Serviços Personalizados
Journal
Artigo
Indicadores
- Citado por SciELO
- Acessos
Links relacionados
- Similares em SciELO
Compartilhar
EconoQuantum
versão On-line ISSN 2007-9869versão impressa ISSN 1870-6622
EconoQuantum vol.5 no.1 Zapopan Jan. 2008
Artículos
Vertical product differentiation and credence goods: Mandatory labeling and gains from international integration1
Ian Sheldon*and Brian Roe**
* Professor of International Trade, Environmental and Development Economics, The Ohio State University, 2120 Fyffe Road, Columbus, Ohio43210, USA. Email: *sheldon.1@osu.edu
** Professor in the Department of Agricultural, Environmental and Development Economics, The Ohio State University, 2120 Fyffe Road, Columbus, Ohio43210, USA. Email:**roe.30@osu.edu
Fecha de recepción: 23/02/2009
Aceptación: 08/05/2009
Abstract
Using a model of vertical product differentiation, we show under what institutional circumstances welfare gains will be maximized as economies integrate and harmonize (mutually recognize) their (each other's) labeling and certification policies for credence goods. Specifically, we show that harmonized mandatory, exclusive discrete labeling will not maximize the gains from economic integration, i.e., the choice of labeling regime can have a negative effect on market structure if firms choose to exit, reducing the range and quality of goods in the integrated market. In contrast, with mandatory, exclusive discrete labeling, there is a higher probability that the gains from integration will be maximized where countries mutually recognize each other's labeling standards.
Keywords: Vertical differentiation, credence goods, labeling, economic integration.
Resumen
A través de un modelo vertical de diferenciación de productos se muestra el marco institucional que maximiza el bienestar cuando dos economías se integran comercialmente y coordinan sus políticas de etiquetado y certificación. En particular, se muestra que el etiquetado discreto, armonizado y obligatorio no maximiza ganancias. Por ejemplo, la elección del tipo de etiquetado puede tener un efecto negativo sobre la estructura de mercado si las empresas deciden salirse, reduciendo así el rango y la calidad de los bienes. Pero con reconocimiento mutuo de estándares de etiquetado existe una probabilidad más alta de ganancia.
Introduction
While goods are often differentiated by process attributes, consumers may be unable to verify such claims, i.e., credence goods (Darby and Karni, 1973). Labeling is one method for addressing the credence good problem, requiring a number of regulatory choices concerning the labeling regime: compulsoriness (mandated or voluntary), explicitness (discrete or continuous), and exclusiveness (only government labeling is available or private firms may also certify). Beyond domestic regulation, labeling rules are also an important issue in trading relations among countries. As countries become more integrated economically, they typically have to agree on either harmonizing or mutually recognizing their rules concerning labeling and certification of credence goods. Harmonization implies that when two countries integrate economically, an agreed upon standard applies in both countries. In contrast, mutual recognition implies a countryoforigin principle is applied, i.e., a standard applied in one country is recognized in the other country. Likewise, any standard set in the latter country is recognized in the former country (Leebron, 1996; Lutz, 2000).
In an earlier paper, we used a model of vertical product differentiation to analyze the efficiency and distributional implications of different approaches to the labeling of credence goods in an economy under autarky (Roe and Sheldon, 2007). In this paper we extend the institutional setting by allowing for the integration of two economies where they agree either to harmonize or mutually recognize their credence good labeling regulations. Specifically, we examine two cases of economic integration: the first involves two countries with identical distributions of income, which we denote as NorthNorth integration; the second involves countries which have overlapping distributions of income, which we denote as NorthSouth integration.2
Using these two cases we show under what institutional circumstances welfare gains will be maximized, as economies integrate, assuming specific rules on credence good labeling are set exogenously by each country's regulatory authorities.3 For NorthNorth and NorthSouth integration, mandatory, exclusive and continuous labeling delivers the same prices and qualities as would occur under perfect information. Alternatively, if there is mandatory, exclusive, and discrete labeling, the probability of lower welfare gains depends on whether countries harmonize or mutually recognize their labeling standards. If a harmonized standard is set too high or too low, higher quality good(s) are pushed out of the market for both NorthNorth and NorthSouth integration, thereby lowering welfare. In contrast, mutual recognition of labeling regimes increases the probability that welfare gains from NorthNorth and NorthSouth integration will be achieved.
The remainder of the paper is structured as follows. We introduce the structure of the basic model in section 2, followed in section 3 by derivation of equilibrium under autarky with perfect information about quality. In section 4, we examine the case of NorthNorth integration with perfect information, followed by an analysis of different possible creden a good labeling regimes. Then in section 5, we conduct the same analysis with respect to NorthSouth integration. Finally, in section 6 we summarize and end with some concluding remarks.
Basic Model
In this section, which draws heavily on our previous paper (Roe and Sheldon, 2007), we outline the basic structure of a model of vertical differentiation with perfect information, first introduced by Gabszewicz and Thisse (1979, 1980) and Shaked and Sutton (1982, 1983), and later extended by Boom (1995).
