Technology Diffusion on the International Trade Network

• DOI: http://doi.org/10.1111/jpet.12186
• Date: 01 April 2016
• Published date: 01 April 2016

TECHNOLOGY DIFFUSION ON THE INTERNATIONAL TRADE

NETWORK

GARY D. FERRIER

University of Arkansas

JAVIER REYES

University of Arkansas

ZHEN ZHU

IMT Lucca

Abstract

Technological innovations generate knowledge spillovers—non-

innovators benefit through the adoption, imitation, and extension of

new technologies. International trade facilitates technology diffusion

by providing importing countries access to technical knowledge that

they can potentially internalize. Previous studies of the effect of

trade on technology diffusion typically only consider the impact of

direct (bilateral) trade on indirect measures of technology (e.g., total

factor productivity). We contend that the analysis of trade’s impact

on technology diffusion would be more accurately assessed by using

direct measures of specific technologies (e.g., intensity levels) and

by allowing for the influence of both the direct and indirect effects

of trade in the analysis. The latter is accomplished by modeling the

international trade system as a weighted network, which quantifies

both direct and indirect trade linkages. Combining trade data with

data on the adoption of specific technologies, we find that the network

effects of trade play a significant role in technology diffusion. In most

cases, countries that are better-connected on the trade network have

higher technology intensities. Further support for the importance

of trade is provided by the finding that for “outdated” technologies,

better-connected countries have lower technology intensities because

of their adoption of newer, substitute technologies.

Gary D. Ferrier, Department of Economics, University of Arkansas, Sam M. Walton College of Business,

1 University of Arkansas, Fayetteville, AR 72701, U.S.A. (gferrier@walton.uark.edu). Javier Reyes, De-

partment of Economics, University of Arkansas, Sam M. Walton College of Business, 1 University of

Arkansas, Fayetteville, AR 72701, U.S.A. (reyes@uark.edu). Zhen Zhu, LIME lab, IMT Institute for

Advanced Studies, Piazza S. Ponziano, 6, 55100, Lucca, Italy (zhen.zhu@imtlucca.it).

The authors thank Gisela Rua of the Federal Reserve Board, two anonymous referees, and the partic-

ipants of the 2013 International Workshop on Networks and Trade at KU Leuven, the 2011 Southern

Economic Association Conference, and the economics seminars at the University of Arkansas and IMT

Lucca for their insightful comments.

Received April 30, 2013; Accepted October 3, 2015.

C2016 Wiley Periodicals, Inc.

Journal of Public Economic Theory, 18 (2), 2016, pp. 291–312.

291

292 Journal of Public Economic Theory

1. Introduction

Technological progress has long been viewed as a critical engine for sustainable eco-

nomic growth and poverty reduction. However, because technological innovation oc-

curs almost exclusively in a few high-income countries (Keller 2004, 2010), the tech-

nological progress at a global level largely depends on knowledge “spillovers” through

technology diffusion.1In principle, there exist several possible economic channels that

facilitate the diffusion of new technologies across countries; the most frequently studied

are international trade and foreign direct investment (FDI).2The international flow of

goods and services through trade and FDI provide first-hand knowledge about the ex-

istence, uses, and benefits of new technologies.3This paper adds to the literature on

the effects of trade on the diffusion of technologies across countries by modeling trade

as a network of both direct and indirect relationships as well as by examining trade’s

effect on the diffusion of specific technologies rather than on an aggregate proxy for

technology such as total factor productivity (TFP).

When empirically testing the effects of trade on technology diffusion, the previous

literature typically considers only the direct (bilateral) trade relationships across coun-

tries (e.g., Coe and Helpman 1995; Keller 1998; Comin and Hobijn 2004). Furthermore,

empirical results have often been at odds, perhaps because different measurements have

been used to assess trade relationships. For example, Coe and Helpman (1995) use im-

ports to weight foreign R&D expenditures and find a significant relationship between

the import-weighted foreign R&D expenditures and domestic TFP, offering evidence of

technology diffusion through trade. Using Coe and Helpman’s (1995) data set, Keller

(1998) employs the summation of foreign produced R&D over time rather than import-

weighted foreign R&D to examine the relationship between technology and TFP. In

contrast to Coe and Helpman (1995), Keller’s (1998) results suggest that knowledge

spillovers are independent of trade and are therefore global, rather than local, in na-

ture. Conversely, Keller (2002) argues that trade is concentrated because of transporta-

tion costs and thus knowledge spillovers are localized. Keller (2004) also concludes that

knowledge spillovers are localized because of the nature of technology—understanding

technology requires interaction between innovator and adopter; trade may provide the

needed interaction.

The hypothesis to be tested is that countries that are better-connected on the trade

network will have higher technology intensities, which implies greater technology dif-

fusion. We contend that the effect of trade on technology diffusion is not confined to

the direct trading partners of the innovator. Once knowledge is received by an inno-

vator’s direct trading partners, it can diffuse further to the trading partners of trading

partners—i.e., there are spillover effects beyond the initial exchange with an innovating

trade partner due to network effects. Thus, the impact of trade would be more accu-

rately assessed by accounting for both direct and indirect trade relationships. This can

be done by considering the international trade system as a (weighted) network of trade

relationships.

1The terms “technology diffusion,” “technology transfer,” and “technology adoption” are often used

interchangeably in the literature.

2For more detail, see Hoffman (2013) who notes that there is a large literature across multiple disci-

plines on the links between technology diffusion and both international trade and FDI.

3Although international trade has received empirical support as an important channel for technology

diffusion, the FDI channel has been harder to validate. This may be because, as Keller (2004) observed,

testing the FDI channel has been largely restricted to company-level case studies. Section 2 briefly

discusses the role of FDI.