The ﬁNewﬂEconomics of Trade Agreements: From Trade ...

The “New”Economics of Trade Agreements:From Trade Liberalization to Regulatory

Convergence?

Gene M. Grossman, Phillip McCalman and Robert W. Staiger

Princeton, University of Melbourne and Dartmouth

July 2019

Grossman, McCalman and Staiger . ( Princeton, University of Melbourne and Dartmouth )New Trade Agreements July 2019 1 / 31

Introduction

Trade negotiations have been remarkably successful at reducing thetraditional barriers to trade in the post-war period

With this success, the trade community has shifted its attention toNTBs that leave world markets still far from integrated

many NTBs arise from differences in domestic regulations or whatSykes (1999a,1999b) has termed ‘regulatory heterogeneity’TBT and SPS Agreements, TTIP negotiations in response

Lamy (2015, 2016) highlights a particular form of internationalexternality that arises from regulatory heterogeneity

firms that must satisfy different regulations across markets may face asubstantial cost in the form of foregone economies of scale

Yet as Sykes argues, international differences in incomes, cultures andtastes generally justify at least some regulatory heterogeneityWhat is the appropriate trade-off in international trade agreementsbetween heterogeneous tastes across international borders and theextra costs imposed by disparate regulations?

Introduction

Trade negotiations have been remarkably successful at reducing thetraditional barriers to trade in the post-war periodWith this success, the trade community has shifted its attention toNTBs that leave world markets still far from integrated

Introduction

Design Modifications Due to Differences between Current U.S. and EU Safety Regulations

A number of vehicle modifications are necessary to allow a vehicle sold in Europe to also be sold in the United States and vice versa. These modifications include changes to componentry, vehicle subsystems, and the underlying design of the vehicles. CAR reviewed differences within the regulations, as well as interviewed several manufacturers to develop a list of vehicle modifications required due to differing U.S. and European regulations. A summary of the identified modifications is provided in Figure 1 and Table 2; the table defines the numbers shown in the figure. In addition to modifications required for regulatory purposes, some of the examples given are due to non-regulatory test standards. A more detailed description of these modifications is located in Appendix A. The financial burden of these modifications varies by application, and a component-level change might or might not carry the same burden as a design-level change. Figure 1. Areas of vehicles requiring component, sub-system, and design-level modifications as a result of differing U.S. and European safety regulations

Introduction

Yet as Sykes argues, international differences in incomes, cultures andtastes generally justify at least some regulatory heterogeneity

What is the appropriate trade-off in international trade agreementsbetween heterogeneous tastes across international borders and theextra costs imposed by disparate regulations?

Introduction

To explore this question, we extend the Venables (1987) model

trade in horizontally differentiated products under monopolisticcompetition in the presence of a competitively produced ‘outside’good

features a “delocation”motive for unilateral policies

We introduce a second dimension of differentiation along which theresidents of different countries have different ideals

an individual pays a utility cost from consuming any good that differsfrom her ideal along this dimension

the loss of utility enters as a ‘demand shifter’in a familiar CESformulation, possibly with an externality

We allow firms in the differentiated product sector to tailor theirbrands to the alternative destination markets

if firms supply different versions of their brands, they bear a fixed costof design adaptation

Introduction

Findings

Without some institutional constraints on standards setting, theincentives to distort standards so as to favor local firms are extreme

can be tempered but not eliminated in an OTA that adopts a “smart”approach to setting effi cient net trade taxes

In the absence of consumption externalities, national treatment canhelp, but mutual recognition dominates as an institutional rule underwhich OTAs can achieve the first bestIn the presence of consumption externalities —even ones that do notcross borders —neither national treatment nor mutual recognitionallows countries to achieve in an OTA what they can in an NTARelated LiteratureVenables (1987), Krugman (1980), Ossa (2011), Bagwell and Staiger(2001, 2012, 2015), Bagwell and Lee (2018a,b), Campolmi et al(2018); Costinot (2008), Sykes (1999a,b, 2000), Staiger and Sykes(2011), Antras and Staiger (2012a,b), Podhorsky (2015)

Introduction

FindingsWithout some institutional constraints on standards setting, theincentives to distort standards so as to favor local firms are extreme

