How Cool is C.O.O.L.?

Transcription

1 How ool is.o.o.l.? Elias Dinopoulos Grigorios Livanis arol West* University of Floria urrent Version: June 27, 2005 bstract: his paper evelops a partial equilibrium moel of a small open-economy proucing an traing an unsafe prouct that is supplie by perfectly competitive proucers. he presence of prouct safety consierations, in this case risks to health, introuces a wege between the market prices proucers receive an the higher risk-ajuste prices consumers respon to. he size of the wege epens positively on the per-unit cost of illness an the proportion of unsafe units emboie in the parent risky prouct. he moel is use to analyze the welfare effects of trae with an without a country-of-origin labeling (OOL) program. ssuming imports are less safe than omestic prouction, the welfare gains from trae in the absence of OOL are ambiguous an may justify the imposition of a trae ban. Even if a full ban oes not improve welfare, some restriction of trae is always welfare-enhancing. hese outcomes erive from an informational istortion that prevents consumers from istinguishing the ifferent countryspecific risks emboie in the foreign an omestic proucts resulting in a pooling equilibrium. he presence of a OOL program removes the informational istortion an generates a welfare maximizing separating equilibrium in which the safer (omestic) prouct commans a higher market price. In the presence of a OOL program, more trae cause by a reuction in protection is better than less trae. Key Wors: ountry-of-origin labeling, protection, prouct safety, welfare. JEL lassification: ress for orresponence: Elias Dinopoulos, Department of Economics, University of Floria, Gainesville, Fl elias.inopoulos@cba.ufl.eu. n electronic version of the paper is available at * Elias Dinopoulos an arol West are Professors of Economics at the University of Floria. Grigorios Livanis is postoctoral research associate at the International gricultural rae an Policy enter, Department of Foo an Resource Economics, at the University of Floria.

2 How ool is.o.o.l.? 1. Introuction he incience of fooborne iseases has ramatically increase in the past fifteen years in the Unite tates an in other inustrialize countries. ccoring to the enters for Disease ontrol an Prevention (D 2004), fooborne infections in the Unite tates annually cause approximately 76 million illnesses, costing $23 billion per year. Wiely publicize outbreaks such as Ma ow isease (Bovine pongiform Encephalopathy or BE), avian influenza ( bir flu ) an the contamination of animal fee with cancer-causing ioxin an polychlorinate biphenyls (PBs) have le to greater consumer awareness of potential foo hazars an increase consumer eman for safer proucts. oncomitantly, these outbreaks have triggere national revisions in trae policies. he efficacy of these policy responses is the focus of this research. he imposition of temporary import bans has been one response. BE outbreaks resulte in a spate of such bans in virtually worlwie ban on anaian beef exports followe the May 20, 2003 announcement that a single breeer cow in lberta ha teste positive for BE. By ugust, anaa s beef export market ha winle from $4.1 billion annually to near zero. In less than ten ays following the December 23, 2003 iagnosis of a BE case in the Unite tates, over 30 countries ha banne U imports, incluing Japan, traitionally the largest buyer of merican beef. More recently, outbreaks of bir flu in Delaware an exas prompte the European Union to ban imports of poultry from the Unite tates. ountry-oforigin labeling (OOL) is another policy measure aressing the problem of potentially unsafe foo imports. OOL allows consumers to ifferentiate proucts that potentially emboy ifferent health risks as a consequence of the uneven geographical origins of fooborne iseases. Japan has manate a OOL for all meat imports since In the U.., the 2002 Foo ecurity an Rural Investment ct calle for voluntary OOL on eptember 30, 2002 an manatory OOL by eptember 30, 2004 for a number of foo proucts such as beef, pork, fresh fruit an vegetables (Feeral Register 2003). Recently, ongress approve a two-year elay for OOL implementation. Juxtapose against the emotional intensity that often surrouns health-relate issues an the sometimes extreme measures that have been implemente to eal with fooborne iseases in 1

3 particular, is a relatively scant literature analyzing the economics of trae policy in risky fooproucts. Many questions remain unanswere. From an economic welfare perspective, are trae embargoes rational when there is a foo safety concern? Perhaps uner some circumstances but not others? What are the welfare effects of policies such as OOL? houl a OOL be augmente by traitional protectionist trae policy instruments (e.g., tariffs)? Prouct safety, an in particular foo safety, issues have been analyze theoretically an empirically, but primarily in the context of close-economy, partial-equilibrium moels. 1 In the international context, there are three relate literatures. First, consierable attention has been given to prouct quality an government intervention to help exporters overcome informational barriers that impee foreign market entry (in particular, averse country-of-origin reputations). 2 While this set of stuies an the current one each emboies a type of enogenous quality etermination, the nature an consequences of prouct quality ifferences, key ecision-making units, international trae context, an pertinent policy analyses iffer substantially. 3 econ, consumer inability to istinguish safe an unsafe proucts in the marketplace resembles consumer inability to istinguish goos by prouction process (eco-frienly, sweatshop, etc.) which, if known, woul affect willingness to pay. 4 ince welfare analyses of trae policy in the latter context lack explicit representation of how prouction process affects consumer utility, it is ifficult to irectly compare these stuies with the present moel. In aition, labeling in this literature is stanar-conforming certification, a process that allows the consumer to efinitively separate proucts of ifferent quality in the marketplace. In contrast, in the current moel, risk of purchasing an consuming an unsafe prouct cannot be 1 ee, for instance, Oi (1973), Epple an Raviv (1978), pence (1977), hapiro (1983), Daughety an Reinganum (1995), an Boom (1998) for theoretical analyses on prouct safety, among many others. For the impact of foo safety on meat eman an for a partial review of empirical stuies on foo safety, see Piggott an Marsh (2004). For extensive theoretical analysis in a omestic context, see Fulton an Giannakas (2004) an references therein. 2 ee Grossman an Horn (1988), Bagwell an taiger (1989), Falvey (1989), Bagwell (1991), Raff an Kim (1999), an hisik (2003) among many others. 3 n interesting extension of the present research woul link to these previous analyses by incorporating the possibility of consumer misperceptions of the safety of a specific country s exports as a consequence of the publicize outbreak of a fooborne isease in that country. Depening upon the nature of the isease, an the feasibility of its plausible incorporation into the potential exporter s explicit choice between high quality an low quality prouction, the situation coul have similarities to examples that motivate the analyses of country-of-origin reputations. 4 Haener an Luckert (1998) an Blen an Ravensway (1999) provie empirical evience of consumer willingness to pay a green premium. heoretical founations of the literature ate from the classic kerlof (1970) stuy of the hien quality problem associate with lemons in the use car market. Recent work by Gaisfor an Lau (2000) an Beaulieu an Gaisfor (2002) aress welfare implications of inistinguishable stanar-conforming an nonconforming goos, incluing effects of certification labeling. 2

