Agricultural policies and migration in a U.S.-Mexico free trade area: A computable general...

Sherman Robinson, University qf California, Berkeley,

Mary E. Burfisher, U.S. Department of Agriculture,

Raiil Hinojosa-Ojeda, University of California, Lo; Angeles,

and Karen E. Thierfelder, U.S. Naval Academy

A United States-Mexico agreement to form a free trade area (FTA) is analyzed using an 11 -sector, three-country, computable general equilibrium model that explicitly models farm programs and labor migration. The model incorporates both rural-urban migration within Mexico and international migration between Mexico and the United States. In the model, sectoral import demands are specified with a flexible functional form, an empirical improvement over earlier specifications, which use a constant elasticity of substitution function. Using the model, we identify trade-offs among bilateral trade growth, labor migration, and agricultural program expenditures under alternative FTA scenarios. Trade liberalization in agriculture greatly increases rural- urban migration within Mexico and migration from Mexico to the United States. Migration is reduced if Mexico grows relative to the United States and also if Mexico retains farm support programs. However, the more support that is provided to the Mexican agricultural sector, the smaller is bilateral trade growth. The results indicate a policy trade-off between rapidly achieving gains from trade liberalization and pro- viding a transition period long enough to assimilate displaced labor in Mexico without undue strain.

1. INTRODUCTION

In June 1990, Presidents Salinas de Gortari of Mexico and Bush of the United States agreed to negotiate the establishmerit of a free trade

Address correspondence to Sherman Robinson, Department of Agriculture and Resource Eco~

nomics. University of Califarniu, Berkeley, CA 94701. We thank Kenneth Hanson, Agapi Somwaru. and Constanza Valdes for comments and assis-

tance. This research was supported by a cooperative agreement between the University of Cal- ifornia and the Economic Research Service, U.S. Department of Agriculture. The work a( the

University of California was also supported by grants from the William and Flora Hewlett Foundation and the UC Mexus Program at the University of California. R. I?.-0. was allso

Journai of Policy Modeling 15(5&6):073-701 ( 1993) 613

0 Society for Policy Modeling, 1993 0161~8938/93/$6.00

674 S. Robimon et al.

area (ETA) between their two countries. An agreement would com- pkxnent the United States-Canada Free-Trade Agreement, which went into effect in January 1988, creating a North American free trade area (NAFTA). The trade bloc will not, in fact, be a “free trade area” in which all trade barriers are removed. If the United States-Mexico negotiations follow the the United States-Canada agreement, tariffs will fall to zero over intervals negotiated sector by sector, but liberalization of non-tariff barriers will be selective. For example, United States-Canadian agricultural trade, although substantially liberalized by the gradual elimination of tariffs, is still not free of barriers. Do- mestic agricultural programs in both countries, and the non-tariff barriers used to support them, remain essentially intact (Goodloe and Link, 1991).

Realistic analysis of a United States-Mexico FTA should consider alternative treatments of agricultural trade, including partial liberalization and the retention or restructuring of domestic agricultural programs. This article provides such an analysis using a three-country, 1 l-sector, computable general equilibrium (CGE) model. Our FTA- CGE model focuses on three modeling issues. First is the explicit modeling of agricultural policies in the two countries in order to capture the linkages, particularly in Mexico, between bilateral agricultural trade barriers and social policy. Mexican agricultural policies modeled include tariffs, import quotas for corn and other grains, input subsidies to producers and processors, and tortilla subsidies to low-income consumers. For the United States, we model the “deficiency payment program” as well as tariffs and quotas. Deficiency payments and the tariff equivalents of quotas are determined endogenously, rather than being treated as fixed ad valorem wedges. Because subsidies to farmers and processors are included in the model, one can analyze the fiscal impacts of changes in agricultural output and trade.

A second issue is labor migration, including both rural-urban migration in Mexico and mig,.... ration from Mexico to the United States. Although not explicitly part of the FTA negotiations, labor migration k sensitive to economic conditions in t’he two countries and to the mix of trade and domestic policies in Mexican agriculture. The ETA-CGE model includes migration explicitly; and 2 is an important feature in the empirical results.

__- supported by a grant from the Secretaria de Relaciones Exteriores de Mexico. Any views expressed

in this article P”: those of the authors and do not necessarily reflect the views of the supporting organizations.

U.S.-MEXICO AGRICULTURAL POLICIES AND MIGRATION 675

The third modeling issue concerns the specification of import demand. The standard approach in open-economy CGE models has been to adopt the Armington assumption of product differentiation coupled with use of a constant elasticity of substitution (CES) import aggregation function. The CES specification has been criticized because it constrains import demand equations to an expenditure elasticity of one, and also implies that every country has market power in its export markets.’ Brown (1987) shows that these assumptions have led earlier multicountry trade models to generate unrealistically large terms-of- trade effects when trade is libera!ized. The ETA-CGE model uses the almost ideal demand system (AIDS) to describe import demand, a flexible functional form that allows nonunitary expenditure elasticities and yields more realistic empirical results while retaining the essential property of imperfect substitutability.

In Section 2, we present the core CGE model and describe how we model import demand, migration, and agricultural programs. In Sec- tion 3, we present model simulations. Our analysis with the FTA-CGE model focuses on the trade-offs among bilateral export growth, migration, and farm program expenditures. Trade liberalization, in which both tariffs and quotas are removed, results in significant bilateral export growth but also large migration flows. We estimate how much Mexican growth is required to absorb the increased rural migration without increased migration to the United States. We show that migration can be reduced by simultaneously lowering trade barriers and increasing agricultural program expenditures in Mexico to support rural employment. Our results indicate that it is feasible to design transition policies so that Mexico can adjust gradually to the structural changes induced by trade liberalization. Then Mexico can reap the benefits from an ETA without a precipitous migration-induced shock to the labor markets in both countries.

2. CORE THREE-COUNTRY CGE MODEL

The ETA-CGE model is an 1 l-sector, three-country, CGE model composed of two single-country CGE models linked through trade and migration flows, plus a set of export-demand and import-supply equations to represent the rest of the world.’ Table 1 and Figure 1 present

‘The CES formulation has also been criticized on econometric grounds [Alston et al. (1990)].

The model is an extension of earlier CGE modeling undertaken at the U.S. Department of Agriculture (USDA), which began with the single-country USDA Economic Research Service

(ERS) CGE model. designed to provide a framework for analyzing the effects of changes in

agricultural policies and exogenous shocks on 1l.S. agriculture (Robinson. Hanson, and Kilkenny

676 S. Robinson et al.

he Aggregate Data, United States and Mexico

Mexico United States

176.7 4.847.4

I.760 19.990

16.5 7.9

10. I 0.4

5.8 I I.6 4.3 0.9

23.8 I.1

14.1 17.7

37.1 48.5

25 .o 32.7

100.0 100.0

49 162

84 246

.50~~w: GDP. per capzta GNP. and populatron data refer to 1988 and come from World Bank.

%“GwU T@FIOZI R~pun IYYO. All other data come from U.S. and Mexican social accounting

aa3mxs &tjef@ by Bufher. Thierfelder. and Hensoc (1392). I

on the two economies and their trade, which are used ~~cbrn~k or base solution of the FT’A-CGE model.

aller and poorer economy than the United States. exico and the United States is wider than that

n and Portugal and the European Community, a potential f trade integration.’ Mexico has a higher share

P than the United States and the W .S. market accounts rcent of Mexican exports. On the other hand, Mexico ut 3 percent of total U.S. exports. As is typical of a


E&pork $28.9 billion hporte: $25.0 billion

Exporte: W3.0 billion Importe: $512.2 billion

Expoti: $499.0 billion / Imports: $343.7 billion , 1 I

Figure 1. United States-Mexico trade, base run.

developing country, a much larger share of the Mexican labor force is rural: 23.8 percent compared with 1.1 percent for the United States.

