planetkdoublehs: March 2008

Chapter 5 The Globalization of Labor

Over the past two decades, labor markets around the world have become increasingly integrated. Political changes and economic reforms have transformed China, India, and the formerEastern bloc countries, effectively involving theirlarge labor forces in open market economies.At the same time, the development of technology,combined with the progressive removal ofrestrictions on cross-border trade and capitalflows, has made it possible for production processesto be unbundled and located farther fromtarget markets for a growing universe of goodsand services. The location of production hasbecome much more responsive to relative laborcosts across countries. There have also beenincreasing flows of migrants across borders,through both legal and informal routes.This ongoing globalization of the labormarket has drawn increasing attention frompolicymakers and the media, particularly in theadvanced economies. The most asked questionis whether the addition of this unprecedentedlylarge pool of labor from emerging marketand developing countries is adversely affectingcompensation and employment in the advancedeconomies.This chapter addresses this important andemotive question. In contrast with most previousstudies, which focus on one country ora single channel of transmission, it takes abroad approach, considering a large sample ofadvanced economies and a full range of transmissionchannels (competing imports of finalproducts, offshoring of intermediate products,and immigration). The chapter focuses on thefollowing issues:• How rapidly has the global labor supplygrown, and which channels of labor globalizationhave been most important?• To what extent can recent trends in laborshares and labor compensation in advancedeconomies be explained by the changingglobal labor supply relative to other factorssuch as technological change and labor marketreform? Has the impact been different inskilled and unskilled sectors?• What policies can help the advanced economiesmeet the challenges of further labormarket globalization?This chapter finds that the effective globallabor force has risen fourfold over the pasttwo decades. This growing pool of global laboris being accessed by advanced economiesthrough various channels, including importsof final goods, offshoring of the production ofintermediates, and immigration. The ongoingglobalization of labor has contributed to risinglabor compensation in advanced economies byboosting productivity and output, while emergingmarket countries have also benefited fromrising wages. Nevertheless, globalization is oneof several factors that have acted to reduce theshare of income accruing to labor in advancedeconomies, although rapid technological changehas had a bigger impact, especially on workersin unskilled sectors. The analysis finds thatcountries that have enacted reforms to lowerthe cost of labor to business and improve labormarket flexibility have generally experienceda smaller decline in the labor income share.Looking ahead, it is important for countries tomaximize the benefits from labor globalizationand technological change, while also working toaddress the distributional impact. To this end,policies should seek to improve the functioningof labor markets; strengthen access to educationand training; and ensure adequate social safetynets that cushion the impact on those adverselyNote: The main authors of this chapter are FlorenceJaumotte and Irina Tytell, with support from Christian deGuzman and Stephanie Denis. Robert Feenstra providedconsultancy support.CHA PTER 5 The Globalization of Labor162affected, without obstructing the process ofadjustment.How Globalized Is Labor?A first question to address is how the openingup of China, India, and the former Eastern bloccountries, together with ongoing demographicdevelopments, has affected the global laborsupply. This is not easy to answer because muchdepends on the assumptions made about howmuch of a country’s labor force is in, or couldpotentially compete in, the global market. Onesimple approach is to weigh each country’slabor force by its export-to-GDP ratio. Bythis measure, the effective global labor supplyquadrupled between 1980 and 2005, with mostof the increase taking place after 1990 (Figure5.1). East Asia contributed about half ofthe increase, due to a marked rise in workingagepopulation and rising trade openness, whileSouth Asia and the former Eastern bloc countriesaccounted for smaller increases. Whilemost of the absolute increase in the globallabor supply consisted of less-educated workers(defined as those without higher education), therelative supply of workers with higher educationincreased by about 50 percent over the last 25years, owing mostly to advanced economies, butalso to China.Advanced economies can access this increasedpool of global labor both through imports ofgoods and services and through immigration.Trade has been the more important channeland has grown more rapidly, not least becauseThis approach, which follows Harrigan and Balaban(1999), is more accurate for developing countries specializedin labor-intensive activities than for advanced economieswhose exports are relatively capital intensive. Inorder to capture the export of labor through emigration,emigration weights could be added to the trade weights.However, these weights are generally very small.This compares to estimates in Freeman (2006) thatthe integration of China, India, and the former Easternbloc countries doubled the number of workers in theglobal economy. The difference is due to the weighingof national labor forces by export-to-GDP ratios in thischapter’s estimates.Figure 5.1. Alternative Measures of Global Labor SupplyEast Asia's marked rise in working-age population and increasing trade opennesshave contributed to about half of the quadrupling of the effective global laborsupply, while South Asia and the former Eastern bloc accounted for smallerincreases.1980 85 90 95 2000 05050100150200250300350Global Labor Supply 400(index, 1980 = 100)Working-agepopulationLabor forceExport-weightedlabor force1Sources: United Nations, Population Prospects: The 2004 Revision Population database;World Bank, World Development Indicators; and IMF staff calculations.National labor forces scaled by export-to-GDP ratios.Includes Western Hemisphere, Middle East and North Africa, and sub-Saharan Africa.More educated labor force is defined by persons with university-level education. Lesseducated is defined by labor force with primary and secondary education plus theuneducated.1231980 85 90 95 2000 050200400600800Export-Weighted Labor Force by Region 1000(index, 1980 = 100)11980 85 90 95 2000 0502505007501000125015001980 85 90 95 2000 050102030405060Working-Age Population 70by Region (millions ofpeople)Non-Oil Export-to-GDP Ratio byRegion (percent)1980 85 90 95 2000 050100200300400500600700Export-Weighted Global Labor Force by Education Level 800(millions of people)1, 3More educatedLess educatedAdvanced economiesEast AsiaSouth AsiaCentral and Eastern Europe andCommonwealth of IndependentStatesOther developing countries2immigration remains highly restricted in mostcountries (Figure 5.2). A similar picture emergesfor developing and emerging market countries,where the export-to-GDP ratio is in generalmuch higher than the ratio of emigrants to thedomestic labor force. Nevertheless, immigrationhas expanded significantly over the pasttwo decades in some large European economies(Germany, Italy, and the United Kingdom) andin the United States. The share of immigrants inthe U.S. labor force is now close to 15 percentand hence comparable to the share of importsin GDP. Elsewhere the share of immigrants isstill substantially less than the share of importsin GDP, but it is not negligible.Focusing on trade, the share of developingcountry products in the manufacturing importsof advanced economies has doubled since theearly 1990s (Figure 5.3). This owes much toChina. Developing countries have also beencapturing an increasing share of world markets.At the aggregate level, however, trade is awin-win game. As China, India, and the Easternbloc countries have opened up, world marketsand opportunities to export have expandedconsiderably for advanced economies anddevelopingcountries alike. Developing countries’imports have been growing faster thanthose of advanced economies and the share ofadvanced economies’ exports going to developingcountries has been rising (though not asrapidly as the share of developing countries intheir own imports). Further, while both importand export prices have been on a decliningtrend relative to output prices, the terms oftrade of advanced economies have improved bya cumulative 7 percent since 1980. Most notably,there was a substantial improvement in theterms of trade of Japan in the first half of the1980s. However, the large fall in import pricesat this time was mainly the result of the strongappreciation of the yen at a time when oil priceswere falling, and was not directly related toglobalization.The stock of emigrants is limited to those emigratingto OECD economies.Figure 5.2. Immigration and Trade(Percent of labor force and GDP, respectively)0510152025303540451990 2005Advanced EconomiesImports of goodsand servicesImmigration2Developing Countries051015202530354045Exports of goodsand servicesEmigration211980 85 90 95 200005101520Stock of Foreign Labor Force in Advanced Economies 25(percent of total labor force)31980 85 90 95 2000 05010203040Imports of Goods and Services in Advanced Economies 50(percent of GDP)Sources: Docquier and Marfouk (2005); OECD, International Migration Data; U.S.Census Bureau; and IMF staff calculations.2000 data for immigration and emigration.Hong Kong SAR, Greece, Israel, New Zealand, Singapore, and Taiwan Province ofChina are not included in average immigration due to data limitations. Russia andSlovak Republic are not included in average emigration due to data limitations.Foreign-born labor force for Australia, Canada, and the United States. For Italy,the Netherlands, Norway, and the United Kingdom, the data refer to the share of foreignemployment in total employment.12304Although immigration has expanded significantly over the past two decades insome large European countries and the United States, trade remains as the moreimportant channel for accessing the large global labor force.United StatesUnited KingdomJapanGermanyFranceItalyCanadaHow Globalized Is Labor?163CHA PTER 5 The Globalization of Labor164As can be seen in Figure 5.4, the strongexport dynamism of emerging market and developingcountries is in skilled as well as unskilledproducts: developing countries’ share in worldexports of skilled goods and services has beenon the rise in recent years. China has led theway, reflecting its very strong growth and a movetoward more skill-intensive goods in its exportbasket. India’s export basket is also changingrapidly toward skill-intensive services, but thecountry’s weight in world trade remains small.One category of trade that has received muchattention in recent years is trade in intermediates.The reduction of barriers to cross-bordertrade and capital flows, combined with technologicalprogress in transport and communication,has made it easier for firms to moveparts of their production to less costly foreignlocations—a process referred to as offshoreoutsourcing or, more simply, offshoring. Nevertheless,and contrary to some popular perceptions,offshored inputs, which account for abouthalf of total imports (the rest being imports offinal products), have grown somewhat moreslowly than total trade (see also OECD, 2006a).Moreover, the scale of offshoring is still quitelimited in the overall economy (Figure 5.5).Imports of intermediate manufacturing andservices inputs (excluding energy) accountedfor about 5 percent of gross output and about10 percent of total intermediate inputs inadvanced economies in 2003, the latest year forwhich data are available. These shares haveSkilled exports are measured as exports of goods andservices produced in skilled sectors, that is, sectors with ahigher share of skilled workers in their labor force. Theresults are generally robust to excluding medium-skill sectorsand focusing instead on low-skill and high-skill sectors(see Appendix 5.1 for details). Using a more refinedclassification of products by skill intensity, Rodrik (2006)concludes that China’s export basket is much more skillintensive than would be expected given China’s level ofdevelopment.It is common to scale imported intermediates bytotal intermediate inputs to estimate the intensity ofoffshoring. However, it seems more appropriate to scaleimported intermediates by total inputs (including laborand capital), since imported intermediates can substitutenot only for domestic intermediate inputs but also forin-house labor and capital.1980 85 90 95 2000 05608010012014016018020080859095100105110115Ratio of Import and Export Price Deflators to Total Absorption 120Deflator, Advanced OECD Economies (index, 2000 = 100)Export price deflator(left scale) Terms of trade(right scale)Import price deflator(left scale)Non-oil terms of trade(right scale)1990 92 94 96 98 2000 02 0402040608010002468Manufacturing Imports of Advanced OECD Economies by Source 10(percent of total)Other advanced economies (left scale)Developing countries (left scale)China (right scale)India (right scale)1990 92 94 96 98 2000 02 0402040608010002468Manufacturing Exports of Advanced OECD Economies by Destination 10(percent of total)India (right scale)China (right scale)Other advanced economies (left scale)Developing countries(left scale)As China, India, and the Eastern bloc have opened up, world markets andopportunities to export have expanded considerably for advanced economies anddeveloping countries alike.Figure 5.3. Share of Developing Countries in TradeSources: OECD, STAN Bilateral Trade Database; and IMF staff calculations.1985 90 95 2000 05010203040501985 90 95 2000 05010203040Imports of Developing 50Countries (percent ofadvanced economies'imports)Non-oil goodsServicesExports of DevelopingCountries (percent ofadvanced economies'exports) Non-oil goodsServicesincreased only moderately since the early 1980s.The share of offshored inputs in gross outputranges from 12 percent in the Netherlands toabout 2–3 percent in the United States andJapan. Offshoring is thus relatively limitedin the United States and Japan, in the sameway that trade openness is usually low in largeeconomies.The manufacturing sector has been mostaffected by offshoring because it is more tradable.For the countries for which long dataseries are available (G-7, Australia, and the Netherlands),the share of imported manufacturinginputs in gross manufacturing output increasedfrom 6 percent in 1980 to 10 percent in 2003,with