Consumers, firms and quality
We assume that consumers in a representative country have a unit demand for a qualitydifferentiated good. Consumer utility is:
(1)
U = u (yp),
where u [u , ∞] is the quality level of the differentiated good, the lower bound u > 0 meets a minimumquality standard, perfectly enforced by government, y is income, and p is the price of the differentiated good, where (y p) is expenditure on a Hicksian composite commodity.4, 5 If the consumer decides not to buy the differentiatedgood, u = 0; hence, the good is always purchased unless price exceeds income. Consumers derive the same surplus from a good of a particular quality, but differ in their ability to pay. Incomes are uniformly distributed on the interval [a, b], a > 0, so that the density function of income y is:
s(b a) being a measure of the size of the representative economy under consideration.6
Firms produce a single differentiated good and all firms share the same production technology characterized by zero production costs and a fixed, qualitydependent cost, F(u), which is sunk by the firm after entry into the market. 7,8 We assume:
(2)
F(u) = ε + α (u u)2,
where ε and α are strictly positive constants. Sunk costs are convex and strictly increasing in quality. Also note that a sunk cost of ε > 0 must be expended to achieve even the lowest quality good; hence, the sunk cost of producing the minimumquality good, is equal to ε. Finally, note that if goods of differing qualities were all priced at marginal cost, all consumers would choose the same (highest) quality, which is the standard definition of vertical differentiation (Tiróle, 1988).
Game structure
Firms maximize profit in the following oneshot, threestage game. At stage 1, each firm decides to enter or not enter the market, incurring sunk costs ε upon entry. At stage 2, firms that have entered simultaneously choose their good's quality level, incurring the additional fixed costs for producing the chosen quality. At stage 3, firms simultaneously set good prices.
Firms are perfectly informed about consumer preferences, the income distribution, existing labeling institutions and all firms' technologies. We invoke the concepts of subgame perfect equilibrium and BertrandNash competition for the price and qualitysetting stages.
Entry and number of firms
Though solutions to multistage games typically begin with analysis of the final stage, and then proceed by backwards induction, we draw upon previous results in the literature on vertical product differentiation to make some initial remarks about the number of firms that will enter this market in the game's first stage. First, we assume the following:
(3)
4a > b > 2a b/4 < a < b/2,
limiting the dispersion of income across the population, i.e., the difference in income between the richest and poorest member of the population. When all quality levels are communicable via labeling, this ensures that exactly two firms will enter this market, so long as fixed costs plus labeling costs are not prohibitively high, and that each entrant experiences a positive market share in equilibrium, i.e., a natural duopoly. Also, this restriction on income dispersion ensures that each consumer either purchases one unit of the differentiated good or is indifferent between purchasing the lowest quality and purchasing none. When a market is structured this way, i.e., such that all consumers always buy a differentiated good, or are at least indifferent to such a purchase, it is called a 'covered' market.
This result, the socalled "finiteness property" (Shaked and Sutton, 1982, 1983), ensures that equilibrium market structure is endogenous. A proof of this result is sketched out as follows, drawing on Gabscewicz et. al. (1981), and Shaked and Sutton (1984). If goods q = 1,…,n are labeled in increasing order of quality, 0 < u1 < …< un , given (1), a consumer is indifferent between good q at price pq and good q 1 at price pq 1 when:
(4)
uq (y pq) = uq 1(y pq 1),
which defines a point in the income distribution:
(5)
yq = (1 rq ) pq 1 + rq pq ,
where rq = uq / (uq uq1). Therefore, consumers with y > yq will strictly prefer good q to good q 1 and the distribution of income, (ba) can be split up into the market shares of successive firms, i.e., the market share of the firm selling the highest quality good will be xn = (b yn ) the market share of the firm selling the next highest quality good will be xn1 = (yn yn1) and so on.
Suppose that a firm offering a good of quality un competes with a firm offering a good of quality un1offered at price pn1 = 0. The choke price for good n is determined by the upper end of the income distribution b. As pn falls, more consumers are willing to purchase good n compared to good n 1 at a zero price, and if pn falls enough, even consumers at the lower end of the income distribution a are willing to pay for good n, i.e., the total demand for good n being (ba), good n covering the market. Note that if pn = 0 even consumers with zero income prefer good n over good n 1.
If the firm offering good n maximizes its profits, and given that its marginal costs of production are zero, it will end up setting a price such that its market share is equal to b/2. From this it follows that if (ba) < b/2, or equivalently a > b/2 the firm offering good n captures the whole market, a natural monopoly, good n1 having zero market share. If (b a) < b/2 or equivalently a < b/2, good n will no longer cover the market, and if b < 4/a only two goods will have a positive market share. Drawing on Gabscewicz et. al. (1981), proof of the latter result draws on the following lemma:
Lemma 1: For any density function f (y), the necessary condition for an equilibrium in which goods 1 to n have a positive market share is,
Applying this to the density function g (y), and for firms selling goods n and n1, it can be stated that their market shares will be s (b yn) > syn and s (yn yn 1) > syn 1 or b > 2yn and yn > 2yn 1, ie., b > 4yn 1. Since by (3), a > b/4, then a > yn 1, so that goods n2, n3,…,1, will have a zero market share.
Price equilibrium
We now solve the final stage of the game under the assumption that two firms have entered and chosen distinct quality levels (0 < u < u1< u2). Higher income consumers will choose the higher qualitygood. From (5), define y' as the income level of a consumer that is indifferent to buying either the high or lowquality good:
(6)
y' = (1r) p1 + rp2 ,
where r = u2 / (u2 u1) and pq is the price of the good with quality level q = 1, 2. Also note that, given these prices, a consumer is indifferent between a good of quality u1 and no good when p1 = y. Given (6), and assuming a covered market, where p1 < a, profits of the two firms are:
(7)
π1 = sp1 (y' a) F (u1)
(8)
π2= sp2 (b' y' ) F (u2)
By differentiating (7) and (8) with respect to p1 and p2, respectively, setting the two resulting expressions equal to zero and solving the two equations for equilibrium prices, we derive,
(9)
(10)
Substituting the definition of r into equation (9) we can derive a restriction on the ratio of quality levels in a covered market:
(11)
These are equivalent to u2 / u1 ≥ (b + a) / (b 2a), or that the ratio of high to low quality in a covered market is limited by aspects of the income dispersion.
