PROTECTIONISM, FREE TRADE AND PREFERENTIAL TRADE: THE MEXICAN EXPERIENCE

by

Pablo Ruiz Nápoles

Graduate Program in Economics, Faculty of Economics, Universidad Nacional Autónoma de México email: ruizna@servidor.unam.mx Visiting Professor at Facultad Latinoamericana de Ciencias Sociales, sede México email: pruiz@flacso.edu.mx

1

PROTECTIONISM, FREE TRADE AND PREFERENTIAL TRADE: THE MEXICAN EXPERIENCE

Pablo Ruiz Nápoles JEL Classification: D57, F43, L60, and O54. Key words: free trade, exports, growth, employment. Abstract This paper is a theoretical and applied analysis of free trade policies and protectionism.

First, the evolution of free trade theories, vis-à-vis protectionist ideas, is described, starting

from the classical economists, their assumptions and implications for attaining welfare

Pareto optimality, full employment and growth. Secondly, a structural analysis of the

Mexican economy’s performance in three successive but different periods, regarding trade

policies for the last thirty-five years is presented. I conclude that extreme free-trade policies

have not been good for economic stability, growth and employment creation in Mexico.

2

INTRODUCTION The Heckscher-Ohlin (H-O) theorem, based on a General Equilibrium model, and the

Purchasing Power Parity (PPP) doctrine are considered the two pillars of the neoclassical

theory of international trade (Krueger, 1983), despite strong criticism from both theoretical

and empirical perspectives. The neoclassical theory of trade claims that free trade and

exchange rate flexibility are a means for achieving trade balance and Pareto optima in

production and consumption. These ideas belong to mainstream economics, which has been

taught as the leading economic paradigm in graduate schools of economics, political

economy and other social disciplines at all major universities around the world for the last

four or five decades. In contrast, active trade policies (protectionism) and fixed or regulated

exchange rates prevailed in most countries, developed and underdeveloped, for most of the

twentieth century. It was in the late seventies and early eighties when liberalism started to

prevail in economic policy both domestically, by reducing government expenses and

regulation, and internationally, by reducing tariffs and eliminating non-tariff barriers to

trade in goods and services.

However, neo-liberal policies have had mixed economic results. While they have

stopped inflation and reduced fiscal deficits, they have not produced economic growth, full

employment, and in some cases, not even balanced trade. This has particularly affected

underdeveloped countries, which have suffered from a lack of economic growth,

widespread unemployment, growing poverty and social inequality since long before these

policies were applied by their governments. Their situation seems to have worsened rather

than improved due to the application of said liberal policies in the last twenty years.

In Mexico, in the middle of the recession that resulted from the 1982 foreign debt

and foreign exchange crises, the government initiated a process of trade liberalization as

3

part of a set of liberal policies aimed at reducing state intervention in the economy. After

forty years of being a highly closed economy, there was an opening of the Mexican

economy as a result of a set of free market policies. The process, which was gradual in the

beginning, accelerated in the mid 1980s when Mexico joined GATT, and it peaked when

Mexico joined the North American Free Trade Agreement (NAFTA) in 1994.

The package of liberal measures was promoted in Mexico and in other developing

countries by the US government, the World Bank and the International Monetary Fund

(Sachs, 1988, p.77). The same type of reforms were also applied in certain Latin American

countries in the eighties and the nineties, under the assumption that mere deregulation

constituted the structural change needed to correct a distorted economy and increase the

level of employment and wages (Weller, 2000, p.13). These liberal policies, based on

exports, were recommended as a new strategy both for recovery and for growth (Sachs,

1988). The set of policies were in line with the “Washington Consensus” (Moreno-Brid, et

al., 2004-5).

This change in development strategy in favor of trade liberalization and state

downsizing is the most significant event in Mexico’s economic history in the last five

decades. It is also recognized that NAFTA was a main cause of the spike in exports and

foreign direct investment (FDI) that has taken place in Mexico. The effects of NAFTA after

ten years in effect have been thoroughly studied by experts, with both favorable and

unfavorable evaluations from the Mexican perspective (see, for example, Blecker, 2006;

Moreno-Brid et al., 2005; Puyana and Romero, 2005; Romalis, 2005; Casares and Sobarzo,

2004; Weintraub, 2004).

Some experts have interpreted NAFTA’s role in Mexico as a corollary of the

Washington Consensus’ set of liberalization, deregulation and privatization measures that

4

were adopted by the Mexican government (Blecker, 2006). This is not necessarily the case,

because while unrestricted, unilateral free trade has been part of trade liberalization policies

in most of Latin America, a three-country preferential trade area like NAFTA has different

operating rules, with different effects on trading partners.

When looking at the various trade policies in perspective, one may rightfully ask

how much they have helped the Mexican economy, considering three different policy

phases: protected trade, unilateral trade, and preferential trade agreements, such as NAFTA.

Thus, some of the questions that may arise are: what were the economic and political aims

of the liberal policies, and of NAFTA, for Mexico? And how well have these aims been

reached? We wish to know what –if anything– went wrong with the various liberal policies.

In this paper we try to address some of these questions by analyzing the

performance of the Mexican economy over three successive periods during the last 35

years: protected trade (1970-1981), unilateral free trade (1982-1993), and the preferential

trade agreement, NAFTA, (1994-2005).1 In the first section we present the main theoretical

arguments in favor of free trade versus protectionism. In particular, we analyze two aspects:

(1) the benefits of free trade and its beneficiaries, and (2) the conditions for Pareto

optimality under free trade. In the second section we present selected data on the Mexican

economy relating a key variable, exports, with other important aggregated variables:

product, employment and imports. The purpose is to compare economic results in the three

periods mentioned.

I. FREE TRADE VERSUS PROTECTIONISM: SOME THEORETICAL ARGUMENTS 1. Free trade arguments of the classical economists

1 Notice that protectionist policies were in effect in Mexico for about forty years, from the late forties to the mid-eighties.

5

When Adam Smith criticized the protectionist ideas of the Mercantilists in 1776,2 his main

argument –based on logical reasoning inspired by rationalism– was that foreign trade must

be balanced rather than positive for any given country, and that this outcome should be

obtained by free trade between trading partners, and not by state intervention through the

establishment of taxes or other restraints on imports. Smith also postulated that when the

trade balance is positive in any given period, currency (gold and silver) accumulates, so all

kinds of goods are not imported and thus, consumption and investment are hindered in

order to reach higher levels, which should only be determined by a country’s productive

capacity as well as by free competition, i.e. laissez faire, laissez passer (Smith, 1937, pp.

398-423).

The basis for the above arguments was that for Smith free trade within a nation –

free competition– is the only mechanism that guarantees the optimal use of each

individual’s abilities which, in turn, leads to the optimal public benefit, the “invisible hand”

argument. When this principle is applied to nations, rather than to individuals, it means that

free trade between nations leads to the highest attainable level of production for the trading

countries considered as a group, that is, the world (Smith, 1937, pp.423-4).

Ricardo criticized the Mercantilists’ ideas too, but on different grounds. For him it

was not through trade that economies grow, because, he argued, foreign trade does not by

itself create any value at all; what a country imports is equal in value to what it exports, and

both flows are always balanced in value terms, since whenever a balance is due, it was

always paid with gold or silver which at the time played the role of both money and

commodities. But for Ricardo there is an important exception to this argument, and that is

2 The protectionist practices prevailed in England since long before Thomas Mun’s ideas in the seventeenth century, until the nineteenth century.

6

when a country’s imports are wage goods, and they may be obtained cheaper abroad than at

home through foreign trade. These imports, in particular, would allow the profit rate to

increase without raising prices in all areas of production in the home country, which

stimulates accumulation and consequently growth, without affecting real wages (Ricardo,

1973, pp.77-80).

Ricardo’s theory of trade is based on the explicit assumption that there is no

mobility of factors of production between countries, so it cannot be assumed that factor

prices (or rates) are equal between countries, and consequently trade advantages cannot be

established on the basis of absolute price comparisons, but rather on the basis of relative

prices within each country. If free trade between any two countries were established on the

basis of relative advantages, after an initial process of Hume’s price-specie flow

movements, a balanced trade in commodities would eventually be reached, with benefits

for both countries. Curiously, in his famous two-country, two-commodity example, the

least productive country would benefit the most. In Ricardo’s argument, the highest level of

productivity, which means the lowest prices for both goods, can only be reached through

economic integration, i.e., free factor mobility between trading countries (Ricardo, 1973,

pp.81-83).

2. The neoclassical free trade theory

Neoclassical theory followed Smith and Ricardo in advocating free trade, both within and

between countries, but classical theories were adapted to fit in the marginal analysis

framework. Within this framework, international or interregional trade is a special case of a

more general theory of trade: the theory of supply and demand. Foreign trade would be the

particular case in which there is no mobility of factors of production between regions or

countries, whatever the cause may be (natural, cultural or political). The crucial effect of

7

the immobility of factors is the existence and prevalence of factor price differences between

countries, as opposed to the usually assumed equality of factor prices within an economy.3

So, even if we assume the same technology, tastes and factor distribution,4 between any

two countries, there will be trade between them if their relative factor endowments are

different and their relative prices too. This is the heart of the Heckscher-Ohlin theorem (see

Ohlin, 1933, p.33).5

But this is not exactly what Ricardo had in mind while explaining relative

advantages in trade. The Ricardian theory stresses the international differences in

technology in conjunction with international differences in real wage levels, while the H-O

theorem assumes the international identity of tastes and technology, tracing the origin of

trade to given differences in factor endowments (Dosi, et al., 1990).

The bottom line of the neoclassical argument is that comparative advantages for any

country in trade will show up only if there is free trade between trading partners. And, if

trade is conducted according to each country’s comparative advantages, it leads to each

country’s specialization in producing and exporting those goods in which its abundant

factor is used intensively relative to other goods and this, in turn, implies the equalization

of prices of all goods traded between countries, assuming free trade, zero transportation

costs and a flexible exchange rate between them. The end of this process is balanced trade

between trading partners.

3 In theory, factor prices would eventually be equalized between trading countries, under strict neoclassical assumptions and free trade (Samuelson, 1949). 4 Those three, plus factor endowments are the data for supply and demand conditions in a Walrasian general equilibrium model (see Hansen, 1971). 5 These are the data sets for supply and demand in the general equilibrium model, Ohlin assumes (1933).

8

At the end of the nineteen century, thanks to the works of, among others, Pareto,

Marshall, Edgeworth and Pigou,6 neoclassical economics –today regarded as mainstream

economics– began to consider the benefits of trade mainly from the point of view of

welfare.7 Thus, starting from a theoretical situation of a country in autarky and equilibrium,

the alternative of trade with other countries, either regulated or free trade, is to be evaluated

by how much welfare is lost and how much welfare is gained. In a classless microeconomic

analysis of static autarky versus with-trade situations, the only possible winners and losers

of welfare are consumers, producers and the government. There are usually two types of

analyses: general equilibrium and partial equilibrium. The closed economy assumptions

are: variable costs, perfect competition (equilibrium), full employment and full mobility of

factors between sectors. By further assuming the immobility of factors between countries,

the with-trade economy results in full employment, balanced trade, equilibrium, expanded

consumption and partial specialization of production, if there is a relative price change.

a) General Equilibrium and the benefits of Trade

The two goods, two factor, small country model, like the one shown in Figure 1, is

traditionally considered a good approach to illustrate the general equilibrium effects of

trade. There are two possibilities: autarky and trade.

