global integraton technology transfer pptx

vi contents 6 The Knowledge Content of Machines: North-South Trade and Technology Giorgio Barba Navaretti, Maurice Schiff, ‘and Isidro Soloaga 7 Exports and Economic Performance: Evid

GLOBAL INTEGRATION

AND TECHNOLOGY

TRANSFER

GLOBAL INTEGRATION

AND TECHNOLOGY

TRANSFER

Edited by Bernard Hoekman and

Beata Smarzynska Javorcik

A copublication of Palgrave Macmillan

and the World Bank

“Telephone: 202-473-1000

En oackawebaang

Alii ore

Acoplcton of The Wd Rk Palgrave Mei,

Totnes ISHS cose Now Yer api apt B21 Sand 110

‘Mac ns oie rae nth id Stas Unit Kinglo hers Pre?

ve can Em the bape Union tie cont

ihe tact Dowco Theil Benker gvcamot ky rgteei TU)

{esominatons nthe nôn we on ay map nhs work any ide on hepa

‘he Well kcncning ep tey rtrd cbdonement or aapane ach our

Terai tpn Cyngor tanning oro aff ink ih

Fr permion Gig cans nt tn De 433 on to poco ay par f his work pln hd gue with comple ior

CONTENTS

Acknowledgments

Contributors

Abbreviations and Acronyms

1 Lessonsfrom Empirical Research on International

Technology Diffusion through Trade and Foreign

Bernard Hockman and Beata Smarzynska Javorcik

Part! LITERATURE SURVEYS

2 Econometric versus Case Study Approaches

Howard Pack

3 Foreign Direct Investment, Linkages,

Part ll FOREIGN TRADE AND PRODUCTIVITY

5 On the Quantity and Quality of Knowledge:

‘The Impact of Openness and Foreign R&D

‘on North-North and North-South Technology

‘Maurice Schiff and Yanling Wang

vi contents

6 The Knowledge Content of Machines:

North-South Trade and Technology

Giorgio Barba Navaretti, Maurice Schiff,

‘and Isidro Soloaga

7 Exports and Economic Performance:

Evidence from a Panel of Chinese Enterprises

Aart Kraay

Part Ill FOREIGN DIRECT INVESTMENT, TECHNOLOGY

TRANSFER, AND PRODUCTIVITY

8 Foreign Investment and Productivity Growth

in Czech Enterprises

‘Simeon Djankov and Bernard Hoekman

9 Technological Leadership and the Choice

of Entry Mode by Foreign Investors

Beata Smarzynska Javorcik

10 Does Foreign Direct Investment Increase the

Productivity of Domestic Firms? In Search

of Spillovers through Backward Linkages

Beata Smarzynska Javorcik

11 Product Quality, Productive Efficiency,

and International Technology Diffusion:

Evidence from Plant-Level Panel Data

‘Aart Kraay, Isidro Soloaga, and James R Tybout

12 Market Discipline and Corporate Efficiency:

Evidence from Bulgari

Simeon Djankoy and Bernard Hoekman

13 Innovation in Mexico: NAFTA Is Not Enough

Daniel Lederman and William F Maloney

Index

Figures

1 Perceived Effects of FDI inthe Czech Republic and Latvia

G1 Average Unit Values in Select Countries, 1989-97

6.1 Persstency of the Technology Gap with One-Year Lag

60.2 Petsistency ofthe Technology Gap with Two-Year Lag

6D Persistency ofthe Technology Gap with Seven-Year Lag,

SA Labor Productivity of Czech Firms with and without

“Total Factor Productivity Growth in Czech Firms

with and without Foreign Partners, 1992-96

“Training and New Technology in Czech Firms

‘with and without Foreign Partners

R&D Intensity and Probability of a foint Venture: Case |

RSD Intensity and Probability ofa Joint Venture: Case 2

RSD Intensity and Probability ofa Joint Venture: Case 3

Net FDI Inflows to Lithuania, 1993-2000

(Change in Horizontal Measure, 1996-2000

‘Change in Backward Measure, 1996-2000

(Change in Forward Variable, 1996-2000

Grovith Rates of Total Factor Productivity

in Selected Countries and Regions, 1960-99

Patents per Million Workers in Selected

Regions, 1960-2000,

[Number of Scientific Publications and Patents in Mexico

Relative to Comparative Countries, 1960-2000

Ratio of R&D to GDP in Selected Countries

Innovation Inputs in Mexico

Efficiency of Research and Development

in Selected Countries, 1985-2000

IRCA Index in Selected Industries

and Countries, 1980-2000

Private Sector Perceptions of Quality of Scientific

Institutions and Univers

‘Technological Capabilities in Manufacturing Industry

Determinants of Total Fctor Productivity

Matching between Machines and Products

Description and Source of Data for Variables

Descriptive Statistics

Summary Statistics for Enterprise Sample

Distribution of Sample by Ownership and Sector, 1990

Summary Statistics on Exporters, 988-92

Summary Statistics on Exporters by Ownership

and Sector, 1990

168

HH

Bi 18s

316 a2 a3

328

36 1Ú

108 1s I6

Size of Exporters andl Nonexporters, 1988-92

Petsistence and Volatility of Export Status

Performance of Exporters and Nonexporters

Basic Model Results

Export Histories

Basic Model Controlling for Export Histories

Descriptive Statistics ofthe Sample

Revenue Shares of Inputs, Mark-Up, and Scale

Estimates by Sector

Panel Regression Estimates

Spillover Effects on Firms without Foreign Linkages

Spillover Effects on Firms without FDL

Entry Modes Chosen by Investors in the Sample, by Country

Entry Modes Chosen by Investors inthe Sample, by Industry

Probit Model with Firm- and Industry-Level

RSD and Advertising intensities

Classification of Industries by Technology Level

R&D Intensity of FDI Projects in Three-Digit SIC Industries

Probit Model with Relative R&D and Advertising Intensities

[Bivariate Probit with Sample Selection

Marginal Effects of Bivariate Probit with Sample

Selection: Entry Mode Equation

FDI Inflows to Central and Eastern European

Countries, 1993-2000

Distribution of Firms with Foreign Capital,

by Industey, 2000

Summary Statistics

‘Additional Summary Statistics for Spillover Variables

(OLS with Lagged and Contemporancous

Spillover Variables

‘Comparison of Coefficients from OL

and Olley-Pakes Regressions

Results of OLS and Olley-Pakes Regressions in First,

Second, al Fourth Differences

Shate of Foreign Ownership and Productivity Spillovers

‘Concentration of Downstream Sectors

and Productivity Spillovers

Colombian Chemical Industries: Dynamic Model

Moroccan Chemical Industries: Dynamic Model

Mexican Chemical Industries: Dynamic Model

Colombian Chemical Industries: Static Model

Moroccan Chemical Industries: Static Model

Mexican Chemical Industries: Static Madel

145 M5 M7

Activity Dynamics inthe Colombian Chemicals

Industry: Descriptive Evidence

Activity Dynamics in the Moroccan Chemicals

Industry: Descriptive Evidence

Activity Dynamics in the Colombian Chemicals

Industry: Econometric Results

Activity Dynamics in the Moraccan Chemicals

Industry: Econometric Results

Performance Statistis, 1991-96

Descriptive Statistics

Pearson Correlation Table

Estimation Results

Estimation Results: Random-Eifects Model

Determinants of Patent Counts

Structure of R&D Effort in Selected

Countries, 1995-2000

a 3m

Among the conclusions emerging from the research contained inthis wolume, 8

‘number stand out First, the evidence suggests that an open trade regime facilitates the difusion of knowledge, Undistorted access to capital equipment and imported inputs that embody foreign knowledge allow firms to acquire know-how; the _greater competition from imports lowers the mark-ups over costs that firms charge customers, At the sige time, given that technology markets are associated with increasing returns, imperfect competition, and externalities, the argument against trade protection is not unconditional The conclusions hinge on the scope of knowledge spillovers International spillovers, for which there is considerable evi dence, strongly til the balance in favor offre trade ff national spillovers are also important, there may be a potential role for intervention Trade policy, however is

xi

‘ver that policies that discriminate between types of FDI are unlikely to promote technology transfer By atemptng to force multinational enterprises to license their technologies or engage in joint ventures, hoạt countries may lower the quality of teh: nologies they receive and reduce incentives for foreign firms to invest at al

‘While the magnitude of international technology diffusion undertaken by

‘multinational enterprises need not be socially optima, evidence presented in this, volume reveals tha such fiems are keen to transfer technology to ther local suppli-

ts Policies that increase incentives to source locally, as opposed to regulation or legislation requiring that multinational enterprises engage in international tech-

‘nology diffusion to local competitors have a greater likelihood of being successful [Examples of auch a policy are supplier development programs that aims to prepare local companies to understand and meet the needs of multinational enterprises

“The services provided under such programs can be effective in asisting firms, pro- Vided they are well designed, mobilize the right type of skills, and ensure that their target audience is aware ofthe services on oer,

‘Many countries actively seek to attract foreign investors through up-front subsi- dies tax holidays, and other grants A rationale for such investment incentives may

be based on positive externalities generated by inflows of FDI Local suppliers may benefit not just through expanded sales but through access to technologies pro- vided by the investors Such postive externalities may be enhanced by the preva- lence of “follow the leader” behavior among multinational enterprises Given the oligopolistc nature of markets within which FDI occurs, new entrant may esult

in addtional investments by both competitors and upstream suppliers of inputs, components, and services An implication is that a host country may be able to

‘unleash a sequence of investments by suecessfully inducing FDI from one or two naj firm

T the local economy lacks a well-developed network of potential suppliers,

‘multinational enterprises may be hesitant to invest, and local suppliers may not develop because of lack of demand In the presence of such interdependence, growth may be constrained by a coordination problem that can partially be resolved by initiating investments from key firms, Such coordination problems

‘aninot be tackled solely through investment incentives, Policy efforts need to focus primarily on improving the investment climate and reducing the costs of absorbing

Foreword tlt

technology, a complex task that involves building human capital and expanding national innovation systems Thus while there may be a case for investment ince: tives, it isa conditional one To be efectiv, the investment climate and absorptive capacity must meet certain conditions Moreover, given competition among coun: tries in atracting FDI through incentive packages, policymakers must carefully

‘examine the magnitude of potential costs and benefits asaciated with such police

Frangois Bourguignon

(Chief Economist and Senior Vice-President

The World Bank

ACKNOWLEDGMENTS

“The chapters in this volume are the product ofa research project undertaken by the World Bank Research Department, The research was partially supported by _grant POS2764 from the Bank’s research support budget, The financial support provided by the UK Department for International Development is also gratefully acknowledged, Some of the chapters have been published in professional journals,

‘others are published here for the fist time

‘he project oves much to the substantial inputs of Professor Jim Tybont (Pennsylvania State University) inthe early phases, Aart Kraay helped manage the projest research, working with Professor Tybout and Bernard Hoekman, The

‘editors are grateful to both of them, as well as to Isidro Soloaga for his tireless

‘effort to put together the datasets used inthe chapters The editors thank Stephen, MeGroarty, Santiago Pombo, Janice Tuten, and Nora Leah Ridolf for their assis- tance in geting the manuscript through the production proces in a timely manne

‘Thanks are also due to Barbara Kaeni for excellent copyediting and to Maribel Flewitt and Rebecca Martin for administrative assistance

All views expressed in this volume are those of the contributors They do not necessarily represent the views of the World Bank, its executive directors, or the countries they represent

Aart Kraay, World Bank, Washington, DC

Daniel Lederman, World Bank, Washington, DC

William Maloney, World Bank, Washington, DC

Giorgio Barba Navaretti, Universiti degli Studi di Milano and Centro Studi Luca

@Agliano, Milan

Howard Pack, Wharton School, University of Pennsylvania, Philadelphia

Kamal Saggi, Southern Methodist University, Dallas Texas

Maurice Schiff, World Bank, Weshington, DC, and Institute for the Study of Labor (IZA), Bonn

Isidro Soloaga, Universidad de las Américas, Puebla, Mexico

computable general equilibrium Council of Mutual Economie Assistance European Bank for Reconstruction and Development European Union

foreign direct investment _gross national product innovative revealed comparative advantage International Standard Industrial Clasiication joint venture

Nomenclature générale des activités économiques dans les Com munautés européennes