So long as quality can be chosen from the continuum of possible qualities and (3) holds, all consumers will have a choice between two distinct qualities offered by the two firms and will always choose a differentiated good. Analysis of the equations (9) and (10) leads to a first remark concerning market behavior.
Remark 1: In a covered market, equilibrium prices for the low and highquality good increase (decrease) as the difference in quality levels between the goods (u2 u1) increases (decreases), i.e., increasing (decreasing) quality differentiation increases (decreases) all prices.
Autarky equilibrium with perfect information
Express the two firms' profit functions as a function of qualities by utilizing the definition of r and by using equilibrium price expressions from (9) and (10):
(12)
(13)
where û1 and û2 are defined in (11).
Remark 2: The lowquality firm chooses the lowest possible quality in equilibrium, i.e., u1* = u.
Consider the quality choice of the lowquality firm. Firstorder conditions yield:
(14)
for u1 > û1 ( u2 )
The profits of the lowquality firm decrease as it raises quality. Increasing quality increases sunk quality costs and increases price competition with the higher quality firm as discussed in Remark 1. Further, a result of the coveredmarket model is that all consumers buy a differentiated good; hence raising quality never pulls more customers into the market.9 This finding concerning the quality level produced by the lowquality firm corresponds with Boom's (1995) equation (21).
The highquality firm's optimal quality decision follows from differentiating (13):
(15)
where the second derivative is
Given the lowquality firm always chooses u1* = u, firm 2's optimal choice of quality is such that u2 induces a covered market price equilibrium:
The equilibrium quality in a covered market is implicitly defined by:
(16)
The quality pairs of u1* = u and (16) represent a Nash equilibrium.10 This is equivalent to Boom's (1995) equation (24). Aggregate consumer welfare in equilibrium is:
(17)
Following this we can state the following proposition, drawing on Roe and Sheldon (2007):11
Proposition 1: As u2 increases (decreases), (a) the welfare of consumers, purchasing the lowquality good decreases (increases), (b) the proportion of consumers purchasing the lowquality good declines (increases), and (c) aggregate consumer welfare increases (decreases).
Autarky equilibrium is described in figure 1. Firms' fixed costs F(u) and revenue sR(.) are plotted on the vertical axis against quality u, where the low and highquality firm's revenue functions can be derived from (12) and (13) respectively:
(18)
(19)
Suppose the lowquality firm chooses u. If the other firm set its quality at this level, price competition drives firms' revenue to zero, given the assumption of zero variable production costs. In addition, due to sunk costs e both firms would incur a loss. Consequently, the optimal choice of the other firm is to increase quality to u2 in order to maximize profits π2. At the same time, this reduces the degree of price competition with the lowquality firm allowing it to maximize its profits π1. If the lowquality firm were to increase its quality from the minimum u to u1 = u2, price competition again results in both firms incurring a loss. Hence, the equilibrium choice of qualities is u , u2.
It is important to note here that the only perfect equilibrium that can exist is for two firms to enter the market and survive in equilibrium with positive prices and positive market shares. If more than two firms enter, given the assumption of zero variable production costs, price competition ensures that all firms will produce the topquality at a zero price, thereby making zero profits. If any firm produces below the topquality, it will have a zero market share, as consumers will only purchase the highquality good at a zero price. Consequently, given sunk costs ε only two firms can enter and make a profit in equilibrium. Following Shaked and Sutton (1982), we state the following proposition:
Proposition 2: Given the income distribution 4a > b > 2a, for any ε >0, and a number of potential entrant firms n >2, (a) there exists a perfect equilibrium where only two firms enter, producing the distinct qualities, and earning positive profits, and (b) no perfect equilibrium exists where q >2 firms enter.
NorthNorth Integrated Equilibrium
Perfect information
Suppose two NorthNorth economies, N=1,2, with the same uniform distribution of income integrate, where a = a1 = a2, and b = b1 = b2, although they may have different sizes of population, i.e., the population of the integrated economy is s i = s1+ s2, where either s1= s2 or s1 ≠ s2. We also assume that firms must incur some additional sunk costs εi in order to enter the integrated market, and that each country has the same minimumquality standard u prior to integration, such that u = ui
Due to the fact that each economy supports only two firms under autarky, the integrated equilibrium will also support only two firms, i.e., two firms will exit. This follows from Proposition 2, with i > 0. However, given that we cannot predict the location of the remaining two firms, we are unable to predict the direction of trade in the integrated equilibrium.
The integrated equilibrium is described in figure 2. With the increase in the population size from s to si, the highquality firm's revenue function rotates upwards, resulting in an increase in the quality of good 2 to u2. Given u2, the lowquality firm's revenue function shifts out and rotates upwards, the quality of good 1 remaining the same at the minimum ui. As a result, in the integrated equilibrium, while the prices and profits of both firms increase, from Proposition 1, aggregate consumer welfare increases, i.e., the gains from integration come from increased quality. This results in the following proposition:
Proposition 3: In the NorthNorth integrated economy with perfect information, (a) highquality increases to , (b) the equilibrium prices and profits of the low and highquality firm increase, and (c) aggregate consumer welfare increases due to increased quality.