Table 2 shows the sectoral structure of GDP, employment, and trade for the two countries, as well as existing trade barriers. The model’s 11 sectors include 4 farm and 1 food processing sector. The food corn sector refers to corn used for human consumption. In Mexico, this includes white corn, the small proportion of domestic yellow corn used for food, and No. 2 yellow corn imports from the United States, which are assumed to enter food use. In the United States, the food corn sector is defined as No. 2 yellow corn that is exported. The composition of the program crops sector corresponds to the other crops eligible for U.S. deficiency payments- feed corn, food grains, soybeans, and cotton. Other agriculture includes livestock, poultry, forestry and fishery, and other miscellaneous agriculture. The fruits and vegetables sector in Mexico includes beans, a major food crop.

The base year for Mexico is mostly 1988.4 For the United States, we use 1987 as the base year because of the severe contraction of agricultural output after the 1988 drought. Bilateral trade flows are from 1988. Because of the volatility in 1987-88 U.S. agricultural output) the model follows Adams and Higgs ( 1986) and Hertel ( 1998)

in the use of a “synthetic” base year for the United States, imposing 1988 United States-Mexican bi~atcra~ tra c flows on a 1987-base

Tab

le

2:

Sect

oral

St

ruct

ure

of

U.S

. an

d M

exic

an

Eco

nom

ies,

B

ase

Solu

tion

Com

mod

ity

Sec

tora

l sh

ares

(per

cent

) in

Bila

tera

l im

port

G

DP

E

mpl

oym

ent

Impo

rts

Exp

orts

ha

rrie

rs

Uni

ted

Sta

tes

Mex

ico

Uni

ted

Sta

ies

Mex

ico

Uni

ted

Sta

tes

Mex

ico

Uni

ted

Sta

tes

Mex

ico

Uni

ted

Sta

tes

Mex

ico

Food

co

rn

0.7

6.3

0.9

0.3

45.0

P

rogr

am c

rops

0.

5 I.

1 0.

3 5.

3 2.

9 3.

3 0.

1 12

.9

Frui

ts/v

eget

able

s 0.

2 I.

1 0.

4 3.

5 0.

1 0.

6 0.

4 3.

0 13

.2

12.5

O

ther

ag

ricu

lture

0.

8 5.

1 1.

4 8.

6 1.

3 I.

7 0.

4 3.

8 0.

6 8.

9 Fo

od

proc

essi

ng

1.7

6.2

1.5

2.5

5.2

2.2

2.9

3.6

3.8

8.2

Oth

er

light

m

anuf

actu

ring

4.

5 5.

5 5.

1 2.

7 4.

3 15

.0

7.0

6.0

4.7

8.1

Oil

and

retir

ving

2.

2 2.

9 0.

5 0.

5 5.

0 12

.0

2.7

10.2

1.

5 8.

8 In

term

edia

tes

5.6

8.2

4.5

3.2

16.8

13

.0

14.0

12

3

2.2

8.0

Con

sum

er

dura

bles

1.

9 2.

5 1.

7 0.

8 14

.5

28.3

10

.0

18.7

1.

8 12

.0

Cnp

ital

good

s 5.

2 3.

4 4.

9 2.

2 25

.6

24.6

31

.8

12.0

3.

6 12

.7

Serv

ices

77

.4

63.3

79

.6

64.4

23

.5

2.6

27.5

30

.3

Sour

ces:

U

.S.

and

Mex

ican

so

cial

ac

coun

tmg

mat

rice

s,

USD

A/E

RS;

19

87 f

or U

nite

d St

ates

an

d 19

88 f

or M

exic

o.

Bila

tera

l im

po

n

barr

iers

ar

e th

e co

mbi

ned

rate

of

tra

de-w

eigh

ted

tari

ffs

and

tari

ff

equi

vale

nts

of q

uota

s on

tr

ade

betw

een

Mex

ico

and

the

Uni

ted

Stat

es.

Perc

ent

com

posi

tion

colu

mns

su

m

to

100

perc

ent,

exce

pt

for

roun

ding

er

ror.

--

_I

--

- -

--

. -

_.

_-

- .

. _

.

.-

- .

---

-


United States economy. This approach has the advslntage of achieving a more representative U.S. base year with minimal adjustment to the data.’

2.A. el Structure

The core model follows the standard theoretical specification of trade-focused CGE models.6 Each sector produces a composite commodity that can be transformed according to a constant elasticity of transformation (CET) function into a commodity sold on the domestic market or as an export. Output is produced according to a CES production function in primary factors and fixed input-output coefficients fok intermediate inputs. The model simulates a market economy, with prices and quantities assumed to adjust to clear markets, All transactions in the circular flow of income are captured. Each country model traces the flow of income (starting with factor payments) from producers to households, government, and investors, and finally back to demand for goods in product markets.

Consumption, intermediate demand, government, and investment are the four components of domestic demand. Consumer demand is based on Cobb-Douglas utility functions, generating fixed expenditure shares. Households pay income taxes to the government and save a fixed proportion of their income. Intermediate demand is given by fixed input-output coefficients. Real government demand and real investment are fixed exogenously.

The model includes six primary factors and associated factor markets: rural labor, urban unskilled labor, urban skilled labor, profes- sional labor, capital, and agricultural land. In factor markets, full employment for all labor categories is assumed. Aggregate supplies are set exogenously. The model can incorporate different assumptions about factor mobility, including labor migration (discussed below). In the experiments reported here, we assume that agricultural land is immobile among crops, but that all other factors are intersectorally mobile, including capital.’ The results should be seen as reflecting

‘A comparison of 1987 and 19X8 United States-Mexico trade shows that Mexican farm imports

increased in 19X8 as U.S. agricultural output fell. Use of a 1987/W split year for the United

States moderates the import‘mce of Mexico in U.S. agricultural trade relative to 198X.

‘See Robinson, CI al. (199la) for a complete equation listing. Robinson (1989) and de Melo (1988) surveyed single-country. trade-focused CGE models. The FTA-CGE model is imple-

mented using the GAMS software, which is described in Brooke. Kendrick, and Meeraus t 1988). ‘However, labor markets arc segmented. Rural labor does not work irl the industrial sectors.

and urban labor does not work in aericulturc. Thrse labor markets are linked through separate JJrig'di;.bJ? tZqlJ:lJirpn~

680

adjustment in the long run, with capital sectors.

There are three key macro balances

S. Robinson et al.

ab!e to leave the agricultural

in each country mode!: the government deficit, aggregate investment and savings, and the balance of trade. Government savings represent the difference between revenue and spending; real spending is fixed exogenously but revenue depends on a variety of tax instruments. The government deficit is therefore determined endogenously. Real investment is set exogenously, and aggregate private savings are determined residually to achieve the nominal savmgs-investment balance.’ The balance of trade for each country (and hence foreign savings) is set exogenously, valued in world prices.

Each country mode! solves for relative domestic prices and factor returns that clear the factor and product markets, and for an equilibrium real exchange rate given the exogenous aggregate balance of trade in each country. The GDP deflator defines the numeraire price in each country mode!; the currency of the rest of the world defines the international numeraire. The mode! determines two equilibrium real exchange rates, one each for the United States and Mexico, which are measured with respect to the rest of the world. The cross rate (United States to Mexico) is implicitly determined by an arbitrage condition.

The mode! specifies sectora! export supply and import demand functions for each country and solves for a set of world prices that achieve equilibrium in world commodity markets. At the sectora! level, in each country, demanders differentiate goods by country of origin and ex- porters differentiate goods by country of destination. Four types of elasticity parameters are used ln the mode!. The production specification requires sectora! elasticities of substitution among primary factors. The CET export supply functions require elasticities of transformation between goods sold on the home and export markets. The import demand functions (described below) require sectora! income elasticities and substitution elasticities for home goods and for goods from each import source. We have drawn on estimates and “guesstimates” from various studies, including Hinojosa and Robin- son (19911, Hanson et a!. (19891, and Reinert and Shiells (1991). In lieu of econometric estimation, sensitivity analysis was carried out to check for the robustness of the mode! results using alternative elasticity parameters.