North American Free Trade Agreement National Statistical Office (of Bulgaria) Organisation for Economic Co-operation and Development original equipment manufacturer

ordinary least squares outward processing trade price-cost margin

research and development total factor productivity

“Trade Related Investment Measures United Nations Industrial Development Organization United States Patent and Trademark Office

unit values index

xix

LESSONS FROM EMPIRICAL RESEARCH ON INTERNATIONAL

TECHNOLOGY DIFFUSION

THROUGH TRADE AND FOREIGN DIRECT INVESTMENT

Bernard Hoekman and Beata Smarzynska Javorcik

This volume examines international technology difusion Its chapters explore the channels through which existing knovledge is transferred across countries, the

‘magnitude of such transfers, and their impact

‘The acquisition and diffusion of knowledge or technology are of grat impor tance for economie development, asthe ad

and production processes is a key deter

ion of new techniques, machines,

f productivity growth Given that

is undertaken in high: rely largely on imported technologies as sources of new productive knowledge This nat say that na R&D

is undertaken in developing countries; considerable amount of follow-on innova tion and adaptation does accur there, contributing to the glabal tock of knowledge

"The determinants of the supply of new knowledge—innovation and invention— have been the subject of a great deal of research The chapters inthis volume ace

‘not concerned with the generation of new knowledge The faces is onthe diffi

of knowledge through trade and direct foreign investment

Improving our understanding of how technology diffuses aross borders is important in order to identify what governments ean do through po

ulate international technology diffusion and its absorption in theie

Numerous ed toa geeater or lesser extent by the ol

acquiring and augmenting or adapting knowledge that will enhance productivity

most research and development (R&D,

income countries, most developing economies m

2 Global integration and Technology Transfer

and economic growth, Such policies range from the general to the specific Examples {include the provision of public gods such as education and infrastructure, the esta lishment of specific funds for the creation of technology, the granting of subsidies for capital investment, action to protect intellectual property rights, design of the strc tute of import protection ta favor the import of equipment and machinery, the granting of tax incentives foe foreign investors, export promotion schemes, the etab- lishment of fece trade or special economic zones, and specialized traning programs

‘The “correct” policy intervention, if any, depend critically on the channels through which technology diffuses internationally and the quantitative effets of the various diffusion processes on allocative efficiency and productivity growth, [either i well understood, Most of the empirical evidence is based on aggregate data or cross-sectional surveys and is subject to multiple interpretations Case stualies, while informative about the case at hand, are difficult ta generalize Hence policymakers have often acted on the bass of theory, anecdotes, or instincts International technology diffusion can occur through numerous channels Trade in goods and services is one All trade beats some potential for transmitting technological information Imported capital goods and technological inputs can directly improve productivity by being used in production processes Alternatively, firms may learn about technologies by exporting to knowledgeable buyers, who share product designs and production techniques with them,

second channel is foreign direct investment (FDI) ofthe pursuit of project- specific joint ventures Multinational enterprises generally transfer technological information to their subsidiaries, directly affecting the productivity of these plants In addition, some of the associated knowledge may “leak” into the host economy In both cases—trade and FDI—technology may diffuse from firms that have acquired it internationally to other firms inthe same industry or region through demonstration effets, abor turnover, mutual input suppliers, or reverse engineering third channel of international technology difusion is direct trade in knowledge through technology purchases or licensing This may occur within firms, among joint ventures, or between unrelated firms Licensing and FDI are often substitutes, but they may also be complements Much ofthe recorded international payments and receipts for intellectual property (royalties) occur within

between parent fiems and affiliates Which form is preferable to technology owners depends on mary Factors, including the strength of intellectual property rights

Tn sum, new technologies can be transmitted internationally theough a variety

of activities They may be embodied in goods and transferred through imports of new varieties of differentiated products or capital goods and equipment, they may

be obtained through exposure to foreign buyers ot foreign investors, or they may bbe acquired through direct trade in "disembodied knowledge”—through contracts supported by polices that protect intellectual property

rms, 38 flows

Lessons from Empirical Research on International Technology Diffusion 3 The chapters in this volume Focus om the frst to channels, trade and EDL!

‘They use cras-country and firm-level datasets to investigate the extent to which trade and FDI serve as channels of technology transfer The chapters rely ing on cros-county data explore the importance af trade openness, absorptive and innovative capacity and the identity of tading partners as determinants of international technology diffusion measured by total factor productivity (TEP) performance The focus i on spillovers theough rad, buling on and extending the original nsight of Coe and Helpman (1995)-In the case of analyses based on fire-level data, the information available allows individual manufacturers to be followed overtime in terms of their productivity and the types of activities they are engaged in (exporting, importing, joint ventures, and s0 on) Data on all potential channels f difusion are generally not available for any of the countries stied, but together the datasets cover many of the activities of interest from the perspective of international technology diffusion

“The fiem-Level analyses span several Cental and Eastern Eucopean transition economies, Chin, two Latin American countries (Colombia and Mexico) and one North African countey (Morocco) They areal quite ferent in terms of intial on ditions and endowments The Central and Eastern European counties are relatively well endowed with human capitl (killed labor) and physia infastrutue They also provide an unprecedented natural experiment of sudden exposure 1o interna- tional trade, investment, and technology In adtion, they are of intrest in terns of the wsethat was made of regional integration agreements, with the prospec of acces sion the European Union amor focal pont fr reform a East si, China has often been cited a a country rapidly absorbing ne technologies andl moving up the technology lider Chinas adoption ofthe “open door policy” in 1979 marked the beginning of a concerted peogram to acquire foreign technology inorder to raise {growth rates and living standards Over the cours ofthe next 25 years, Chinas spectacular growth was accompanied by its emergence asthe largest developing country ecipient of FDI, one of the word’ largest trading nations, and a major iewportr of capital goods The semi-industil countries of Latin America, where pening to trade i relatively recent aftr endownents are quite diferent (natal resource abundance), offer an alternative opportunity for analyzing technology ditfsion patterns given the availabilty of high-quality Sem-lvel dats, Morocco is simile to Latin America in terms of relatively high endowments of natural resources, but it difers in tems of endossments of unskilled labor and the use of

3 eee trade agreement withthe European Union to improve export market aces

“This introductory chapter sets the stage forthe individual analyses t reviews the majo findings of the chapters in the volume aswell as related reseaech—with

an emphasis on that undertaken bythe World Bank—that focuses on contributing tothe understanding of international technology diffusion Much ofthe chapter

