The Review of Economics and Statistics

VOL. CIV

MARCH 2022

NUMBER 2

BETTER, FASTER, STRONGER: GLOBAL INNOVATION
AND TRADE LIBERALIZATION

Federica Coelli, Andreas Moxnes, and Karen Helene Ulltveit-Moe*

Abstract—This paper estimates the effect on innovation of increased market
access facilitated by trade liberalization. We use a novel empirical design
that exploits tariff cuts during the 1990s, along with detailed data on inno-
vation among ﬁrms from 65 Länder. Our results reveal a large effect of
tariff cuts on innovation as measured by patent data, suggesting that mul-
tilateral liberalization has promoted innovation and growth. These effects
are not driven by the deterioration of innovation quality, and the results are
robust to controlling for changes in the patent system and to industry-wide
trends in innovation.

ICH.

Einführung

TRADE policy liberalization opens up new markets abroad

and therefore increases the effective size of the market.
Economists have long known that the amount of invention
is governed by the extent of the market.1 However, there is
currently no comprehensive empirical study of how improved
market access through trade liberalization has affected world-
wide innovation.

This paper seeks to ﬁll this gap in the literature. Rather than
focusing on a speciﬁc country, we present a novel global data
set as well as empirical methodology that allows us to produce
broad and systematic evidence on the impact of improved
market access on worldwide innovation. Our approach en-
ables us to disentangle this effect of trade liberalization on
innovation from other institutional changes that often go hand
in hand with trade policy.

Tariffs in both developing and developed countries came
down substantially in the 1990s, leading researchers to name
the period the Great Liberalization (Estevadeordal & Taylor,
2013). This was partly due to the completion of the GATT
Uruguay Round in 1994, which resulted in substantial tariff
cuts over the period 1995 Zu 2000. On average, developed

Received for publication September 7, 2018. Revision accepted for pub-

lication June 16, 2020. Editor: Amit K. Khandelwal.

∗Coelli: University of Zurich; Moxnes and Ulltveit-Moe: Universität

Oslo and CEPR.

We thank Swati Dhingra and Rachel Grifﬁth, numerous seminar and con-
ference participants, anonymous referees, and the editor for helpful sug-
gestions and valuable discussion. We also thank Bjarne J. Kvam from the
Norwegian Patent Ofﬁce (Patentstyret) for guidance related to the patent
Daten. This project has received funding from the European Research Council
under the European Union’s Horizon 2020 research and innovation program
(grant agreement 715147 and grant agreement 805007).

A supplemental appendix is available online at https://doi.org/10.1162/

rest_a_00951.

1An early contribution is Schmookler (1966).

country tariffs were cut from around 6% Zu 3%, while de-
veloping country tariffs were cut from almost 20% Zu 13%
zwischen 1990 Und 2000.2 This paper uses the Great Liberal-
ization as a quasi-natural experiment to estimate the causal
impact of improved market access on innovation among ﬁrms
aus 65 Länder.

A major empirical concern in any study of the effect of
market access on innovation is the endogeneity of tariffs,
Zum Beispiel, that tariff reductions are likely to reﬂect cross-
industry differences in lobbying intensity and industry con-
centration. We overcome this issue by exploiting variation in
applied most favored nation (MFN) tariff cuts across a ﬁrm’s
export markets, which are likely exogenous to other determi-
nants of innovation in the home country and industry of the
ﬁrm. Konkret, we link tariff data to the initial industries
and foreign countries the ﬁrm is exposed to through patent
ﬁling, in order to compute the average tariff cut faced by the
ﬁrm. Intuitively, a ﬁrm x located in Germany and selling to
the United States and Mexico is affected differently from a
ﬁrm y in the same industry selling to China and South Korea
because tariff cuts vary across those export markets. Condi-
tional on industry-country trends, those MFN tariff cuts are
unlikely to be caused by those two ﬁrms.

The data requirements for this exercise are large; one would
ideally need a ﬁrm-level panel data set on innovation, along
with detailed information on where ﬁrms are located and in
which markets they sell in. To achieve this, we construct a
global and comprehensive microlevel data set on patenting
based on the global database PATSTAT, recently developed
by the European Patent Ofﬁce. We observe nearly every ﬁrm
worldwide that ﬁles a patent, in which country (patent ofﬁce)
each ﬁles, along with its industry and home country afﬁlia-
tion, over four decades. We follow Aghion et al. (2016) Und
construct ﬁrm-level measures of market exposure by using
information on patent ﬁling in the years prior to the Uruguay
negotiations. Compared to weights based on exports, diese
weights based on patent ﬁling are potentially superior mea-
sures of market exposure because they are likely to reﬂect
the ﬁrms’ expectations of where their future markets will
Sei. Darüber hinaus, we provide evidence that these patent weights

2Applied most favored nation (MFN) tariffs. See online appendix E for

Einzelheiten.

The Review of Economics and Statistics, Marsch 2022, 104(2): 205–216
© 2020 The President and Fellows of Harvard College and the Massachusetts Institute of Technology
https://doi.org/10.1162/rest_a_00951

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

206

THE REVIEW OF ECONOMICS AND STATISTICS

are strongly correlated with weights constructed from ﬁrm-
destination level exports.

Our ﬁrm-level approach has a number of advantages. Erste,
because initial foreign market exposure varies signiﬁcantly
within a country and narrowly deﬁned industries, our global
approach allows us to sweep out all home country-industry
trends in innovation by ﬁxed effects. Dabei, we tackle
a set of well-known challenges. Erste, tariff cuts also lead to
greater import competition in the ﬁrms’ home market, welche
also affects innovation (Bloom, Draca, & Van Reenen, 2016).
Zweite, the likelihood of patenting depends on a host of time-
varying factors such as legal framework and technological
characteristics of an industry.3 Third, changes in tariff pol-
icy often go together with other government reforms, solch
as product market deregulation. Our empirical methodology
allows us to isolate the impact of improved export market
access and sidestep all of the above issues.