“Enterprise savings ratc5 ;IW asunlcd ICI adjl;ht to achieve the necessary level of aggregate

sirwngs in each country. This treatment is known BS Johanhen rnac‘ro closure

U.S.-MEXICO AGRlCULTi';PriL !dLICIES AND MIGRATION 681

2. emand Equations

The standard approach in Armington trade models is to specify an “import aggregation” function, usually a CES function.’ In the case of a multicountry model, the function is extended to include goods from many countries. In the CES case, the substitution elasticity is assumed to be the same for all pairwise comparisons of goods by

j country of origin. “’ The first-order conditions define import demand as a function of relative prices and the substitution elasticity.

As noted earlier, the use of CES functions in multicountry Armington trade models has led to empirical problems due to the restrictive nature of the CES functions. Instead of the CES import aggregation function, we use import demand equations based on the almost ideal demand system (AIDS). ” The AIDS function is a flexible functional form in that it can generate arbitrary values of substitution elasticities at a given set of prices. It also allows expenditure elasticities different from one.

In the AIDS approach, the expenditure shares are given by

S ,.I..< I =oL r.k.r I + c Y,.k.c I.,? l"g(pM,,k.,,) + fkk.<, log r2

(1)

where subscript i refers to sectors; subscript k refers to the United States and Mexico, and subscript c refers to the United States and Mexico, and the rest of the world. Si.k,c., is the expenditure share on imports of good i into country k from country cl. ci.k is nominal expenditure on composite good i in country ik, PMi,m.(.2 is the domestic price of imports, and P,,I; is the aggregate price of the composite good. The Greek letters are parameters.

We adopt the notation convention that when k = cl,

M 1.1.1 = D,.kr PM,,,,, = Ph. and &,,,I = PD,,k - D,.kfz’,,k

where Mi,k.r., is the import of good i into country k from country cl, D,,k is the domestically produced good sold on the domestic market,

‘The properties of CGE models incorporating CES import aggregation functions have been extensively studied. See, for example, de Melo and Robinson (1989) and Dev:arajan. Lewis.

and Robinson ( 1990). “‘Other generalizations of the CES function could allow different, but tixed. elasticities of substi-

tution between goods from different countries. See, for example, the Constant Ratios of Elasticities

of Substitution Homogeneous (CRESH) function described in Dixon et al, (1981). It is also common IO use nested CES functions, with a two-good CES function specifying substitution between domestically produced goods and a composite of imports, which is itself a CES function of goods

from *tiirious countries of origin. “‘I he AIDS specification in this model draws heavily on work by Robinson, Sotde. and

Wcyerbrock ( IWiR). The discussion hclow is based on their paper.


PD,,, is the price of D,.k, and the aggregate price index, P,.k, is defined as a translog price index (Deaton and Muellbauer , 1980). ”

Various restrictions on the parameters are required to have the system satisfy standard properties of demand systems such as symmetry, ho- mogeneity, adding up, and local concavity. We calibrated the parameters for the FTA-CGE model by starting from a set of expenditure elasticities and substitution elasticities for each sector in each country We assume that substitution elasticities are the same for goods from any pair of countries, so our AIDS functions are effectively simple extensions of the multi-country CES functions to include expenditure elasticities different from one.

2.C. Migration

The FTA-CGE model specifies three migration flows: rural Mexican to rural U.S. labor markets, urban unskilled Mexican to urban unskilled U.S. labor markets, and internal migration within Mexico from rural to unskilled urban labor markets. Migration is assumed to be a function of wage differentials across the linked labor markets. In equilibrium, international migration adjusts to maintain a specified ratio of real average wages, wgdf ,,,,, $, for each labor category in the two countries, measured in a common currency. Similarly, internal migration in Mex- ico maintains a specified ratio of average real wages between the rural and unskilled urban markets. The wage condition is given by

where the index mig refers to the three migration flows, WF,,,,,,* is the real average wagt, and EXR, is the exchange t-rite, which is included in the equations for international migration. The domestic labor supply in each skill category in each country is then adjusted by the migrant labor flow.

An implication of this specification of international migr,ation flows is that real wages measured in a common currency are equated, but they can grow at different rates measured in domestic currency. It is therefore possible to obscrvc migrants moving from a labor market where real wages are rising, in domestic currency terms, to one in

“Robmv)n CI al. ( 190 If-11 analyze the empirical properties of the different import aggrcgalion

U.S.-MEXICO &GRICULTI_'RAL POLICIES AND MIGRATION 683

which they are falling. e issue is in the specification of what mo- tivates international mi ts. For example, if they are motivated by the desire to accumul vings that they intend to repatriate, then migration will be sen o the exchange rate. On the other hand, if they are motivated ervations on relative changes within the two economies, base mestic prices, then migration could be expected to be insensitive to the exchange rate. On balance, we favor the first view, but feel that the model probably overstates the sensitivity of migration to changes in the exchange rate.

Migration flows generated by the FTA-CGE model refer to changes in migration from a base of zero. They should be seen as additional migration flows due to the policy change, adding to current flows. Current migration flows are substantial, both within Mexico and between Mexico and the United States.13 In addition, the net migration flows generated by the model represent workers, or heads of households. In recent years, a substantial share of migrants have been family members. The model thus probably understates total increased migration due to a policy change, because family members will tend to migrate with workers.

2.D. Agricultural Program

In both the United States and Mexico, the agricultural sector is characterized by a complex set of government interventions. These policies distort production. consumption. and trade. and involve significant fiscal expenditures 1 both countries. In 1988, Mexico’s farm progri6:n costs represented er half of total national expenditure on subsidies and equaled almost 1 percent of GDP. Mexican agricultural progmrn expenditure in the mode! base yeas totaled $1. ! bihn (T&k

3). I4 In the United States, deficiency payment expenditures in 1987 totaled $10.6 billion, or 1 rcent of govemnment spending and about 10 percent of the fiscal de

In the FI’A-CGE model, agricultural policies are modeled either as price wedges, which affect output decisions, or lump-sum income transfers. The wedges and transfers are either specified exogenously or determined endogenousl , based on the institutional characteristics

‘3Various researchers have placed t net increase of undocumented Mexican immigrants in

the United States to be around 100. a year during the 1980s. See Bean, Edmonston, and

Passe! (IWO). ‘JThls total represeuts agricultural qut subsidies for I and subsidies to food processi

for 1988. Mexican sgicultura grams and YadC policies havC changed dParnaticab’ 0~

past few yCar~. so it is im~“ort 10 use the latest data avai

684 S. Robinson ct al.

Table 3: Mexican Agricultural Program Expenditures

Subsidy

Other Food program Fruits and Other FOOd corn crops vegetables agriculture processing Total

Billion Pesos

Credit (CSUB) 126.8 137.5 5x.4 42.3 364.2

Fertilizer (FSUB) 55.8 147.2 17.4 220.3

Irrigation (IRSUB) 92.7 244.9 2x.9 366.5

Feed (FDSUB) 35.7 35.7

Direct payment (DSUB) 325. I 32s. 1

Price (PSUB) I ,030.3 1.08s. I Tortilla (LOSUB) 223.8 323.8

Total (billion pesos) 275.2 529.5 104.7 78.0 1.579.21 2.620.7

Total ($U.S. millionsi 121.2 233.3 46. I 34.4 695.7 l.lS4.S

Producer incentive

equivalent (PIE).

percent 7.9 0.0 2.1 0.2 0.4 -

SOWce: USDA/ERS ( IYY I ). Food corn refers to corn used for human consumption (85% of total corn output). Fruits and

vegetables include beans (frijoles). CSUB, FSUB. IRSUB. and FDSUB refer to 190 subsidies

in 1988 prices. DSUB. PSUB, and LOSUB refer to 1988. The PIE rates are given d rwlorenr. although they are modeled as specific subsidies (per unit output).

of the program being modeled (Table 5). Border policies (tariffs, quotas, and export subsidies) affect producers through their effect on the output price, PX,,k, which is effectively a weighted average of the prices of output sold in the domestic market, PDi,L, and in each export market, PEi.k,c.,. Similarly, they affect consumers through the price of the composite good, Pi,L, which is effectively a weighted average of the domestic currency price of the imported good, PMi,A, and the

Table 4: U.S. Agricultural Program Expenditures (in $ Billion)

Food corn

Other program

crops TOtal

Detici:*ncy payments 0.76 0.x.5 !0.62

Producer incentive equivalent (PIE), percent 25.7 20. x

Sourws: A~ri~~ulrurd 014rkwk. April I9Y I , and unpublished USDA data.

