European Integration and GATT: The Impact on United States Agriculture

November 1998                                          SCSB# 390

TRADE, POLICY AND COMPETITION:
FORCES SHAPING AMERICAN AGRICULTURE PROCEEDINGS

Chapter 4
European Integration and GATT:
The Impact on United States Agriculture

Cemal Atici and P. Lynn Kennedy

Introduction

International trade negotiations in agriculture, such as the General Agreement on Tariffs and Trade (GATT), reflect the linkage between agricultural trade policies at the international level and domestic farm policies. Because the Common Agricultural Policy (CAP) of the European Union (EU) has encouraged agricultural production, surpluses have developed in recent years and the EU has become a major exporter of several commodities. U.S. agricultural exports have suffered a decline in both volume and value during the 1980s (Congressional Research Service, 1986). EU exports have taken a greater share of world agricultural trade, and U.S. agricultural exports to the EU have begun to decline as well (GATT, 1993). This has resulted in trade disputes in between these countries.

Because agricultural policies affect both domestic and international markets, it is essential to know the objectives and consequences of different policies in agriculture. In order to accomplish this, the research presented here examines the effects of liberalized trade resulting from the Uruguay round of GATT, combined with the integration of Turkey into the European Union. Particular emphasis will be placed on the impact of these policy changes on the United States in terms of trade in agricultural products. The empirical analysis will involve ten agricultural products that each play a significant role in the European Union, Turkey, and the United States in terms of production and consumption. They are beef and veal, dairy milk, corn, wheat, rice, soybeans, cotton, sugar, tobacco, and pork and poultry.

To accomplish these objectives, this study employs a partial equilibrium trade simulation model, Modele International Simplifie de Simulation (MISS) (Mahé et al., 1988). MISS is a partial equilibrium trade model that simulates in a comparative static framework the effects of various policy decisions. Once the model is initialized, simulations will be conducted that mirror the effects of the Uruguay Round agricultural agreement. In addition, consumer, producer, and government budget weights will be estimated. These weights, when combined with the net gains or losses to producers, consumers, and government, reflect the net gains or losses to the economies as viewed by policy-makers.

Theoretical Framework

The framework underlying this analysis is similar to that used by Johnson et al. (1993) and Kennedy et al. (1996). In this model, three large countries and the rest of the world (an aggregation of many small countries) produce, consume, and trade N commodities. Aggregate production, consumption, and trade in country i is described by vectors of supply, demand, and excess demand. Farms in country i produce some subset of the N traded commodities, taking prices, technology, and endowments as given in order to maximize profit. Aggregate supply is shown as:

(1) Y₁ ( P_f ; Z_f ), . . . , Y_N ( P_f ; Z_f ),

where P_f= ( P_f1 , . . . , P_fN ) is the vector of the producer prices of the N traded commodities, and Z_f is a vector of exogenous factors, such as prices of inputs and factor endowments. Demand for the N agricultural commodities is summarized by the vector of demand functions

(2) X ( P_c ; Z_c ) = ( X₁ ( P_c ; Z_c ) , . . . , X_N ( P_c ; Z_c ) ),

and the corresponding indirect utility function

(3) U ( P_c ; Z_c ),

where P_c = ( P_c1, . . . , P_cN ) is the vector of consumer prices for the N commodities, and Z_c is a vector of exogenous variables. Trade in the N commodities is summarized through excess demand

(4) E ( P_f, P_c ; Z_f, Z_c ) = X ( P_c ; Z_c ) - Y ( P_f ; Z_f ).

Governments intervene in their domestic markets through price instruments and demand/supply shift instruments. Price instruments, denoted A^p_fn for producers, f, and A^p_cn for consumers, c, of commodity n, directly or indirectly affect the farm and consumer prices of the N commodities. Letting P_wn be the world price of commodity n, the following domestic price functions are defined

(5) P_fn = P_fn ( A^p_fn, P_wn ), and P_cn = P_cn ( A^p_cn, P_wn ), "n=1,...,N.