Autarky’s equilibrium is at the tangency of the production possibilities (or

transformation) curve between good X and good Y, with the domestic price line p* at Q. The

meaning of this equilibrium is that the quantities of X and Y (Xp and Yp) produced and

consumed are the same, supply and demand are satisfied simultaneously, all factors of

production are being fully utilized, and a community indifference curve is also tangent to

6 See Arrow & Scitovsky (1969) and Corden (1974). 7 All American and European textbooks on international trade deal with trade policy from this viewpoint (see Heller, 1968; Södernsten, 1970; Kindleberger, 1973; Markusen et al.,1995; Krugman & Obstfeld, 2005).

9

that point, so that producers and consumers are at their best according to the Pareto

criterion.8

Figure 1 Open economy General Equilibrium

Source: Markusen, et al., 1995, p.55

Now, if we move from autarky to free trade two possibilities may occur: (1) the

country’s price ratio will be the same as the trading partner’s, and these prices are

represented by the price line p*, or (2) the trading partner’s price ratio is different from the

home country’s, so a new price ratio is established between trading partners, and this

implies a movement along the transformation curve to another production position.

Case (1) is described in Figure 1 where there is no price line change, but with-trade

consumption takes place at point C, reaching a higher community indifference curve. In

other words, all the benefits from free trade go to consumers; thereby social welfare is

optimized. The same mixed product (X and Y) is produced domestically, a part of it is

domestically consumed (segments 0–Xp and 0–Yc), the remaining part of Y (segment Yc–

Yp) is exchanged for the imported part of the X good consumption (segment Xp–Xc). This

case describes precisely the gains from exchange under free trade conditions. With-trade

optimization conditions for general equilibrium are defined as: 8 “A situation is efficient or Pareto-optimal if it is impossible to make one person better off except by making someone else worse off” (Layard and Walters, 1978, p. 7).

10

Producer optimization: MRT = px* / py

*

Consumer optimization: MRS = px* / py

*

Trade balance: px(Xc – Xp) + py*(Yc – Yp) = 0

where: MRT = Marginal Rate of Transformation (in production); MRS = Marginal Rate of

Substitution (in exchange); px* = foreign price of good X; py

* = foreign price of good Y.

In Figure 2 we have case (2) in which there is a price line change, moving the

production point from A (autarky production of X and Y) to Qt with-tax-trade production

and to Qf, with-free-trade production, points. It shows that the new price line shifts

production in the direction of partial specialization for the home country in good Y

production. This clearly favors the expansion of consumption to a higher indifference curve

tangent to the new price line p* at point Cf. By geometry, exports and imports can be

exchanged at world prices, keeping trade in balance. A tariff imposed on the imported good

would reduce the level of consumption of both X and Y, although this level would be higher

than in autarky, as shown in Figure 2, by means of community indifference curves Ut and

Uf, for social utility with tariff trade and with free trade, respectively.

In this case free trade is better than autarky and better than protected trade. It is

better in the welfare sense that the economy is, with free trade, at a maximum point of

efficiency, both in production and in consumption. The imposition of an import tariff to one

of the goods produced, imported and consumed, good X in this model as shown in Figure 2,

hinders movement further along the transformation curve, that is, from Qt to Qf, so as to get

more specialized on the greater comparative advantage good. It also prevents attaining the

highest possible indifference curve at Uf. Domestic production of good X is, in this case,

stimulated by a tariff on its imports.

11

Figure 2 Transformation Curve and Price Ratios

Autarky, Protected and Free Trade

Source: Markusen, et al., 1995, p. 247. By construction, the transformation curve of a two-good, two-factor, general

equilibrium model, like the one above, assumes full employment of the two factors of

production along all its points, and since the curve depicts the greatest possible mixed

production of both goods, any point on it is Pareto-optimal. The movement from autarky to

free trade is one along the transformation curve, so it does not imply any increase in total

production but rather a change, from one production line to the other, without getting full

specialization in either one. It is clear, therefore, that the improvement caused by this

change affects consumption entirely: a country may consume more of one good or of both,

by means of trading, than was previously the case under autarky. This is the meaning of

efficiency: each trading country, through a change in its domestic price ratio, will end up

producing and exporting the good which it produces best, in exchange for the good it does

not; full employment of both factors does not change at all. If we assume in this change that

12

prices move between trading countries to reflect reciprocal demand conditions, and they

always match with domestic supply conditions, equilibrium is maintained. Therefore, free

trade seems to be, in theory, a static “all win” game, as far as countries involved are

concerned.9

As we see it, there are two ways of measuring efficiency. On the one hand, if the

question is viewed from the subjective angle, it needs a community indifference curve map

for estimating consumption optimization after trade. But building up a community

indifference curve map requires interpersonal comparisons of preferences. This issue has

been seriously questioned, and is the subject of an endless debate, with no concrete results

(see Arrow & Scitovsky, 1969, Parts IV and V). On the other hand, gains in efficiency can

also be viewed as obtaining some goods relatively cheaper with free trade than without it.

The reason for this is that under free trade, they are either produced more efficiently at

home, or imported by exchanging them for efficiently produced export goods. Free trade is

supposed to make home producers in general more competitive by lowering their input

costs and by breaking up domestic monopolies. Trade policies have, from the point of view

of efficiency, two simultaneous targets: production and consumption. If we cannot establish

the best consumption point for all possible levels of efficient production and trade –and we

can’t without a social utility function– then we cannot be sure that free trade consumption is

Pareto-optimal. However, as long as the maximum efficiency in production is reached by

free trade, Pareto optimality is achieved. This has one particular implication: the increase of

social welfare derived from the increase in total production (the increase in employment,

the equalization of net social products and the equalization of domestic prices to marginal

costs) (Kaldor, 1938, p. 551). 9 Exactly as stated in Smith and Ricardo’s free trade theories (see infra).

13

On the other hand, as far as welfare is concerned, besides efficiency there is the

question of equity. This can be viewed from factor income distribution. For this distribution

to be considered Pareto-optimal, the distribution of income between factors of production

with free trade must be better, or at least as good as that which existed under autarky. If, in

the movement from autarky to free trade, one of the factors reduces its real income at the

expense of the other, then the free trade situation cannot be regarded as Pareto-optimal. In a

famous 1941 article, Wolfgang Stolper and Paul Samuelson demonstrated in a two-good,

two-factor model that free trade necessarily harms the relative scarce factor’s real income

while favoring the other’s (Stolper and Samuelson, 1970, p.261). However, the argument is

qualified by pointing out that the harm inflicted by free trade to one factor is less than the

gain of the other; that is, the real income of the abundant factor is relatively increased, so

instead of ceasing trade, it is better to compensate the suffering factor (Stolper and

Samuelson, 1970, p. 267).10

If, in case (2) illustrated in Figure 2, we look at the change in production from A to

Qf along the transformation curve, we see an increase in the production of Y and a decrease

in the production of X. If this change occurs due to comparative advantages, it means a

movement from scarce factor-intensive good X towards the abundant factor-intensive good

Y. This increased utilization of the relatively abundant, and therefore cheaper, factor of

production will necessarily make its price go up and, since we are assuming full

employment and equal factor prices across sectors, real income increases for the abundant

factor in both lines of production at the expense of the other factor. There is a change in

income distribution in favor of the pre-trade abundant factor. In the case of many countries

10 A subsidy for the suffering factor is thereby justified as a compensation for making the owner of the gaining factor better with free trade.

14

this factor is expected to be labor, of course. So it may be said that an increase in the real

wage bill could be expected for a labor abundant country, which opens up to free trade.11

The assumption of full employment seems crucial for getting to the point of

maximum social welfare, whether in autarky or a with-trade situation, for any given

country. This is so because all possible less than full employment (LFE) production points

are inside the full employment concave transformation curve in a two-good, two-factor

model. Of all possible LFE mixed production points, not one is Pareto-optimal by

definition, and only one of them is the actual point for a country in an LFE situation. In

short, within the neoclassical analysis, less than full employment is not compatible with

general equilibrium and Pareto-optimality.

Free trade in the two-good, two-factor general equilibrium model tends to stimulate

production and exports of those goods, which are factor-intensive on the relatively

abundant factor, on the condition that this factor is relatively cheap with respect to the

other, based on its relative abundance. Now, if we are in some pre-trade “equilibrium” –

supply and demand balance–, which happens to be an LFE position, the after-trade position

must be on the full employment transformation curve. In other words, efficiency implies

full employment. Greater efficiency in economics means cheaper goods, either produced at

home or imported. This, in turn, may stimulate saving and thus investment, in a

neoclassical equilibrium.

In summary, free trade generates greater efficiency than autarky: higher productivity

and lower prices, which translate into more savings and investment, as long as full

11 This is what was expected in all Latin American countries when they adopted trade liberalization measures in the eighties and nineties following the “Washington Consensus” package.

15

employment exists. Protected trade seems to be a situation that is somewhere between

autarky and free trade, so it could be a second best according to Pareto criteria.12

An export expansion will clearly be brought about by free trade, but since full

employment is assumed at the beginning, there can be no increase in overall production in

real terms, but rather a switch from domestic importable goods production to domestic

exportable goods production to pay for the increase in imports.

One important aspect of the market mechanism that is sometimes overlooked in

international trade and which is needed in order to arrive at Pareto-optimal results is the

fact that this mechanism implies money prices –not relative ones– being the signals for the

decisions of individual optimizing economic agents inside and across countries. The

original Ricardo’s theory of comparative advantages, in his two-country two-commodity

example, required that money prices showed clear advantages for either country’s

producers and consumers, so that both groups move in the right direction. According to

Ricardo, the most efficient country beginning with lower prices is paid for its exports with

money. The net inflow of money makes its prices to go up until one of the two absolute

advantages disappears. Simultaneously, the net outflow of money in the least efficient

country makes its prices go down until one of them is relatively lower so as to make the

good attractive for import from abroad. It is the money flow, which does a sort of

transformation from absolute into relative advantages (Shaikh, 1980, p. 205). In

neoclassical economics this same effect may be obtained by free exchange rate movements,

to the point of trade balance (Ohlin, 1933; Friedman, 1973).

12 The second best policy works when the first best is politically or institutionally impossible (see Kindleberger, 1973, p.200).

16

This, however, is different from the relative price change in the home country that

necessarily occurs when it opens up to trade. This change favors exports rather than

imports; that is, export prices should go up relative to import prices in money terms as

measured in the currency in which international transactions take place. This change is

expected from an increased total demand of exportable products (domestic plus foreign

demand) and an increased total supply of importables.

b) Partial Equilibrium, Free trade and Tariffs

In partial equilibrium (one good market), free trade means to directly compare domestic

prices to foreign prices in the same currency, thus it turns a country into either a net

exporter or a net importer of that good, according to whether its home price is lower or

higher than the foreign price for that particular good. In the case of a net importer’s

domestic market, foreign producers may eliminate domestic producers from their own

domestic market with lower prices. But even if they compete, home production is limited

according to how domestic demand responds to a price reduction caused by the opening of

that market to foreign competition.

The microeconomic theory of demand was created with specific assumptions

regarding consumer’s behavior in mind, so that the quantity demanded of a particular

homogeneous good in the market is a well behaved decreasing function of its money price,

that is, the demand curve. Any point along this curve is a consumers’ maximizing point.

