Agricultural Trade in a North American Free Trade Agreement

Agricultural Trade in a North American Free Trade Agreement Thomas Grennes and Barry Krissofl 1. INTRODUCTION

a

GRICULTURE has special interest for trade policy analysts. Because agricultural protectionistshave successllly resisted liberalisationin the past, the potential benefits from future liberalisation are greater than in other sectors. The policy instruments used to protect agriculture are different from those used in other sectors. In addition to tariffs and import quotas at the border, agricultural trade is influenced by a complex set of domestic policies that regulate prices and quantities of finalproducts and inputs. Protracted disputes over agricultural issues have delayed progress on other trade issues in both the Uruguay Round and the North American Free Trade Agreement (NAFTA) negotiations. Given the distinctive role of agriculture in trade policy, this paper will analyse the effect on trade in agricultural products of the proposed NAFTA. Although the subject of NAFTA has been addressed previously in this journal (January 1992), the agricultural sector has received little attention. By providing more detail about individual agricultural products and product-specific policies, this paper will complement the more aggregative analyses. Several specific issues will be addressed. First, what will be the effect of NAFTA on the pattern of agricultural exports and imports? Second, how would the effects '9

THOMAS GRENNES is from North Carolina State University. BARRY KRISSOFF is from the Economic Research Service, US Department of Agriculture. The views expressed are those of the authors and not necessarily those of NC State University or the US Department of Agriculture. The authors appreciate the constructive suggestions of an anonymous referee. I A NAFTA agreement in principle was announced by the negotiators on 13 August, 1992 (Lawrence, 1992). Strictly speaking, North America includes the Caribbean islands and the countries of Central America. However, it has become conventional to refer to the trilateral arrangement among the United States, Canada, and Mexico as NAFTA. However, spokesmen for the excluded countries have already expressed concern about the possible adverse effects of excluding them (Rodriguez, 1992).

0 Basil Blackwell Ltd. 1993, 108 Cowley Road, Oxford OX4 IJF, UK and 238 Main Street, Cambridge, MA 02142, USA.

483

484

THOMAS GRENNES AND BARRY KRISSOFF

of NAFTA differ from the effects of unilateral trade liberalisation by Mexico? Third, how much trade diversion would there be and what products would be diverted? Fourth, how would macroeconomic changes such as income and exchange rate adjustment affect the agricultural sector?

2. THE IMPORTANCE OF DISAGGREGATION

Other models dealing with NAFTA have emphasised interaction among the major sectors of the economy and have treated agriculture as a single aggregate (Brown, Deardorff and Stem, 1992). However, aggregation over many diverse products and product-specific policies conceals important information that is relevant to trade. Consider two situations that are both consistent with a-constantaverage price for agricultural products. In case one, prices of grain and horticultural products are constant, but in case two, grain prices rise by 50 percent and horticultural prices fall by 50 percent (assuming equal weights). The two situations have radically different implications for resource allocation and trade within the agricultural sector. Disaggregation allows for a more precise treatment of substitutes and complements, as well as input-output relationships (for example, feed grain and meat). Disaggregation also permits a more precise matching of policies and products. Some products are subject to price supports, input subsidies and land use restrictions, but other products are free from some or all of these features. Quantification of levels of protection is more precise for narrow product categories. Also when parameters are not econometrically estimated, more reliable parameter values can be obtained for m o w product categories (like wheat) than for broad, heterogeneous aggregates (all agricultural goods). To illustrate the potential usefulness of a model of the agricultural sector, two results from more aggregate models should be noted. Brown, Deardorff and Stem treat agriculture as a single aggregate, and an implication of their model (1992, p.29) is that NAFTA would cause a reduction in output and employment in United States agriculture. A study by Peat Marwick provides some greater commodity detail by dividing agriculture into four sectors: animal products, field crops, fruits and vegetables, and other products. An implication of their analysis is that the US field crops sector would contract as a result of NAFTA. These results should be met with some scepticism in view of the widely held opinion that United States producers have a comparative advantage in agriculture in general and in grains and oilseeds in particular. The pre-NAFTA pattern of trade shows the US exporting grain and oilseeds to Mexico and importing horticultural products from Mexico. In addition, there is two-way trade in livestock and meat. Unless current trade barriers severely distort the pattern of trade in addition to reducing the volume of trade, one might expect a NAFTA to largely magnify the current trade pattern, i.e., increase US exports of grain and oilseeds, @ Basil Blackwell Ltd. 1993

AGRICULTURAL TRADE IN A NORTH AMERICAN FTA

485

TABLE 1 Commodities Used in the US-Mexico Analysis Grain/OiLseeds

Livestock/Meat/Dairy

Wheat Corn Other coarse grains Soybeans Soymeal Soyoil Other oilseeds Other meals Other oils

Live cattle Beef Pork Poultry Eggs Butter Cheese Milk powder Fluid milk

Horticultural

Other

Melons Frozen orange juice concentrate Cucumbers Onions Peppers Tomatoes

Sugar Cotton Tobacco Coffee Dry beans

imports of horticultural products, and two-way trade in livestock and meat. These are the findings of the US International Trade Commission (ch. 4), and they are also the results of our own model. Casual evidence from the political process also supports this view. The Mexican government has attempted to protect its corn farmers from the adverse effects of greater imports from the United States, and north of the border some of the strongest opposition to NAFTA has come from representatives of fruit and vegetable producers.

