AGEC 640 – Agricultural Policy Thursday, September 18 th , 2014 Nutrition and Food Markets

1. AGEC 640 – Agricultural Policy Thursday, September 18th, 2014Nutrition and Food Markets Today: Imperfect information & food demand Reading: Masters and Sanogo, 2002 in AJAE Homework #2 on food choices due Thr., 9/25

2. What’s behind consumers’ price and income response? Quantity of food consumed Price of food “demand curve” “Engel curve” (=income-consumption curve) P1 price elasticity of demand: %∆Q/ %∆P Q2 income elasticity of demand: %∆Q/ %∆Y Q1 P2 Q1 Q2 Y1 Y2 Quantity of food consumed Consumers’ income …we need to think very carefully about what generates these curves!

3. To understand food demand,we’ll want to consider… • consumers’ optimization (“Econ 101” effects) • preferences: indifference curves and welfare • price effects: demand curves and elasticity • income effects: Engel curves and elasticity • really constrained optimization (Econ 102, 103…) • what else might be useful to understand food intake? • benefits are delayed, and often not observable • credit/insurance constraints (poor can’t borrow to buy food) • “behavioral” effects (predictable violations of rationality) • weak self-discipline (addiction, obesity, etc.) • distorted perceptions (anxiety, obsession, etc.) • information asymmetries (role of 3rd party quality assurance)

4. Optimization and consumer preferences The points in this quadrant offer more of both goods, so any optimizing consumer would prefer them to “O” Quantity of “a”, all other goods Initial observed point “O” Qa The points in this quadrant offer less of both goods, so any optimizing consumer would prefer “O” to them All combinations amongst which the consumer is indifferent must fall along a downward sloping line. Qb Quantity of “b” goods

5. Optimization and substitution possibilities The “indifference curve” Eventually, one becomes less willing to reduce “b” in exchange for more of all other things, so the indifference curve becomes steeper here Quantity of “a”, all other goods There is an indifference curve, drawn smooth for simplicity. Qa Eventually, one becomes less willing to reduce all other things in exchange for more of “b”, so the indifference curve becomes flatter here Qb Quantity of “b” goods

6. Constrained optimization: Indifference curves and the “expenditure line” Indifference curve through initial point Quantity of “a”, all other goods higher indifference levels lower indiff. levels Qa Exp. = PaQa + PbQb Expenditure level at the initial point Qa = Exp./Qa – (Pb/Pa)Qb Slope of expenditure line = -Pb/Pa Qb Qb Quantity of “b” goods

7. Constrained optimization: When the price of “b” rises, how do consumers adjust? the price of b has no effect on this point Quantity of “a”, all other goods Indifference level at the initial point The new expenditure line is steeper slope = -Pb’/Pa The new indifference level is lower Slope of expenditure line = -Pb/Pa Quantity of “b” goods higher prices induce substitution and reduce “real income”

8. Price effects The Demand Curve Price When price changes, consumers move along their demand curve. Welfare is lower at higher prices (later, we’ll see this as “consumer surplus”) Quantity Consumed

9. Income effects The Demand Curve When income rises, consumers’ demand curve shifts (usually to the right, Price as consumers buy larger quantities at each price) Quantity Consumed

10. Price Elasticity of Demand To measure the “steepness” of demand curves in a more useful way than with its slope, we use Price ($/lb) +5 the elasticity of demand (ε): 1.25 = percentage change in quantity for a percentage change in price = %ΔQ / %ΔP = 5/10 / -.25/1.25 = - .5 / -.2 = - 2.5 -.25 1.00 Quantity Consumed (lbs/yr) 10 15

11. Elasticity and expenditure:“Price-elastic” vs. “price-inelastic” demand Price ($/lb) “Inelastic demand” : demand curve is relatively steep when price rises, expenditure (P×Q) rises | ε| < 1 “Elastic demand” : demand curve is relatively flat when price rises, expenditure (P×Q) falls | ε| > 1 Quantity Consumed (lbs/yr) “Unit-elastic demand” : demand curve is such that when price rises, expenditure (P×Q) stays constant: | ε| = 1

12. Income Effects on Food Consumption Remember that when income rises, consumers’ demand curve shifts (usually to the right) Price ($/lb) 1.25 It’s helpful to draw a curve of consumption on income, for a given price 1.00 Quantity Consumed (lbs/yr) 10 15