4 Global Integration and Technology Transfer

‘Technology Diffusion through Trade

‘rade can contribute to international technology diffusion by providing local firms acess to new technologies embodied in imported machinery and equipment and by creating opportunities for reverse engineering of products developed abroad, It can also create incentives to adopt and improve technologies through exporting opportunities

‘Access to Technologies through Imports

Coe and Helpman (1995) and Coe, Helpman, andl Hoflimaister (1997) find that foreign knowledge (R&D) embodied in traded goods has a statistically significant positive impact on aggregate TFP of importing countries.’ Subsequent research finds that this impact is greater the more open a country i, the more skilled i its labor force, and, in the case of developing countries, the more trade takes place with industrial economies (Schiff and Wang forthcoming)

Knowledge accumulation means new knowledge (an increase in its quality), greater accesso existing knowledge (an increas in its quantity), orboth Chapter 5,

by Maurice Schiff and Yanling Wang, examines the relative contribution ofthese {wo components of knowledge to (otal TEP for North-North and North-South trade-related knowledge diffusion They proxy quantity by openness (as messured

by the trade to GDP ratio) and quality by the R&D content of trade (the type of products traded) The literature assumes equal contributions to TFP of openness and the R&D content of trade Schiff and Wang show that this assumption does not correspond to reality They find that R&D hasa greater impact on TEP than openness for North-North trade and that openness has greater impact on TFP than R&D for North-South trade These results imply that the impact of openness on TEP in {developing (developed) countries is larger (smaller) than previously obtained in this literature and that developing countries may obtain larger productivity gains from trade liberalization than previously thought

As variations in capital goods trace can better explain cross-country dillerences {in productivity than can overall trade (se, for example, Eaton and Kortum 1999),

Lessons from Empirical Reseach on International Technology Difusion 5

the distinction between quantity and quality is further examined inthe context of imported capital goods by Barba Navatet, Schiff, and Soloaga (chapter 6) Thelt chapter focuses on the impact of imported technologies on the productivity of countries in Central and Eastern Europe and the Southern Mediterranean Comparing the unit values of machinery imported by each countey with the unit, values of machinery imported by the United States, the authors calculate the gap between the technology purchased and the technological frontier (defined by the United States) They find that the gap is persistent and in some cases even increasing,

‘They conclude that the productivity growth in manufacturing depends on the types of machines imported (quality) and not on the share of imported equipment

in total investment (quantity)

‘These results suggest that open tade policies are critical for developing countries

in attracting technology Openness alone, however, is not sufficient—both absorptive capacity and the ability to adapt foreign technology need to be in place, both of which are related to human capital endowments and investment in RAD- intensive industries In developing countries, technology acquisition often amounts to adapting existing methods to local circumstances (Evenson and

‘Westphal 1995) This process takes time Gradual adoption of new techniques or

‘ne inputs is optimal for risk-averse producers in the face of costly adoption and

"uncertain returns Producers need to learn how to apply the new technology: they often begin by applying itt a small past of their output and, if profitable,

‘gradually increasing its application over time (Tybout 2000)

‘The farther the “technological distance” of a county from the global frontier (best practice), the more difficult it isto absorb information effectively into peo: duction systems (Keller 2002) As Kelle (1996) argues, accesso foreign technologies alone does not increase growth rates of developing countries Countries tend to acquire international technology more readily if domestic firms have local R&D programs, there are domestic private and public research laboratories and universities, and there exists a sound basis of technical skillsand human capital In chapter? ofthis volume, Howard Pack surveys and synthesizes much ofthe relevant literature on this issue, concluding that the existence of such “complementary inputs” reduces the costs of imitation, adaptation, and follow-on innovation

‘Access to Technologies through Exports

While it is quite plausible that repeated contacts with knowledgeable foreign buyers provide greater awareness of and access to foreign technologies th literate survey

by James Tybout (chapter 4) concludes thatthe jury ssi out on whether firms learn from exporting, This san important policy as wells analytical question, as many governments have export promotion programs that are atleast in part

{6 Global Integration and Technology Transfer

premised om the existence of such learning, as well as on the presumption that such new knowledge will spill over to other firms inthe local economy

Two studies—Clerides, Lach, and Tyhout (1998), using plant-level data from Columbia, Mexico, and Morocco, and Bernard and Jensen (1999), based on US

<data—analyze the causal links between exporting and productivity They observe that firms that are relatively efficent become exporters, while plants that cease exporting experience increasingly high costs, They find no evidence of learning from exporting and conclude that the well-documented positive association between exporting and greater efficiency i explained by the self-selection of more ficient firms into export markets

‘These results contrast with the conclusions of Aart Kraay (chapter 7)-Using data

‘0n 2,105 Chinese enterprises during the period 1988-92 and controlling for past, performance and unbserved firm characteristics, Kraay finds that past exports lead

to significant improvements in enterprise performance Interestingly, these learning effects are most pronounced among established exporters For new entrants t0 export markets, learning effects appear tobe insignificant or even negative

“The realty may be more nuanced than a simple dstinetion between self-selection {nto an earning fom exporting, In another World Bank study, Hallward-Driermayer, Iarossi, and Sokoloff (2002), using firm-level information from a sample of East, Asian countries, conclude that frms make deliberate decisions to raise productivity

in order to serve export markets Its not simply that more produstive fim sl selec nto exporting but that fms that explicily target export markets consistent _make diferent decision regarding investment training technology, ad the selection

of inputs, thereby raising their productivity Thus the “exporter selection” process is

‘not necessarily driven hy exogenous shocks, such as trade reforms, but reflects invest- men made by firms in anticipation of accessing foreign markets

Foreign Direct Investment as a Conduit

Several chapters in Roberts and Tyhout (199%) examine the potential impacts

‘of FDI, Subsequent rescach, including several chaptersin thie volume, has continu

to investigate the sign and magnitude ofthe effects of PDI, However, the major

Lessons from Empirical Reseach on International Technology Difusion 7

contribution of the more recent literature is to distinguish between horizontal spillover effects within an industey and vertical spillovers generated by linkages

in the production or value chain Research has focused on the incentives for

‘multinational enterprises to deliberately transfer technology to upstream local firms that are (potential) suppliers of intermediates that the multinational enterprise uses in production,

