International Economics I

INTERNATIONAL
ECONOMICS I
Daniel Dueñas
Index
Introduction and Main Facts about Trade
The Neoclassical Framework: The Specific Factor Model
The Heckscher-Ohlin Model
The Ricardian Model
Increasing Returns to Scale (The Krugman Model)
Firm Heterogeneity (The Melitz Model)
FDI (Multinationals)
Topic in Trade Policy

International Economics I
Introduction and Main Facts about Trade
International Economics has basically two branches:

• International Trade: it studies trade in goods and services, factor movement across space,
economic integration, exporters and importers.
• International Finance: studies exchange rate, international financial markets, trade in assets.
We will focus on the first one.
International Trade is the exchange of goods and services across borders. It’s part of the process of
globalization, which it also includes movements of capital, labour, ideas, etc.
Although incentives are the same, International Trade differs from Interregional because there are
more barriers and restrictions in trade across countries than within. Barriers to international trade
are imposed by governments, but there can also be natural (geography), cultural (language), or
institutional (legislation issues, contract enforcement).
Once the introduction is done, some facts will be exposed to start.
-Since WWII, there has been a huge increase of International Trade. Some potential causes:
• Technological Improvements in communication and transports.
• Better political context for cooperation and integration after WWII, increase of
multilateralism (alliances between different countries with common goals; e.g. EU).
• Liberalization of trade: fall in tariffs and non-tariff barriers.
-Size and distance matters for international trade.

• The richer the countries, the more percentage of exports/GDP (USA, China, Germany, Japan
and France are who trade the most). Therefore, size matters, and also distance.
• Actually this can be proved with the gravity model, which explains surprisingly well trade
across countries: the more size in GDP and the less distance between two countries, the
more trade between them. The best way to bring it to the data is to take logs to both sides,
adding an error, and use a linear regression to estimate the equation:
• Structure of international trade: North-North 40%, North-South 40%, South-South 20%.
Most of exports in North America, EU, and Asia are done “inside” (EU to EU, Asia-Asia…).
-Developing export manufacturing, developed export “high-tech”.

• 80% are goods (53% manufactures, 18% fuels & mining, 8% agriculture), 20% are services.
• The patterns changed a lot over time: before (until 1980) agriculture>manufactures,
developing countries were main exporters in agriculture, now in manufacturing (while
developed are in technological advanced goods and services).
• Developed countries trade similar (different goods) with other developed (developing)
countries. That is called intraindustry (interindustry) trade.
-Not all firms export, just the most productive ones, of which only the most productive are
multinationals.
-Since early 90’s, there has been a rise of offshoring and global chain value, consisting on the
reallocation of production and/or supporting services processes to foreign countries.
During this subject, we will try to answer some questions about trade patterns, firms, exports, effect
of trade to unemployment and inequality, trade barriers, trade policies, and much more.
The Neoclassical Framework: The Specific Factor Model. Intuition.
Closed Economy
Production Possibility Frontier (PPF):

-Tells how much of each good an economy can produce given its resources and technological level.
-Represent the technological feasibility of the economy, which is given by the productivity
(productions functions: FM and FA) and quantities of inputs (L, K and T).
-A movement in the PPF show the reallocation of inputs from M to A, representing the opportunity
cost of producing one good in terms of the other good.
Analytically: Opportunity cost of M relative to A is equal to the slope of the PPF: the MRT =
If MRT = 2, we would give up 2 units of M to obtain 1 of A. Intuition: MPLA = 2 > MPLM =1, so as
we obtain more productivity with A, we prefer it.
Supply:
-Tells how much of A and M the economy will produce. It is obtained by firms maximize profits.
-Wages are equal because labour is mobile between sectors, as:

• If QM increases and firms demand more labour: ↑wM .
• As L moves from A to M, MPLA↑ and MPLM ↓.
• Therefore, as wi=pi × MPLi , wA↑ and wM↓, leading to wM=wA=w and MPLA/MPLM=p’M/pA.
Note: the decrease in MPLM allows us to explain why wM increases less than pM .
-There is a positive relationship between relative price and relative quantity: the higher is the price
of a good, the higher are the incentives to produce that good (relative to the other good).
Demand:
-Tells how much a consumer will demand the quantities of M and A. It is obtained by maximising
the utility function subject to the budget constraint.
-There is negative relationship between relative price and relative demand: if the price of a good
increases, we will consume less of that and more of the other.
Equilibrium:
-Is the relative price (pM/pA) where Relative Supply is equal to Relative Demand (RD = RS).
-Analytically, PPF (supply) and indifference curve (demand) have to be tangent to the relative price
line at the same point.
Open Economy
Equilibrium:
-Is the international relative price (pIM / pIA ) where International Demand = International Supply.
-Now, a country can consume different quantities from the produced ones.
-Although, it has to be in trade balance: imports have to be equal to exports.
-Moreover, imports must be equal to the value of exports i.e. we buy imports with exports.
Comparative Advantage:
-When in a country the relative price of M is lower in closed economy than in open economy, it
means that can produce it cheaper, so has a comparative advantage in the production of M.
-When we open to trade, M becomes more expensive. Therefore:
• Firms will reallocate L from production of A to M. Relative Supply of M increases.

• Consumers will demand less of M and substitute for A. Relative Demand of M decreases.
The quantity produced (QM) of M is larger than the quantity demanded (DM). Thus, M is exported.
Instead, the QA of A is lower than DA (comparative disadvantage). Therefore, A is imported.
There are gains from trade when there is a comparative advantage.
Gains from Trade and Redistributive Effects
Gains From Trade: when we open the economy to trade, economy’s choice expands. Therefore:
-Economy move to a higher indifference curve and representative consumer achieves higher utility.
-The higher the ToT, the more we are taking profit of our comparative advantage, the more welfare.
Redistributive effects: trade leads to changes in real income on the different agents in the economy:
Workers:
-Real wage equals to marginal products (w/pi = MPLi) and MPL are decreasing in labour.
-The labour increases in the exporting sector (M) and decreases in the importing sector (A).
-Therefore, workers gain in terms of A and lose in terms of M.
Capitalists:
-As their income comes from M, and production of M increases, profits will increase.
-Because the increase of pM/pA and πM/pM , πM/pA will increase too.
-Therefore, capitalists gain with respect to both goods.
Landowners:
-As their income comes from A, and production of A decreases, profits will decrease.
-Because the decrease of pA/pM and πA/pA , πA/pM will decrease too.
-Therefore, landowners lose with respect to both goods.
Conclusion: owner of the factor that produces exporting good gains, owner of factor that produces
importing goods loses. The effect on workers is ambiguous.
The Neoclassical Framework: The Specific Factor Model. Formulas.
Closed Economy
Production Possibility Frontier (PPF): how much of each good an economy

can produce given its resources and technological level.
Production Functions: QM=FM (K,LM), QA=FA (T,LA)
Labour Endowment: L=LA+LM
MPLM and MPLA are positive but decreasing for an additional unit of labour:
Supply: is obtained by finding the optimal production (maximising firm profits):
Equilibrium production: when Opportunity Cost of M = Relative Price of M
Relative Supply Curve:
Demand: is obtained by finding the optimal consumption (maximising utility subject to budget):
Equilibrium Consumption: when MRS = Relative Price of M
Relative Demand Curve:

Equilibrium: is obtained at relative price (pM/pA) where Relative Supply is equal to Relative Demand
Conditions:
• Produced Value = Consumed Value (Income = Expenditure):
• Produced Quantities = Consumed Quantities (of each good):
Open Economy
Equilibrium: is obtained at pIM / PIA such that International Demand = International Supply.
Comparative Advantage: when international relative price of M is higher than in autarky.
Gains From Trade

The Neoclassical Framework: The Specific Factor Model
Why countries trade? What are the gains from trade? Who gains and who loses from trade in each
country? To answer, it will be needed a general equilibrium neoclassical model.
We will depart from a simple starting point: the Specific Factors Model. In it, there are:
• 2 goods: manufactures (M: clothes, computers, etc.) and agricultural (A: food).
• 3 factors of production: labour (L, mobile between sectors), used to produce A and M; land
(T, specific factor), used to produce A; and capital (K, specific factor), used to produce M.
• Technology with decreasing marginal productivity of factors.
• Convex and homothetic preferences.
• Perfect competition and full employment.
Closed Economy
First, we will study the equilibrium in a closed economy, which occurs when demand=supply.
Firms demand labour and produce goods while workers supply labour and consume goods.
Everything is consumed and produced nationally. Let’s see how production works.
The economy produces M and A combining labour with its specific factors:
• The production function of manufactures is the following: the quantity of M produced
depends on labour employed in manufactures and capital. QM=FM (K,LM).
• The production function of agricultural goods is the following: the quantity of A produced
depends on labour employed in agriculture and land. QA=FA(T,LA).
• MPLM and MPLA are positive but decreasing for an additional unit of labour.
There are two assumptions:

• 1. An increase on labour leads to an increase of
production. In other words, the MPL is positive.
• 2. Every additional worker adds less production
than the previous ones. In other words, there are
decreasing MPL, as said above.
The Production Possibility Frontier (PPF) tells us how much of each good an
economy can produce given its resources and technological level.
The labour endowment is allocated between the production of M and A: L=LA+LM.
Then, for a given allocation, the economy produces:
• Q1M = FM (K,L1M) units of manufactures.
• Q1A = FA (T,L1A) units of agricultural goods.
The PPF tells us how much quantities Q2M and Q2A change if we move one worker from A to M.
This graphic puts together labour endowment, the productions

functions of agricultural goods and manufactures, and the PPF;
and allows us to see the impact of moving a worker from A to
M through the PPF, which is concave due Assumption 2.
Consequently, the PPF represent the technological feasibility of the economy, which is given by the
productivity (productions functions: FM and FA) and quantities of inputs (L, K and T). The
movement in the PPF show the reallocation of inputs from M to A, representing the opportunity cost
of producing one good in terms of the other good.
The decreasing MPL and PPF concavity can be understood by the following: if QA>>>QM, then a
small reallocation of workers from A to M increases QM more than QA falls. Thus, increasing
production of M represents a low opportunity cost. Analytically, the opportunity cost of M relative
to A is equal to the slope of the PPF: the Marginal Rate of Transformation (MRT)=
Once we know the feasible production of the economy thanks to PPF, now we need to know the
optimal production: how much of A and M the economy will produce. We assume that there are
perfectly competitive markets in which firms maximize profits with prices and wages already given.
From the F.O.C., it is obtained that MC (wi) = MR (pi*MPLi) in each sector:
As labour is mobile between sectors, wages are equal. If wA>wM, workers in M would rather work in
A. For this reason, it is supposed that a reallocation leads to an equilibrium:
From this, we obtain the following:
Thus, equilibrium production of A and M is determined by the relative prices

that lies at the tangency between the PPF and the line of the relative prices, as
it can be seen here:
If the price of M would increase (p’M > pM):

• Firms in M want to increase production and demand more labour ↑wM.
• As L moves from A to M, MPLA↑ and MPLM ↓. (As MPL is decreasing,
every additional worker lead to less marginal product. The more L, the less MPL).
• From wi=pi*MPLi we deduce: if MPLA↑ and MPLM ↓, then wA↑ and wM↓. That will lead to
the reallocation mentioned above, produced until wM=wA=w and MPLA/MPLM=p’M/pA.
• From wi=pi*MPLi we also deduce that w increases less than pM, as MPL decreases.
The Relative Supply Curve shows the positive relationship

between relative price and relative quantity: The higher is
the price of a good, the higher are the incentives to produce
that good (relative to the other good).
After observing relative supply, now we will see relative demand. A consumer will demand the
quantities of M and A that maximizes utility subject to the budget constraint, where:
• The utility function: U=U(DA, DM).
• The budget constraint: pMDM+pADA≤V where V is the income level of the consumer.
Assumptions on the utility function, U=U(DA, DM):
• Is increasing in the quantities consumed: The more goods, the better.
• Has decreasing marginal utilities: an additional good adds less utility than the previous good.
• Is described by convex indifference curves in the space of (DM, DA). It
means that consumers prefer consuming a bit of both goods rather than a
lot of one and a little of the other.
• Is homogenous of degree one. Everyone have same relative demand: rich
have the same relative demand as poor, they just consume more goods.
So, to find the optimal consumptions, utility has to be maximized subject to the
budget constraint:
Then, the problem can be solved using Lagrange:
Taking the first conditions, we obtain:
Finally, equalizing the FOCs, we obtain:
So, Marginal Rate of Substitution (amount of M that consumer is

willing to trade for one unit of A) is equal to the relative price. In
other words, indifference curve is tangent to the line of the relative
price. . Concluding, Relative Demand (DM/DA) is decreasing in
price. If pA↑, consumer demand less of A and more of M.
To end up, let’s define equilibrium in a closed economy (autarky).

All quantities consumed must be produced:
Therefore, Relative Supply should be equal to Relative Demand. As a result, equilibrium will be the
relative price (pM/pA) where RD=RS. So that equilibrium can be obtained, PPF and the indifference
curve have to be tangent to the relative price line in the same point.
We can compute GDP in this economy using various equations:

• GDP using production approach: V=pMQM + pAQA
• GDP using expenditure approach: V=pMDM + pADA
• GDP using income approach V=πA + πM + w × L (where π = p Q − wL are the profits of firms in sector i=M,A).
i i i i
Therefore, to achieve equilibrium in a closed economy two conditions must be hold:

• Produced value equals consumed value (or income = expenditure):
• Produced quantities equal consumed quantities of each good:
Open Economy
In an economy open to trade with the rest of the world, good prices are determined in equilibrium
with the international markets: pIM and pIA are such that international demand=international supply.
If an open economy is small, its economy take international prices as given.