Similarly, the supply curve that was built on, among other assumptions, the law of

diminishing factor returns, represents an increasing function of the supplied quantity of a

particular homogeneous good with respect to its money price; any point along this supply

curve represents for producers a profit maximizing (or cost minimizing) point. Given that

quantities and prices of supply and demand refer to same market at the same moment in

17

time it is, therefore, possible to find a point of intersection that simultaneously satisfies

consumers and producers, yielding maxima benefits for both groups. Still, there is a

consumer surplus in terms of welfare, for those consumers originally willing to pay a

higher price but who instead receive a lower one.

Figure 3

Welfare Surplus Before and After Trade

In Figure 3 welfare consumer surplus is illustrated, assuming pre-trade equilibrium

price P* and quantity Q*. Consumer surplus is defined as the total value to consumers of

their consumption of a good minus the amount they have to pay for it; this is the shadowed

area below the demand line, above the price line P* in diagram (b). If trade opening drives

the price down to P’, consumer welfare surplus is increased by areas R and T in diagram

(c). A part of this welfare gain for consumers is taken from local producers, so the net

welfare gain is the area T. This gain is due to the increase in total quantity supplied from Q*

to Q’ and the reduction of prices from P* to P’. In the case of a net importer country for a

particular good, consumer welfare surplus increases and producer welfare surplus decreases

in the domestic market, still there is a net welfare gain for local consumers.

18

Figure 4 Welfare Producer Surplus

Before and After Trade .

P

P*

Q* Q

S

(a)

P

P*

Q* Q

S

(b)

P

P*

Q* Q

S

(c)

R T

Q**

P**

In Figure 4 welfare producer surplus is illustrated. It is defined as the surplus in

revenue for those producers able to supply various quantities at prices lower than the

market equilibrium price. This is the total revenue received minus the cost of production.

This surplus is represented by the shadowed area above the supply line and below the

equilibrium price line in diagram (b). When trade opens up and it raises the price from P* to

PP

**, welfare producer surplus is increased by areas R and T in diagram (c) as total supply

increases from Q* to Q**. In short, for net exporting countries, welfare increases for

producers and is reduced for consumers in the domestic market of the good it exports.

There is a net gain for producers.

From the diagrams in Figures 3 and 4 it is clear that a change from autarky to free

trade produces welfare surplus gains to consumers by way of import expansion at lower

prices and to producers by the way of export expansion at higher prices. In the first case,

diagram (c) of Figure 3, the supply curve S moves along the demand curve D as it is

increased by new (foreign) producers entering the domestic market. In the second case it is

19

the demand curve D that moves along the supply curve S, indicating the increase in

(foreign) consumers.

If, for any reason, a tariff is imposed on goods that are either i) imported by the

home country, or ii) exported by the home country, welfare gains are reduced, as though

there had been a shift back to autarky. However, these welfare surpluses are not entirely

lost, but redistributed in favor of local producers and the government in the first case, and

foreign producers and the foreign government in the second one.

Figure 5 illustrates the general case of an import tax on a particular good market.

Lines S and D represent domestic supply and domestic demand, respectively of good X.

p*X is the free trade price, domestic production of good X is initially the segment 0–Q1,

and imports are represented by the segment Q1–Q2; the value of total consumption is the

rectangle 0, p*X, B, Q2. The consumer surplus for good X under free trade is the area β p*

X

B; correspondingly, the producer surplus is the area α p*X A. If the home country imposes a

tariff on X imports, this raises the price of X for domestic consumers from p*X to pX,

reduces total consumption from point B to point D, and imports are then limited to the

segment Q3–Q4, and domestic production is increased from Q1 to Q3.

The new consumer welfare surplus is reduced to the area β pX D. Part of the surplus

is transferred to local producers and the rest is transferred to the government in the form of

taxes. There has been a transfer of welfare surplus between the local economic agents –

consumers, producers and the government– with no profits or losses for the economy.

If we start from a situation where trade is protected and tariffs are removed as a

result of a new “liberal” trade policy, the graph in Figure 5 illustrates the same movements

as before but in a reverse direction. Now the supply curve moves to the right because there

is an increase of suppliers (local plus foreign), while the demand curve remains the same.

20

This will imply a new equilibrium point at a greater quantity Q2, and at a lower price per

unit, p*X.

Figure 5 Import Tax Effects in Partial Equilibrium

Source: Markusen, et al., 1995, p.283.

It also implies ipso facto a gain in welfare surpluses for consumers and a loss of

welfare surpluses for local producers and the government.13 One might assume that local

producers are also consumers not only of the same goods they produce, but are also

consumers of inputs for production, so that if there is a general removal of tariffs, local

producers are indirectly benefited as consumers.

The problem remains of how much losses and how much profits result from a free

trade situation after opening up the market for foreign competition, i.e. a tariff elimination

in that market, and this depends on the elasticity of the local demand curve for goods.

13 This means the displacement of local workers by foreign ones, not in terms of welfare but in terms of production, In other words, it means a drop in employment in all net importing local industries.

21

In the case of home country exports, there is a welfare surplus for producers as seen

in Figure 4, when tariffs are lifted, or the foreign market opens up to free trade. The extent

of this gain depends on the price elasticity of export supply.

This problem is similar to that of measuring the effects of an exchange rate’s

depreciation, on the trade balance, through the change in prices of imports and of exports.14

Since we do not know a priori the absolute values of price-elasticities of the demand for

imports and the supply of exports, there is no basis for assuming that a free trade situation

is, on average, better or worse than a protection situation.

The other complication for evaluating the possible benefits of free trade as opposed

to protectionism is related to the fact that export and import prices move not only in

response to tariffs, whether imposed or lifted, but assuming no technical change, they vary

directly with exchange rate movements, affecting ipso facto supply or demand curves,

depending upon the currency used to measure exports or imports.

For instance, if tariffs are lifted in all sectors of the home country economy, we

expect that prices of imported goods to go down in local currency for local consumers as in

the diagrams above, but if we assume a flexible exchange rate market and the exchange rate

goes up because more foreign exchange is demanded to pay for the additional imports, the

price per unit of every particular imported good will go up in local currency by the

percentage of the depreciation, since local producers have costs that are already higher than

foreign producers and are affected by depreciation, depending on the quantity of imported

inputs they have in their respective production line. In the end, depreciation turns out to be

a measure that only protects exporters, by transferring welfare benefits from local import

substituters to exporters. However, it is of course possible that the increase in exports 14 Formalized as the “Marshall-Lerner-Robinson Condition” (see Kindleberger, 1973, p. 328-9).

22

offsets this exchange rate rising tendency, if there is reciprocity in tariff reductions in the

trading partner country.

From the point of view of partial equilibrium, a free trade policy may, caeteris

paribus, guarantee a welfare net gain for home country consumers of imported goods, but

only for some, and they are those local consumers willing to pay a higher price and who

receive a lower one in the exchange. Who are these local consumers? In the case of

developing countries it seems they are not low-income consumers, given the

characteristically uneven income distribution. The opportunity cost of this welfare gain is a

welfare loss for local producers and for the government. In the case of local producers it

means the shutdown of plants and the firing of workers. It is assumed, however, that new

plants and jobs will be created in the exporting sectors as exports grow, and also in other

sectors by the expanded local inputs demanded from exporting sectors. The real net gain is

obtained only if, on average, employment increases are higher than job losses, and

aggregate demand grows faster than imports, under the new conditions produced by free

trade.

In partial or general equilibrium analysis one might expect that, based on

neoclassical orthodox theory, free trade will bring about greater efficiency in domestic

production and lower prices for domestic consumption, but economic growth is not

necessarily considered, and if there is full employment from the outset, there won’t be any

growth except for that created by increased productivity.

In other words, the theory is not wrong, but stubborn reality is not adequate for the

free trade theory implications. One other way to address this problem is to assess the

assumption that free trade will permit the optimization of production and consumption, and

ultimately of social welfare, if not investment and growth as well, for all trading parties. In

23

that respect, it seems that certain basic assumptions make it difficult for this theory to

produce the expected results if they are simply not fulfilled. In a famous article regarding

assumptions and theory, Friedman argued that it is not whether assumptions are realistic or

unrealistic which make associated theory correct but whether its predictions are proven

right. This is based on the idea that it is not important if economic agents follow rational

behavior in their individual economic decisions, but rather if the overall results were as if

these individuals had acted rationally (Friedman, 1953). If we apply this principle to free

trade theory, we should expect welfare improvements in production and consumption, in

both efficiency and equity, in any country engaged in free trade, regardless of whether this

country fulfills the necessary assumptions established by the theory.

3. New theories of trade and free trade

As mentioned above, the Comparative Costs theory based on the Heckscher-Ohlin theorem

has prevailed as the neoclassical theory of international trade, despite all criticisms since

the theorem’s publication in 1933. However, during the last fifty years, there have been

some departures from the traditional assumptions of neoclassical trade theory that have led

to particular theories of trade which, together, are known as the “New Theories” of

International Trade (Krugman, 1983).

The full employment assumption has been challenged by the “vent-for-surplus”

approach, which deals with the question of unused resources and trade patterns in

underdeveloped countries (Myint, 1958). Three other necessary assumptions of perfect

competition –an essential part of the neoclassical trade theory– have also been questioned:

the equal availability of technology across countries, the existence of identical production

functions with constant returns to scale, and the homogeneity of products. These criticisms

gave rise to new theories, which considered the following: technology gaps (Posner, 1961;

24

Dosi et al., 1990), the product cycle (Vernon, 1966) and increasing returns to scale

(Krugman, 1979).

The main implications of the comparative costs principle and the Heckscher-Ohlin

theorem have also been questioned. One criticism is related to the trade pattern

determination by relative factor proportions, and there are two critical aspects: the non-

coincidence of the expected factor intensities ranking, with export and import ranking

(Deardoff, 1979) and the absolute availability of resources as the determinant of trade

patterns (Kravis, 1956). There is another that is related to the expected intra-industry trade

as a result of specialization based on the Heckscher-Ohlin model, which contrasts the actual

growing importance of intra-industry trade (Dixit and Norman, 1980; Caves, 1981;

Krugman, 1981). However, most of these new theories have been made to “fit” the

orthodox theory by using the appropriate model (Gandolfo, 1987, I Ch.8).

In his early writings, Keynes was supportive of free trade on the grounds that

comparative advantage would lead to optimal specialization, but by 1930 he had changed

his views and considered the free trade doctrine no longer valid in the case of a high

unemployment economy (Milberg, 2002). Like some other experts, Keynes was critical of

the neoclassical assumption of full employment, which is basic to neoclassical comparative

costs theory. The importance of the criticism of this theory is that, “Under conditions of

persistent unemployment the theory of comparative advantage is irrelevant because the

mechanism which would otherwise transform a situation of differential comparative costs

into one of differences in absolute money costs and prices no longer operates” (Milberg,

2002, p. 241).

Raul Prebisch (1959), and other structuralists, questioned the neoclassical trade

theory on its implications for a free trade policy, in the case of underdeveloped countries.

25

Besides criticizing the full employment assumption, which does not apply in most

underdeveloped countries, Prebisch also tried to prove that free trade was not optimal in

these countries.