Our results are robust to a number of potential concerns.
Erste, ﬁrms within the same industry with different export
market exposure may also be differentially exposed to import
competition from these countries. We show that controlling
for the tariff cuts in a ﬁrm’s home market does not change our
main results. Zweite, one may worry that MFN tariff cuts in
the destination country are contemporaneous to other policy
changes in that country. An example of this is market size (oder
stricter patent enforcement, such as TRIPS). Being exposed
to a high-tariff-cut country may be correlated with innovation
simply because that country grows fast and increased market
size fosters innovation. We show that we can either introduce
a vector of destination ﬁxed effects or, alternatively, use a
control function approach that will eliminate this concern.
Dritte, our long time period allows us to perform placebo tests;
we test if treated ﬁrms exposed to high-tariff-cut countries
typically always patent more.

Our results show that the Great Liberalization of the 1990s
had a large, positive net impact on innovation. The overall
estimates mask considerable heterogeneity across countries
and industries. The impact of market access on innovation
appears to be greater in developed compared to developing
Länder. Darüber hinaus, the effect is greater among countries
that were initially more closed to trade.

One may question whether increased patenting reﬂects
mehr Innovation. The literature typically ﬁnds a strong cor-
relation between patenting and research and development
and between patenting and other measures of innovation.
Jedoch, the concern remains that more trade could induce
the need for greater protection of intellectual property rights
(IPR), das ist, that more patenting can simply be attributed
to a “lawyer effect.” To deal with this, we calculate citation
counts for all ﬁrms in our data set to control for the quality
of a patent and check whether average citations are falling
in response to trade liberalization. The data reject this hy-
pothesis; if anything, average citations are rising in response

3Typical examples are regulatory changes in the patent system and differ-

ences across patent ofﬁces.

to better market access. Alternative measures correlated with
the economic value of patents conﬁrm that market access has
not led to a reduction in patent quality.

The contributions of this paper are as follows. Erste, we de-
velop a simple theory on trade and innovation and a novel em-
pirical methodology consistent with the model. This allows
us to isolate the effect of improved market access and produce
broad and systematic evidence of the impact of trade liber-
alization on worldwide innovation over a decade with steep
global tariff declines. daher, our analysis goes beyond the
current literature that has primarily focused on unilateral or
regional trade liberalization. Zweite, there is a large literature
on the impact of trade policy on ﬁrm performance (z.B., TFP
or labor productivity), but there is much less direct evidence
on observable output-based measures of innovation such as
patents; see Steinwender and Shu (2019) for a thorough re-
view. We provide comprehensive evidence on the role of one
of the mechanisms through which trade liberalization fosters
improved ﬁrm performance and productivity growth. Dritte,
we construct and analyze a novel, comprehensive, and global
ﬁrm-level patent data set that has so far not been analyzed
within the context of international trade.

Our analysis speaks to different strands of literature. Unser
work is related to the empirical analyses of ﬁrm-level data
on the impact of trade liberalization on ﬁrm performance
such as Amiti and Konings (2007), Goldberg et al. (2010),
Khandelwal and Topalova (2011), and Loecker et al. (2016).
Our work also indirectly relates to the literature on trade,
import competition, and technology adoption. Bloom et al.
(2016) analyze the effect of Chinese import competition on
technology upgrading in Europe, while Autor et al. (2019)
examine the impact of China competition on patenting in the
United States.4

We also relate to Bustos (2011) and Lileeva and Treﬂer
(2010), who analyze complementarities between trade liber-
alization and technological upgrading and innovation. Was
distinguishes our paper from these contributions is that (A) Wir
develop a new identiﬁcation strategy, (B) we offer evidence
from global tariff cuts, Und (C) we use patents as a direct
output-based measure of innovation rather than input-based
or survey information.5

Endlich, our empirical approach is related to Aghion et al.
(2016) and Calel and Dechezleprêtre (2016), who also use
PATSTAT data but focus on very different questions, nämlich,
the impact of environmental policies on technical change. Unser
choice of approach and results inform not only the literature
on trade policy but also the broader literature on the effects of
the drivers of innovation (Acemoglu & Linn, 2004; Aghion
et al., 2005; Bloom et al., 2016 and Grifﬁth, Harrison, &
Simpson, 2010).

4Aghion et al. (2018), Boler, Moxnes, and Ulltveit-Moe (2015), Gopinath
and Neiman (2013) and Halpern, Koren, and Szeidl (2015) also examine the
link between ﬁrm performance and trade, but do not analyze trade policy.
5Steinwender (2015) also documents the relationship between access to

export markets and productivity increases in the case of Spain.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

BETTER, FASTER, STRONGER

207

Der Rest der Arbeit ist wie folgt gegliedert. Section II
presents our theoretical framework. Section III lays out the
empirical model and highlights econometric issues. Abschnitt
IV describes the data and descriptives. Section V presents and
discusses the empirical results, and section VI concludes.

II. Theoretical Framework

We aim to investigate the effect of foreign market access
on ﬁrms’ innovation. To do so, we start by presenting a basic
economic framework to support the analysis and proceed by
developing testable predictions for the relationship between
market access and innovation.