Imperfect information
We now assume consumers in the integrated NorthNorth economy do not believe any firmbased communication concerning quality due to the unverifiable nature of process attributes. All communication of quality occurs through a mandatory label that is administered and verified exclusively by a public agency. We assume the public certifier perfectly monitors and communicates the quality of individual firms ex ante for a fee paid by the firms.12 The fixed cost of certifying and labeling the good is given as:
(20)
I j (u) = I j for u > u
= 0 otherwise,
where j {t, d} and t and d stand for continuous and discrete labeling, respectively. Continuous labels communicate the exact level of quality while discrete labels merely communicate if quality meets or exceeds a particular quality threshold. Firms claiming quality meeting the minimumquality standard are never charged a fee, I j ( u ) = 0, because a firm has no incentive to produce a higherquality good and market it as the minimum quality. We assume there are no economies of size in the costs of public certification, and that such costs are the same throughout the integrated economy. We also assume discrete certification is less costly, I t (u) > I d (u) u > u.13 Then finally, we assume that there are no variable costs of labeling.14
We now consider four cases where quality information is unverifiable by consumers in the integrated NorthNorth economy: (XL) no labeling is possible; (MEC) mandatory, exclusive continuous labeling; (MEDh) harmonized mandatory, exclusive discrete labeling; (MEDmr) mandatory, exclusive discrete labeling with mutual recognition.15
(i) Case XL no labeling
First, suppose quality is opaque to the consumer and that no labeling program exists in the integrated economy.
Proposition 4: In the presence of credence attributes and the absence of labeling, (a) a single firm in the integrated NorthNorth economy supplies the lowest quality level ( ui), charges pNL = b/2 and earns profits π XL = b2 / 4 ε i , (b) at least some consumers purchase no goods, (c) there are no gains from integration.
The sunk cost of entry, ε i, combined with the threestage game supports the entry of a single firm into the integrated market, while the opaqueness of quality and lack of labeling leads to production of the minimumquality standard u . The resulting price and profit levels are simple monopoly outcomes given the linear demand structure that emerges from a uniform distribution of consumers within the given income interval.16 On the consumer side, because pXL = b/2 and, by the restriction (3) on income distribution, the poorest consumer has an income smaller than this, a < b/2. Therefore, some consumers will not consume the good under monopoly.
(ii) Case MEC mandatory, exclusive, continuous labeling
Next consider the case where, in the integrated economy, any firm that claims quality higher than the minimum has to participate in a continuous labeling program. Firm profit functions under this labeling regime become:
(12')
for u1 > û1 (u2)
(13')
for u2 > û2 (u1),
where the only change from profit functions (12) and (13) under perfect information is the addition of the cost of continuous labeling, which is a step function triggered by the sale of a good with quality higher than the minimum, plus firms have to incur the additional sunk costs, ε i, of entering the larger integrated NorthNorth market.
Under our assumptions, firms are able to communicate their desired quality level perfectly via the mandated continuous label. This results in the following propositions.
Proposition 5: For NorthNorth MEC, if I then two quality levels will be produced; otherwise, case MEC results are identical to case NL results.
Proposition 6: For NorthNorth MEC, if I , then , .
Proposition 5 outlines a labeling cost threshold, . Costs above the threshold, which is the entirety of profits less labeling costs earned by the highquality firm, cause the market to collapse to the monopoly analyzed in case XL because no highquality firm would enter. Otherwise, two firms enter and produce distinct qualities.
Proposition 6 points out that, as long as two firms enter, the labeled market is identical to the perfect information market with respect to prices, qualities and profits for the lowquality firm. Only the profit of the highquality firm is different because it incurs labeling costs. Hence, continuous labeling does not distort firm choices so long as it is not too expensive. Consumers experience no change in welfare compared to the perfect information case so long as two qualities are produced, as labeling leaves price and quality unchanged in equilibrium, i.e., the gains from NorthNorth economic integration are still realized under MEC labeling.
(iv) Case MEDh harmonized, mandatory, exclusive, discrete labeling
In the case of harmonized mandatory, exclusive, discrete labeling, we assume that in the integrated market, firms claiming higher than minimal quality have to implement a single harmonized, discrete standard, = , for NorthNorth countries, N = l,2, and firms are forbidden from certifying and communicating any other standard.
Proposition 7: For NorthNorth MEDh, the integrated market will support two qualities if the harmonized standard, where both γ(.) and δ(.) are nonnegative, decreasing functions of I dand . Otherwise MEDh results are identical to XL results.
Proposition 7 outlines an interval in which the harmonized discrete mandatory labeling standard must fall in order for two qualities to be produced. If the authorities choose a standard outside this interval, one or both firms earn negative profits and will not enter the integrated market. Hence, for a standard outside this interval, only one firm enters and the market collapses to the monopoly outcome of case NL. Proposition 7 also points out that, as labeling costs rise, the interval the harmonized standard must fall within shrinks. In other words, as the cost of labeling increases, the authorities in the integrated economy have less room for 'error' (in the eyes of the highquality firm) in setting the harmonized standard because the highquality firm will have less residual profit remaining to entice its entry.
The welfare implications of harmonized standards set lower (higher) than firmpreferred standards are stated in the following proposition:
Proposition 8: For NorthNorth MEDh and , < (>) (a) decreases (increases) aggregate consumer welfare, (b) improves (diminishes) the welfare of consumers purchasing the lowquality good, (c) diminishes (improves) the welfare of consumers purchasing the highquality good, (d) decreases (increases) the profits of the lowquality firm, and (e) decreases the profits of the highquality firm.
If the harmonized standard is lower than then the two qualities are closer together and, as we point out in Remark 1, price competition becomes more intense between the two firms. This bodes well for consumers who purchase the lowquality good, who now pay a lower price. Consumers of the highquality good also pay a lower price, but as was shown in Proposition 1, these consumers would rather have the higher quality and pay the higher price. In aggregate, consumers lose due to the lowering of quality. The more intense price competition harms both firms.