Let the large countries be denoted as countries 1, 2, and 3 and let the rest of the world be country 4. The vector of excess demand functions in the rest of the world is E₄ (P_W; Z₄ ), where Z₄ is a vector of exogenous variables in the rest of the world. World markets are competitive by assumption; i.e., world prices adjust to clear world markets. Hence,

(6) E₁ ( P_f1 (A^p_f1, P_w ), P_c1 ( A^p_c1, P_w ), A^s_f1, A^s_c1 ; Z₁ ) +
E₂ ( P_f2 (A^p_f2, P_w ), P_c2 ( A^p_c2, P_w ), A^s_f2, A^s_c2 ; Z₂ ) +
E₃ ( P_f3 (A^p_f3, P_w ), P_c3 ( A^p_c3, P_w ), A^s_f3, A^s_c3 ; Z₃ ) + E₄ ( P_w ; Z₄ ) = 0.

where 0 is an N by one vector of zeros, and A^s_fn and A^s_cn are shift instruments that shift supply and demand functions (input subsidies, acreage reduction, etc.). Assume that world prices are implicitly defined as functions of the actions of the two governments. Then,

(7) P_w = P_w (A^p_f1,A^p_c1,A^s_f1, A^s_c1,A^p_f2, A^p_c2, A^s_f2, A^s_c2, A^p_f3, A^p_c3, A^s_f3, A^s_c3 ; Z₁, Z₂, Z₃, Z₄ ).

In choosing agricultural policies, governments consider the effects of their policies on the welfare of various groups, namely, producers, consumers, and taxpayers. Since agricultural policies often improve the welfare of certain groups at others' expense, governments must weigh the welfare gains of one group against the welfare losses of others. These trade-offs are represented by a political payoff function (PPF), a weighted, additive function of producer quasi-rents, indirect utility of consumers, and the costs of agricultural policies. In order to simplify these expressions, let -i represent the other countries, let A_i = ( A_fi, A_ci ) = ( A^p_fi, A^s_fi, A^p_ci, A^s_ci ), and suppress Z₁, Z₂, Z₃, and Z₄. Producers are aggregated by commodity group. The welfare of each producer group is the profit obtained from the production and sale of the commodity. Therefore, assuming differentiability, the welfare associated with the production of the n^th commodity is the line integral

(8) P_n ( P_n ) = ∫₀^Pn P_nY_n (p) dp

as commodity n is a net output or net input respectively. Let

(9) P ( P_f ; Z_f ) = ( P₁ ( P_f ; Z_f ) , . . . , P_N ( P_f ; Z_f ) ),

be the vector of quasi-rents as a function of the policies of governments, substitute for P_f by using equation 5, suppress Z_fas mentioned previously, and substitute for P_w by using equation 7 to obtain

(10) P ( A_i, A_-i ) = P_i ( A^p_fi, P_w ( A_i, A_-i ), A^s_fi ).

Similarly, substituting into equation 3 the domestic price function, indirect utility is obtained as a composite function of the actions of the governments:

(11) U ( A_i, A_-i ) = U_i ( P_ci ( A^p_ci, P_w ( A_i, A_-i ) ), A^s_ci ).

To define the government budget in the N agricultural commodities, let T denote a transpose. Then, aggregate consumer expenditures are P_c * X^T , producers receive P_f * Y^T, and excess demand is purchased in world markets at prices P_w for P_w * E^T. Hence, using equations 1 and 3 and substituting for E with equation 4, the budget is:

(12) B ( P_f, P_c, P_w ; Z ) = ( P_c - P_w ) * X^T (P_c ; Z_c ) ( P_f P_w ) * Y^T ( P_f ; Z_f ).

Making the proper substitutions for P_f, P_c, P_w, and Z as before, the budget of government i becomes a function of both government's agricultural policies:

(13) B (A_i, A_-i) = B_i (P_fi (A^p_fi, P_w (A_i, A_-i ), P_ci (A^p_ci, P_w (A_i, A_-i) ), P_w (A_i, A_-i), A^s_fi, A_ci).

Finally normalizing on the budget and using equations 10, 11, and 12, the PPF is:

(14) V_i ( A_i, A_-i ) = P ( A_i, A_-i ) * l_fi + U ( A_i, A_-i ) * l_ci + B ( A_i, A_-i ),

where l_fi is an N by one, strictly positive vector and l_ci is a positive scalar. The (l_fi, l_ci) are the political weights of the respective commodity groups and the aggregate consumer group in country i.