Consider a ﬁrm i with productivity zi, located in country m,
and producing in industry j with constant returns to scale us-
ing only labor. Goods sold from m to a foreign country n in in-
dustry j are subject to an ad valorem tariff Tjmn = τ jmn − 1 ≥
0. Preferences across varieties within an industry are CES
with an elasticity of substitution σ. This gives rise to a demand
function Ain p−σ
imn in country n, where pimn is the price charged
by ﬁrm i in n and the demand shifter Ain may vary across ﬁrms
and countries, and is exogenous from the point of view of an
individual ﬁrm.6 Producers engage in monopolistic competi-
tion, so that the price charged by ﬁrm i in market n is pimn =
[σ/ (σ − 1)] τ jmnwm/z, where wm is the wage of country m.
For expositional clarity, we normalize the wage to 1, als es
will be inconsequential for the remaining analysis. The prof-
its from serving country n are πimn =
Bin = (1/σ) [(σ − 1) /σ]σ−1 Ain. Global proﬁts are then

(cid:3)σ−1 Bin, Wo

z/τ jmn

(cid:2)

(cid:2)i =

(cid:5)(cid:6)

(cid:4)

(cid:7)σ−1

(cid:8)

Bin

zi
τ jmn

The ﬁrm faces the problem of how much to innovate. Con-
sider the simplest possible case where productivity z is pro-
portional to the ﬁrm’s stock of knowledge Ki, zi = ξKi. Wir
discuss the measurement of Ki in sections III and VB. Gain-
ing new knowledge is costly, and we assume that the cost of
obtaining a stock of knowledge Ki is c (Ki) = ψK k
ich , Wo
ψ determines average innovation cost and k > σ − 1 abhalten-
mines how quickly those costs rise with knowledge. The ﬁrm
then chooses the optimal Ki that maximizes global net proﬁts,
(cid:2)i − c (Ki). In online appendix A, we show that the optimal
knowledge stock is
(cid:9)

(cid:10)

1/[k−(σ−1)]

Ki = κ

τ1−σ
jmn Bin

(1)

(cid:4)

where κ is a positive constant.7

6Given the CES structure, Ain = Ein/Pσ−1
and Pjn is the CES price index of industry j.

7κ ≡

(cid:11)

(cid:12)
ξσ−1 (σ − 1) / (kψ)

proﬁt maximization is satisﬁed when k > σ − 1.

1/[k−(σ−1)]. The second-order condition for

, where Ein is a demand shifter

Now consider a change in τ jmn from one equilibrium to
the next. Using the exact hat algebra approach as popularized
recently by Dekle, Eaton, and Kortum (2008), we get
(cid:10)

(cid:9)

1/[k−(σ−1)]

ˆKi =

ωin ˆBin ˆτ1−σ

jmn

(2)

(cid:4)

(cid:13)
Ö

where ωin = πin (z) /
πio (z) is the share of global proﬁts
coming from market n in the initial equilibrium, and the hat
notation denotes the value in the counterfactual relative to the
initial equilibrium, i.e. ˆx ≡ x(cid:5)/X. Gleichung (2) highlights two
important economic mechanisms. Erste, all else equal, foreign
market access (lower τ jmn) leads to both higher proﬁts and a
greater knowledge stock. Intuitively, a larger effective market
means that a marginal improvement in productivity or quality
yields a higher return. Zweite, our theory shows that tariff
cuts in large markets matter more for innovation compared
to tariff cuts in small ones and that the theoretically correct
weight is the initial share of proﬁts in that market.8

Before concluding this section, we brieﬂy discuss three
possible extensions of the model. Erste, in our model, tariff
cuts matter only if the ﬁrm is already exporting to a desti-
nation, das ist, if ωin is strictly positive. In der Praxis, ﬁrms
may choose to both start exporting to country n and inno-
vate as a response to tariff cuts in n. We investigate this case
theoretically in online appendix B and empirically in online
appendix I. Darüber hinaus, we show in online appendix G that
there is a striking degree of persistence in ωin over time, sug-
gesting that exit or entry into new markets is limited in our
data set.

Zweite, the approach chosen here means that we only an-
alyze the impact of market access on innovation among ex-
isting ﬁrms, das ist, we do not consider ﬁrm entry. Das ist
guided by the empirical analysis and identiﬁcation strategy.
As will become clear, our unit of analysis is the ﬁrm, and we
require presample data on ωin for all the ﬁrms in the data set
(which by construction excludes entrants from the analysis).
Dritte, we abstract from other economic factors that also
affect innovation. Wichtig, it is well known that a more
competitive marketplace (e.g. coming from import competi-
tion) has an impact on innovation (Aghion, Harris, & Vick-
ers, 1997, and Aghion et al., 2005). In diesem Papier, we iden-
tify only the effect of market access on innovation; Jedoch,
we ﬂexibly control for the impact of import competition on
Innovation.

III. The Empirical Model

Based on the theoretical framework presented above, Das
section develops our main empirical model and discusses the
identiﬁcation strategy.

8Note that tariff cuts will also affect the price index and therefore ˆBin.
Our empirical approach will capture both the direct impact of ˆτ jmn and the
indirect impact of ˆBin.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

208

THE REVIEW OF ECONOMICS AND STATISTICS

Online appendix C shows that equation (2) can be approx-

imated by

(cid:3) ln Ki = β(cid:3) ¯Ti + εi,

(cid:3) ¯Ti ≡

(cid:4)

ωin(cid:3)Tjmn

(3)

(4)

(cid:13)
N

is the weighted average of tariff changes across all of
ﬁrm i’s export markets, Tjmn is the ad valorem tariff
that country n levies on exports from country m in in-
dustry j, β ≡ (1 − σ) / [k − (σ − 1)], and εi ≡ [k − (σ −
1)]−1
ωin(cid:3) ln Bin. We proceed with this approximation
because it is empirically more convenient to work with.9
According to our framework, we expect that the knowledge
stock is changing when weighted average tariffs in export
markets decline or when weighted average demand (εi) rises.
As demand shocks are unobserved in our data, εi will enter
into the regression residual.

