3 - Liconsa and Social Assistance for Milk - Lecture Notes

Case Snapshot

In 2003, Mexico debated how to help poor families. Liconsa, an agency under the Ministry of Social Development, runs a milk subsidy program: eligible families buy up to 4 liters per week per beneficiary at 3.5 pesos per liter, roughly half the retail price (about 7.2 pesos/liter urban, 6.8 pesos/liter rural). Eligible beneficiaries are children under 12, adults 60 and over, and pregnant women in families below the official poverty line.

The rival program is Oportunidades (formerly Progresa), a conditional cash transfer: cash to poor mothers, conditional on keeping children in school and attending health clinics.

The policy fight:

• The government had already phased out Liconsa's free-tortilla program (2002), arguing the money was better spent expanding cash grants.
• Critics call the milk program archaic and redundant — cash is more flexible.
• Defenders say milk and Oportunidades are complements, not substitutes: one is a nutrition program, the other is income support. And the milk program is self-financing, so why cut it.
• Families can enroll in Liconsa or Oportunidades, but not both.

The exam question type: compare an in-kind transfer (or price subsidy) with a cash transfer; analyze who benefits, the distortion involved, and when each is justified.

Background - Mexico's Anti-Poverty Programs

Mexico's poverty context in 2001: income per capita about USD 5,540, yet 16% of 100 million people lived on USD 1 per day or less. The sequence of programs:

• Conasupo (state trading firm, founded 1963; Liconsa began under it). Distributed subsidized beans, tortillas, milk. Had a history of corruption and was used by the ruling PRI party as an electoral weapon. Dismantled in 1999; Liconsa moved under the Ministry of Social Development and cleaned up.
• Solidaridad (President Salinas, 1988-1994): grants to poor neighborhoods to build community infrastructure, with communities choosing the projects ("empowerment"). Became politicized.
• Progresa (President Zedillo, 1994-2000): replaced Solidaridad. A conditional cash transfer targeting the root causes of poverty - low education and poor health.
• Oportunidades (President Fox, 2000 on): Progresa renamed and expanded.
• Liconsa milk program: the survivor under debate in the case.

Core Theory - In-Kind Transfers Versus Cash Transfers

This is the central theory of the case. Government help to the poor can take three forms:

• In-kind transfer: give a specific good (milk), or subsidize its price.
• Cash transfer: give money, no strings.
• Conditional cash transfer: give money conditional on a behaviour (school attendance, clinic visits).

The Consumer-Choice Framework

Model the family as choosing between milk and all other goods, subject to a budget constraint (income and prices) and trying to reach the highest indifference curve (highest utility).

"All other
 goods"
   |
   |\
   | \         <- original budget line
   |  \
   |   \  \
   |    \   \      cash transfer:  budget line shifts OUT, parallel
   |     \    \    price subsidy:  budget line PIVOTS out (milk cheaper,
   |      \  \  \                  so the line gets flatter)
   +------------------ Milk

• A cash transfer shifts the budget line outward, parallel - the family is richer at unchanged relative prices.
• A price subsidy on milk makes milk cheaper, so the budget line pivots/rotates outward - it gets flatter. Relative prices change.

The Cash-Dominance Result

The fundamental result: for the same cost to the government, a cash transfer leaves the recipient at least as well off as an in-kind transfer, and usually strictly better off.

Why: with cash the family can buy any bundle it could afford under the in-kind transfer, including the exact bundle the in-kind transfer pushes them to - plus many other bundles. More choice cannot make a chooser worse off. This is the consumer-sovereignty argument: the family knows its own needs best.

A price subsidy in particular always introduces a substitution effect (it distorts the milk-versus-other-goods choice), so an equal-cost cash transfer strictly dominates it for the recipient's own welfare.

Inframarginal Versus Marginal Recipients

There is an important exception that limits the distortion:

• An inframarginal recipient would buy more milk than the subsidized allotment even at the full price. For them, the subsidy on those allotment liters is just a pure income transfer - it acts exactly like cash, with no distortion, because they were going to buy that milk anyway.
• A marginal recipient is one whose milk choice is actually changed by the subsidy or by the 4-liter cap. For them the subsidy creates a genuine substitution effect and a distortion.

The 4-liter-per-week cap matters here: by capping the subsidy, much of the benefit falls on inframarginal units, which limits how distorting the program is.