Prebisch begins by observing that underdeveloped countries produce and export

mainly commodities and import manufactured goods. Comparative costs and free trade

would confine these countries to this specialization and trade pattern. This situation has two

important implications for employment and growth. In the first place, technical progress,

recommended for increasing productivity in export activities, is a way to hasten economic

growth. Prebisch argues that if technical progress takes place in primary activities for

export and domestic consumption, but industry is not sufficiently developed to absorb the

redundant work force, the result will be increased unemployment. He instead favors

simultaneous industrialization and technical progress. The problem is then that the

comparative costs principle rules out the production of manufactured goods in these

countries. For Prebisch the usual neoclassical argument does not apply. Instead he argues,

“It is not really a question of comparing industrial costs with import prices but of

comparing the increase in income obtained in the expansion of industry with that which

could have been obtained in export activities, had the same productive resources been

employed there” (Prebisch, 1959, p.253).

Prebisch also observed that there is a tendency for underdeveloped countries’ terms

of trade to decline. He argued that this is caused by the type of products these countries

export and the type of products they import. Primary products are said to have a declining

income-elasticity of demand (inferior goods) while manufactured goods have rising income

elasticity (superior goods). Therefore, as income increases, the demand for commodities

does not rise as fast as the demand for manufactured goods. For those countries that highly

26

depend on imports for carrying out production and who export primary goods, this means a

major constraint for economic growth. Therefore, these countries are left with two

alternatives for their economies: to grow, running a trade deficit and thus a rising foreign

debt, or to maintain a foreign trade balance but a stagnant economy. This situation is

defined as the “external restriction” for growth. These two implications constituted

Prebisch’s arguments for justifying the establishment of protectionism and import

substitution policies in underdeveloped countries, to pursue both industrialization and

growth.

Joan Robinson (1973) criticized two of the main assumptions of the neoclassical

theory of trade: full employment and perfect competition, with their implied assumptions of

identical production functions across countries. Robinson concludes that the assumptions of

the neoclassical theory of trade exclude the discussion of any current trade issues,

especially those regarding free trade and investment in underdeveloped countries

(Robinson, 1973, p.221).

The history of international trade in capitalist economies and their trade policies

shows that no country in the world has committed itself to free trade on a permanent basis.

Instead, a wide variety of protectionist policies have been followed by today’s developed

countries like the US and European countries for many years as strategies for growth and

development (Chang, 2002).

In modern times the rationale for the second best policies –as they have been

called– has been that theories on full employment, perfect competition within and among

countries, fully developed sectors, and availability of resources, among others, simply do

not apply to existing economies. On the contrary, market distortions like unbreakable

oligopolies, income distribution disparities, factor prices not reflecting factor scarcities, and

27

gaps in technology prevail. (Corden, 1974), and it looks as though they will not disappear

because of freer trade. In other cases, the theoretical first best policy is politically or

institutionally impossible to apply (Kindleberger, 1973).

More recently, the existence and prevalence of increasing returns to scale for firms

and industries and of imperfect competition among firms and across countries has been

stressed in discussions of modern global trade, as opposed to perfect competition and

constant returns to scale. This calls for state intervention through establishing trade

strategies for growth and development (Krugman, 1987).

Nonetheless, in the eighties the World Bank, IMF officials and other policy advisers

for Latin America recommended the implementation of trade liberalization policies in

underdeveloped countries, hoping that freer trade would produce structural changes, such

as a switch in production from inefficient sectors to productive ones; that is, where relative

advantages are higher for the home country. These policies are synthesized in the

Washington Consensus package which includes privatization of state owned firms and

deregulation of economic processes in order to reduce costs for private firms and stimulate

competition (see Moreno-Brid et al., 2004-5). But developing countries in Latin America

are still far away from fulfilling the necessary conditions for benefiting from free trade, like

full employment and equilibrium. There are also great differences between developed and

underdeveloped countries, like technology gaps, infant industries, a lack of strategic

resources, etc. All of the aforementioned account for real asymmetries between developed

and underdeveloped trading partners, which cannot be dealt with without the use of trade

and industrial policy measures.

28

4. Export-led growth strategies

To see how foreign trade –though not necessarily free trade– produces growth and

increases employment, we have to use the Keynesian argument that in a model that takes

into account price stability and fiscal and external equilibrium, a dynamic factor that does

not create imbalance is foreign demand (export demand) (see Cornwall, 1977, ch.7;

Davidson, 1997). The increase in exports has a direct effect on the level of production and

employment, and it also expands the demand for intermediate goods and thereby causes an

indirect increase in employment. On the other hand, the foreign exchange inflows generated

by increased exports allows imports to increase by that same amount, without creating a

deficit. There exists, at least in theory, a virtuous circle in export-led growth (Cornwall,

1977, p.165). This kind of growth is supposed to increase productivity and, therefore,

competitiveness, which again reinforces the export trend. It is important to note that an

export led growth strategy is always based on manufacturing exports, because demand in

this sector is highly elastic with respect to income and prices in the global market.15

Recent historical evidence shows that export-led growth strategies have been very

successful in some Asian countries; however, none of these countries adopted full

liberalization policies simultaneously as a strategy to promote exports. On the contrary, the

East Asian experience shows that successful export-promoting policies have been regularly

accompanied by import controls and rigid regulations of the movement of capital (Sachs,

1988, p.78). And in South Korea and Taiwan, they were preceded by assorted industrial

policies (Rodrik, 1995). It has been argued that the outward orientation of growth in these

countries occurred as a consequence of an investment boom and not the other way around

15 This argument is coincidental with Prebisch (1959) who argued in favor of protectionist policies for industrialization in underdeveloped countries.

29

(Rodrik, 1995). The success of export-led growth strategies was clearly based not on free

trade but on regulated trade and industrial policies.

II. FREE TRADE VERSUS PROTECTION. SOME EXAMPLES FROM THE MEXICAN ECONOMY 1. The Mexican liberal strategy

In the early eighties, some experts in Mexico thought that manufacturing exports would

become the engine for growth in the rest of the economy, especially for increasing

employment. This, however, required changes in foreign exchange and foreign trade

policies that were in line with the aforementioned mainstream neo-liberal theories (see

Levy, 1981; Clavijo and Valdivieso, 1983). One of the main reasons for justifying the

abandonment of protectionism was that it was producing a bias against exports (Lustig,

1992).

Trade policy was changed so that exporters could import low cost, high quality

intermediate goods (inputs), in order to use the relatively cheap local labor to produce

manufacturing goods for export at a competitive level. However, such a strategy required

investment from abroad, and there were still obstacles to this at the end of the eighties.

The wage differential between Mexico and the US in manufacturing, along with

their geographical proximity, has always been an attraction for foreign firms. Nonetheless,

foreign investment was strongly limited and regulated, up to the eighties; and this is why it

did not represent an important share of total installed investment in Mexico.

Therefore, the opening process could not be completed until the late eighties or the

early nineties, when financial opening took place in Mexico, aided by a boom in capital

investment in emerging markets around the world. In effect, the liberalization of finance

and investment in Mexico completed the process of trade liberalization and stimulated

export-oriented manufacturing production.

30

However, at the end of the eighties unilateral free trade was not producing the

desired results. The clear failure of the neo-liberal economic policies led to a nationwide

public distrust in those policies by 1988. Thus, once the ruling party managed to win the

presidential election, thereby holding onto political power, it was a declared goal of the new

administration to deepen liberal reforms and to politically lock up the process through a

free trade agreement with the US. It took the form of a preferential trade agreement

including Canada which was signed in 1993 and which took effect in 1994. It was, of

course, the North American Free Trade Agreement, NAFTA.

Thus, in the last 35 years Mexico has experienced three different trade policies in

three subsequent periods: protected trade (1970-1981), unilateral free trade (1982-1993)

and preferential trade (NAFTA) (1994-2005). The results have also been different. Observe

the economic performance of the Mexican economy through selected economic indicators

in Table 1.

Table 1 Selected Economic Indicators

Average annual rates of growth in per cent Indicator 1970-1981 1982-1993 1994-2005 GDP in constant pesos 6.9 1.6 2.9 Population 2.6 2.6 1.4 Per capita GDP 4.1 -0.9 1.5 Total Exports (US$) 29.1 4.1 9.8 Oil Exports 103.5 -1.5 15.8 Manufacturing non maq. 19.8 17.0 13.0 Maquiladoras 27.2 18.2 13.6 Gross Fixed Investment 9.3 0.2 4.6 Paid Employment 4.8 1.3 1.5 Consumer Prices* 196.0 710.8 163.6 Terms of Trade (US$)* 28.3 -64.1 20.4 Real Exchange Rate Index* 30.3 5.3 1.6 Exchange Rate (Mex/US)* 78.3 696.7 155.1 *Acumulated variation in the period Sources: Instituto Nacional de Estadística Geografía e Informática INEGI and Banco de México, S. A., México.

31

Looking at Table 1, it is clear that the protected trade period (1970-81), according to

all of these indicators, was much better than the unilateral free trade period (1982-93). All

that was expected from trade liberalization in terms of export boom, investment, growth,

inflation reduction, and an improvement in the terms of trade, simply did not occur. The

neo-liberal policy package applied thus far clearly failed. In contrast, economic

performance improved during the NAFTA period. In some specific sectors, the effects of

the change from unilateral free trade to preferential trade were rather dramatic: the

manufacturing and maquiladora export boom (especially in the first half of the last period),

the reduction in the inflation rate and, consequently, in the currency depreciation rate, and

the improvement in the terms of trade. Nevertheless, the GDP growth rate, especially per

capita GDP, during the NAFTA period is lower than in the first period, and also compared

to what would be the growth rate needed to absorb the growing work force. Some of the

exports boom in the second half of the last period was due to oil exports, which have

benefited from increases in global prices. Employment also increased more in the protected

trade period than in the free trade period and the NAFTA period.

2. Exports and Growth

Figure 6 shows the trend of GDP growth rates in real terms and manufacturing exports –

including maquiladoras– in current US Dollars for the whole period 1970-2005. It is

striking to observe that there seems to be no correlation between these two variables,

during the late seventies, the eighties and part of the nineties. In fact, if we consider the

1982 to 1997 period, there is even a negative correlation. For example, whenever there is a

low or negative GDP growth rate, the growth rate in manufacturing exports is positive and

high. This can be explained by the fact that reductions in GDP growth were brought about

to a large extent by the adjustment programs put into effect by the government in 1982-83,

32

33

1986 and 1995, which included government spending cuts, currency depreciations,

domestic credit restrictions and real wage reductions. All of this affected domestic

consumption and investment but not manufacturing exports. This negative relationship

between exports and output can also be explained by the fact that, until the early 1990s,

most production was oriented towards the domestic market; only residual production was

exported, with the exception of oil and maquiladora products. This situation changed in the

NAFTA period during which, with the exception of 1995, export and output growth began

to correlate.

The increase in exports in the first half of the last period is, undoubtedly, the result

of an increasing US demand for imports, and also of the increased access to the US market

as a result of NAFTA. An important factor in the liberalization process was that, since

1992, the Mexican economy opened up its capital account, so that foreign investment of all

types, as well as Mexican capital flights, started to flow in and out of the country without

regulation. Some of the foreign direct investment probably matured around 1994 and some

plants started to operate, both export oriented and assembly plants (maquiladoras).