Endogenous tariffs. A potential concern is that (cid:3)Tjmn (Und
(cid:3) ¯Ti) is endogenous because ﬁrms can lobby for improved
market access through bilateral or regional trade negotiations.
We solve this by instrumenting (cid:3)Tjmn (Und (cid:3) ¯Ti) mit dem
applied MFN tariff rate cut (cid:3)T MF N
(see section IV), so that
the instrumental variable is

(cid:4)

(cid:3) ¯T MF N
ich

≡

ωin(cid:3)T MF N

The intuition for our instrument is as follows. The applied
MFN tariff rate of country n is the rate that applies to all
countries except the ones n has signed a trade agreement
mit. Als solche, it is unlikely that a ﬁrm i from country m has
any inﬂuence over the MFN tariff of country n.

Sample period. The years 1992 Zu 2000 are deﬁned as our
baseline sample period. daher, the change in average tar-
iffs facing ﬁrm i is (cid:3) ¯Ti = ¯Ti2000 − ¯Ti1992, and the change in the
knowledge stock of ﬁrm i is (cid:3) ln Ki = ln Ki2000 − ln Ki1992.
The choice of sample period is motivated by the fact that tariff
reductions agreed on during the Uruguay Round were gradu-
ally phased in from 1995 Zu 2000. In the data, we also observe
tariff cuts before 1995; starting our sample in 1992 ensures
that we capture the full impact of tariff reductions.10 Our data
also conﬁrm that the 1990s were unique: the overall reduction
in tariffs was much greater during the latter half of the 1990s
compared to both earlier and later periods (see ﬁgure 2). Fi-
schließlich, we choose to work with long differences, 1992 Zu 2000,
in our baseline speciﬁcation because we want to allow for
long time lags in the innovation response to trade liberaliza-

tion. Long differences also eliminate serial correlation in the
Fehler, since the averaging over periods ignores time-series
Information (see Bertrand, Duﬂo, & Mullainathan, 2004).

Outcome variable.

In the model presented above, the out-
come variable (cid:3) ln Ki is the change in the log knowledge
stock. Our empirical counterpart is the cumulative patent
count of a ﬁrm i until year t,

Kit ≡

T(cid:4)

s=1965

pis,

(5)

where pis is the number of unique granted patents ﬁled by ﬁrm
i in year s. The outcome variable (cid:3) ln Kit gives the change in
the log cumulative patent count between 1992 Und 2000 Und
provides a measure of the innovation that takes place during
this time period. Focusing on the change in the stock over
a long time period smooths out lumpiness and 0s in the pit
Variable. In der Tat, in a given year, the median pit is 0 while
the maximum pit is very large, suggesting that linear models
are not adequate to model the data-generating process at the
annual level.

ich

Econometric concerns. Estimating equation (3) is chal-
lenging for a number of reasons. The ﬁrst econometric con-
cern is that the weighted average tariff reduction (cid:3) ¯Ti (Und
the instrument (cid:3) ¯T MF N
) may be correlated with unobserv-
able ﬁrm characteristics. Zum Beispiel, ﬁrms exposed to high-
tariff-reduction countries may innovate more even in the ab-
sence of trade liberalization. We address this in three ways.
Erste, we present a falsiﬁcation test regressing knowledge
growth during the 1980s on the same 1990s market access
Variable (cid:3) ¯Ti. Section VC shows that the estimated coefﬁ-
cient in this case becomes close to 0, suggesting that there
are no pretrends driving our results.

Zweite, we address the concern by including home
country-industry pair ﬁxed effects (η jm) in the regressions as
well as controlling for a vector of preperiod ﬁrm characteris-
Tics (Ci).11 Intuitively, we compare ﬁrms within the same nar-
rowly deﬁned industry, with the same home country, and with
similar observed characteristics during the preperiod, but that
differ in terms of their exposure to international markets, Und
we ask whether ﬁrms exposed to high-tariff-cut countries in-
novate more than ﬁrms exposed to low-tariff-cut countries.
This approach also ensures that changes in the patent system
or industry-speciﬁc trends in patenting are all differenced
out. daher, our baseline approach will be based on the
estimation of

(cid:3) ln Ki = η jm + β(cid:3) ¯Ti + C(cid:5)

ich

Phi + εi.

(6)

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

9Online appendix, section C, evaluates the performance of the approxi-

mation.

10The year 1992 is also the ﬁrst year for which the tariff data used in the

analysis are available.

11Industries are deﬁned at the NACE three-digit level. Presample covari-
ates are home weights ωH
ich , the number of countries the ﬁrm is patenting
in during the preperiod, ni,Pre, and the log knowledge stock of the ﬁrm in
1985, ln Ki,Pre.

BETTER, FASTER, STRONGER

209

A third way of addressing the concern that tariff reductions
may be correlated with unobservable ﬁrm characteristics is
by differencing out idiosyncratic ﬁrm trends. Konkret, Wir
split the sample into our main sample period, (t = 1), and add
a second period, (t = 2), and estimate the equation

(cid:3) ln Ki2 − (cid:3) ln Ki1 = η jm + β

(cid:2)
(cid:3) ¯Ti2 − (cid:3) ¯Ti1

(cid:3)

+ C(cid:5)
ich

Phi + εi.
(7)

Idiosyncratic growth trends in innovation that may be corre-
lated with (cid:3) ¯Ti are then differenced out. This is reminiscent of
a triple differences model, as we compare the growth in the
change in tariffs (two differences) across ﬁrms (third differ-
enz). We choose t = 1 as the baseline period 1992 Zu 2000
and t = 2 as the years 2000 Zu 2004.