the rise being somewhat stronger in the latteryears of the sample (Figure 5.6). In 2003, theoffshoring intensity in manufacturing rangedfrom 4 percent in Japan to a high of about25 percent in Canada. Imports of services inputsby the overall economy remain low at 1 percentof gross output, although the offshoring intensityin services has increased in recent years in anumber of countries, including Canada, Germany,and the Netherlands.Interestingly, the rise in offshoring inadvanced economies has been driven mostly byimports of skilled rather than unskilled inputs.Several factors may help explain this finding.First, in line with advanced economies’ comparativeadvantage in skill-intensive production,goods traditionally produced in unskilled sectors(e.g., textiles) are more likely to be imported asfinal goods rather than intermediates. Sectorsinvolved in the rise in the imports of intermediariesare electronic equipment; other machineryand equipment; and chemical, rubber, andplastic products. It should be noted, however,The flattening in 2001–02 is temporary and reflectsthe slowdown in world trade associated with the globalrecession.See Jensen and Kletzer (2005) and Amiti and Wei(2005) for more details on offshoring of services. Thelatter also find that offshoring of services remains verylimited, although it has grown in recent years.The share of imported intermediates in total imports ofunskilled products is lower than the comparable share forskilled products, at 37 percent and 68 percent, respectively.1992 94 96 98 2000 02 040102030405060701995 96 97 98 99 2000 01 02 0305101520253035Figure 5.4. Developing Countries: Exports of SkilledManufacturing Goods and Services(Percent)As a Share of Advanced Economies' Skilled ExportsChinaIndiaSouth AfricaPolandBrazilEmerging EuropeSources: OECD, International Trade in Services Database and STAN IndustrialDatabase; World Bank, World Integrated Trade Solution database; and IMF, Balance ofPayments Statistics.Bangladesh, China, India, Malaysia, the Philippines, and Thailand.Czech Republic, Hungary, Poland, Romania, Slovak Republic, and Turkey.Argentina, Brazil, Chile, Colombia, Mexico, Peru, and Venezuela.Egypt, Ethiopia, Morocco, South Africa, Sudan, and Tanzania.12341992 94 96 98 2000 02 04020406080100ChinaIndiaSouth AfricaPolandBrazilAs a Share of Country's Total Exports in ManufacturingAs a Share of Country's Total Exports in ServicesThe recent robust growth in developing countries' share in world exports, particularlyin skilled products, owes much to China.Developing countriesDeveloping AsiaLatin America1 Middle East and Africa234How Globalized Is Labor?165CHA PTER 5 The Globalization of Labor166that offshoring is likely to involve the least skill-intensivestages of production in these skilledsectors, although the available data do notallow confirmation of this. Second, the bulk ofadvanced economies’ imports (of both final andintermediate products) still comes from otheradvanced economies and likely includes moreskilled rather than unskilled products. Third,as mentioned earlier, the global supply of laborwith higher education has increased relative tolabor with lower education.How Has the Globalization of LaborAffected Workers in Advanced Economies?The rapid growth of the global labor supplyand its manifestation through increasing exportsof emerging market and developing countriesleads to the question of how these trends haveaffected workers in advanced economies. Withexports from emerging market and developingcountries being intensive in labor, especiallyunskilled labor, traditional trade theory wouldpredict that the integration of these countriesinto the world economy would exert downwardpressure on the wages (corrected for productivity)of workers in advanced economies. Hence,the share of national income received by labor—the so-called labor share—would be expectedto decline. To see this, it is worth noting thatthe labor share can be expressed as the ratioof labor compensation per worker to averageworker productivity.Nevertheless, workers in advanced economiescould still be better off if the positive effectsfrom enhanced trade and productivity on theeconomy’s income (the size of the total “pie”)are larger than the negative effect on the shareof this income that accrues to labor. The vastliterature documenting gains from trade (see,for example, Lewer and Van den Berg, 2003;Berg and Krueger, 2003) suggests that theincrease in the economy’s income may indeedbe substantial. Recently, Grossman and Rossi-Hansberg (2006) have argued that the productivity-enhancingeffect from trade in intermediatescould be even larger than from trade in final1980 85 90 95 2000036912Total Offshoring in Advanced OECD Economies 15(weighted; percent)Figure 5.5. Offshoring by Advanced EconomiesThe extent of offshoring is still quite limited in advanced economies. In 2003, theoffshoring of nonenergy manufacturing and services inputs averaged about 5percent of gross output and roughly 10 percent of total intermediate inputs. The lowlevel of offshoring is particularly pronounced in the world's largest economies, theUnited States and Japan.1970 75 80 85 90 95 2000036912Total Offshoring in G-7 Economies 15(percent of gross output)Italy CanadaUnited StatesUnited KingdomFranceGermanyJapanSources: OECD, Input-Output Tables (1995, 2002, and 2006 editions), International Tradein Services Database, and STAN Industrial Database; Groningen Growth and DevelopmentCentre, 60-Industry Database (September 2006); and IMF staff calculations.Offshoring measures calculated using Input-Output Tables from OECD; resulting seriesextended from 2001 to 2003 by estimating extent of offshoring using a combination of datafrom the OECD STAN Industrial Database and the Groningen 60-Industry Database; onlyoffshoring of nonenergy manufacturing and services inputs considered.Advanced OECD economies used in calculations for long time series include Australia,Canada, France, Germany, Japan, the Netherlands, the United Kingdom, and the UnitedStates; weighted using series on GDP at current U.S. dollars from the World EconomicOutlook database.Advanced OECD economies used in calculations for short time series include Australia,Austria, Canada, Finland, France, Germany, Greece, Italy, the Netherlands, Japan, Korea,Portugal, Spain, Sweden, the United Kingdom, and the United States; weighted using serieson GDP at current U.S. dollars from the World Economic Outlook database.1231985 90 95 2000304050607080Total Offshoring in G-7 Economies(percent of total imports of manufacturing goods and services)United StatesUnited KingdomFranceGermanyJapanShare of gross output (long time series)Share of gross output (short time series)Share of total intermediate inputs (long time series)Share of total intermediate inputs (short time series)23231goods because, in addition to a competitioneffect for producing sectors, trade in intermediatesalso reduces the costs of production of usingsectors. The empirical evidence on the productivityeffects of offshoring is, however, mixed.What do the data show? Looking first at thelabor share, there has been a clear decline sincethe early 1980s across the advanced economies(Figure 5.7).10 The decline is stronger for thelabor share than for the share of employees’compensation, reflecting a reduction in theshare of other categories of workers in thetotal workforce (other categories of workersinclude self-employed, employers, and familyworkers).11 A part of this decline is a reversalof the rise in labor shares that took placein the 1970s, especiallyin Europe and Japan(Blanchard, 1998).12There is little empirical evidence on the productivityeffects of offshoring to date (see Olsen, 2006). Thereare some indications that positive productivity effectsof manufacturing offshoring depend on the degree towhich firms are already globally engaged. However, theirglobal engagement may be already close to optimal levelsin advanced economies, suggesting that the potential forproductivity gains from services offshoring may be larger.Positive productivity effects of services offshoring to dateappear to be generally small in manufacturing plants, butsomewhat bigger in service-sector firms. Amiti and Wei(2006) find a significant positive effect of services offshoringand a somewhat smaller positive effect of manufacturingoffshoring on productivity in the United States.10National accounts provide the share of employees’compensation in total income but do not identify separatelythe labor income of other categories of workers(self-employed, employers, and family workers). Severalcorrection procedures are available (Gollin, 2002) and,for data availability reasons, the employees’ compensationwas augmented with compensation of other categoriesof workers by assuming that the latter command similarwages per worker as employees. The results are robust ifother procedures are used (see Appendix 5.1).11Focusing on the United States, for which data areavailable since 1930, the share of employees’ compensationin national income does not appear to be at a historicallow (though this may be partly related to the risein the share of employees in the total workforce).12Blanchard (1998) argues that the rise of the laborshare in Europe in the 1970s was driven by a negativeshift in labor supply as wages did not adjust fast enoughto the slowdown in underlying factor productivity growth.Over time, though, employment adjusted downward,exerting downward pressure on wages and returning thelabor share toward its previous level (though at a higherunemployment rate). The further decline that has taken1980 85 90 95 20000510152025Offshoring of Manufacturing Inputs by the Manufacturing Sector 30(weighted; percent of manufacturing gross output)Figure 5.6. Advanced Economies: Offshoringby Category of Inputs1The manufacturing sector has been more affected by offshoring because it is moretradable, although there are considerable differences across countries (the verticalline shows the range of country outcomes). Skilled inputs have also played a moresignificant role in the growth of offshoring in advanced economies than unskilledinputs.Weighted average Range of country outcomes1980 85 90 95 20000123Offshoring of Services Inputs 4(weighted; percent of gross output)21980 85 90 95 20000123Offshoring of Unskilled and Skilled Inputs 4(weighted; percent of gross output)UnskilledSkilledSources: OECD, Input-Output Tables (1995, 2002, and 2006), International Trade inServices Database, and STAN Industrial Database; Groningen Growth and DevelopmentCentre, 60-Industry Database (September 2006); and IMF staff calculations.Offshoring measures calculated using Input-Output Tables from OECD; resulting seriesextended from 2001 to 2003 by estimating extent of offshoring using a combination of datafrom the OECD's STAN Industrial Database and the Groningen 60-Industry Database; onlyoffshoring of nonenergy manufacturing and services inputs considered. Advanced OECDeconomies used in calculations for long time series include Australia, Canada, France,Germany, Japan, the Netherlands, the United Kingdom, and the United States; weightedusing series on GDP at current U.S. dollars from the World Economic Outlook database.Excludes the United States since import data are reported as inclusive of "cost,insurance, and freight"; thus, values that normally accrue to business services are includedin associated goods sectors.12How Has the Globalization of Labor Affected Workers in Advanced Economies?167CHA PTER 5 The Globalization of Labor168The decline in the labor share since 1980has been much more pronounced in Europeand Japan (about 10 percentage points) thanin Anglo-Saxon countries, including the UnitedStates (about 3–4 percentage points).13 WithinEurope, the strongest decline is observed in Austria,Ireland, and the Netherlands. Further, mostof the decline in the labor share can be attributedto the fall in unskilled sectors, which wasmore pronounced in Europe and Japan than inthe Anglo-Saxon countries. This decline reflectsa combination of the reduction in the withinsectorlabor share and the shift of output fromunskilled toward skilled sectors (see Figure 5.7).The income share of labor in skilled sectors, onthe other hand, has been on the rise, especiallyin Anglo-Saxon countries where it has increasedby about 5 percentage points. It is important toemphasize that due to the nature of the availabledata, these results relate to income sharesof workers in skilled and unskilled sectors,rather than to income shares of skilled andunskilled workers themselves.Despite the fall in the overall labor share, reallabor compensation has expanded robustly inall advanced economies since 1980, with growthaccelerating since the mid-1990s. This trendreflects both employment growth and increasesin real compensation per worker, with a strongerweight on employment in the Anglo-Saxoncountries and on real compensation per workerin Europe (Figure 5.8). Since the mid-1990s,however, employment growth has picked up inEurope, outpacing the growth in real compensationper worker. Growth in labor compensationof unskilled sectors, however, has been very sluggish(Figure 5.9). While unskilled employmenthas held steady in the United States, increasesplace in the labor share since the mid-1980s is the resultof an adverse labor demand shock: at a given wage andcapital stock, firms have steadily decreased employment.Such a shift may have various sources: the adoption oftechnologies biased against labor and toward capital or ashift in the distribution of rents away from workers.13For the purpose of this chapter, Europe includes theeuro area countries, Denmark, and Norway, while Anglo-Saxon countries include Australia, Canada, the UnitedKingdom, and the United States.Figure 5.7. Advanced Economies: Labor Income Shares(Percent of GDP unless otherwise noted)Over the past two decades, there has been a continued decline in the share ofincome that accrues to labor, especially in Europe and Japan. The income share ofworkers in unskilled sectors has dropped strongly while that of workers in skilledsectors has generally made small gains.United Sates EuropeOther Anglo-Saxon4 Japan51930 45 60 75 90 2005455055606570751980 85 90 95 2000 0545505560657075Income Share of Labor byGroup of CountriesHistorical U.S. Income Shareof Employees1970 75 80 85 90 95 2000 05455055606570751980 84 88 92 96 2000 0445505560657075 G-7 Economies(weighted)Advanced Economies(weighted)3Income share of employees1 Income share of labor2in real compensation per worker have beenmeager in unskilled sectors and the earningsgap between skilled and unskilled sectors haswidened by 25 percent. In Europe, real compensationper worker in unskilled sectors grewbroadly in line with that in skilled sectors, butemployment in unskilled sectors lost groundto