Introduction

In the absence of consumption externalities, national treatment canhelp, but mutual recognition dominates as an institutional rule underwhich OTAs can achieve the first best

In the presence of consumption externalities —even ones that do notcross borders —neither national treatment nor mutual recognitionallows countries to achieve in an OTA what they can in an NTARelated LiteratureVenables (1987), Krugman (1980), Ossa (2011), Bagwell and Staiger(2001, 2012, 2015), Bagwell and Lee (2018a,b), Campolmi et al(2018); Costinot (2008), Sykes (1999a,b, 2000), Staiger and Sykes(2011), Antras and Staiger (2012a,b), Podhorsky (2015)

Introduction

In the absence of consumption externalities, national treatment canhelp, but mutual recognition dominates as an institutional rule underwhich OTAs can achieve the first bestIn the presence of consumption externalities —even ones that do notcross borders —neither national treatment nor mutual recognitionallows countries to achieve in an OTA what they can in an NTA

Related LiteratureVenables (1987), Krugman (1980), Ossa (2011), Bagwell and Staiger(2001, 2012, 2015), Bagwell and Lee (2018a,b), Campolmi et al(2018); Costinot (2008), Sykes (1999a,b, 2000), Staiger and Sykes(2011), Antras and Staiger (2012a,b), Podhorsky (2015)

Introduction

Demand (absent externalities)

In the absence of externalities, the bundle of differentiated goods thatcomprise C JD in country J ∈ H,F is defined by

C JD =

∑i∈ΘJ

(aJi − aJ

)2] (cJi)β 1

with aJi ∈ [0, 1] and 1 > aH > aF > 0

With PJ the appropriate industry-level price index for C JD and IJ

per-capita disposable income in J,

V(PJ , I J

)= I J − logPJ ; and C JD =

NJ citizens in country J

In the absence of externalities, the bundle of differentiated goods thatcomprise C JD in country J ∈ H,F is defined by

C JD =

∑i∈ΘJ

(aJi − aJ

)2] (cJi)β 1

with aJi ∈ [0, 1] and 1 > aH > aF > 0

With PJ the appropriate industry-level price index for C JD and IJ

per-capita disposable income in J,

V(PJ , I J

)= I J − logPJ ; and C JD =

NJ citizens in country J

⇒ Per capita demand for brand i in country J is

cJi =(AJi)σ (

pJi)−σ (

PJ)σ−1

where AJi ≡ A−(aJi − aJ

)2> 0, σ = 1/ (1− β) and

PJ ≡[

∑i∈ΘJ

(AJi)σ (

pJi)1−σ

]− 1σ−1

Supply

A constant marginal cost of production λ

Fixed costs of a firm selling brand i depend on design choices

K + κ(aHi − aFi

Trade costs are composed of iceberg costs, and ad valorem exporttaxes and import tariffs

ιJ = 1+ φ+ eJ + τJ

Supply

K + κ(aHi − aFi

ιJ = 1+ φ+ eJ + τJ

Supply

K + κ(aHi − aFi

ιJ = 1+ φ+ eJ + τJ

Supply

Each country may also subsidize/tax consumption of differentiatedproducts at an ad valorem (nondiscriminatory) rate sJ

The profit-maximizing factory-gate price for all firms, regardless oflocation, is

σ− 1λ

Hence the consumer price of a typical local brand in country J is

pJJ = q(1− sJ

), J = H,F

while the consumer price of an imported brand in country J is

pJJ=(1− sJ

)qιJ , J = H,F

Supply

σ− 1λ

pJJ = q(1− sJ

), J = H,F

pJJ=(1− sJ

)qιJ , J = H,F

Supply

σ− 1λ

pJJ = q(1− sJ

), J = H,F

pJJ=(1− sJ

)qιJ , J = H,F

Supply in Unregulated Markets

Suppose firms have free rein to design their products

How does a firm choose its product design for the goods it will sell onits local and export markets?