Equation 14 explicitly links the policies of the governments with their objectives. What is lacking is a means for the governments to choose agricultural policies. More precisely, how is the game between the governments resolved, and how are production, consumption, and domestic and world prices subsequently determined? The following describes an equilibrium in which a government, given the policies of the other countries, chooses policy to maximize its PPF. Formally, a best response correspondence is defined for each government. Then, the equilibrium is defined using the best response correspondence. For a given A_-i, government i chooses A_i, a best response to A_-i, such that

(15) V_i ( A_i, A_-i ) ≥ V_i ( A_i, A_-i )" A_i e A_i.

where A_i is the set of actions (policies) available to government i. Thus, every A_-i in A_-i has a set of actions in A_i that satisfy equation 15. The set defines the best response correspondence of A_-i. A pair of actions ( A_i, A_-i ) is an equilibrium if it satisfies equation 15 for all i; that is, A_i* is a best response to A_-i* for all i. Differentiating equation 14 with respect to A_fi and A_ci, the first order necessary conditions for a maximum are,

∂V_i
——
∂A_fi
=
∂V_i
  ——
∂A_fi P_i
  ——
∂A_fi

∂P_i
  ——
∂A_fi ∂U_i
  ——
∂A_fi
.
∂U_i
  ——
∂A_fi
l_fi

+

l_fi ∂B_i
  ——
∂A_fi
=
∂B_i
  ——
∂A_fi
0

0

(16)

For a given A_-i, if V_i is concave in A_i, then any A_i* that solves equation 16 maximizes V_i, so it is a best response to A_-i. Thus equation 15 implicitly defines the best response correspondence (of interior points) A_i* ( A_-i). A_i* ( A_-i ) is single valued, a function, if and only if V_i is strictly concave in A_i for all values of A_-i. ( A_i, A_-i ) is a Nash equilibrium if

∂V_i
——
∂A_fi

∂V_i
  ——
∂A_fi

=

| ( A_i, A_-i )

0

0

(17)

Suppose governments negotiate to improve their positions relative to the one-period equilibrium that they currently pursue. If governments are sovereign and rational as described above, then no treaty will be signed or complied with that does not make both governments at least as well off as prior to the agreement. Furthermore, a necessary condition for a treaty to be signed and complied with would be the existence of actions ( A_i+, A_-i+ ) such that

(18) V_i ( A_i⁺, A_-i⁺ ) &ge: V_i ( A_i, A_-i ) for all I.

The set of actions that satisfy equation 18 are called the treaty action space and the elements of this space treaty actions.

Empirical Analysis

The base year for the empirical analysis is 1992. Ten commodity groups are distinguished: beef and veal; dairy; corn; wheat; rice; soybeans; cotton; sugar; tobacco; and pork and poultry. To initialize the model, protection ratios are calculated for producers and consumers in the United States, EU, and Turkey for the base year 1992. These protection ratios, combined with production and consumption levels, are used as a base from which all simulations will be conducted. Nominal protection ratios were calculated by the ratio of domestic price to border price. Prices have been calculated in terms of commodity values for all products. Nominal protection ratios for producers and consumers can be seen in Table 4.1. A description of the quantity and price data, exchange rates, and elasticities used to initialize the model can be found in Atici (1996).

The PPF weights for the United States and EU are derived through the evaluation of incremental changes in the observed policies from their base year levels. These changes are then used as approximations of the partial derivatives in equation (16). When equation (16) is solved for l_fi and l_ci , the PPF weights are obtained. These approximated weights are normalized such that the budget weight is one. They are presented in Table 4.2.

To find a Nash Equilibrium for the countries involved, a game theoretical framework is used. The normal-form representation of a game is specified by the following: the players in the game, the actions available to each player, and the payoffs corresponding with each action combination. In this case there are three players in the game, the United States (US), EU, and Turkey (TUR). Let A_k denote the set of actions available to player k, for k=US, EU, TUR. Let (A_US, A_EU, A_TUR) denote combination of actions, and let P_k denote player k's payoff function where P_k(A_US, A_EU, A_TUR) is player k's payoff resulting from actions (A_US, A_EU, A_TUR). The normal-form representation of a three player game specifies the player's action spaces ( A₁, A₂, A₃ ) and their payoff functions ( P₁, P₂, P₃ ). This game is denoted by G = { A₁, A₂, A₃; P₁, P₂, P₃ }.

In the normal-form game, G = { A₁, A₂, A₃; P₁, P₂, P₃ } let A_k1 and A_k2 be feasible strategies for player k; i.e., they are members of A_k. Action A_k1 is strictly dominated by A_k2, if for all combinations of actions available to the other players, k's payoff from playing A_k1 is strictly less than k's payoff from playing A_k2, such that P_k ( A_k1, A_-k ) < P_k ( A_k2, A_-k ) for all A_-ke A_-k. Rational players will not play strictly dominated strategies, a concept which is useful in finding solutions to bimatrix games. If a unique solution to a three-player, normal-form, noncooperative game between the United States, EU, and Turkey is to be found, it must be self-enforcing. Each player's predicted action must be that player's best response to the predicted action of the other player. This is the concept of Nash equilibrium.