employment in skilled sectors (and actuallycontracted by a cumulative 15 percent).14Turning to emerging market countries, theorywould predict that the globalization of laborwould bring large benefits for workers in theform of wage convergence toward the levels inadvanced economies. Data from the manufacturingsector confirm that real wages in emergingmarket countries, particularly in Asia, have beencatching up with those in the United States(Figure 5.10). Real wages (corrected for purchasingpower) have been converging rapidly andare relatively high in Asian countries that starteddeveloping earlier (Hong Kong SAR, Korea, Singapore,and Taiwan Province of China). Wagesin other Asian countries, including China, havebeen converging at a slower pace, though this hasaccelerated in recent years.15 Studies confirm thatboth trade and emigration have contributed torising incomes of nationals of developing countries,although the evidence on their impact oninequality is mixed (see Box 5.1 for a discussionof the evidence on the implications of globalizationfor labor markets in developing countries).Labor Compensation and the Globalization ofLabor: An Empirical ExaminationWhile striking, the globalization of labor isbut one of the forces that has been affecting thelabor markets of advanced economies over thepast two decades. Rapid technological changeis another central development with potentiallyimportant implications for labor market14Katz and Autor (1999) find similar changes in thegap between high- and low-income earners for the UnitedStates and European countries.15Asia’s labor productivity has also been convergingtoward the U.S. level (see the September 2006 WorldEconomic Outlook).1980 85 90 95 2000152025303540451980 85 90 95 200015202530354045Income Share of Labor inUnskilled Sectors(percent of economy-widevalue added)Income Share of Labor inSkilled Sectors(percent of economy-widevalue added)Income Share of Labor WithinUnskilled Sectors(percent of unskilled sectors'value added)1980 85 90 95 200050556065707580851980 85 90 95 20005055606570758085Income Share of Labor WithinSkilled Sectors(percent of skilled sectors'value added)United States EuropeOther Anglo-Saxon4 Japan5Figure 5.7 (concluded)Sources: Haver Analytics; International Labor Organization, Labor Statistics Database; OECD,Employment and Labor Market Statistics, National Accounts Statistics, and STAN IndustrialDatabase; United Nations, National Accounts Statistics (2004); and IMF staff calculations.Income share of employees is the ratio of employees' labor compensation to value added.The income share of labor estimates the share of labor compensation of employees and"nonemployee" workers in value added.Advanced economies include Australia, Austria, Belgium, Canada, Denmark, Finland, France,Germany, Ireland, Italy, Japan, the Netherlands, Norway, Portugal, Spain, Sweden, the UnitedKingdom, and the United States; weighted using series on GDP in U.S. dollars from the WorldEconomic Outlook database.Anglo-Saxon economies include Australia, Canada, and the United Kingdom. Australia isexcluded from the analysis by skill level due to lack of data.Europe includes Austria, Belgium, Denmark, Finland, France, Germany, Ireland, Italy, theNetherlands, Norway, Portugal, Spain, and Sweden. Ireland, the Netherlands, and Spain areexcluded from the analysis by skill level due to lack of data.5432 1How Has the Globalization of Labor Affected Workers in Advanced Economies?169CHA PTER 5 The Globalization of Labor170outcomes (Figure 5.11 ). The information andcommunications technology (ICT) revolution, anall-purpose technological revolution that Blinder(2006) has compared to the third industrial revolution,has stimulated capital accumulation (seethe September 2001 World Economic Outlook) andfavored skilled labor—with which it is more complementary—over unskilled labor. Technologyhas also progressed in other areas as reflected inthe strong rise in patent applications in OECDeconomies, especially since the early 1 990s.There have also been changes in labor andproduct market policies. Reforms have proceededin several areas, but generally in thedirection of lowering the cost of labor to businessand enhancing the flexibility of markets. Fourmain developments in labor market policies areparticularly worth noting (see Figure 5.11 ): (1) amarked increase in the generosity of unemploymentbenefits in Europe (as measured by averagereplacement rate of income), in contrast witha slight decline in Anglo-Saxon countries; (2) ageneral decline in the tax wedge, especially inthe United States where it has fallen by about10 percentage points since 1 995; (3) substantialdeclines in legislated employment protection andproduct market regulation, especially in Europeand Japan, both of which started with particularlyrestrictive stances; and (4) persisting large crosscountrydifferences in the degree of employmentprotection, with low protection in the UnitedStates and other Anglo-Saxon countries andrelatively high protection in Europe and Japan.Recent studies (Bassanini and Duval, 2006; andAnnett, 2006) have highlighted reductions in thetax wedge, reductions in unemployment benefits,deregulation of product markets, and more limitedemployment protection as the main factorsthat have contributed to employment growthand declining unemployment.16 Disentangling16Some of these variables may also affect the laborshare in similar ways, especially if the elasticity of substitutionbetween labor and capital is high. For instance, anincrease in the unemployment benefit replacement rateincreases the reservation wage of workers and leads in thevery short run to a rise in the labor share. But as employmentadjusts downward, the labor share declines and canFigure 5.8. Advanced Economies: Labor Compensationand Employment(Index, 1980 = 100)Real total laborcompensationReal labor compensationper workerEmploymentDespite the fall in the overall labor share, real labor compensation has grownrobustly in advanced economies, with a stronger weight on employment inAnglo-Saxon economies.Sources: Haver Analytics; International Labor Organization, Labor Statistics Database;OECD, Employment and Labor Market Statistics, National Accounts Statistics, and STANIndustrial Database; United Nations, National Accounts Statistics (2004); and IMF staffcalculations.Advanced economies include Australia, Austria, Belgium, Canada, Denmark, Finland,France, Germany, Ireland, Italy, Japan, the Netherlands, Norway, Portugal, Spain, Sweden,the United Kingdom, and the United States; weighted using series on GDP in U.S. dollarsfrom the World Economic Outlook database.Anglo-Saxon economies include Australia, Canada, and the United Kingdom.Europe includes Austria, Belgium, Denmark, Finland, France, Germany, Ireland, Italy,the Netherlands, Norway, Portugal, Spain, and Sweden.1231980 85 90 95 2000 05801001201401601802001980 85 90 95 2000 0580100120140160180Advanced Economies United States 200(weighted)11980 85 90 95 2000 0580100120140160180Europe3 2001980 85 90 95 2000 0580100120140160180200 Other Anglo-Saxon21980 85 90 95 2000 0580100120140160180200 Japanthe influence of these variables is difficult, inparticular because technological change and theglobalization of labor may be expected to affectcompensation and the labor share in similarways. The influence of policy variables is complex,particularly because they may also affect thelabor share indirectly by facilitating or obstructingthe adjustment of the economy to laborglobalization and technological progress.This section uses an econometric model toanalyze the relationship between labor compensationand labor globalization—measuredin terms-of-trade prices, offshoring, andimmigration—controlling for technologicalprogress and changes in labor market policies.The basic model, which has solid microeconomicfoundations and is widely used in the recenttrade literature (see, for instance, Feenstra, 2004;Harrigan, 1998; and Kohli, 1991), relates thelabor share to the capital-labor ratio and importand export prices (expressed relative to domesticprices).17 The two latter variables capturethe effects of globalization of trade: declinesin import prices are expected to decrease thelabor share, as imports that come increasinglyfrom developing countries are labor intensive; incontrast, declines in export prices should benefitlabor relative to capital because of the highcapital intensity of advanced economies’ exports.The basic model is augmented to include theintensity of offshoring, the share of immigrantsin the domestic labor force, the share of ICTcapital in total capital, measures of labor marketpolicies, and country fixed effects.18 The modelfall below its initial level if the elasticity of substitutionbetween capital and labor is high enough (Blanchard,1998). Other shocks that increase the cost of labor, such asan increase in the tax wedge or an increase in employmentprotection, can be expected to have similar effects.Although strict product market regulation creates rents, itis not clear that it should affect the distribution of theserents between labor and capital and hence the labor share.17The factor share equations are derived from the maxi-mization of an (economy-wide) revenue function, taking asgiven the factor endowments and sectoral prices (import,export, and absorption). See Appendix 5.1 for more details.18The theoretical rationale for including these variablesis that they may act as shift factors in the revenue (GNP)function (Feenstra, 2004).Despite strong growth of labor compensation overall, the growth in laborcompensation of unskilled sectors has been very slow. In the United States, theearnings gap between skilled and unskilled workers has widened by about 25 percentsince 1980, while in Europe, employment in unskilled sectors has contracted.1980 85 90 95 200090100110120130140150160Labor Compensation in Unskilled Sectors (weighted)1 170Figure 5.9. Advanced Economies: Labor Compensationand Employment in Skilled and Unskilled Sectors(Index, 1980 = 100)Real total laborcompensationReal labor compensationper workerEmploymentOther Anglo-Saxon2 JapanUnited States Europe 3Sources: Haver Analytics; International Labor Organization, Labor Statistics Database;OECD, Employment and Labour Market Statistics, National Accounts Statistics, and STANIndustrial Database; United Nations, National Accounts Statistics (2004); and IMF staffcalculations.For analysis by skill level, advanced economies include Austria, Belgium, Canada,Denmark, Finland, France, Germany, Italy, Japan, Norway, Portugal, Sweden, the UnitedKingdom, and the United States; weighted using series on GDP in U.S. dollars from theWorld Economic Outlook database.For analysis by skill level, Anglo-Saxon economies include Canada and the UnitedKingdom.For analysis by skill level, Europe includes Austria, Belgium, Denmark, Finland, France,Germany, Italy, Norway, Portugal, and Sweden.1231980 85 90 95 2000901001101201301401501601701980 85 90 95 200090100110120130140150160Index of Skilled to Unskilled 170EmploymentIndex of Skilled to UnskilledReal Labor Compensationper Worker1980 85 90 95 200080901001101201301401980 85 90 95 20008090100110120130140 Index of Unskilled Real LaborCompensation per WorkerIndex of Unskilled EmploymentHow Has the Globalization of Labor Affected Workers in Advanced Economies?171CHA PTER 5 The Globalization of Labor172was estimated on a panel of 18 advanced OECDeconomies over 1982–2002, for the overalllabor share and for the income shares of laborin skilled and unskilled sectors (see Appendix5.1 for more details). At the outset, it shouldbe noted that the effects of globalization canonly be imperfectly disentangled from those oftechnology, especially for technological progressin transport and communication, whichvastly expands the opportunities for globalizedproduction. Similarly, part of the decline inimport (and, in some cases, export) prices maybe attributable to productivity improvements inthe production of information and communicationstechnology.The results from estimating this modelsuggest that labor globalization, technologicalchange, and labor market policies have allaffected labor shares over the past two decades(Figure 5.12).19 Both labor globalization andtechnological progress have acted to reduce thelabor share, with the impact of technologicalprogress being somewhat larger, while changesin labor market policies have generally had asmaller but positive impact on the labor share.2019The contribution of a factor to the average annualchange in the labor share over the sample period is theproduct of its coefficient and of its own average annualchange over the same period.20Most studies have focused on explaining the declinein the relative wage (or labor share) of unskilled workersin the United States (see Freeman, 1995; and Feenstra,2004, for a survey). Studies that attempt to explain theevolution of the overall labor share are more scarce.Most studies conclude that skill-biased technologicalchange is a more important cause of wage inequalitythan trade (e.g., Harrigan, 1998; and Harrigan andBalaban, 1999). Feenstra (2004 and 2007) finds thatthe role of trade and technological progress are equallyimportant in explaining rising wage inequality. In arecent contribution, Guscina (2006) finds that laborshares across countries are equally affected by technologicalprogress and openness. Harrison (2002) alsofinds that globalization tends to reduce the labor share.Another strand of the literature examines whetherglobalization increases the elasticity of labor demand towages and finds mixed results (see, for instance, Slaughter,2001; and OECD, 2006a). Studies of immigrationtend to find that its effects on wages and employment ofnatives are small (Greenwood, Hunt, and Kohli, 1996;and OECD, 2006b).Figure 5.10. Catch-Up by Emerging Markets'Manufacturing Wages(Percent of U.S. manufacturing wages in constant PPP dollars)Sources: UNIDO, Industrial Statistics Database (2006); CEIC Asia Database; InstitutoBrasileiro de Geografia e Estatística, Pesquisa Industrial Mensal; and IMF staff calculations.1970 75 80 85 90 95 2000010203040506070Emerging Asia1970 75 80 85 90 95 2000020406080100Emerging Latin AmericaWhile relative manufacturing wages in emerging Asia are generally increasing,catch-up has been most evident in the newly industrializing economies, such asKorea, Singapore, and Hong Kong SAR. In contrast, Latin American economies havenot experienced much convergence.China Korea MalaysiaIndiaIndonesiaSingaporeHong Kong SARThailandBrazil ColombiaChile MexicoVenezuelaFigure 5.11. Information and CommunicationsTechnology (ICT) Capital, Patents, and Labor MarketIndicatorsUnited StatesOther Anglo-Saxon1 JapanEurope2The globalization of labor is but one of the forces that have influenced labormarkets in advanced