A firm producing brand i in country J earns profits of

πiJ = (q − λ)[NJcJiJ

(aJiJ)+ (1+ φ)N Jc JiJ

(aJiJ)]

−[K + κ

(aJiJ − aJiJ

)2]The unregulated firm maximizes profits by designing its offerings sothat aH > aHiJ > a

FiJ > a

(aJiJ)]

−[K + κ

(aJiJ − aJiJ

The unregulated firm maximizes profits by designing its offerings sothat aH > aHiJ > a

FiJ > a

(aJiJ)]

−[K + κ

(aJiJ − aJiJ

)2]The unregulated firm maximizes profits by designing its offerings sothat aH > aHiJ > a

FiJ > a

Equilibrium

Suppressing the dependence on prices and therefore trade taxes andconsumption subsidies, the zero-profit conditions are

NJcJJ(aJJ ,P

J(n, aJH , a

))+ (1+ φ)N Jc JJ

(aJJ ,P

J(n, aJH ,a

))=K + κ

(aJJ − aJJ

)2q − λ

, J = H,F

Solving the zero-profit conditions ⇒

nJ = nJ (a) , PJ = PJ (a) , J = H,F

Finally, for unregulated markets, we solve for the choices of aJJ and aJJ

by firms producing in J

where the firms take the number and composition of competitors, n,and the industry-level price indexes PH and PF , as given

Equilibrium

NJcJJ(aJJ ,P

J(n, aJH , a

))+ (1+ φ)N Jc JJ

(aJJ ,P

J(n, aJH ,a

))=K + κ

(aJJ − aJJ

)2q − λ

, J = H,F

nJ = nJ (a) , PJ = PJ (a) , J = H,F

Equilibrium

NJcJJ(aJJ ,P

J(n, aJH , a

))+ (1+ φ)N Jc JJ

(aJJ ,P

J(n, aJH ,a

))=K + κ

(aJJ − aJJ

)2q − λ

, J = H,F

nJ = nJ (a) , PJ = PJ (a) , J = H,F

Properties of the Unregulated Equilibrium

Lemma (1)Consider the unregulated equilibrium with the profit-maximizing choices ofcharacteristics. Beginning at this equilibrium, a small increase in anyproduct characteristic induces exit by home firms and entry by foreignfirms.

Properties of the Unregulated Equilibrium

Lemma (2)Consider the unregulated equilibrium with the profit-maximizing choices ofcharacteristics. Beginning at this equilibrium, a small change in anyproduct characteristic has no first-order effect on the home or foreign priceindex, i.e., dPH/daJJ ′ = 0 and dP

F /daJJ ′ = 0 for all J ∈ H,F andJ ′ ∈ H,F.

National Welfare Measures

For the representative consumer in country J, we have

V J = I J − logPJ with I J = LJ/NJ + RJ/NJ

where RJ is the aggregate tax revenues in country J

We can write aggregate welfare in country J as

ΩJ (a,p, ρ) ≡ NJV J = LJ + RJ (a,p, ρ)−NJ logPJ (a,p)

where ρ is a vector of world prices and a can be evaluated underregulation or in unregulated markets

National Welfare Measures

For the representative consumer in country J, we have

V J = I J − logPJ with I J = LJ/NJ + RJ/NJ

where RJ is the aggregate tax revenues in country J

We can write aggregate welfare in country J as

ΩJ (a,p, ρ) ≡ NJV J = LJ + RJ (a,p, ρ)−NJ logPJ (a,p)

where ρ is a vector of world prices and a can be evaluated underregulation or in unregulated markets

Global Welfare

As usual, world prices drop out of global welfare

also, absent externalities, total spending on differentiated goods equalsone: ⇒

Ω (a,p) ≡∑JLJ +∑

z J︷︸︸︷(τJ + e J

)MJ (a,p)

−∑JNJ logPJ (a,p)−∑

1− sJ

Trade Agreements

In an “old trade agreement” (that includes a subsidies agreement)

the two govs choose the net trade taxes, zH and zF , and the subsidies,sH and sF , to maximize Ω (a,p)the sovereign choices of standards might be unconstrained, or subjectto institutional rules such as national treatment or mutual recognition

Under a “new trade agreement,” the govs negotiate a set of productstandards along with net trade taxes and consumption subsidies