Food corn refers IO No. 2 yellow corn exports Ikticicncy payments for food corn are conrputcd from the share: of No 2 Lorn expclrts in tot;tl I1.S. corn oulput (I IV ). I’ll< rates arc given rrtl wlrwcwl. slthrwgh they arc rnodclctl ;I\ \lwc.ilic r;ltc\ Ilwr unit output).

LT.S.-MEXICO AGRICULTURAL POLICIES AND MIGRATION 685

Table 5: How Agricultural Programs at-e Modeled

Program instruments

Fixed Endogenous

Program

United States Mexico

Price wedges

PX (output price)

PD

PWM

itax

pSUb

tin

PVA

Income transfers Households

PIE Deficiency payment

program CSUB (credit)

FSUB (fertilizer) lRSUB (irrigation)

FDSUB (feed) DSUB (direct)

TM2

itax

Tariffs and quotds

on imports

psub

Tariffs and quotas

on imports

Value-added tax

losub

PIE refers IO producer incentive equivalent. In the model, the PIE variable equals the sum of all price-wedge instruments that affect the output price (PX). Tariff rates (tm) are fixed parameters.

The tariff equivzlem of a quota. TM?. is a variable determined endogenously. given the fixed

impori quota.

domestic good price, PD,.k. ” Given the CET and AIDS functions, the link between trade policy and domestic prices is weaker than in a mode! -where al! goods are perfect substitutes.

MEXKXN AGRICULTURAL PROGRAMS. Six Mexican policies are modeled. If’ In the four agricultural sectors, these are input subsidies, tariff:, ;Ind quotas. In the food processing sector, we model direct subsidies and price subsidies, in addition to tariffs and quotas. The sixth Mexican policy is the low-income, or tortilla, subsidy.

Mexico provides its farmers with input subsidies on credit, fertilizer, irrigation, and feed. Input subsidies are represented in the model as a per-unit markup on output price mcasurttd as a fixed number of pesos

“PX is a CIIT agprepalion of PD and PE, while P is a translog aggregation of Pi) and Phi.

“‘Mexican agricultural p4icics arc dcscnhcd in Krissol’f. Ncl’f. and Sharplcs ( lYY2). Grcnnes cl al. (IYYI ). Burfi\her (IYYZ). Mielkc (lYHY. IYYO), O’Maril and Ingc~ (IW)), I<ch~%\ iid M~clkc f IYXO). ;mtl IlSDA (IWI)


per unit of output. ” Reflecting their effect on the producer’s output decision, input subsidies are summed into 3 producer incentive equivalent (PIE) in pesos per unit of output. For the United States and Mexico, the producer incentive equivalents in ad valorem terms range from 0.2 to 26 percent (Table 3 and 4). Given the assumption of fixed input-output coefficients, the profit maximization problem uses the value added-price (PVA) in computing the marginal revenue product as an argument to determine demand for primary factors. PVA is the price received by producers (PX, defined net of indirect taxes) minus the cost of intermediate inputs (given by input-output coefficients, IOjJ plus all subsidies (PIE)

PVA,., = PX,., - 2 W’,,,,,+‘,,,) + PIE,., (3)

Increasing the producer’s PVA with a positive PIE increases factor returns and induces a resource pull of factors toward the subsidized sector, causing output in the sector to expand.

Import quotas in agriculture are used by the Mexican government as a supply management tool to maintain targeted domestic farm prices. Import licenses are generally issued after the domestic crop has been harvested and purchased. To acquire a license, private importers or Mexico’s food parastatal, the Compania National de Subsistencias Populares (CONASUPO), must show that domestic supplies are being purchased for not less than the government target price. The tariff equivalent of the quota is calculated as the price gap between the world price and the domestic price. Following Dervis, de Melo, and Robinson (1982) and Kilkenny (1991), the tariff equivalent is determined endogenously, so that the quota’s ad valorem equivalent (and hence the value to license holders of the import premia) changes with the price gap. 7% capture the significant changes in Mexico’s trade policies since 1988, quotas are assumed only in sectors still under quota in 1991; 1991 tariff rates are assumed in sectors in which import licenses have recently been removed.

Premium income from each sector is distributed to the holders of import licenses. Because only Mexicans are awarded licenses, the rent is retained domestically. In the FTA-CGE model, the rent is allocated between government and enterprises according to their shares of im-

“Input subsidies can be tied directly IO output because intermediate demand is modeled with

fixed input-output coefficients. With more complex production functions, input subsidies should

be directly tied to input usage. A “u” as the final letter in the name of a subsidy signifies that it is provided per unit of output.


ports. ‘* Tariffs are modelled with fixed cd valorem rates and tariff revenues are paid by consumers to the government.

Since December 198?!, Mexico controlled prices of almost all basic foods, including corn products, wheat products, dairy, eggs, poultry, and pork. To enable food processors to sell their output at fixed low consumer prices, the government offsets their high input costs with two types of subsidies. One is a direct subsidy (DSUB). This is a modeled as a fixed budgetary transfer from the government to the processing sector, with a unit value (DSUBU) given by

DSVBV = ~svB’ _- m (4)

The unit subsidy represents a wedge on the PVA, whereas the subsidy expenditure is treated as a fixed component of Mexican farm program expenditures.

The Mexican government also provides processors with an input price subsidy (PSIJB) to compensate them for the high purchase price of agricultural inputs and enable them to sell their output on the domestic market at the controlled retail price. Price subsidies increase a processor’s domestic sales price to PDA, the “actual*’ domestic sales price received by each producer:

PDA, = PD, + PSUBU, (5)

where PSUBU is the input price subsidy in pesos per unit of output. In a model with more sectoral disaggregation, the unit price subsidy

can be modeled endogenously as the price wedge on a processor’s domestic sales price that is required to maintain the fixed retail price of the composite good.” Because the 1 l-sector model aggregates air food processing into a single sector, PSUBU is represented as a fixed price wedge on the domestic sales price and consumer food prices are permitted to vary. To simulate the effect of the removal of farm sector quotas under an PTA on input prices in the processing sector, we simply remove the wedge rather than allow the model to determine the change in PSUBU. The cost of the price subsidy to the government increases with an increase in domestic sales and is included in agricultural program expenditures.

Mexico provides low-income consumers with s Gdized corn tor-

‘Tariffs and quotas are modeled identically for the United States and Mexico. except that in

the United States, quota premia accrue to capital income. “Mexico’s subsidies to the food processing sectors are modeled endogenously in Burfisher.

Robinson, and Thierfelder (1992), which includes distinct agricultural processing sectors.


tillas. Under one program, low prices are offered in CONASUPO- owned retail outlets located in low-income neighborhoods. More recently, the government has provided low-income households with one kilogram per day of tortillas, approximately half the daily average household consumption (Levy and van Wijnbergen, 1991). Because the FTA-CGE model has only a single aggregate household, with no differentiation by income, the tortilla subsidy is represented as a lump- sum income transfer to the single household. Similar to direct subsidies to processors, expenditure by the government on low-income corn subsidies, (LOSUB) is fixed and enters into Mexican agricultural program expend Htures .

