International spillovers of R&D and marginal social returns

• DOI: http://doi.org/10.1111/roie.12404
• Author: Elmer Sterken,Kazuo Ogawa,Ichiro Tokutsu
• Published date: 01 August 2019
• Date: 01 August 2019

936

|

© 2019 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/roie Rev Int Econ. 2019;27:936–954.

1

|

INTRODUCTION

The Global Financial Crisis of 2008/2009 has had a serious negative impact on the majority of devel-

oped economies. Especially in Europe there is a growing and serious concern for the development of

total factor productivity (TFP) growth. Stiglitz (2012) considers the slowdown and uneven distribution

of productivity growth as the important cause of the financial crisis. Van Ark et al. (2013) claim that

TFP has emerged as the Achilles' heel of Europe's growth performance. This typically applies to the

“older” members of the European Union, the EU‐15, and maybe less to the accession countries. It is

generally believed that investment in R&D could stimulate TFP growth, as it is typically the case for

the U.S. economy. Investment in R&D can be transmitted easily by modern IT, leading to international

spillovers. On the one hand this could, for instance, lead to a larger growth potential in the EU, where

the innovation space has been enhanced, but on the other hand game‐theoretic considerations could

Received: 16 June 2018

|

Accepted: 4 April 2019

DOI: 10.1111/roie.12404

ORIGINAL ARTICLE

International spillovers of R&D and marginal social

returns

KazuoOgawa1

|

ElmerSterken2

|

IchiroTokutsu3

1College of Foreign Studies,Kansai Gaidai

University, Hirakata City, Japan

2University of Groningen, Groningen, The

Netherlands

3Graduate School of Business

Administration,Kobe University, Kobe,

Japan

Correspondence

Kazuo Ogawa, College of Foreign

Studies, Kansai Gaidai University, 16‐1

Nakamiyahigashino‐cho, Hirakata City,

Osaka 573‐1001, Japan.

Email: kogawa@kansaigaidai.ac.jp

Funding information

This research was financially supported

by KAKENHI Grant‐in‐Aid for Scientific

Research (S) #15H05728, (B) #25285068

and the program of the Joint Usage/

Research Center for Behavioral Economics

at ISER, Osaka University.

Abstract

This study analyzes marginal social and private returns of

R&D investment through the impact of international spillo-

vers of R&D stocks. We compare the marginal social with

marginal private returns using data of 27 OECD and EU

countries from 1995 to 2008. We consider two channels of

R&D spillovers: embodied in trade flows and disembodied

by bilateral technological proximity. We find that marginal

social returns on R&D are much larger than the marginal

private returns for R&D‐intensive countries, in the embod-

ied spillover channel. We also find that the embodied spillo-

ver channel through import flows is more important than the

disembodied channel.

JEL CLASSIFICATION

E22; O32; O33; O47

|

937

OGAWA et Al.

lead to underinvestment: why should a small open economy invest in R&D if spillovers of investments

abroad could be absorbed? In order to understand the role of R&D investment in economic develop-

ment it is highly relevant to analyze the drivers of investment and, more importantly, the impact of

own and imported impact of R&D stocks.

There is a large body of studies that have examined the impact of international spillovers on recip-

ient countries from a variety of angles. Our study is novel in that we estimate the marginal social re-

turns on R&D by incorporating the channels through which a change in the R&D stock of one country

is propagated into the R&D activities of other countries. To this end we endogenize the accrual of the

domestic R&D stock by estimating R&D investment functions. R&D investment decisions generally

depend on the available own domestic R&D stock as well as foreign R&D stocks. Therefore, an exog-

enous shock to the R&D stock in one country is propagated into the R&D stocks of other countries by

way of the interplay of R&D investment across countries. This eventually affects the growth path of

all countries. To that end we also estimate the development of TFP. We compute the marginal social

returns on R&D investment as well as marginal private returns for each country by estimating this

propagation process. To the best of our knowledge, this is the first attempt to estimate marginal social

returns on R&D, taking account of the interplay of R&D activities across countries. Specifically, we

follow the procedure taken by Bloom, Schankerman, and Von Reenen (2013), who develop a method-

ology for computing the marginal social and private returns on private R&D capital stocks, measured

in terms of the output gains generated by a marginal increase in R&D over heterogeneous firms.

Our sample countries consist of 27 OECD and EU countries during the years 1995 to 2008. Most

of the past studies have covered only OECD countries or industrialized countries1

. Industrialized

countries are of course active in R&D investment and to a large extent potential suppliers of technol-

ogy. We expand the coverage of sample countries to include relatively new EU accession states that

the previous studies have not shed light on2

. These countries are less active in R&D investment, but

might benefit more from international R&D spillovers. So the EU accession countries are an ideal

set to analyze R&D spillovers in more detail. The World Input–Output Database (WIOD) of the

University of Groningen is a perfect database that provides basic data to describe R&D spillovers

across borders through bilateral trade flows for all the EU states3

.

Let us preview our main findings. We construct “composite” foreign R&D stocks under two differ-

ent assumptions of channels through which technology is transmitted across countries. In one channel

we assume that imports embody technological knowledge of trade partners and bilateral import shares

are used as relative weights of domestic R&D stocks constructing the foreign R&D stock. This is in

line with pioneering work of Coe and Helpman (1995). Specifically, we use the total import as weight

variables in constructing foreign R&D stocks4

. In the other channel R&D spillovers are transmitted

directly in disembodied form using bilateral technological proximity between countries á la Jaffe

(1986).

In our analysis we find that international technology spillovers are better explained by bilateral im-

port shares than by disembodied direct spillovers. The marginal social returns on R&D capital through

bilateral import flows are more than twice as large as the marginal private returns for R&D‐intensive

countries. In contrast, the marginal social returns are slightly above the marginal private returns for

less R&D‐intensive countries. We also find that the embodied spillover channel is more important

than the disembodied spillover channel in the sense that the marginal social returns are much larger

than the marginal private returns in the embodied channel. R&D‐intensive countries moreover have a

larger gap between the marginal social and private returns. It implies that R&D‐intensive countries are

expected to generate more spillovers at the margin, but the observed R&D is smaller than the—from

a global perspective—socially optimal level.