so the measured differences (0.7 extra years of schooling, about 8% higher lifetime earnings, 12% fewer illnesses, no fall in adult labor) can be read as caused by the program, not as mere correlation. It turns the evaluation into credible causal evidence.

Part 4 - The Effects of Rice Subsidies in Thailand

Q7 (11 pts). Why is a price floor unsustainable unless someone buys up the surplus? Why does the resulting over-production count as a deadweight loss?

Tests: explaining the mechanics and welfare effect of a price support.

[!success]- Model answer A binding price floor sets the price above equilibrium, so quantity supplied exceeds quantity demanded - there is a surplus. If nothing else happens, sellers competing to offload that surplus would bid the price back down to equilibrium and the floor would collapse. The floor can hold only if a buyer removes the surplus - in Thailand, the government became the buyer of last resort and stockpiled the excess rice.

Why over-production is a deadweight loss: the support price Ps pulls farmers to expand output until their marginal cost equals Ps. But the extra units are only worth the world price Pw to buyers. Society spends real resources (land, water, labor, fertilizer) producing rice worth less than it cost - that gap is pure waste. The loss on the extra output is roughly the triangle 1/2 x (Ps - Pw) x (extra quantity), on top of storage, spoilage, and resale losses.

Q8 (11 pts). Thailand's rice scheme lost enormous sums of public money, yet it was politically popular and hard to stop. Using the idea of concentrated benefits and diffuse costs, explain why a fiscally ruinous policy can still be politically durable.

Tests: applying public-choice reasoning to government failure.

[!success]- Model answer This is a classic government failure explained by public-choice theory.

The benefits are concentrated on farmers - a large, well-organized group, over-represented in the voting system because rural districts are weighted heavily, and backed by lobbyists. Each farmer gains a lot, so each has a strong incentive to defend the policy and to vote on it.

The costs are diffuse - spread thinly across millions of taxpayers. Each individual loses only a little, so no one has an incentive to organize against it; the cost is barely noticed.

This asymmetry means politicians face loud, organized pressure to keep the policy and only quiet, unorganized pressure to end it. Add vote-buying (the scheme was launched to win the 2011 election) and rent-seeking by the agricultural lobby, and an inefficient, money-losing policy becomes politically durable - until a fiscal or political crisis (here, a coup) forces the issue.

Part 5 - Synthesis

These questions are unweighted and are the full-length, syllabus-style items - practice writing complete, structured arguments.

Q9 (open, ~12 min). For any two of the four cases, state whether government intervention was the cure for a market failure or itself the cause of the problem, and explain.

Tests: classifying interventions and structured comparison.

[!success]- Model answer The four cases deliberately span the range:

EpiPen - market failure, intervention largely absent. The failure is monopoly power; government is mostly not acting, and the case argues it arguably should (faster generic approval, importation, negotiation). Here intervention would be a cure.

Rice - intervention is the cause. The price support created over-production, stockpiles, fiscal ruin, and environmental damage. A textbook government failure.

Elevator - intervention as well-designed cure. The free-rider problem would block a worthwhile public good; the pivot mechanism is a designed institution that solves it. A cure - when the design is good.

Liconsa - a debatable intervention. A price subsidy distorts choice, but it may correct a nutrition externality or paternalistic concern. Cure or distortion depends on whether you accept the merit-good rationale.

A strong answer picks two, classifies each, and notes that the deciding factor is institutional design, not "intervene or not."

Q10 (open, ~10 min). A classmate concludes: "All four Module 3 cases show that governments should stay out of markets." Do you agree? Construct a reasoned response.

Tests: argument construction and resisting an over-simple thesis.