Trade Agreements

In an “old trade agreement” (that includes a subsidies agreement)

the two govs choose the net trade taxes, zH and zF , and the subsidies,sH and sF , to maximize Ω (a,p)the sovereign choices of standards might be unconstrained, or subjectto institutional rules such as national treatment or mutual recognition

Under a “new trade agreement,” the govs negotiate a set of productstandards along with net trade taxes and consumption subsidies

A New Trade Agreement when Externalities are Absent

An effi cient NTA maximizes Ω (a,p) with respect to z, s and a

Global effi ciency requires zH = zF = 0 and sH = sF = 1/σ

effi cient consumption subsidies offset the monopoly distortion

net trade taxes different from zero can only harm world welfare oncethe optimal consumption subsidies are in place

A New Trade Agreement when Externalities are Absent

An effi cient NTA maximizes Ω (a,p) with respect to z, s and a

Global effi ciency requires zH = zF = 0 and sH = sF = 1/σ

effi cient consumption subsidies offset the monopoly distortion

net trade taxes different from zero can only harm world welfare oncethe optimal consumption subsidies are in place

New Trade Agreement

With z = 0 and s = 1/σ in place, global welfare for any a is

Ω(a,pE

)≡∑

LJ −∑J

NJ logPJ(a,pE

)−∑

σ− 1

where pE are the effi cient prices implied by z = 0 and s = 1/σ.

To see how the globally-effi cient product characteristics aredetermined, we borrow Figure 1 from Venables (1987)

Our Figure 1 drawn with NH(PH)σ−1

and NF(PF)σ−1

on the axes

fixes a at the levels that would emerge without gov regulation and withz = 0 and s = 1/σ

New Trade Agreement

Ω(a,pE

)≡∑

LJ −∑J

NJ logPJ(a,pE

)−∑

σ− 1

and NF(PF)σ−1

on the axes

New Trade Agreement

Ω(a,pE

)≡∑

LJ −∑J

NJ logPJ(a,pE

)−∑

σ− 1

and NF(PF)σ−1

on the axes

Figure 1: Optimal NTA

distort consumersallocation of spending between domestic and imported di¤erentiated goods.

To see how the globally-e¢ cient product characteristics are determined, we borrow Figure 1

from Venables (1987). Our Figure 1 is drawn with NHPH1

and NFPF1

on the axes, and

our rendition xes the product characteristics at the levels that would emerge without government

regulation and with zH = zF = 0 and sH = sF = 1=. The downward-sloping line labelled H = 0

gives the combinations of NHPH1

and NFPF1

that are consistent with zero prots for

home rms, in the light of (12). It has a slope equal to

dhNF (PH)1

= (1 + )1AHHAFH

where the inequality follows from the fact that > 0 and that AHH > AFH at the prot-maximizing

choices, say ~aHH and ~aFH . Similarly, the downward-sloping line labelled

F = 0 gives the combinations

of NHPH1

and NFPF1

that are consistent with zero prots for foreign rms when their

two versions have characteristics ~aHF and ~aFF . This line has a slope equal to

dhNF (PH)1

(1 + )1

AHFAFF

Also depicted in the gure are combinations of NHPH1

and NFPF1

that imply

nH = 0 and nF = 0, respectively. These combinations are readily derived from the expres-

sions for PF and PH . As shown in the gure, the nH = 0 locus is a ray from the origin with

New Trade Agreement

Is the point Q in Figure 1 effi cient?

Figure 1: Optimal NTA

distort consumersallocation of spending between domestic and imported di¤erentiated goods.

To see how the globally-e¢ cient product characteristics are determined, we borrow Figure 1

from Venables (1987). Our Figure 1 is drawn with NHPH1

and NFPF1

on the axes, and

our rendition xes the product characteristics at the levels that would emerge without government

regulation and with zH = zF = 0 and sH = sF = 1=. The downward-sloping line labelled H = 0

gives the combinations of NHPH1

and NFPF1

that are consistent with zero prots for

home rms, in the light of (12). It has a slope equal to

dhNF (PH)1

= (1 + )1AHHAFH

where the inequality follows from the fact that > 0 and that AHH > AFH at the prot-maximizing

choices, say ~aHH and ~aFH . Similarly, the downward-sloping line labelled

F = 0 gives the combinations

of NHPH1

and NFPF1

that are consistent with zero prots for foreign rms when their

two versions have characteristics ~aHF and ~aFF . This line has a slope equal to

dhNF (PH)1

(1 + )1

AHFAFF

Also depicted in the gure are combinations of NHPH1

and NFPF1

that imply

nH = 0 and nF = 0, respectively. These combinations are readily derived from the expres-

sions for PF and PH . As shown in the gure, the nH = 0 locus is a ray from the origin with