Implicit to the analysis of knowledge spillovers from EDI isthe assumption, that foreign ownership per se conveys some intangible advantage and that prox

mi to foreign firms or plants ean be beneficial to domestic firms, Until recently, there was relatively little robust empirical confirmation thatthe first part of this assumption holds, particularly inthe context of developing counties In chapter 8 Simeon Djankov and Bernard Hoekman use firm-level panel data from the C2ech: Republic to show that during 1992-6, EDI had a positive impact on the TEP

‘growth of recipient firms, correcting forthe fact that foreign companies may (and Aid) invest in fiems with above-average initial productivity More recent evidence

is provided by another output of the World Bank research program in this area mold and Javorcik (2005) use the Indonesian Census of Manufacturing to demonstrate that foreign acquisitions increased the productivity of the acquired Indonesian plans, To control forthe possibility that foreign investors may select plans that are more productive to begin wit, the study employs a matching tech: rique and the difference in differences approach The analysis indicate that benefits

of foreign ownership are realized very quickly and that acquired plants experience lange increases in productivity, investment outlays, employment, and output For

«ign ownership is also found to lead to greater integration of plants into the global economy through exports and imports

In studying FDI spillovers, itis important to distinguish between horizontal (Gntraindustry) and vertical (interindusty) spillovers In principle, vertical spillovers are mote likely to occu, insofar as multinational enterprises can be expected to take actions to prevent knowledge from leaking to their competitor in the same industry

‘They can do so by paying efficiency wages, enforcing intellectual property rights, and imposing noncompete clauses on key stall Vertical spillovers are moreliely to occu, Ihecause foreign affiliates have an incentive to reduce sourcing costs by encouraging productivity improvements in (potential local suppliers of inputs and services However the local upstream industries concerned may not always benefits foreign entey ina downstream sector may also give ris a greater imports of intermediates oF result in entry of foreign input suppliets Such “follow the client” FDI is quite common in sectors such as automobiles More generally it can be expected to be prevalent for services suppliers, for whom trade soften not technically possible Case studies suggest that substantial technology diffusion occurs due to FDI (Blomstrom and Kokko 1998) Econometric studies are more diverse Some find

8 Global Integration and Technology Transfer

that firms in sectors with a relatively high moltinational enterprise presence tend

to be mote productive (Kokko, Tansini, and Zajan 1997), while others conclude that domestically hel firms may actually do worse a the foreign presence in their industry increases (for example, Aitken and Harrison 1999) Negative spillover effects may occur in the short run if multinational enterprises siphon off domestic demand or bid away high-quality labor As Gérg and Strobl (2001) point out in a recent meta-analysis ofthe literature on intraindustry spillovers studies employing panel (rather than cross-sectional) data, which can and thus do control for unob- served firm characteristics (or unobserved industry characteristics inthe ease of esti

‘mations at the sectoral level), ae les likely to finda positive effect

(Chapter 8 falls into ths category It finds that FDI has a positive impact on TEP growth of recipient firms, coneolling for selection bis, but it ako concludes that joint ventures and foreign afiiates havea negative spillover effect on purely domes tic irms in the same industry lntecstingly, foreign affiliates alone do not havea sig nificant effect on the performance of domestic ims and joint ventures taken together,

‘which the authors attribute to the heter absorptive capacity of joint ventures

“The results of chapter 9, by Beata Javorcik, suggest that the findings of Djankow snd Hoekman in chapter 8 could also be due to differences in technological content and thus in the spillover potential of joint ventures and foreign subsidiaries Javorcik’s analysis is hased on data from 22 transition economies forthe early 1990s, She shows that foreign investors with the most sophisticated technologies (relative tothe industry mean) are more likely to engage in wholly owned projects

‘than in joint ventures These elfects are present in high- and! medium-technology sectors but not in low-R&D industries

In contrast to the mixed results on intraindustry (horizontal) spillovers from FDI, the evidence on interindustry (vertical) technology transfer from multinational enterprises has been consistently p

documented such transfers taking place through firms from industrial coun- tries buying the output of Asian firms to sell under their own brand names Mexico's maquiladona sector is an example of vertical international technology diffusion Most maguiladonas began as subsidiaries of U.S firms that shifted labor-intensive assembly operations to Mexico (Tan and Batra 1985) However, lover time the maguiladoras adopted more sophisticated production tech niques, many of which were imported from the United States (Sugai 2002), In chapter 10 Favorcik analyzes firm-level panel data feom Lithuania, She pro: duces evidence tha is consistent withthe hypothesis that postive productivity spillovers from FDI occur through contacts between foreign affiliates and local suppliers in upstream sectors Her results suggest that a one standard deviation increase in FDI in the sourcing (buying) sector is associated with a 15 percent

ive The literature has

Lessons from Empirical Research on International Technology Difusion 9

‘sein output of each domestic firm in the supplying industry Vertical spillovers

ae associated with projects with shared domestic and foreign ownership but not with filly owned foreign investments

‘Openness to Trade and FDI: Interactions

and Procompetitive Effects

“The studies described looked at each channel of technology transfer in isolation A more systematic approach requires the simultaneous consideration ofthe effets

of various international transactions on firm productivity This is the focus ofthe research undertaken by Aart Kraay Isidro Soloaga,and James Tybout in chapter LI, whieh uses plant-level panel data from Colombia, Mexico, and Morocco Theie Findings suggest that international activities (such as exporting, importing, and FDI) do not help much in predicting current fiem performance, once pas realiza- tions on quality and marginal costs are controlled for Inthe minority of cases, where significant associations emerge, international activities appear to move costs and product quality in the same direction,

“The survey by Pack (chapter 2) points out that both theoretical studies and empirical evidence indicate that host-country absorptive capacity is crucial for obtaining significant benefits fom FDI This suggests that liberalization of trade and FDI policies should be complemented with appropriate measues with respect to education, R&D, and human capital accumulation Similaely,empieical evidence supports the agument that intlleetual property rights ate trade related and that asymmetric intellectual property rights protection across countries dis torts the pattern of world trade and alters the composition of FDI inflows (Fink and Masks 2005; Javorcik 2004)

Lowering barriers to imports and FDI increases competition, as foreign goods and foreign producers enter the domestic market The magnitude of the effect, may be larger in developing countries, which (in part due to previous protection {fom international competition) lag behind industrial nations in terms of techno- logical sophistication, quality and variety of products, and productivity While in the short run the loss of marketshare to foreign goods or firms may push local firms up their average cost curve and thus lower ther productivity, in the medium, ran weaker firms will be forced to exit and survivors will lower thei cost base,