This is obvious for the highquality firm because the harmonized standard deviates from its preferred (profitmaximizing) choice of quality. For the lowquality firm, the loss of profits from a decrease in u2 is obvious after differentiation of (12') with respect to u2.
If the standard is higher than price competition is relaxed. This harms consumers of the lowquality good, who now pay higher prices. Consumers of the highquality good welcome the increase, as they value the quality increase more than they are harmed by the price increase. The relaxed price competition inflates the lowquality firm's profits as they gain a higher price with no increase in production costs. The highquality firm does charge a higher price, but the convex, fixed cost of producing quality comes to dominate and drive the highquality firm's profits down. The highquality firm suffers regardless of the direction of the harmonized labeling standard's deviation from the perfectinformation quality choice.
(iv) Case MEDmr mandatory, exclusive, discrete labeling with mutual recognition
In the case of mandatory, exclusive, discrete labeling with mutual recognition, we assume that in the integrated market, firms claiming higher than minimumquality have to implement either one of two discrete standards, or . If each country sets the same standard, . Proposition 7 applies, but if ≠ , the following proposition can be stated:
Proposition 9: For NorthNorth MEDmr, and ≠ , the integrated market will support two qualities if at least one of the standards, where both γ(.) and δ(.) are nonnegative,decreasing functions of I d and . Otherwise MEDmr results are identical to XL results.
Proposition 9 outlines an interval in which at least one of the discrete labeling standards must fall in order for two qualities to be produced, the interval shrinking as labeling costs rise. Obviously if both countries' labeling standards fall in the required interval, with mutual recognition, the standard chosen will be that closest to the firmpreferred standard. In either case, the distributional implications of the chosen standard are the same as stated in Proposition 8. If both countries set a labeling standard outside this interval, one or both firms earn negative profits and will not enter the integrated market. As a result, only one firm enters and the market collapses to the monopoly outcome of case XL.
The key to the MEDmr case is that, because the firm choosing highquality has two standards to choose from through the principle of mutual recognition, there is a higher probability that the gains from economic integration will be realized due to one of the standards being close to the highquality firm's preferred standard.
NorthSouth Integrated Equilibrium
Perfect information
Suppose two economies, North and South, each have incomes uniformly distributed over the range [ak, bk], and 4ak > bk > 2ak , where subscript k refers to either North (N) or South (S). In addition, assume that aN > aS , bN > bS , and bN < 2bS , aN < 2aS , and that the same technology is available in North and South. Under autarky, both North and South will be able to sustain two firms in equilibrium selling distinct qualities. Also assume that the North sets and enforces a higher minimumquality standard than the South, such that Consequently, in the North, given the higher minimumquality standard, the highquality firm, in order to escape the pressure of price competition, will also produce and sell a higherquality good in equilibrium, which follows from differentiation of (16):
(21)
resulting in the low and highquality goods in the North under autarky being of higher quality than their counterparts in the South.
We now allow North and South to integrate, assuming as before that firms must incur some additional sunk costs εi in order to enter the integrated market. In addition, assume that North and South mutually recognize each other's minimumquality standard. Following Gabscewicz et. al. (1981), the conditions postulated on the income distribution imply:
(22)
,
such that in the integrated equilibrium, the following inequalities must hold , where yn is the income of the consumer indifferent between a good of quality q offered at pq, and a good of quality q1 offered at pq1 , q=1,…,n. Since the income of the consumer who is indifferent between consuming the minimumquality good from the North and the minimumquality good from the South, yn2 , is less than or equal to the lowest income in the integrated economy, aS , the integrated economy can only support three goods in equilibrium. In other words, the minimumquality good in the South will be eliminated due to economic integration. There will be intraindustry trade, if the mediumquality good is produced in the South, and the minimum and highquality goods are produced in the North.
The benefit to consumers of economic integration follows from the reduction in prices of the remaining three goods, qn , qn1 , and qn2. In other words, for the lowestquality good qn3to be eliminated there must be a reduction in the price of qn2 that makes even consumers of income aS better off than before. In addition, as consumers with income y > aS can do at least as well as those with aS , all consumers with income above aS must also gain due to the fact that pq and pq1 are also reduced. Consequently, we can write the following proposition:
Proposition 10: If NorthSouth have incomes uniformly distributed over the range [ak, bk], and 4ak > bk > 2ak, where each economy supports two goods under autarky, then if aN > aS , bN > bS , and bN < 2bS , aN < 2aS , the integrated economy supports only three goods in equilibrium, with qualities, > > . Aggregate consumer welfare increases due to lower prices in the integrated market.
Following Gabscewicz et al. (1981), the proof of this result proceeds in stages showing that the market share of the highestquality good qn extends beyond that of the mediumquality good qn1 extends below aN , while that of the minimumquality good qn2 extends below aS . ( i ) yn < bS, i.e., the market share of the highestquality good extends below bS . Suppose yn ≥ bS , from Lemma 1, we know SN dy > SN yn, i.e., bN yn > yn or bN /2 > yn. By assumption bS > bN /2, therefore, bS > yn, which is a contradiction, hence, yn < bS . (ii) yn < bN /2, i.e., the market share of the highestquality good extends beyond bN /2. Suppose that aN ≤ yn < bS . From Lemma 1 we know that (sN + ss ) dy + sN dy > (SN+ SS ) yn , hence, (sN bN + ssbs) / [2(sN + sS )] > yn , but by assumption bN > bs , so that yn < bN /2. If yn < aN , which is a contradiction, and by assumption aN < bN /2, then yn < bN /2 as required. (iii) yn1 < aN, i.e., the market share of mediumquality good qn1 extends beyond aN . Suppose instead yn1 ≥ aN that, from Lemma 1 we know for qn1 that yn_1 < yn /2, and that yn /2 < bN /4 < aN , so yn1 < aN , which is a contradiction, hence yn1 < aN . (iv) yn2 ≤ as, i.e., the market share of the minimumquality good qn2 extends below aS . Suppose instead that yn2 ≥ aS , from Lemma 1 we know for qn2 that yn2 < yn1 /2, and that yn1 /2 < aN /2 < aS , so yn2 < aS which is a contradiction, hence yn2 ≤ aS.