If an open economy is large, the national supply/demand of goods affect international prices.
(Here, size refers to the importance in the market: even in GDP, Spain>Qatar, Spain is a small open
economy in oil market and Qatar is large). Now, we will assume we are a small open economy.
In an open economy, a country can consume different quantities from the produced ones:
The only condition that must hold is that produced values equals consumed value:
This condition entails that we assume “trade balance” (imports=exports).
The equation can be rearranged to obtain the following (remember: D-Q=consumption-production):

The imported A goods, (DA-QA) must be bought using the exports of M, (QM-DM):
In other words, the extra consumption of A (DA-QA) is bought using extra production of M (QM-DM).
The international price determines if a country exports or imports a

given good: if the international relative price is different from the
closed economy, production and demand will not be equal.
For now, we will suppose that the price of M in the international markets are higher that
the local price in the autarky.
That will generate incentives to reallocate labour from the production of A to the production
of M: firms will choose a new equilibrium production (where PPF is tangent to the relative
price line).
Meanwhile, consumers will decrease demand of M and substitute for A; leading to a new
equilibrium demand (indifference curve tangent to the relative price line), given the value of
production (i.e. income).
When we are in autarky and then we open the economy (and the price of M in the
international markets are higher than the local price in the autarky):
• As M becomes relatively more expensive, it will be more
produced but less consumed.
• As A becomes relatively cheaper, it will be less produced but
more consumed.
Opening the economy leads to a higher utility to the consumers, as now they will be consuming in a
point that is beyond (or above) the PPF. In other words, trade expand the possibilities of consuming.
The increase on the relative price leads to an increase in the relative

supply and a decrease in the relative demand.
In other words, an extra production is being generated, leading to

exportation.
We say that there is comparative advantage in the production of a good (in this case, M)
if its relative price in closed economy is lower than in open economy, as it would mean
that we can produce M cheaper. This leads to the following pattern of trade:
• A country exports the goods in which has comparative advantage (e.g., M), as the quantity
produced is larger than the quantity demanded: XM=QIM – DIM > 0.
• A country imports the goods in which has comparative disadvantage (e.g. A), as the quantity
demanded is larger than the quantity produced: XA=DIA – QIA > 0.
There is a comparative disadvantage when pIM/PIA < pM/pA.
If relatives prices in open and closed economy coincide (pIM/PIA = pM/pA), there is no comparative
advantage (nor disadvantage) for any country. Therefore, there are no gains from trade.
There is absolute advantage in the production of a good if its price in closed economy is lower than
in open economy (pIM>pM), but absolute advantage does not imply comparative advantage.
Trade requires comparative advantage, not necessarily absolute advantage.
There is comparative advantage if countries are different.
The sources of comparative advantages are the determinants of the prices in a closed economy:
• Technology [FM(K, LM), FA(T, LA)]: a country more productive in QM=FM(K, LM) would be
able to produce more M (lower pM/pA).
• Factor endowment, K, T, L: the more K, the more production of M (lower pM/pA).
• Preferences of consumers, U(DM, DA), (which are not really a source of comparative
advantage, but might affect relative prices): a country that likes more M than A would have
higher relative prices pM/pA.
Gains from Trade and Redistributive Effects
Without trade, the economy is constrained to consume at the PPF. Trade expands the economy’s
choice, as the economy move to a higher indifference curve and a representative consumer of this
economy achieves higher utility. That is the gain of trade.
One way to evaluate the economy’s gains from trade is to look at the Terms of Trade, which are
the relative price of exports in terms of imports. In an economy that exports M and imports A:
In that case, a decrease of our terms of trade leads to producing less M and more A, partially
stopping exportation of M and importing A. As a result, welfare would drop, and the economy
would get closer to autarky. Therefore, the better the ToT, the more welfare.
ToT helps us to analyse tariffs: supposing τ% of the value of imports, so the price for the domestic
consumer becomes pIA(1+τ). That would incentivise production of more A and less M and the
economy would be producing less of the good in which has comparative advantage. Now:
The Terms of Trade will deteriorate which leads to a decrease of the welfare.
Although, this is a very simplistic analysis; there are many other issues to consider.
Now, let’s imagine three different agents in the economy:

• Workers: they own L, supply labour and earn wages.
• Capitalists: own K, produce M and earn πM=pMQM – wLM.
• Landowners: own T, produce A and earn πA=pAQA – wLA.
Who wins and who loses from trade? (we will use real income in terms of both goods as a measure of welfare).
-Workers: recalling from profit maximizations, real incomes are equal to marginal products:
As we are exporting M, trade implies:
Therefore, there is an ambiguous effect: workers gain in terms of A and lose in terms of M.
-Capitalists: their income comes from profits from M:
QM and LM ↑, while w/pM 🡣. As all of them affect the πM/pM, what will happen?
The additional demand of M incentivise the production of M. Therefore, (recall: capitalist take pM
and w as given), capitalists will hire until the “last one” that makes MC=MR.
Hence, due the increase of production, profits will increase.
Moreover, because the increase of pM/pA and πM/pM , πM/pA will increase too.
Capitalists gain with respect to both goods.
-Landowners: with the same argument, landowners lose with respect to both goods (πM/pM , πA/pA 🡣)
Therefore, in the Specific Factors Model:

• The ones who own the factor specific (K) of the exporting sector (M) gains.
• The ones who own the factor specific (T) of the importing-competing sector (A) loses.
• The welfare change for the mobile factor is ambiguous.
But trade is not detrimental, as it makes the “pie” bigger. Therefore, we can always design a policy
that redistributes the gains from trade to the loser so everybody benefits. However, implementing
such policies are hard because involves political negotiation.
Summary:
After studying the implications of trade in a neoclassical model with specific factors, we observe:
-If the domestic relative price is different than the international prices, trade occurs.
-Trade expands the consumption possibility of the economy and increase overall welfare.
-Nevertheless, gains from trade are heterogeneous: owner of the factor that produces exporting good
gains, owner of factor that produces importing goods loses.
-Meanwhile, the effect on the mobile factor is ambiguous.

The Heckscher-Ohlin Model. Intuition
In the H-O Model, the comparative advantage comes from differences in factor endowments.
That difference will generate a specific pattern of trade.
Relative Demand:
-We obtain consumption (DT, DA) given the prices (PT and PA) and its their choices (b).
Relative Supply:
-We obtain production (QT, QA) in our economy given technology (aKi , aLi) and endowment (L, K).
-Labour and capital ratio (L/K) depends on the cost of labour relative to the cost of capital (w/r).
-Thus, if the cost of labour increases relatively (↑w/r) firms will substitute labour for capital (↓L/K).
-There is also different relative intensity in factor utilization for the production of different goods
-If T uses more L/K than A, then aLT / aKT > aLA / aKA.
-The more the relative endowment, the more production of the good intensive in that factor:
↑L/K → ↑QT , ↑K/L→ ↑QA.
-Rybczynski effect: if L↑, both sectors will grow, but as T grows more (as L is the factor which uses
more intensively), some K will be reallocated from A to T. Therefore: %ΔQT > %ΔL > 0 > %ΔQA.
Equilibrium Relative Prices:

-Lead to the equilibrium in which Relative Demand (DT / DA) = Relative Supply (QT / QA)
-Relative price of a good is decreasing in the relative endowment of the factor it uses intensively.
-The more abundant is a factor, the cheaper will be the good that uses it intensively.
-So, the more L/K, the more QT/QA, the lower PT/PA.
Factor Prices:
-The price of a factor is increasing in the relative price of the good intensive in that factor.
-For instance, if T becomes relatively more expensive than A, production of T will increase.
-Then, the demand for L will increase relatively more than K.
-So the price of L will increase relatively more than K’s, and so will the factor price.
-Stolper-Samuelson effect: an increase in the price of a good (e.g. PT) increases more than
proportionally the price (w) of the factor it uses intensively (L). %Δw > %ΔPT > 0 > %Δr.
Conclusion: if the relative endowment of a factor increases (L/K↑), the relative price of
the good that uses intensively falls (PT/PA ↓) and the relative price of that factors falls (w/r↓).
HO: Open Economy
Price Convergence:
-When we open to trade the price of both goods is equal to the international price in both countries.
-Why equilibrium relative price lies between the closed-economy ones? (PT/PA < PTI/PAI < PT*/PA* ).
Because L/K > L+L*/K+K* > L*/K*.
Heckscher-Ohlin:
-When we open to trade, the good in which we have a higher relative endowment (L/K > L*/K*)
will become relatively more expensive (QT/QA > QT*/QA* → PT/PA < PTI/PAI < PT*/PA* ).
-Equilibrium RD implies that in each country the RD of the good in which there is comparative
advantage will be lower than RS → T: RS*<RDI<RS. A : RS<RDI<RS*.
-Therefore, home exports T and imports A while foreign exports A and imports T.
-In other words, each country exports the good intensive in its relative abundant factor.
Factor Price Equalisation:

-As technology (a’s) and prices PTI/PAI are the same in each country...
-The prices factor will also be the same (w=w*=wI and r=r*=rI).
-Trade benefits the abundant factor and hurts the scarce factor.
-So, PTI/PAI > PT/PA → wI/rI > w/r and PTI/PAI < PT*/PA* → wI/rI < w*/r*.
Applications
Effects of immigration:
-When the comparative advantage is reinforced (weakened), trade increases (decreases).
-When a third country is involved, (L+L*)/(K+K*) changes, and also QTI/QAI and PTI/PAI.
Thus, it has an effect on wI/rI.
-If immigration is between home and foreign country, (L+L*)/(K+K*) remains unchanged, so will
QTI/QAI, PTI/PAI, and wI/rI.
Skill premium:
-If US and Mexico open to trade, the relative price of the good in which they have a comparative
advantage will increase, and so will the factor price of that good.
-Therefore: relative price of PCs↑→relative remuneration of H↑→skill premium↑→inequality ↑.
Empirical Evidence
The evidence in favour of the HO is mixed.
-But the pattern of trade between developed and developing are well reflected in the HO model:
-Vietnam (L-abundant country) exports L-intensive goods, while Germany (K-abundant country)
exports K-intensive goods.
International Economics
The Heckscher-Ohlin Model. Formulas
-Relative Demand: -Resource constraint of the economy:
-MRST is equivalent to the relative input prices:
-Factor requirements of production function:
-If T is more relatively intensive in L: - -If (↑w/r), (↓L/K).
-Supply of QT and QA:
-Relative supply: -Equilibrium in closed economy (RD=RS):
-Relative prices (which lead to equilibrium where RD=RS):
-Factor prices: -Relative price of factors:
-Rybczynski effect: -Stolper-Samuelson effect:
HO: Open Economy
-If Home is relatively L-abundant, in closed economy:
-By equalising IRS=IRD… we obtain that with the open trade,

. PT=PT*=PTI and PA=PA*=PAI
.
and PT/PA < PTI/PAI < PT*/PA*
-As PTI>PT , home has a comparative advantage with the production of T.
-As PAI>PA , foreign has a comparative advantage with the production of A.

-Heckscher-Ohlin Theorem: in open economy (provided that no perfect specialization occurs), a
country exports the good intensive in its relatively abundant factor.
-Factor price equalisation: as PTI/PAI and “a’s” are the same, w=w*=wI and r=r*=rI .
-PTI/PAI > PT/PA → wI/rI > w/r . In the same way, PTI/PAI < PT*/PA* → wI/rI < w*/r* .
Applications
-When labour changes:
-When NAFTA occurs (i.e. open the economy): (we assume HUSA/LUSA > HMEX/LMEX).
Empirical Evidence
HO-Vanek (net factor f content of country c’s trade = factor f endowment – factor f demand)
In the neoclassical frameworks, there is trade when there is comparative advantage. While in
Ricardian model the comparative advantage comes from technological differences across countries,
in the HO model it comes from differences in factor endowments (capital, labour, land…).
In the Heckscher-Ohlin model, a country’s comparative advantage not only depends on its relative
factor abundance, but also on the relative intensity in factor utilization for the production of
different goods.
We will study how differences in resources (factors) generate a specific pattern of trade. In fact, the
HO model is very similar to the Specific Factor Model. The crucial difference is that while in SFM
there is only one mobile factor and works better for the short run, in HO all factors are mobile
across sectors and is more useful at a long run.
We will have:
• 2 countries: home and foreign (represented with *), both with same preferences.
• 2 goods: textiles (T) and automobiles (A), with same technology to produce each good in
both countries, and with T using labour more intensively than A.
• 2 factors of production: labour (L) and capital (K), which are mobile between sectors but not
between countries. There are different relative endowments of labour (e.g. if L/K > L*/K*,
home has relative abundance of L).
First, we will define relative demand and relative supply to find an equilibrium in closed economy:
There will be standard utility with usual assumptions: increasing in both goods, homogenous of
degree one, preferences are identical in both countries.
We will assume that representative consumer’s income is rK + wL, and that consumer spends a
fraction b of her income in good T, and (1-b) in good A: PTDT=b(rK+wL) and PADA=(1-b)(rK+wL)
Combining and rearranging it, we obtain the relative demand:
Production is carried by combining both inputs K and L using a technology (a production function):
QT=FT(KT, LT) & QA=FA(KA,LA) where Ki and Li are the quantities of capital and labour in each
sector (for i=T, A). There is some degree of substitution between inputs.
Given the total quantities of K and L, the resource constraint of the economy is:
The firms decide how much of each input uses by the following:
Combining the two optimality conditions:
Therefore, the Marginal Rate of Technical Substitution should be equal to the relative input prices.
Because of factor mobility, factor prices (r, w) are the same in both sectors, but MPL and MPK are
not the same in both sectors. This means that the labour and capital ratio (L/K) depends on the cost
of labour relative to the cost of capital (w/r). Thus, if the cost of labour increases relatively (↑w/r)
firms will substitute labour for capital (↓L/K).
Production function has the following factor requirements:
As seen, aKi and aLi are unit factor demands, which show
how much K or L is needed to produce 1 unit of T or A.
In general, depend on factor prices (w and r). For now, we consider them constant and exogenous.
As we said before, T is relatively intensive in L. Therefore:
This graph summarizes it: there is only one quantity of w and r in the
economy, but different amounts of L and K. Sector T is relatively intensive
in L compared to A. Moreover, we can see that if w/r increases, L/K will
decrease in both sectors too, as firms will substitute L for K.
Equilibrium in closed economy will be such that RD(=DT/DA) = RS ⮕
We already know that RD (=DT/DA=(b*PA/(1-b)*PT)), now we need to obtain RS (=QT/QA).
As production of A and T has to achieve full employment of L and K:

Then, we can use these two equations to derive the relative supply QT/QA.
After a lot of calculation, we obtain the RS:
Consequently, with the relative supply we obtain how much are we going to produce (QT, QA) in our
economy given the technology (aKi , aLi) and the endowment (L, K).
To produce both goods, two conditions are required for home:

• Different factor intensities across sectors (T is intensive in L):
• Relative labour endowment within the “cone of diversification”:
i.e. L/K is lying between the relative labour intensities of T and A
It means that a country produces at least some of both goods.
As we have seen
·The more the relative endowment of L, the more production of the L-intensive good: ↑L/K → ↑QT .
·The more the relative endowment of K, the more production of the K-intensive good: ↓L/K→ ↑QA.
This is formulated in the Rybczynski effect, which says that: an increase in the
endowment of a factor (e.g., L) raises disproportionately the production of the
good intensive in that factor (QT).
It can be better understood graphically: as LT ↑ (while LA=), the PPF expands (while pA/pT =). In
order to absorb the increase of LT , we need to employ more K, which will be reallocated from A to
T. Therefore, the ↑QT implies QA ↓, but ↑QT > QA ↓.
Once we have the RS, we can obtain the PT/PA that leads to the equilibrium where RD=RS.
After a lot of formulation and rearranging, we obtain the following:
From the formula, we can deduce that PT/PA is a decreasing function of L/K: the relative price of a
good is decreasing in the relative endowment of the factor it uses intensively. So, the more L/K, the
more QT/QA (RS), the lower PT/PA.
Another way to see it: different relative endowments leads to different relative
prices. That is what makes a country have a comparative advantage. Thus:
• In K-abundant countries, the K-intensive good is cheaper.
• In L-abundant countries, the L-intensive good is cheaper.
Then, what about the factor prices (r, w)? As we are perfect competition, p=MC 🠚
As aKi = Ki/Qi and aLi = Li/Qi , we can rearrange the last equation so that Qi×Pi = Ki×r + Li×w,
which is the same as Total Costs = Total Revenue.
Then, if T becomes relatively more expensive than A (PT/PA):

• Production of T will increase.
• Demand for L will increase relatively more than K.
• Therefore, the price of L will increase relatively more than K’s 🠚 w/r ↑
After that, by developing operations, we can obtain the

factor prices:
· which are increasing in the price of the good intensive in that factor:↑PT 🠚 ↑w , ↑PA 🠚 ↑r
· which are decreasing in the price of the other good: ↑PT🠚↓r , ↑PA 🠚 ↓w
Therefore, we can obtain the relative price of a factor:
which is increasing in the relative price of the good intensive in that factor: PT/PA↑, w↑. PT/PA↓, r↑.
This is explained by the Stolper-Samuelson effect, which says that: an increase in the price of a
good (e.g. PT) increases more than proportionally the price (w) of the factor it uses intensively (L).
Summary: Equilibrium in Closed Economy:
Finally, all is connected: if the relative endowment of a factor increases (L/K↑), the relative price of
the good that uses intensively falls (PT/PA ↓) and the relative price of that factors falls (w/r↓).
HO: Open Economy
Let us consider 2 large economies (what happens “inside” has effect on international prices), home
and foreign (*), with same tastes (RD=RD*), same technology, but different relative endowments:
aLT/aKT > L/K > L*/K* > aLA/aKA . A and T: produced in both countries, but Home is relatively L-
abundant. Thus, Home has comparative advantage in T, Foreign has it in A. In closed economy:
Open to trade (assuming no trade frictions/costs) leads to price convergence, so the price of both
goods has to be equal to the international (I) price in both countries: PT=PT*=PTI and PA=PA*=PAI .
We obtain by equalizing International RD with International RS:
Given that RS are different among countries (QT/QA > Q*T/Q*A), the equilibrium
relative price will lie between the closed-economy ones: PT/PA < PTI/PAI < PT*/PA* .
This happens due L/K > L+L*/K+K* > L*/K*.
Therefore, we obtain the mentioned price convergence.
To open to trade (and achieving the equilibrium relative price) implies that:
• For home, T becomes relatively more expensive (than when we were in closed economy:
PTI>PT). Thus, home has a comparative advantage with the production of T.
• To foreign, A becomes relatively more expensive (than when we were in closed economy:
PAI>PA). Thus, foreign has a comparative advantage with the production of A.
Consequently, the equilibrium relative demand implies that in each country, the relative demand of
the good in which there is comparative advantage will be lower than relative supply:
Textile: RS*<RDI<RS. Automobile: RS<RDI<RS*. As a result, this extra production and
consumption will lead to home exporting T and importing A, foreign exporting A and importing T.
This leads us to the Heckscher-Ohlin Theorem: in open economy (provided that no perfect
specialization occurs), a country exports the good intensive in its relatively abundant factor.
As Terms of Trade (Pexports/Pimports) increases (such as PT/PA and PA*/PT* ↑) and ToT has a positive
effect in the wealth, both countries will have gains from trade.
If both countries produce both goods, factor price equation must hold in both countries:
As international prices are the same in both countries (PTI/PAI) and we assume that technologies are
the same (the “a’s”), factor prices should equalize across countries: w=w*=wI and r=r*=rI .
This leads to factor price equalization. If no barriers to trade, technologies are the same in both
countries, and there is no complete specialization: factor prices equalize.
That will also entail some consequences for income distribution:

• In Home, the increase of the relative price of T in home makes L gain relative to K.
PTI/PAI > PT/PA → wI/rI > w/r
• In Foreign, the increase of the relative price of A in foreign makes K* gain relative to L*.
PTI/PAI < PT*/PA* → wI/rI < w*/r*
Trade benefits the abundant factor (w and in Home, r in foreign) and hurts the scarce factor.
Applications
First, let us study what happens when factor endowment change by studying labour change:
immigration. In this case, from a third country into the foreign country (K-abundant):
L*/K*↑ → QT*/QA*↑ → PT*/PA*↓ →PTI/PAI ↓ → wI/rI ↓
This will have many consequences: the comparative advantage will be weakened in both countries,
leading to less trade. Part of gains from trade will be lost, and workers in both countries are going to
lose while capitalists will gain relatively.
If instead, the immigration would be from a third country to home country (L-abundant):
L/K ↑ →QT/QA↑→ PT/PA↓ →PTI/PAI ↓ → wI/rI ↓
Here there will be opposite consequences: the comparative advantage will be reinforced in both
countries, leading to an increase of trade. There will be more gains from trade, but again, workers
will lose relative to capitalists.
Now let us consider that migration goes from home (L-abundant) to foreign (K-abundant) country:
L*/K*↑ → QT*/QA*↑ and L/K↓ → QT/QA↓, while (L+L*)/(K+K*) unchanged, as also QTI/QAI .
Again, the comparative advantage will be weakened in both countries, leading to less trade. Part of
gains from trade will be lost, but since RSI is unchanged, there will be no effect on PTI/PAI and wI/rI:
wealth from workers and capitalists will also remain unchanged.
Now let us consider that migration goes from foreign (K-abundant) to home (L-abundant) country:
L/K↑ → QT/QA↑ and L*/K*↓ → QT*/QA*↓, while (L+L*)/(K+K*) unchanged, as also QTI/QAI .
Now, the comparative advantage will be reinforced in both countries, leading to more trade. There
will be larger gains from trade, but since RSI is unchanged, there will be no effect on PTI/PAI and
wI/rI: wealth from workers and capitalists will also remain unchanged.
The other application we are going to do will be focused on the effects of trade in the skill premium.
One particular type of inequality within a country is the skill premium, which is the wage gap
between skilled and unskilled workers (Ws/WL), which has risen worldwide in the last 40 years.
Some possible determinants: trade, drop in relative supply of skilled worldwide (probably not true),
and technological change.
Let us focus on the US-Mexico case (where it has been an increase of trade due NAFTA).
First, we will assume that technology is the same and WH > WL in both countries, and that there are:
• 2 factors: high-skilled labour (H) and low-skilled labour (L).
• 2 goods: textiles (intensive in L) and PCs (intensive in H).
• H is relatively more abundant in the US: HUSA/LUSA > HMEX/LMEX.
Due this last assumption, if we apply the theoretical results learnt in class, PCs are exported by USA
and imported by Mexico, while textiles are exported by Mexico and imported by USA.
If suddenly NAFTA happens (and therefore there is more trade):

• In US: relative price of PCs↑→relative remuneration of H↑→skill premium↑→inequality ↑.
• In Mexico: relative price of PCs↓→relative remuneration of H↓→skill
premium↓→inequality ↓. (when NAFTA happens):
In the data, we have seen that skill-premium has increased in US, but it has also increase in Mexico.
The basic HO model fails for Mexico; therefore, it should be something else!
A possible explanation is the skilled-biased technological change, which means that H has become
more and more productive over time in both countries, which implies a higher demand for H
relative to L in both countries. Consequently, skill-premium increases because of changes in
technology. There is evidence in favour of this argument: production in the US has become more
intensive in H in all sectors.
Empirical Evidence
The essence of the HO model is that trade is driven by differences in factor abundance across
countries. Goods trade is a substitute for factor trade: India exporting T is a way of exporting “L”.
To test the predictions of the model, we should look at the factor content of the goods traded.
Although, in reality there is many goods, factors and countries, so sometimes can be difficult to
know which good is intensive in which factor or factor abundance relative to other factors.
The Leontieff paradox (1953) showed that the US export were much more labour-intense than its
imports despite the fact that US had much more capital per worker than the other countries. Some
explanations were that HO ignores the difference of technology among countries, land (important
input!), disaggregation of labour by skill, and trade barriers.
Leamer’s critique (1980) exposed that we should not look at L/K of exports/imports, but to the net
factor content of all trade instead.
That critique lead to the HO-Vanek model, an alternative version of the HO. Consist on looking at
net factor, such as: net factor f content of country c’s trade = factor f endowment – factor f demand.
• Vcf and Vcw : country c and world (w) endowment of factor f.
• sc = country share c in world income →demand of f in country c= sc Vwf
• Fcf = net factor f content of country c’s trade.
Therefore: if net factor of f (e.g. labour) is positive,

that country will export labour-intensive goods. If net factor of f is negative, it will import them.
For instance, if country c has a GDP equal to 5% of world GDP (sc=0.05) and:
• An endowment of factor j equal to 10% of world endowment of j (Vcj/Vwj =0.1), it will be a
net exporter of j: 0.1-0.05= 0.05 →it will export the 5% of world endowment of j).
• An endowment of factor h equal to 2% of world endowment of h (Vch/Vwh =0.02), it will be a
net importer of h: 0.2-0.05= -0.03 →it will import the 3% of world endowment of j).
Provided no perfect specialization occurs, a country is net exporter of the services of its abundant
factor and net importer of its scarce factor.
Bowen et al. (1987) considered 27 countries and their endowment of 12 factors, and counted for
how many countries the net exporter of each factor follows the predicted pattern. The result was
that the HO-Vanek theory had no predictive power concerning the direction of trade.
Trefler (1995) pointed out that the reality showed very little factor trade compared to HO
predictions (“missing trade”). Davis and Weinstein proved that HO works if you add different
technology (factor productivity), no factor price equalization across countries, and non-traded goods
+ trade costs.
Romalis (2004) showed the validity of a “quasi-H-O” prediction: “countries abundant in skilled
labour and capital capture a higher share of US imports in sectors intensive in those factors”.
The intuition was that given the set of exporters to a certain destination (the US), skill-abundant
countries are “better” at exporting skill-intensive goods, hence capture a higher import share the
higher the skill intensity of the good.
The advantages of that model were that it was no needed to assume same technology and factor
price equalization, and also used high-quality and homogenous data. Therefore, data hold.
Concluding, the evidence in favour of the HO is mixed: trade in goods does not necessarily reflect
trade in factors: volume of trade is substantially lower than predicted.
Moreover, the main missing point was the technological differences across countries.
Although, the “main pattern of trade” between developed and developing are well reflected in the
HO model: Vietnam (L-abundant country) exports L-intensive goods, while Germany (K-abundant
country) exports K-intensive goods.
The Ricardian Model. Intuition
In the Ricardian Model, the comparative advantage comes from technological differences across
countries, and therefore, on the relative productivity and relative wage across countries.
These differences generate a specific pattern of trade.
The Ricardian Model
Supply:
-When we compare the productivities in goods in different countries, we care about relative
productivity. We will order all goods by decreasing home relative productivity (from more to less).
Efficient Specialisation:
-When a country produces the goods of its comparative advantage.
-The country that sells each good cheaper will end up producing it. In other words, every country will
specialise in goods whose sectors have a relative productivity higher than the relative wage.
Demand:
-The relative wage allows us to find:
• The z goods that produce the home country (supply).
• The Relative Demand of home goods (for a given z, relative wage must clear the market).
-So, relative wage is determined by: supply (A[z]) = demand of home produced goods.
-The Relative Demand shows the demand in the “home” goods market for given specialization z in trade
balance. Thus, the equation must be always equal.
Equilibrium:
-Home produces z goods at a given value of relative wage (w/w*).

-This equilibrium will depend on:
• Technological differences between home and foreign (the slope of A(i)).
• The size of both countries (L*/L).
Equilibrium Properties and Gains from Trade
Gains From Trade:
-When we move from autarky to open economy, GFT come from efficient specialisation, as each
country can consume cheaper the goods in which the other has a comparative advantage.
-The higher are the differences between relatives productivities, the larger the Gains from Trade.
When a Foreign Country get larger (L*/L ↑)
-As now there is more population in foreign, the demand for home goods increases (for a given z).
-Therefore, relative wage increases (↑w/w*). Thus, home loses comparative advantage in the goods in
which relative productivity is lower to relative wage, and foreign will produce them.
-Home specialises in less goods (↓z→average a↑) while Foreign specialises in more (↑z→avrg a↓)
-Conclusion: ↓z, ↑w/w*
-To see the changes in welfare, we consider real wage.
-At Home:
• ↑ Due efficient specialisation, goods that shift to “foreign” (z’ < i ≤ z) become cheaper.
• ↑ As there is higher relative wage, goods that remain “foreign” (z ≤ N) become cheaper.
-At Foreign (mirror effect):
• ↓As there is lower relative wage, goods that remain “home” (i ≤ z) become more expensive.
• ↓As foreign productivity is lower, goods that shift to “foreign (z’ < i ≤ z) become more
expensive.
Conclusion:
• Small countries gain more from trade: specialise in few sectors with high productivity.
• Advanced countries benefit from China and India opening to trade, allowing them to specialise
in high-tech sectors while importing cheaper in low-tech sectors.
When Foreign Relative Productivity increases in all sectors (a*Li)’ < a*Li →A(i) for all i
-There will be a larger set of goods in which Foreign’s relative productivity will be higher than their
relative wage. Due to the loss of relative productivity, Home produce less goods (z↓).
-Thus, the relative demand of home goods decreases, and so will the Home’s relative wage (↓w/w*).
-Mirror effect: meanwhile, foreign country specialises in more sectors and enjoys higher relative wage
due technological improvement.
-Conclusion: z↓, ↓w/w*.
Effects on Welfare (real wage):
At Home:
• ↑ Due efficient specialisation, goods that shift to “foreign” (z’ < i ≤ z) become cheaper.
• ↑ As the positive effect on the Foreign productivity is larger than the negative effect on the
Home relative wage, goods that remain “foreign” (z ≤ i ≤ N) become cheaper.
At Foreign:
• ↑ As there is higher relative wage, goods that remain “home” (i ≤ z) become cheaper.
• ↑ As the positive effect on the Foreign productivity is larger than the positive effect on the
Foreign relative wage, goods that switch to “foreign” (z’ < i ≤ z) become cheaper.
• ↑ Due to higher foreign productivity goods that remain “foreign” (z ≤ i ≤ N) becomes cheaper.
Conclusion: both trade partners gain if one experiences technological progress in all sectors, but the
country that experiences it gains more.
Transportation Costs
-With Transportations Costs, now each country has a different function for which buys a good i from the
other country, leading to two different A(i) function and two different equilibrium value z.
-At i ≤ z, Foreign buys from Home. At z*≤ i, Home buys from Foreign. z < i < z* are non-tradeable.
Empirical Evidence
Transfers:
-They allow us to relax the trade balance assumption. T = I – X.
-It has no effect on w/w* nor z if we assume all goods are traded. If we assume non-tradeable:
-Transfer increases relative demand for Home goods and hence its relative wage, as the money
transferred is not used on Foreign goods (tradeable) but Home’s (non-tradeable).
The Ricardian Model. Formulas
The Ricardian Model
-Production:
-Resource constraint: L=L1 + L2 + … +LN = aL1Q1 + aL2Q2 + … + aLNQN
-Relative productivities of N goods:

(ordered by decreasing home relative productivity)
-Condition to specialisation:
(home will specialises in goods whose
s sectors have a relative productivity higher than the relative wage)
-Distribution of productions:
(Home produces all goods i=1, 2, …, z,
while Foreign produces i=z+1, z+2, …, N.
-Comparatives advantages:
(Home has a comparative advantage in z-1 while foreign has it in z+1.)
-Expenditure in good j: (wL + w*L*) × 1/N
-Total expenditure in home-produced goods: (wL + w*L*) × z/N
-Equilibrium in the “home” goods market (for given specialisation z): w*L* × z/N = wL × (N-z)/N
(there must be trade balance: home exports = home imports).
-Relative Demand (of home goods): -Equilibrium (supply=demand):
-Real wage: -GFT in Home Real Wage:
-GFT in Foreign Real Wage:

-Effect of (L*/L↑):
-Effect of technological progress (a*Li)’ < a*Li → A(i) ↓ for all I:
-If we assume transport costs:
- Home only buys foreign goods if:
-Foreign only buys home goods if:
-To find z and z* (we use goods markets equilibrium with balanced trade: home imports = home exports):
Empirical Evidence and Applications
-Transfers: -Transfers with non-tradeables:

The Ricardian Model
In the neoclassical frameworks, there is trade when there is comparative advantage.
While in Heckscher-Ohlin comparative advantage depends on differences in factor endowments

(capital, labour, land…), in the Ricardian Model it relies on the technological differences across
countries, and therefore, on the relative productivity and relative wage across countries.
We will see study how differences in productivity generate a specific pattern of trade, among other
things. Concretely, we will revise the Ricardian model developed by Samuelson et al (1977).
The Ricardian Model
We have:
• 2 countries: home and foreign (*)
• Many goods: indexed by i=1, …, N, N large (→∞)
• One factor of production: labour (L and L*), mobile between sectors and immobile between
countries.
• Technologies with constant returns to scale: different across sectors and across countries
(being the source of comparative advantage).
• Same preferences in both countries.
• Perfect competition.
The units of i produced (Qi) are defined by the units of labour employed in sector i (Li) and the
home labour productivity in sector i (aLi) i.e. how much of L needed to produce one unit of i.
a’s can be different across goods and across countries: aLi ≠ aL1 ≠ aL2 ≠ aLN ≠ aLi* ≠ aL1*≠ aL2* ≠ aLN*
The resource constraint with many goods is defined by the (full) employment of all workers in all
the different sectors: L=L1 + L2 + … +LN = aL1Q1 + aL2Q2 + … + aLNQN
As the production function is linear, PPF will be a straight line.

If we have 2 goods, L=aL1Q1 + aL2Q2, we can rearrange it to obtain:
Once we know how this economy produces, now we will see how it decides to produce:
Suppose that Home and Foreign have different technologies for all goods i=1, 2, …, N, and Home is
more productive in sector j. (That implies aLj < a*Lj , as it takes less L to produce the same Q).
When comparing two goods, Home is relatively more productive

than Foreign in producing good 1 than in producing good 2.
To understand it better: Germany (H) can be more productive that Vietnam (F) at producing cars
and clothes, but relatively speaking, is much more productive in cars than in clothes.
For convenience, we will order goods by decreasing home relative

productivity, which can be graphically represented by the following
(where A(i)=aLi*/aLi):
To determine Pi and wi , we use the maximization function to firms for each i:
From that, we obtain that Pi = aLi × w. As L is mobile across sectors, we assume same wage in all
sectors. In contrast, as L is immobile across countries, there will be different wages.
Then, who will end up producing good i? The country that sells it cheaper. With free trade, the
country that produces the good at the cheaper price “captures” the market. Hence, PiI= min{Pi, Pi*}.
Therefore, every country will specialises in good whose sectors have a relative productivity higher
than the relative wage: Home will specialise in goods i with
Let us imagine that there is a good z in which Home and Foreign are equally productive:
Therefore, z will represent the turning point in which Home stops producing and Foreign starts:
Home produces all goods i=1, 2, …, z, while Foreign produces i=z+1, z+2, …, N.
We can rewrite that in terms of relative prices, so we see the comparative advantage.
The autarky price of z-1 relative to z+1 is:
• In Home: P = Pz-1/Pz+1 = aLz-1/aLz+1 rearranging from
• In Foreign: P* = P*z-1/P*z+1 = a*Lz-1/a*Lz+1
Therefore, Home has a comparative advantage in z-1 while foreign has it in z+1.
So, there will be an efficient specialisation when a country produces the goods of its comparative
advantage, meaning that the country’s relative productivity is high enough so to compensate its
relative cost of labour.
However, a country does not need to have higher labour productivity (absolute advantage) to
specialize in a good. Even if aLi > a*Li, can have Pi<Pi* if the relative wage is low enough so that
w/w* < a*Li/aLi < 1. In other words, absolute advantage may not be sufficient.
If the relative wage increases, home specialises in the production of fewer goods (w/w*↑→↓ z), so
produces only in the sectors where its relative productivity is higher, and becomes on average more
productive relative to the foreign country.
The relative wage is determined by: supply = demand of home (or foreign) produced goods.
The A(i) schedule represents the supply for given relative wage w/w*, so the supply side will
depend on technology (the “a’s”), while demand will depend on preferences of consumers.
In the demand side, we will assume that preferences are the same in both countries and consumers
spend on each good the same share (1/N) of their income (wL).
Therefore, the total expenditure in good j: (wL + w*L*) × 1/N .
As all the goods between i=1 and i=z are produced by Home, z represents all the home-produced
goods. Consequently, the total expenditure in home-produced goods: (wL + w*L*) × z/N.
So, to obtain the equilibrium in the “home” goods market for given specialization z:
• There must be trade balance (home exports = home imports): w*L* × z/N = wL × (N-z)/N
• Rearranging, we obtain the Relative Demand in terms of marginal commodity z and w/w*:
for given specialisation pattern (z), the relative wage must

c clear the market.
Then, if Home specialises in more goods (z↑):

• Relative demand of home goods (and labour) increases
(z/(N-z))↑.
• To compensate it and restore the equilibrium, home
income must increase (w/w*)↑.
To draw the demand-side, B(i, L*/L), we assume any i to be the

marginal good. Then B(i, L*/L) is increasing in i.
Concluding:
-Home produces z goods and has a relative wage of w/w*.
-Equilibrium values depend on:

• Technological differences between home and foreign (the slope of A(i)).
• The size of both countries (L*/L).
We will see how equilibrium and gains from trade varies with the change of variables.
Real wage, or in other words, the purchasing power of home wage in terms of good i is:
• In closed economy (recall Pi = w × aLi):

• In open economy:
So, home real wage will be home productivity, while foreign real wage will be foreign productivity
multiplied by relative wage.
To see where gains from trade come, we compare real wage in autarky and in open economy.
Home real wage:
• Is unchanged in terms of the goods that remain “home” (i ≤ z):
• Increases in terms of the goods of foreign specialization (z < i ≤ N):
Therefore, home has the following gains from trade: it can consume the same units of goods i=1,2,
…, z, and more of the goods i=z+1, …, N, since they are produced cheaper in foreign country.
In foreign real wage, the opposite happens:
• Increases in terms of “home” goods (i ≤ z):
• Is unchanged in terms of goods that remain foreign (z < i ≤ N):
Therefore, foreign has the following gains from trade: it can consume the more of goods i=1,2, …,
z, since they are produced cheaper in home country, and the same of the goods i=z+1, …, N.
So, the Gains from Trade for both countries depend on the difference
between relative productivity and the relative wage (aLi*/aLi & w/w*):
• The higher the difference, the larger the GFT.
• In the graph, the gains are proportional to the areas between
the A(i) schedule and the equilibrium relative wage.
• The more diverse the trading partners (steeper A(i)), the
larger their GFT.
Now, let us suppose that foreign country becomes larger (L*/L ↑). Then, Curve B (i, L*/L) becomes
steeper (or moves to the left), leading to ↓z and ↑w/w*.Consequently:
• For a given specialisation (z), the demand for home goods increases.
• As a result, relative wage increases (↑w/w*)→home loses comparative advantage in the
goods with lower relative productivity (a*Lz’/aLz’ > a*Lz/aLz).
• Thus, home specialises in less goods (↓z) and home average
relative productivity increases.
• Meanwhile, due relative wage increase (↑w/w*)→foreign
acquires comparative advantage in the sectors of lower relative
productivity (a*Lz’/aLz’ > a*Lz/aLz).
• Therefore, foreign specialises in more goods (↓z) and foreign
average relative productivity decreases.
Conclusion: ↓z, ↑w/w*
To see which is the effect of L*/L ↑ to the welfare, we will consider home real wage.
[Recall: z’<z and (w/w*)’>(w/w*)].
• It is unchanged in terms of the good that remain “home” (i ≤ z’)
• Increases in terms of the goods that shift to “foreign” (z’ < i ≤ z), since
they become cheaper due to efficient specialization.
• Increases in terms of the goods that remain “foreign” (z ≤ N), since they
become cheaper due to (w/w*)’/a*Li > (w/w*)/a*Li
Now let us consider foreign real wage. [Recall: (w*/w)’<(w*/w)].

• Falls in terms of the goods that remain “home” (i ≤ z), since home goods
become more expensive due to lower relative wage.
• Falls in terms of the goods that shift to “foreign (z’ < i ≤ z), since they
become more expensive due to low
foreign productivity.
• Is unchanged in terms of the goods that remain “foreign”.

Conclusion:
• Small countries gain more from trade, since they can specialize in few sectors, where they
have the highest productivity and can import more goods at a cheaper price.
• Advanced countries benefit from China and India opening to trade (the South becomes
larger), since they can specialise in sectors of higher technological advantage (e.g. PCs, ICT,
etc…) and import cheaper in sector with less advanced technology (e.g. textiles).
Now we will analyse what are the effects of technological progress.
Let us suppose that foreign relative productivity increases in all sectors:

• For a given relative wage, home stops producing the goods in
which it loses enough relative productivity z↓ (1→2).
• For given w/w* and specialisation in less goods, relative
demand falls.
• The relative wage has to fall reflecting the fall in home
productivity (w/w*)’ < w/w* (2→3).
• The foreign country specialises in more sectors and enjoys
higher relative wage due technological improvement.
Conclusion: z↓ and w/w*↓.
The effect is going to be similar as when L*/L↑ (actually, home real wage does not change):
-Home welfare:
• Is unchanged in terms of the goods that remain “home” (i ≤ z’).
• Increases in terms of the goods that switch to “foreign” (z’ < i ≤ z)
that become cheaper due efficient specialisation.
• Increases in terms of the goods that remain “foreign”:
it becomes more expensive due to lower relative wage (w/w*↓) but also
cheaper due to higher foreign productivity (a*Li↓). Overall, becomes cheaper as
|∆ (w/w*↓)| < |∆a*Li|: the relative wage decrease less because higher demand for home goods
raise wages at home as well.
-Foreign welfare:
• Increases in terms of the goods that remain “home” (i ≤ z’) as become
cheaper due to higher relative wage (w/w*)’ < w/w*.
• Increases in terms of the goods that switch to “foreign” (z’ < i ≤ z)
as become cheaper as foreign outweighs higher relative wage.
• Increases in terms of the goods that remain “foreign” (z ≤ i ≤ N) as
become cheaper due to higher foreign productivity.
Both trade partners gain if one experiences technological progress in all sectors, but the country that
experiences it gains more.
Now let us suppose (as in Samuelson [2004]) that home (North) starts with an absolute advantage in
all sectors (a*Li > aLi) for all i, but then foreign (South) catches up in technology in all sectors
(a*Li = aLi), so the A(i) becomes flat.
That will maximise world production and South will reach the level of GDP per capita of the North.
But there will be no reason for trade any more: North loses all gains from trade, as it goes back to
its closed-economy equilibrium (e.g. the US stop enjoying cheap textiles import from China).
When we suppose that there is a transport cost (and some goods could be lost on the way),
to get one unit of Foreign good, Home consumers have to buy t > 1 units (t = iceberg cost).
Home only buys Foreign goods if they are cheaper than Home’s:
At the same time, Foreign only buys Home goods if they are cheaper than Foreign's.
Before, we had an equilibrium z in which the relative productivities were

the same. As now:
There will be a region of z with goods that are not traded in the
international markets.
In this situation, the efficient specialization for given relative wage:
• Home produces goods i ≤ z. In z:
• Foreign produces goods z* ≤ i. In z*:
• As A(z∗)×t > A(z)/t , goods z < i < z* are non tradeable.
In order to find both z and z*, we will use the goods markets equilibrium
(balanced trade: home imports = home exports).
Therefore, we have two different A(i) functions and a B(i) which depends on two different z’s.
Empirical Evidences and Applications
The model has some limits:

• It predicts perfect specialisation, but that is not observed.
• As there is only one factor, we are unable to see the redistributive effects of trade.
• Is difficult to apply the baseline model to a world with many countries. How do we stablish
the relative productivities?
Anyway, all these assumptions are made for simplicity.
However, has some validity, as predicts that comparative (and not absolute) advantage determines
the pattern of trade: evidence from 1963 on the US and the UK in 26 sectors showed that US had
absolute advantage in all sectors but exported in only the half of them; and the US exported more in
the sectors where its relative productivity was higher.
Eaton and Kortum (2002) did some extensions: add many goods and countries, transport costs, and
instead of relative productivity, focus on the probability that a country is the most efficient producer.
It is very useful to evaluate quantitatively many different things, as we will see:
First, about quantifying the gains from trade. Due to trade barriers, countries used to consume much
more in domestic products. Nonetheless, the decrease of trade barriers from globalization lead to an
increase of the gains from trade. Actually, there was an effect of a 25% drop in trade costs that led
to gains positive and decreasing in country size.
Some countries transfer (lend) wealth to others (e.g., via trade surplus or capital flows), leading to
global imbalances (as China to US). But we were assuming trade balance on our model!
Suppose China (*) transfers T to the US. Also suppose:

• All goods are traded.
• Technology [A(i)] is unchanged.
• B (i, L*/L) in unchanged (as China and US spend T in the same way as before: 1/N in each
good).
Now, our trade balance equilibrium assumption requires

that US import – export equal the transfer.
That would not imply an effect on w/w* nor z, so there would be no effect on welfare.
We were assuming that all goods were traded, but if now we assume that goods z < i < z* are non-
tradeables, the trade balance equilibrium will require:
The result is that the transfer increases relative demand for US goods and hence its relative wage.
The intuition is the following: if the money transferred from Foreign to Home is not used on
tradeable goods (from Foreign) but on non-tradeable goods (from Home), the demand from Home
will increase relative to Foreign’s. However, transfers are unsustainable at a long term.
To eliminate global imbalances, some wage adjustment is needed.
The last application will be referring to technological progress. Some theories stated that many
countries benefited from China’s unbalanced technical progress. The reason was the following: as
China had a technology equal to the world average, if China adopts the more advanced US
technology, world technological diversity increases, which lead to gains from all.
Conclusion:
-Comparative advantage is based on technological diversity: a country specialises in the sectors

where its relative productivity more than compensates its relative wage.
-Trade benefits both trading partners.
-Trade is more beneficial for small countries.
-Technological progress in a country, if uniform, is beneficial for both trading partners. If biased
towards the sectors of its comparative disadvantage, may hurt the trading partner.
Increasing Returns to Scale (The Krugman Model). Intuition
In the neoclassical models (H-O, Ricardian), there is trade when there is comparative advantage.
That allows us to explain inter-industry trade between different countries (North – South).
Now we have an additional reason to trade: access to international markets allows to increase
production and decrease costs (IRS), as some goods are only viable in large scale because of large
fixed costs. That allows us to explain intra-industry trade between similar countries (North – North).
There will be new gains from trade:
• Consumers have more “varieties” to choose.