In the three player normal-form game G = { A₁, A₂, A₃; P₁, P₂, P₃ } the actions ( A₁, A₂, A₃* ) are a Nash equilibrium if, for each player k = 1, 2, and 3, A_k* is player k's best response to the actions specified for the other player's -k, such that:

P_k ( A_k, A_-k ) P_k ( A_k, A_-k* ) for all A_k e A_k.

Game Simulations

This analysis incorporates game theory to identify optimal strategies for the countries involved. The United States and EU choose among four strategies: status-quo (ST); base year reductions until the final year according to the Uruguay Round of GATT (GT); 50 percent reduction from their base year protections (50); and a 100 percent reduction from their base year protections (FT). Turkey chooses between two strategies: the application of its agreed upon Uruguay Round protection reductions without joining the European Union ( GT_TUR ); and joining the European Union and setting agricultural protection levels equal to those of the EU ( EU_TUR )

Each country k chooses some action A_k e A_k in order to maximize its political payoff function given the action choices of the other country. The games are presented in the following way. If the players' action spaces are specified as A_US, A_EU, and A_TUR and their payoff functions as P_US, P_EU, and P_TUR, then this game is denoted by G = { A_US, A_EU, A_TUR ; P_US, P_EU, P_TUR }. This simulation will determine the Nash equilibrium between the United States and EU when their actions are ST, GT, 50, and FT, while the actions available to Turkey are GT and EU. Thus, the action space is A_k = { ST_k, GT_k, 50_k, FT_k } for k = US, EU and A_k = { GT_k, EU_k } for k = TUR.

Two separate games will be analyzed. The first incorporates the weights approximated previously. The second game utilizes Political Preference Functions with weights of one, to reflect the outcome if policy-makers view all groups equally. It is hypothesized that the first game will better reflect the process occurring between the three countries.

Results

As can be seen from Table 4.3, both the United States and the EU have strictly dominant strategies by choosing the fifty percent reduction from their protection levels. As a result the unique Nash equilibrium occurs at { 50_US, 50_EU, GT_TUR }. At this action, PPF values are 749, 725, and -88 for the United States, EU, and Turkey respectively.

The Nash equilibrium is identified in the following manner. Suppose Turkey chooses GT_TUR. When the United States chooses SQ_US, the best response for the EU is to choose action 50_EU since the EU's PPF is higher than that for each of its other available actions. It is obvious that the EU is better off choosing the action 50_EU. Regardless of the action chosen by the United States, the EU, as a rational agent, will choose 50_EU. On the other hand, regardless of the strategy chosen by the EU, the United States will respond by choosing action 50_US since its PPF is highest at this point. Thus, the United States and EU have strictly dominant strategies of 50_US and 50_EU, respectively. The same result occurs when Turkey chooses EU_TUR. Thus, since 50_US and 50_EU are strictly dominant strategies, the iterative elimination of strictly dominated strategies narrows the game solutions to Turkey's choice between GT_TUR and EU_TUR at { 50_US, 50_EU }. As a rational agent, Turkey compares its payoff from GT_TUR (-88) to that from EU_TUR (-389) and chooses GT_TUR in order to maximize its PPF. Thus, the Nash equilibrium solution to this game occurs at { 50_US, 50_EU, GT_TUR }.

Now consider the scenario in which all weights are equal to one. The PPF results for this scenario can be seen in Table 4.4. As can be seen from these results, the unique Nash equilibrium occurs at free trade in this { FT_US, FT_EU, EU_TUR }. It must be remembered that, since Turkey chooses to join the European Union in this game, Turkey effectively chooses free trade since the EU's optimal strategy involves the adoption of free trade. This outcome is consistent with theory in terms of gains from trade.

Other Indicators

The adoption of policies consistent with the non-cooperative game solution presented in Table 4.3, { 50_US, 50_EU, GT_TUR }, will have consequences for agricultural producers and consumers in each of the three countries. Not surprisingly, producers with high protection levels, relative to those in other countries, during the base period will be worse off due to the resulting customs union protection levels, while those with relatively low initial protection levels will benefit. This section will provide selected indicators which result from the formation of the customs union. These include the following: changes in domestic producer and world prices; changes in production and consumption; changes in self-sufficiency; and changes in welfare.

Changes in domestic prices resulting from the customs union are presented in Table 4.5. This solution, when viewed from a price perspective, will hurt all United States and European Union producers, with the exception of the U.S. soybean sector. Conversely, all producer prices in Turkey increase, with the exception of corn, rice, and pork and poultry. This solution results in across-the-board increases in world prices, with the exception of pork and poultry.