4 completely eliminate. If the only labeling possibility is country of origin, an if the consumer knows the proportion of imports that are stanar-conforming versus non-stanar conforming, then straightforwar representation of how prouction process affects utility reners the situation a special case of this stuy s more general moel. thir stran of literature analyzes rules of origin (ROO) that prevent transshipment in a Free rae rea (F). he effects of ROO on trae, welfare an istribution of rents in the supply chain uner various market structures have been extensively examine in the literature. 5 While both ROO an OOL involve country labeling, there are critical istinctions for policy analysis moeling. ROO impacts the consumer irectly via price (higher or lower epening upon eligibility for tariff-free shipment). OOL, in contrast, irectly influences consumer behavior by expaning information on prouct attributes he/she associates with expecte prouct safety. Price consequences are only inirect, as the change in consumer information alters eman conitions. More basically, the unobservable prouct-quality ifferences inherent in a OOL analysis give rise to the possibility of ifferent prices for omestic an foreign prouction that o not characterize homogeneous ROO markets. his paper evelops a partial equilibrium moel to analyze the welfare effects of a OOL program in the presence of risky foos supplie by omestic an/or foreign proucers uner perfect competition. he theoretical moel uses builing blocks from the seminal stuy by Oi (1973), who establishe that in the presence of insurance markets, the uncertainty associate with the risk of consuming an unsafe prouct is reflecte in the risk-ajuste price (RP) 6. Higher than the market price, the RP inclues the proportion of unsafe units in the parent prouct an expecte amage costs of consuming those hazarous units. In this paper, we consier a prouct with an inherent health risk that is measure by the proportion of its unsafe units supplie in the market. he consumer knows with some exogenous probability (equal to the proportion of unsafe units) that the prouct is risky, but cannot etermine whether the consumption of any particular unit will lea to averse health outcomes. he primary focus of the research is the welfare effects of international trae in such a prouct an international trae policies with regar to such a prouct, when safety varies by country of origin. 5 Krueger (1993, 1997), Lloy (1993), Lopez-e-ilanes, Markusen an Rutherfor (1996), Roriguez (2001), Falvey an Ree (2002). For a recent literature review see Krishna (2005). 6 In this paper we use the terminology of risk-ajuste price instea of full price, because it is more self-explanatory. he full price concept was evelope by Becker (1965) to analyze the ultimate consumption flow. 3

5 We first analyze the effects of prouct risk an severity of the isease (as measure by the per-unit opportunity cost plus monetary cost of illness) on social welfare for the autarky (closeeconomy) equilibrium. Because the presence of safety risk creates a ifference between the consumer (RP) an the proucer (market) price, the market for a risky prouct might not exist (Proposition 1). s intuitively expecte, an increase in the riskiness of the prouct, or in the cost of illness, leas to a ecrease in social welfare (Proposition 2). For a small country that imports a riskier prouct than it prouces, movement from the autarky equilibrium to free-trae in the absence of a OOL program (i.e., un-ool trae) leas to a ecrease in the prouction of safer omestic units an to a ecrease in proucer surplus. he effect of un-ool free trae on expecte consumer surplus is ambiguous an the welfare ranking between un-ool trae an autarky is also ambiguous (Proposition 3). his result is consistent with the theory of istortions (Bhagwati 1971): Un-OOL trae involves an informational istortion associate with the inability of consumers to assign the correct risk level to omestic an foreign goos that leas to a pooling equilibrium an ambiguous gains from trae. his result allows for the possibility of welfare-enhancing import bans. Even in the case where an import ban oes not ominate un-ool free trae, some restriction of un-ool trae is always welfare-enhancing. We then analyze the effects of introucing a OOL program that permits the consumer to ifferentiate safer omestically prouce goos an less safe imports. Equilibrium requires equalization of the RPs between the omestic an foreign goos resulting in an increase in the price an quantity of the healthier omestic prouct an an increase in the proucer surplus (Proposition 4). imultaneously, the implementation of OOL leas to a ecrease in aggregate safe quantities of the prouct consume an a ecline in the expecte consumer surplus. OOL removes the informational istortion associate with ifferential risk levels an reestablishes the traitional gains from trae (Proposition 5): In the presence of OOL, more trae (cause by a reuction in a tariff) increases the welfare of a small country even if it imports riskier goos. More OOL trae is better than less OOL trae an welfare uner OOL trae excees that of autarky or un-ool trae. While no moel can thoroughly aress the multiplicity of issues regaring foo safety an global commerce, the current theoretical moel shes some light on the efficacy of trae policies commonly propose to eal with those issues. It is a first step in eveloping a rational approach 4