This is an interesting result in that even though North and South mutually recognize each other's minimumquality standard, price competition ensures that while the lowestquality good is driven from the integrated market the poorest consumers in the South, are now able to purchase the minimumquality good produced in the North. Of course, North and South could harmonize their minimum quality standard to that of the North, in which case, the South's minimumquality standard would be driven from the market by flat. However, there will still be intensified price competition between the three remaining goods.17
Imperfect information
Based on the credence good labeling scenarios laid out in the NorthNorth case, we can also draw some conclusions about the gains from economic integration where North and South either harmonize or mutually recognize their labeling regulations.
(i) Cases XL and MEC
In the XL case, Proposition 4 still holds, other than the upper end of the income distribution in the integrated economy is now bN not b, i.e., with no labeling, there are no gains from integration, market structure being characterized by a monopoly selling the minimumquality good in North and South. For the NorthSouth MEC case, the only difference from the perfect information case of Proposition 10 is that if mandatory continuous labeling is not too costly, the market will support three goods with qualities u3 > u2> , the firms supplying the high and mediumquality goods earning lower profits due to labeling costs, while consumer welfare remains the same.
(ii) Case MEDh
In the case of NorthSouth MEDh, we assume that in the integrated economy, one harmonized labeling standard is set, ug = uk, where k = N, S.
The impact of the labeling standard depends on its location relative to what would be optimal for the firms choosing qualifies , and :
if ug ≤ , this will force the highestquality good from the market, and it may force the mediumquality good out of the market as well if ug is set too low, thereby intensifying price competition too much between the medium and minimumquality goods. This bodes well for consumers who purchase the lowquality good, who now pay a lower price. Consumers of the mediumquality good may also pay a lower price, but they would rather have "higher" mediumquality and pay a higher price, while consumers of the highquality good clearly suffer a loss of welfare. In aggregate, consumers lose due to the lowering of quality;
if ≤ ug ≤ either the medium or the highestquality good will be driven from the market, depending on the location of the harmonized public standard between the mediumquality and the highquality goods. Essentially, if the standard is set not too far from the optimal level of quality, ()the highquality (mediumquality) good will be driven from the market, as only one good can survive at that level of quality. This will of course diminish competition between the remaining goods, because whether the medium or highquality good survives, it is the case that ≤ ≤ ug ≤ . Consumers of the lowquality good will lose from paying a higher price, while consumers of either medium or highquality goods will lose if their preferred good is forced out of the market. Consumers of the mediumquality good will benefit if the standard results in a quality increase, while consumers of the highquality good lose if the standard results in a quality decrease;
if ≤ ug, the mediumquality good will be forced from the market as it will be unprofitable for two firms to compete at a standard set higher than that preferred by the highquality firm. This harms consumers of the lowquality good, who now pay higher prices, and also consumers of the mediumquality good who are unable to purchase the highquality good. Consumers of the highquality good benefit as they value the quality increase more than they are harmed by the price increase.
These results are summarized in the following propositions:
Proposition 11: For NorthSouth MEDh:
(i) if ug ≤ and ug [ γ (I d), + δ (I d)], (a) the highestquality good will be driven from the market, and (b) the integrated market will only support two qualities, u2 , :
(ii) if ≤ ug ≤ and either, ug [ γ (I d), + δ (I d)] or ug [ γ (I d), + δ (I d )], (a) either the highestquality or mediumquality good will be driven from the market, and (b) the integrated market will only support two qualities, u2 , or u3 , ;
(iii) if ≤ ug and [ γ(I d), + δ (I d)], (a) the mediumquality good will be driven from the market, and (b) the integrated market will only support two qualities.
Otherwise MEDh results are identical to XL results. In all cases, γ(.) and δ(.) are nonnegative, decreasing functions of I d and .
Proposition 12: For NorthSouth MEDh
(i) if ug [ γ (I d), + δ (I d)], and ug ≤ , (a) aggregate consumer welfare decreases, (b) the welfare of consumers purchasing the minimumquality good decreases, while the welfare of those who purchase the mediumquality good falls, (c) the welfare of those who prefer the highquality good falls, and (d) the profits of the medium and lowquality firms decrease;
(ii) if either [ γ(I d), + δ (I d)] or ug [ γ(I d), + δ (I d)], and ≤ ug ≤ , (a) aggregate consumer welfare decreases, (b) the welfare of consumers purchasing the minimumquality good decreases, while the welfare of consumers who prefer the prefer the medium or highquality good decreases if they cannot purchase that good,(c) the welfare of those who purchase the mediumquality (highquality) good increases (falls) if the standard raises (lowers) quality, and (d) the profits of either the medium or the highquality firm increase, and the profits of the lowquality firm increase;
(iii) if ug [ γ (I d), + δ (I d)], and ≤ ug (a) aggregate consumer welfare decreases, (b) the welfare of consumers purchasing the minimum and mediumquality goods decreases, (c) the welfare of consumers purchasing the highquality good increases, and (d) the profits of the lowquality firm increase while those of the highquality firm fall.
