ECO364 - International Trade - Chapter 4 - Trade and...

ECO364 - International TradeChapter 4 - Trade and Wages

Christian Dippel

University of Toronto

Summer 2009

Christian Dippel (University of Toronto) ECO364 - International Trade Summer 2009 1 / 33

Trade and Wages

I Across many developed countries, real average hourly earnings havestagnated or fallen for less skilled workers while wages for skilledworkers continue to rise since the 1970s.

I Consequently, the “skill premium” has increased leading to increasedwage inequality.

I At the same time, trade volume (as a proportion of GDP) hasincreased rapidly since the early 1980s.

I Trade Agreements.I “Death of distance”

I Is there a connection? If so, how strong is it?


Trade and Wages

I Source: DeNavas-Walt, Carmen; B. D. Proctor, R. J. Mills (08 2004)


Trade and Wages

I Inequality has grown by 40% based on the measures above.

I Inequality has increased by approximately 25% based on morecommon survey measures.

I This summarizes the discussion in Freeman (1995), Bivens (2007),and Krugman (2008).

I Read Freeman (1995) and skim Bivens (2007) and Krugman (2008).

I From the perspective of developed countries opening up to trade withdeveloping countries, the Heckscher-Ohlin model makes a sharp(testable) prediction.

I Instead of capital and labor, suppose that the two factors ofproduction are skilled and unskilled labor.


Trade and Wages

I For simplicity....

I Two countries: China and Canada.I Two factors: skilled and unskilled labor.

I China possesses a relative abundance of unskilled labor.

I Two goods: Skilled and unskilled labor intensive goods.

I This augments the previous HO diagram to:


Trade and Wages


Trade and Wages

I Canada’s abundant factor (skilled labor) will win and its scarce factor(unskilled labor) will lose.

I In the real world, complete factor price equalization is unlikely tooccur

I Transportation costsI Productivity differencesI Among others....

I However, the direction of the change is likely to hold.


Trade and Wages

Effect of Trade on Wages

I HO Mechanism:I Canada has a Comparative Advantage in skill-intensive sectors because

it is skill-abundant.I The price of skill-intensive goods is increased through trade and skilled

wage increase.I Both sectors use less skilled labor because of the wage change.I But the skill-intensive sector expands and the unskilled-labor-intensive

sector contracts.

I Key comparative static: If wage-changes are due to trade, thenS/U and wS/wU are negatively correlated at the industry level.

I Firms substitute away from more expensive factor.


Trade and Wages

I Recall our basic Lerner diagram with computers being skilled-laborintensive and textiles unskilled-labor intensive.


Trade and Wages

I As wswu↑ due to trade, firms will substitute toward more unskilled labor

intensive techniques such that as S/U ↓.


Trade and Wages SBTC

I Now let’s instead consider Skill-Biased Technological Change(SBTC) as a source of inequality.

I Let’s first make some important distinctions:I Technological Change is Factor-Biased if it makes one factor more

productiveI Technological Change is Sector-Biased if it makes one sector more

productiveI Technological Change is Factor-Augmenting if it raises the return to

one factor

I If technological change explains the increasing wage-inequality, itneeds to be skill-augmenting.

I When is technological change skill-augmenting?



Consider the following Cobb-Douglas production functions for textiles andcomputers:

QT = ATSαTU1−αT QC = ACSβCU1−β

C

Firms producing X maximize PXQX − wSSX − wUUX

First Order Conditions (FOC) give: :

(1) wS = pTATαSα−1T U1−α

T (3) wS = pCACβSβ−1C U1−β

C

(2) wU = pTAT (1− α)SαTU−αT (4) wU = pCAC (1− β)SβCU−βC

Dividing (1)/(2) and (3)/(4) gives us:

wSwU

= α1−α

UTST

wSwU

= β1−β

UCSC



Setting (1)=(3) and plugging in the two expressions for wSwU

gives us:

(wUwS

)β−α = pTpC

ATAC

(1−α)(1−α)αα

(1−β)(1−β)ββ

I Let’s consider skill-biased and sector-biased technological change:I skill-biased TC raises α and β while sector-biased TC changes AT

AC.