6 to the economic cost-benefit analysis of OOL programs that several inustrial countries, incluing the U, are consiering implementing or have recently implemente. ection 2 of the paper evelops the moel an stuies the properties of the closeeconomy equilibrium. ection 3 introuces the economics of un-ool free-trae, an erives its welfare implications, for a small country importing a riskier prouct than it prouces omestically. ection 4 analyzes the economic effects of OOL trae. onclusions are provie in the last section an some proofs are relegate to appenixes. 2. lose-economy Equilibrium onsier an economy proucing an unsafe (risky) goo enote by an an outsie composite safe goo Y, which will be use as the numeraire. ssume that labor is the only factor of prouction, an that each unit of goo Y requires one unit of labor, implying that wages are equal to unity. o focus on the analysis of prouct safety we assume that perfect competition prevails in all markets an consumers have ientical preferences. he risk associate with a purchase of the unsafe goo is capture by the assumption that it emboies a certain proportion, λ, of safe units, = λ, an a remaining unsafe portion, (1 λ) with 0 λ 1. onsumption of safe units yiels positive utility, but consumption of unsafe units not only results in no aition to utility, but simultaneously incurs a cost L per unit of unsafe goo consume. While the consumer knows the risk of becoming ill, capture by parameter λ 7, he/she cannot ifferentiate between a safe an an unsafe unit. For instance, accoring to the Foo afety an Inspection ervice (FI 2004) of the UD, a consumer faces a (1 λ) = probability of becoming ill from almonella, if he/she eats one egg. his probability is obtaine by iviing 174,356, the estimate number of annual illnesses attribute to almonella for 2000, by the U.. population to obtain the per-capita chance of becoming ill an then iviing the resulting expression by 178, the annual per-capita consumption of eggs. n alternative interpretation of the risk emboie in is as follows: Rather than an expecte proportion, λ, of safe units in any purchase an a resultant (1 λ) expecte 7 In our formulation the probability of the averse health outcome is treate as objective information. hus, any consumer who faces the same problem will assign the same probability. One coul introuce the case where λ epens on self-protection actions an on a set of information (i.e. past experience) that each consumer uses in forming risk perceptions. his is an interesting generalization that is left for future research. Notice, though that any valuation of a public policy change shoul be base on objective risks. 5

7 proportion of unsafe units, the consumer making a purchase of faces an exogenous probability (1 λ) the purchase will make him/her ill. oncomitantly, with probability λ, the purchase can be expecte to be consume without averse health consequences. If a particular expeniture on turns out to be a ba lot, then the associate cost of illness is proportional to the volume of purchase an consume. Both interpretations yiel the same key relationships that are use in the subsequent moel analysis, but formal presentation is restricte to the first interpretation. We postulate the existence of a competitive health (meical) insurance market that provies insurance to all consumers in the market against the loss L cause by the consumption of the unsafe prouct. Loss L is given exogenously an captures the irect (i.e., meical treatment) costs an the inirect (i.e., lost wages) costs of illness per-unit of unsafe consume. Parameter L can be as large as the economic cost of life (as in the case of the Ma ow isease) an in principle epens on the quality of the health system. Following the insurance literature, we further assume that the insurance offere to the consumers is full an actuarially fair in the sense that the insurance premium equals the expecte value of the insurance claims. 8 In the case of eggs, one can measure the expecte amage cost to the consumer using the cost of illness (OI) ata available on the ER website Fooborne Illness ost alculator (FOI 2004). he OI metho inclues both irect an inirect costs of an illness. In the case of almonella in eggs, the ER website ata imply that the average cost of illness is $2,126. ssuming the representative consumer erives utility only from the safe units of prouct an from the outsie (safe) goo Y ; an, following the stanar approach to partialequilibrium analysis, suppose that the utility function is separable in an Y where (, ) ( ) U Y = u + Y (1) u( ) is an increasing an concave function of the safe quantity of the risky goo inicates that the consumer oes not receive any utility from the unsafe units an 9. In this formulation, the price of prouct Y is equal to unity (numeraire), while the market price of 8 he absence of an actuarially fair an full insurance complicates the analysis. ee Oi (1973) an especially Epple an Raviv (1978) among others for more etails on this issue. Oi (1973) aopts the assumption of full an actuarially fair insurance, while Epple an Raviv (1978) provie also results for the case of partial insurance. 9 he analysis can be generalize to the case of severe risky proucts that result in a negative utility level if one risky unit is consume. his novel extension is beyon the scope of the present paper. 6

8 prouct is enote by P. ince Y enters the consumer s utility linearly, equation (1) allows us to focus on partial-equilibrium analysis, while the assumption of a single unsafe prouct will be relaxe in section 4 that analyzes the economic effects of OOL. he above notation an assumptions imply the following maximization problem for the representative consumer. For a given amount of purchase, only λ will yiel positive utility. Insuring against the expecte cost of illness (or setting asie the funs to pay for it) requires expeniture of (1 λ)l. ssuming a total buget M, an a price P of consumer s maximization problem is as follows: ( λ ) + (1 λ), the Max u M P L (2) he first-orer conition for (2) can be written as P (1 λ) u ( ) = + L (3) λ λ where a prime superscript enotes a partial erivate an the argument in the left-han sie of (3) is equal to the amount of safe foo consume ( = λ ). oncavity of u. guarantees that the secon-orer conition for (2) is satisfie. Recalling that the consumer erives utility only from goo Y an the safe units of prouct, it is obvious from (3) that the solution to the utility maximization problem (2) is ientical to maximizing (, ) ( ) where U Y = u + Y subject to the buget constraint M = P ˆ + Y, ˆ P (1 λ) P = + L (4) λ λ is the risk-ajuste price (RP) of an unsafe goo. In the presence of actuarially fair insurance, the economic interpretation of (4) is escribe elegantly by Oi (1973) 10 : price (expecte cost) of obtaining a safe unit of a risky prouct, () ˆP is the risk-ajuste P / λ is the warranty price, an the term (1 λ) L / λ is the actuarially fair insurance premium rate per safe unit. We illustrate the relative magnitue of the RP for the case of eggs emboying the risk of contracting almonella. ubstituting the risk of becoming sick ( 1 λ = ), the cost of illness ( L = $2,126 ), the market price of one grae shell egg ( P = $0.081), the RP of eggs 6 10 ee also Becker (1965) who evelope the technique of ecomposing the full price of an ultimate consumption flow. Notice that the full price concept in these stuies is terme as risk-ajuste price in our paper. 7