• Larger markets (i.e. more countries to sell) reduce costs.
• Pro-competitive gains: more competition (from abroad) reduces prices than if firms have
some monopolistic power.
Increasing Returns and Monopolistic Competition
Monopolistic Competition:
-Due to the large fixed costs, perfect competition (p = MC) implies losses.
-Therefore, we need another market structure. Why is called Monopolistic Competition.
-Competition: different firms produce the same good, but a different variety of it.
-Monopolistic: firms choose price given demand curve and have market power over their variety.
Consumption (ci):
-We obtain the demand for each variety i maximizing the utility function restrained to the budget
constraint (Lagrange).
Let us analyse the formula intuitively:

• The more the price of the variety i (pi), the less the consumption of i (ci).
• The more the “other” prices of the economy apart i (P), the more the consumption of i (ci).
• The more the wage (w), the more the consumption of i (ci).
Elasticity:
-Price-elasticity of demand (εp) represents the sensibility of the consumption of i (ci) to an increase
of the price (pi). The higher the price-elasticity (↑εp), the lower the market power (↑α).
-Elasticity of substitution between any two varieties (εij) represents the substitutability between
varieties. The higher, the more willing you are to substitute (↑εij), the lower the market power (↑α).
Price (pi):
-We obtain the price by maximizing the profits function. Once we derivate, we equalise MC=MR.
-Not only we obtain the price but also the quantity produced of each variety.
Let us analyse the formulas intuitively:

• Marginal cost (↑β) and market power (α↓) are increasing with price (pi) and decreasing with
quantity (qi).
• Price index (P) and market size (L) are increasing with quantity.
Profits (π). Free Entry:
-We assume that new firms (i.e. varieties) enter the sector as long as π>0.
Market Clearing. Varieties:

-We assume labour market clearing (supply of L workers = demand of workers in total production).
-That allow us to obtain the number of varieties. Let us analyse the formulas intuitively:
• Larger economies produce more varieties (L↑→n↑).
• The higher the fixed cost the fewer varieties (F↑→n↓).
• The higher the elasticity of substitution the fewer varieties (α↑→n↓).
Equilibrium:
-Even though we already found, the price for each firm and the number of varieties, we could also
find another equilibrium.
-Combining labour market clearing and free entry we obtain a price function increasing in varieties.
-The intersection with the two prices function, we would find another varieties equilibrium.
-Anyway, when plotting this formulas in an exercise we will obtain the same price (Problem Set 4).
Open Economy
-In this model, opening to trade leads to the following gains:

• As firms need to compensate the large fixed cost, producing becomes viable only with large
scale. This is accomplished thanks to trade, as it leads to achieving a larger market.
Thus, larger market, higher profits, more entry, leads more firms and varieties to produce.
• Consumers can consume more varieties so the utility becomes larger.
Pro-Competitive GFT
-Now we will assume that α is increasing in n: more firms in the market decrease market power.
-Let us imagine that there is an increase of market size (L):

-As there is more demand, the number of varieties increases.
-That leads to a decrease of the market power. Therefore, price decreases.
-As a consequence, profits will decrease, so some firms will exit.
Conclusion:
• The increase of n is lower (than without competitive gains). Less gains from variety.
• Prices fall. Consumers can consume more of each variety. That is the pro-competitive GFT.
Empirical Evidence
From the IIT index, we see that Intra-Industry Trade is higher for differentiated and high-tech
goods, and therefore, between advanced countries.
Increasing Returns to Scale (The Krugman Model). Formulas
Notation:
• w: wage. We will take it as a numeraire (w=1).
• F: fixed costs (e.g. building a plant implies F units of labour).
• β: variable costs (every unit of output costs β units of labour).
• α: it is related with elasticity and market power

(the higher α, the higher the sensibility of consumption, the higher the
substitutability between varieties, the lower the mark-up and market power).
• P: price index (price weighted average of the consumption implied by the utility
function. In other words, is all the “other” prices of the economy apart i).
Consumption:
We obtain the demand for each variety i maximizing the utility function restrained to the budget constraint (Lagrange).
Price:
We obtain the price maximizing the profits function. Once we derivate, we equalise MC=MR.
Profits:
By equalising profits to zero, we know the free entry point. In there, we obtain the scale of production by isolating q.
Varieties:
By the labour market clearing, we obtain the equilibrium number of n varieties.
Price:
Combining the labour market clearing and free entry we obtain the price equilibrium.
Open Economy:
Why there are Gains From Trade? Because UFT/UA > 1
Pattern of Trade:
Imports and exports
Pro-Competitive GFT:
To go from β/α to β/α(n) leads to lower increase of varieties and lower price.
IIT:
Implications on intra-industry.
Increasing Returns to Scale (The Krugman Model)
In the early neoclassical frameworks, there is trade when there is comparative advantage. Therefore,
trade exists because countries have differences in technology or factor endowments.
Although this is very useful to explain inter-industry and trade between “North” and “South”, in the
data most of the trade is between countries that are very similar (developed countries: intra-
industry).
To account for these empirical regularity we must have additional reasons to trade: access to
international markets allows to increase production and decrease costs i.e. there are Increasing
Returns to Scale (IRS).
To explain that, the New Trade Theory (Krugman et al, late 70’s and early 80’s) relies on increase
returns to scale at the firm level (combined with monopolistic competition). It makes sense,
because, some goods are only viable in large scale because of large fixed costs, so it requires
specialization to take advantage of large scale production.
It allows us to consider intra-industry trade (each country imports and exports different varieties of
the same good) and trade between similar countries (e.g., North-North).
We can model two different types of IRS:

• External Economies of Scale: the larger the industry, the lower the costs of that industry
(decreasing in cost depends on the size of the industry).
• Internal Economies of Scale: the larger the firm, the lower the costs of that firm (decreasing
in cost depends on the size of the firm). Thus, few large firms producing differentiated
products.
Now: let us consider differentiated goods within a sector (e.g. iPhone and Galaxy S, varieties
perceived as imperfect substitutes). Internal economies of scale imply that each variety is cheaper if
production is concentrated in one large firm, and that Apple is located at US and Samsung in Korea.
We will have new gains from trade:

• More “varieties” to choose (Americans [Koreans] preferring Samsung [Apple] are happier).
• Larger markets (i.e. more countries to sell) reduce costs.
• Pro-competitive gains: more competition (from abroad) reduces prices than if firms have
some monopolistic power.
Increasing Returns and Monopolistic Competition
Let us assume that production requires:

• Fixed input of F units of labour (e.g. building a plant).
• Variable input of β units of labour per unit of output.
• w = wage; Fw = fixed costs; βw = marginal cost (MC).
Now, take wage as the numeraire (w=1).

We obtain the average production cost (TC/q), which is decreasing in q. AC = F/q + β
Perfect competition requires that the price is equal to the marginal cost (p = β).
The profits obtained would be the following: π(q) = pq − F − βq = βq − F − βq = −F.
Selling at MC, firms would make losses. Therefore, no firm would want to produce.
If not competitive, what market structure?
-Monopoly: aggressive hypothesis, but only realistic in very specific markets.

-Oligopoly: interesting, but it involves complicated modelling strategies interaction between firms.
-Monopolistic competition, where:
• There is a monopoly price: firms choose price given demand curve.
• Each firm produces an individual and differentiated variety of the same good, and they have
market power over that variety.
• As there are many firms, there are no strategic interaction: although demand for every
variety depends on all prices, each individual firm is atomistic and ignore the decision of the
others.
Once that is said, let us assume there are n firms in the sector, with:
• Each firm producing a different variety of the same good (imperfect substitutes) and having
market power over it (monopoly).
• Each firm chooses price to maximize profit taking the demand for its variety as given (as in
monopoly) and without considering the effect of its price on market conditions (as in perfect
competition).
Consumption
We will consider a country with L consumers, which draw utility from the n varieties:
The utility function is the sum of the consumption of all goods.
We assume consumption of each variety ci = c = C/n (equal shares of total consumption).
Then, U = n(C/n)α = Cα n1-α, increasing in n since α ∈ (0, 1).
Therefore, the individual demand of each variety i is the solution to:
To obtain the demand, we set the Lagrangian:

(with the utility function and the budget constraint).
Then, rearranging the First Order Condition, we obtain: αciα-1 = λpi . (For a good j: αcjα-1 = λpj ).
Dividing both expressions (α and λ are gone), we obtain the demand of variety i relative to j:
After a lot of rearranging (not needed to know), we obtain the demand for each variety i:
(where P =[ piα/(1-α)]-(1-α)/α is the price index (note: P decreasing in n).
The price index is the price weighted average of the consumption implied by the utility function. In
other words, is all the “other” prices of the economy apart i.
The more varieties n, the less the index price P. The less P, the less ci (consumption of i), as if the
price of “other” varieties decrease you substitute the variety i for the others.
α is related with the elasticity:
-The price-elasticity of demand is:

(if α↑, εp↑: it represents the sensibility of the consumption of i (c ) to an increase of the price (p )).
i i
-The elasticity of substitution between any two varieties is:

(if α↑, εij↑: it represents the substitutability between varieties. The higher, the more willing you are to substitute.)
Price:
Once we have seen the demand, now we will study the supply at the equilibrium.
All firms have the same technology (i.e. fixed cost F and marginal cost β) .
As every firm has a monopoly over one variety, firm producing variety i will choose its quantity or
price in order to maximise profit πi, given aggregate demand for their variety qi = Lci and w=1.
As is a monopoly, we can maximise profits over price or quantity.

The F.O.C. requires that MR = MC. As Revenue is p(q)×q (due monopoly)→
In Perfect Competition we take price as given and revenue is p×q, so MR=p.
It can be seen graphically:

→If q↑, MR↓ and p↓→ ‘p/’q < 0. Therefore, MR decreasing.
→If q↑, AC will tend to β (and to MC).
→TR = p*×q*. TC = AC(q) × q.
After a lot of rearranging, we obtain that MR = MC

Hence:
So, the price is going to be equal to perfect competition price multiplied by the
mark-up, which is a measure of market power.
Imagine the elasticity of substitution increases: firm will have less market power (α↑), therefore will
be able to charge lower mark-up, and the price will be lower (p↓).
If we assume same technology (β) and same isoelastic demand (constant elasticity, same α) in all
firms, there will be the same price (pi=p) and same scale (qi=q=(P/p)1/(1-α) L/P).
Profits. Free Entry

Therefore, substituting q and p in profits:
(β (marginal cost) , F (fixed cost) ↑; π↓. L (market size) and P (price index)↑, π↑).
As p= β/α, we can rewrite profit as:
Now, we assume the free entry condition: new firms (i.e. varieties) enter the sector as long as π>0.
In equilibrium, entry drives profit to zero (π = 0).
That leads us to obtain the scale of production q:
Market Clearing. Varieties. Then, to obtain the equilibrium number of varieties n, we impose labour
market clearing (supply of L workers is equal to the demand of workers in the overall production).
Even though there is a lot of maths, the more important part is the intuition:
Equilibrium:
Optimality conditions of consumers and firms:

(optimal consumption/demand for each variety i and optimal price set by firms)
We have assumed two conditions: goods market and labour market clearing:
(first: production of the firm = total demand of the firm. second: demand of workers = supply of workers)
We also assume free entry until π = 0:
Combining labour market clearing and free entry we obtain:
Therefore, we obtain the equilibrium price.

Even though, there are other ways of plotting the Krugman equilibrium.
Open Economy
Now we will consider two countries, which have:

• Same technology (F, β) and preferences (α). Therefore, all firms charge same price (p=p*)
and produce same quantities (q=q*).
• Possible different country size (L≠L*) and different number of varieties:
That is the intuition of IRS: larger market, higher profits, more entry, leads more firms and varieties
to produce.
Departing from the usual neoclassical model, there is no comparative advantage. Therefore, there is
no reasons from trade! Where are the gains from trade in this model?
When we open to trade, the variety i is consumed in both countries. As now firm has to produce to
to serve both countries, there will be a larger scale.
Also, as consumers can consume more varieties, the utility becomes larger:
If n↑→U↑ and 1/L↑→U↓, how we know utility becomes larger?