A ﬁnal concern is that the error term εi, a weighted average
of country-speciﬁc demand shocks, may be correlated with
trade liberalization. A case in point is the TRIPS agreement
that strengthened intellectual property rights (IPR) among
WTO members in the aftermath of the Uruguay Round.
A positive correlation between tariff reductions and IPR
strengthening could therefore produce biased estimates.12
We solve this by using a control function approach and the
fact that we observe aggregate patenting by industry and
country, and this measure is itself determined by the un-
observed shocks Bin. Konkret, we calculate the aggre-
gate knowledge stock by industry j and home country h,
Kh jt =
Kit , Wo (cid:4)h j is the set of ﬁrms in industry j
with h as home country, and construct the weighted average,

i∈(cid:4)h j

(cid:13)

˜εi ≡

(cid:4)

ωin(cid:3) ln Kn j,

(8)

Wo (cid:3) ln Kn j = ln Kn j2000 − ln Kn j1992. While home
country-industry pair ﬁxed effects control for innovation
trends in ﬁrm i’s home market, by adding ˜εi to our baseline
estimation equation, we control for innovation trends in ﬁrm
i’s destination markets. Zum Beispiel, if a U.S. ﬁrm primarily
exposed to the Indian market is innovating more because
the Indian market is growing quickly (hoch (cid:3) ln BiIndia),
then including ˜εi will control for this effect. An alternative
approach is to use a vector of ﬁxed effects for each of ﬁrm i’s
destination markets. We explore this approach, along with
other robustness checks, in section VC.13

12TRIPS established minimum and common standards of IP protection to
be adopted by all WTO members. While the institutions in the developed
countries were little affected due to already strong IP protection, Entwicklung
countries had to reform and strengthen their IP protection system to comply
with new WTO rules.

13A potential remaining concern is that tariff cuts in an industry may
be correlated with lowering nontariff barriers in the same industry. Unser
estimates would then reﬂect improved market access coming from both the
former and the latter. Jedoch, data on nontariff barriers at disaggregated
levels over a long-time horizon are not available, and it is therefore hard to
separate the two mechanisms.

IV. Data

A. Patents

Our main data set is based on the European Patent Ofﬁce’s
(EPO) Worldwide Patent Statistical Database (PATSTAT).14
PATSTAT offers bibliographic data, family links, and cita-
tions of 90 million applications from patentees of more than
100 Länder. It contains the population of all patents glob-
ally since the mid-1960s. The patent documents as provided
by PATSTAT are a rich source of information. We observe
the name of the applicant (patentee) and date of ﬁling, Kneipe-
lication, and if and when the patent was granted. Wir haben
information on citations, technology areas (IPC codes), Und
the research teams behind the inventions. A unique patent is
typically ﬁled in more than one country. We know the geog-
raphy of the patent in the sense that we have information on
both the home country of the patentee and the other coun-
tries in which the patent has been ﬁled. Home country is the
residence country of the applicant. Patentees can be private
business enterprises, universities or other higher education
institutions, governmental agencies, or individual inventors.
In our ﬁnal sample, 57% of patentees are ﬁrms, 40% are indi-
vidual inventors and only 3% of the patentees belong to other
types. We use all types in the analysis but commonly refer
to them as ﬁrms throughout the paper.15 More details on the
sample and the construction of the data set are provided in
online appendix D, while Abramovsky et al. (2008) provide
a thorough review of the PATSTAT data and the patenting
Verfahren.

Firm-speciﬁc knowledge stocks. PATSTAT allows us to
construct an international ﬁrm-level data set and to follow
the patenting activity of a ﬁrm through time. To measure the
innovation activity of a ﬁrm i in year t, we use the number
of granted patents dated by the earliest ﬁling year, pit .16 Dat-
ing the patents by application ﬁling date is conventional in
the empirical innovation literature as it is much more closely
timed with when the R&D process took place than the patent
publication and grant date.17

Patenting is known to be highly correlated with innovation
and R&D (Griliches, 1990). The advantages and limitations
of patenting as a measure of innovation have been extensively
discussed.18 For our purpose, there is one major advantage

14The April 2015 Ausführung.
15We expect that individual inventors, who may typically be entrepreneurs
about to start a new business, also respond to a change in foreign market
Zugang. daher, we choose not to limit the sample just to the patentees
registered as ﬁrms.

16Not all ﬁled patents are granted. We limit the analysis to patents that
are granted to account for differences in quality. To be granted a patent, ein
innovation must satisfy three key criteria: it must be novel or new, it must
involve an inventive step, and it must be industrially applicable.

17Patent applications are usually published 18 months after the ﬁrst

application.

18See OECD (2009), Griliches (1990), and Nagaoka, Motohashi, and Goto
(2010) for reviews and discussion of patent data as innovation indicators.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

210

THE REVIEW OF ECONOMICS AND STATISTICS

of using patents: they are the only source of information that
allows for a comprehensive ﬁrm-level analysis of innovation
at a global scale. In section VB, we use different measures to
control for the quality of patents as innovation indicators.

In our analysis, a patent corresponds to a unique invention,
so ﬁling the same patent in multiple locations does not in-
ﬂate the patent count (pit ). Konkret, PATSTAT organizes
patents into “patent families” that identify identical inven-
tions ﬁled in multiple countries.19 An additional advantage
of PATSTAT is that names of applicants are harmonized over
the entire sample period, alleviating the concern that slight
differences in the spelling of patentee names generate multi-
ple patentee IDs.

Firm-speciﬁc weights. The empirical analysis relies on
observing the ﬁrm-destination speciﬁc weights ωin. Diese
weights reﬂect the relative importance of a country n in the
ﬁrm’s total proﬁts. Proﬁts and sales are unobserved in our
patent data, but we do observe in which markets a ﬁrm is
patenting. As Aghion et al. (2016) pointed out, a patent-based
weighting scheme may potentially be a superior measure be-
cause it reﬂects the ﬁrms’ expectations of where their future
market will be. We calculate these weights based on patent
ﬁlings over the preperiod years 1965 Zu 1985. Wir gebrauchen 1965 als
the starting year because the number of patents in PATSTAT
is limited in earlier years. The ﬁnal year of 1985 was cho-
sen because the Uruguay Round negotiations started in 1986;
somit, the weights are not themselves affected by trade lib-
eralization of the 1990s. Konkret, we deﬁne

ωin ≡ xin(cid:13)
k xik

(9)

where xin is the number of patents issued by ﬁrm i in market
n during the preperiod. Seeking intellectual property rights
in a country is typically motivated by (future) proﬁts in that
Markt. There is strong empirical support that patent weights
are highly correlated with sales weights (see Aghion et al.,
2016). We provide additional empirical evidence on this in
online appendix G. The weights are also remarkably persis-
tent over time, even over a period of 20 Jahre; see online
appendix H. This suggest that time-invariant ﬁrm and coun-
try characteristics (z.B., country-speciﬁc entry costs on the
supply side or idiosyncratic taste differences on the demand
Seite) are limiting where ﬁrms export goods and ﬁle patents.