U.S. AGRICULTURAL PROGRAMS. The U.S. deficiency payments program provides payments to farmers who participate in feed grain. wheat, rice, or cotton programs. Following Kilkenny and Robinson (1990) and Kilkenny (1991), we model the payment endogenously as the difference between a fixed target price (TP) and the market price (PX). The total payment a farm receives (DEFPAY) is the payment rate multiplied by eiigible base production (XP).20

DEFPAY, = (TP, - PX,),XP, (6)

We calculate the initial unit value of the deficiency payment from data on total government expenditures on deficiency payments, base output, and participation rates. The unit value of the deficiency payment is a component of the producer-incentive equivalent. We fix the eligible production at the base year levels, implying that any increase in U.S. program crop output comes from outstde the deficiency payments program. The total payment a farm receives is the payment rate multiplied by eligible base production2’

3.A. Scenarios

We analyze the effects of a United States-Mexico FI’A under six scenarios, which are summarized in Table 6. All the scenarios involve

“The initial value of the TP is calculated from base year data tin the aggregate cost of deticiency

payment which is then used to estimate the markup on the market price. The model also implicitly

fixes participation rates at the base year rate, implying that any increase in U.S. program crop

output comes from outsidc the deficiency payments program. In recent years. the market price

has been above the loan rate. so we have not had to model the nonrecourse loan program.

“Alternatively, the participation rate can be made endogenous; changes in deficiency payment

expenditures would result from changes both in the deficiency payment and base acreage.


Table 6: Description of Scenarios

No. Scenario Description

I.

2.

3.

3a.

4.

Industrial trade liberalization

Trade liberalization

Trade and Mexican agricultural

liberalization

Trade and Mexican corn liberalization

Trade liberalization plus common agricultural policies

Partial trade liberalization

Partial liberalization plus capital

growth in Mexico

Remove all nonagricultural bilateral tariffs and quotas.

Remove all bilateral tariffs and quotas.

Scenario 2 plus eliminate all agricultural

support programs in Mexico.

Scenario 2 plus eliminate input subsidies for

corn sector in Mexico.

Scenario 2 plus add a deficiency payment program for corn and other program crops

in Mexico.

Tariffication of Mexican agricultural import

quotas at 50 percent of tariff equivalent.

Add deficiency payment program to Mexico in corn. Leave all existing agricultural

programs intact.

Tariffication of Mexican agricultural import quotas at 50 percent of tariff equivalent.

Mexican capital stock IO percent higher. Mexican agricultural subsidies (PIE) cut in

half. No deficiency payment program in Mexico.

changes in policies between the United States and Mexico, leaving unchanged their policies with respect to the rest of the world. The first scenario is the removal of all nonagricultural protection between the United States and Mexico, leaving all agricultural protection and programs intact. The second scenario is removal of tariffs and quotas in all industrial and agricultural sectors, but again leaving all agricultural programs intact. ** A third scenario considers trade and agricultural liberalization, removing all Mexican subsidies to farmers and food processors, in addition to full trade liberalization. We also consider a variant of Scenario 3, 3a, in which input subsidies to the food corn sector is eliminated, but all other Mexican agricultural programs remain

*‘U.S. quotas on sugar imports (in the processed food sector) were also left intact. As a net sugar importer, Mexico is unlikely to become a sugar exporter to the United States under an FTA except insofar as it attempts to increase its quota rents from arbitrage sales to the U.S.

market. In any event, trade in sugar and items of high sugar content was not liberalized under the United States-Canada VA and is likely to be excluded from the United States-Mexico FTA.

S. Rnbinson et al.

urrh and fifth scenarios explore the effects of restruc- domestic agricultural policies in conjunction with an

scenario, in addition to fir11 United States-Mexico on, Mexico adopts a deficiency payments program for

crops similar to the U.S. program and continues its sing subsidies. This mix of policies has the effect

producers through domestic programs rather the fifth scenario, a corn deficiency payments

ined with the tariffication of Mexican agricultural level of the tariff equivalents of base year quotas. liberalization scenario reduces the fiscal burden of

ucers because some trade protection is main- o, the Mexican aggregate capital stock is

nted by 10 percent, simulating the effects of the capital inflows under an FTA.

ts the macro results from the six scenarios. GDP in both countries, except in Mexico under partial , in Mexico with free trade and the elimination of

rograms, and in the United States with Mex- _ Although there are efficiency gains from the

ons. migration effects are far larger and account for nes in GDP in some experiments.

ases in all the scenarios. Mexico increases its mted States and to the rest of the world. We find

os 2 to 5, an FTA results in very small trade diversion imports from the rest of the world decline by 1

FTA with Mexican growth (Scenario 6), however, of imports from the rest of the world. An FTA he United States, with U.S. imports increasing

the rest of the world in all scenarios. The 10 capital, combined with partial trade lib=

!o Mexico to increase 16 percent and version of U.S. exports away from the rest of the

li&xralization of Mexican agriculture are designed van Wijnbergen (19?41) in their single-country

7:

Agg

rega

te

Res

ults

of

a F

ree-

Tra

de A

rea

unde

r A

lter

nati

ve S

cena

rios

(P

erce

ntag

e C

hang

e fr

om B

ase

Mod

el

Solu

tion

)

1.

Indu

stry

3.

T

rade

4.

C

omm

on

5.

Par

tial

6.

G

row

th

trad

e 2.

A

U t

rade

pl

us a

ll

3a.

Tra

de

agri

cult

ural

tr

ade

plus

par

tial

li

bem

liza

tim

li

berp

lizr

stio

n ag

ricu

ltur

e pl

us c

orn

poli

cy

libe

rali

zati

on

libe

rali

zati

on

Rea

l G

DP

Uni

ted

Stat

es

M&

!XiC

O

Rea

l ex

chan

ge

s-at

e U

nite

d S

tate

s

Me

xic

o

!3xp

otts

(w

orld

pri

ces)

U

nite

d St

ates

to

Mex

ico

Uni

ted

Stat

es

to r

est

of w

orld

ex

ico

to U

nite

d St

ates

eX

iC0

t0

RS

t O

f W

Oi!

d

Rea

r of

wor

ld

to U

nite

d St

ates

R

est

of w

orld

to

Mex

ico

Rea

B w

ages

: U

nite

d St

alss

R

ural

U

un

skill

ed

cf

skill

ed

Prof

essi

onal

re

ntal

C

apita

l re

ntal

-0.1

7.8

0.1

0.2

0.2

-0.1

0.

1 0.

2 -

0.6

0.3

0.2

0.1

2.1

0.1 1.9

0.1

2.1

2.0

1.7

4.2

1.4

6.0

8.6

9.1

8.7

7.9

7.6

16.1

0.

1 0.

2 0.

3 0.

2 0.

1 0.

1 -0

.1

4.0

5.1

5.3

5.2

4.9

4.8

6.8

3.0

4.6

4.9

4.7

4.0

3.7

17.1

0.1

0.2

0.3

0.3

0.1

0.2

0.3

0.3

-0.7

-1

.0

-0.8

-0

6 -0

.3

6.4

- 1.

2 -

1.2

0.1

0.1

0.7

0.1

- 0.

4 -0

.4

-0.5

-0

.5

0. I

0.1

-0.1

-0.1

0.7

-0.2

-0.4

-0

.4

- I.0

- 1.

0 -

1.7

- 1.

7 0.

1 0.

1 1.0

0. I

0.7

0.4

0.4

--

_-

- -

_-

._

-_-

- .I

- -

~~

-.

- -

-.

_.

.-

- -

- -

Tab

le

7:

(con

tin/

4ed)

1. I

ndus

try

3.

Tra

de

4.

Com

mon

5.

P

arti

al

6.

Gro

wth

tr

ade

2.

All

tra

de

plus

all

3a

. T

rade

ag

ricu

ltur

al

trad

e pl

us p

arti

al

libe

rali

zati

on

libe

rali

zati

on

agri

cult

ure

plus

cor

n po

licy

li

bera

liza

tion

li

bera

liza

tion

Rea

l w

ages

: M

exic

o

Ru

ra!