After a lot of rearranging, we obtain that the utility is higher under
free trade, and consequently, there are Gains From Trade:
Therefore, we obtain a new type of GTF: gains from variety. From the formula, we deduce that:
• Gains are lower if varieties are better substitutes:
• Gains are higher for smaller countries:

The pattern of trade consist on each country exporting its varieties and importing the foreign ones.
where
As a result, all firms in both countries are exporters.
Now, trade will not happen due to comparative advantage, but to achieve a larger market.
Trade will be intra-industry trade: exports and imports of same good, but different varieties.
Pro-Competitive GFT
Let us consider another possible gain from trade. Now, in a pro-competitive situation, we will
assume that α is increasing in n: α(n). Therefore, changes in n affect prices: p = β / α(n)
Thus, more firms in the market decrease market power of monopolists (mark-up)→ ↑n → ↓ 1/α.
For instance, smartphones become better substitutes as more varieties enter the market. The
monopoly power will erode as n increases, so mark-ups and prices will fall: n↑ → α↑ → p↓.
As a consequence, profits will fall and there will be less entry, and, as a result, a decrease in variety.
Now, imagine that there is an increase of L. Theoretically, that increases the number of varieties.
With competitive gains, though, the increase of n is lower.
Thus, even though nFT + nFT* > nA, nFT < nA and n*FT < n*A . So, there will be less gains from variety,
but prices fall and we are able to consume more of each variety. That is the pro-competitive GFT.
It can be all pretty summarized

with this two graphs. Highlights:
• p = β / α(n) instead of p = β / α
• n’’<n’ (lower increase of n)
• Decrease of p* with PC GFT
The lower gains from variety is compensated with the fact that consumers can consume more of
each variety. Therefore, the gains can be even larger.
Empirical Evidence
To measure Intra-Industry Trade (IIT) between two

countries (or home vs rest of the world), we use the
Grubel and Lloyd index:
• If e and i are very similar, IIT is close to 1 and there is lots of intra-industry trade.
• If one of e or i is very large and the other close to 0, IIT≈0. No intra-industry trade.
In general, IIT is higher for differentiated and high-tech goods.
Nonetheless, the IIT index has two potential limitations:

• The less disaggregated (agg.: manufacture > disagg.: wood manufacture) the sectors, the higher IIT.
• IIT does not distinguish between intermediates (engines) and final goods (cars) within a sector.
The solution would be to compute two different indexes computed on super-disaggregated data:
• “Vertical” IIT: intermediate goods imported and exported in the same industry.
• “Horizontal” IIT: similar final goods imported and exported in the same industry.
Both IIT are very predominant between similar (advanced) countries. In contrast, between
developed and developing countries (North-South), there is predominance of unidirectional trade.
Summary
-This model combines IRS and differentiated goods, leading to monopolistic competition.
Monopolist’s price will be decreasing in substitutability.
-Now, the reason for trade is not comparative advantage, but the following: larger markets lead to
more varieties, which makes consumer happier; and the effect of trade is to increase the market size.
Therefore, the gains from trade comes from the gains from variety.
-The pattern of specialization will be the following:
• Each country specializes in a number of different varieties depending on its size.
• Each country exports all domestic and imports all foreign varieties: intra-industry trade.
• Smaller countries benefit more from trade.

Firm Heterogeneity (The Melitz Model). Intuition
The last model seen predicts that all firms export, but in reality, only more productive export.
Melitz Model take into account that firms trade (not countries) and to access a new market is costly.
The results show that a trade liberalization leads to a “selection of the best fit”.
A Model with Firm Heterogeneity
Productivity (φ):
-Firms will differ in productivity (φ), and therefore, in MC (1/φ). There is always a fixed cost, but
the one from exporting is larger than the one from only serving the domestic market.
Characteristics of the market (A):

-We can rearrange the profits with A. Price index (P) and market size (L) are increasing with profits.
Makes sense: the higher they are the more profitable the market.
Note: the formula shows that α is also increasing with profits. But, how a decrease of market power
leads to an increase of the profits?
Productivity cut-off (φ*D):

-Is the level of productivity required (φ > φ*D ) so a firm can survive in the market by having π ≥ 0.
If your productivity is below the cut-off (φ < φ*D), the firm will exit.
By analysing the formula, we observe that:

• If A is high, the market is profitable: for little productivity, you can have profits. A↑→φ*D ↓.
• If the fixed cost is high, you have to be productive to pay it. For this reason, FD ↑→ φ*D ↑.
Free Entry Condition:

-Firms will enter the market until the expected profits are equal to the entry cost.
Open Economy
Additional costs:
-Serving to a foreign market implies:
• A new fixed cost (fX), larger than the fixed cost of serving at the domestic market (fX > fD).
• An iceberg cost: ship τ > 1 units to deliver 1 unit of the good. Therefore, MC = τ / φ.
-As costs are higher, the profits are lower. Both πD and πX increase with φ, but πX by less due to τ.
-Therefore, the productivity cut-off of exporting is higher than the one from serving at the domestic
market (φ*X > φ*D). Thus, only the most productive firms export, as they must be big enough to
profitably cover the fixed export cost.
Selection effect:
-Open to trade leads to the more productive (so they can export: φ*X > φ) foreign firms entering in
the domestic market. Due to foreign competition:
• Competition increases, more productivity will be required, φ*D will increase.
• The market becomes more profitable and profits are lower than in autarky AA>AT, πDA>πDT .
-Consequently:
• Some domestic producers and small exporters (exporting will not compensate the losses
from competition) will lose. Thus, some firms exits.
• Big exporters win.
-What happens when we move to autarky to open economy has the same interpretation if there is a
trade liberalisation in which trade costs (τ and/or fX) fall.
Results from opening to trade:
-We gain from foreign (high-productive) varieties but lose domestic (low-productive) varieties.
-Therefore, the average of productivity of our country increases.
-As high productivity means lower marginal cost (and viceversa), consumers now will consumer
cheaper varieties.
-Overall, welfare increases.
Empirical Evidence
There are two studies that test the veracity of the theory.
In one, it is seen that trade liberalisation:

• Raises productivity of each sector due to import competition by inducing the least
productive firms to exit, as we saw in the theory.
• Raises productivity of exporting firms: A↑ justifies investment in technology upgrading φ↑.
It was not predicted, as we assumed constant φ in firms.
• Reduces employment (not predicted, as we assumed frictionless labour market).
In the other, as theory exposed:

-More productive firms (φ↑) can cover more fixed costs (fX), and consequently, export to more
countries, export more products, and export disproportionately more.
-Larger markets (L↑) deliver higher profits (L↑→AX↑→πX), as they attract more exporters, exporters
choose them first, and less productive exporters decide to just export to them.
Therefore, this model is powerful to explain many empirical regularities.

Firm Heterogeneity (The Melitz Model). Formulas
-Marginal Cost: MC = 1 / φ where φ is productivity, different for each firm.
-Profits (q chosen given the demand):
-Domestic profits (of a firm with productivity φ, using p = 1 / αφ) :
-Productivity cut-off (i.e. zero profit cut-off, where π(φ*D) = 0):

(at which level of productivity φ a firm make profits (πD ≥ 0), and survives).
-Expected profits (when to enter we must pay an entry cost f E for given cut-off productivity φ*D. Before we did not.):
-Free entry condition (firms enter in the markets until the expected profits are equal to the entry cost):
Open Economy: two new costs.→ New fixed cost (fX) and iceberg cost τ > 1. Therefore: MC = τ / φ.
-Productivity cut-off (i.e. zero profit cut-off, where π(φ*X) = 0):

(at which level of productivity φ an enter make profits by exporting (πX > 0), so exports)
-Winners and losers from trade:

Firm Heterogeneity (The Melitz Model)
The New Trade Theory presented in the previous topic showed the intra-industry trade and the GFT
which where result of the IRS, consequence of the decrease on prices due to larger markets.
There was:
• n varieties of differentiated goods with total cost function: TC = (F + βq)w
• L consumer with “love of variety” and demand: qi = (P/pi)1/(1-α)wL/P with α: (0,1)
And the equilibrium was:

• Price: from profit maximization: pi = p = wβ / α (MC* mark-up)
• Quantities: from free entry: qi = q = αF/ [β(1 – α)]
• Number of varieties: from labour market clearing n = L(1 – α) / F
The model with IRS and differentiated goods predicts that all firms export, but in reality, only few
of them export. In US (2007), an 18% of manufacturing firms export, and these export a 14% of
total shipment. Moreover, exporters are bigger and more productive (among other things) than non-
exporters, and larger exporters export more.
Melitz took these facts into account to modify the previous model accompanied by the idea that to
access a new market is costly and countries do not trade, firms do. The new result consist on the
fact that trade liberalization benefit the large firms and harms the small, leading to some exit.
Therefore, there is a “selection of the best fit”.
A Model with Firm Heterogeneity
There will be the same preferences and market structure as the previous model: monopolistic
competition (firms have market power in the variety that produce), n-goods, constant elasticity of
substitution between them and we will take again wage as a numeraire (w=1).
But we will have new assumptions:
• There is differences of productivity across firms: that generates heterogeneity.

• There is a fixed export cost, which only the most productive firms are willing to pay.
• There is also a variable (iceberg) trade cost, but it is not crucial.
We will study the reduction in trade costs when there is trade liberalization between two symmetric
countries.
Firms will differ (exogenously) in productivity φ and in MC = 1/φ (as we will integrate the
productivity in the marginal cost). Thus, the more productive, the less marginal cost.
Therefore, the total cost of a firm with productivity φ:
As firms are monopolistically competitive, they choose the quantity q to maximise profits given the
demand for the variety (as in Krugman’s model).
Therefore, the more productive firms (↑φ) will

charge lower prices, sell more, make higher profits.
-As consumers have same preferences, the Marginal Revenue
is the same for all firms (optimal profits at MC=MR).
-Instead, the MC is different.

It depends on productivity: ↑φ →↓MCi .
-As Firm 1 is more productive (↑φ →↓MCi), it will be able to charge

a lower price (↓pi), sell more (↑qi) and have more profits (↑πi).
Let us assume that firms can shut down at no cost (free exit) given profits:
Firms will stay if π ≥ 0, otherwise exit. Therefore, only the most productive firms (φ > φ*D ) survive.
There will be a productivity cut-off where π(φ*D) = 0, where zero profit cut-off is:
It can be seen graphically:

-The more φ, the more profits. If φ=0, π= FD.
-If A is high, the market is profitable.
Thus, for little productivity, you can have profits. A↑→φ*D ↓.
-If the fixed cost is high, you have to be productive to pay it.
For this reason, FD ↑→ φ*D ↑.
We still need to find A: recall that A depends on the price index P;

and therefore, on n. In Krugman we could find n by the free entry condition.
Similarly, we need an additional condition to jointly determine A and φ*D.
Now, the free entry condition is over the expected profits. The intuition is the following: if we want
to know if a project is good or not, we will test it “in the market”. If it is bad, we exit (φ < φ*D).
Then, suppose to enter we must pay an entry cost fE for given cut-off productivity φ*D .
The expected profits would be:
Logically, the free entry condition will consist on the following: firms enter in the markets until the
expected profits are equal to the entry cost:
As said before, cut-off equilibrium is:

• Increasing in fD : if fixed costs increases, more difficult to survive, more productivity needed.
• Decreasing in fE: if entry cost increases, higher expected profits are needed for firms to
enter, for which is needed a low cut-off.
• Negatively correlated with A: more profitable markets, easier to survive, less φ needed.
A priori, we have the n that we were needing to find A. But we would need to define distribution of
ideas (g), which we will not. Instead, we will focus on the cut-off (φ*D) and how trade changes it.
Open Economy
Suppose now that a firm can sell into a foreign market denoted by (x) and assume there are two
symmetric countries with same size, technology and preferences → A = AX.
Moreover, there are two additional costs to serve a foreign market:
• A new fixed cost (distribution and servicing costs): fX .
• An iceberg cost: ship τ > 1 units to deliver 1 unit of the good. Therefore, MC = τ / φ.
The price of exported goods now are:
The profit from serving the foreign country will be:
Rearranging, we obtain that profits from exporting are:
Again, a firm decides to export only if πX > 0. Therefore, there

is a cut-off φ*X such that firms with φ > φ*X decide to export:
We will assume that the profits will be higher in the domestic than the
foreign market, even though the level of productivity is the same
[πD(φ) > πX(φ)], as τα/(1- α) fX > fD . Consequently, no firm prefers
exporting than serving the domestic market.
This can be seen graphically with this graph:
The intuition is the following: when exporting, the costs are

higher, the profits are lower, and the productivity required is larger.
Both profits functions increase with productivity, but πX by less (due to τ).
We divide the firms in three groups:

• Firms with productivity below φ*D exit. They do not enter in none market.
• Firms between φ*D and φ*X , which produce in the domestic market.
• Firms above φ*X, which also export.
Thus, only the most productive firms export, as they must be big enough to profitably cover the
fixed export cost.
From φ*D to φ*X , there are the firms which produce in domestic market.
Their profits will be πD(φ).
From φ*X to φα/(1- α), there are the firms which also export.
Their profits will be πD(φ) + πX(φ).
Therefore, profits increase in productivity, and profits of exporters increase faster.
Open to trade also increases competition: as more productivity will be required, φ*D will increase.
As foreign exporters enter the domestic market, some will be more productive than domestic non-
exporters. As market will become more competitive, A↓.
Due to that foreign penetration, non-exporters will lose domestic sales, while they do not gain
market shares in the foreign market. Therefore, marginal (least-productive) firms are forced to exit.
Consequently, there will be a selection effect: fewer firms per country and higher average
productivity of survivors.
Something similar would happen if the costs of export (τ and/or fX) fall.
In this graphic, the red line are domestic profits in autarky (πDA) while blue are
the ones obtained when we open to trade (πDT first, πDT+ πX at φ*X).
Due to foreign competition:

• The profits are lower when we open to trade: as AA > AT, πDA > πDT.
• More productivity is required when we open to trade [(φ*D)T > (φ*D)A].
Consequently:
• Domestic producers and small exporters (exporting will not compensate
the losses from competition) will lose. Thus, some firms exits.
• On the other hand, the big exporters will win.
(As they have to increase their production labour will be reallocated from losers to winners).
The result of opening to trade is the following:
-Varieties effect: now we will be able to import foreign varieties (which is positive), but some
domestic varieties (firms) will disappear (negative).
-Welfare: while the new imported varieties have lower marginal cost (higher productivity), and
therefore, cheaper; lost domestic varieties have higher marginal cost (lower productivity and as a
result, are more expensive. Thus, the average of productivity of our country increases. So, overall,
welfare improves.
Summary
When firms are heterogeneous in productivity (φ):

• More productive firms charge lower prices, sell more, and make more profits.
• Least productive firms exit the market.
When we switch to free trade (τ and/or fX);