economies over the past two decades. Rapid technologicalchange and changes in labor and product market policies are other significantdevelopments with potentially important implications for labor market outcomes.1982 87 92 97 200201234561982 87 92 97 20020123456 Employment ProtectionLegislation IndexProduct Market Regulation Index1985 90 95 20000204060801001201401601801980 85 90 95 20000123456 ICT Capital(percent of total capital)Number of Patents perWorking-Age Person 31982 87 92 97 200201020304050601982 87 92 97 20020102030405060 Average Replacement Ratefor Unemployment Benefits(percent of income)Tax Wedge(percent of labor cost)4Sources: Bassanini and Duval (2006); Jorgenson and Vu (2005); OECD, Science andTechnology Statistics; and IMF staff calculations.Australia, Canada, New Zealand, and the United Kingdom.Austria, Belgium, Denmark, Finland, France, Germany, Ireland, Italy, the Netherlands,Norway, Portugal, Spain, Sweden, and Switzerland. Greece is not included due to datalimitations.Patents that have been filed at the European Patent Office, Japanese Patent Office, andgranted by the United States Patent and Trademark Office (measured by priority year, thatis, year of first application).Difference between the labor cost to the employer and the net take-home pay of theemployee, in percent of the labor cost.1234Each channel of labor globalization (tradeprices, offshoring, and immigration) individuallyplays a relatively small role in explaining thedecline in the labor share.Labor globalization contributed to thedecline in labor shares in most countries, withbroadly similar effects in both Anglo-Saxoncountries and Europe.21 Nevertheless, the laborglobalization effect in the Anglo-Saxon countriesand Europe is driven by different factors.Europe’s labor share has been affected bothby offshoring and immigration, while, in theAnglo-Saxon countries, offshoring was a somewhatless important factor. Similarly, withinEurope, large economies were affected moreby immigration than by offshoring, while theopposite holds for small economies. Anothercomponent of globalization—the change intrade prices—generally had only a small neteffect on the labor share. Hence, while globalizationexerted downward pressure on the laborshare through declines in import prices, thiseffect has been broadly compensated by similardeclines in export prices, which have boostedthe labor share since exports of advancedeconomies are capital intensive. In large Europeancountries and Japan, the net effect fromchanges in trade prices was actually to boostthe labor share, likely reflecting a strongerconcentrationof exports in capital-intensivegoods.The reasons for the milder decline of thelabor share in the Anglo-Saxon countries thanin Europe are found in the role of technologicalchange and labor market policies rather than inthe differences in the impact of labor globalization.Technological change has contributed tothe reduction of the labor share in both groups,but less so in the Anglo-Saxon countries. Inparticular, in the United States, ICT capital evencontributed to raising the labor share, possiblyreflecting the fact that the United States is most21Some caution is needed when interpreting theseresults, since they are based on the regression coefficientsthat are the same for all the countries and averageannual changes in variables that are country specific.How Has the Globalization of Labor Affected Workers in Advanced Economies?173CHA PTER 5 The Globalization of Labor174-0.50 -0.25 0.00 0.25Europe largeFigure 5.12. Contributions to the Annual Change in Labor Share(Percentage points)1Labor globalization and technological progress have acted to reduce the labor share, with the impact of technological progress being somewhat larger,while changes in labor market policies have generally had a smaller but positive impact on the labor share. Trade prices, offshoring, and immigrationindividually play a relatively small role in explaining the decline in the labor share.Decomposing Changes in Labor Share Decomposing the Contribution of Labor GlobalizationSource: IMF staff calculations.1982–2002 or longest period available. 1986–2001 for Japan, as changes in the relative import price in earlier years reflected the yen's strong appreciation ratherthan globalization. The contributions are based on estimated regression coefficients and average annual changes in the respective variables by country (see Appendix5.1).The annual change in the labor share in this figure corresponds to the sample period for which all the regression variables were available and may thus differ fromthe one shown in Figure 5.7.Europe large includes France, Germany, Italy, and Spain.Europe small covers Austria, Belgium, Denmark, Finland, Ireland, the Netherlands, Norway, Portugal, and Sweden.Anglo-Saxon countries include Australia, Canada, the United Kingdom, and the United States.12345EuropeEurope smallAnglo-SaxonJapanAll advancedeconomiesEuropeJapanAll advancedeconomiesContribution of Technological ChangeEuropeJapanAll advancedeconomiesContribution of Labor GlobalizationEuropeJapanAll advancedeconomiesContribution of Labor Market PoliciesChange in the Labor Share-0.50 -0.25 0.00 0.25EuropeJapanAll advancedeconomiesEuropeJapanAll advancedeconomiesContribution of Trade PricesEuropeJapanAll advancedeconomiesContribution of OffshoringEuropeJapanAll advancedeconomiesContribution of ImmigrationContribution of Labor Globalization3452Europe largeEurope smallAnglo-Saxon345Europe largeEurope smallAnglo-Saxon345Europe largeEurope smallAnglo-Saxon345Europe largeEurope smallAnglo-Saxon345Europe largeEurope smallAnglo-Saxon345Europe largeEurope smallAnglo-Saxon345Europe largeEurope smallAnglo-Saxon345This box reviews the evidence on the effectsof emigration and trade on labor markets andincomes in developing countries.EmigrationWhile a vast theoretical and empirical literatureconsiders the impact of immigration on destinationcountries, little work has been done onemigration and its impact on source countries.This is surprising because the shares of the laborforce leaving many individual source countriesas emigrants are considerably higher than theproportionate changes in the labor force ofmany receiving countries due to immigration. Tocite a few examples, the labor force in Barbados,Belize, El Salvador, Guyana, and Jamaicahas been reduced by 20 percent or more due toemigration to the OECD countries. Meanwhile,immigrants constitute about 15 percent of theU.S. labor force, and the share is considerablylower in most other OECD countries.In general, source countries do not recordinformation on those who emigrate. However,Mexico and other Latin American countries—from where immigration is mostly to the UnitedStates—offer ideal case studies because U.S. datasources can be used to analyze the impact on thesource countries. Along these lines, Cardarelliand Ueda (2004) assess the impact of migrationto the United States on the welfare of sourcecountries. Using as a yardstick the incomeproduced by the nationals of the country irrespectiveof where they live, they estimate thatthe well-being of Mexican-born people was, onaverage, 20 percent higher than the country’sGDP alone would suggest over 1994–2003.Cardarelli and Ueda also conclude that immi-Note: The main author of this box is Prachi Mishra.See Borjas (1994 and 1995) for surveys of theempirical literature on immigration.The outflow of workers is largely to the UnitedStates, and took place between 1970 and 2000. In1965, the United States implemented the Immigrationand Nationality Act, which changed the basis of entryinto the United States from country quotas to familybasedreunification. This brought about a drasticchange in the composition of immigration, increasingthe share of migrants from developing countries.gration opportunities to the United States haveraised the well-being of nationals born in severalother developing countries, particularly in LatinAmerica and the Caribbean (e.g., Jamaica, Haiti,Nicaragua, and El Salvador) and in the Philippinesand Vietnam. One channel of incomegains for developing country residents, includedin these calculations, is the large flow of remittancesback into the country from emigrantsliving abroad (see the April 2005 World EconomicOutlook). While in Mexico annual remittanceswere about 3 percent of GDP over 1 990–2003,they amounted to over 10 percent of GDP in ElSalvador and Jamaica over the same period.Focusing on workers who have stayed home,Mishra (2007) examines the effect of emigrationto the United States on wages in Mexico,using data from the Mexican and U.S. censusesfor 1 970–2000. She finds a strong and positiveeffect of emigration on Mexican wages: a 10 percentdecrease in the number of Mexican workersin a given skill group (defined by schooling andexperience) increases the average wage in thatskill group by about 4 percent (Aydemir andBorjas, 2006, find a similar result). The impacton wages differs dramatically across schoolinggroups, with the greatest increase being for thehigher wage earners (those with 12–15 yearsof schooling) owing to the higher emigrationrate of this group. Hence, while all categoriesof workers who stay home benefit in terms ofhigher wages, emigration could serve as a partialexplanation for the increasing wage inequalityin Mexico.The positive effect of emigration on wages inMexico is confirmed by Hanson (forthcoming).He examines changes in the distribution oflabor income across regions of Mexico duringthe 1990s, a period of rapid globalizationof the Mexican economy. He finds that overthe decade, average hourly earnings in high-Emigration accounts for approximately 37 percentof the increase in relative wages of high school graduates(12 years of schooling) and 14 percent of theincrease in relative wages of those with some collegeeducation (13–15 years of schooling) between 1990and 2000.Box 5.1. E migration and Trade: How Do They Affect Developing Countries?How Has the Globalization of Labor Affected Workers in Advanced Economies?175CHA PTER 5 The Globalization of Labor176migrationstates rose by 6–9 percent relative tolow-migration states.While workers benefit from higher wages andfamilies from remittance inflows, capital ownerswho hire these workers lose. Overall, however,estimates suggest that there is a small aggregateannual welfare gain in the case of Mexico.Nevertheless, emigration can lead to loss ofwelfare if the fact that emigration of high-skilledworkers leads to a decline in the productivityof those who have stayed behind is takeninto account. For example, qualified doctors,researchers, and engineers confer a positiveexternality on the rest of the population, andthis is lost when they emigrate. Mishra (2006)estimates substantial productivity losses for thosewho stay behind because of the very high ratesof high-skilled emigration from the Caribbeancountries. Gupta, Pattillo, and Wagh (2007)also report a high rate of migration of skilledworkers from sub-Saharan African countries.One consequence of this is a human resourceshortage in the health sector of these countries,as skilled health care professionals get hired inthe high-demand OECD countries.TradeA large body of research shows that tradeopenness in developing countries has raisedaggregate incomes and growth rates (see Bergand Krueger, 2003, for a survey). Using crosscountryand panel regressions, many studieshave found that openness to trade is a significantexplanatory variable for the level or growthrate of real GDP per capita, with the weightof evidence suggesting that this result holdseven when the endogeneity of trade opennessis taken into account and after controlling forother important determinants, such as the qualityof institutions and geography.In contrast, the internal distributional consequencesof trade reform in developing countriesare still the subject of intense debate (see Goldbergand Pavcnik, forthcoming, for a survey).The workhorse model to analyze the labor marketconsequences of trade liberalization—theStolper-Samuelson theorem—predicts that tradeliberalization will shift income toward a country’sabundant factor. For developing countries,this suggests that liberalization will principallybenefit the abundant unskilled labor. Yet manydeveloping countries, including Argentina, Brazil,Colombia, China, India, and Mexico experienceda widening wage gap between skilled andunskilled labor during periods of trade reformduring the 1980s and 1990s.Of course, rising wage inequality does notnecessarily imply a causal impact of tradereforms (since typically trade reforms wereaccompanied by significant domestic reforms inmost countries). Hence, the literature in thepast decade has focused on trying to identifythe causal link between trade liberalization anddistributional outcomes. Two key methodologiesused are the industry-level and the regionalapproaches that examine whether industriesor regions that were more exposed to tradeliberalization experienced larger changes inlabor market outcomes. However, a drawback ofboth these approaches is that they can directlyDavis and Mishra (2007) discuss a variety of reasonsfor why the assumptions underlying the Stolper-Samuelson model may be too simplistic to hold in thereal world. One possible reason is that the patternof trade depends on a country’s “local” rather thanglobal factor abundance: a country’s factor abundanceneeds to be compared with that of others thatproduce the same set of goods. For example, Mexicois less skill abundant relative to the United States butmore skill abundant relative to China. When Mexicojoined the General Agreement on Tariffs and Trade inthe mid-1980s, it opened its borders to the less-skill-abundant world, which could explain the risingwage inequality in the late 1980s.The definition of skill varies across specific countrystudies. Studies using household survey data defineskill based on education of the household head,whereas studies using plant- or firm-level data typicallydifferentiate between production and nonproduction,or blue-collar and white-collar, workers.Other explanations of the rising wage gaps includeskill-biased technological change or increased offshoringof activities that are relatively skill intensive fromthe point of view of developing countries (thoughoffshoring may itself be triggered by a free tradeagreement with an advanced economy, leading toreduced tariffs) (Feenstra, 2007).Box 5.1 (concluded)advanced in the use of ICT. The adverse labordemand effects of ICT appear to be strongerat the early stages of ICT adoption, before theneeded adjustments in workers’ education havetaken place.Changes in labor market policies have hada positive effect on the labor share in Anglo-Saxon countries, but a much more modesteffect on average in Europe, particularly inlarge European economies where labor