I You can check that the effect of changes in α and β on (wU

wS) is

non-monotonic: it does not always raise the return to skill!I But the effect of an increase in AC

ATon ( wS

wU) is always positive.

I What’s the message? We need TC to be at least partly sector-biasedto be skill-augmenting.



I Purely Factor-Biased Technological Change (both isoquants tilttoward skill) can leave w/r unaffected



I Sector-Biased Technological Change always changes w/r



I There is a big debate on whether SBTC in the past decades has beenmore sector- or more factor-biased.

I Key for us: As long as it was partly sector-biased, ws/wu and S/Uwill go up together

I We have therefore arrived at two alternative explanations of the risews/wu but the two explantions give us different predictions on thecorresponding changes in S/U at the industry level



I Trade theory predicts, counter-factually, that S/U must be decliningat the industry level when ws/wu goes up.

I This gives us an ability to identify one story from anotherI Maybe now the “Two Cheers for Formalism” in the introductory slide

are more compelling?

I Data show that S/U has been rising dramatically within industriessince the 1970s.

I Skill-Upgrading has happened across the board in all sectors.

I Whatever the exact form of SBTC, the pattern of wages looks a lotmore like the SBTC story than the trade story.


Trade and Wages Factor Content

I Now we know that SBTC was a much stronger force in changingwage-inequality relative to trade .

I But we also want to know how strong of a force trade was in absoluteterms.

I Two types of empirical studies have been used to assess themagnitude of the effect of trade on wages: Factor Content andPrice Change studies.

I Factor Content studies look at trade and calculate how much labor isbeing exported/imported and how much this shifts the relative supplycurve of skilled to unskilled labor for a country.

I Price Change studies exploit the fact that both factor prices andgoods prices should be changing (Stolper-Samuelson Effects).

I Intuition for Factor-Content Studies: Everyone consumes on thediagonal of the Edgeworth-box. The larger the factor-content oftrade, the more different a country is from the world averageendowment mix and the larger are the changes in relative wagesimplied by trade.



I Ask how much trade in factors is implied by trade in goods.

I Start with the volume of exports and imports.

I Use unit labor requirements to calculate how much labor (net)imports embody.

I Designed to see how much the relative supply of skilled workers hasshifted.

I Assumes an aggregate relative factor demand curve for the economyand asks how much a change in the RS curve for factors affectsrelative wages.



I Assume that the North specializes in computers and the Southspecializes in textiles. Let XN

C and X ST be Northern exports of

computers to the South and Southern exports of textiles to theNorth, respectively.

I aNS,C is the unit skilled labor requirement computers in the North and

aNU,C is the unit unskilled labor requirement computers in the North.

I aSS,T is the unit skilled labor requirement textiles in the South and

aSU,T is the unit unskilled labor requirement computers in the South.

I Define SNtrade and UN

trade as the skilled and unskilled labor that isembodied in the exports of the North. Similar terms apply to theSouth.

I Define SN and UN as the absolute skilled and unskilled laborendowments in the North.



SNtrade = aN

S,CXNC + aN

S ,TXNT SS

trade = aSS ,CX S

C + aSS,TX S

T

UNtrade = aN

U,CXNC + aN

U,TXNT US

trade = aSU,CX S

C + aSU,TX S

T

The new amounts of skilled and unskilled labor in the North will be

SN = SN − SNtrade + SS

trade UN = UN − UNtrade + US

trade

SS = SS − SStrade + SN

trade US = US − UStrade + UN

trade

SN , UN denote the factor content of the North’s Consumption bundle.



Factor Content Studies: Example

I Suppose that the SN = 200 and UN = 100. For the South, SS = 100and US = 200.

I Suppose that the North exports 10 computers and the South exports10 units of textiles. Assume that all other trade values are zero.