Why Governments Still Use In-Kind Transfers

If cash is better for recipients, why give milk? An exam answer must give both sides. Reasons to prefer in-kind:

• Paternalism and merit goods: the government may value child nutrition more than the family does, or distrust how cash would be spent. An in-kind transfer steers consumption toward the good society wants consumed.
• Externalities: better-nourished children produce social benefits (healthier, more productive adults; lower future public health costs) that the family does not fully count.
• Self-targeting (screening): if the transferred good - or the hassle of obtaining it - is unattractive to the non-poor, the non-poor will not bother to claim it. The ordeal of queuing at a Diconsa shop at fixed hours acts as a screening device that keeps benefits flowing to the genuinely poor and reduces leakage.
• Political economy: "milk for poor children" is politically easy to fund and builds supportive constituencies (including dairy producers). Cash transfers are politically more fragile.
• Administrative control: the milk program used barcoded ID cards and elected beneficiary committees, and had far less fraud than the tortilla program (where shopkeepers billed for tortillas never handed out).
• Intra-household targeting: giving the benefit in a controlled form, or to mothers, can change who in the household controls the resource.

Price-Subsidy Analysis and Deadweight Loss

Income Effect and Substitution Effect

A price subsidy lowers the price of milk. Its effect on the family splits into two parts:

• Income effect: the subsidy 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 an income effect. A price subsidy produces both. The substitution effect is the distortion: it pushes the family to consume milk it values at less than milk's true (unsubsidized) cost.

Where the Deadweight Loss Comes From

The substitution-induced extra milk is the deadweight loss of the price subsidy: society spends the full resource cost to produce milk that the family values at less than that cost. The size of the loss grows with how strongly quantity responds to the price cut (the elasticity of milk demand).

The standard ranking, by recipient welfare per peso of government spending:

Cash transfer        >=   Quantity-based in-kind (voucher)   >=   Price subsidy
(pure income effect)      (cash-equivalent if inframarginal)      (always has a
                                                                   substitution
                                                                   distortion)

The catch: this ranking is by the recipient's own utility. If the government's goal is nutrition itself (the merit-good view), then the "distortion" - extra milk - is the intended outcome, not a loss.

Reading Exhibit 1 - The Evidence

The Data Table

The evaluation compared Liconsa families with similar non-Liconsa families, split by urban/rural and by all-poor versus poorest-half. Key figures from Exhibit 1:

What the Numbers Mean

• Liconsa families drink much more milk than comparable non-Liconsa families - the consumption difference is large and statistically significant in most cells.
• They often spend less on milk while drinking more - the row-3 differences are negative - because the subsidy lowers the price.
• The decisive row is the last one. "Savings available for other goods" = 52 times the weekly milk-spending difference. It is small for average families (153 urban, 445 rural) and even negative for the poorest rural families (-69) - they actually spent more on milk than before.
• Interpretation: almost all of the "savings" from the subsidy were poured back into buying more milk.

This is the crux of the debate:

• The defenders' reading: this "proves" the milk program is a genuine nutrition program that changes consumption, distinct from income support (Oportunidades). If it were just disguised income, families would have spread the savings across many goods. So the two programs are complements.
• The economist's caution: families pouring the windfall into milk is precisely the substitution effect of a price subsidy. Whether that is "good" depends on whether you accept the merit-good / externality rationale for milk. If child nutrition has real external benefits, the extra milk is the point; if not, it is deadweight loss and equal-cost cash would have served the families better.

Targeting performance, also exam-relevant:

• Exclusion error (under-coverage): the program reached only about one-third of those eligible. The Ministry of Finance capped enrollment at 5 million while eligible beneficiaries numbered roughly 14.5 million (12.6 million children under 12 plus 1.9 million adults over 60).
• Inclusion error (leakage): some enrolled families were not actually poor, because politicians had pressed Liconsa to enroll constituents (notably in the capital and Hidalgo).

Conditional Cash Transfers and the Progresa Evaluation

How Conditional Cash Transfers Work

Progresa/Oportunidades gives cash to poor mothers conditional on children attending school and the family visiting health clinics. Design features:

• Grants go to mothers, believed more likely to spend on children (intra-household targeting).
• Education grants rise with grade and are larger for girls, because dropout risk rises with grade and is higher for girls - the grant is calibrated to offset the opportunity cost of staying in school. Range: about 90 pesos/month (boy, grade 3) to 335 pesos/month (girl, grade 9).
• Rationale: it keeps the flexibility of cash but uses the conditions to correct a specific failure - households under-investing in children's human capital because of poverty, credit constraints, or short-sightedness.