The changes in production and consumption, presented in Table 4.6, are consistent with those expected as a result of the price changes, with a few exceptions. In particular, a decrease in the U.S. wheat price results in an increase in wheat production and a decrease in wheat consumption. This instance is likely due to the cross-price elasticity effects, perhaps with respect to corn. Another notable exception occurs in the case of pork and poultry production in the EU. A decrease in the producer price results in an increase in pork and poultry production. This corresponds with significant reductions in feed prices.

Self-sufficiency ratios, domestic production divided by domestic consumption, are presented in Table 4.7. The United States experiences self-sufficiency decreases in beef and veal, dairy, rice, cotton, sugar, tobacco, and pork and poultry, with increases in corn, wheat, and soybeans. In the EU, self-sufficiency decreases in all products, with the exception of pork and poultry. In Turkey, a decline in self-sufficiency is experienced in rice, soybeans, and pork and poultry, while increases will occur in beef and veal, dairy milk, wheat, cotton, sugar. Tobacco does not change.

Changes in producer welfare, as presented in Table 4.8, are also consistent with those expected based on the change in prices. As a result, the United States and European Union consumers gain from this solution, while consumers in Turkey are worse off. Viewing this outcome from the producers point of view, Turkish producers gain from the outcome while U.S. and EU producers are losers. As a result of the reduction in their protection levels, the government budget sector is better off in the United States and EU; i.e., these countries experience budget savings. At the same time Turkey suffers a budget loss. It must be remembered that in order to determine the game solution in Table 4.3, these gains and losses are weighted by the governments according to the political preference function weights presented in Table 4.2.

Conclusions

The results of this analysis suggest that it is in the best interest of Turkey, from an agricultural standpoint, to adopt agreements made in the Uruguay Round of GATT as a developing country rather than joining the EU and applying EU protection. On the other hand, the Turkish agricultural sector can benefit by joining the EU and convincing the EU to adopt free trade. Results show that producer surpluses in the United States and EU will decrease due to the fact that these countries decrease their protection levels. Since these decreases in protection increase world prices, the consumer surplus in Turkey will decrease as well. The Uruguay Round of GATT has eliminated quotas in many products by replacing them with tariff equivalents. However, these tariff equivalents were higher than those previously used in the United States, while the EU also used higher protection bases for protection reduction.

The Nash equilibrium occurred in the level of 50 percent reduction for both US and EU using the estimated weights. When all sectors are weighted equally, the Nash equilibrium occurred at free trade. Some important results are related to Turkey's integration and the Nash equilibrium. The results show that Turkey's integration into the EU is not in Turkey's best interest from an agricultural point of view. Turkey can apply its GATT commitments and be better off than with integration.

Among the significant products of this analysis is the result that the Nash equilibrium for the United States and EU occurred at 50 percent reduction from their base period protections with the estimated weights when Turkey is not in the EU. If Turkey were to join the EU, the Nash equilibrium would again be a 50 percent reduction from the base level protection for the United States and EU. When weights are equal one, the Nash equilibrium occurs at free trade levels for the United States and EU, with Turkey choosing to join the EU. This result is consistent with theory, and suggests that the countries involved weight the sectors analyzed in a manner consistent with the weights used in this analysis.

Free trade is not an optimal solution using the estimated weights. Both the United States and EU benefit from reducing protection levels to a point between the existing protection levels and free trade. Although free trade is not the optimal solution in agriculture, simulations show that there exists an optimum with freer trade. Future negotiations can identify areas of further protection reductions. This seems likely since the Nash equilibrium occurs at a protection reduction (50 percent) that is greater than GATT commitments for the United States and EU.

The results have several implications. Turkey's loss in agriculture from joining the EU may be compensated for by the potential gains in manufacturing and the service sector as well as the EU's funding for various sectors. Turkish policy-makers should evaluate these gains and losses, and decide whether it is in the country's best interest to integrate with the EU. Comparisons of this type can be made in a similar framework to include the manufacturing and services sectors. The framework and results of this study can contribute to future analyses that consider various welfare aspects of trade liberalization and economic integration.

References

Document Prepared by:
Leigh H. Stribling, lstribli@acesag.auburn.edu
Alabama Agricultural Experiment Station
Auburn University

return to top of page

return to contents

Chapter 4 European Integration and GATT: The Impact on United States Agriculture

Introduction

Theoretical Framework

Empirical Analysis

Game Simulations

Results

Other Indicators

Conclusions

References

Chapter 4
European Integration and GATT:
The Impact on United States Agriculture