I Assume that unit labor requirements are as follows:

aNS,C = 1/5 aS

S ,T = 1/10

aNU,C = 1/10 aS

U,T = 1/5

I Verify that trade leads to the North “importing” 1 unit of unskilledlabor and “exporting” one unit of skilled labor.

I This leads to: SN = 199, SS = 101, UN = 101, SU = 199.



I A country that imports unskilled-labor-intensive goods is importingunskilled labor through the factor content

I The larger the factor content of trade, the stronger the forces offactor price equalization

I The factor-content of trade turns out to be much smaller thanpredicted by theory.



Price Change Studies

I We should also see convergence in goods prices.

I Use goods prices and assumptions about the shape of the SS curve tosee how much trade might have changed wages.

I We should also see the price of unskilled-labor-intensive goods fallingin developed (skill abundant) countries.



Price Change Studies

I Let’s go back and look at the HO graph:



I Both types of studies find that increased trade has led to increasedinequality in the OECD.

I However, both find that the magnitudes only explain approximately4%-6% of the 24% change in inequality since the mid 1970s.

I Recall that imports from China and India have surged over the past10 years and the study by Freeman (1995) is 13 years old. Do thesecalculations still hold up?

I Bivens (2007) shows that trade has contributed little to inequalityeven up to today.

I Krugman (2008) shows that this is somewhat of a paradox. Importsare coming from more unskilled labor abundant countries than tenyears ago. Why is there no noticeable change?



Conclusion

I Wage inequality has been increasing in the OECD in the 80s and 90s.

I At the same time, the trade volume has expanded enormously.I HO Theory tells us that trade with developing countries has

contributed to the deterioration of the relative wage of unskilledworkers. However:

1. Trends at the industry level are inconsistent with trade models.2. Factor Content and Stolper Samuelson studies suggest that trade only

explains approximately 20− 25% of the total deterioration (6% ofabout 25%) based on data from the Current Population Survey.

I But SBTC is (by far) the most important determinant of theincreased inequality.



What about developing countries?



I Factor-Proportions Theory predicts that the skill premium indeveloping countries should fall.

I The skill premium has increased leading to increased wage inequalityin many developing countries

I HO gets it exactly wrong in these developing countries.I Data for East Asia is still coming out.

I Runjuan Liu (2008) finds that Chinese provinces more subject to tradehave seen inequality grow more.

I In Latin America, we see the opposite of the HO predictions


Trade and Wages Recent Developments in the Debate

I The Consensus in the late 90’s was that SBTC was the mostimportant driving force in changing inequality across the globe.

I Two recent theoretical and empirical research agendas haverevitalized this debate:

1. Directed Technological Change (DTC)2. Trade in intermediate products



Directed Technological Change

I Increasing evidence that TC is the result of firms’ R&D effortsI TC is “directed”!

I It can be shown that the TC will in equilibrium always be directedtowards the sectors that use the abundant factor intensively.

I This means that firms in the North innovate to increase ACAT

and firms

in the South innovate to reduce ACAT

.

I This has interesting implications:



DTC in the North

I It can be shown that firms’ incentives to innovate “toward” theabundant factor are stronger with trade (as the return to theabundant factor increase)

I This implies in the North that:

1. DTC is consistent with the SBTC explanation of rising wage-inequality2. but that trade might cause part of this SBTC so that trade might be

to blame after all



DTC in the South

I In the South, DTC poses a puzzle:

I If HO gets it exactly wrong and SBTC is the only alternativeexplanations, then DTC gets it exactly wrong because firms shouldinnovate “towards” unskilled labor.

I An important piece of the puzzle probably lies in the fact thatdeveloping countries are not on the technology innovation frontier,instead they are on a technology adoption frontier that lies inside theinnovation frontier.

I However, Southern firms would still have an incentive to adopt themost unskilled-labor-biased innovations.

I The real explanation probably lies with the importance of trade inintermediate products and outsourcing/offshoring.

I we will explore this issue at the end of this course.


ECO364 - International Trade - Chapter 4 - Trade and...

Documents

Transcript of ECO364 - International Trade - Chapter 4 - Trade and...