3. DESCRIPTION OF THE MODEL

The methodology for our analysis is to use a partial equilibrium 3-region, 29-commodity static model that emphasises specific products in the agricultural sector (Table l).3The regions in the model are: the United States, Mexico, and Rest-of-World (ROW). Canada is included in ROW to simplify the analysis and because agricultural trade between Mexico and Canada is very small. The model is based on national product differentiation. Each country produces a product that can be distinguished from other country producers. For example, US, Mexican, and ROW corn are assumed not be to homogeneous but, instead, are imperfect A specific application to NAFTA appears in Grennes et al. (1991). The model uses the Static World Policy Simulation (SWOPSIM) framework developed by Roningen (1986) and extended by Roningen, Sullivan and Dixit (1991). 0 Basil Blackwell Ltd. 1993

486

THOMAS GRENNES AND BARRY KRISSOFF

substitutes. In this way, we allow for possible expansion of trade between the United States and Mexico and trade contraction between ROW and the United States and between ROW and Mexico. The partial equilibrium model contains commodity supply and demand equations which are parameterised to reproduce 1988 data for the United States, Mexico, and ROW. We call this our BASE solution and we refer to the 1988 data as the base period. In order to simulate a NAFTA, two-way border protection between the United States and Mexico is removed. To obtain this solution, the base model is modified by removing price wedges representing two-way border protection between the United States and Mexico. The model then recalculates prices and domestic supply and demand levels, and the volume of trade. The pattern of prices and quantities observed in the base period is then compared with the pattern that emerges from the revised solution. For example, the removal of a Mexican price wedge equivalent of an import licence lowers the price of US produced corn that Mexicans consume. The lower price encourages demand for US produced corn and reduces demand for Mexican produced corn depending on the elasticity of substitution4between them. NAFTA is represented by removing all tariffs and the tariff equivalent of licences and quotas that were in place in 1988. All domestic policies are assumed to remain in place. For purposes of comparison, a second experiment consists of unilateral free trade by Mexico. The results presented here describe the impact of the scenario in a typical year after each scenario is fully implemented and the agricultural sectors of the participating countries (and ROW) have had several 1988-like years in which to adjust. Thus, we assume that world agricultural markets were in intermediate-run equilibrium under 1988 conditions and all other exogenous variables pertinent to agricultural markets remain the same over the simulation period, ‘Intermediaterun’ means that the supply and demand elasticities in the model represent about a 5-year period of adjustment to changes in policies and prices. The scenario represents an idealised case, and the precise features of the actual NAFTA may differ from the one being modelled here. In the US-Canada agreement, tariffs are being phased-out over a 10 year period that began in 1989. In the announced NAFTA, border policies are to be phased-out over a period of 0-15 years depending on the sensitivity of the sector in question. For example, for Mexican corn and dry beans, and US concentrate orange juice and cantaloupe, duties will be eliminated over a 15 year period beginning in 1994 (Lawrence, 1992). No attempt

Details on the model structure and the database can be found in Liapis, Krissoff and Neff (1992). Briefly, demand and supply equations are specified for each commodity within each country. Demand depends on consumer prices, the quantity of an intermediate input (a meat supply quantity for the demand for feed products), and an income term. Supply depends on producer prices (or consumer prices of feed grains). 0 Basil Blackwell Ltd. 1993

AGRICULTURAL TRADE IN A NORTH AMERICAN FTA

487

TABLE 2 Changes from BASE in Agricultural Exports due to NAFTA

Importers: Exporter

us

United States Mexico Rest-of-World

-

Total

482

166 3

- 39

169

443

Total Exports

ROW

Mexico

5

423 171 - 36

- 54

558

- 59

-

is made here to model trade during a transition period before a fully implemented NAFTA.’

4. IMPLICATIONS FOR US AGRICULTURE

Based on 1988 conditions for a US-Mexico NAFTA, the model indicates an increase in US agricultural exports to Mexico of $482 million and an increase in US agricultural imports from Mexico of $166 million (see Table 2). This increase represents over a 15 percent expansion from the $4 billion trade in 1988. The greater increase in US exports relative to imports is partly due to the higher initial level of border protection in Mexico (25 percent) compared to the United States (5 percent) for modelled products. It is instructive to compare the effects of NAFTA with the effects of unilateral free trade by Mexico. In this case Mexico removes its trade barriers against the United States and ROW, but neither trading partner reciprocates. In fact, Mexico did relax many of its trade barriers before and during the negotiations on NAFTA. The results are shown in Table 3. In the case of unilateral liberalisation by Mexico, US exports to Mexico rise by $435 million or 94 percent of the increase with NAFTA. Thus, for the United States the effects of unilateral action by Mexico are not substantially different from NAFTA (see also Sobarzo, 1992). However, with unilateral action by Mexico, US imports only increase by $25 million or 15 percent of the increase with NAFTA. Retaining US barriers against Mexican products would deny Mexican producers many of the benefits from NAFTA.