(iii) Case MEDmr
In the case of mandatory, exclusive, discrete labeling with mutual recognition, we assume that in the integrated market, firms claiming higher than minimumquality have to implement either one of two discrete standards, uS or uN, for which the following proposition can be stated:
Proposition 13: For NorthNorth MEDmr:
(i) if and one of uS or uN [ γ(I d), + δ (I d)], and one of uS or uN [ γ(I d), + δ(I d)], the integrated market will support three qualities, where γ(.) and δ(.) are nonnegative, decreasing functions of I d and ;
(ii) if uS uN, MEDmr results are identical to MEDh results.
Otherwise MEDmr results are identical to XL results.
Proposition 13 outlines intervals in which the discrete labeling standards must fall in order for three qualities to be produced, the intervals shrinking as labeling costs rise, assuming uS ≠ uN. In terms of welfare effects, compared to the perfect information case, the firms supplying the high and mediumquality goods will earn lower profits, while consumer welfare remains the same. If, however, uS uN, and the labeling standards fall in one of the required intervals, the standard chosen will be that closest to the firmpreferred standard, either the medium or highquality good being offered, the welfare effects being the same as those outlined in Proposition 12. Finally, if both countries set a labeling standard outside the required intervals, firms wishing to sell above the minimumquality earn negative profits and will not enter the integrated market. As a result, only one firm enters and the market collapses to the monopoly outcome of case XL.
The key to the MEDmr case is that, because the firms choosing medium and highquality have two standards to choose from through the principle of mutual recognition, there is a higher probability that the gains from economic integration will be realized due to one of the standards being close to the medium and highquality firms' preferred standards.
Summary and Conclusions
In an earlier paper, we used a model of vertical product differentiation to analyze the efficiency and distributional implications of different approaches to labeling of credence goods in an economy under autarky (Roe and Sheldon, 2007). In this paper we extend the institutional setting by allowing for the integration of two economies where they agree to either harmonize or mutually recognize their credence good labeling regulations. With perfect information about qualities, we show that integration of two NorthNorth economies with identical income distributions, results in increased quality in equilibrium, while NorthSouth integration, which allows more goods to be viable in equilibrium, results in lower prices in equilibrium.
The propositions derived in the paper hold some important implications for labeling programs when economies integrate in the presence of credence goods. In our framework, the market structure is fundamentally altered in that two firms in the NorthNorth case and three firms in the NorthSouth case, rather than one, may enter when labeling is present and labeling costs are not too high. Consumers are given greater choice, and competition between firms helps push down prices and, hence, improves welfare.
When there is mandated, exclusive, and continuous labeling (MEC), the labeling regime in both the NorthNorth and NorthSouth cases delivers the same prices and qualities as would be delivered under perfect information in the integrated economy, i.e., the labeling regime is nondistorting, and the gains from economic integration are realized. In contrast, if the authorities use harmonized mandated, exclusive, and discrete labeling (MEDh), quality distortion may occur. Quality distortion has distributional implications, with lower standards preferred by lower income consumers and higher standards preferred by higher income consumers and by the lowquality producing firm. Hence, discrete labeling offers the authorities a means to influence the distribution of welfare in the integrated economy.
If the authorities have exclusive authority to certify and label a quality dimension, in the NorthNorth case, they risk pushing out the highquality good if the harmonized standard is too high or too low to yield positive profits for the highquality producing firm, while in the NorthSouth case, they run the risk of pushing either one of or both the medium and highquality goods out of the market.
In our earlier paper, we discussed in some detail the extent to which the assumptions of the underlying model affect the results of using this type of vertical differentiation model (Roe and Sheldon, 2007). We do not repeat that discussion here, instead we conclude by noting that the results of the current paper are sensitive to the assumption that on integration, economies harmonize their labeling regulations, when in fact they may mutually recognize each other's existing labeling regimes.
This does not matter in the case of mandatory, exclusive continuous labeling (MEC), because there is no divergence between countries' standards, i.e., no standards are set as labeling is continuous. However, mutual recognition of standards can affect the results in the case of mandatory, exclusive, discrete labeling (MEDmr). Specifically, in the NorthNorth case, if one standard is closer than the other to what is optimal for the highquality firm, mutual recognition of standards may ensure that a highquality firm enters the integrated market. Likewise, in the NorthSouth case, if one standard is closer to what is optimal for the mediumquality firm, and one is closer to what is optimal for the highquality firm, then under mutual recognition, either one or both firms will have an incentive to enter the integrated market. However, if there is little divergence between the standards of the integrating countries, then the previous results for the MEDh case will hold even with mutual recognition, i.e., the highquality firm may be driven out in the NorthNorth case if the standards are set too low, while either the mediumquality and/or the highquality firm may be driven out in the NorthSouth case if the standards are either set too low, or too high. The key point is that compared to harmonization of standards, mutual recognition by countries of each other's labeling regimes actually increases the probability that the benefits of integration will be achieved.