13. Income Effects on Food ConsumptionA hypothetical “Engel” curve Quantity Consumed (lbs/yr/pers) Engel curve for all uses 700 Engel curve for food use only 500 200 0 250 500 1000 5000 10,000 Income ($/yr/pers)

14. Income Elasticity of Demand Quantity Consumed (lbs/yr/pers) Income elasticity (e) : % change in Q / % change in Y varies widely by income level, and by type of use 700 500 200 0 250 500 1000 5000 10,000 Income ($/yr/pers) (see slide 31a)

15. Qty. Consumed (kg/year) Elasticity along the Engel Curve 30 inelastic or “normal” negative or “inferior” “necessary” 20 Income elasticity(e=%ΔQ/ %ΔY) is closely linked to income level : income-elastic (“luxury”) goods: e > 1 income-inelastic (“normal”) goods: 0 < e < 1 negative-elasticity (“inferior”) goods: e < 0 elastic or “luxury” 10 0 500 1000 1500 2000 2500 3000 3500 Income ($/year) no effect

16. Average income and price elasticities of demand in Indonesia (estimated in the 1970s) “inelastic” “inelastic” “elastic” “elastic” Reminder: elasticity is %ΔQ/%ΔY (income) or %ΔQ/%ΔP (price). An extreme case: can price elasticity ever go positive? Such “Giffen” goods have finally been documented, by Jensen & Miller (2007)

17. Income elasticities by income group, rural Brazil, 1974-75 (“inferior” for everyone) (“luxuries” for the poor) Effect of income growth among the poorest 30% in Brazil, 1974-75

18. Calorie intake by nutrient group and income level income level in 1962 (log scale) The poorest eat mainly carbohydrates; income growth permits an increase in fats and proteins calories from each nutrient group (percent of total)

19. At high (U.S.-level) incomes, we’ve been switching back to more carbohydrates.

20. Now… what else might be usefulto understand food intake? • Food is a public concern, but it is not a “public good” • Food demand generate subtle kinds of market failures • One reason is that its benefits are important but often delayed and often not observable, so we have: • credit and insurance constraints (the poor can’t borrow) • “behavioral” effects (people are predictably irrational) • weak self-discipline (addiction, obesity, etc.) • distorted perceptions (anxiety, obsession, etc.) • asymmetric information (need quality assurance) • example of Masters and Sanogo (2002) study of infant foods in Bamako, Mali

21. Weight-for-height (WHZ) and height-for-age (HAZ) of children in Mali, relative to international norms The most severe nutritional deficits occur in a relatively brief period

22. Child mortality is closely linkedto their weight-for-height Source: Reprinted from Fawzi, W.W., M.G. Herrera, D.L. Spiegelman, A.E. Amin, P. Nestel, and K.A. Mohamed. 1997. “A Prospective Study of Malnutrition in Relation to Child Mortality in the Sudan.” The American Journal of Clinical Nutrition 65(4): 1062-9.

23. Why are nutritional deficits so severe between 4 and 24 months of age? • exposure to disease • due to initial introduction of water and solids, or • perhaps also more mobility and contact with others; • deficits in nutrient intake • due to differences between infant and family foods, • and failure to provide enough infant-quality foods.

24. Characteristics of infant foods • Infants need foods of higher nutrient density than the family diet, from when they outgrow the nutrients in breastmilk to when they can digest enough of the family diet. • High density is obtained from high-cost ingredients (oilseeds or animal products), mixed with low-cost staples (cereal grains, etc.) • but parents cannot observe an infant food’s ingredient ratios and density, even after consumption, because other factors affect growth

25. Economics of credence goods Akerlof (1970) and others: If quality is unobservable, quantity will be… zero! optimizing people will not respond to price … except as remedied by trust in a brand. ==> sellers use advertising and high prices as visible commitments to quality, so • buyers must pay a premium over known costs, to buy guaranteed quality, • e.g. premiums paid for “brand-name” lawyers, pharmacists, auto mechanics, etc.