policiesare estimated to have actually contributed to adecline in the labor share. The contribution oflabor market policies is driven primarily by thechanges in the tax wedge and unemploymentidentify only relative differences across regionsor industries and not identify the impact on thenation as a whole.The econometric evidence from differentcountries is mixed on how trade reforms affectrelative labor market outcomes across regionsor industries. Topalova (2005) and Edmonds,Pavcnik, and Topalova (2007) find that districtsin India that were more exposed to importliberalization experienced a slower reductionin poverty, which was coupled with lower investmentin human capital and a lower declinein child labor. On the other hand, using abroader measure of openness, Hanson (2007)finds that states in Mexico with high exposureto globalization (measured by the shares offoreign direct investment, imports, and exportsassembly in state GDP) experienced a rise inlabor incomes relative to low-exposure states inthe 1990s.The empirical evidence on the effect of tradeliberalization on wages at the industry level isPorto (2006) is one study that uses a generalequilibrium model of trade to answer the ambitiousquestion of the overall effect of trade liberalizationon inequality, in the context of Argentinean tradereforms. The model is used to simulate the effect oftrade policy changes on the distribution of householdwelfare (household expenditure per capita). He findsevidence of a pro-poor bias caused by the reform. Onaverage, poor households gained more from reformsthan did middle-income households. However, thedrawback of this approach is that predictions of themodel depend crucially on parameter estimates thatare typically not known (e.g., wage-price elasticities)and are difficult to estimate consistently with timeseriesdata on wages and prices when many otherpolicies changed along with trade (see Goldberg andPavcnik, forthcoming, for a discussion of this paper).also mixed. For example, studies find no significantrelationship between trade policy andindustry wages in Brazil and Mexico (Pavcnikand others, 2004; and Feliciano, 2001), whilethe reduction in tariffs within a sector is foundto be associated with a significant reduction inwages in that sector in Colombia (Goldbergand Pavcnik, 2005) but with an increase inwages in Poland (Goh and Javorcik, 2007). Theevidence from India on the effects of changesin tariffs on wages is mixed (see Topalova, 2005;Dutta, 2004; and Kumar and Mishra, forthcoming).Given that the sectors that experiencedthe largest tariff reductions were those withthe largest share of unskilled workers, theindustry-level studies thus suggest mixed effectsof trade liberalization on the overall wage gapbetween skilled and unskilled workers: tradereforms were associated with a higher wage gapin Colombia, possibly with an unchanged wagegap in Brazil and Mexico, and with lower wageinequality in Poland.In conclusion, on the one hand, emigrationand trade both increase the aggregate incomesof developing countries (once the income ofemigrants is included). On the other hand, theexisting evidence on the impact of globalizationon inequality is mixed, particularly in the caseof trade. Further research efforts are needed tofully understand these important issues.These studies use a two-step methodology. First,they use household survey data to estimate “industrywage premia,” defined as the part of worker wagesthat is explained by a worker’s industry affiliationafter controlling for observable worker characteristics(e.g., schooling, experience, and so on). Second,the estimated industry wage premia are regressed onmeasures of trade reform by industry.How Has the Globalization of Labor Affected Workers in Advanced Economies?177CHA PTER 5 The Globalization of Labor178benefit replacement.22 The decline in the taxwedge in Anglo-Saxon countries, especially inthe United States, benefited the labor share,while in Europe the labor share was hurt bya rise in unemployment benefit replacementrates.Turning to look at the skilled and unskilledsectors separately, the main factor affecting theincome share of labor in unskilled sectors overthe sample period, beyond the shift of employmenttoward skilled sectors, is technologicalchange (Figure 5.13). This result is consistentwith the belief that computers and other ICTequipment act as a substitute for unskilled labor,but they tend to complement skilled labor. Onthe other hand, labor globalization contributedto a decline in the income share of labor inskilled sectors, much more so than in unskilledsectors. This is in line with earlier findings thatthe increase in offshoring was mostly drivenby the offshoring of skilled inputs rather thanunskilled inputs. However, this was more thanoffset by the shift of employment from unskilledsectors to skilled sectors and the income shareof labor in skilled sectors actually increasedmoderately.23Of course, the effects on labor shares do notby themselves give the full picture of how workers’well-being is affected by forces of globalizationand technological change. These factorsalso influence output and total labor compensa-22The other variables, namely, employment protectionlegislation, product market regulation, and uniondensity, did not have significant effects. The analysis wasexpanded to investigate whether some labor market institutionstend to amplify or attenuate the impact of laborglobalization and technological progress. Although strictemployment protection legislation does not appear tohave any effect on its own, there is some evidence that ittends to increase the adverse effects of labor globalizationon labor shares. A more flexible labor market may thuscontribute to limiting the decline in the overall laborshare caused by globalization.23Workers in unskilled sectors have also benefitedsomewhat less from labor market policy changes.Although product market regulation has a negligibleimpact on the overall labor share, it seems to benefit theincome share of labor in unskilled sectors. Hence, thereduction in product market regulation over the sampleperiod had a negative effect on this income share.Labor in Skilled Sectors-0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3Change in the labor shareContribution of labor globalizatonContribution of technological changeContribution of labor market policiesContribution of employment shifts toskilled sectorsFigure 5.13. Advanced Economies: Contributions tothe Annual Change in the Labor Share by Skill Level(Percentage points)1Source: IMF staff calculations.1982–2002 or longest period available. 1986–2001 for Japan, as changes in therelative import price in earlier years reflected the yen's strong appreciation ratherthan globalization. The contributions are based on estimated regression coefficientsand average annual changes in the respective variables by country (see Appendix 5.1).1-0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3Change in the labor shareContribution of labor globalizatonContribution of technological changeContribution of labor market policiesContribution of employment shiftsaway from unskilled sectorsWhile technological change affected mostly the income share of labor inunskilled sectors, the labor income share in skilled sectors was moreaffected by labor globalization.Labor in Unskilled Sectorstion. The model results imply that on average,in advanced economies, the decline in tradedgoods prices yielded about a 6 percent increasein both output and total labor compensation, inreal terms, over 25 years.24 Thus, although thelabor share went down, globalization of laboras manifested in cheaper imports in advancedeconomies has increased the “size of the pie”to be shared among all citizens, resulting in anet gain in total workers’ compensation in realterms (Figure 5.14).25In sum, the econometric analysis suggeststhat both labor globalization and technologicalchange have been important factors behind theobserved decline in labor shares in advancedeconomies. The rapid progress in ICT has hada particularly strong effect on the unskilledsectors. The role of labor market policies hasdiffered across countries, with positive effectslargest in the United States and much moremodest on average in Europe (and negative insome countries). Finally, global competitionhas brought down international trade prices.Cheaper imports have increased the size of realtotal labor compensation, implying that workershave participated in the benefits of the biggereconomic “pie,” although their share of it hasdeclined.24This result was calculated as follows. The modelallows for deriving elasticities of labor compensationto trade prices: on average, a 1 percent declinein the relativeprice of imports raises real total laborcompensationby 0.5 percent, while a 1 percent declinein the relative price of exports lowers it by a somewhatsmaller 0.4 percent. Combining these elasticities withthe actual average changes in relative export and importprices implies an average annual increase in labor compensationof about 0.2 percent on average in advancedeconomies (or about 6 percent if compounded over 25years). The increase in output implied by the change intrade prices is just the difference between the percentchange of total labor compensation and the percentchange of the labor share (which is very small in thiscase).25It should also be noted that a comprehensive evaluationof the impact of globalization on workers’ financialmeans needs to go beyond labor compensation andto take into account an increase in direct and indirectasset ownership (see the September 2006 World EconomicOutlook).EuropeEurope largeEurope smallAnglo-SaxonJapan-0.1 0.0 0.1 0.2 0.3 0.4 0.5Impact of Change in Trade Prices on Annual Change in Labor Share2Although on the whole the labor share went down, the globalization of laborincreases the size of the pie to be shared among all citizens, resulting in a net gainin workers’ compensation in most countries.34Figure 5.14. Effects of Changes in Trade Prices on LaborShare, Output, and Labor Compensation(Percent)1EuropeEurope largeEurope smallAnglo-SaxonJapan-0.1 0.0 0.1 0.2 0.3 0.4 0.5Impact of Change in Trade Prices on Annual Change in Output234EuropeEurope largeEurope smallAnglo-SaxonJapan-0.1 0.0 0.1 0.2 0.3 0.4 0.5Impact of Change in Trade Prices on Annual Change in Labor Compensation234Source: IMF staff calculations.1980–2004 or longest period available. 1986–2004 for Japan, as changesin the relative import price in earlier years reflected the yen’s strongappreciation rather than globalization. The effects are based on estimatedregression coefficients and average annual changes in the respectivevariables by country (see Appendix 5.1).Europe large includes France, Germany, Italy, and Spain.Europe small covers Austria, Belgium, Denmark, Finland, Ireland, theNetherlands, Norway, Portugal, and Sweden.Anglo-Saxon countries include Australia, Canada, the United Kingdom, andthe United States.1234All advancedeconomiesAll advancedeconomiesAll advancedeconomiesHow Has the Globalization of Labor Affected Workers in Advanced Economies?179CHA PTER 5 The Globalization of Labor180Summary and Policy ImplicationsThere has been a dramatic increase in the sizeof the effective global labor force over the pasttwo decades, with one measure suggesting it hasrisen fourfold. This expansion is expected tocontinue in the coming years. The UN projectsa 40 percent rise in the world’s working-agepopulation by 2050, and trade openness willcontinue to grow, especially in services. Indeed,tentative projections suggest that the effectiveglobal labor supply could more than doubleagain by 2050.26The global pool of labor can be accessedby advanced economies through imports andimmigration. Trade is the more important andfaster-expanding channel, in large part becauseimmigration remains very restricted in manycountries. Contrary to popular perceptions,the intensity of offshoring of the productionof intermediates is still small in the overalleconomy, although the manufacturing sector ismore affected because of its greater tradability.Imports of offshored intermediates have alsobeen growing somewhat more slowly than totaltrade.The integration of workers from emergingmarket and developing countries into the globalworkforce has produced important benefits foradvanced economies. Export opportunities haveexpanded considerably. It has provided accessto cheaper imported goods and has enabledcompanies to operate more efficiently. This hasboosted productivity and output, and contributedto rising real labor compensation. Foremerging market economies, the ongoing integrationof labor into the global marketplace hasbenefited workers, with manufacturing wagesrising rapidly.26This projection is based on the medium variant ofthe UN projections of working-age population and on theassumption that the ratio of non-oil exports to GDP willcontinue expanding at the rate observed in recent years(see Appendix 5.1). World Bank (2006) also provides projectionsof the world’s workforce until 2030 and projectsthat although the vast majority of the world’s workforcewill remain unskilled, the supply of skilled workers islikely to grow faster than that of unskilled workers.Nevertheless, labor globalization has negativelyaffected the share of income accruing to laborin the advanced economies (the labor share).It is, however, only one of several factors thathave affected the labor share over the past twodecades. Rapid technological change—especiallyin the information and communications sectors—has had a bigger impact, particularly on thelabor share in unskilled sectors. This is broadlyconsistent with findings highlighted in a recentjoint study by the International Labor Officeand the World Trade Organization (2007).Against this background, the increasing globalizationof labor and ongoing technologicalchanges raise important challenges for policymakersin the advanced economies. They mustseek to harness the benefits that the growingpool of global labor is creating. This meanscontinuing along the path of trade liberalization,while ensuring that domestic economiesare sufficiently flexible to be able to adjust andrespond to the pressures of globalization. At thesame time, it is important to be fully cognizantof adjustment costs, and policies do need to supportthose people who are negatively affectedby labor market globalization and technologicalchanges. In broad terms, policies need torespond along three dimensions:• Improve the functioning of labor markets. Steps toreduce tax wedges to enable workers to takehome a larger proportion of their gross payand to ensure that unemployment benefitreplacement rates do not deter workers fromseeking employment have helped a numberof countries adjust to the pressures of globalization.The duration of unemploymentbenefits and the work availability requirementsare also important (see Annett, 2006;and Bassanini and Duval, 2006). Moreover,policies that increase the flexibility of theeconomy and thereby enable workers to movemore easily from declining to expanding areasof the economy help the process of adjustment.A variety of country-specific approachesare possible, as demonstrated by the range ofexperience of successful reformers in westernEurope (see Box 2.2). Reform packages alsohave to be designed with fiscal consequencesin mind.