• Most productive firms become exporters
• Most productive foreign firms enter domestic market.
• Least productive domestic firms exit the market (selection).
• Gains from variety: gain foreign (high-productive) varieties, lose domestic (low-productive)
varieties.
Empirical Evidence
Summarizing, if there is a trade liberalization, and consequently, trade costs (τ and/or fX) fall:
• There will be higher profits from exporting (↑πX) and a lower cut-off i.e. less productivity required
to “survive” (φ*X↓). New and old exporters gain.
• There will be lower profits in the domestic market due to more foreign competition (↓πD) and higher
cut-off i.e. more “selection” as more productivity required (φ*D↑). Least productive non-exporters
lose.
We can get some empirically relevant predictions:
• The average productivity will increase at industry level due to selection.
• Increases in sales (output) of exporters.
• Drop in sales (output) of import-competing firms.
Let us see if this predictions coincide with the empirical evidence. We will observe the Trefler (2004) paper,
which considers the Canada-US free trade agreement (CUFSTA, 1989) and uses data on sectors and plants
(firms) in Canada (80-96). For different reasons, is a good example.
To identify the variation on yi due to the liberalization

(reduction of the tariffs τ), the paper runs a regression.
That led to different results in Canada:
-Employment: at first there was a reduction of the 5% at plant and at industry level due to the competition
from US exporters (consequence of τCAN ↓). But that was a transitory effect, guiding to a shift in the work
force: the skilled labour increased relative to the unskilled (skl/unskl↑).
-Labour productivity: due to the “selection” effect (exit of the least productive) there was a +15% at industry
level (0 at plant level as φ remained constant). It also seems that there was an increase on productivity due to
τUS ↓. Maybe because it led to import of better machinery?
-Production: the possibility of exporting at US (due to τUS ↓) led to +6% production at plant level.
So, trade liberalization:

• Raises productivity of each sector due to import competition by inducing the least productive firms
to exit, as we saw in the theory.
• Raises productivity of exporting firms: A↑ justifies investment in technology upgrading φ↑. It was
not predicted, as we assumed constant φ in firms.
• Reduces employment (not predicted, as we assumed frictionless labour market).
On the other hand, Bernard et al. (2015) observed that in US (in 2007) few firms export to more than one
market (1) and more than one product (2). But those who did (2) also did (1), and represented an important
share of total export.
Therefore: larger firms are much more likely to export, and export much more. Moreover, larger markets
attract more exporters and imply more sales per exporter.
Recalling this with the theory:
More productive firms (φ↑) can cover more fixed costs (fX), and consequently, export to more countries,
export more products, and export disproportionately more.
Larger markets (L↑) deliver higher profits (L↑→AX↑→πX), as they attract more exporters, exporters choose
them first, and less productive exporters decide to just export to them.
Taking stock: countries do not trade, firms trade! In this model firms are heterogeneous in their productivity:
trade benefits the most productive, and the least productive do not survive and exit. This model is powerful
to explain many empirical regularities.
FDI (Multinationals). Intuition
In the last topic, we saw how only the larger and more productive firms could export.
Now, we will see how multinationals are larger and more productive than usual exporters.
The expansions of multinationals are done with Foreign Direct Investment (FDI): by setting up a
new firm abroad (“Greenfield”) or by investing in existing plants (“Brownfield”).
Data Facts about FDI and Multinationals
-FDI inflows have increase a lot over time (since 90’s), mostly to developed countries.
-Between advanced countries there is high FDI inflows and outflows (bilateral) while developing
countries are more likely to receive (from rich countries) FDI than to do it (unilateral).
-The more K/L and the more innovation, the more intra-trade firm there will be.
-MNC’s subsidiaries are larger and more productive relative to local firms.
-The more technological-advanced or capital-intensive the sector, the more horizontal FDI.
Horizontal FDI
-Consists on replicating the entire production process in an affiliate to sell in a foreign market.
-It is usually done between similar countries (North – North FDI).
Proximity vs Concentration Trade-Off:

-Is the dilemma that emerges when we choose between export and FDI.
-Multinationals face higher fixed costs (and lose “scale”) but save transportation costs.
Horizontal FDI Productivity Cut-Off (φ*I):

- πI depart from a lower level (fI > fX), but grows faster (πX‘s slope < πI’s due to τ).
-Therefore, at first πX > πI, but at some point there will be a level of productivity (φ*I) for which FDI
is more profitable than exporting (πI > πX).
-As φ*I > φ*X > φ*D, the multinationals will be the most productive firms.
The empirical evidence show that this is true.
Vertical FDI
-Consists on moving part of the production process to another country in order to save costs.
-It is usually done between different countries (North – South FDI).
Trade-Off:
-Firms decide if they produce only in their home country (North) or they also do it partially in a
foreign country (South), which would be doing FDI.
-Again, FDI implies a trade-off between a lower variable cost and a higher fixed cost.
-So, firms will choose FDI if they save more in variable cost than in fixed cost, and thus, πS(FDI)>πN.
Vertical FDI Productivity Cut-Off (φ*VI):
-πS(FDI) depart from a lower level (fS > fN), but grows faster (πS(FDI)’s slope > πN‘s due to wS < 1).
-Therefore, at first πN > πS(FDI), but at some point there will be a level of productivity (φ*VI) for which
producing partially in the South is more profitable than only producing in the North (πS(FDI) > πN).
-Multinationals which do Vertical FDI are more productive (φ > φ*VI), since doing Vertical FDI
requires a lot of productivity to cover the large fixed costs of setting a subsidiary abroad (fS).
Vertical FDI is more likely (φ*VI ↓) if:

• The comparative advantage is strong (as low wS).
• Set-up cost abroad is relatively slow (low fS – fN).
• Share of variable cost in production (β) is high e.g. Indian workers can do a lot of
production that is done in the US, so there can be more profit from the cheap labour in India.
• If the countries are closer so τ is lower (we could multiply τ by the variable cost).
Offshoring, FDI and Licensing/Outsourcing
-There is an alternative to FDI: the internalization decision. It can be made:
• Instead of Horizontal FDI: licensing. Implies saving in fixed production cost, but the risk of
imitation/loss of exclusive technology overweights it. Therefore, we observe little licensing.
• Instead of Vertical FDI: outsourcing. Again, allows savings cost, but there are critical factors
(e.g. hold-up problem). As is much more beneficial, we observe much more outsourcing.
The more advanced the sector, the more FDI and less outsourcing (and vice-versa).
FDI (Multinationals). Formulas
Horizontal FDI
-Characteristics:
• Higher fixed cost (fI > fX)
• No additional variable cost (i.e. no τ).
• MC = 1/φ.
-Cut-off exporting vs FDI:

(determines if a firm exports or becomes a multinational)
Vertical FDI
-Characteristics:
• WN = 1, wS < 1.
• fS > fN
• 1 – β: tasks performed only at North.

• β: tasks that can be performed at South.
-Marginal Costs:
-Total Costs of one unit of q:
-Profits:
-Cut-off: no FDI vs FDI

FDI (Multinationals)
In the last topic we saw that trade is concentrated in very large firms, which export more quantity,
more products and to more countries (top 1% exporters account for 81% of US exports in 2000).
More importantly, a lot trade (90% in US exports) is accounted by multinational firms.
Moreover, one-half of US imports are transacted within the boundaries of multinational firms.
A multinational is an enterprise that controls and manages production establishments located in at

least two countries. It is usually defined by the:
• Parent firm or the headquarter, located at the source country.
• Affiliates, located in the host country.
Then, the expansion of multinationals are done basically with Foreign Direct Investment, which is
an investment made by a firm or individual in one country into business interests located in another
country. It can be:
• “Greenfield”: set up a new firm abroad. Generally, is costlier.
• “Brownfield”: investing in existing plants (mergers and acquisitions).
Data Facts about FDI and Multinationals
There has been a large increase in FDI inflows over time, mostly to developed
countries.
Advanced countries do (1: outflows) much more FDI than developing, but just receive (2: inflows)
slightly more FDI.
(1)
(2)
Therefore, between advanced countries there is high FDI inflows and outflows (bilateral) while
developing countries are more likely to receive (from rich countries) FDI than to do it (unilateral).
There is a lot of intra-firm trade in MNCs between headquarters and

subsidiaries. The pattern is:
• Intermediate goods for low capital intensive sectors (textiles and
shoes) are imported from external firms.
• Intermediate goods for high capital intensive sectors
(pharmaceuticals and machinery) are imported from a subsidiary.
The more K/L and the more innovation, the more intra-trade firm there will be.
Relative to local firms, MNC’s subsidiaries are larger (by employment

and sales), are more productive, invest more in R&D and are more
export oriented, as it can be seen in this table.
While headquarters specialise in R&D and product design, subsidiaries:
• Mainly sell in their own markets (horizontal FDI) e.g. Italian firm sets up
in Spain and sells there (done between developed countries).
• Second, export to other countries (hub for horizontal FDI) e.g. from Vigo,
export to Portugal (done between developed countries).
• Third, export back to the home country (vertical FDI) e.g. Nike sets up a factory in
Vietnam, and from there, the shoes are exported back to the US.
The more technological-advanced or capital-intensive the sector, the more horizontal FDI, and vice-versa.
“Greenfield” and “Brownfield” represent an equal percentage of FDI flows at world level, but
mergers and acquisitions (B) are majority (68%) in advanced countries while setting up a new firm
abroad (G) is more common (82%) in least developed countries (LDCs).
Why do some firms find it optimal to operate in more than one country? The answer depends on the
production activities that the affiliate carry on. Therefore, taking into account that:
• In Horizontal FDI firms replicate the entire production process in an affiliate to sell in a
foreign market, it is usually done between similar countries (North-North FDI).
• In Vertical FDI firms break-up the production in different countries to save cost, it is usually
done between different countries (North-South FDI). Furthermore, production sometimes is
exported back to headquarter countries.
Horizontal FDI
As exporting can be potentially costly (high transportation costs, tariffs, etc.), set up production
closer to the consumer base (saves the iceberg costs) can be an improvement.
The problem is that is costly to set a plant in a new country due to high fixed costs (new factory,
bureaucracy, etc.) and we lose “scale” (with exporting we had IRS due to fixed costs savings).
This is called the proximity-concentration trade-off.
We will extend the Melitz model by formalizing the choice between export and FDI (horizontal:
firms settle in the foreign country to produce & sell there) within the same model.
Horizontal FDI will entail:

• A fixed cost (fI > fX) to settle production in the foreign country.
• No additional variable cost (i.e. no τ). MC = 1/φ.
Therefore:
The proximity vs concentration trade-off is the following: multinationals face higher fixed costs, but
save transportation costs. A firm will choose FDI rather than exporting when πI > πX .
We assume that price of factors (i.e. wages) are the same since countries are similar (North-North
FDI. Therefore, if the firm makes the decision of FDI or not is not due to the costs savings.
We also assume same preferences α.
We have profits from domestic sales, from export, and from FDI:
Under the assumption that fI is large enough (recall: fI > fX > fD), we divide firms in these groups:
• Firms with productivity below φ*D: they will exit.

• Firms between φ*D and φ*X: they will only produce in the
domestic market. Profits: πD.
• Firms above φ*X : they will also export (apart of producing
in the domestic market). Profits: πD + πX.
• Firms above φ*I : they will become multinationals,
producing in more than one market. Profits: πD + πI .
πX has a flatter slope than πI and πD due to τ.
The cut-off which determines if a firm exports or becomes a multinational can be rearranged as:
To sum up, productivity of multinationals (exporters or not) is φ*I > φ*X > φ*D .
Therefore, the multinationals will be the most productive firms.
To see if that is true or not, Helpman et al. (2004) estimated on US data for firms of 52 sectors.
It is true: data showed that multinationals are 53.7% more productive (have larger labour
productivity) than non-exporters, while exporters non-multinationals are 38% more productive than
non-exporters, so the fact of being a multinational implies 15% more productivity than not being it.
Now let us see what happens with sales. It depends on trade iceberg cost (τ) and fixed FDI cost (fI).
The theory shows that if τ is low and fI, there will be more export relative to multinational sales.
To see if that is true, we will use (again) empirical evidence of Helpman et al. (2004) that consists
on estimation on the same data and also subsidiaries in 38 countries. Again, it is true.
Vertical FDI
Horizontal FDI is something that happens between advanced countries when a firm wants to
produce in another country. On the other hand, we have Vertical FDI, where multinationals wants to
save costs by moving part of the production process to another country.
Vertical FDI has:

• A large fixed cost of setting a subsidiary abroad.
• The benefit of lower variable production cost due to the comparative advantage on labour of
the country to which the offshoring is done (workers have lower salaries).
We will extend the Melitz model to analyse the choice of vertical FDI.
In Horizontal FDI, we assumed that wages were equivalent between countries.

Now, the wage in the advanced country is WN = 1 and in the developing country is wS < 1.
The production process will be divided in:
• 1 – β: share of “headquarters” tasks (e.g. R&D, marketing, design, etc.), which can be
performed only in the headquarters in the North with local labour with wN = 1.
• β: share of manual production tasks, which can be performed also in a subsidiary in the
South with local labour with wS < 1. When the product is “sent back” to the headquarters,
the cost is τ × wS .
We were assuming that MC = 1/φ, because MC = w/φ and w = 1.

Therefore, the cost of producing in North (N) vs South (S) is now:
The fixed cost of producing in the North is lower than do it in the South: fS > fN.
Then, the total cost of one unit of q (taking wN = 1) is:
So, FDI implies a trade-off between a lower variable cost and a higher fixed cost. Therefore, firms
choose FDI if it saves more in variable cost than in fixed cost, and consequently, πS > πN.
We can find a cut-off such that if φ > φ*VI , firms choose FDI:
The cut-off can be seen graphically:
• If φ < φD , the firm will exit.