9 becomes ˆ P = + 2,126* = $ In other wors, the consumer behaves as if the risk of almonella generates an 8.6% increase in the market price for safe eggs with a corresponing ecrease in the quantity of eggs consume. Equation (3) efines the eman function for the safe quantity of a risky prouct as a D function of its RP ˆP, an is enote by ( P ˆ ). his relationship will be use in calculating the expecte consumer surplus in the welfare analysis. Recalling the relationship between the safe an unsafe quantities of a risky prouct, D be obtaine then by ( ˆ D ) ( ˆ) = λ, the eman for the risky prouct can P = P λ. ubstituting λ = into the left-han sie of (3) yiels the inverse market eman function of the risky goo, where the epenent variable is the market price (as oppose to the RP): ( λ ) (1 ) P= λu λ L (5) Equation (5) yiels the first result of our moel which is state in the following proposition: Proposition 1: market for an unsafe prouct oes not exist if the following conition hols: ( ) P = λu 0 (1 λ) L 0, i.e., the vertical intercept of its market eman curve is non-positive. he conition in proposition 1 efines a lower boun of prouct safety λ L L u ( ) 0 = /( + 0 ) < 1 which varies positively with L an negatively with the marginal utility of consuming the first safe unit of the goo. he supply sie of the economy is moele as follows: We assume that proucers maximize profits with respect to a given market price output of goo supplie is given by ( ) P of the risky prouct an that the P (6) where ( P)/ P 0, an > ( ) 0 P = for a non-negative price P 0 : he supply curve is upwar sloping an has a non-negative vertical intercept. he assumption of a non-negative vertical intercept is not critical for the analysis. Implicit in (6) is the assumption that the supply of a risky goo oes not epen on the proportion of safe units, but simply on per unit market 8

10 price of. 11 For welfare analysis purposes, it is useful to invert (6) an efine the inverse supply of a risky goo ( ) ( / ) P= P = P λ (7) where P ( )/ > 0 an P( 0) = P>0. he supply of safe units ( λ, P) is straightforwarly obtaine by multiplying (6) by the proportion of safe units λ, i.e., ( λ, ) λ ( P = P. ) he autarky (close-economy) equilibrium conition requires equality between the quantity supplie an the quantity emane for the traeable goo : D ( ) (, λ, ) P P L = (8) onition (8) etermines the market equilibrium price P (where subscript enotes the autarky equilibrium) an the equilibrium quantity of, of equation (8) by λ, evaluate at, yiels P ( ) ( λ, ) λ (, λ, ) ˆ (, λ ) D D,. In aition, multiplying both sies P = P L P P L. (9) which implies equality between the equilibrium level of safe units of goo prouce an consume. Having etermine the proucer price P, one can calculate the consumer RP, Pˆ, irectly by setting P= P in (4). Figure 1 illustrates the close-economy equilibrium. he horizontal axis measures the amount of safe an total units, an respectively, an the vertical axis enotes prices (both market an risk ajuste). he upwar-sloppe curve ( ) P illustrates the supply (an inverse D supply) curve of the risky goo, an the ownwar-sloppe curve ( P,, L) λ is the market eman curve for the risky goo, which is implicitly efine by equation (5). 12 intersection of these two curves illustrates geometrically the solution of (8), which yiels the close-economy market-equilibrium price P equilibrium quantity of the risky goo prouce an quantity prouce he. Having etermine the, one can reaily etermine the equilibrium 11 One coul introuce the assumption that the supply of a risky goo is a ecreasing function of λ an analyze the effects of policies that provie irect incentives to proucers to increase the safety of their proucts. his generalization is beyon the scope of the present paper an constitutes an interesting topic for further research. 12 hese curves are not straight lines in general, but the use of linear curves in all figures of the paper is base on expositional consierations. 9

11 amount of safe units = λ by subtracting horizontally the amount of unsafe units (1 λ) from (i.e., the intersection of P an ( λ, P) ). P ( λ, P) P ˆ α ( P ) P P β γ D ( P ˆ ) P D ( λ, LP, ) 0 Figure 1: lose-economy Equilibrium D he ownwar-sloppe curve ( ˆ ), P illustrates the eman curve for safe units as a function of the risk-ajuste price ˆP (efine in equation (3)). Evaluating this inverse eman curve at the equilibrium level of safe units price. rea (α) that is locate below curve to the expecte consumer surplus. curve yiels the equilibrium RP, ( ˆ ) Pˆ, which excees the market D P an above the equilibrium RP,, is equal rea (β+γ), which is locate below the market equilibrium price ( P) Pˆ an above the supply, measures proucer surplus. In our analysis proucer surplus captures the rents to specific factors (or inustry profits) associate with the supply of P risky units, given our assumption that consumers bear all the risks an so proucers are not concerne with the istinction between safe an unsafe units. onsequently, the close-economy equilibrium level of social welfare is measure by area (α+β+γ). his geometric property will be utilize later in the welfare analysis of various public policies. We nee to emphasize that the use of this 10