Table 2 Mexican Export Structure

Percentages 1970 1981 1993 2005 Total (FOB) 100.0 100.0 100.0 100.0 Oil products 2.5 62.5 4.8 14.9 Manufacturing 33.0 14.4 37.1 36.4 Maquiladoras 17.6 13.8 42.1 45.3 Other Primary 46.9 9.3 6.0 3.4 Sources: Instituto Nacional de Estadística, Geografía e Informática, INEGI and Banco de México, S. A., México.

Figure 6GDP and Manufacturing Exports 71-05

Annual Rates of Growth

-7.0

-5.0

-3.0

-1.0

1.0

3.0

5.0

7.0

9.019

71

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

-20.0

-10.0

0.0

10.0

20.0

30.0

40.0

50.0

60.0

Real GDP Manufacturing Exports

34

Source: Own elaboration with data from INEGI and Banco de México, S.A.

The data in Table 2 indicate that the composition of exports has changed

dramatically. The Mexican economy switched from commodities and oil exporting in the

seventies to manufacture exports in the nineties. In other words, one of the aims of free

trade policies seems to have been accomplished: structural change.

Table 3 Mexico’s Structure of Final Supply and Demand

in Mexican Pesos –Percentages–

1970 1981 1993 2005 Aggregate Supply 100.0 100.0 100.0 100.0 GDP 91.2 87.7 83.9 70.4 Imports 8.8 12.3 16.1 29.6 Aggregate Demand 100.0 100.0 100.0 100.0 Consumption 72.2 65.2 69.6 58.0 -Private 90.8 86.3 86.7 89.3 -Government 9.2 13.7 13.3 10.7 Gross Fixed Investment 18.2 23.2 15.6 14.6 Change of Inventories 2.5 2.0 2.0 0.3 Exports 7.1 9.6 12.8 27.1 Sources: Instituto Nacional de Estadística Geografía e Informática INEGI, and Banco de México, S. A.

In Table 3, the structure of aggregate demand shows that the export boom occurred

in the NAFTA period, in which the exports of goods and services represented 27 per cent of

total demand, more than two times greater than its share in 1993, the final year of the

unilateral free trade period. In contrast, consumption and investment have been declining

their respective shares of total demand. The cost paid for the export increases during the

free trade and NAFTA periods was the sharp increase of imports in those periods as per

centages of aggregate supply –greater than the export share of demand– while Gross

Domestic Product represented a lower percentage of total final supply. To summarize, a

structural change in foreign trade in Mexico was the main and only real achievement of the

free trade policy; other positive results have come from the preferential trade policy,

NAFTA.

35

Figure 7 Growth Rate and Trade Balance

Source: Own elaboration with data from INEGI and Banco de México, S.A.

Figure 7 illustrates how trade liberalization policies and the integration process via

NAFTA did not place Mexico on a path to real export-led growth. It shows that the

relationship between trade performance and economic growth has been deteriorating.

Between 1970 and 1981, Mexico’s real GDP expanded at an annual average rate of 6.9 per

cent and registered a trade deficit of 2.4 per cent of GDP. The international debt crisis and

the collapse of the oil boom resulted in an economic slowdown starting in 1982 and was

worsened by free trade policies; the average growth rate was 1.6 per cent in the period of

1982-1993, accompanied by a trade surplus of 2.1 per cent of GDP. In the NAFTA period

1994-2005 real GDP expanded at a 2.9 per cent annual average rate and registered a trade

deficit of minus 1 per cent of GDP. Therefore, with relatively greater amounts of foreign

resources as a proportion of GDP like it received in the seventies, the Mexican economy

was now able to grow, on average, at less than a half of the annual rates it experienced

between 1970 and 1981.

36

2. Exports and Employment The unemployment problem has undoubtedly been the most acute in the Mexican economy

for many years. Migration flows of Mexicans to the US, due to the lack of jobs in Mexico,

are by no means a new problem. During the NAFTA period these flows not only did not

stop but actually increased, despite the relatively high growth rate of the Mexican economy

from 1995 to 2000 (Cornelius, 2002).

The economically active population (EAP) in Mexico represents little more than

half of the population over12 years old and has been growing at a rate of 3.5 per cent a

year.16 The average number of new jobs demanded each year in the last 15 years is about

one million. The Mexican economy has not been able to generate this number in any given

year in this period. Hence, there is a cumulative job deficit, which is difficult to assess

accurately.17

Although the lack of jobs before this 15 year period was not any better, the problem

now is the higher amount of new labor people are looking for, and which they are not

finding which causes all sorts of social problems in Mexico as well as in the US, regarding

illegal migrant workers. Today the remittances of US dollars from Mexican workers (legal

and illegal) to their families in Mexico represent the second highest source of revenues in

the current account of Mexico’s balance of payments.

For these reasons, job creation has been a priority in the economic policy agenda of

many administrations in the Mexican government. Thus, the main idea for opening up the

16 In 1990 the figure of estimated EPA was about 24 million people and the number of new jobs required to keep these people occupied was then about 900 thousand per year. In 2000 the EPA was around 34 million people, so the number of new jobs needed per year was about one million 200 thousand (data from Censos Nacionales de Población y Vivienda, INEGI, México). 17 The data reported by official national surveys to the government agencies, includes the underemployed as employed people, therefore, it is useless as an indicator of the real job deficit.

37

economy was promoting exports and thereby creating jobs in the supposed comparative

advantage in labor-intensive sectors.

This process, however, implied an adjustment in the labor market. Despite the

neoclassical presumption that any job adjustment is automatic and smooth, it is widely

recognized that, in practice, a trade opening causes short-term unemployment (Cox and

Edwards, 1997, pp.8-9). Still, it is expected that, if free trade prevails in the medium and

long terms, comparative advantages appear, and in the case of countries with relatively

abundant labor like Mexico, employment will rise in labor-intensive industries and so

growth may be based on trade expansion and comparative advantages (Dowrick, 1997).

Thus far, the structural change in exports has not corresponded to change in the

total employment structure: while primary activities have declined, manufacturing has not

changed, and the only activities that have significantly increased their shares in total

employment have been construction and trading (mostly linked to domestic activities).

Table 4 Structure of Paid Labor by Sector in Mexico

–Percentages– Sector 1970-1981 1982-1993 1994-2004 Agriculture, forestry, and fishing 31.9 27.0 21.9 Mining and oil extraction 1.2 1.0 0.4 Manufacturing 13.5 12.5 12.4 Construction 8.0 9.7 12.2 Electric, gas, and sanitary services 0.3 0.5 0.6 Trade, hotels and restaurants 15.5 16.8 19.7 Transportation and communications 4.2 5.4 6.2 Financing, insurance and rentals 1.9 2.2 2.1 Personal and social services 23.5 24.9 24.5 Total 100.0 100.0 100.0 Source: Sistema de Cuentas Nacionales, Instituto Nacional de Estadística, Geografía e Informática, INEGI, México.

With the aim of measuring the specific effects of dynamic export activity on

domestic employment, we used an input-output model, which includes the technical

38

coefficients matrix, the direct employment vector, and the final demand vector (in this case,

of exports). The basic idea was to measure the impact of final demand on gross output,

estimate employment coefficients by industry and, using the estimated gross output,

calculate the employment generated by the level of gross output that is required by the

actual level of exports. Based on this method we obtained the gross output required to

produce the actual level of exports each year, and we could calculate the level of

employment in each industry corresponding to the level of gross output for the period of

1978-200018 (see Appendix 1). As a result of applying this model, we obtained the level of

employment associated with the gross output required to produce the actual level of exports

each year for said period.

The results shown in Figure 8 and Table 5 indicate that while total employment had

been growing at an average annual rate of 2.6 per cent during the whole period, in the free

trade and NAFTA periods, the level of employment associated with total exports has

increased at higher rates, so that the percentage of employment generated by exports had a

spike in 1995 and reached its highest level (15.4 per cent of the total) in 2000. Most of this

employment is domestic production associated with exports (11.2 per cent) while the

Maquiladora industry represents 4.2 per cent of total.

18 We have an unwanted restriction in the availability of data for other years, especially regarding input-output tables of the Mexican economy, which are available only for the years 1980, 1985, 1990, 1993 and 1996, and are all projections of the 1980 matrix.

39

Figure 8 Total Employment and Relative Shares of Employment Generated by Exports and Maquiladoras

0.0

5.0

10.0

15.0

20.0

25.0

30.0

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Mill

ions

of W

orke

rs

0.0

2.5

5.0

7.5

10.0

12.5

15.0

17.5

20.0

22.5

25.0

27.5

30.0

Perc

enta

ge o

f Tot

al

Total Employment Generated by Exports Employment in Maquiladoras Source: Own elaboration with data from the input-output model of Appendix 1

Table 5 Mexico: Estimated Employment Structure

Domestic Market and Exports 1980 1993 2000 Total, thousands of workers 21,356 26,040 30,613

Percent structure Total Employment 100.0 100.0 100.0 Domestic Market 95.5 91.1 84.6 Exports with Maquiladoras 4.5 8.9 15.4 Exports w/o Maquiladoras 4.0 6.9 11.2 - Direct 2.2 3.9 5.9 - Indirect 1.8 3.0 5.3 Maquiladoras 0.6 2.1 4.2 Source: Own elaboration based on data from Sistema de Cuentas Nacionales, INEGI, and Stata Matrix, México.

The relative share of export employment tripled from 1980 to 2000, which is by no

means a minor achievement. However, considering that total employment has not increased

40

to a significant extent and also that in 1995, when export employment grew, the total level

of employment diminished due to that year’s crisis, one can infer that export activities have

attracted jobs from domestic market activities, especially from those which were displaced

by imports.

Surprisingly enough, the ratio of jobs indirectly generated by exports to export

direct jobs –a little less than one in the three periods– shows that the multiplying effects of

exports on employment remains roughly the same regardless of the rupture of domestic

industry linkages caused by trade liberalization in the eighties.19

In order to see how much of this multiplying effect had actually been lost, we

estimated the level of gross output required for exports in the 1994-2000 period (the first

half of NAFTA) and the corresponding employment associated (directly and indirectly)

with exports with two matrices (from 1980 and 1993), and compared the obtained results.

The differences between the two estimates for gross output and employment generated by

exports are all positive and favorable for the 1980 matrix. It means that, if the economic

structure in the eighties had remained the same in the nineties, the level of employment and

gross output associated with exports would have been higher.

Table 6 Differences of Export gross output and employment as estimated with 1980 matrix,

from values estimated with 1993 matrix, Year Gross Output Employment 1994 6.14 2.07 1995 8.47 3.03 1996 8.52 3.46 1997 8.73 3.68 1998 10.24 4.26 1999 5.63 3.65 2000 5.62 3.69

Source: Own elaboration with data from INEGI, and CIESA.

19 This is an important issue that has been widely debated (see Ruiz-Nápoles, 2001 and 2004).

41

3. Exports and Imports

Mexico’s foreign trade (both exports and imports) increased notably since the trade opening

in the eighties, but it did not diversify, as is shown in Table 7. Most exports and imports are

concentrated in the North American market, that is, the US and Canada, even before

NAFTA.

From the initial period in 1970-1981, exports and imports tripled in the free trade

period 1982-1993 and increased five times for the period 1994-2005. The trade balance

improved in the free trade period when there was a cumulative trade surplus of eight billion

dollars, while it deteriorated for the NAFTA period, turning into a deficit of 65 billion

dollars.