• If φD < φ <φI , the firm will produce at home: no Vertical FDI.
• If φI < φ, the firm will produce at foreign: Vertical FDI.
Recall: slopeD: A, slopeI: A/wS (wS < 1).
Therefore, multinationals which do Vertical FDI (by sending the less advanced part of its
production process to a developing country) are more productive (φ > φ*VI), since doing Vertical
FDI requires a lot of productivity to cover the large fixed costs of setting a subsidiary abroad (fS).
Vertical FDI is more likely (φ*VI ↓) if:

• The comparative advantage is strong (as low wS).
• Set-up cost abroad is relatively slow (low fS – fN).
• Share of variable cost in production (β) is high e.g. Indian workers can do a lot of
production that is done in the US, so there can be more profit from the cheap labour in India.
• If the countries are closer so τ is lower (we could multiply τ by the variable cost).
Offshoring, FDI and Licensing/Outsourcing
Previously, we discussed the location decision: to produce in the home country or in the foreign
when the parent firm owns the affiliate.
There is an alternative: the domestic firm could write a contract with a foreign firm, which is the
internalization decision. It can be done by:
• Licensing (instead of horizontal FDI) e.g. a Catalan firm write a contract with an Italian firm in
which the C builds the infrastructure, I sells to its market and sends part of the profit back to C.
• Outsourcing (instead of vertical FDI) e.g. Nike pays a Vietnamese factory to produce shoes,
which are sent back to Nike who sells them.
How to choose licensing vs horizontal FDI. There is a trade-off, as licensing implies:

• Benefit: saving in fixed production cost.
• Cost: risk of imitation/loss of exclusive technology, which is negative.
In general, risk overweights cost savings.
On the other hand, when choosing between outsourcing or vertical FDI there is a trade-off, as
outsourcing implies:
• Benefits: there are cost saving as foreign providers can specialise and serve more costumers
(economies of scale).
• Cost: there are critical factors/processes, as it is more likely costly ex-post renegotiation
with the provider (hold-up problem).
Offshoring means acquiring intermediate goods and/or services from abroad. Therefore, offshoring
groups together both outsourcing and vertical FDI. It is important, as trade in intermediate goods
represent the 40% of world trade in manufacturing.
Few firms do licensing, but the decision of outsourcing vs vertical FDI is challenging. It depends on
the sector:
• There is FDI in sectors with high capital and R&D intensity.
• There is outsourcing in textiles, accessories, shoes…
Taking Stock:
Multinationals are larger and more productive than usual exporters.
We extended the Melitz Model to account for this regularity. The Model predicts that large firms do
more horizontal and vertical FDI.
The determinants of horizontal FDI are:

• Transportation costs/tariffs.
• Fixed cost of production in the foreign country.
The determinants of vertical FDI are:

• Differences of factor price (wS), technology (β) and the fixed cost of production in the
foreign country.
Topics in Trade Policy. Intuition
The models we study do not include many elements that matter in reality, which limits our analysis.
Dynamic Gains and Losses
-We have seen that the effects of any action happens immediately, so are static. In reality, those
effects happen along time, so are dynamic.
-For instance, with a trade liberalisation. We expect that everything adjusts quickly and the effects
that we were trying to achieve are obtained very soon.
-In the “China shock” (import competition) static models predicts that labour should reallocate from
these industries to other industries.
-In reality, the cost of adjustment is high: results show that there has been lower wages and higher
unemployment, with social consequences. In fact, an efficient reallocation can take a lot of years if
we are lucky enough to achieve it.
-Solution: speed up the adjustment (flexible labour markets needed) and help the losers.
Global Value Chains and Other Policy Instruments
Preferential Trade Agreements (PTA) are any multilateral trade agreement.
-Can entail an increase of trade between member countries (trade creation) but also...
-Induce to countries to stop importing a good from countries outside the PTA (trade diversion).
-This can be negative if leads to import from a less productive country than before.
Rules of Origin (RoO) attribute a country of origin to a product in order to determine its “economic
nationality” in order to reduce the perverse incentives.
-As in some PTA you need to comply with RoO to enjoy the benefits of the PTA, if you are a final
good producer located in the PTA, you have two options:
• Comply with RoO: export to PTA members at preferential tariff rates. In that case, RoO
distort sourcing and lead to trade diversion in the intermediate goods.
• Not comply with RoO: source inputs from the most efficient producers around the world,
but faces high tariffs when exporting to PTA members.
Political Economy of Trade Policy
Due to the fact that gains are diffused but losses are concentrated, lobby, and populism; many times
politicians do not apply free trade policies and choose a “bad” trade policy.
Topics in Trade Policy
We have seen that, in general, trade leads to gains: although exceptions and losers, is beneficial.
In many models, the optimal trade policy is almost always free trade.
But this does not happen in reality: what are the other elements that are not in the previous theory
but could matter in reality? Therefore, what are the limits of our analysis?
We will consider three examples and illustrate them with some empirical analysis.
Dynamic Gains and Losses
In all the models we saw, the effects of a trade liberalization are static, which means that happen
immediately. In reality, markets have frictions and most of workers/firm’s decisions are dynamic:
• The investment decisions that firms have to take time to be build.
• It is costly to hire and fire workers.
• There are frictions in labour market: unemployed individuals cannot find jobs immediately.
• Human capital (e.g. training and experience) is not fully transferable across industries.
• Location decisions (e.g. industries gaining in other regions than industries losing).
Therefore, trade liberalization can be costlier than what is shown in the models.
To observe at the empirical evidence, we will observer at the China shock.
Due to market reforms, urbanization, and a strong comparative advantage in labour-intensive

manufactures, there was a trade expansion of manufacturing goods along the 90s and the 00s.
Even though it has represented a large positive net global supply shock for manufacturing, some
regions or countries which were specialized in manufacturing goods lost.
Employment in the US decreased in industries that competes with Chinese imports.

Standard trade theory (i.e. static models) predicts that labour should reallocate from these industries
to other industries, but, what really happened to these workers?
To see it, we could look at the regional approach or at the worker level approach (not shown here).
At the regional approach, we will look at a shift-share variable that

captures how much the region i is affected by Chinese imports.
IP: Import Penetration where:

Intuition: more exposed regions (i.e. high ΔIPWi) have relatively
higher employment share (i.e. high Lij/Lj) in industries that
suffered from Chinese imports (i.e. high ΔMj).
In other words: the relatively higher the employment share of

region i in an industry j (i.e. high Lij/Lj)...
the more impact on the region i (i.e. high ΔIPWi) will have the
increase of Chinese imports in industry j (i.e. high ΔMj).
Then, we can use ΔIPWi as an explanatory variable to see the effects.
An increase of Chinese imports leads to:

• Increase of total unemployment (↑).
• Increase of population out of the labour force (↑).
• Decrease of wages (↓).
• To offset, governmental transfers increases (↑).
• No effect on migration (=).
The cost of adjustment to trade are relevant: in the most exposed U.S. regions to import competition
from China there has been lower wages, higher unemployment, less stable marriages and political
polarization, among others.
Sometimes, the effects of a trade reform is felt 15 or 20 years later!

And the effects can be felt beyond the affected industries: the entire region suffers.
The institutional context matters for the adjustment speed: rigid labour markets tend to do worse.
Let us recall that there are still gains: consumer benefits from cheap Chinese goods.
So, there are still a lot of discussion on what are the optimal policies to remedy the adjustment.
Theory says that we should speed up the adjustment and help the losers.
There are good examples of policies: Trade Adjustment Assistance (US) or the European
Globalisation Adjustment Fund (EU).
It involves retraining/job search assistance for workers and credit/recovery plans for firms.
Global Value Chains and Other Policy Instruments
An important trend in international trade is the emergence of global value chains: different stages of
a production process (e.g. R&D, design, production of parts, marketing, etc.) are increasingly
fragmented across firms and countries. In fact, intermediates inputs account for 2/3 of total trade.
Therefore, tariff on an intermediate good can be very costly: if it increases the price of an important
input, it increases the price of all the goods that use that input.
But tariffs is just a little part of trade policy, which is much more complex. For example, many
countries hold Preferential Trade Agreements (any multilateral trade agreement). Types of PTA:
• Free Trade Agreement: low tariffs between members, but each country sets its own tariffs
with other countries (e.g. NAFTA).
• Custom Unions: members set common tariffs with other countries (e.g. MERCOSUR).
• Common Area: custom unions + free movement of factors (e.g. EU).
PTA entail Trade Creation: an increase of trade between member countries.
But also brings Trade Diversion: it can induce to countries to stop importing a good from countries
outside the PTA, and it can be negative if leads to import from a less productive country than before.
PTA often comes with other trade instruments and interacts with more complex trade policies in
non-trivial ways: recently, the Regional Comprehensive Economic Partnership (RCEP) not only
dealt with tariffs but also with unifying rules, introducing of e-commerce, Rules of Origin…
With so many intermediate inputs, how we can define what is produced inside a PTA member and
what is not? To solve that, there are the Rules of Origin (RoO), which are the rules to attribute a
country of origin to a product in order to determine its “economic nationality”.
There can be two types of Rules of Origin:

• Value-added requirements: at least X% of the value of the final good must be “domestic”.
This can be very arbitrary and complicated to determine.
• Change of tariff classification: some inputs cannot be sourced (at all) from outside the PTA.
It is simpler but also stricter.
As a result, if you are a final good producer located in the PTA, you have two options:
• Comply with RoO: export to PTA members at preferential tariff rates.
• Not comply with RoO: source inputs from the most efficient producers around the world,
but faces high tariffs when exporting to PTA members.
Theoretically, RoO distort sourcing and lead to trade diversion in intermediate goods.
In fact, in a large survey by the International Trade Center, RoO emerge as the most problematic
non-tariff measure faced by manufacturing firms. Moreover, there is very little evidence since legal
complexity of RoO makes extremely challenging to study.
To identify the effects of RoO, we will look at a study made about NAFTA (AER, 2018), which is
specially useful because the RoO are written very disaggregated (e.g. men’s trousers have different
RoO than women’s) and they are well defined in terms of tariff classification (e.g. to produce men’s
trousers some restricted fabrics must be sourced from within PTA).
To see the effects of RoO of NAFTA on Mexican Imports, we will observe:

• A time difference by comparing Mexican imports between 1991 and 2003.
• A goods difference by comparing treated intermediate goods (subjected to RoO) to non-
treated goods.
• A countries difference by comparing non-treated countries (outside NAFTA) to treated
(NAFTA partners).
The goal was to check if an intermediate good subject to RoO (e.g. fabric used to produce men’s
trousers) may suffer trade diversion, meaning a change in Mexican imports (from outside NAFTA
countries to NAFTA partners).
The results were that the effects from NAFTA’s RoO were the following:
• NAFTA RoO decreased the growth rate of Mexican imports from third countries relative to
NAFTA partners by around 45%.
• It was larger when RoO are stricter and when Mexican producers had stronger incentives to
comply, i.e. greater importance of North America export market.
Summing up:
• Trade policy is very complex and tariffs are generally just a small part of the picture.
• Trade policy between two countries can affect trade between other countries (trade diversion).
Political Economy of Trade Policy
Why politicians do not apply free trade policies and choose a “bad” trade policy? Some reasons:
-Gains are diffused but losses are concentrated: for instance, in Trump’s steel and aluminium tariffs
there are:
• Winners: steel/aluminium workers and shareholders.
• Losers: everybody who consumes a good made of aluminium.
If we compare both effects, the increase of the price of (for instance) an aluminium can is negligible
relative to a job loss or fall in share prices.
-Lobby: while sectors subject to import competition will lobby for protection (i.e. the steel company
owners, the steel union), the consumers will likely not waste their time fighting against it (at most,
can producers).
-Populism: if “median voter” benefits from these policies, politicians will propose it to get elected.
For example, following the HO model with high and low skill workers: opening to trade with a
foreign country with relative abundance with low-skill workers will lead to a decrease in the wages
of low skill workers (Stolper-Samuelson). Therefore, low skill workers will be against free trade
and politicians would support them to obtain their votes.
There is some empirical evidence about that:
Feigenbaum and Hall (2015) studied legislators votes after the increase of Chinese imports in the
US (between 1990-2015) by using the regional approach of the China Shock (seen above) and
looking at the increase in import penetration at congressional districts.
I will omit the empirical strategy for simplicity, as it is not very relevant.
The results were the following:
The districts affected by Chinese import competition (i.e. high ΔIPWi) voted more protectionist
positions on trade bills, but had no effect in the voting of other bills.
They also saw that there was stronger effects (politicians “were more protectionists”) in more
electoral competitive districts (e.g. 51-49%), and the effect persist no matter the incumbent was
Democratic or Republican.

International Economics I

Uploaded by

Copyright:

Available Formats

International Economics I

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

International Economics I

Uploaded by

Copyright:

Available Formats

INTERNATIONAL

Introduction and Main Facts about Trade

The Neoclassical Framework: The Specific Factor Model

The Heckscher-Ohlin Model

The Ricardian Model

Increasing Returns to Scale (The Krugman Model)

Firm Heterogeneity (The Melitz Model)

Topic in Trade Policy

International Economics has basically two branches:

Once the introduction is done, some facts will be exposed to start.

-Size and distance matters for international trade.

-Developing export manufacturing, developed export “high-tech”.

Production Possibility Frontier (PPF):

-Wages are equal because labour is mobile between sectors, as:

• Firms will reallocate L from production of A to M. Relative Supply of M increases.

There are gains from trade when there is a comparative advantage.

Gains from Trade and Redistributive Effects

Production Possibility Frontier (PPF): how much of each good an economy

Production Functions: QM=FM (K,LM), QA=FA (T,LA)

Labour Endowment: L=LA+LM

Supply: is obtained by finding the optimal production (maximising firm profits):

Equilibrium production: when Opportunity Cost of M = Relative Price of M

Relative Supply Curve:

Equilibrium Consumption: when MRS = Relative Price of M

Relative Demand Curve:

Comparative Advantage: when international relative price of M is higher than in autarky.

Gains From Trade

There are two assumptions:

This graphic puts together labour endowment, the productions

From the F.O.C., it is obtained that MC (wi) = MR (pi*MPLi) in each sector:

From this, we obtain the following:

Thus, equilibrium production of A and M is determined by the relative prices

If the price of M would increase (p’M > pM):

The Relative Supply Curve shows the positive relationship

Then, the problem can be solved using Lagrange:

Taking the first conditions, we obtain:

Finally, equalizing the FOCs, we obtain:

So, Marginal Rate of Substitution (amount of M that consumer is

To end up, let’s define equilibrium in a closed economy (autarky).

We can compute GDP in this economy using various equations:

Therefore, to achieve equilibrium in a closed economy two conditions must be hold:

If an open economy is small, its economy take international prices as given.

The equation can be rearranged to obtain the following (remember: D-Q=consumption-production):

The international price determines if a country exports or imports a

The increase on the relative price leads to an increase in the relative

In other words, an extra production is being generated, leading to

Gains from Trade and Redistributive Effects

Now, let’s imagine three different agents in the economy:

Therefore, in the Specific Factors Model:

-Meanwhile, the effect on the mobile factor is ambiguous.

The Heckscher-Ohlin Model

Equilibrium Relative Prices:

Factor Price Equalisation:

-When the comparative advantage is reinforced (weakened), trade increases (decreases).

The evidence in favour of the HO is mixed.

The Heckscher-Ohlin Model

-Relative Demand: -Resource constraint of the economy:

-MRST is equivalent to the relative input prices:

-Factor requirements of production function:

-If T is more relatively intensive in L: - -If (↑w/r), (↓L/K).

and PT/PA < PTI/PAI < PT/PA

We already know that RD (=DT/DA=(bPA/(1-b)PT)), now we need to obtain RS (=QT/QA).

-Expenditure in good j: (wL + wL) × 1/N

-Total expenditure in home-produced goods: (wL + wL) × z/N

-Effect of technological progress (aLi)’ < aLi → A(i) ↓ for all I:

Now let us consider foreign real wage. [Recall: (w/w)’<(w/w)].