~

~

~~~~~

Some models incorporate a multi-period analysis by phasing-in border policy changes and adding intermediate or long-term projections of income and population growth on the demand-side and productivity growth on the supply-side based on trend or other exogenous information. Although these factors are surely relevant for the future effects of a NAFTA, they often require very strong assumptions making it difficult to distinguish the direct effects of removing border protection from the exogenous demand and supply-side projections. Basil Blackwell Ltd. 1993

488

THOMAS GRENNES AND BARRY KRISSOFF TABLE 3 Unilateral Mexican Trade Liberalisation ($ millions) Importers:

Exporter

us

United States Mexico Rest-of-World

25 16

Mexico

41

435

-

30 465

Total Exports

ROW

- 46 24

-22

389 49 46 484

a. Implications for Commodity Groups The United States’ main farm exports to Mexico are feed grains, oilseeds, live animals, meat, and dairy products. These exports are likely to expand with liberalised trade. We estimate that grains and oilseeds exports would increase by $430 million, nearly 90 percent of the expansion in US agricultural exports (Table 4). With the rise of exports to Mexico, total US agricultural exports to all countries would increase, but by less than 2 percent. Mexico’s main exports to the United States are tropical and speciality crops such as coffee, fruits, and vegetables, as well as live animals. Horticultural exports would increase by $104 million or over half of Mexico’s expansion of exports to the United States. There would also be an increase in Mexican exports of feeder cattle. The NAFTA examined here implies a small (less than one-percent) net expansion in US agricultural production. Producers of some commodities such as feed grains would expand production. Production of certain horticultural products would decrease slightly. No single horticultural product in the model showed a production decline in excess of 2 percent. Also, some exports of US horticultural products such as deciduous fruits (apples, pears) are expected to expand. One reason for the small effect per year is that certain Mexican products (for example tomatoes) compete with US production mainly during the winter months. The total increase in production of export-oriented commodities would be small because agricultural exports to Mexico represent a small portion of US production. Corn is a good example. Corn exports to Mexico are estimated to increase about 65 percent due to the removal of Mexican border restrictions. However, this large percentage change only represents about a 3 percent increase in total US corn exports and less than a 1 percent increase in production.

b. Grains and Oilseeds When Mexican border protection vis-a-vis the United States is removed, prices in Mexico of US crops decline. Corn and other coarse grains experience price 0 Basil Blackwell Ltd. 1993

AGRICULTURAL TRADE IN A NORTH AMERICAN FTA

489

TABLE 4 Changes from BASE in Agricultural Exports by Commodity Group

Exporter

us

Importers: Mexico

United States Grainsioilseeds LivestocWmeats/dairy Horticulture Other Total Mexico Grains/oilseeds LivestocWmeats/dairy Horticulture Total Rest-of-World Grainsloilseeds LivestocWmeats/dairy Horticulture Other Total

ROW

Total Exports

430 49 1 3 -

-61

0 2

369 49 3

482

- 59

423

-2 56 104 __

9 0 -6 __

11 57

5 -

166

5

171

2

-1

12 5 - 16 0

- 38

-25

0 0

5 15 -1 -

3

- 39

- 36

-1

declines of over 30 percent (Table 5). Mexican consumers increase their demands for US corn and other coarse grains by more than 50 percent and decrease their demands for comparable goods produced in Mexico. The increase in Mexican demand of US products has a marginal effect on US pricies. For example, corn and other coarse grain producer prices increase by 1 and 2 percent, respectively. The higher prices encourage a slight production response; corn and other coarse grain supplies expand by 0.3 and 1.7 percent, respectively. TABLE 5 Changes from BASE in Agricultural Production, Consumption and Prices for Select Grains and Oilseeds in Mexico and the United States'

Country/Commodity Mexico US corn Mexican corn US other coarse grain Mexican. other coarse grain United States US corn US other coarse grain

'

Production

-7.3

- 10.9 0.3 1.7

Consumption

Price

64.0 -7.3 50.1 - 13.9

-33.2 - 15.9 -32.3 - 15.8

-0.8 -2.1

1.1 2.3

Note: As mentioned earlier in the paper, US, Mexican, and ROW corn are assumed not to be homogeneous but, instead, are imperfect substitutes. Thus all three types of corn are available in each country or region, but the United States produces only US corn and Mexico produces only Mexican corn.

0 Basil Blackwell Ltd. 1993

490

THOMAS GRENNES AND BARRY KRISSOFF FIGURE 1 Changes from BASE in the Value of US Grain and Oilseed Exports

wheat

Corn Other coarse grains Soybeans Soymeal Soyoil Other oilseeds Other meals Other oils 50

0

100

150

200

250

million $

Corn

...........................................

....,.. ....:.:.:. ........................ .......,.. ........

Other coarse grains Soybeans Soymeal Soyoil Other oilseeds. Other meals. Other oils

I

~

0

10

20

30

40

50

60

70

percent Total exports

Exports to Mexico

The United States increases its corn and other coarse grain exports by $186 and $100 million, respectively, shipping $219 and $123 million more to Mexico and $33 and $23 million less to ROW (Figure 1). The US exports $84 million more of oilseeds and products and $4 million more of wheat to Mexico. Mexican coarse grain imports from the United States expand by about 60 percent relative to our 1988 base period (Table 4). This increases the interdependence between the United States and Mexico as Mexican imports from the United States comprise around 40 percent of domestic consumption (Table 6). The share of US corn in the Mexican market rises from 23 percent to 34 percent with NAFTA. Historically, coarse grain imports have comprised 25 to 30 percent of domestic @ Basil Blackwell Ltd. 1993