Firm characteristics.

Information about patentees in PAT-
STAT is restricted to what can be retrieved from the patent
applications. Our basic ﬁrm characteristics are industry af-
ﬁliation (NACE Rev. 2 three-digit) and home country of the
ﬁrm. Industry afﬁliation is assigned based on the technology
Bereich (IPC codes) of the patents ﬁled by a ﬁrm. See online
appendix D for more details.

19We use the DOCDB patent family.

B. Tariffs

jmt

The main source of tariff data is the UNCTAD Trade Anal-
ysis and Information System (TRAINS), which contains tar-
iffs at the most disaggregated level of the Harmonized System
(HS) for more than 150 Länder. From this database, we ex-
tract the average applied MFN industry-level tariff (NACE
three-digit) for the period 1992 Zu 2004, T MF N
, mit 1992
being the ﬁrst year for which a complete data set is available.
We use these to calculate the ﬁrm-speciﬁc weighted average
MFN tariffs, ¯T MF N
, which vary across ﬁrms, both because
ﬁrms are exposed to different markets and because they be-
long to different industries. Online appendix E describes the
procedure followed to calculate industry-level tariffs, while
the online appendix F provides details about the historical
background for tariff reductions during the 1990s.
The bilateral tariff, Tjmnt , is calculated as T MF N

× RT Amnt ,
where RT Amnt takes the value 0 if there exists a trade agree-
ment between m and n in year t and 1 ansonsten. Our measure
is an approximation of the true bilateral tariff; for some coun-
try pairs (z.B., EU countries), Tjmnt will equal the true bilateral
tariff, while for others (z.B., the United States and Mexico),
Tjmnt will be measured with error because trade agreement
tariffs are not always 0. We use Tjmnt to calculate the ﬁrm-
speciﬁc weighted average tariffs ¯Ti. The information on RTAs
comes from the comprehensive data set on RTAs that is part
of the CEPII gravity data.20

jnt

C. Final Sample of Firms

Our point of departure is a data set constructed on the basis
of PATSTAT described in section IVA and online appendix D
matched with industry-level tariff from UNCTAD TRAINS
and information on regional trade agreements (RTAs).21 Aim-
ing to investigate the impact of the Great liberalization of the
1990S, our point of departure is the sample of ﬁrms resid-
ing in WTO countries with patent activity during the sample
Zeitraum, 1992 Zu 2000. We refer to this as the initial sample.
During the 1992–2000 period, there were 763,581 ﬁrms ﬁling
3,644,556 granted patents. Firms from WTO member coun-
tries were responsible for 85% of global patenting over this
Zeitraum.

Our empirical approach investigates the impact of market
access at the intensive margin. Somit, we need to limit the
analysis to ﬁrms that existed by 1992. The empirical strategy,
darüber hinaus, requires information about the ﬁrms’ patent ﬁling
abroad in the preperiod, as well as the trade barriers they face
in their foreign markets. Given these requirements, we con-
struct the ﬁnal sample that forms the basis for the empirical
Analyse. The ﬁnal sample consists of ﬁrms that (A) resided in
countries that had become members of WTO (GATT) prior to
1995; (B) had applied for at least one granted patent by 1992

20See Head, Mayer, and Ries (2010) and Head and Mayer (2014) für

details on the data set.

21We drop ﬁrms for which industry or home country afﬁliation were

missing.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

BETTER, FASTER, STRONGER

211

TABLE 1.—INITIAL VERSUS FINAL SAMPLE

(cid:13)
ich
(cid:3)Kit

(cid:3)Kit

Mean
Median
Standard deviation

Number of ﬁrms

Initial

3,644,556

4.77
1
143.57
763,581

Final

663,252

19.60
2
329.63
41,058

The table shows the aggregate increase in the knowledge stock from 1992 Zu 2000 along with the mean,

median, and standard deviation of (cid:3)Kit for the initial and the ﬁnal sample of ﬁrms.

to ensure that the ﬁrm exists at the beginning of the sample
Zeitraum); (C) had been observed at least once in the preperiod
(1965–1985) in order to be assigned weights ωin;22 (D) hatte
patent activity outside their home country and thus a positive
weight ωin in at least one foreign country; Und (e) had issued
patents only in countries where tariff data for their industry
and export market is available.

The ﬁnal sample consists of 41,058 ﬁrms, mit 663,252
unique patents being granted between 1992 Und 2000. Die fi-
nal sample captures roughly one-ﬁfth of total patenting in the
WTO over the sample period, 1992 Zu 2000. Our ﬁnal sample
consists of ﬁrms from 65 different countries and 54 anders
industries. Tisch 1 summarizes the difference between the
initial and ﬁnal samples, comparing total patenting in WTO
countries to total patenting in the ﬁnal sample. The reduction
in the number of ﬁrms and patents in the ﬁnal sample relative
to the initial sample is primarily driven by restrictions (B) Und
(C); das ist, the analysis is limited to ﬁrms that already exist.
Tisch 12 in the online appendix shows additional moments
from the initial and ﬁnal samples. Wichtig, der Median
number of citation-weighted patents in the initial sample is 0.
This suggests that that majority of patents in the initial sam-
ple have negligible economic value; Außerdem, the ﬁnal
sample is more representative of patents that generate eco-
nomic value. Online appendix K provides additional details
on countries and industries in the ﬁnal sample and descrip-
tives that show that the ﬁnal sample is representative along
these dimensions.