The actual trend of the trade balance from the period of 1970-2005 is shown in

Figure 9. It must be observed that while there was a growing trade surplus with the North

American market as a whole, there was a growing deficit with the rest of the world in the

NAFTA period.

Table 7

Trade Balance of Mexico Millions of US Dollars

1970-1981 1982-1993 1994-2005 Total Exports 90,138.4 100.0 315,999.2 100.0 1,653,178.6 100.0 North America 54,518.5 60.5 233,417.3 73.9 1,472,096.5 89.0 Rest of the World 35,619.9 39.5 82,581.9 26.1 181,082.1 11.0Total Imports 114,293.0 100.0 307,963.1 100.0 1,718,581.6 100.0 North America 74,007.0 64.8 219,041.4 71.1 1,184,427.6 68.9 Rest of the World 40,286.1 35.2 88,921.7 28.9 534,154.1 31.1Balance -24,154.7 8,036.1 -65,403.1 North America -19,488.5 14,375.9 287,668.9 Rest of the World -4,666.2 -6,339.8 -353,072.0 Source: Own elaboration with data from Estadísticas de Comercio Exterior, Instituto Nacional De Estadística Geografía e Informática, INEGI, México.

42

Figure 9 Trade Balance of Mexico

Millions of US Dollars

-80,000

-60,000

-40,000

-20,000

0

20,000

40,000

60,000

80,00019

70

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

Total Trade Balance North America Rest of the World

It is precisely between 1994 and 2005 when the accumulated trade deficit with the

rest of the world reached the high figure of 353 billion dollars which offset the accumulated

trade surplus with the NAFTA area of 287 billion dollars, so there was an overall

accumulated trade deficit of little more than 65 billion dollars.

Most of the trade deficit in the NAFTA period comes from trading with Asian

countries, mainly Japan, China, Hong Kong, Taiwan and South Korea. This is surprising,

since the Mexican government has established and maintained high tariffs on the import of

consumer goods from these countries, with which there is no free trade agreement. These

growing Asian imports consist of growing inputs for manufacturing export industries and,

43

on the other hand, final goods imported –legally or illegally– as if they were coming from

the NAFTA area.20

4. Free trade and equality in Mexico

One of the desired results of free trade is an improvement in equality. As previously noted,

this takes place in the form of a redistribution of income among the owners of factors of

production, which, by necessity, implies a redistribution of income among population.

To be precise, the Pareto criterion applied for income distribution among factors

consists in improving one factor of production income without harming the other. This is

impossible in the Heckscher-Ohlin-Samuelson free trade model because if there is full

employment in both factors, the resulting (partial) specialization based on relative factor

endowment leads to increasing the abundant factor income in real terms, while

simultaneously reducing the scarce factor’s real income (Stolper and Samuelson, 1941). If

there is no full employment in both factors, which is the case for many Latin American

countries, we might expect that free trade would lead to an increase of both factor’s real

incomes, resulting from value added real increases. Yet these increases should be

distributed in a different way that in the pre free trade situation, the greatest increase going

to the abundant factor, and the smallest to the scarce one. In the case of Mexico the relative

abundant factor is, by all means, labor. This is why we expected total wages to increase

their share in the total value added in a free trade situation as the employment of labor is

increased.

In Mexico the proportion of wages to GDP (the wage bill) in real terms, shown in

Figure 9, was better in the protected trade period, 1970-81 (36.2 in average), than in the

20 There are of course no official data on these latter imports but they are sold in formal and informal markets in most cities around the country.

44

free trade period, 1982-93 (30.4) and in the preferential trade area period 1983-2004 (30.3).

In fact, the wage bill declined, though it was expected to increase as a result of the

expansion of exports and the overall increase in productivity. So we can say that free trade

policy in the case of a non-full employment country has actually harmed the abundant

factor of production in real terms (that is, labor).

Figure 9 Total wages as a proportion of GDP in real terms

Percentages

10

15

20

25

30

35

40

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

Perc

enta

ge

Total Wages/GDP

Source: Own elaboration based on data from Estadísticas de Cuentas Nacionales, INEGI, México.

In the same graph one can observe, however, that there was a sub-period (1988-

1994) in which the wage bill recovered in real terms. In this sub-period there was a signed

agreement between nationwide workers unions, chambers of commerce and trade, and the

federal government for controlling the increase of prices and wages, thereby stopping

inflation. It was a non-liberal policy that actually worked.

45

A measure of inequality, the Gini Coefficient,21 shows a trend that goes exactly in

the opposite direction of what was expected. This inequality coefficient, which fell from

1968 to 1984, shot up in 1985 and maintained an upward trend thereafter, according to

official data (Székely, 2005) and the graph in Figure 10 based on it.

Figure 10 Gini Inequality Coefficient 1968-2004

0.375

0.400

0.425

0.450

0.475

0.500

0.525

0.550

0.575

0.600

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

Source: Széleky, M., 2005, p.17.

5. Some contradictory results of the “structural change”

a) Production, labor and trade

To obtain the effects of labor and production shifts caused by free trade and NAFTA on the

structure of the Mexican economy, we analyzed the national income accounts for the 1970-

2004 period. We considered four variables: workers employed by industry, Gross Domestic

Product (GDP), exports and imports at constant 1980 prices by industry. We considered 72

industries, leaving out the federal government and military services industry. Then, we 21 The Gini Coefficient is a measure of inequality of distribution, defined as the ratio of the area between the Lorenz curve of the distribution and the curve of the uniform distribution, to the area under the uniform distribution. It is used to measure income inequality. It is a number between 0 and 1 where 0 corresponds to perfect equality and 1 corresponds to perfect inequality. The data and the graph for Mexico were taken from Székely (2005).

46

distinguished between tradable and non-tradable goods industries. The former were those

that have had exports or imports in any single year during the period of analysis. We

concentrated on these industries. Then, we calculated the number of workers per 1 million

of GDP at constant prices and its inverse, the amount of GDP produced per worker for the

whole period, but we chose only four years for comparison: 1970, 1980, 1990 and 2000.

After analyzing the ratio of workers per 1 million of GDP among industries, we divided

them in two groups: labor-intensive goods and input-intensive goods industries,

establishing as a limit two thousand workers per 1 million of GDP at constant prices, so

that any ratio above this limit was considered labor-intensive and below the limit, input-

intensive. The results of these calculations are shown in Table 8.

When the results of these four years were compared, we noticed some striking

results that seemed to contradict in principle what one would expect from a free trade

policy; that is, the expected shifts in production and labor predicted by the neoclassical

theory of trade: the Heckscher-Ohlin theorem.

If it is assumed that Mexico is more abundant in labor than in other factors of

production relative to the US, its main trading partner, one would expect relatively lower

wages in Mexico than in the US. As a consequence, if there is any comparative advantage

under free trade conditions, it should be in labor-intensive goods. So when free trade opens

up, comparative advantages appear, and Mexico should end up specializing in labor-

intensive goods production. Full specialization is ruled out by the assumption of variable

costs, but domestic production is expected to move towards labor-intensive goods and away

from input-intensive goods. Consequently, the export of labor-intensive goods is also

expected to grow, as is the amount of labor used in the production of these goods (both

47

exported and not exported), relative to the input-intensive production of goods, which is

supposed to decrease. An increase in the imports of these latter industries is expected.

The data in Table 8 shows that the number of tradable goods industries was constant

from 1970 to 2000 (61 out of 72), while their respective GDP declined from 40.2 per cent

in 1970 to 37.2 in 1980, and subsequently has stayed at 37.9 per cent in 1990 and 2000.

The number of workers employed in these industries has declined relatively from 52.2 per

cent of the total of workers in 1970 to 44.1 in 1980, 44 in 1990 and 37.5 per cent in 2000. It

seems strange that the biggest share of total employment today corresponds to non-tradable

goods industries (62.5 per cent), especially taking into account the structural change that

has occurred in the economy, and the fact that it is among the most open economies in the

world.

If we look at the differences between labor-intensive and input-intensive goods

industries, one can see that while the former group has kept the majority of employed

workers (90 per cent in average), it has decreased its share in the Gross Domestic Product

from 86.2 in 1970 to 52.5 in 2000. Also, this group of industries has diminished in number

from 48 in 1970 to 31 in 2000, and its share of exports from 85.9 per cent in 1970 to 37.7 in

2000. The opposite is also true for the input-intensive goods group which now produces

47.5 per cent of tradable goods GDP, and exports 62.3 per cent of the total, but absorbs

only 10.2 percent of that sector’s workforce.

48

Table 8 Relative Importance of Labor and Inputs Intensive Goods industries

Constant 1980 Pesos All Private Industries (72) 1970 % 1980 % 1990 % 2000 % Employed workers (thousands) 12,446.2 100.0 19,434.9 100.0 23,400.6 100.0 30,519.5 100.0GDP (millions) 2,302.3 100.0 4,382.8 100.0 5,248.5 100.0 7,421.6 100.0Workers per 1 million 5,406 4,434 4,684 4,112 GDP per worker 184.99 225.51 213.49 243.18

Tradable goods industries 61 61 61 61 Employed workers (thousands) 6,498.7 52.2 8,565.4 44.1 10,288.5 44.0 11,441.6 37.5GDP (millions) 925.5 40.2 1,630.7 37.2 1,991.0 37.9 2,812.8 37.9Exports (millions) 90.1 347.5 708.7 1,637.1 Imports CIF (millions) 173.3 445.9 584.6 1,791.9 Workers per 1 million 7,022 5,253 5,167 4,068 GDP per worker 142.41 190.38 193.53 245.84 Price Ratio Input-I/Labor-I 0.844 1.000 0.836 0.616 Labor-intensive industries 48 78.7 38 62.3 38 62.3 31 50.8Employed workers (thousands) 6,297.0 96.9 7,872.6 91.9 9,262.4 90.0 10,269.6 89.8GDP (millions) 797.4 86.2 1,133.5 69.5 1,226.9 61.6 1,476.6 52.5Exports (millions) 77.4 85.9 74.1 21.3 221.9 31.3 616.4 37.7Imports CIF (millions) 142.0 81.9 300.5 65.9 418.0 71.5 1,143.5 63.8Workers per 1 million 7,897 6,945 7,549 6,955 GDP per worker 126.63 143.98 132.46 143.79 Input-intensive industries 13 21.3 23 37.7 23 37.7 30 49.2Employed workers (thousands) 201.7 3.1 692.7 8.1 1,026.0 10.0 1,171.9 10.2GDP (millions) 128.1 13.8 497.2 30.5 764.1 38.4 1,336.2 47.5Exports (millions) 12.7 14.1 273.3 78.7 486.8 68.7 1,020.7 62.3Imports CIF (millions) 31.3 18.1 155.4 34.1 166.6 28.5 648.4 36.2Workers per 1 million 1,574 1,393 1,343 877 GDP per worker 635.16 717.72 744.74 1,140.13 Source: Own elaboration with data from Sistema de Cuentas Nacionales, Instituto Nacional de Estadística, Geografía e Informática, INEGI, México

In the year 1990 when Mexico was already open to freely trade its exports, the

figures for labor-intensive goods were 31.3 percent and 68.7 percent for input-intensive

goods, while imports were 71.5 percent of labor-intensive goods and 28.5 percent of input-

intensive goods.