AGRICULTURAL TRADE IN A NORTH AMERICAN FTA

491

TABLE 6 The United States’ Share of Mexican Grains and Oilseed Market, BASE and NAFTA Scenario

Commodity Wheat Corn Other Coarse Grains Soybeans Soybean Meal Soybean Oils Soybean Oilseeds Other Meals Other Oils

BASE Percent

NAFTA Scenario Percent

9 23 31 83 26 12 12 2 27

10 34 44 84 28 13 16 3 30

use, although there has been substantial variation from year to year. The United States’ share of the Mexican oilseed market also expands. c. Livestock and Dairy

The removal of Mexican import protection increases the Mexican demand for

US livestock and meats and has a marginal effect in increasing US livestock and meat prices. This encourages a small US supply response in cattle and poultry. As a consequence, the United States increases its exports to Mexico by nearly $18 million in slaughter cattle, a little over $5 million in beef, and about $25 million in pork and poultry (Figure 2). This represents an 8 percent increase in US cattle exports and 3 to 4 percent increase in US pork and poultry exports. On the Mexican consumer side, the removal of import protection allows increased availability of meats and dairy products lowering domestic consumer prices of US products from 5 to 20 percent (Table 7). Mexico also increases its feeder cattle exports to 1065 thousand head from the 1988 base of nearly 850 thousand head, approximately a 25 percent increase. This represents a $55 million rise in Mexican exports to the United States, nearly onethird of the increase in all Mexican agricultural exports6 Model results are sensitive to changes in consumers’ perceptions of a domestic product compared to a foreign product. (The elasticities of substitution across product groups for each good are assumed to equal 3, following the practice by Brown and Stern, 1989.) For instance, decreasing the elasticity of substitution for corn in Mexico (from 3 to 1) implies that Mexican consumers are more discerning with respect to domestic and foreign corn, i.e., Mexicans are less willing to substitute yellow (US) corn for white (Mexican) corn in their diets. Under this scenario, US corn exports to Mexico expand 20 percent less than indicated in the test. In another example, the behavioural assumptions of Mexican consumers with regard to purchasing poultry are altered; US and Mexican poultry are assumed to be more similar. In this case, the removal of Mexican import barriers on feed grains reduces the production cost of poultry and Mexicans buy nore poultry from domestic relative to US producers. Thus, the increase in US poultry exports to Mexico is less than shown in the text. 0 Basil Blackwell Ltd. 1993

THOMAS GRENNES AND BARRY KRISSOFF

492

FIGURE 2 Changes from BASE in the Value of US Livestock, Meats, Eggs, and Dairy Exports

Cattle Beef

I

Pork Poultry

10

5

0

15

million $ Cattle Beef Pork Poultry Eggs

Butter Cheese

0

10

20

30

40

50

per cent

d. Horticultural Commodities There is an increase in US demand for Mexican melons, frozen orange juice concentrate (FCOJ), cucumbers, onions, green peppers, and tomatoes with the removal of US import tariffs (Table 8). As a result, Mexico expands production in vegetables, by approximately 2 to 7 percent and in FCOJ, by nearly 19 percent. 0 Basil Blackwell Ltd. 1993

493

AGRICULTURAL TRADE IN A NORTH AMERICAN FTA TABLE 7 Changes from BASE in Agricultural Production, Prices for Select Meat and Livestock Commodities

CountrylCommodity Mexico US cattle Mexican cattle US beef Mexican beef US pork Mexican pork US poultry meat Mexican poultry meat US poultry eggs Mexican poultry eggs United States Mexican cattle US cattle

Production

Consumption 11.2

0.2 -0.2

-0.5 15.0

-0.2

-

-0.5 2.1 2.5

-7.4

- 15.7 -5.0

25.3

-0.3 -8.1

0.5

-1.1

23.9 2.1 4.8

-9.1 3.0 - 10.0 -9.3

2.5 26.0

0.0

Price

-0.1

-9.5

-0.2

TABLE 8 Production, Consumption, and Price Responses for Horticultural Commodities

Country/Commodity Mexico (Mexican produced) Melons FCOJ Cucumbers Onions Green Peppers Tomatoes United States (US produced) Melons FCOJ Cucumbers Onions Green Peppers Tomatoes United States (Mexican produced) Melons FCOJ Cucumbers Onions Green Peppers Tomatoes 0 Basil Blackwell Ltd. 1993

Production Percent

Consumption Percent

Price Percent

2.4 18.8 6.6 3.8 1.6 1.8

-0.6 - 13.4 - 1.3 -0.7 -0.2 -0.4

1.3 12.2 3.6 2.1 0.9 1.o

-0.5

-0.6

-0.3

-0.4

-0.7 -0.3

-2.0

-2.2 -0.1 - 1.8 -0.9

-2.4 -1.0 - 1.9 - 1.9

10.8 32.1 10.8 13.7 I .2 10.2

-4.0 -9.2 -5.2 -5.8 -5.1 -6.1

-0.8 - 1.5 -0.7

-

-

-

THOMAS GRENNES AND BARRY IUUSSOFF

494

FIGURE 3 Changes from BASE in the Value of Mexican Fruit and Vegetable Exports

FCOJ

Cucumbers Onions Peppers Tomatoes

0

20

10

30

50

40

million $

Melons

FCOJ Cucumbers Onions Peppers Tomatoes

0

5

10

15

20

25

30

35

per cent

Its export volume expands in the range of 5 to 15 percent for vegetables and 20 percent for FCOJ, equal to $45 million (Figure 3). The removal of US tariffs on horticultural products results in price decreases in the United States ranging from 0 to 2 percent for US produced products and 4 to 9 percent for Mexican produced products. The increased share in the US market by Mexican fruits and vegetables is very small, 1 to 3 percentage points (Table 9). Q Basil Blackwell Ltd. 1993

AGRICULTURAL TRADE IN A NORTH AMERICAN FTA

495

TABLE 9 Mexico’s Share of US Fruit and Vegetable Market, BASE and NAFTA Scenario Commodity