New Trade Agreement

Proposition

Let a be the vector of product characteristics that result fromprofit-maximizing design choices in an unregulated equilibrium whenzH = zF = 0 and sH = sF = 1/σ. Then the maximum world welfare isachieved in a monopolistically-competitive equilibrium when zH = zF = 0,sH = sF = 1/σ, and a = a.

How could the globally effi cient outcome be achieved with an NTA?

First, the agreement must stipulate zero net trade taxes on all goods

prevents govs from using import tariffs for delocation and export taxesfor delocation and TOT manipulation

Second, as long as consumption subsidies abide by nationaltreatment, the agreement need not stipulate that s = 1/σ

in this model, consumption subsidies have no impact on the TOT andunder national treatment are unattractive as tools for delocation

New Trade Agreement

Proposition

First, the agreement must stipulate zero net trade taxes on all goodsprevents govs from using import tariffs for delocation and export taxesfor delocation and TOT manipulation

New Trade Agreement

Proposition

First, the agreement must stipulate zero net trade taxes on all goodsprevents govs from using import tariffs for delocation and export taxesfor delocation and TOT manipulation

[NF(PF)(sigma – 1)]

[NH(PH)(sigma – 1)]

New Trade Agreement

Finally, how might the NTA address product standards?

It could require the home and foreign govs to set product standards(aHH , a

)=(aHH , a

(aFH , a

)=(aFH , a

)respectively

Or it could require the home and foreign govs to permit the range ofproduct characteristics

[aHF , a

]and [aFF , a

FH ] respectively

or both govs could promise not to regulate product characteristics at all

⇒ An NTA designed in this way would maximize joint welfare

Could an OTA maximize joint welfare while being silent on issues ofproduct standards?

New Trade Agreement

)=(aHH , a

(aFH , a

)=(aFH , a

)respectively

[aHF , a

]and [aFF , a

FH ] respectively

New Trade Agreement

)=(aHH , a

(aFH , a

)=(aFH , a

)respectively

[aHF , a

]and [aFF , a

FH ] respectively

New Trade Agreement

)=(aHH , a

(aFH , a

)=(aFH , a

)respectively

[aHF , a

]and [aFF , a

FH ] respectively

An Old Trade Agreement without National Treatment

Suppose an OTA calls for free trade (τ = e = 0), and consumptionsubsidies (but not standards) that conform to national treatment

we find that the unilateral choices of consumption subsidies are choseneffi ciently to offset monopoly pricing (s = 1/σ)

What regulatory standards will each gov choose?

With τJ = eJ = 0 and sJ = 1/σ, the gov of each country J seekswith its choice of aJH and a

JF to maximize

ΩJ (a,p, ρ) = LJ −NJ logPJ (a,p)−NJ 1σ− 1

⇒ the objective of each gov is simply to minimize the local price index

JF to maximize

Would the home gov regulate its local firms?

No: any regulation that requires a discretely different productcharacteristic than the profit-maximizing choice would shift theπH = 0 line to the right at point Q, leading to a higher PH

Might the home gov regulate imports?

Yes: any regulation that requires a discretely different productcharacteristic than the profit-maximizing choice would shift theπF = 0 line to the right at point Q, resulting in a lower PH

Notice that home gains by regulating imports in either direction

How can we understand this welfare improvement?