GDP per worker –a gross measure of productivity– was five times higher in input-

intensive goods industries than in labor-intensive goods industries in 1970, and has

increased sevenfold in 2000. The changes in technology through time have affected both

49

groups, but this is more so the case for input-intensive goods, where the average number of

workers per 1 million of GDP has come down to 877, almost half of what it was in 1970.

If we take the trade balance by industry as a gross measure of revealed comparative

advantage, we observe that from 1980 to 2000, Mexico has had a positive balance in input-

intensive goods industries and a negative balance in labor-intensive goods industries.

So, according to the national income accounts data, we see that the Mexican

economy has been moving from labor-intensive activities to non-labor intensive activities,

and it is in these activities where comparative advantages are revealed. These two results

are in plain contradiction to what neoclassical trade theory predicts for free trade

conditions, if we assume that Mexico is labor abundant relative to its trading partners.

b) Production Possibilities, price ratios and mixed-production between industry groups

We also used the national income accounts data to elaborate a model of production

possibilities for the years 1970, 1980, 1990 and 2000. The real data were the implicit price

deflators of GDP by industry and the GDP in constant prices by industry. With these we

estimated the mixed output of labor-intensive and input-intensive, and the price ratio of

both groups of industries.

Using a standard equation for production possibility functions, taking the intercepts

of the price ratio lines, we calculated a fictional PP curve for each of the four years. We

established the domestic production-mix point in line with the price ratio between labor-

intensive and input-intensive industries, which lies in a less than full employment position,

that is, inside the concave PP curve for each of the four years (see Appendix 2). The

resulting graphs are shown in Figure 11. This Figure illustrates the data from Table 8

regarding the production-mix of input-intensive and labor-intensive industries in the years

1970, 1980, 1990 and 2000; they also show the price ratio of both groups of industries.

50

By comparing the data from 1970 and 1980, we noticed that relative prices became

more favorable for input-intensive goods in that period and that production also changed

relatively in that direction. This was the period in which there was a complex trade

protection system, and non-oil exports represented less than 5 percent of Aggregate

Demand. This confirms the previous findings of Boatler (1974), that in the seventies there

was a trend for capital-intensive exports to increase, which does not correspond to what

was expected according to the relative factor proportions of the Heckscher-Ohlin theorem.

Figure 11 Production Possibility Curves and Price Ratios

Millions of GDP in constant 1980 Pesos

P 1970

P 1980

P 1990

P 2000

0

250

500

750

1,000

1,250

1,500

1,750

2,000

2,250

2,500

0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 3,000 3,250 3,500 3,750

Labor Intensive GDP

Inpu

ts In

tens

ive

GDP

Source: Own elaboration with data from Sistema de Cuentas Nacionales, INEGI, México

But it was expected that when the Mexican economy was opened to free trade,

comparative advantages would show up in favor of labor-intensive goods, which would be

exported in great quantities, thereby stimulating growth, employment and wages. But if we

compare the PPC and Price Lines for 1980 and 1990, we can observe that although relative

prices moved in favor of labor-intensive goods, production moved towards input-intensive

51

goods and so did the trade balance, that is, the revealed comparative advantages. This was

the period of unilateral free trade, undervalued currency, high inflation and zero growth, the

so-called “lost decade” for many Latin American countries. A study of foreign trade

patterns in the manufacturing sector in Mexico in this period (Casar, 1989) shows that some

of the neoclassical assumptions of comparative costs theory and perfect competition do not

apply to the manufacturing sector: inter-industry trade is declining and intra-industry trade

is growing, while economies of scale, usually ruled out by assumption, are very important

in explaining trade advantages. In the next decade, from 1990 to 2000, the same production

trend of moving towards input-intensive industries not only remained the same but also

increased. GDP in these industries represents almost half of total GDP for tradable goods

(see Table 8).

These production shifts through time clearly show a trend towards more modern

(capital-intensive) production in Mexico, regardless of its relative abundance of labor; that

is, independently and even in opposition to what the H-O theorem dictates for a country

like Mexico.

CONCLUSIONS

After this long review of currently accepted theories about the welfare benefits of free trade

as opposed to protectionism for all periods and countries in the world, and after the analysis

of the performance of the Mexican economy in three different periods, each characterized

by a different trade policy, several conclusions can be drawn.

To get all the benefits derived from free trade, a country’s economy must already

have full employment and equilibrium. The welfare benefits of reducing tariffs are but a

redistribution of pre-existing welfare, from the local government and producers to

importers, be they high income consumers or producers. These welfare benefits, however,

52

can be offset by currency depreciation that operates in favor of exporters. In any case, the

welfare gains depend on the price-elasticity of demand in each market, which cannot be

predicted. The efficiency gains of free trade in overall consumption cannot be measured

without a community indifference map, whose very existence is still under debate. The

efficiency gains in production arise from the partial specialization of the economy in the

areas of production where it has comparative advantages derived from its relative factor

endowment, but these advantages must show up in terms of relative prices. In other words,

production will shift from sector A to sector B, only when the ratio of the prices of B to A

rise; that is, line of production B becomes attractive enough for producers to want to switch

over.

This point requires some additional remarks. It does not matter how we compare

local relative prices to foreign ones, the price of the comparative advantage local good must

be low relative to its foreign equivalent good, but it must rise when trade opens up, to a

point between the foreign price and the local one. And, according to the Law of One Price,

it will become equal for all trading countries when measured in a common currency,

assuming no transportation costs. If this process is generalized to all tradable goods

(exportable and importable), the expected result is that the ratio of the price of abundant

factor-intensive goods over the price of scarce factor-intensive goods in the local economy

must rise, and this means that its terms of trade will increase unless local currency

depreciation offsets these price changes. Still, production efficiency gains and partial

specialization do not signify growth or an increase in employment.

But under a Keynesian argument, economic growth and full employment can be

achieved by a market expansion brought about by exports. This, however, does not imply

either the use of free trade policies nor assumes full employment.

53

Absolute free trade has not prevailed in reality anywhere in the world, only in

certain underdeveloped countries and for short periods of time. Free trade policies have not

produced either export led-growth or full employment. In Mexico’s case, the economy’s

performance has been much worse with free trade than with protected trade in almost every

indicator studied. NAFTA has been beneficial to Mexico to the extent that it is not a

unilateral free trade agreement, but rather a limited free trade area agreement; in fact, it

could be improved if it were utilized in full, while compensating for the asymmetries

between Mexico and the US.

The movement towards “more efficient” production in Mexico is not the result of

neo-liberal policies, but of the general advance of technology and industrial production

throughout the world in the last three decades. In this sense, the neoclassical theory of

international trade (H-O theorem) is far behind the contemporary realities of trade. It does

not mean, though, that modern production and trading is better for achieving social ends

like full employment, better education, or more equality in income distribution (all of these,

with regards to economic growth, are considered development indicators). These are

objectives of political economy, which cannot be reached without state policy measures;

free market forces cannot achieve them. On the contrary, unregulated movement of capital,

goods, services and people between countries in the modern world may lead to unstable

markets in which the only beneficiaries are speculators.

54

APPENDIX 1: Input-Output analysis To estimate the level of Gross Output generated by aggregate demand, we start from

Leontief’s system of equations:

x = Ax + f (1)

where: x = gross output, A = technical coefficients matrix, f = final demand vector

(Pasinetti, 1977; Bulmer-Thomas, 1982; Dervis, et al, 1982; Aroche and Rupra, 1991), the

solution for which is:

x = (I – A) –1 f (2)

where: (I – A) –1is a matrix known as “Leontief’s inverse”.

A specific application of this model is:

x d = (I – Ad ) –1 (f d + e d) (3)

where: xd = domestic gross output, Ad = technical coefficients matrix for domestic

transactions, f d = domestic final demand vector and, e d = vector of domestically produced

exports (Bulmer Thomas, 1982; Dervis, et al, 1982).

If we split the demand according to its source, we have:

xdf = (I – Ad ) –1 f d (4)

where: xdf = domestic gross output exclusively associated with domestic demand, and

xde = (I – Ad) –1 e d (5)

where: xde = domestic gross output exclusively associated with exports.

We estimated equation (5) for Mexico with annual values from 1978 to 2000, taking

as Ad the technical coefficients matrix for domestic transactions from 1975, 1980, 1985,

1990, 1993 and 1996. The matrix year was the midyear for these five-year periods, from

1978 to 1992; 1993 and 1994 were estimated with the 1993 matrix; and the period of 1995-

55

2000 was estimated with the 1996 matrix. The variable ed was the vector of exports of

goods and services, not including maquiladoras, in real terms for the period 1978-2000.

The first step is to obtain the labor coefficients vector, according to the following

equation:

λ = n Ŷ -1 (6)

where: λ = labor coefficients vector, n = employment by industry vector, Ŷ = diagonal

matrix of gross output by industry,

Ŷ 1 = x d (7)

where: 1 = unit vector of order m, and x d is determined by equation (3).

The estimated coefficients express in each industry the following ratio:

λi = ni / yi

where: ni = employment of industry i, yi = gross output in industry i. λ is the vector of

industry labor coefficients, where i = 1,2,..., m.

Labor generated by exports in each industry is calculated by the following equation:

ne = λ Ŷe (8)

where: ne = vector of industry employment associated to gross output generated by exports,

λ = vector of labor coefficients, estimated by equation (6) and, Ŷe = diagonal matrix of

gross output generated by exports estimated by equation (5).

Export direct employment is a vector estimated by:

le = λ Ê (9)

where: le = direct employment associated to exports, λ = vector of labor coefficients,

estimated by equation (6) and, Ê = diagonal matrix of exports by industry,

Ê 1 = e d (10)

56

APPENDIX 2: Production Possibilities Curves and Price Ratios

The general equation for the PP curve22 is:

Y 2 + a X 2– b = 0 (1)

where: Y = Inputs intensive industries real GDP; X = Labor-intensive industries real GDP;

and b = highest production of Y at X = 0, determined by the price ratio of input-intensive

industries implicit GDP deflator/Labor-intensive industries implicit GDP deflator. The

equation for PP curves derived from equation (1) above was:

2aXbY −= (2)

where: Y = Inputs intensive industries real GDP, X = labor-intensive industries real GDP.

The estimated values for a,and b are:

Year A b 1970 0.7369 665,000 1980 1.0000 2,659,182 1990 0.6988 3,203,384 2000 0.3795 5,043,617

The straight line price ratio was calculated with the equation:

Y = c – dX (3)

where: Y = inputs intensive industries real GDP, X = labor-intensive industries real GDP,

c = highest value of GDP of inputs-intensive industries, and d = price ratio of input-

intensive/labor-intensive industries. The estimated values of c and d are:

Year c d 1970 801.1 0.844 1980 1,630.7 1.000 1990 1,789.8 0.836 2000 2,245.8 0.616

22 See Layard and Walters, 1978, Ch. I.

57

DATA SOURCES: Input-Output, domestic transactions matrices for Mexico, 72 entries, for years 1980 and 1985, from: Instituto Nacional de Estadística, Geografía e Informática, INEGI, Sistema de Cuentas Nacionales, México. Input-Output, domestic transactions matrices for Mexico, 72 entries for years 1990, 1993 and 1996 from: Consultoría Internacional Especializada, S.A. de C.V.(CIESA), Stata Matrix, Versions 1.0 and 2.0, 1994 y 1998, México. Other data from: Indicadores Económicos y Financieros, Banco de México, S. A., www.banxico.org.mx Sistema de Cuentas Nacionales de México, Instituto Nacional de Estadística Geografía e

Informática, INEGI, México, www.inegi.gob.mx Censos Nacionales de Población y Vivienda, Instituto Nacional de Estadística Geografía e

Informática, INEGI, México, www.inegi.gob.mx REFERENCES Aroche, F., and Rupra, I., 1991, “Comercio y empleo: el caso mexicano,” Investigación

Económica 195, Jan-Mar, pp. 21-42. Arrow, K., and T. Scitovsky, 1969, Readings in Welfare Economics, Irwin, Homewood Ill. Blecker, Robert A., 2006, “Macroeconomic and Structural Constraints on Export-Led

Growth in Mexico,” unpublished paper prepared for the 2006 LASA Conference, San Juan, Puerto Rico. Center for North American Studies, American University, Washington, D. C.