Note that we cannot distinguish between ﬁrm exit and
zero innovation in our data. Zum Beispiel, if we observe zero
patenting from 1995 and onward, then the stock of patents,
Kit , will be constant for the remaining years of our sample.
daher, our baseline result will capture the impact of mar-
ket access on both the intensive margin (change in innovation
among continuing ﬁrms) and extensive margin (ﬁrms that
stop innovating).

D. Descriptives

Figur 1 shows the distribution of ﬁrms across home coun-
tries and industries (NACE 2-digit) in our sample. We note the
dominance of Japan and the United States and by the indus-

22Both granted and nongranted patents are used for the construction of
the weights ωin. The weights reﬂect expectations on future markets. Dort-
Vordergrund, it is the action of seeking intellectual property protection in a foreign
country that is relevant rather than the ﬁnal outcome of the application
Verfahren.

tries machinery and equipment (28), computers, electronic
and optical products (26), and other manufacturing (32).
Tables 15 Und 16 in online appendix K provide more de-
tails on patent counts and patenting ﬁrms across industries
and countries.

Figur 2 shows the mean weighted average MFN tariffs,
¯T MF N
, for ﬁrms with the United States, Deutschland, Japan, Und
Es
the United Kingdom as home country. There is a strong de-
cline during the latter half of the 1990s; the average ﬁrm ex-
perienced a decline in weighted tariffs of around 3 Prozentsatz
points during the 1990s. Auch, the decline almost stops in the
Jahr 2000, consistent with the fact that Uruguay Round con-
cessions were phased in until that year. The averages mask
a considerable amount of heterogeneity. Figur 3 zeigt, dass
the whole distribution of weighted tariffs across ﬁrms shifts
markedly to the left from 1992 Zu 2000. We summarize the
data in table 2, which shows the median, median, and stan-
dard deviation of ¯T MF N
, ¯Tit , and the log knowledge stock,
ln Kit , im Laufe der Zeit.

V. Ergebnisse

Innovation and Trade Liberalization

ich

We proceed by estimating the model presented in equation
(6) and the alternative speciﬁcation equation (7) using 2SLS.
All speciﬁcations include home country-industry (NACE
three-digit) pair ﬁxed effects, which will control for aggregate
(country and industry) trends in patenting. Columns 1 Zu 3 In
table 3 show the results for our baseline speciﬁcation. As de-
scribed above, we instrument the tariff cut variable (cid:3) ¯Ti with
applied MFN tariff cuts (cid:3) ¯T MF N
. Column 1 has only ﬁxed ef-
fects, and column 2 adds presample ﬁrm characteristics (Die
home weight, ωiH , the number of countries the ﬁrm is patent-
ing in during the preperiod, ni,Pre, and log knowledge stock
In 1985, ln Ki,Pre), while column 3 also controls for aggre-
gate destination trends ˜εi, as explained in section III. Column
4 presents the results for the model described in equation
(7), where we difference out idiosyncratic ﬁrm trends. Der
results are highly signiﬁcant across speciﬁcations, with an
estimated coefﬁcient in the range of −1.68 to −3.22. Diese
results strongly suggest that foreign market access leads to
signiﬁcantly higher innovation.

Tisch 13 in online appendix J presents the estimated coef-
ﬁcients on the various controls, ﬁrst-stage estimates as well
as reduced-form results of the speciﬁcation with a complete
set of controls. The instrument is strongly correlated with the
endogenous variable. Figur 10 in the online appendix shows
a binned scatterplot between the dependent variable and the
instrument.

A semilog elasticity of −2.6 (in column 3) implies that a 1
percentage point reduction in tariffs causes a 2.6% increase in
the knowledge stock of a ﬁrm over a period of eight years.23

23Acemoglu and Linn (2004) ﬁnd that a 1% increase in potential market
size leads to approximately a 4% increase in the entry of new nongeneric
drugs.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

212

THE REVIEW OF ECONOMICS AND STATISTICS

FIGURE 1.—SHARE OF PATENTING FIRMS BY COUNTRY AND INDUSTRY

The ﬁgure shows share of ﬁrms by home country and NACE Rev. 2 two-digit industry for the period 1992 Zu 2000. Only the top ten countries and industries are shown.

FIGURE 2.—AVERAGE FIRM-SPECIFIC TARIFFS, ¯T MF N

The ﬁgure shows the annual average ¯T MF N

across ﬁrms according to home country.

FIGURE 3.—DENSITY OF FIRM-SPECIFIC TARIFFS, ¯T MF N

, IN 1992 AND 2000

For expositional purposes, the histogram is truncated at ¯T MF N

= 20.

As a simple back-of-the-envelope exercise, we ask how large
our estimates are compared to the mean growth in the knowl-
edge stock over the sample period. Our data show that over
the period 1992 Zu 2000, the mean knowledge stock globally

grew by 41%, while the mean reduction in the ﬁrm-speciﬁc
tariff measure was 2 percentage points (mean of (cid:3) ¯Ti). Our re-
sults therefore suggest that roughly 13% (2.6 × 2/41) of the
observed increase in the knowledge stock can be explained by
improved market access induced by trade liberalization. Der
back-of-the-envelope exercise holds all general equilibrium
outcomes ﬁxed; das ist, the assumption is that the industry and
country ﬁxed effects do not themselves change in response to
trade policy. General equilibrium responses are likely large,
and therefore this exercise cannot identify the aggregate im-
pact of the tariff cuts on innovation (which is outside the
scope of this paper).

Is Patenting a Good Measure of Innovation?