Urb

an

un

skill

ed

Urb

an

skill

ed

Pro

fess

ion

al

Lan

d

ren

tal

Cap

ital

re

nta

l

Net

fa

rm

pro

gra

m

exp

end

itu

re

Un

ited

S

tate

s

Mex

ico

Mig

rati

on

(p

er

I .C

JOO

p

eop

le)

Mex

ican

ru

ral-

Un

ited

S

tate

s ru

ral

Mex

ican

u

rban

-Un

ited

S

tate

s u

rban

Mex

ican

ru

ral-

Mex

ican

u

rban

I.0

1.0

0.1

0.1

- 0.

4

0.2

0.3

0.3

0.3

0.4

0.6

0.3

7.6

22.0

-9

4.2

9.9

67.6

24

.8

- 8.

9

4 21

31

25

II

IO

8

142

406

685

5t?

0 13

0 18

8 -6

1

68

343

637

452

37

104

-94

1.0

0.1

0.8

1.6

1.0

0.1

0.8

1.6

0.3

-0.3

0.

3 0.

5

03

-0.3

0.

3 0.

5

-7.7

-

IS.8

-

10.3

0.

3

0.2

0.3

0.2

0.1

1.2

4.3

1.2

4.3

0.2

3.8

0.2

3.9

-2.1

3 3.

9

Th

e re

al

exch

ang

e ra

te

is t

he

pri

ce-l

evel

d

efla

ted

ex

chan

ge

rate

, u

sin

g

the

GD

P

def

lato

r.

A

po

siti

ve

chan

ge

rep

rese

nts

a

dep

reci

atio

n.

Exp

ort

s ar

e

valu

e at

w

orl

d

pri

ces

(in

d

olla

rs).

N

et

farm

p

rog

ram

ex

pen

dit

ure

eq

ual

s fa

rm

pro

gra

m

exp

end

itu

res

min

us

tari

ff

reve

nu

e a

nd

im

po

rt

qu

ota

p

rem

ium

reve

nu

e ac

cru

ing

to

th

e g

ove

rnm

ent

fro

m

agri

cult

ure

an

d

foo

d

pro

cess

ing

.

,, ,,

,,,

,,,

Tab

le

8:

Sec

tora

l R

esu

lts.

Sce

nar

ios

1 to

3a

(Per

cen

tage

Ch

ange

fro

m B

ase

Mod

el S

olu

tion

)

1. I

ndus

try

trad

e 2.

A

ll t

rade

S

. T

rade

an

d ai

l li

bera

liit

ion

libe

rali

itio

n ag

ricu

ltur

e 3.

T

rade

an

d co

rn

outp

ut

Exp

orts

ou

tput

E

xpor

ts

outp

ut

Exp

orts

O

utpu

t E

xpor

ts

Uni

ted

Stat

es

Food

cor

n 0.

2 6.

7 13

4.7

7.8

149.

8 7.

7 15

3.1

Prog

ram

C

rops

0.

1 0.

7 43

.7

I.0

69.4

0.

7 43

.3

Fmits

/veg

etab

les

0.1

- 2.

9 0.

4 9.

6 0.

6 7.

7 0.

5 9.

4 O

ther

ap

rkul

ture

0.

1 -2

.4

0.2

6.9

0.3

4.5

0.2

6.8

Food

pr

oces

sing

0.

1 7.

4 0.

1 6.

4 0.

3 6.

7 0.

2 6.

4 O

ther

lig

ht

man

ufac

turi

ng

0.1

5.8

0. I

5.

2 0.

2 5.

0 0.

2 5.

2 O

il an

d re

fini

ng

13.5

13

.4

13.2

13

.4

Inte

rmed

iate

s 0.

2 6.

7 0.

2 6.

0 0.

3 5.

5 0.

2 5.

9 C

onsu

mer

du

rabl

es

0.1

7.5

0.2

6.8

0.2

6.5

0.2

6.8

Cap

ital

good

5 0.

1 7.

2 0.

1 6.

5 0.

2 6.

0 0.

2 6.

5 Se

rvic

es

-0.6

0.

1 -0

.8

0.2

-0.8

02

-0

.8

Mex

ico

Food

co

rn

- 1.

3 -

13.1

-

19.3

-

18.9

Pr

ogra

m

crop

s -0

.4

- 5.

6 -

11.6

-5

.5

Frui

ts/v

eget

able

s 0.

4 1.

3 5.

2 18

.1

2.6

16.0

5.

3 18

.4

Oth

er

agri

cultu

re

-0.5

1.

3 3.

0 -

1.9

1.4

3.2

Food

pr

oces

sing

-0

.6

6.0

-0.5

7.

3 -3

.1

4.5

- 0.

5 7.

3 O

ther

lig

ht

man

ufac

tw%

g 0.

5 7.

2 0.

6 8.

1 0.

7 8.

6 0.

6 8.

2 O

il an

d re

fini

ng

3.8

3.9

4.0

3.9

Inte

rmed

iate

s I .

4 4.

1 1.

4 4.

8 1.

4 5.

3 1.

4 4.

9 C

onsu

mer

du

rabl

es

3.3

5.4

4.7

7.0

5.9

8.4

4.8

7.2

Cap

ital

good

s 2.

4 6.

7 2.

8 7.

7 3.

2 8.

4 2.

9 7.

8 Se

rvic

es

-0.3

0.

3 -0

.3

0.5

-0.2

0.

8 -0

.3

0.6

Rea

l ou

tput

an

d ex

port

s.

Exp

orts

ar

e to

the

par

tner

co

untr

y (U

nite

d St

ates

or

Mex

ico)

.

Tab

le

9:

Sect

oral

R

esul

ts,

Scen

ario

s 4

to 6

(Pe

rcen

tage

C

hang

e fr

om

Bas

e M

odel

So

lutio

n)

4.

Com

mon

eg

ricu

ltur

e 5.

P

arti

al t

rade

6.

G

row

th

and

part

id

poli

cy

libe

rali

zati

on

Sib

erak

atio

a

outp

ut

EX

pol-

tS

outp

ut

Exp

orts

ou

tput

E

xpor

ts

Uni

ted

Stat

es

Food

co

rn

4.7

!02.

8 2.

0 39

.6

3.6

82.9

pr

ogra

m

crop

s 0.

4 24

.4

0.5

37.4

0.

8 75

.5

Frui

ts/v

eget

able

s 0.

2 10

.6

0.3

10.3

0.

1 28

.7

Oth

er

agri

cultu

re

0.1

7.3

0.1

7.1

0.2

21.8

Fo

od

proc

essi

ng

6.5

6.8

-0.2

17

.9

Oth

er

light

m

anuf

actu

ring

0.

I

5.3

0.1

5.5

15.9

O

il an

d re

fini

ng

13.5

13

.4

22.

I in

term

edia

tes

0.2

6.3

0.2

6.3

0.3

12.5

C

onsu

mer

du

rabl

es

0.1

7.1

0.1

7.1

-0.1

10

.0

Cap

ital

good

s 0.

1 6.

8 0.

1 6.

8 0.

1 9.

5 Se

rvic

es

-0.8

0.

1 -0

.7

-0.1

M

exic

o 1.

1

Food

cor

n -

1.7

- 1.

7 -2

.9

Prog

ram

cr

ops

- 1.

9 -4

.9

-1.6

Fr

uits

/veg

etab

les

4.9

17.4

5.

0 17

.5

12.6

20

.3

Oth

er

agri

cultu

re

0.5

3.4

-0.1

2.

5 9.

2 7.

9 Fo

od

proc

essi

ng

7.8

-0.6

6.

6 9.

9 13

.8

Oth

er

light

m

anuf

actu

ring

0.

5 7.

7 0.

5 7.

6 9.

4 15

.5

Oil

and

refi

ning

3.

8 3.

8 -1

.1

Inte

rmed

iate

s 1.

4 4.

4 1.

4 4.

4 6.

7 9.

7 C

onsu

mer

du

rabl

es

3.9

6.1

4.0

6.2

14.1

C

apita

l go

ods

15.6

2.

6 7.

2 2.

6 7.

2 7.

6 13

.5

Serv

ices

-0

.4

0.3

-0.3

0.

4 6.

8 6.

6

Rea

l ou

tput

an

d ex

port

s.