12 stanar measure of proucer surplus implies that proucers are not liable for the prouction of unsafe units an our analysis abstracts from moral hazar an principal-agent consierations associate with the prouction of unsafe proucts. hese issues have been analyze in close economy moels an constitute an important irection for future research. 13 It is apparent from Fig. 1 that, in the presence of fooborne risk, the close-economy equilibrium involves two istinct types of welfare istortions. First, consumer behavior epens on the RP, which excees the market price. his iscrepancy between the two prices is similar to the welfare effects of a specific tax incience that reuces total welfare. econ, unlike a tax incience, in the present moel the tax revenue is proportional to the value of the unsafe units which oes not yiel any utility to the consumer. In other wors, the corresponing tax revenue is not a transfer but a pure welfare loss associate with the prouction an consumption of unsafe units. his welfare loss is measure by the area ( P ˆ P ) in Fig. 1 an epens on the market quantity of, the proportion of unsafe units, an the per-unit cost of illness. In aition, for any given parameters of the moel an in the presence of perfect competition, the market solution maximizes social welfare (efine as the sum of expecte consumer plus proucer surplus). In other wors, unless the social planner can alter the risk parameter λ (perhaps though testing) or the per-unit cost of illness L (through health care reforms or evelopment of better treatments), the market solution coincies with the maximization of social welfare. o establish this property, enote with social (an private) costs of proucing ( ) the aggregate units of the risky goo. ssuming a total buget of M, the social planner erives utility from the amount of safe units consume = λ an incurs two types of costs, health insurance costs (1 λ)l an prouction costs ( ). Hence, the social planner s problem is Max u ( λ ) + M (1 λ) L ( ) which yiels the first-orer conition ( ) ( ) (1 λ) λ λ u = + L (10) 13 pence (1977), Epple an Raviv (1978), an Boom (1998) among others have evelope close-economy moels that explicitly analyze proucer liability issues. 11

13 In the competitive market equilibrium, P = ( ). omparing (10) with (3) yiels the esire result, namely that, uner consumer liability an full insurance, the market equilibrium maximizes the level of national welfare. Differentiating the equilibrium conitions (8) an/or (9) totally, one can erive the comparative statics properties of the close-economy equilibrium. Proposition 2: he close-economy equilibrium is characterize by the following properties: (a) n increase in per-unit amage cost L shifts the market eman curve in Fig. 1, D ( λ, LP, ), ownwar an generates: a ecline in the market-equilibrium price ; a fall in the equilibrium quantities an ; an increase in the risk-ajuste price Pˆ ; an a reuction the social welfare, measure by expecte consumer plus proucer surplus. (b) ecline in prouct safety, measure by a reuction in parameter λ, reuces social welfare. (c) If the market price elasticity for prouct is not numerically small, i.e., D D ε P = ( P)( P ) > P ( P+ L), then a ecline in prouct safety, measure by a D reuction in parameter λ, shifts the market eman curve in Fig. 1, ( λ, LP, ), ownwar generating eclines in the market-equilibrium price the amount of safe quantity P P, the market-equilibrium quantity an a rise in the equilibrium risk-ajuste price Pˆ. an Part (a) follows straightforwarly from (5) an Figure 1. For part (b), ifferentiate welfare (W = u( λ ) + M (1 λ) L ( ) ) with respect to λ an simplify using (10) to verify W λ > 0. For (c), substitute λ for in (10) an then ifferentiate totally to etermine λ whose sign epens upon the sign of u (.)λ + u (.) + L. Use (5) ifferentiate with respect to (λ fixe) to establish (c). For the rest of the analysis we assume 2.c hols. hese results are consistent with the empirical evience that consumer eman is susceptible to any new information concerning the way consumers perceive objective (or subjective) threats to foo safety as measure by the parameters λ an L. For instance, accoring to Piggott an Marsh (2004) the public will generally respon to a fooborne outbreak by ecreasing its consumption, at least in the short run. Moreover, if the foo-safety problem is recurring, it can result in an inwar shift of consumer eman for a specific goo. In the case of 12

14 the 1996 outbreak of BE ( Ma ow isease) in the Unite Kingom, both the prouct risk (1- λ ) an loss L (equal to the statistical value of life) were large an cause a substantial ecline in the eman for beef. However, if the prouct risk (1- λ ) is very small then even for large L the ifference between the RP an the market price will be small an the eman for a risky prouct will be etermine by its market price. his is consistent with the finings of a report by the Foreign gricultural ervice of the UD (F 1998), which inicate that while E.U. consumers are concerne with foo safety, price of beef is also important in eciing whether or not to purchase beef. hat is, if the price of beef is low enough, consumers may buy it espite any remaining concerns over BE. 3. Un-OOL Free rae Having establishe the welfare properties an comparative statics of the close economy equilibrium, we now analyze the benchmark free-trae equilibrium in the absence of country-of- origin labeling (OOL). We assume that the consumer cannot istinguish imports an omestic goos in the marketplace although he/she knows all the parameters of the moel an the market equilibrium values of the relevant enogenous variables. onsequently, this section analyzes the pooling equilibrium associate with an informational istortion: he inability of the omestic consumer to ifferentiate between imports an omestic proucts. o facilitate the economic intuition an the clarity of the geometric analysis, we will illustrate the free-trae equilibrium for the case of a country that imports an unsafe goo at a fixe international market price small-country case). his implies that the home country faces a horizontal supply curve of imports at the international market price ill equal to 1 λ *. 14 Denote with P * P * (the an each unit of imports carries a risk of becoming the market quantity of the omestic risky prouct an with * the corresponing quantity of risky imports coming from the rest of the worl, where subscript will be use to inicate function an variables associate with the (free) trae equilibrium. Let = λ be the omestic safe quantity consume that oes not result in any averse health outcomes. imilarly, assume that * = λ * * is the corresponing quantity of the foreign 14 Here we abstract from analyzing the case in which a country s imports originate from a variety of countries with ifferent safety parameters λ *. his case can be analyze, but it is beyon the scope of the present paper. 13