"upgrade their production, or both The result is that the average productivity of Indigenous fiems will increase This view is reflected in Tybout’s survey (chapter 4),

‘which concludes that mark-ups generally fall with import competition, import- competing firms cut back their production levels when foreign competition intensifies (atleast in the short run), increased trade rationaizes production in

10 Global Integration and Technology Transer

the sense that markets for the mos efficient plants are expancled but large import-

‘competing firms tend to simultaneously contract, and exposure to foreign compe- tition often improves intraplant efficiency

Policies ta liberalize trade and FDI ae, of course, not the only policy instru

‘ments for increasing competition Complementary policies are important as wel including in particular hard budget constraints to ensure that financial resources sce allocated to activites with the highest expected return and that firms ling to perform are forced t ext Dankov and Hoekman explore the importance of such

‘complementary sources of market discipline on TEP in chapter 12 They investi= gate the effects of demonopolization and restructuring of conglomerates, the

‘imposition of harder budget consteains, and trade liberalization on the perform- ance of Bulgarian firms between 1991 and 1995 They find 2 positive relationship between the increased market discipline associated with these reforms and subse- {quent productivity growth, result thats consistent with the general finding inthe literature, But they also conclude that different types of market discipline have dif- ferent impacts on exporting and nonexporting (import-competing) enterprises Trade policy reform has less of an impact on the exporters in the sample—indeed, trade reforms are found to have a significant effect only once firms that export most of thei output are removed from the simple, Moreover, acess to subsidies (Goft budget constraints, as measured by the magnitude of arrears) and more con- centrated markets reduce the impact of trade iberalization, Not considering such factors can lead to ether over- oF underestimation of the effects of trade reforms

on TEP

“The impact of increased competition asa result of FDI inflows is documented {in several chapters in this volume, as well s in many of the investment climate related surveys undertaken in developing countries and transition economies since the late 1990s In a 2003 survey of 391 domestic enterprises in the Czech Republic, almost half of respondents reported that foreign entry had increased the level of competition in their sector Twenty-nine percent of respondents claimed

to have lost market share as a result of foreign entry Similar responses were obtained from a survey of 396 enterprises in Latvia (Javoreik and Spatareanu, 2005), Many studies finding of a negative correlation between the presence of for ign firms and the productivity of domestic enterprises in the same sector may be tributed tothe fet that the increased competition resulting from foreign entry

‘outweighs the effect of knowledge transfer (Aitken and Harrison 1999) The per centage of Czech firms reporting multinational siphoning off the demand for their products exceeded the share of Czech firms reporting leening from foreign firms through the demonstration effect (figure 1.1) These findings help explain the negative spillovers obtained by Djankov and Hoekman in chapter 8

Lessons from Empirical Research on Internationa Technology Difusion 11 FIGURE 1.1 Perceived Effects of FDI in the Czech Republic

Policy Implications of Research Findings

“Market transactions in technology are affected (constrained) by three major prob- lems: asymmetre information, market power and externalities (Hoekman, Maskus, and Saggi 2005) These market failures imply that there isa potential for policies to increase welfare by encouraging international technology diffusion insofar as these failures result in a (globally) suboptimal amount of international technol- ogy diffusion Such policies may be constraining (for example, a prohibition on trade) or encouraging (for example, subsidies) To be effective a policy must alter the incentives of private agents that possess innovative technologies inthe right” way In practice, the potential for welire-improving policy may not be realized ddue to mistakes or rent-seeking activity In this area of poly, perhaps more than

in others, the potential for goveenment failure is high

Asymmetric Information

‘Technology transfer involves the exchange of information between those who have it and those who do not, Firms that have information eannot fully reveal their knowledge without destroying the basis for trade, creating a problem of asymmetric information: buyers cannot flly determine the value ofthe information

12 Global Integration and Technology Transer

before buyingit Ths can lead to large transactions costs, which stifle market-based technology transfer In the international context, information problems are more severe, are and the enforcement of contracts more dificult to achieve The theory

of the multinational firm posits that one motivation for such firms to establish foreign subsidiaries is the desive to overcome the difficulty of using markets to profit rom their proprietary technologies

Market Power

‘Owners of (new) technologies typically have substantial market power, result- {ng from lead time, fixed (sunk) costs, and learning by doing or from the grant- ing of intellectual property rights This necessarily implies that the price of technology will exceed the socially optimal level (that is, the marginal cost While this divergence between price and cost allows innovators to profit from their innovation, i also implies a reduction in the (static) national welfare of those importing technologies

Externalities

Externalities may arise if the costs and benefits of technology exchange ate not {ally internalized by those involved A major share of benefits to recipient coun- {ties of international technology diffusion is likely to arse from uncompensated spillovers Positive spillovers exist whenever technological information is diffused into the wider economy and the technology provider cannot extract the economic value ofthat diffusion Spillovers can arise from imitation, trade, licensing, FDI, and the movement of people

Import-competing industries register productivity gains following trade and EDI reforms (with much of the gains due primarily to factor use reallocation effects), and price-cost margins fall asa result of greater competition But the policy implications of these phenomena are not straightforward, as Erdem and

‘Tybout (2004) note, hecause itis not clear what drives improved performance Isit due to better management incentives? To greater returns to innovation? To incentives to shed redundant labor asa result of grester competitive pressures? Understanding the drivers is important from a policy perspective, as policies need to target the specific incentives that matter, There is significant uncertainty here One conclusion that seems robust is that domestic distortions (market failures) and general economywide policies motivated by public good rationales (slucation, infrastructure) may be mare important than trade and investment policies in determining long-run outcomes However, trade and investment policies area key determinant of international technology diffusion The analyses

in this volume of the effects of trade and FDI suggest that the impacts of trade

Lessons from Empirical Research on International Technology Diftsion 13

and investment policies differ depending on circumstances in the importing or host country such as the ability to supply intermediate inputs, the country’s absorptive capacity, and the extent of any bias toward or against types of imports and types of FDI