Melons Frozen Orange Juice Concentrate Cucumbers Onions Peppers Tomatoes

BASE

NAFTA Scenario

12 3 36 7 22 19

13 4 39 8 24 20

5. TRADE DIVERSION

The effect of NAFTA on trade diversion is expected to be small. One reason is that most of Mexico’s agricultural imports already come from the United States, in the absence of a NAFTA. Mexico’s imports from ROW are expected to decrease by $39 million as a result of an agreement, mostly in grain and oilseeds. Since US imports from ROW are expected to increase by $3 million, the net effect of NAFTA is a reduction in imports from ROW of $36 million. The reason for the increase in US imports from ROW is that Mexico purchases more US grain, oilseed, and meat products and US consumers, in turn, marginally increase their purchases of similar products from ROW. The only important case of trade diversion is the substitution by the US of FCOJ imports from Mexico for those from Brazil.

6. INCOME GROWTH

NAFTA trade and investment reforms across all sectors can affect income growth, exchange rate movements, capital flows, and factor markets. These ‘macroeconomic’ variables, in turn, can influence agricultural markets and, therefore, the agricultural trade findings reported above. Unlike macroeconomic trade models, the sectoral model of this paper cannot show the macroeconomic effects of NAFTA on income growth and exchange rate movements, for example. However, sensitivity anaiysis to different assumptions regarding changes in income growth and exchange rates can be provided. Conventional macro models have attempted to represent the effect of a preferential trade agreement on income growth, and a typical result is a one time increase in real GDP of 1.6 percent in Mexico, 0.7 percent in Canada, and 0.1 percent in the United States (Brown, Deardorff and Stem, 1992). A robust result is that the total gains would be small and most of the gains would accrue to Mexico. However, Kehoe (1992) has claimed that traditional models may seriously understate the income growth induced by NAFTA. He appeals to the dynamic benefits captured 0 Basil Blackwell Ltd. 1993

496

THOMAS GRENNES AND BARRY KRISSOFF

by endogeneous growth models. In particular, Mexico would be expected to benefit from increased specialisation and additional intra-industry trade (measured by the Grubel-Lloyd index). For example, additional two-way trade is expected in livestock and meat. The potential gains to Mexico can be seen by asking what would happen to worker productivity if specialisation and intra-industry trade increased to about the level that Ireland achieved, seven years after it joined the EEC. If that occurred, the growth rate of manufacturing output per worker would increase by 1.43 percent per year. After 30 years, output per worker would be more than 50 percent higher than it would otherwise be (Kehoe, 1992, p. 27). Thus, the dynamic benefits from increased openness could swamp the static benefits measured by conventional general equilibrium models. One might expect greater gains from intra-industry trade m manufactured goods, which would stimulate the agricultural sector indirectly by faster growth in GDP. Intra-industry trade also occurs for agricultural products (at the three digit SITC level), especially trade in agricultural inputs (Hart and McDonald, 1992). The effects of more rapid income growth in Mexico can be shown in terms of our model. Let Mexican real income grow by 10 percent as a result of NAFTA.’ Mexican imports of agricultural products from the United States will increase by $630 million compared to $482 million without the income effect. Most of the additional imports will be meat products and feedgrains. Higher income in Mexico will stimulate demand for domestic as well as foreign goods, and all Mexican exports to all sources will decline by 65 percent relative to the case of constant income.

7. EXCHANGE RATES

The exchange rate policies of member countries are relevant to free trade agreements because exchange controls can be used to substitute for tariffs and import quotas. There is evidence from other countries that exchange rate policy can have a greater impact on a particular sector of the economy than sector-specificpolicies such as taxes, subsidies, and price controls (Krueger et al, 1988). In the case of NAFTA, a change in exchange rate policy can affect the entire agricultural sectors of the member countries. In 1992 and early 1993 there has been discussion about the possibility of an overvalued peso. Government policy has allowed a rate of nominal depreciation of the peso less than the inflation differential between Mexico and the US. The result has been real appreciation of the peso. Government officials have stated that a greater nominal depreciation of the peso would undermine the credibility The income elasticities for Mexico are in the range of 0.4 to 0.9 for livestock products, 0.4 and 0.6 for horticultural products, and 0.2 to 0.3 for food grains and dry beans. @ Basil Blackwell Ltd. 1993

AGRICULTURAL TRADE IN A NORTH AMERICAN FTA

497