15 (importe) safe foo so the aggregate consumption of safe units is given by + *. ssume, for simplicity that parameters λ, λ* [0,1] are exogenous an may iffer from each other. Without loss of generality, assume ( ) max λ *, λ < λ* < λ < 1, which implies markets exist for 0 0 each of the two proucts an home prouces a safer prouct than the rest of the worl. ssume finally that amage costs L associate with consumption of unsafe omestic an importe goos are equal i.e., the illness imports carry is the same as that of omestic prouction. ince consumers cannot istinguish between the two risky goos, free-trae will result in equalization of the omestic an worl prices, i.e., P = P*. he market quantity supplie will be + *, where the omestic quantity supplie is given by the omestic supply curve evaluate at the worl market price. he omestic supply of safe units is consequently given by ( P *) = λ. o etermine the market-equilibrium quantity of imports * note that importe an omestic goos are inistinguishable in the marketplace an their costs of illness per unsafe unit consume are ientical. his implies consumer eman will epen on the average probability of becoming ill λ = ( + *) /( + *). In this case, the solution to the consumer problem can be obtaine by assuming that he/she maximizes utility ( *) u + by choosing the aggregate quantity consume + * subject to the nonstochastic buget constraint M = Y + P ˆ ( +. he first-orer conition of the consumer s maximization problem is then given by ( ) ˆ *) u + * = P = [ P* + (1 λ) L]/ λ (11) where P * is the common market price, P is the free-trae risk-ajuste price, λ Ζ + * + * = = = (1 s*) s* * * * λ + Χ + λ + λ λ* ˆ is the free-trae consumption safety level, an s* = */( + *) is the consumption share of imports. Un-OOL free-trae yiels a common market price (11). ubstituting (12) an ( *) λ P P * (12) an a common RP given by = into (11) etermines the un-ool free-trae equilibrium value of safe imports *, which can reaily be transforme into the marketequilibrium quantity of imports * = * / λ *. 14

16 Fig. 2 illustrates the etermination of the un-ool free-trae equilibrium values of safe imports an λ λ in the (, *) ( *) λ space. pecifically, equation (11) can be written as P* + L = (13) u P + + L ( ( *) *) an, for clarity of exposition, we replicate equation (12) below λ = + *. (14) * + λ λ * Because the omestic quantity of safe units ( P* ) λ ( P ) = is a function only of the worl * price an λ, (13) an (14) constitute a system of two simultaneous equations in two unknowns λ an *. he solution is illustrate in Fig. 2. he upwar-sloppe curve is the graph of equation (13). It has a positive vertical intercept efine by setting * = 0 in equation (13), an a positive slope: λ λ Eq.(13) λ P* + L ( λ ( *)) u P + L 0 Eq. (14) * * Figure 2. Determination of Free-rae Equilibrium of Imports an Foo afety s the quantity of safe imports increases the marginal utility eclines, an the enominator of (13) ecreases. Equation (14) efines a ownwar-sloppe curve in the ( λ, *) space uner 15

17 the assumption that omestically prouce goos are safer than importe ones ( λ* < λ < 1). he vertical intercept of the ownwar-sloppe curve equals λ : For any level of safe omestic units ( P *), as the amount of imports increases, the level of prouct safety eclines. he intersection of these two curves etermines the free-trae equilibrium quantity of safe imports * an the level of λ. he total quantity of imports is given by * = */ λ *. Note that if P* P, the graph of equation (13) in Fig. 2 lies above that of equation (14) for non-zero * an the free-trae equilibrium reuces to the autarky equilibrium. If ( ) P * = 0, (14) in Fig. 2 is unefine at *=0 an otherwise horizontal at λ = λ *. he freetrae equilibrium reuces to one of solely purchasing imports. We abstract from these uninteresting egenerative equilibria by assuming: P * P ( ) < an P * > 0 (15) Fig. 3 illustrates the welfare effects of unsafe foo imports an the autarky welfare level. It oes this by superimposing on the close-economy equilibrium in Fig. 1, the market supply of imports, which is a horizontal line intersecting the vertical axis at P*. P P ˆ P ˆ α δ ( P ) P P * β γ D ( P ˆ ) P 0 + *, Figure 3. Un-OOL Free-rae Equilibrium 16