‘Technology-Related Trade Policies and International

Technology Diffusion

Many countries have historically engaged in infant industry protection, but the evidence suggests that diffusion of Knowledge is facilitated by an open trade regime (Hoskman, Maskus, and Sag 2005) Undistorted aces to capital equip sent and imported inputs that embody foreign knowledge allows irs to acquire know-how The cas for open markets extends to other pradcts swell as greater competition wll reduce pric cont markups At the sre ime, given that techno: ogy markets are associated with increasing returns, imperfect competition and exteraalics the argument aginst trade protection isnot unconditional The conclisions hinge othe scope of knowledge spillovers International spillovers, for which theres considerable evidence (see chapter 5 and the references cited the), srongl tt the balance in favor of fee trade If national spillovers are also inmportans, however, there may bea potential oe for intervention Fr example, i productivity improvements depend only om a country’s own RSD, a case can be tmade for policies that ensue that industries in which such improvements oeeurat

4 rapid at are not allocated elsewhere The relative magnitude of interstional

as opposed to national spillovers are not knoven, and much will depend om the specific circumstances of countries, What matter is that research indicates there

is geographic component to knowledge lows, suggesting that agglomeration ype effets can be important (Jae Tatenberg, and Henderson 1993; Sjgholm 1996; Andretsch and Feldman 1996) From a policy perspective, restricting trade is neither effiient nor effective, Instead general polis are needed to increase the incentives of agents to undertake activites that generate social benefits that exceed private rtums, without simultaneously crating additonal distortions uch pol cies inclade messutes to improve the investment climate and reduce the “costs of doing business” as well as investments in education Trade policy docs litle to encourage local R&D and necessarily leads to other distortions

‘One se of trade polices that is often motivated on the bats of international technology difusion objectives trade related investment measures (TRIMS) Examples are local content and technology transfer requirements for foreign investors, TRIM discriminate against imports by erating incentives (oer than import taifs) to source inputs from domestic producers In the international technology difusion context, a motivation for TRIMS is often that foreign firms

14 Global integration and Technology Transer

‘might be expected to transfer knovedge to ensee that local inputs satisfy their specifications TRIMS act as an implicit tariff on intermediate goods imports, because manufacturers ate forced to use higher-cost local inputs They are inferior

to tarifs in welfare terms, as no tariff revenue is generated More important from the perspective of international technology diffusion isthe fact that TRIMS pro- vide little oF no incentive for the protected producers of intermediate goods to

EDE-Related Policies

Historically, restrictive trade policies were often complemented by restrictions on EDI,in part to prevent the use of investment a-a way of getting round trade pro: tection (so-called tarif-jumping investment) Japan, the Republic of Korea, and

“Taiwan (China) have all imposed restrictions on EDI Policies were often more welcoming toward other modes of international technology difusion, including liberal trade policies affecting imports of machinery and equipment (through zer0 duty ratings or drawback mechanisms) and the licensing of foreign technol-

‘ogy: In recent years national FDI policies have generally become more liberal in developing countries, but governments often differentiate between joint ventures and fully owned subsidiaries of multinational enterprises Chins, amvong others, thas been more encoueaging of joint ventures than inward FDI, although it bas bbecame the leading destination for FDI Such a policy stance may reflect a desire

to maximize technology transfer to local agents

‘The iteratute suggests that such discriminatory policies are unlikely to achiev the desired effects with espect to technology transfer Survey evidence indicates that tech- nologies transferred to wholly owned subsidiaries ace of « newer vintage than licensed technologies or those transferred to joint ventures (Mansfield and Romeo 1980) For-

«ign investors alo tend to devote more resources to technology transfert their wholly owned subsidiaries than to partially owned afiiates (Ramachandaram 1993) More- lover the econometric analysis conducted by Javorcik (chapter 10) shows that multina- tional enterprises with advanced technologies tend to enter a host countey through

‘wholly owned subsidiaries rather than joint ventures By forcing multinational enterprises to license their technologies oF engage in joint ventures, host countries maybe lowering the quality of technologies they receive as wells reducing the incens tives to invest in the county atl

While the magnitude of international technology diffusion undertaken by

‘multinational enterprises need not be socially optimal, considerable evidence sug- gests that such firms are keen to transfer technology to their local suppliers." Pal cies that Facilitate this process, as opposed to regulation or legislation requiring that multinational enterprises engage in international technology diffusion to

Lessons from Empirical Research on International Technology Diftsion 15 local competitors havea greater likelihood af being cessful A concrete example

of such policy would be a supplier development program that aims to prepare local companies to understand and meet the needs of multinational enterprises Such programs have been successfully implemented in Ireland and mone recently the Czech Republic Participation is offered to promising small and medium-size enterprises that undergo an evaluation proces fllowed hy intense work with out- side consultants to make improvement to areas in which they are licking, Pactici- pants ae given opportitcs to start new business relationships by meeting with

‘multinationals looking for local suppliers,

IF they aze well designed, the services provided in this way can he elective in assisting firms, especialy smaller enterprises, mobilize the right type of skills (consultants) and ensure that their target audience is awate ofthe services on offer In practice these conditions are frequently not met As result such pro {grams often hae limited impact at best, because the services provided do not [generate much value-added or because the “wrong” fems—those with limited potential for improvement or those that had access to market-based consulting services (in the process creating a potential negative spillover on the private consulting industry)—get help.”

Many countries actively sek to attrac foreign investors through up-front sub- sides, ox holidays, and other grants Fr thereto bea eationale fr such inves ment incentives host countries must enjoy positive externalities from inward FDL

“The prevalence of lla the leader" behavior among multinational enterprises provides another potential case for FDI incentives Given the oligopolstic ature

of markets within which FDI occurs, new entrant may atrat investments by both competitors and upstream supplies fit does, competition at multiple stages of production may increas, raising efciney, overall output, and employ- sent ba implication is that host country may be abe to unleash a sequence of investments by successfully inducing FDI from one or two major Firms

IF the local economy lacks a well-developed network of poteotial supplies, snultnational enterprises may be hesitant a invest, and local supplies may not develop because of lack of demand Inthe presence of such interdependence, {growth may be constrained by a coordination problem that can he partially resolved by initiating investments from key firms Ofcourse, such coordination problems cannot be tackled solely theough investment incentives Policy efforts need to focus primarily on improving te investment climate and reducing the costs of absorbing technology complex task that involves building human capital and expanding national innovation systems

"Thus while there may bea case fr incentives it sa conditional one To be effective, the investment climote and absorptive capacity must meet certain conditions.” Moreover given competition among countries in providing FDI