of its anti-inflation policy. Since a real appreciation of the peso is equivalent to a tax on Mexican exports and a subsidy to Mexican imports, the effect of alternative exchange rate policies on agricultural trade can be analysed in terms of our trade model. One possible rule for Mexican exchange rate policy is to allow a rate of nominal peso depreciation that just offsets the Mexican-US inflation differential i.e., follow Purchasing Power Parity. A base year for applying the rule must also be chosen, and different base years could result in different real exchange rates and different patterns of trade. How responsive is US-Mexican agricultural trade to a change in the real exchange rate between the dollar and the peso? Here we take the real exchange rate to be the nominal rate relative to the price levels in the United States and Mexico. The real rate is taken as exogenous to the model, and it is calculated as the bilateral rate using consumer prices.8 Suppose Mexican government officials consider 19629and 1986 as alternative base years. To satisfy PPP in 1988 relative to a base year 1962 would have required a 4 percent appreciation (fewer pesos per dollar) of the peso (see Fleissig and Grennes, 1991). However, satisfying PPP in 1988 relative to a base year of 1986 would have required a 19 percent depreciation (more pesos per dollar) of the peso. What would be the effect of these two alternative exchange rate policies on agricultural trade? The effects of the two exchange rate policies are compared with the effects of the actual 1988 exchange rate policy in Table 10. The 4 percent appreciation of the peso combined with NAFTA would increase US agricultural exports to Mexico to $543 million rather than $482 with the actual 1988 exchange rate. Grains and oilseeds exports increase by an additional $50 million. US imports would be $149 million rather than $166 million, and most of the decline would be in horticulture. Conversely, if the peso had been depreciated by 19 per cent in 1988 in the presence of NAFTA, US agricultural exports would have been only $242 million rather than $482 million. Grain and oilseed exports would have been diminished by over $200 million. At the same time US imports from Mexico would have been $231 million rather than $166 million.

8. NAFTAANDTHEURUGUAYROUND

The effect of a NAFTA on trade depends on the level of protection facing nonmember countries at the time the regional arrangement is implemented. NAFTA See Fleissig and Grennes (1991) for a discussion of measurement issues. Ten Kate (1992) presented PPP calculations for Mexico using 1978 calculations as the base year. This is equivalent to using 1962 as the base year, since the real value of the peso was the same in 1962 and 1978 (Fleissig and Grennes, 1991). 0 Basil Blackwell Ltd. 1993

498

THOMAS GRENNES AND BARRY KRISSOFF

TABLE 10 Exchange Rate Effects: Changes from BASE in Agricultural Exports by Commodity Group NAFTA Exporter/Commodity

us

Mexico

NAFTA and 4% Peso Apprec.

NAFTA and 19% Peso Deprec.

us

us

Mexico

Mexico

million dollars United States Graindoilseeds LivestocWmeatddairy Horticulture Other Total

430 49 1 3 -

-

482

-

-

__

480 59 1 4 -

-

543

-

~

12.5%increase Mexico Graindoilseeds LivestocWmeatsldairy Horticulture Other

2 56 104 3 ~

Total

166

1 54 92 2 149

-

10.0%decrease

242

50.0%decrease

-

-

232 9 0 1 -

~

4 62 156 9 23 1

-

__

-

39.0%increase

Note: Base year is 1988

is a discriminatory arrangement and the margin of discrimination is the difference between the most-favoured nation level of protection facing non-members and the zero level of protection against members. In the limit if the world adopts free trade, NAFTA will have no effect. Our analysis is based on levels of protection in place in 1988. By reducing the margin of discrimination between members and nonmembers, a successful Uruguay Round should reduce the impact of NAFTA and all other preferential arrangements, including those already in place. For example, the large amount of agricultural trade diversion that has occurred since Greece, Spain and Portugal joined the EEC would be reduced (Plummer, 1991a and 1991b; and Hamilton and Winters, 1992). A successful Uruguay Round might also eliminate some domestic policies (price supports, land diversion, input subsidies) that are assumed to remain in place in the present analysis. In addition, unilateral liberalisation of domestic policies by Mexico will reduce the effect on trade of NAFTA.

9. ADDING COUNTRIES TO THE NAFTA

The United States government has invited other countries in the Americas to participate in a broader regional free trade agreement. President Bush’s Enterprise @ Basil Blackwell Ltd. 1993

AGRICULTURAL TRADE IN A NORTH AMERICAN FTA

499

for the Americas proposal opens up the possibility for a larger regional bloc (sometimes called Western Hemisphere free trade), and several governments have expressed interested in immediate participation, including Chile and Costa Rica (Saborio, 1992). Each potential entrant could carry out separate negotiations with NAFTA members, but it would be possible to reduce the costs of many separate negotiations by including in the original NAFTA agreement some standards of accession that new members would be expected to meet (Rodriguez, 1992). It would therefore be instructive to show the effect of expansion on the new members and on the initial members of NAFTA. We have not carried out a full analysis of expansion of membership, but the effects will vary by country and product. For example, in the case of wheat, the effect of adding Latin American members depends heavily on whether Argentina and Brazil are included. Brazil is the largest wheat importer in the region and Argentina is the largest wheat exporter, and Brazil imported 40 percent of its wheat from Argentina in 1988. If Brazil joins but Argentina does not, much of the Brazilian market will be supplied by Canada and the United States. Erzan and Yeats have claimed that most of the benefits from Western Hemisphere free trade would accrue to Mexico and Brazil. They also compare the effects of exclusive free trade agreements with the US and broader agreements that include Mexico and the ten major South American countries. The benefits to Mexico are not very sensitive to adding other countries. The gains to Mexico from an agreement with the United States and eleven other countries would be 98 percent of the benefits of an exclusive agreement with the United States (Erzan and Yeats, 1992, p.34). The effect of adding countries depends on how import quotas (for example, sugar) are treated. If the quota for imports into the US from all sources is held constant, a NAFTA including Mexico would divert US imports from other sources.