Bulmer-Thomas, Victor, 1982, Input-Output Analysis in Developing Countries, Sources,

Methods and Applications, John Wiley & Sons, London. Boatler, Robert W., 1974, “Las predicciones de la teoría del comercio internacional y el

crecimiento de las exportaciones manufacturadas de México,” El Trimestre Económico, No.164, Oct-Dec, pp. 811-841.

Casar, José I., 1989, Transformación en el patrón de especialización y comercio exterior

del sector manufacturero mexicano 1978-1987, Instituto Latinoamericano de Estudios Transnacionales ILET, Nacional Financiera, Mexico.

Casares, E., R. and H., Sobarzo, 2004, Diez años del TLCAN en México: una perspectiva

analítica, Fondo de Cultura Económica, México.

58

Caves, Richard E., 1981, “Intra-industry Trade and Market Structrure in the Industrial Countries,” Oxford Economic Papers, No. 33, pp.203-223.

Chang, Ha-Joon, 2002, Kicking Away the Ladder Development Strategy in Historical

Perspective, Anthem Press, London. Clavijo, F., and S. Valdivieso, 1983, “La creación de empleos mediante el comercio

exterior, el caso de México,” El Trimestre Económico 50, No. 2, Apr-Jun, pp. 873-916.

Corden, William M., 1974, Trade Policy and Economic Welfare, Clarendon Press, Oxford. Cornelius, Wayne, 2002, “Impacts of NAFTA on Mexico-to-US Migration,” in Chambers,

E. J. and Smith, P. (eds.) NAFTA in the New Millennium, The University of Alberta Press/Center for US-Mexican Studies, La Jolla, San Diego, Cal., ch.12, pp. 287-304.

Cornwall, J., 1977, Modern Capitalism: Its Growth and Transformation, St. Martin Press,

New York. Cox, A., and S. Edwards, 1997, “Trade liberalization and unemployment: Policy issues and

evidence from Chile” in Borkakoti J. and Milner Ch. (eds.), International Trade and Labour Markets, St. Martin Press, New York.

Davidson, Paul, 1997, “The General Theory in an Open Economy Context.” In Hartcourt,

G. and Riach, P. A. (eds.), A ‘second edition’ of The General Theory vol. 2, Routledge, London.

Deardoff, Alan V., 1979, “Weak Links in the Chain of Comparative Advantage,” Journal

of International Economics 9, no.2, May, pp.197-209, reprinted in J.N. Bhagwati (ed.), International Trade: Selected Readings, The MIT Press, Cambridge, MA., 1981, pp.76-87.

Dervis, K., J. De Melo and S. Robinson, 1982, General Equilibrium Models for

Development Policy, The World Bank Research Publications, Washington, D.C. Dixit, A. and V. Norman, 1980, Theory of International Trade, Oxford University Press Dowrick, Steve, 1997, “Trade and growth: a survey” in Fagerberg, J., et al. (eds.),

Technology and International Trade, Edward Elgar, Cheltenham, U.K. Dosi, G., K. Pavitt and L. Soete, 1990, The Economics of Technical Change and

International Trade, New York University Press, New York. Friedman, Milton, 1953, “The case for Flexible Exchange Rates,” Ch. IV in Essays in

Positive Economics, Chicago: The University of Chicago Press, reprinted in American Economic Association, Readings in International Economics, R. Caves and H. Johnson (eds.), 1968, Irwin, Homewood, Ill., pp.413-417.

59

Gandolfo, Giancarlo, 1987, International Economics I, Springer-Verlag, Heidelberg. Gregory, M., and C. Greenhalgh, 1997, “International Trade, Deindustrialization and

Labour Demand: An Input-Output Study for the UK (1979-90)” in Borkakoti J. and Ch. Milner (eds.), International Trade and Labour Markets, , St. Martin Press, New York.

Hansen, Bent, 1966, “The Walrasian System” in Price Theory, H. Townsend (ed.), 1971,

Penguin, New York, pp.107-127. Heller, Robert, H., 1973, International Trade Theory and Empirical Evidence, 2nd ed.,

Prentice Hall, New Jersey. Kaldor, Nicholas, 1938, “Welfare Economics propositions and Interpersonal Comparisons

of Utility”, The Economic Journal, 49, reprinted in Arrow, K. and T. Scitovsky, 1969, Readings in Welfare Economics, Irwin, Homewood Ill.

Kindleberger, Charles P., 1973, International Economics, 5th ed., Irwin, Homewood Ill. Kravis, Irving, B., 1956, “Availability and Other Influences on the Commodity

Composition of Trade,” Journal of Political Economy, 64, pp.143-155. Krueger, Anne, 1983, Exchange Rate Determination, Cambridge University Press,

Cambridge UK. Krugman, Paul, R., 1987, “Is Free Trade Passé?”, Economic Perspectives, Vol. 1, No.2,

Fall, pp.131-144. Krugman, Paul, R., 1983, “New Theories of Trade Among Industrial Economies,” The

American Economic Review, Vol. 73, No.2, May, pp. 343-347. Krugman, Paul, R., 1981, “Intraindustry Specialization and the Gains from Trade,” Journal

of Political Economy, 89, October, pp. 959-973. Krugman, Paul R., 1979, “Increasing Returns Monopolistic Competition and International

Trade,” Journal of International Economics, 9, No.4, November, pp.469-479. Krugman, P. and M. Obstfeld, 2005, International Economics: Theory and Policy, 6th

Edition, Addison Wesley Pearson, New York. Layard, P.G.R. and A. Walters, 1978, Microeconomic Theory, McGraw-Hill, New York. Levy, Santiago, 1981, “Foreign trade and its impact on employment. The Mexican Case,”

Journal of Development Economics 9, pp. 47-65.

60

Leontief, Wassily W., 1953, “Domestic Production and Foreign Trade: The American Capital Position Re-examined,” Proceedings of the American Philosophical Society, Vol. 97, reprinted in International Trade selected readings, J. Bhagwati (ed.), Penguin, Harmondsworth, UK, pp.93-139.

Lustig, Nora, 1992, “Tipo de Cambio, Protección Efectiva y Exportaciones

Manufactureras. México 1983-1987,” Investigación Económica, no.200, Apr-Jun, UNAM, pp.75-109.

Milberg, William, 2002, “Say’s Law in the open economy: Keynes’s rejection of the theory

of comparative advantage”, in Sheila C. Dow and John Hillard, Keynes, Uncertainty and the Global Economy: Beyond Keynes, Volume Two, Edward Elgar, Northampton: pp. 239-253.

Myint, Hyla, 1958, “The ‘Classical Theory’ of International Trade and the Underdeveloped

Countries,” Economic Journal, Vol.68, pp. 317-337. Markusen J. R., J. R., Melvin, W. H. Kaempfer and K. E. Maskus, 1995, International

Trade Theory and Evidence, McGraw-Hill, New York. Moreno-Brid, J. C., P. Ruiz-Nápoles and J. C. Rivas-Valdivia, 2005, “NAFTA and the

Mexican Economy: a look back on a ten year relation”, The North Carolina Journal of International Law and Commercial Regulation, University of North Carolina, Vol. 30, No. 4, pp. 997-1023.

Moreno-Brid, J. C., E. Pérez Caldentey and P. Ruiz-Nápoles, 2004-5, “The Washington

Consensus: A Latin American perspective fifteen years later,” Journal of Post Keynesian Economics, Vol. 27, No.2, pp.345-365.

Ohlin, Bertil, 1933, Interregional and International Trade, Harvard University Press

Cambridge, MA. Pasinetti, Luigi, 1977, Lectures on the Theory of Production. Cambridge University Press,

Cambridge, UK. Prebisch, Raul, 1959, “Commercial Policy in the Underdeveloped Countries”, American

Economic Review, Vol. XLIX, No.2, May, pp.251-273. Posner, Michael, V., 1961, “International Trade and Technical Change,” Oxford Economic

Papers, No. 13, pp. 323-341. Puyana A. and J. Romero, 2005, Diez años con el TLCAN. Las experiencias del sector

agropecuario mexicano, FLACSO-El Colegio de México, México. Ricardo, David, 1973, The Principles of Political Economy and Taxation, (1821), New

York: Everyman’s Library.

61

Robinson Joan, 1973, “The Need for a Reconsideration of the Theory of International Trade” in International Trade and Money, M.B. Connolly (ed.), Allen and Unwin, London.

Rodrik, Dani, 1995, “Trade Strategy, Investment and Exports: Another Look at East Asia,”

Working Paper 5539, National Bureau of Economic Research, Cambridge, MA, November.

Romalis, John, 2005, “NAFTA’s and CUSFTA’s impact on International Trade, Working

Paper 11059, National Bureau of Economic Research, Cambridge, MA, January. Ruiz-Nápoles, Pablo, 2004, “Exports, growth and employment in Mexico, 1978-2000”,

Journal of Post Keynesian Economics, Vol. 27, No.1, pp.105-124. Ruiz-Nápoles, Pablo, 2001, “Liberalisation, Exports and Growth in Mexico 1978-1994: a

structural analysis,” International Review of Applied Economics Vol.15, No. 2, pp. 163-180.

Sachs, Jeffrey, D., 1988, “Políticas comerciales y de tipo de cambio en programas de ajuste

orientados al crecimiento,” Estudios Económicos, Vol. 3, No.1, Jan-Jun, El Colegio de México, pp. 77-108.

Samuelson, Paul, A., 1949, “International Factor-Price Equalisation Once Again”, The

Economic Journal, June, pp.181-10 Shaikh, Anwar, 1980, “The laws of international exchange” in Growth, Profits and

Property, E.J. Nell (ed.), Cambridge, Cambridge University Press. Smith, Adam, 1937, The Wealth of Nations, (1776) Cannan Edition, Modern Library, New

York. Södernsten, Bo, 1970, International Economics, Harper and Row, New York. Széleky, Miguel, 2005, Pobreza y Desigualdad en México entre 1950 y el 2004, Serie:

Documentos de Investigación, Secretaría de Desarrollo Social, México. Vernon, Raymond, 1966, “International investment and international trade in the Product

Cycle,” Quarterly Journal of Economics, Vol. 80. Weintarub, Sidney (editor), 2004, NAFTA’s Impact on North America The Fist Decade,

The Center for Strategic and International Studies, Washington, D.C. Weller, Jürgen, 2001, Reformas Económicas, Crecimiento y Empleo, Fondo de Cultura

Económica - ECLA, U.N., México.

62