Consider first a standard aHF that pushes foreign firms closer to aH

home consumers benefit directly from goods with aHF closer to aH

but nH falls and nF rises (Lemma 1) — anti-delocationdirect effect dominates costs of anti-delocation and PH must fall

Now consider a standard aHF that pushes foreign firms further from aH

home consumers loose directly from goods with aHF further from aH

but nH rises and nF falls (Lemma 1) —delocationnow benefits of delocation dominate direct effect and PH must fall

Either way, govs that are free to regulate products differentlyaccording to their source will apply pernicious standards to imports

⇒ the globally effi cient outcome cannot be achieved with a free-tradeagreement that is silent about regulation

Where does the process of non-cooperative regulation lead us?

For every pair of standards that applies to local products, each govhas a unilateral incentive to push its import standard to an extreme

until it reaches a boundary of the product space and can go no further

or one of the govs captures the entire world market for its local firms

PropositionSuppose ξ = 1, τ = e = 0 and s = 1/σ. Suppose governments are free tochoose any standards for local products and for imported products,without need for national treatment. Then, in the Nash equilibrium of thestandard-setting game, either (i) nJ = 0 for some J ∈ H,F, or (ii)aFH ∈ 0, 1 and aHF ∈ 0, 1. The equilibrium level of global welfare is lessthan that attained under an NTA.

A Smarter OTA without National Treatment

Suppose the OTA maintains zero net tariffs, but instead of settingτ = e = 0 as in an FTA, sets τH = τF = −eH = −eF ≡ τ > 0

equilibrium prices and quantities are unchanged, but home welfare now

ΩH = LH + τq(MH − EH

)−NH log

(PH)−NH 1

σ− 1

so trade-tax revenue considerations arise

Beginning from aHF = 1, easing standards increases PH , but MH rises

and EH falls, increasing home tax revenues

first effect is independent of τ; second effect rises linearly with τ

if slope of πH = 0 and πF = 0 are steeper and flatter than −1, globalwelfare higher under a smarter OTA with τ > 0 than under FTA

Still, a smarter OTA cannot get all the way to the first best (Figure 2)

)−NH log

(PH)−NH 1

σ− 1

)−NH log

(PH)−NH 1

σ− 1

)−NH log

(PH)−NH 1

σ− 1

)−NH log

(PH)−NH 1

σ− 1

Figure 2: Numbers of Firms as Function of aHF

foreign rms are delocated to the home-country market. So the incentive for the home country to

defect toward the left from the e¢ cient prot maximizing aHF is due to delocation. But moving

right from the prot maximizing level aHF , PH falls despite the fact that initially nF is rising and

nH is falling. So the incentive to defect toward the right from the e¢ cient prot maximizing aHF is

initially in the interval ((aHF ; aH1F ) not due to delocation; it is due instead to the direct impact

on PH of having imports adopt a characteristic that is a little closer to the Home ideal aH , and

this direct impact dominates the (anti-) delocation e¤ects here. Once we move into the interval

(aH1F ; aH2F ), both nH and nF are falling with further increases in aHF , so again the incentive for the

home country to keep raising aHF in this interval to lower PH is not due to delocation, but must

still be due to the domination of the direct impact on PH of having imports adopt a characteristic

that is a little closer to the Home ideal aH . In the interval (aH2F ; aH), we now have delocation and

the direct impact described above both helping to push PH lower. But for the interval (aH ; 1), the

direct e¤ect is now going the wrong way so it is the delocation e¤ect that dominates at this point

and keeps PH falling.

This illustrates why setting tari¤s in a way that perfectly o¤sets the PH -reducing incentives

of the home government with countervailing revenue incentives will not be possible, because the

PH -reducing incentives themselves are not tied monotonically to the trade volume e¤ects and

hence the potential trade tax revenue e¤ects of standards choices, and only reect trade volume

e¤ects in a consistent way as aHF approaches the extremes of 0 or 1. So while the judicious choice

of (e¢ cient) trade tax/subsidies can reduce the Nash distortions in standards from their extreme

levels, it cannot eliminate these distortions completely, an observation we formalize in Proposition

An FTA with National Treatment

Can national treatment applied to standards solve the prisoner’sdilemma that arises from the urge to delocate?