10. MOBILITY OF CAPITAL AND LABOUR

Analysing the effect of NAFTA on migration and investment is a difficult task. Mexican migration to the US and foreign investment in Mexico (especially maquiladoras) have been important before NAFTA was discussed. Some additional foreign investment has occurred in Mexico recently and more will occur even if NAFTA is not completed. A problem is distinguishing between migration and investment that are induced by NAFTA and those flows that would have taken place in its absence. This problem is particularly important because, in most models, the effects of factor mobility dominate the effects of changing tariffs and import quotas. NAFTA is beneficial to Mexico partly because of the direct gains from increased trade. An additional benefit is that a free trade treaty would make the economic reforms of the Salinas administration binding on subsequent administrations in 0 Basil Blackwell Ltd. 1993

500

THOMAS GRENNES AND BARRY KRISSOFF

Mexico. This commitment would give a sense of permanence to the recent reforms, and it would send a favourable signal to domestic and foreign investors. In our model factors are mobile within the 29-good agricultural and nonagricultural sectors. Thus, employment is expected to decline in Mexican grains (especially corn) and expand in Mexican horticulture as relative output prices adjust. Since the increase in employment in the horticultural sector is not expected to adequately compensate for the lost employment in the traditional agricultural sectors, employment expands in the nonagricultural sectors. lo In the past there has been little foreign investment in Mexican agriculture. In 1989 less than 0.1 percent of total foreign direct investment was in agriculture (Cook et al., 1991, p.471). Investment was discouraged by several factors including restrictions on foreign ownership, limits on farm size, and complex restrictions applicable to ejidos (Heath, 1992). Several of these restrictions have been relaxed or removed recently, including a constitutional change that allows ejidatarios to rent land and participate in joint ventures with domestic and foreign corporations. As a result of more liberal investment rules, two new tomato processing plants opened in Mexico in 1991, and the industry is expected to expand whether a NAFTA is completed or not (USDA, 1992, p.96). Some models explicitly treat migration between agriculture and the rest of the domestic economy (Robinson et al ., 1991;and Levy and Van Wijnbergen, 1991). Displacement of Mexican workers by corn imports is a politically sensitive issue. What happens to these low wage workers depends on employment opportunities in the expanding horticultural sector as well as non-agricultural opportunities in Mexico. Migration also depends on the consumer price of corn in cities. In the past the producer price of corn has been kept artificially high and the urban consumer price has been kept artificially low. Employment opportunities for rural workers displaced by corn imports might also be improved by NAFTA if additional foreign investment occurs, especially if investments are located near the regions of declining corn production, rather than in the maquiladora zone along the US border. Some models have attempted to deal with the effects of NAFTA on foreign investment, but authors usually resort to some ad hoc procedure such as increasing Mexico’s capital stock by some fixed percent (usually 10 percent). This procedure was used by Brown, Deardorff and Stem (1992), by the Peat-Marwick study (1991), and in the paper by Robinson et al. (1991). The exercise is instructive because in all three cases the effect on Mexican income of increasing the capital stock dominated the effect of reducing trade barriers. For example, in the Brown, Deardorff, Stem model the effect on Mexican income of additional capital was l o There is an interaction between mobility of capital and labour in the sense that additional employment in Mexican horticulture depends on investment in transportation and water projects. This is an implicit assumption of the analysis, since factor markets and the non-agricultural sector are not explicitly modelled.

0 Basil Blackwell Ltd. 1993

AGRICULTURAL TRADE IN A NORTH AMERICAN FTA

501

2-3 times the effect of relaxing tariffs and import quotas (p. 11). The effect of NAFTA on international investment may be its most important effect, but it has been difficult for modellers to represent it. An additional complication is that the profitability of foreign direct investment depends on the rules of origin that determine what products are duty-free in member countries. Rules of origin are relevant for a free trade agreement, but not a customs union, because of the possibility of importing a product into a member country with a low external tariff and reexporting their product free of duty to a member country with a high external tariff.

1 1 . CONCLUSIONS