• Improve access to education and training. Developingworkers’ skills is necessary for keepingup with rapid technological change andfor continuing innovation. Skilled sectorshave been better able to adapt to changingconditionscaused by the ICT revolutionthan unskilled sectors. Further, countriesthat started adopting ICT and trainingworkers in this area earlier experienced lessdecline in their labor share. Workers mustalso be ready to compete with the growingpool of skilled workers in emerging markets,especially those in Asia. Beyond increasesin spending on education and training, thequality of this spending is crucial. Experienceshows that evaluation and targetingof trainingare important to maximize itsimpact.• Ensure adequate social protection for workers duringthe adjustment period. This includes providingadequate income support to cushion,but not obstruct, the process of change,and also making health care less dependenton continued employment and increasingthe portability of pension benefits in somecountries. The latter would also enhancethe flexibility of the economy by facilitatingthe move of workers from declining sectorsto expanding sectors. Whether measuresspecifically targeted at workers who have beendisplaced by international trade are desirableis less clear (see, e.g., OECD, 2005). The factthat these workers may face special hurdlesreintegrating into the labor market as theyare often older and less educated, and theirskills are specific to declining industries oroccupations, argues in favor of such measures.Also, minimizing losses for such workers mayincrease support for the international economicintegration process. However, it may bedifficult (even conceptually) to differentiatebetween job losses caused by globalizationand those caused by other factors, since mostlabor markets are characterized by high ratesof turnover and year-to-year earnings variability.If trade-displaced workers are treatedmore generously, including, for instance, bybeing provided supplementary wage subsidies,such compensation should be structured toavoid dulling incentives to search actively fornew jobs.Appendix 5.1. Data Sources and MethodsThe main authors of this appendix are FlorenceJaumotteand Irina Tytell.Variable Definitions and Data SourcesThis section provides further details on theconstruction of the variables used in this chapterand the sources of the data.Sectoral ClassificationThroughout the chapter, the analysis is carriedout both for the aggregate economy andfor a disaggregation of the economy by skillcategory. The classification of trade and laborinto skill categories is based on the skill intensityof the sector. Hence skilled exports are exportsof goods and services typically produced by skillintensivesectors. The skilled labor share is theshare of national income that accrues to workersin skill-intensive sectors. One drawback of thisapproach is that it does not capture changesthat occur between skilled and unskilled workerswithin sectors. A more refined approach was,however, not feasible because of the lack ofcross-country data on the wages of production(unskilled) and nonproduction(skilled) workers,which would have been needed to calculatelabor shares and labor compensation of skilledand unskilled workers.The classification of sectors into skilled andunskilled is based on the share of skilled workersin the labor force of the sector, where a personis considered skilled if he or she has at leastupper secondary education. Data on the averagefraction of skilled labor in each sector (across 16OECD economies from 1994 to 1998) are fromJean and Nicoletti (2002). The chapter classifies18 sectors (from the International StandardAppendix 5.1. Data Sources and Methods181CHA PTER 5 The Globalization of Labor182Industrial Classification, Revision 3) into twobroad aggregates, namely, unskilled and skilledsectors, as reported in Table 5.1. In order totest the robustness of the results, an alternativethree-category split was also used, whichdistinguishes between low-skill, medium-skill,and high-skill sectors. Figure 5.15 shows thatthe patterns of the labor shares (and real laborcompensation per worker and employment) forthe narrower high- and low-skilled aggregatesare similar to those for the broader skilled andunskilled aggregates.Labor Compensation and Labor SharesLabor compensation was calculated byaugmentingthe compensation of employeesfor the income of other categories of workers(self-employed, employers, and family workers).Following Gollin (2002) and for dataavailabilityreasons, it was assumed that othercategories of workers earn the same averagewage as employees. Labor compensation ishence the product of the compensation ofemployees and the ratio of total employmentTable 5.1. Classification of Sectors by Skill IntensityMain Classification Alternative ClassificationUnskilled Low skilledAgriculture AgricultureMining MiningFood and tobacco Food and tobaccoTextiles, apparel, and leather Textiles, apparel, and leatherWood WoodOther nonmetal products Other nonmetal productsMetals and metal products Metals and metal productsTransport equipment ConstructionOther manufacturingConstruction Medium skilledTrade, hotels, and restaurants Paper and publishingTransport equipmentSkilled Other manufacturingPaper and publishing UtilitiesFuel, chemicals, and rubber Trade, hotels, and restaurantsMachinery and equipment Transport and communicationsUtilitiesTransport and communications High skilledBusiness services Fuel, chemicals, and rubberSocial and personal services Machinery and equipmentBusiness servicesSocial and personal servicesSources: OECD; and IMF staff estimates.Figure 5.15. Advanced Economies' Labor Income Share,Labor Compensation, and Employment: Robustness toAlternative Skill ClassificationIncome Share of Labor(weighted; percent of GDP)1980 85 90 95 2000010203040501980 85 90 95 200001020304050High skilledLow skilledSkilledUnskilledReal Labor Compensation per Worker(weighted; index, 1980 = 100)1980 85 90 95 2000801001201401601980 85 90 95 200080100120140160SkilledUnskilledHigh skilledLow skilledEmployment(weighted; index, 1980 = 100)1980 85 90 95 2000801001201401601980 85 90 95 200080100120140160High skilledLow skilledSkilledUnskilledSources: Haver Analytics; International Labor Organization, Labor Statistics Database;OECD, Employment and Labour Market Statistics, National Accounts Statistics, and STANIndustrial Database; United Nations, National Accounts Statistics (2004); and IMF staffcalculations.For the analysis by skill level, advanced economies include Austria, Belgium, Canada,Denmark, Finland, France, Germany, Italy, Japan, Norway, Portugal, Sweden, the UnitedKingdom, and the United States; weighted using series on GDP in U.S. dollars from theWorld Economic Outlook database.11and employees.27 Other correction procedures(see Gollin, 2002, for a review), for which thedata are not widely available, yield similar patternsover the subset of the sample used in thischapter for which the data are available. Thiscorrection was applied at both the aggregateand the sectoral level of data. When sectoraldata on employees or total employment werenot available, the following procedure was used:• the ratio of total employment to total employeeswas assumed to be the same as in previousyears or, if it was not available for any year, itwas assumed to be equal to the average forthis sector across other OECD economies; and• the sum of “nonemployee” workers acrosssectors was constrained to add up to the totalfor the aggregate economy by scaling theimputed number of nonemployee workersproportionately.The variables are defined as follows. Reallabor compensation is labor compensationdeflated by the CPI index from the WorldEconomic Outlook database. The labor shareis calculated as the ratio of labor compensationand value added at basic prices.28 The share oflabor in skilled (unskilled) sectors is the ratio oflabor compensation in skilled (unskilled) sectorsto the economy-wide value added.The main data source is the OECD’s StructuralAnalysis (STAN) Database. However,several other sources were used to fill in missingdata and extend the series to the most recentyear possible. For employees’ compensation andvalue added, these include the OECD’s NationalAccounts Statistics, the UN’s National AccountsStatistics, and Haver Analytics (for Japan). Fordata on total employment and employees, theadditional sources were the OECD’s Employmentand Labor Market Statistics Database andthe ILO Labor Statistics Database. Due to data27Korea was excluded from the sample because someof the income of the self-employed is already in theemployees’ compensation, making it impossible to applythe correction (see also Young, 1995 and 2003).28The exceptions are Japan, where value added is measuredat producer prices, and the United States, where itis measured at market prices.availability reasons, the calculations were limitedto advanced OECD economies.Manufacturing WagesManufacturing wages for advanced and developingeconomies are from the UNIDO IndustrialStatistics Database. They were convertedinto constant purchasing power parity (PPP)dollars using CPI indices and PPP exchangerates from the World Economic Outlook database.The data for China are from the CEIC Asiadatabase.ImmigrationThe data on foreign labor force are from theOECD’s Trends in International Migration (2003edition for all countries except the UnitedStates). For the United States, the data are fromthe U.S. Census Bureau and the U.S. Bureauof Labor Statistics. Data for Italy, the Netherlands,Norway, and the United Kingdom arefor foreign employment instead of labor force.Data for Australia, Canada, and the UnitedStates refer to foreign-born labor force insteadof foreign labor force. The available series wereextended backward using growth rates from thestock of foreign (or foreign-born, in the caseof Australia, Canada, and the United States)population when available, and the missing yearswere interpolated.Data on emigration for 1990 and 2000 arefrom Docquier and Marfouk (2005) and refer tothe stock of emigrants to the OECD economies.Trade and OffshoringData on trade used in the chapter are from avariety of sources. Aggregate data on trade quantitiesand prices are from the World EconomicOutlook database, including for the non-oilgoods and services aggregates. Sectoral tradedata for advanced economies (used to constructskilled and unskilled trade) are from the OECDSTAN Industrial Database (for manufacturing)and from the OECD International Trade inServices Database (for services). For developingcountries, sectoral trade data were obtainedfrom the World Integrated Trade Solution (forAppendix 5.1. Data Sources and Methods183CHA PTER 5 The Globalization of Labor184manufacturing) and from the IMF’s Balance ofPayments Statistics (for services). Data on manufacturingtrade of advanced OECD economies bysource country are from the OECD STAN BilateralTrade Database. The services data for Indiawere extended using the CEIC Asia database.Offshore outsourcing is the outsourcingof intermediate production to companies inlocations outside the country, which can beforeign affiliates or independent companies.It is measured by the imports of intermediateinputs, as provided in the OECD Input-OutputTables (1995, 2002, and 2006 editions). Thesetables assume that an industry uses an importof a particular product in proportion to its totaluse of that product (“the import proportionalityassumption”), and this proportion is theeconomy-wide share of imports in domesticdemand. The measure used in the chapter onlyincludes nonfuel manufacturing and servicesinputs. Imported intermediate inputs of a sectorare scaled by either the sector’s gross output orits total use of intermediates. Sectoral offshoringintensities are then aggregated based on sectoralgross output weights. Finally, the data on theoverall offshoring intensity are interpolated formissing years.For years beyond 2000, the OECD data wereextended using the latest input-output tableavailable (2000 for most countries) and updatingthe data on the import proportions for eachcategory of intermediate input. The latter wasapproximated by the share of imports in domesticabsorption (consumption and investment)for that category of products (sector). Data onimports by sectors are from the OECD STANIndustrial Database for manufacturing and fromthe OECD International Trade in Services Databasefor services. Data on value added by sector(used to calculate absorption) are from a combinationof the OECD STAN Industrial Databaseand the Groningen 60-Industry Database.2929Sectoral offshoring intensities were aggregated usingsectoral value-added weights, due to the lack of data onsectoral gross output for the later years. The historicaland extended series were then spliced using growth rates.Imports of final goods and services are constructedas a residual by subtracting importedintermediate inputs from total imports.Global Labor SupplySeveral measures of the global labor supplyare calculated, including the world’s workingagepopulation, the world’s labor force, and anexport-weighted world’s labor force. The latterattempts to measure the presence of the countries’labor supply in the international marketand is calculated as the sum across countriesof national labor forces, each weighted by thecountry’s ratio of non-oil exports to GDP (Harriganand Balaban, 1999). The export-to-GDPratio is capped to one to limit the weight ofcountries specialized in re-export trade. Dataon working-age population and labor forceare from various sources, including the WorldEconomic Outlook, the World Bank’s WorldDevelopment Indicators, the United NationsPopulation Projections, and the CEIC Asia databases.The global labor supply by education levelis calculated using the Barro-Lee (2000) dataset on educational attainment of the populationaged 15 or more. It is assumed that the share ofthe labor force with higher education is aboutthe same as the share of the population aged 15or more with higher education. For the years2001–05, this share was extrapolated linearly foreach country.The projections of the global labor supplyfor 2006–50 are based on the UN projectionsof the working-age population. The labor forceparticipation rate in each country is assumed toconverge by 2050 to the current rate of laborforce participation in the United States. Assuminginstead that labor force participation ratesremain at their current levels does not havemuch effect on the global labor supply projections.Projections for the export-to-GDP ratioare based on country-specific World EconomicOutlook projections until 2012, and on the trendincrease observed in the world export-to-GDPratio for later years. Under these assumptions,the cumulative growth in the export-weightedglobal labor force over 2005–50 could rangefrom a low of 120 percent (under the low variantof the population projections) to a high of190 percent (under the high variant).Capital Stock and ICT CapitalFajnzylber and Lederman (1999) are thesource of the capital stock series for the entireeconomy. This data set extends the capital stockseries estimated by Nehru and Dhareshwar(1993) by adding the annual flow of gross fixedcapital formation and assuming a 4 percentdepreciation rate to the preexisting stock ofcapital.Jorgenson and Vu (2005) provide series onIT investment using national expenditure datafor computer hardware, software, and telecommunicationsequipment. A perpetual inventorymethod applies varying depreciation rates toestimate IT capital stock. This method assumesa geometric depreciation rate of 31.5 percentand a service life of 7 years for computer hardware,31.5 percent and 5 years for software, and11 percent and 11 years for telecommunicationsequipment.Labor Market Policy IndicatorsThe indicators of labor and product marketpolicies were provided by Bassanini and Duval(2006). The indicators are defined as follows:• Average unemployment benefit replacementrate is the average of the unemploymentbenefit replacement rates corresponding tomultiple income, family, and unemploymentduration situations. These include two incomesituations (100 percent and 67 percent ofthe average production worker earnings),three family situations (single, with dependentspouse, with spouse in work), and threeunemployment durations (1st year, 2nd and3rd years, and 4th and 5th years of unemployment).The original data are from theOECD’s Benefits and Wages Database.