Finally, it should be noted that if private certification is permitted, there are two implications for the results. First, with mandatory continuous labeling (MEC), the welfare gains from integration are unaffected as there will be no incentive for firms to hire a private certifier. This follows from the fact that firms are already able to communicate their desired quality level perfectly via the mandated continuous label. Second, with mandatory discrete labeling (MED), the welfare gains from integration may be greater if regulators permit either private certification of a standard different to the harmonized standard, or private certification where there is minimal divergence between mutually recognized standards private certification lowers the risk that higherquality goods are pushed out of the market if standard(s) are set too low or too high.18
References
Boom, Anette (1995). "Asymmetric international minimum quality standards and vertical differentiation", Journal of Industrial Economics, 43(1), 101119. [ Links ]
Darby, Michael R. and Edi Karni (1973). "Free competition and the optimal amount of fraud", Journal of Law and Economics, 16(1), 6788. [ Links ]
Gabszewicz, John J. and Jacques F. Thisse (1979). "Price competition, quality and income disparities", Journal of Economic Theory, 20(3), 340359. [ Links ]
Gabszewicz, John J. and Jacques F. Thisse (1980). "Entry (and Exit) in a Differentiated Industry." Journal of Economic Theory 22(2): 327338. [ Links ]
Gabszewicz, John J., Avner Shaked, John. Sutton, and Jacques F. Thisse (1981). "International trade in differentiated products", International Economic Review 22(3), 527534. [ Links ]
Leebron, David W. (1996). "Lying down with Procrustes: An analysis of harmonization claims", in Jagdish N. Bhagwati and Robert E. Hudec eds., Fair Trade and Harmonization: Prerequisites for Free Trade? Massachusetts, MA: MIT Press, pp.41117. [ Links ]
Lutz, Stefan H. (2000). "Trade effects of minimum quality standards with and without deterred entry", Journal of Economic Integration, 15(2), 314344. [ Links ]
McCluskey, Jill J. (2000). "A game theoretic approach to organic foods: An analysis of asymmetric information and policy", Agricultural and Resource Economics Review, 29(1), 19. [ Links ]
Roe, Brian and Ian Sheldon (2007). "Credence good labeling: The efficiency and distributional implications of several policy approaches", American Journal of Agricultural Economics, 89(4), 10201033. [ Links ]
Ronnen, Uri (1991). "Minimum quality standards, fixed costs, and competition", The Rand Journal of Economics, 22(4), 490504. [ Links ]
Scarpa, Carlo (1998). "Minimum quality standards with more than two firms", International Journal of Industrial Organization, 16(5), 665676. [ Links ]
Shaked, Avner and John Sutton (1982). "Relaxing price competition through product differentiation", Review of Economic Studies 49(1), 313. [ Links ]
Shaked, Avner and John Sutton (1983). "Natural oligopolies", Econometrica, 51(5), 14691483. [ Links ]
Shaked, Avner and John Sutton (1984). "Natural oligopolies and international trade", in Henryk Kierzkowski ed., Monopolistic Competition and International Trade. Oxford: Oxford University Press, pp.3449. [ Links ]
Sutton, John (1991). Sunk Costs and Market Structure. Cambridge, MA: MIT Press. [ Links ]
Tirole, Jean (1988). The Theory of Industrial Organization. Cambridge, MA: MIT Press. [ Links ]
1 The authors would like to thank seminar participants at the University of Guadalajara, the University of Nebraska, and Ohio State University for helpful comments on an earlier version of this paper.
2 This classification is designed to capture integration of similar advanced economies (NorthNorth), and integration of advanced and developing economies (NorthSouth).
3 By assuming rules on credence good labeling are set exogenously, we ignore the possibility of regulatory capture. Consequently, no explicit politicaleconomic framework is included in our analysis.
4 A separate literature already exists focusing specifically on minimumquality standards, e.g., Ronnen (1991), Boom (1995), Scarpa (1998), and Lutz (2000).
5 See Roe and Sheldon (2007) for a discussion of the multiplicative form of utility.
6 See Shaked and Sutton (1983) on relaxation of the assumption on the shape of the income distribution.
7 The assumption of zero variable production costs can also be relaxed without altering the main results of the paper.
8 Many goods can be characterized by a vertical quality that is dominated by fixed costs.
9 If the income distribution were broader such that the market was uncovered, increasing quality could draw more consumers into the market and may cause firm one's optimal quality to be interior.
10 More technically, this represents a Nash equilibrium only if the lowquality firm has no incentive to leapfrog the highquality firm, and, hence become the highquality provider given that the highquality firm has already chosen (16). Boom (1995) has shown that such an incentive never exists if (16) holds; hence, a unique Nash equilibrium exists.
11 See Roe and Sheldon (2007) for a detailed discussion of Proposition 1.
12 The assumption of perfect monitoring, while strict, allows the market to be converted from one of credence goods to one of search goods. If monitoring were noisy, deduction of equilibrium would require a repeated game structure as in McCluskey (2000). Because monitoring is assumed to be perfect, repeating the current game would not change the resulting equilibrium.
13 Monitoring a discrete standard is likely to be cheaper as it merely requires checking that processes meet or exceed a given threshold, i.e., going over a check list, while continuous labeling may require additional monitoring equipment to calibrate and report exact performance.
14 Allowing for variable costs of labeling would be similar to allowing for variable costs of production. As discussed in note 5, such an expansion of the model would not alter our fundamental results.
15 In the case of continuous labeling, the issue of harmonization versus mutual recognition simply does not arise.
16 The only circumstance under which multiple firms selling the lowquality good enter is when ε i = 0, i.e., a perfectly contestable market (Sutton, 1991).
17 Alternatively, if North and South harmonize to the minimum quality standard of the South, as long as the cost of labeling the higher minimum quality is not too high, the lower minimum quality good is still likely to be driven from the market.
18 This discussion draws on the results of our earlier paper, where we refer to the possibility of private certification as nonexclusive labeling (Roe and Sheldon, 2007).