16 Global Integration and Technology Transer

incentive packages, one must carefully examine the magnitude of potential costs and benefits associated with such policies, Before considering FDI incentive, it probably makes sense to inves in marketing a country asa profi

investment and ensuring that potential investors can easly access all relevant information (Rodeiguez-Clare 2004),

to take full advantage of international technology diffusion

Daniel Lederman and William Maloney (chapter 13) examine the evolution of

‘Mexican technological progres in the past few decades, devoting special attention

to the role of trade, FDI, and the national innovation system Their main message

is that trade liberalization (North American Free Trade Agreement) was helpful but insullicient to help Mexico catch up tothe levels of innovation and the pace of technological progress in its North American partners The evidence suggests that, given its level of development, Mexico suffers from low levels of R&D expenditure and low levels of patenting activity It severely underperforms when compared with sucessful developing economies, such asthe Republic of Korea The authors argue that Mexico’ national innovation system—how the private sector, universi- ties, and public polices interact to produce economically meaningful innova- tion —is ineficient Without addressing this weakness, Mexica is unlikely to be ale to converge an the pace of innovation in North America, Improving the qual- ity of research institutions; providing incentives for researchers to get involved with the productive sector, in particular incentives for appropriating innovations cmanating from technical research; negotiating the cofinancing of research

«exchange progeams with NAFTA partners; and developing domestic credit mar- kets are among the policy responses suggested

Of great relevance to inteenational technology diffusion isthe role that subsidies can play to facilitate learning, technology acquisition, and dynamic comparative

Lessons from Empirical Research on International Technology Diftsion 17 auvantage where returns to such activities eannot be appropriated by private agents and hence will not be undertaken by any individual fim Amsden (1989) and others argue that policy interventions, incudingipliitor explicit subsidies, lay behind the economic“miralesin the Republic of Korea and Taiwan (China) They claim that carefully ageted subsidies allowed these governments to tim late key sectors that hesame efficent in their own right and provided positive spillovers fr the economy as a whole

In considering this infant industry argument for government support itis inportant to diferetite between sector specific subsidies and general policies facilitating learning and the development of private enterprise, ln a recent com prehensive retrospective on the East Asian development experenes, Noland and Pack (2008) argue that sectorspecific policies did not result in high rates of TEP

‘growth for manufacturing In the Republic of Korea and Taiwan (China), TEP

‘growth was not much higher than inthe Organisation for Esonomie Co-operation and Development (OECD) In india selective interventions polices were associated wit detning TEP growth rates, whilthe opening ofthe economy led to an increase

in TEP growth rates (Keshna and Mitra 1998) Speci interventions will often got

it wrong in part as a result of rent-scking, nd in part a a eslt of general equi- vin effects (a subsidy for one ett usualy implies taxon allothers)

“The cae foe general policy support far certain types of activity, including ino: vation, education, transport infrastructure and similar publi goods, is unconto versal The same is true fr policies aimed to promote socially enefcial activites Markets do fl, and there may be good rationales for governments to provide incentives for firms and agents to undertake activities that would otherwise be undersupplied An important example that has a direct bearing on the subject at hand isthe learning externality analyzed by Hausmann and Redrik (2003) The market undersupplies investment by frm in new (nontraditional) activites because of appropriabilty problems: as soon as an entrepreneur i succesful in identifying a profitable new production opportunity entry by imitators prevents him of her from recouping coms [n this eas, subsidy or similar incentive can help expand innovation and risk taking* As Rodriguz-Clare (2004) note, this argument provides rationale foe supporting innovating projects leading to new knowledge about the country’s comparative advantage but i sna a justification for general poliy of supporting small and medium-size enterprises No sta

‘ae for trade policy which i too blont an instrument (Baldwin 1969) Any support program shouldbe aimed at young firms, as fens that have emained smal for ran years most likely exhibit lowe productivity and are therefore not promising targets for intervention (se, for example, Srinivasan 2004),

“The elisient use of support policies requires that governments be elective at identifying cases that justly intervention, designing elfectve instuments, and

18 Global Integration and Technology Transer

{implementing these instruments appropriately In practic, governments may fil {in doing ll thre: the policy problem isto assess the relative sizes of government failure and the failures ofthe market for knowledge Potential problems include

‘the fact that subsidies can serve to support inefficiency firms may behave strategically (by underinvesting, for example) in order to win subsidies; and subsidies can result in corruption, bad corporate governance, and rent-seeking behavior The greatest challenge of implementing subsidies is that they are dificult to control Government needs to establish an effective and credible exit strategy that weeds

‘out unsuccessful efforts from successful ones, The capabilities and autonomy of the state play a fundamental role in implementing subsidy policies effectively (Rodrik 1993), suggesting that ths isa policy path that countries with weak gov- ertanee and institutions pursue at their peril Greater policy consistency and credibility may be obtained through international cooperation (for example, trade agreements) which can help government be more credible in ensuring that

«policy to assist a particular activity is temporary (Hoekman 2005)

Intellectual Property Rights and Technology Transfer

Intellectual property rights can support markets in technology including interna- tional technology diffusion, and they ar likely to interact with both trade and FDI flows (Arora, Fosfuri, and Gambarlela 2001; Javordik 3004; Maskus 2009) Although the chapters in ths volume do not address intellectual property rights issues, we discuss briely some of the stylized facts and findings fom the literature fiom the perspective of international technology diffusion.”

Absent intellectual property rights, firms would be less willing to engage in technology transactions, Patents and tade secrets provide legal basis for revealing the proprietary characteristics of technologies to subsidiaries and licensees, supporting the formation of licensing contracts” Patent peotection both increases flows of international technology diffusion to countries with techno- logical capacity and shifts incentives for investors between FDI and licensing

“The empirical literature yields several results Fest, patent applications from forcign nations are strongly associated with productivity growth in recipient counties (Eaton and Kortum 1996) Except inthe United States, more than half

of productivity growth in OECD countries comes from importing technologies (patents) This proportion is highe for small economies Thus tade in ideas isa

‘major factor in world economic growth

Second patent citations reflect knowledge Flows across borders While there is

a limited amount of eifasion overall (because of distance, border, and differences across regions in technological specialization [Peri 2003]), the most significant patents and knowledge in highly technological sectors are widely diffused There isa strongly positive impact of knowledge flows on international innovation,