26. Table 1. Infant foods for sale in Bamako, Mali (1999) Brand name Packaging Retail Prices (FCFA/unit)* Mkt. Stores Pharmacy Cérélac (wheat) 400 g. can 1400 1500 1615 Cérélac (wheat) 200 g. box 600 850 Cérélac (rice) 400 g. can 1600 Cérélac (wheat/Banana) 400 g. can 1750 Cérélac (wheat +3 fruits) 400 g. can 2240 Blédilac** (wheat) 250 g. can 1270 Blédina** lactée fruits 250 g. box 1830 Farinor** (maize/soy)400 g. box 1690 1750  Source: Masters and Sanogo, 2002. MISOLA 500 g. bag 300 UCODAL (e.g. Sinba) 200 g. bag 200

27. Economics of credence goods(continued) Akerlof (1970) saw an alternative remedy to lower the cost of credence goods: third-party certification • mandatory testing => complete coverage examples: FDA, USDA limitations: tax-funded, incentive for fraud • fee-for-service => self-sustaining, self-policing examples: ISO, UL limitations: scale economies, may not emerge spontaneously

28. Benefits and Costs of Certification in Mali • benefits of certification • what is food-quality information really worth? • do some mothers value information more than others? • costs of certification • what are set-up and marginal costs? • what volume would be certified? • net gains from certification • do they justify the investment risks? • would the program be sustainable?

29. To detect what certification is worth, we need a experimental economics.. Vickrey (1961) and others: Preference-revealing auctions mimic real markets: price is fixed by others; choice is to accept or decline M&S experiment is aimed at very, very low-income people -- avoids using money, so choice is among infant foods only; avoids calculations, so choice is easy and natural: • give mothers a can of Cerelac • offer to swap for increasing quantities of substitute products • record whether they accept or decline each choice • give them one of their choices, selected at random The design is not quite strategy-proof; respondents may hold out for slightly more of each substitute than in ideal auction.

31. Summary statistics for willingness-to-pay results WTP (FCFA per 400 g.) ave. s.d. min. max. WTP by product 1. Cerelac 1500.00 (market price) 2. Certilac 1159.83 357.77 667 1500 3. Anonymous product 704.59 210.36 462 1500 4. Raw ingredients 119.14 35.07 75 150 WTP for premium for 2 over 3: Certification 455.24 251.22 0 1038 for 3 over 4: Processing 585.45 213.97 312 1425

32. So certification has large benefits…but what would be its cost? • 10 million FCFA/month for advertising • 1 m. FCFA/mo. per 50 tests/mo. for staffing • 0.6 m FCFA/mo. per 20 tests/mo. for transport • 3,520 FCFA/mo. per test for consumables • 1,000 x 400g. sold per test done => cost falls below ave. WTP at 30,000 bags/mo. => cost falls below 100 FCFA at 215,000 /mo.

33. How much food would be certified? • 50,000 children aged 6-24 months in Bamako • 100 g./day average consumption per child => 8 bags of 400 g. per month • low scenario: 39% use (now using Cerelac) • high scenario: 89% use (now using any food) => total potential is 155,000 – 350,000 bags/mo.

34. Would certification’s benefits exceed its costs? • Two ways to estimate: (1) Each respondent has a different WTP • Some respondents’ WTP below cost • Construct demand curves & optimal quantity • Find economic surplus (2s) All respondents have identical WTP • Average WTP is way above cost • Full market size is certified • Find total benefits - costs

35. Estimated certification demand and average-cost curves

36. Table 3. Consumer surplus and net benefit from certification Case 1 Case 2 Consumer surplus approach Equilibrium qty. certified(400 g. bags/mo.)150,000345,000 Net economic surplus gain(FCFA/month)51,543,984134,029,438 Net economic surplus gain(US$/year)951,5812,474,390 Cost-benefit analysis approach Total estimated market size(400 g. bags/mo.)154,746353,138 Net economic benefit/month(FCFA/month)51,713,454132,138,264 Net economic surplus gain(US$/year)954,7102,439,476

37. OK, so quality information is needed…should testing be voluntary or mandatory? • Private solutions work if a “certifier” can • force users to pay (that is, exclude free-riders), e.g.: • Underwriters’ Laboratories (sellers pay for label) • Consumer Reports, CarFax (buyers pay each time) • and thereby capture enough revenue • to pay the cost of obtaining the information; while • maintaining a credible quality signal of their own! • Often it’s cheaper to use the government: • “value capture” from consumers can occur through taxation (e.g. USDA grades) or user fees (e.g. air traffic) • quality-destroying competition can be barred by mandatory standards and licensing • but the “quality” being upheld may not be valued by consumers; standards may be just an entry barrier to protect insiders

38. …Some conclusions • Food intake decisions involve • not only observable attributes (taste, texture, color), • but also unobservable qualities and delayed effects. • Food choices can be hard to understand using economics on observable prices and quantities only • so sometimes we need inferences and experiments • but much of the “irrationality” in food consumption can be understood as constrained optimization • this allows us to plan public interventions that raise welfare, even though food is not a public good.