We return to an FTA, with τJ = eJ = 0 and sJ = 1/σ for J = H,F

We show that national treatment can mitigate the urge to delocatewith standards, but cannot achieve the first best

under national treatment imported and locally produced goods must beoffered the same options ⇒ each gov chooses a single standardfirst best requires four standards, national treatment delivers two

Proposition

Suppose ξ = 1, zH = zF = 0 and sH = sF = 1/σ. Suppose eachgovernment is free to choose any standard or set of standards as long asthey are offered to all firms irrespective of origin. Then, in the Nashequilibrium of the standard-setting game, the outcome is equivalent to onein which each government names a single standard, and it does notachieve the maximal level of global welfare that is attained by an NTA.

An FTA with Mutual Recognition

Can mutual recognition solve the prisoner’s dilemma that arises fromthe urge to delocate?

Under mutual recognition, each gov respects the legitimacy of theother country’s regulatory aims

therefore, any product that meets standards in an exporting country isconsidered acceptable for sale in the importing country

⇒ exporting firms have the choice of whether to meet the standards ofthe destination market or their local country

Suppose the gov of country J announces a1J or a2J

mutual recognition implies that country-J firms must produce a versionwith a1J or a

2J for local sales

but they can choose to meet any of the four legal standards for theirsales in country J

What standards will the govs set in a noncooperative equilibrium, ifsubject to an FTA with mutual recognition?

2J for local sales

We show: each gov will choose the product characteristics thatmaximize profits for its own representative national firm

But these are the standards that would emerge under a globallyeffi cient NTA

A caveat: under terms of the European Union treaty, a firm caninvoke mutual recognition in its export market only if it also suppliesa similar good to its local market

Proposition

Suppose ξ = 1, τH = τF = eH = eF = 0 and sH = sF = 1/σ. Supposethat each government is free to choose two or more standards for localsales and that firms can invoke mutual recognition for export sales of anyproduct that can legally be sold in its native market. Then, in the Nashequilibrium of the standard-setting game, each government will set two ormore standards and the outcome is the same as in the globally effi cientNTA.

A New Trade Agreement when Externalities are Present

What changes with ξ < 1?

The bundle of differentiated goods that comprise C JD in countryJ ∈ H,F is defined by

C JD =

∑i∈ΘJ

[A− ξ

(aJi − aJ

)2] (cJi)β− (1− ξ)

(aJi − aJ

)2 (cJiµ)β 1

with aJi ∈ [0, 1] and 1 > aH > aF > 0

ξ does not impact the total disutility that the representativeconsumer bears from local consumption of less-than-ideal varieties

ξ only impacts the composition between own and others’consumption

What changes with ξ < 1?

The bundle of differentiated goods that comprise C JD in countryJ ∈ H,F is defined by

C JD =

∑i∈ΘJ

[A− ξ

(aJi − aJ

)2] (cJi)β− (1− ξ)

(aJi − aJ

)2 (cJiµ)β 1

with aJi ∈ [0, 1] and 1 > aH > aF > 0

ξ does not impact the total disutility that the representativeconsumer bears from local consumption of less-than-ideal varieties

ξ only impacts the composition between own and others’consumption

⇒ The effi cient product characteristics, quantities of per-brandconsumption and numbers of home/foreign firms don’t change with ξ

rather, ξ determines the policies needed to achieve these outcomes

when ξ < 1, profit maximizing choices of standards are no long effi cient

∴ An FTA with mutual recognition cannot achieve first best whenξ < 1, because firms may not self-select into first-best standards

and even if firms would self-select as desired, govs could delocate bygranting favorable standards to their exporting firms while bearing noneof the externality generated cost

In the presence of consumption externalities —even ones that do notcross borders — the requirements for cooperation are more severe

Conclusion

We have studied the need for NTAs when goods are horizontallydifferentiated and ideal product attributes differ across countries

Without some institutional constraints on standards setting, theincentives to distort standards so as to favor local firms are extreme

In the absence of consumption externalities, national treatment canhelp, but mutual recognition dominates as an institutional rule underwhich OTAs can achieve the first best

In the presence of consumption externalities —even ones that do notcross borders —neither national treatment nor mutual recognitionallows countries to achieve in an OTA what they can in an NTA

The ﬁNewﬂEconomics of Trade Agreements: From Trade ...

Documents

Transcript of The ﬁNewﬂEconomics of Trade Agreements: From Trade ...