Exp

or&

are

to

the

part

ner

coun

try

(Uni

ted

Slir

tes

or M

exic

o).


Sectoral results are given in Tables 8 and 9. U.S. export growth to Mexico is highest in the agricultural sectors where Mexican protection has been relatively high. In all five agricultural trade liberalization scenarios, U.S. export growth correspunds to a decline in Mexican food corn and program crop output. Full liberalization of the Mexican food corn sector (Scenarios 3 and 3a) results in a nearly 20 percent fail in output, whereas U.S. food corn output rises 8 percent and exports to Mexico soar by almost 150 percent. A policy mix in Mexico that maintains some trade barriers for agriculture (Scenario 5) results in much lower, although still significant, growth in U.S. exports of corn to Mexico.

Mexico’s fruit and vegetable sector undergoes less spectacular but significant export growth (ranging from 16 to 20 percent) under an FTA that includes agriculture, reflecting the high initial U.S. tariff rates in this sector. Agricultural trade liberalization leads to a significant increase in two-way trade in fruits and vegetables, with exports ex- panding in both countries. Mexican fruit and vegetable output expands significantly, while U.S. output expands only slightly.

3.C. Migration and Farm Program Expenditure

Complete trade liberalization and the removal of subsidies to Mex- ican agriculture and food industries (Scenario 3) has a major effect on migration. About 11 percent of Mexico’s rural labor force (668,000 workers) migrate either to the United States or to Mexican urban areas (Table 7). These workers come from the corn, program crop, and other agricultural sectors, which contract sharply with quota and program removal. Expansion of fruit and vegetable output, spurred by export growth to the United States, is insufficient to absorb the displaced agricultural workers. With the exception of the Mexican growth scenario, real wages of rural and urban unskilled workers fall in the Wnited States by up to 2 percent because of increased migration. In Mexico, rural and unskilled wages rise for the same reason

A total of 716,000 Mexican workers migrate to the United States, 3 I ,000 directly from the Mexican rural sector to the U.S. rural sector; another 685,000 urban unskilled migrants move to the United States from Mexican cities. There is a domino effect at work, with rural- urban migration in Mexico leading to migration from Mexican to U.S. urban areas. Isolating the impact of Mexican food corn liberalization

effects, generates more realistic international price effects than does the CES formulation used

in earlier multicountry trade models.

696 S. Robinson er al.

(Scenario 3a) indicates that about 70 percent of the total migration from Mexico to the United States associated with complete agricultural trade and program liberalization is due to liberalization of the corn sector.

Scenarios 4 and 5 were designed to ameliorate the impact of trade liberalization on Mexican migration. They are successful in reducing the migration flows, but they also increase Mexican agricultural program expenditures. Scenario 4, which adds a deficiency payments program in Mexico similar to that in the United States, supports corn and program crop output through domestic subsidies instead of quota protection. Mexican agricultural output falls only 2 to 5 percent in these sectors, but Mexican agricultural imports from the United States still increase sharply because removal of trade barriers lowers the relative price of imports. The deficiency payments program leads to a much smaller increase in migration, but incurs a 68 percent increase in Mexican net farm program expenditures (which take account of change in import tariff and premium revenues, as well as budgetary outlays).

Scenario 5, which replaces agricultural quotas with tariffs set at half of the tariff equivalents of base year quotas, supports the Mexican corn producer price and results in only a small contraction in output. Only 114,000 workers leave Mexican agriculture. While Scenarios 4 and 5 both reduce Mexican migration flows, the increase in Mexican agri- cf:!t~ural orogram expenditures is much lower when partial trade barriers arc maiitained (Sceiiario 5), increasing only 25 percent.

A comparison of Scenarios 3 to 5 illustrates the trade-offs among labor migration, farm program expenditure, and bilateral trade growth. Free trade in corn in particular can be expected to generate massive rural outmigration in Mexico. Continued support to Mexican corn through either subsidies or protection will reduce migration flows, but increase farm program costs and reduce growth of U.S. exports to Mexico.‘5

Scenario 6 results in a reduction in Mexican migration flows and only a slight contraction in output in the corn and program crops sectors (2 to 3 percent). Expansion of other sectors absorbs Mexican rural labor and eliminates any new net increases in the migration flow to the United States (indeed, reversing it by 53,QOO). There are significant real wage increases for all labor groups in Mexico. Mexic‘j’s net

l_i.S.-MEXKO AGRICt_‘LTCRAL POLICIES AND MIGRATION 697

Migration (1 CKXIs)

800-

200-

lOO-

-. -_

-100 -

-206

-300 -

-v---~ -~~- --~-

0% 5% 10% 15% 20% 25% Mexican capital stock gmwth

Figure 2. Migration and c+pitaI stock in Mexico.

F agricultural program t qenditures fall 9 percent, due both to decreased ;nput suhsidiex and -t *Icreased tariff revenue. This scenario indicates the importance of Mexican economic growth to the success of the FTA.

Figures 2 and 3 show the trade-offs between migration and Mexican agricultural program spending and growth. Both figures use the fLIl1 trade and Mexican agricultural program Bibcralization (Scenario 3) as their base. Figure 2 shows the sensitivity of different types of migration to increased growth. To counteract completely the increases in migration resulting from Scenario 3, Mexican capital stock would have to grow 17 percent relative to the United States. Figure 3 demonstrates the sensitivity of migration to spending on agricultural input subsidies. With a 100 percent reinstatement of input subsidies, the Bevel of migration un,.ier an FTA becomes that of Scenario 2, which is still significant. For both increased capital growth and agricultural support policies. the migration relationship is almost linear. Each percentage point increase in Mexican capital stock reduces migration to the United States by roughly 44,000 workers and each percentage in agricultural input subsidies reduces migration by 10.0


Migration (10003)

750 1

SO- From rural Mexico

5OO-

450-

I

20% I I

40% 60% Percent of base subsidy rate

I I

80% 100%

Figure 3. Migration and agricultural input subsidies in Mexico.

This article analyzes the effects of a United States-Mexico freetrade agreement using a multicountry CGE model in which labor migration and domestic agricultural programs are modeled explicitly. A flexible functional form is used for import demand equations. The model is used to analyze six scenarios. These represent complete bilateral trade liberalization and Mexican agricultural program elimination, two combinations of Mexican agricultural programs that would reduce the labor migration caused by an FTA, and trade liberalization with a capital inflow into Mexico.

Our results show that both countries gain under an FTA, even when some production and trade-distorting policies are maintained. Bilateral trade increases significantly with removal of trade barriers. An FTA creates trade for both countries in all scenarios, but some scenarios lead to trade diversion for Mexico, with slightly reduced imports from the rest of the world. As exico grows, however, its trade with both the United States and the st of the world

We show that alternative st As generate trade-offs


between the growth in exports that is stimulated by lower trade barriers vwsw the cust srlch liberalization generates in agricultural program expenditures and new net increases in labor migration flows. Free trade increases bilateral trade but induces large rural outmigration from Mex- ico. Mexico can reduce labor migration through the adoption of a deficiency payments program that maintains agricultural income, but the fiscal effects are prohibitive. Retaining some trade barriers in agriculture reduces bilateral trade growth, but also reduces migration and growth in agricultural program expenditures. Increased capital inflows into Mexico result in expanded bilateral trade, much lower migration flows, and a large reduction in Mexican agricultural program expenditures. Dynamic effects are clearly very important in achieving the full benefit of an FTA.

These findings suggest that Mexico will need a lengthy transition period and must allocate resources to agriculture during the transition. Trade liberalization leads to an immediate increase in rural outmigration, whereas the increased growth needed to absorb the displaced labor takes longer. The rapid introduction of free trade in agriculture and the elimination of agricultural support programs may not be dc- s’irable for either country when the social and economic costs associated with increased migration are weighed against the benefks of increased trade growth.

REFERENCES

Adams, P.D., and Higgs, P.J. (1986) Caiibratlon of Computable General Equilibrium Models from Synthetic Benchmark Equilibrium Data Sets. IMPACT preliminary working paper No. GP-57, Melbourne, Australia.