• Labor tax wedge is the difference betweenthe labor cost to the employer and thecorrespondingnet take-home pay of theemployee for a single-earner couple withtwo children earning 100 percent of theaverage production worker earnings. It isthus the sum of personal income tax and allsocial security contributions expressed as apercentageof the total labor cost. The originaldata are from the OECD Taxing WagesDatabase.• Employment protection legislation is theOECD summary indicator of the stringencyof Employment Protection Legislation. Theoriginal data are from the OECD EmploymentOutlook (2004).• Product market regulation is the OECD summaryindicator of regulatory impedimentsto product market competition in sevennonmanufacturing industries (gas, electricity,post, telecom, passenger air transport, railwayspassenger and freight services, and roadfreight). The original data are from Conwayand others (2006).• Union density measures the share of workersaffiliated with a trade union. The originaldata are from the OECD Employment Outlook(2004).Econometric ApproachThis section presents the model used to examinethe relationship between globalization andlabor shares and reports the results from theestimations.MethodologyThe econometric approach used in thischapter is based on a model used frequently inthe trade literature (see Feenstra, 2004; Harrigan,1998; and Kohli, 1991). The model uses arevenue function with fixed factor quantities (oflabor and capital) and exogenous product prices(of exports, imports, and domestic absorption).It assumes that firms are maximizing profits, allmarkets are competitive, and factors can movefreely between firms. The revenue function islinearly homogeneous and concave in factorquantities and convex in product prices. It istypically specified as a flexible translogarithmic,or translog, form (Christensen, Jorgenson, andLau, 1975).Appendix 5.1. Data Sources and Methods185CHA PTER 5 The Globalization of Labor186Using the translog revenue function, productand factor shares can be obtained asS i = – p–i–yi– = ai + Σaij lnpj + Σgij lnvj + ΣϕijzjGR i = –w–i–v–i = bi + Σgij lnpj + Σbij lnvj + Σφijzj ,Gwhere S are product shares, R are factor shares,p are product prices, v are factor quantities, andz are shift variables. The shift variables captureany factors that could be expected to shift therevenue function, for example, measures oftechnological progress or offshoring, as suggestedby Feenstra (2004). The share equationsare subject to a number of cross-equation restrictionsthat follow from symmetry and linearhomogeneity of the corresponding revenuefunction.Given the theme of this chapter, the estimationsfocused primarily on the labor share equations.The following equation was adopted asthe basic specification for the analysis:pE pM LRL = bL + gELln––– + gMLln––– + bLLln–––pA pA K+ φ LM KICT KICT LXX + φLM ––– + φLC –––– + φLC2(––––)2+ eL,L K Kwhere PE, PM, and PA are prices of exports,imports, and absorption; L is labor; K is capital;X is offshoring; LM is immigrant employment;and KICT is ICT capital. The relative prices andquantities are used to impose the necessaryhomogeneity restrictions. Labor shares are correctedfor the income of other (nonemployee)categories of workers, prices and the capitalstock variables are measured in 2000 U.S. dollars,and labor stock variables are representedby employment. Offshoring is measured as ashare of imported intermediate inputs in totalintermediate inputs, immigration is capturedas a share of immigrant employment in totaldomestic employment, and ICT capital ismodeled as a share of ICT capital in the totalcapital stock (more detail on these measurementsis provided above). The effect of ICTcapital is represented by a quadratic functionto reflect potential nonlinearities associatedwith the need for learning the new technology:the adverse effect on wages and employmentis likely to be greatest before workers acquirethe skills necessary to effectively handle thenew equipment. The model was estimated ona panel of 18 countries over 1982–2002 usingcountry fixed effects. The basic specificationwas extended to include several measures oflabor market policies, including the tax wedge,the replacement rate of unemployment benefits,indices of product market regulation,employment protection legislation, and uniondensity.30A potential concern with the accuracy ofthe estimation is that the variables related tolabor globalization—trade prices, offshoring,and immigration—may be endogenous. Tradeprices are unlikely to be exogenous for countrieswhose economic size is sufficiently large.31Reverse causality or common third factors maybias the effects of offshoring and immigrationon the labor share. To address this concern, aninstrumental variables estimation was used withvariables reflecting domestic and foreign supplyand demand conditions, as well as lags of thepotentially endogenous variables as instruments.Specifically, the list of instruments included theshare of government consumption in GDP; theconsumption tax rate; the (log of) total population;the (log of) export-weighted real GDP oftrading partners; the distance-weighted exportadjustedemployment in the rest of the world (ameasure of the global labor supply); and lags of(logs of) relative trade prices, offshoring, andimmigration.In addition to the aggregate labor shareequation, a system of labor share equations forskilled and unskilled workers was also estimatedas follows:30A specification including interaction terms betweenthese policy variables and measures of labor globalizationand technological progress was also explored.31The price of absorption could also be affected bychanges in the labor share, which reflect changes in unitlabor costs.pk LRS = b k S +k=Σ E,MgkS ln––– +k=Σ S,UbSk ln––– + φSXXpA K+ φ LM KICT KICT SM ––– + φSC –––– + φSC2(––––)2+ eL K K SR U = b pk Lk U +k=Σ E,MgkU ln––– +k=Σ S,UbkU ln––– + φUXXpA K+ φ LM KICT KICT UM ––– + φUC –––– + φUC(––––)2+ eU,L K Kwhere S and U denote skilled and unskilled,respectively, and the other variables are thesame as above. A symmetry restriction postulatesthat the coefficients on the (log of)labor-capital ratio of the unskilled in the firstequation and the skilled in the second equationare the same. This system was augmented toinclude country fixed effects and the measuresof labor market policies, and was estimatedby iterated three-stage least squares using theinstruments listed above.Estimation ResultsThe estimation results from the aggregatelabor share equation are shown in Table 5.2.Most of the variables are statistically significantand have expected signs:• Higher relative export prices and lower relativeimport prices are associated with thelower labor share. This is consistent withadvanced economies’ exports being relativelycapital intensive and their imports, whichincreasingly come from developing countries,being relatively labor intensive.• Offshoring and immigration are negativelyrelated to the labor share, consistent with therising global labor supply exerting a negativeeffect on domestic labor demand. The coefficientson these variables in the instrumentalvariables regression are somewhat larger inabsolute value, suggesting the presence ofreverse causality: a lower labor share, whichreflects lower unit labor costs, makes offshoringless appealing for domestic firms andmakes immigration less attractive for foreignworkers.• Technological progress appears to have a nonlineareffect on the labor share, consistent withthe idea that labor-saving innovations initiallycreate the need for extra learning on the partof workers, but enhance their productivitylater on as the necessary skills are acquired.• Among the policy variables, only higher taxwedges and unemployment benefit replacementrates are associated with a lower laborshare, reflecting labor market rigiditiesstemmingfrom these policies.32 A nonlinearspecification including interaction terms withlabor globalization and technological progressvariables suggested, in addition, that employmentprotection legislation tends to increasethe effects of these variables on the laborshare.32Other labor and product market variables, specifically,the index of employment protection legislation, theindex of product market regulation, and the union densitymeasure, were not statistically significant and were,therefore, excluded from the final specification.Table 5.2. I mpact of Labor Globalization andTechnological Change on Labor SharesFixed EffectsEstimation(excluding Fixed InstrumentalDependent Variable: labor market Effects VariablesLabor Share policies) Estimation EstimationRelative export price (log of) –0.117*** –0.113*** –0.165***Relative import price (log of) 0.076** 0.087*** 0.138***Labor-capital ratio (log of) 0.055** 0.015 –0.025Offshoring –0.196* –0.156* –0.285***Immigration –0.627*** –0.553*** –0.746***ICT capital –2.871*** –2.643*** –3.517***ICT capital squared 56.407*** 44.962*** 55.598***Tax wedge . . . –0.002* –0.002***Unemployment benefits . . . –0.001*** –0.001***Fixed effects Yes Yes YesObservations 231 225 208R-squared 0.61 0.62Anderson test . . . . . . 151.63***Hansen test . . . . . . 6.61Source: IMF staff calculations.Note: * denotes statistical significance at the 10 percent level; **denotes statistical significance at the 5 percent level; and *** denotesstatistical significance at the 1 percent level. Standard errors areheteroscedasticity and autocorrelation robust. ICT = information andcommunications technology.Appendix 5.1. Data Sources and Methods187CHA PTER 5 The Globalization of Labor188The findings are generally robust to theexclusion of outliers (identified in terms oftheir influence on predicted values and thevariance-covariance matrix of the estimates) andof individual countries.33 They are also robustto splitting the import price into that of oiland non-oil imports (while the oil price has astatistically significant effect on the labor share,it is small in magnitude). The coefficients onthe ICT capital stock, its square, and offshoringbecome statistically insignificant when timeeffects are included. This is not surprising sincetime effects are often used in empirical studiesto capture the effect of worldwide technologicalprogress and other broad global trends. Thetime effects show a declining pattern over time,consistent with the negative effect of the growth33Partial correlation plots, showing the correlationbetween the labor share and each regressor after controllingfor the other explanatory variables, confirm that theestimated relationships are quite robust.in the ICT capital stock and offshoring on thelabor share. The coefficients on the share of ICTcapital are more robust to the inclusion of timeeffects when measured as a share of investment,rather than of capital stock. Similarly, the coefficienton offshoring of skilled inputs is morerobust to the inclusion of time effects than thaton total offshoring.The estimation results from the labor shareequations for skilled and unskilled sectorsare shown in Table 5.3. The first two columnscontain independent fixed effects estimationsof the two equations, the middle two columnspresent independent instrumental variablesestimations, and the last two columns showthe system estimation with the cross-equationrestriction imposed. Labor globalization andtechnological progress appear to have somewhatdifferent effects on the labor shares ofworkers in skilled and unskilled sectors. Laborglobalization has a somewhat stronger effecton the skilled sectors, in line, for example,Table 5.3. I mpact of Labor Globalization and Technological Change on Skilled and Unskilled Labor SharesInstrumental Three-Stage___F_ix_e_d_ _E_ff_e_c_ts_ _E_s_ti_m_a_t_io_n_ __ ____V_a_r_ia_b_l_e_s_ E_s_t_im__a_ti_o_n_ ___ ___L_e_a_s_t_ S_q_u_a_r_e_ E_s_t_im__a_ti_o_n___Skilled labor Unskilled Skilled labor Unskilled Skilled labor UnskilledDependent Variable share labor share share labor share share labor shareLog of:Relative export price –0.072*** –0.049*** –0.117*** –0.060*** –0.115*** –0.058***Relative import price 0.053*** 0.031** 0.089*** 0.041*** 0.097*** 0.044***Skilled labor-capital ratio 0.093** –0.203*** 0.075** –0.210*** 0.156*** –0.163***Unskilled labor-capital ratio –0.089*** 0.181*** –0.098*** 0.177*** –0.163*** 0.143***Offshoring –0.134* –0.016 –0.203*** –0.052 –0.191*** –0.043Immigration –0.507*** –0.162** –0.678*** –0.225** –0.663*** –0.216***ICT capital –0.808 –0.922* –1.413* –1.099** –2.046*** –1.409***ICT capital squared 22.358* 10.458 29.792*** 13.346* 38.688*** 17.860***Tax wedge –0.001 –0.001** –0.002*** –0.001*** –0.002*** –0.001***Unemployment benefits –0.001** –0.000* –0.001*** –0.000** –0.001*** –0.000***Product market regulation 0.000 0.002 0.001 0.002** 0.000 0.002**Fixed effects Yes Yes Yes Yes Yes YesObservations 219 219 202 202 202 202R-squared 0.53 0.94 . . . . . . . . . . . .Anderson test . . . . . . 