18 he close-economy equilibrium correspons to the market price P, an the RP Pˆ D where Pˆ ) λ. Price P etermines the proucer surplus (β+γ) an RP Pˆ etermines ( = expecte consumer surplus (α). he opening of trae establishes a lower proucer price results in the importation of P* an * = λ * * units of safe imports an (1 λ*) * units of unsafe imports. Uner un-ool free trae, omestic proucers face a lower price an reuce the quantity of unsafe an safe foo prouce, an therefore the proucer surplus is equal to area (β). he move from autarky to free trae results in a ecline in the proucer surplus which equals area (γ) in Fig. 3. he effects of un-ool free trae on the expecte consumer surplus (an the total welfare) are in general ambiguous. In orer to calculate the RP associate with the free-trae equilibrium, one has to a the safe quantity emboie in imports *, which is etermine in Fig. 2, to the omestic quantity of safe units. he RP uner free trae the total safe quantity consume Pˆ correspons to D + *. he area (α+δ), which is below curve ( P ˆ ) an above the RP Pˆ correspons to the expecte consumer surplus uner un-ool free trae. In general, the consumer welfare ranking between autarky an free-trae is ambiguous an epens on the ranking of the RPs uner the two regimes. he ranking is unclear because the move from autarky to free trae lowers price, but increases risk of illness. Fig. 3 illustrates a case in which the move from autarky to un-ool free trae leaves the economy s welfare unchange: Un-OOL free-trae reuces the market price from P to P* an reuces the proucer surplus by area (γ). his reuction in welfare is the same as the increase in expecte consumer surplus measure by area (δ), cause by a reuction in the RP from Pˆ to P. hus, even if imports are more risky than omestic proucts, the economy is ˆ inifferent between imposing an import ban an engaging in free trae. Of course, in this case consumers like free trae more than the import ban (free trae results in higher consumer surplus than autarky), while proucers woul avocate an import ban base on the effects of trae on proucer surplus. It is straightforwar to show the existence of cases for welfare improving import bans. For example, starting at an equilibrium of inifference between a ban an free trae, a ecline in import safety λ * oes not affect the graph of equation (13) in Fig. 2, but rotates clockwise from its intercept the graph of equation (14) an results in a lesser safe quantity of 17

19 imports *, for any given market price P*. his implies that the total safe quantity available to consumers + * eclines in Fig. 3, an the expecte consumer surplus uner free un- OOL trae falls. ince the expecte proucer surplus epens on the market price reuction in P*, a λ * leaves this component of national welfare unaffecte. onsequently, the expecte gain from trae measure by area (δ-γ) is negative an an import ban improves welfare. One can reaily construct other scenarios that justify the ban of unsafe imports base on other parameter changes. Note finally that even if free trae is preferre to a ban, there always exists a restriction of trae that improves welfare i.e., free un-ool trae is never welfare maximizing for a small country. (Proof is given in the ppenix ). he effects of moving from autarky to free-trae are summarize in the following proposition: Proposition 3: tarting at the autarky equilibrium an assuming that the omestically prouce goo is safer than the importe prouct ( λ > λ *), the introuction of free-trae by a small country results in: (a) ecline in the market price an market quantity of the safer omestic prouct. (b) n increase in the market quantity an safe quantity of the less safe importe prouct. (c) n ambiguous effect on the total safe quantity consume an on the expecte consumer surplus. () ecrease in the safe quantity of the omestic goo an a ecrease in the proucer surplus. (e) n ambiguous effect on national welfare measure by proucer plus expecte consumer surplus, but if welfare improves as a result of free trae, it can always be raise further by some restriction of free un-ool trae. he reason for the ambiguous welfare ranking between the autarky an pooling equilibria can be trace to the theory of market istortions: While free trae introuces stanar efficiency gains, it simultaneously introuces an informational istortion forcing consumers to act on a common (average) safety risk. he moel is consistent with the evience of import bans following outbreaks of fooborne isease abroa. hese bans can be moele as a move from un-ool free trae (with safe omestic an risky importe goos, λ = 1, λ* < 1) to the autarky equilibrium. he welfare consequences of this move are in general ambiguous, an epen on the magnitues of eman an supply elasticities, the severity in the reuction of foo safety 18

20 capture by the risk parameter λ * an the amage costs L. For example, the larger the ifferential of safety risk between omestic an importe goos measure by likely it is that a trae ban will be welfare improving. λ λ *, the more he reason for the existence of a welfare enhancing trae restriction (even if free trae ominates an import ban) also erives from the informational istortion introuce by the imports. In particular, from the iniviual consumer perspective, the marginal unit of carries a risk (1 λ ) < (1 λ*). However, for society, the marginal unit of is importe an carries the higher illness risk (1 λ*). lternatively, the RP associate with a marginal unit of for the consumer is less than the RP associate with a marginal unit of for society. onsequently, consumers over-consume in the free-trae equilibrium, setting up the conitions for welfare-enhancing trae restrictions. 4. OOL rae We are now in a position to analyze the economic effects of introucing country-of -origin labeling (OOL). In orer to keep the analysis as simple as possible, we will not formally analyze the effects of costs associate with implementation of a OOL program. If the costs of instituting an maintaining a national OOL system are fixe or sunk costs, they constitute an aitional welfare cost that can reaily be incorporate in the cost-benefit calculations without altering the qualitative conclusions of the analysis. We will also treat OOL as a government policy introuce after the country has engage in free trae an will maintain the small-country assumption for comparison purposes. he introuction of OOL removes the informational istortion associate with the inability of consumers to incorporate the safety risk ifferential between imports an omestic goos. In the presence of OOL trae, the consumer can istinguish whether a goo is importe or omestic, allowing the two types of to have ifferent prices. Denoting OOL values by subscript, maximizing the consumer s utility function ( ) u + * + Y subject to a eterministic buget constraint M = P ˆ + Pˆ* * + Y yiels the following first-orer conitions for an interior solution : (1 ) ( *) ˆ P λ u + = P = + L λ λ (16) * (1 *) ( *) ˆ P λ u + = P* = + L λ* λ* (17) 19