A model of the agricultural sector containing product-specific policy information can be a useful complement to economy-wide models. Several conclusions emerge from the 29 commodity model. A NAFTA will increase agricultural exports of both the United States and Mexico. A larger increase in US exports is expected partly because the initial level of protection is higher in Mexico. More specialisation is expected as the United States will export more grain (especially corn) and import more horticultural products. There will also be more intra-industry trade in livestock, and meat products. Trade diversion is not expected to be important, the only significant example is FCOJ. Most of the benefits of NAFTA to American exporters would also be realised if Mexico liberalised its trade unilaterally. The indirect effects of economic growth, exchange rate changes, capital flows, and labour mobility on trade can rival the direct effects of tariffs and quotas. Thus, the results from the 29-agricultural sector specific model should be interpreted judiciously.

REFERENCES Brown, D., A. Deardorffand R. Stem (1992), ‘A North American Free Trade Agreement: Analytical Issues and a Computational Assessment.’ Zhe World Economy, 15, 1 , 11-30. Brown, D.K. and R.M. Stem (1989), ‘Computable General Equilibrium Estimates of the Gains from US-Canadian Trade Liberalization,’ in Economic Aspects of Regional Trade Arrangements, D. Greenaway, T. Hyclak and R.J. Thornton (eds.), (New York University Press). Cook, R.L. et al. (1991), North American Free Trade Agreement: Effects on Agriculture. Volume IV, Fnd and VegetabfeIssues (Park Ridge, IL: American Farm Bureau Research Foundation). Erzan, R. and A. Yeats (1992), ‘Free Trade Agrements with the United States: What’s In It For Latin America?’ World Bank Working Paper WPS 827 (January). Fleissig, A. and T. GreMeS (1991), ‘The Real Exchange Rate Conundrum: The Case of Central America’, APAP II International Paper 1990-1, North Carolina State University, Raleigh (July). GreMeS, T. et al. (1991), An Analysis of a United States-Canada Mexico Free Trade Agreement, IATRC Commissioned Paper Number 10 (November). Hamilton, C. and L.A. Winters (1992), ‘Opening Up International Trade with Eastern Europe,’ Economic Policy (April), 78-1 16. 0 Basil Blackwell Ltd. 1993

502

THOMAS GRENNES AND BARRY KRISSOFF

Hart, T. and B. McDonald (1992), Intra-Industry Trade Indexes for Canada, Mexico, andthe United States, 1962-87, ATAD-ERS-USDA, Staff Report No. AGES 9206 (Washington, February). Heath, J.R. (1992), ‘Evaluating the Impact of Mexican Land Reform on Agricultural Productivity. ’ World Development 20, 695-7 11. Josling, T. (1992), ‘The Treatment of National Agricultural Policies in Free Trade Areas,’ IATRC Working Paper 92-7 (June). Kehoe, T.J. (1991), ‘Free Trade and Economic Growth’, Paper presented at the Conference on North American Trade (Center for Strategic and International Studies, Washington, DC, July). Kehoe, T.J. (1992), ‘Assessing the Economic Impact of North American Free Trade,’ University of Minnesota (May). KPG - Peat Marwick (1991), ‘The Effects of a Free Trade Agreement Between the US and Mexico: Executive Summary’, Policy Economics Group (Washington, 27 February). Krissoff, B., L. Neff and J. Sharples (19M), ‘Estimated Impact of a Potential US-Mexico Preferential Trading Agreement for the Agricultural Sector,’ IATRC Working paper 92-1 (March). Krueger, A.O., M. Schiff and A. Valdes (1988), ‘Agricultural Incentives in Developing Countries: Measuring the Effects of Sectoral and Economywide Policies, ’ World Bank Economic Review, 2, 255-12. Lawrence, R. (1992), ‘North American Trade Pact Reached: Agreement Ends Most Tariffs in Ten Years,’ Journal of Commerce, 13 (August), l a and 4a. Levy, S . and S. van Wijnbergen (1991), ‘Maize and the Mexico-United States Free Trade Agreement’ (Washington, World Bank, February). Liapis, P., B. Krissoff and L. Neff (1992), Modeling Preferential Trading Arrangements for the Agricultural Sector: A US Mexico Enample, ERS-USDA Staff Report No. AGES 9212 (Washington, July). Low, P. (1992), ‘Trade Measures and Environmental Quality: The Implications for Mexico’s Exports,’ in P. Low (ed.), International Trade and the Environment, World Bank Discussion Paper No. 159 (April). Plummer, M. (1991a), ‘Ex-Post Empirical Estimates of the Second Enlargement’, Weltwirtschaftliches Archiv, 127, 171-82. Plummer, M. (1991b), ‘Efficiency Aspects of the Accession of Spain and Portugal to the EC,’ Journal of Common Market Studies, 29, 311-23. Robinson, S., M.E. Burfisher, R. Hinojosa-Ojeda and K. Thierfelder (1991), ‘Agricultural Policies and Migration in a US-Mexico Free Trade Area,’ Giannini Foundation Working Paper No. 61 7 (December). Rodriguez, M.A. (1992), ‘Don’t Lock the Caribbean Basin Out of Free Trade Zone,’ Wall Street Journal, 7 (August), A13. Roningen, V. (1986), A Static World Policy Simulation (SWOPSIM) Modelling Framework. US Department of Agriculture, Economic Research Service, Staff Report AGES860625 (July). Roningen, V., J. Sullivan and P. Dixit (1991), Documentation of the Static World Policy Simulation (SWOPSIM)Modelling Framework, US Department of Agriculture, Economic Research Service, Staff Report No. AGES 9151 (September). Saborio, S. (ed.) (1992), lhe Premise and the Promise: Free Trade in the Americas (New Brunswick, NJ: Transaction Publishers). Sobarzo, H.E. (1992), ‘A General Equilibrium Analysis of the Gains from Trade for the Mexican Economy of a North American Free Trade Agreement,’ Zhe World Economy, 15, 83-100. Ten Kate, A. (1992), ‘Trade Liberalization and Economic Stabilization in Mexico: Lessons from Experience,’ World Development, 20, 659-72. US Department of Agriculture (1992), ‘Agriculture in a North American Free Trade Agreement: Analysis of Liberalizing Trade Between the United States and Mexico’ (Washington, July). United States International Trade Commission (1991), lhe Likely Impact on the United States oJ a Free Trade Agreement with Mexico, USITC Publication 2353 (Washington, February).

0 Basil Blackwell Ltd. 199