140.83*** 140.83*** . . . . . .Hansen test . . . . . . 7.7 8.4 . . . . . .Source: IMF staff calculations.Note: * denotes statistical significance at the 10 percent level; ** denotes statistical significance at the 5 percent level; and *** denotes statisticalsignificance at the 1 percent level. Standard errors are heteroscedasticity and autocorrelation robust. ICT = information and communicationstechnology.with more offshoring occurring in the skilledsectors. Technological change affects both skillgroups negatively, but the effect is less strongfor the skilled, consistent with the nonlinearitydue to learning requirements, as suggestedabove. These results should be treated assomewhat more tentative, however, given thatthe classification by skill is based on broad economicsectors.The contributions of the various factors to thechange in the labor shares shown in the maintext are calculated as the average annual changein the respective variable multiplied by the correspondingcoefficient estimate. The averagesacross country groups are weighted by the numberof years of data available for each country,so that countries with more data receive a largerweight in these averages. These contributionsallow introducing cross-country differences inthe role of various factors, although they do notfully reflect cross-country heterogeneity, sincethe estimated coefficients are the same for allcountries in the sample.Elasticity CalculationsThe econometric model used in this chapterallows going beyond the effects of various factorson the labor share by computing the elasticitiesof labor compensation per worker and employmentto these factors (Kohli, 1991).The elasticities of labor compensation perworker to trade prices (given employment) areobtained as follows:e ( W , p i) = – g–iL– + Si ,RLwhere i = E, M, and the output shares RL and Siare evaluated at the mean for each country.The employment elasticities with respect totrade prices (given labor compensation perworker) are obtained as follows:e(W,pi)e(L,pi) = – ––––––,e(W,L)bwhere e(W,L) = ––LL– + RL – 1 and i = E, M.RLCombining these elasticities gives the elasticityof the total labor compensation to trade prices:e(WL,pi) = e(W,pi) + e(L,pi),where i = E, M. It is important to point out thatthese elasticities are derived from the modelthat assumes fixed prices, hence possible priceadjustments are not taken into account in thesecalculations.34To compute the actual percent changes in thetotal labor compensation resulting from changesin trade prices, these elasticities are multipliedby the average percent changes in relative tradeprices in each country. The averages acrosscountry groups are weighted by the number ofyears of data available for each country, so thatcountries with more data receive a larger weightin these averages. The results are shown in themain text.ReferencesAmiti, Mary, and Shang-Jin Wei, 2005, “Fear of ServiceOutsourcing: Is It Justified?” Economic Policy, Vol. 20(April), pp. 308–47.———, 2006, “Service Offshoring and Productivity:Evidence from the United States,” NBER WorkingPaper No. 11926 (Cambridge, Massachusetts:National Bureau of Economic Research).Annett, Anthony, 2006, “Lessons from SuccessfulLabor Market Reformers in Europe,” in Euro AreaPolicies: Selected Issues, IMF Country Report No.06/288 (Washington: International MonetaryFund).Aydemir, Abdurrahman, and George J. Borjas, 2006,“A Comparative Analysis of the Labor MarketImpact of International Migration: Canada, Mexico,and the United States,” NBER Working Paper No.12327 (Cambridge, Massachusetts: National Bureauof Economic Research).Barro, Robert J., and Jong-Wha Lee, 2000, “InternationalData on Educational Attainment: Updates34Further, the elasticities of the total labor compensationto import and export prices might be somewhatoverestimated since the elasticity of the wage (employment)assumes constant employment (wage). However,the net effect of trade prices on total labor compensationshould not be affected much.References189CHA PTER 5 The Globalization of Labor190and Implications,” CID Working Paper No. 42(Cambridge, Massachusetts: Center for InternationalDevelopment).Bassanini, Andrea, and Romain Duval, 2006, “EmploymentPatterns in OECD Countries: Reassessing theRole of Policies and Institutions,” OECD EconomicsDepartment Working Paper No. 486 (Paris:Organization for Economic Cooperation andDevelopment).Berg, Andrew, and Anne O. Krueger, 2003, “Trade,Growth, and Poverty: A Selective Survey,” IMFWorking Paper 03/30 (Washington: InternationalMonetary Fund).Blanchard, Olivier, 1998, “Revisiting European Unemployment:Unemployment, Capital Accumulationand Factor Prices,” NBER Working Paper No. 6566(Cambridge, Massachusetts: National Bureau ofEconomic Research).Blinder, Alan S., 2006, “Offshoring: The NextIndustrial Revolution?” Foreign Affairs, Vol. 85(March/April).Borjas, George J., 1994, “The Economics of Immigration,”Journal of Economic Literature, Vol. 32 (December),pp. 1667–717.———, 1995, “The Economic Benefits from Immigration,”Journal of Economic Perspectives, Vol. 9 (Spring),pp. 3–22.Cardarelli, Roberto, and Kenichi Ueda, 2004, “Domesticand Global Perspectives of Migration to theUnited States,” in United States: Selected Issues, IMFCountry Report No. 04/228 (Washington: InternationalMonetary Fund).Christensen, Laurits R., Dale W. Jorgenson, and LawrenceJ. Lau, 1975, “Transcendental LogarithmicUtility Functions,” American Economic Review, Vol. 65(June), pp. 367–83.Conway, Paul, Donato de Rosa, Giuseppe Nicoletti,and Faye Steiner, 2006, “Regulation, Competitionand Productivity Convergence,” OECD EconomicsDepartment Working Paper No. 509 (Paris:Organization for Economic Cooperation andDevelopment).Davis, Donald R., and Prachi Mishra, 2007, “Stolper-Samuelson Is Dead, and Other Crimes of BothTheory and Data,” in Globalization and Poverty,ed. by Ann Harrison (Chicago: University ofChicago Press and National Bureau of EconomicResearch).Docquier, Frédéric, and Abdeslam Marfouk, 2005,“International Migration by Educational Attainment,1990–2000,” in International Migration, Remittancesand the Brain Drain, ed. by Ç. Özden and M.Schiff (Washington: World Bank).Dutta, Puja Vasudeva, 2004, “Trade Protection andInter-Industry Wages in India,” PRUS WorkingPaper No. 27 (Sussex: Poverty Research Unit,Department of Economics, University of Sussex).Edmonds, Eric, Nina Pavcnik, and Petia Topalova,2007, “Trade Adjustment and Human CapitalInvestments: Evidence from Indian Tariff Reform,”NBER Working Paper No. 12884 (Cambridge,Massachusetts: National Bureau of EconomicResearch).Fajnzylber, Pablo, and Daniel Lederman, 1999,“Economic Reforms and Total Factor ProductivityGrowth in Latin America and the Caribbean (1950–95)—An Empirical Note,” Policy Research WorkingPaper No. 2114 (Washington: World Bank).Feenstra, Robert C., 2004, Advanced International Trade:Theory and Evidence (Princeton, New Jersey: PrincetonUniversity Press).———, 2007, “Globalization and Its Impact on Labor,”Global Economy Lecture 2007, Vienna Institute forInternational Economic Studies, Vienna, February8. Available via the Internet: http://www.econ.ucdavis.edu/faculty/fzfeens/papers.html.Feliciano, Zadia M., 2001, “Workers and TradeLiberalization:The Impact of Trade Reforms inMexico on Wages and Employment,” Industrialand Labor Relations Review, Vol. 55 (October),pp. 95–115.Freeman, Richard B., 1995, “Are Your Wages Setin Beijing?” Journal of Economic Perspectives, Vol. 9(Summer), pp. 15–32.———, 2006, “Labor Market Imbalances: Shortages,or Surpluses, or Fish Stories?” paper presented atthe Boston Federal Reserve Economic Conference,“Global Imbalances—As Giants Evolve,” Chatham,Massachusetts, June 14–16.Goh, Chor-Ching, and Beata S. Javorcik, 2007, “TradeProtection and Industry Wage Structure in Poland,”in Globalization and Poverty, ed. by Ann Harrison(Chicago: University of Chicago Press and theNational Bureau of Economic Research).Goldberg, Pinelopi, and Nina Pavcnik, 2005, “Trade,Wages, and the Political Economy of TradeProtection: Evidence from the Colombian TradeReforms,” Journal of International Economics, Vol. 66(May), pp. 75–105.———, forthcoming, “Distributional Effects ofGlobalization in Developing Countries,” Journal ofEconomic Literature.Gollin, Douglas, 2002, “Getting Income SharesRight,” Journal of Political Economy, Vol. 10 (April),pp. 458–74.Greenwood, Michael J., Gary Hunt, and Ulrich Kohli,1996, “The Short-Run and Long-Run Factor-MarketConsequences of Immigration to the UnitedStates,” Journal of Regional Science, Vol. 36 (February),pp. 43–66.Grossman, Gene M., and Esteban Rossi-Hansberg,2006, “The Rise of Offshoring: It’s Not Winefor Cloth Anymore,” paper presented at FederalReserve Bank of Kansas City symposium, “The NewEconomic Geography: Effects and Policy Implications,”Jackson Hole, Wyoming, August 24–26.Available via the Internet: http://www.kc.frb.org/PUBLICAT/SYMPOS/2006/PDF/Grossmanand-Rossi-Hansberg.paper.0831.pdf.Gupta, Sanjeev, Catherine Pattillo, and Smita Wagh,2007, “Impact of Remittances on Poverty andFinancial Development in Sub-Saharan Africa,” IMFWorking Paper 07/38 (Washington: InternationalMonetary Fund).Guscina, Anastasia, 2006, “Effects of Globalization onLabor’s Share in National Income,” IMF WorkingPaper 06/294 (Washington: International MonetaryFund).Hanson, Gordon, forthcoming, “Emigration, LaborSupply, and Earnings in Mexico,” in MexicanImmigration to the United States, ed. by George Borjas(Chicago: University of Chicago Press and NationalBureau of Economic Research).———, 2007, “Globalization, Labor Income, andPovertyin Mexico,” in Globalization and Poverty,ed. by Ann Harrison (Chicago: University ofChicago Press and National Bureau of EconomicResearch).Harrigan, James, 1998, “International Trade andAmerican Wages in General Equilibrium, 1967–1995,” Staff Report No. 46 (New York: FederalReserve Bank of New York).———, and Rita A. Balaban, 1999, “U.S. Wages inGeneral Equilibrium: The Effects of Prices, Technology,and Factor Supplies, 1963–1991,” NBERWorking Paper No. 6981 (Cambridge, Massachusetts:National Bureau of Economic Research).Harrison, Anne E., 2002, “Has Globalization ErodedLabor’s Share? Some Cross-Country Evidence”(unpublished; Berkeley, California: University ofCalifornia at Berkeley).International Labor Office and World Trade Organization,2007, Trade and Employment: Challengesfor Policy Research (Geneva: International LaborOrganization).Jean, Sébastien, and Giuseppe Nicoletti, 2002, “ProductMarket Regulation and Wage Premia in Europeand North America: An Empirical Investigation,”OECD Economics Department Working PaperNo. 318 (Paris: Organization for Economic Cooperationand Development).Jensen, J. Bradford, and Lori G. Kletzer, 2005, “TradableServices: Understanding the Scope and Impactof Services Outsourcing,” IIE Working PaperNo. 05-9 (Washington: Institute for InternationalEconomics).Jorgenson, Dale W., and Khuong Vu, 2005, “InformationTechnology and the World Economy,” ScandinavianJournal of Economics, Vol. 107 (December),pp. 631–50.Katz, Larry F., and David H. Autor, 1999, “Changesin the Wage Structure and Earnings Inequality,”in Handbook of Labor Economics, Vol. 3A, ed. byOrley C. Ashenfelter and David Card (Amsterdam:North-Holland).Kohli, Ulrich, 1991, Technology, Duality, and ForeignTrade: The GNP Function Approach to Modeling Importsand Exports (Ann Arbor: University of MichiganPress).Kumar, Utsav, and Prachi Mishra, forthcoming, “TradeLiberalization and Wage Inequality: Evidence fromIndia,” Review of Development Economics.Lewer, Joshua J., and Hendrik Van den Berg, 2003,“How Large Is International Trade’s Effect on EconomicGrowth?” Journal of Economic Surveys, Vol. 17(July), pp. 363–96.Mishra, Prachi, 2006, “Emigration and Brain-Drain:Evidence from the Caribbean,” IMF Working Paper06/25 (Washington: International Monetary Fund).———, 2007, “Emigration and Wages in Source Countries:Evidence from Mexico,” Journal of DevelopmentEconomics, Vol. 82 (January), pp. 180–99.Nehru, Vikram, and Ashok Dhareshwar, 1993, “A NewDatabase on Physical Capital Stock: Sources, Methodology,and Results,” Revista de Análisis Económico,Vol. 8, No. 1, pp. 37–59.Olsen, Karsten Bjerring, 2006, “Productivity Impactsof Offshoring and Outsourcing: A Review,” OECDScience, Technology and Industry Working PaperNo. 2006/1 (Paris: Organization for EconomicCooperation and Development).Organization for Economic Cooperation and Development(OECD), 2004, OECD Employment Outlook(Paris).References191CHA PTER 5 The Globalization of Labor192———, 2005, “Trade-Adjustment Costs in OECDLabour Markets: A Mountain or a Molehill?” inOECD Employment Outlook (Paris).———, 2006a, “The Internationaliation of Production,International Outsourcing and OECD LabourMarkets” (Paris).———, 2006b, “Migration in OECD Countries: LabourMarket Impact and Integration Issues” (Paris).Pavcnik, Nina, Andreas Blom, Pinelopi Goldberg,and Norbert Schady, 2004, “Trade Liberalizationand Industry Wage Structure: Evidence fromBrazil,” World Bank Economic Review, Vol. 18, No. 3,pp. 319–44.Porto, Guido, 2006, “Using Survey Data to Assessthe Distributional Effects of Trade Policy,” Journalof International Economics, Vol. 70 (September),pp. 140–60.Rodrik, Dani, 2006, “What’s So Special About China’sExports?” NBER Working Paper No. 1947 (Cambridge,Massachusetts: National Bureau of EconomicResearch).Slaughter, Matthew J., 2001, “International Trade andLabor-Demand Elasticities,” Journal of InternationalEconomics, Vol. 54 (June), pp. 27–56.Topalova, Petia, 2005, “Factor Immobility andRegional Impacts of Trade Liberalization:Evidence on Poverty and Inequality from India”(unpublished; Cambridge, Massachusetts: MassachusettsInstitute of Technology). Available viathe Internet: http://www.isid.ac.in/~planning/Topalova.pdf.Young, Alwyn, 1995, “The Tyranny of Numbers: Confrontingthe Statistical Realities of the East AsianGrowth Experience,” Quarterly Journal of Economics,Vol. 110 (August), pp. 641–80.———, 2003, “Gold into Base Metals: ProductivityGrowth in the People’s Republic of China Duringthe Reform Period,” Journal of Political Economy,Vol. 1 (December), pp. 1220–61.World Bank, 2006, Global Economic Prospects2007: Managingthe Next Wave of Globalization(Washington).

planetkdoublehs

Thursday, 27 March 2008

Shaking, Waking, and Which Dreamed It?

Tuesday, 25 March 2008

A smily face from a friend of mine--

Monday, 24 March 2008

Easter Afterwards--

Microsoft Office Live Help

Wednesday, 19 March 2008

United Colors of Benetton Afterwards...

Monday, 17 March 2008

Labor Globalization Afterwards

Blog Archive

About Me