21 where P enotes the proucer price of omestically prouce, P* the proucer price of imports an ^ inicates associate RP. Uner OOL trae, the consumer buys the prouct with the lower risk-ajuste price, since a safe unit gives the consumer the same utility, whether it is prouce omestically or importe. he ifferent country-specific health risks generate perceive quality ifferences that are reflecte in ifferent market prices. oexistence of both goos in the market requires that consumers erive the same marginal utility from the two risky proucts (that is, the consumer must be inifferent between consuming a safe unit of the omestic goo an a safe unit of the importe goo). his implies that the introuction of OOL results in equalization of RP between the omestic an importe prouct. Formally, equations (16) an (17) imply that Pˆ = [ P + (1 λ) L]/ λ = [ P* + (1 λ*) L ]/ λ * (18) which etermines P an equilibrium RPs, Pˆ = Pˆ*. Unlike the equilibrium analyze in the previous section, OOL trae introuces a market price ifferential in favor of the safer prouct. olving equation (18) for the proucer price of the omestically prouce goo yiels P P* λ λ* = λ + L λ * λ * λ (19) ccoring to equation (19), the goo with the lower safety risk (in this case the omestic prouct since λ > λ * by assumption) commans a higher market price at equilibrium because the consumer perceives it as a higher quality (healthier) goo. ubstituting (19) into the omestic supply of the risky goo yiels the equilibrium safe omestic quantity prouce ( λ, ) λ ( ) = P = P (20) ince the introuction of OOL raises the market price of the omestic prouct relative to the omestic price of imports ( P > P* ), the introuction of OOL generates a higher proucer surplus compare to the free-trae equilibrium without OOL. herefore, abstracting from implementation costs, the introuction of OOL will be supporte by proucers of omestic goos that are safer than importe ones. OOL effects on expecte consumer surplus epen on OOL effects on the RP. From (12) it is obvious that λ = (1 s*) λ+ s* λ* > λ*. It follows from (11) an (17) that 20

22 ( *) u + = ˆ P* (1 λ*) P* (1 λ ) P ˆ = + L > + L = P = (21) λ* λ* λ λ ( *) u + he total amount of safe quantity consume uner OOL free trae is strictly less than the corresponing quantity uner un-ool free trae, i.e. + * < + *. his inequality follows from the concavity of the consumer s utility function. he result implies that, starting at the un-ool free-trae equilibrium, the introuction of OOL reuces the expecte consumer surplus, which is an increasing function of the aggregate safe quantity consume. ince the consumption of the safer omestic goo increases with the introuction of OOL, (i.e., > ), the safe (an market) quantity of imports eclines (i.e., * < *). hus, OOL increases the omestic market price of the safer (omestic) prouct, reuces the quantity of the less safe (importe) prouct by more than the increase in omestic prouction, an results in a reuction of expecte consumer surplus. hese results lea to the following proposition that summarizes the economic effects of introucing a OOL program. Proposition 4: tarting at the un-ool free-trae equilibrium an assuming that the omestically prouce goo is safer than the importe prouct ( λ > λ *), the introuction of OOL by a small country results in: (a) n increase in the market price an market quantity of the safer omestic prouct. (b) ecline in the market quantity an safe quantity of the less safe importe prouct. (c) ecline in the total safe quantity consume an a ecline in the expecte consumer surplus. () n increase in the safe quantity of the omestic goo an an increase in the proucer surplus. imports, ( We are now in a position to establish the optimality of OOL. With a price *, an P for omestically prouce,, an a omestic cost of proucing ), the level of national welfare in general is given by: ( λ λ ) λ λ ( ) W = u + * * + M [(1 ) L + (1 *) L*] P* * n interior maximum for W ( > 0, * > 0), requires: ( ) ( ) λu λ + λ* * (1 λ) L = 0 an (23) P* for, (22) 21

23 ( ) λ* u λ + λ* * (1 λ*) L P* = 0 (24) he concavity of u an > 0 assure secon orer conitions are satisfie. Given that both an u are monotonic, if an interior maximum exists, it is unique. ince ( ) = P, (16) an (17) imply an interior OOL equilibrium is this unique maximum. Note that un-ool trae can never maximize welfare since joint satisfaction of (23) an (24) at ( > 0, * > 0 ) requires ( ) > P* (assuming λ > λ *) an in the un-ool trae equilibrium ( ) = P *. ppenix B establishes that if an interior maximum exists, it ominates corner solutions of = 0 an * = 0. Hence, if an interior OOL equilibrium exists, it ominates both autarky an the un-ool trae equilibrium. corner OOL equilibrium at all imports is preclue by (15). 15 corner OOL equilibrium at the autarky solution can be consistent with (15). In that case, P + (1 λ) L = u ( λ λ ) < P * + (1 λ*) L λ * i.e., at the autarky solution, the marginal utility of an aitional unit of is less than the RP of buying it as an import an hence, there is no market for imports. In this situation, clearly the equivalent autarky an corner OOL equilbria ominate the un-ool trae equibrium because no interior maximum exists. Hence, welfare uner a OOL regime always excees that of un- OOL trae an it excees that of autarky except in cases the two are equivalent. Furthermore, it is straightforwar to establish that any restriction of OOL trae reuces welfare 16. If a non-prohibitive specific tariff, t, is impose on imports, the OOL equilibrium can be etermine as above by replacing P * by P* + t. Equations (17), (19) an (20) with P* replace by P * + t etermine the market-equilibrium values of P,, an *. ubstituting yiels = λ an * = λ * * in (17) an ifferentiating totally the system of these equations P* + (1 λ*) L (0) + (1 λ) L 15 corner OOL equilibrium at all imports can exist only if < which is preclue λ* λ by (15) an (.) > 0 which imply (0) < P*, an λ > λ *. 16 For this erivation we assume an interior OOL solution since imposing tariffs in the corner OOL solution of autarky is uninteresting. 22