The Randomized Evaluation

A key methodological point: the government randomly assigned communities to Progresa, creating a randomized controlled trial (RCT). Random assignment makes the treated and untreated groups comparable, so measured differences are causal, not just correlation.

Results of the evaluation (carried out by an international non-profit, IFPRI):

• School attendance rose enough to add about 0.7 years of schooling (from 6.2 to 6.9 years), implying roughly an 8% increase in lifetime earnings.
• Families getting health and nutrition grants had 12% fewer illnesses.
• No reduction in adult labor supply - the cash did not make adults work less. (Fewer children worked, because they were in school.)

The Hidden Cost - Trade Protection and the Quota Rent

The milk program is described as self-financing - it needs no budget support. Exam point: "self-financing" is not the same as "costless." Trace the hidden cost.

• Mexico protects its dairy industry by limiting imports of powdered milk. The world price of bulk milk powder is only about 45% of the Mexican price. Private importers must pay a 125% tariff.
• Liconsa has the right to import bulk powdered milk at the low world price (free of that tariff). It then manufactures and sells liquid milk.
• That right is an implicit subsidy: Liconsa captures the gap between the cheap world price and the protected domestic price - a quota rent.
• So the program is financed by the trade barrier. The true cost is borne by Mexican milk consumers (who pay protected high prices) and by forgone tariff revenue. It is an opportunity cost, just not a line in Liconsa's budget.

Illustrative scale (2002): Liconsa imported 104.5 million kilos of powder for 1,517 million pesos, produced 1,028 million liter-equivalents of milk, and sold about 95% at the subsidized price to beneficiaries and 5% at unsubsidized prices.

The Debate - Both Sides

Exam Toolkit - Solving a Transfer or Subsidy Case

1. Classify the instrument: cash, conditional cash, in-kind quantity (voucher), or in-kind price subsidy.
2. Draw the budget constraint and indifference curves. Show that, for equal government cost, cash weakly dominates in-kind for the recipient's own utility.
3. Decompose a price subsidy into income effect (also produced by cash) and substitution effect (the distortion / deadweight loss).
4. Check inframarginal versus marginal: is the quantity cap binding? Inframarginal benefit behaves like cash.
5. List the in-kind justifications: paternalism / merit goods, externalities, self-targeting, political economy, administrative control.
6. Assess targeting: inclusion error (leakage to non-poor) versus exclusion error (under-coverage).
7. Find the hidden cost: "self-financing" usually hides a quota rent, a tax expenditure, or an opportunity cost.
8. If given evaluation data: note whether assignment was random (causal evidence); separate income-driven from substitution-driven responses.

Formula Sheet

Subsidy per liter      = retail price - subsidized price
   Urban: 7.2 - 3.5 = 3.7 pesos       Rural: about 6.8 - 3.5 = 3.3 pesos

Annual savings to family = subsidy per liter x liters of subsidized milk consumed
   Example (urban, all poor): 3.70 x 460.20 liters  =  about 1,703 pesos

Savings available for other goods = 52 x (weekly milk-spending difference)
   Example (urban, all poor): 52 x 2.95  =  about 153 pesos per year

Government cost of a price subsidy = subsidy per unit x subsidized quantity

Recipient welfare ranking (equal government cost):
   cash  >=  in-kind voucher  >=  price subsidy

Total effect of a price subsidy = income effect + substitution effect
Total effect of a cash transfer = income effect only

Practice Questions

1. A family receives milk at a subsidized price. Using a budget constraint and indifference curves, explain why an equal-cost cash transfer would leave the family at least as well off. Then give two reasons a government might still prefer the milk subsidy.

2. The retail milk price is 7.2 pesos and the subsidized price is 3.5 pesos. A family buys 460 liters of subsidized milk per year. Compute the annual savings.

   Answer

   Subsidy 3.7 per liter times 460 = about 1,702 pesos.

3. Explain the difference between an inframarginal and a marginal recipient, and why the distinction determines whether a price subsidy is distorting.

4. Liconsa families spent their "savings" mostly on more milk. Give the defenders' interpretation and the economist's more cautious interpretation of this fact.

5. Why is the milk program called "self-financing," and why is that label misleading? Identify who really bears the cost.

6. Distinguish inclusion error from exclusion error, and give an example of each from the Liconsa program.

7. What does it add to an evaluation when communities are assigned to a program at random, as with Progresa?