Alston, J.M., Carter, C.A., Green, R., and Pick, D. (1990) Whither Armington Trade Models? American Journal of Agricultural Economics 72(2$: 455467.

Bean, F.D., Edmonston, B., and Passe], J.S., Eds. (1990) Undocumented Migrarion fo the

United Stales: IRCA and rhe Experiences of the 1980s. Washington, DC: The Urban Institute Press.

Brooke, A., Kendrick, D., and Meeraus, A. (1988) GAMS: A User’s Guide. Redwood City, CA: The Scientific Press.

Brown, D. (1987) Tariffs, Terms of Trade and National Product Differentiation. Journal of

Potlicy Modeling 9: 503-526.

Burtisher. M.E. (1992) The Impact of a U.S.-Mexico Free Trade Agreement on Agriculture: A Computable General Equilibrium Model with Agricultural Trade Policies and Farm Pro- grams. PhD dissertation, University of Maryland, College Park, MD.

Burfisher, M..E., Thierfelder, K.E., and Hanson, K. (1992) Data Base for a Computable General Equilibrium Analysis of a U.S.-Mexico Free Trade Agreement. Washington, DC: U.S. Department of Agriculture.

Burfisher, M-E., Robinson, S., and Thierfelder, K.E. (1992) Agricultural Policies in a U.S.- Mexico Free Trade Agreement. North American Journal of Economics and Finance 3:

187-139.


Deaton, A., and Mueibauer. J. (1980) Economics and Consumer Behavior. Cambridge, UK:

Cambridge University Press. de Melo, J. (1988) Computable General Equilibrium Models for Trade Policy Analysis in

Developing Countries: A Survey. Journal of Policy Modeling IO: 469-503. de Mclo, J., and Robinson, S. (1989) Productivity and Externalities: Mou’eis of Export-Led

Growth. Washington, DC: World Bank. Dervis, K., de Melo, J., and Robinson, S. (1982) General Equilibrium Mode1.q for Development

Policy. Cambridge, UK: Cambridge University Press. Devarajan, S. Lewis, J.D., and Rob:nson, S. (1990) Policy Lessons from Trade Focused, TWO-

Sector Models. Journal of Policy Modeling 12:625 625-657. Dixon, P.B., Parmenter, B., Sutton, J., and Vincent, D. (1982) ORANI: A Multi-sector Model

of the Australian Economy. Amsterdam: North Holland. Goodloe, C., and Link J., (1991) The Relationship of a Canadian-U.S. Trade Agreement to a

Mexican-U.S. Trade Agreement. Paper presented at the XXII meeting of the International Association of Agricultural Economists, Tokyo, Japan, August 27, 1991.

Green, R., and Alston, J.M. (1990) Elasticities in AIDS Models. American Journul of Agri-

cultural Economics 72(2): 442445. Grennes, T., Krissoff. B., Sharples, J., Estrada, J.. Gardea, J., and Valdes, C. (1991) An

Analysis of a United States-Canada-Mexico Free Trade Agreement. International Ag- ricultural Trade Research Consortium commissioned paper. Department of Agricultural Economics, University of Minnesota, Minneapolis.

Hanson, K., Robinson, S., and Tokarick, S. (1989) United States Adjustment in the 1990’s: A CGE Analysis of Alternative Trade Strategies. Working paper No. 510, Department of Agricultural and Resource Economics, University of California, Berkeley.

Hertel, T. (1990) Applied General Equilibrium Analysis of Agricultural Policies. Staff paper #90-9, Department of Agricultural Economics, Purdue University.

Minojosa-Ojeda, R., and Robinson, S. (1991) Alternative Scenarios of U.S.-Mexico Integration: A Computable General Equilibrium Analysis. Working paper No. 609, Department of Agricultural and Resource Economics, University of California, Berkeley, Published in Spanish as Diversos escenarios de la integracicin de 10s Estados Unidos y Mexico: Znfoque de equilibro general computable. Economiu Mexicana l(1): 71-144 (1992).

Kilkenny, M. (1991) Compu:able General Equilibrium Modeling of Agricultural Policies: Doc- umentation of the 30-Sector FPGE GAMS Model of the United States. Staff report No. AGES 9125, US Department of Agriculture, Economic Research Service.

Kilkenny. M., and Robinson, S. (1988) Modeling the Removal of Production Incentive Distortions in the US Agricultural Sector. In Agriculture andGovernments in an Interdependent World

(A. Maunder and A. Valdes. Eds.). Aldershot, England: Dartmouth Publishing Co. Kilkenny, M., and Robinson, S. ( IWO) Computable General Equilibrium Analysis of Agricultural

Liberalization: Factor Mobility and Macro Closure. Journal of Policy Modeling 12: 527-

556.

Kravk 1.B.. Heston, A., and Summers, R. (1982) World Product and in ..me: International

Compurisons of Real Gross Product. Baltimore, MD: The Johns Hopkins University Press for the World Bank.

Krissoff. B., Neff, L., and Sharples. J. (1992) Estimated impacts of a Potential U.S.-Mexico Preferential Trading Agreement for the Agricultural Sectors. Mimeo. Washington, DC: U.S. Department of Agriculture, Economic Research Servi,,e.

Levy. S.. and van Wijnbergen, S. (1991) Agriculture in the Mexico-U.S. FreeTrade Agreement. Paper prepared for the CEPR-OECD Conference on International Dimensions to Structural Adjustment, Paris, France April 22 and 23, 1991.

U.S.-MEXICO AGRICULTURAL POLICIES AND MIGRATION

Mielke, M. (1989) Government Intervention in the Mexican Crop Sector. Staff report NO.

AGES89-40. US Department of Agriculture, Economic Research Service.

Mielke, M. (l%% The Mexican Wheat Market an& Trade Prospects. Staff report No. 9052. US Department of Agriculture, Economic Research Service.

O’Mara, G.T., and Ingco. M. (1990) MEXAGMKTS: A Model of Crop and Livestock Markets in Mexico. Working paper No. WPS 446. Washington, DC: World Bank. Policy Research

and External Affairs.

Reinert, K.. and Shiells, C. (1991) Trade Substitution Elasticities for Analysis of a North American Free Trade Area. Washington. DC: US International Trade Commission.

Roberts, D., and Mielke. M. (1986) Mexico: An Export Market Profile. FAER No. 220. US Department of Agriculture, Economic Research Service.

Robinson, S. (1989) Multisectoral Models. In Srinivassn. Eds. Handbook ofDevelopmenr Eco-

nomics (H. Chenery and T.N. Srinivason. Eds.). Amsterdam: North-Holland.

Robinson, S.. Hanson, K., and Kilkenny. M. (1990) The USDAIERS Computable General Equilibrium Model of the United States. Staff paper No. AGES9049. US Department of

Agriculture, Economic Research Service.

Robinson, S.. But-fisher, M.E.. Hinojosa-Ojeda. R. and Thierfelder. K. (1991a) Agricultural Policies and Migration in a U.S.-Mexico Free Trade Area: A Computable General Equi-

librium Analysis. Working paper No. 617. Department of Agricultural and Resource Economics, University of California. Berkeley.

Robinson, S., Soule. M., and Weyerbrock. S. (1991b) Import Demand Functions. Trade Volume, and Terms-of-Trade Effects in Multi-Country Trade Models. Unpublished manuscript. Department of Agricultural and Resource Economics, University of California at

Berkeley.

Summers. R. and Heston. A. (1991) The Penn World Table (Mark 5): An Expanded Set of Intematiolral Comparisons, 1950-1988. Quarrerly Journal ojEconomics 106: 327-368.

US Department of Agriculture, Economic Research Service. (1991). Estimates of Mexican

Producer and Consumer Subsidy Equivalents. US Department of Agriculture internal

document.

Agricultural policies and migration in a U.S.-Mexico free trade area: A computable general...

Documents

Transcript of Agricultural policies and migration in a U.S.-Mexico free trade area: A computable general...