As noted in in section IVA, one may argue that patents are
an imprecise measure of knowledge and innovation. Patent-
ing is not the only way to protect innovations. Another prob-
lem is that patent quality is highly heterogeneous. Nach
to Nagaoka et al. (2010), roughly half of the patents owned
by a ﬁrm are used either by that ﬁrm internally or licensed to
Andere. The remaining patents are used for strategic reasons,
such as attempts to block inventions by competitors. Dort-
Vordergrund, it is possible that ﬁrms take out more patents without
innovating more, Zum Beispiel, in response to import compe-
tition. If this were the case, one would expect that ﬁrms are
taking out patents on their marginal innovations, so that the
average quality of their patent stock is decreasing.

To address this issue, we use four proxies for patent quality:
the number of citations, the size of the research teams behind
a patent, the number of technology areas (IPC codes) to which
a patent is attributed (patent breadth), and family size. Family
size refers to the number of markets in which a patent is ﬁled.
We use citations because high-value inventions are more ex-
tensively cited than low-value patents (Harhoff et al., 1999).
We include the size of research teams since a set of studies
has associated the number of inventors listed in a patent with
the economical and technological value of the patent (OECD,

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

BETTER, FASTER, STRONGER

TABLE 2.—MEAN, MEDIAN, AND STANDARD DEVIATION OF ln Kit , ¯T MF N

, AND ¯Tit

1992
2000
2004

Mean

1.41
1.83
1.98

ln Kit

Median

1.10
1.61
1.79

1.26
1.33
1.33

Mean

0.049
0.029
0.026

¯T MF N
Es

Median

0.043
0.021
0.019

0.046
0.036
0.032

Mean

0.032
0.019
0.016

¯Tit

Median

0.020
0.010
0.010

213

0.047
0.033
0.028

TABLE 3.—MARKET ACCESS AND KNOWLEDGE CREATION (2SLS)

(cid:3) ln Ki
(1)

(cid:3) ln Ki
(2)

(cid:3) ln Ki
(3)

(cid:3) ln Ki2 − (cid:3) ln Ki1
(4)

Dep. Variable:

Change in tariff ((cid:3) ¯Ti)

Dep. Variable: (cid:3) ln ¯Qi
Change in tariff ((cid:3) ¯Ti)

Home country-industry FE
Firm controls
Destination market controls ( ˜εi)
1st stage F -statistic
Number of ﬁrms
Standard errors clustered by home country-industry in parentheses. Firm controls are presample ﬁrm characteristics: the home weight, ωiH ; the number of countries the ﬁrm is patenting in during the preperiod,
. Destination market controls controls for industry-speciﬁc innovation trends in a

ni,Pre; and log knowledge stock in 1985, ln Ki,Pre. The change in tariffs (cid:3) ¯Ti is instrumented with the change in MFN tariffs, (cid:3) ¯T MF N
ﬁrm’s destination markets. ∗∗∗ p < 0.01, ∗∗ p < 0.05, and ∗ p < 0.1. i −2.81*** (.47) Yes No No 2,372 41,058 −.68* (.37) Yes Yes Yes 774 26,886 −3.22*** (.43) Yes Yes No 1,204 41,058 −.30*** (.09) Yes Yes Yes 1,344 37,539 −2.62*** (.45) Yes Yes Yes 1,029 40,805 −.16* (.10) Yes Yes Yes 1,030 40,747 −1.68*** (.35) Yes Yes Yes 600 40,805 .16** (.07) Yes Yes Yes 1,029 40,805 TABLE 4.—MARKET ACCESS AND INNOVATION QUALITY (2SLS) Citations (1) Research Team (2) IPC Codes (3) Number of Markets (4) Home country-industry FE Firm controls Destination market controls ( ˜εi) 1st stage F -statistic Number of ﬁrms Standard errors clustered by home country-industry in parentheses. Firm controls are presample ﬁrm characteristics: the home weight, ωiH ; the number of countries the ﬁrm is patenting in during the preperiod, . Destination market controls controls for industry-speciﬁc innovation trends in a ni,Pre; and log knowledge stock in 1985, ln Ki,Pre. The change in tariffs (cid:3) ¯Ti is instrumented with the change in MFN tariffs, (cid:3) ¯T MF N ﬁrm’s destination markets. ∗∗∗ p < 0.01, ∗∗ p < 0.05, and ∗ p < 0.1. i 2009). The number of technical classes attributed to a patent application (patent breadth) has been found approximate the value of a patent portfolio (Lerner, 1994). We calculate average quality of the knowledge stock as follows. Let qp denote the number of citations three years after a patent p was ﬁled, the number of inventors, the number of IPC codes, or the family size associated with patent p. The cumulative sum is then Qit = t(cid:4) (cid:4) qp, s=1965 p∈(cid:5)is (10) where (cid:5)is is the set of ﬁrm i’s patents ﬁled in year s. The average quality of the knowledge stock is then calculated as ¯Qit = Qit /Kit . We proceed by using (cid:3) ln ¯Qi = ln ¯Qi2000 − ln ¯Qi1992 as the dependent variable and estimate our baseline model again. The results using all four proxies for quality are reported in columns 1 to 4 of table 4.24 The results suggest that market access did not affect the quality of patents, that is, there is no evidence of a lawyer effect. If anything, the point estimates indicate that better market access may have increased the quality of patents, since it increased the average number of citations, average size of research teams, and patent breadth. C. Robustness Falsiﬁcation test. A potential concern is that ﬁrms be- ing exposed to countries with high tariff cuts always have higher patent growth compared to other ﬁrms. To address this concern, we perform a placebo test and regress knowl- edge growth during the 1980s, ln Ki1988 − ln Ki1980, on trade policy changes during the 1990s, (cid:3) ¯Ti2000 − (cid:3) ¯Ti1992.25 The results are shown in the ﬁrst column of table 5: the coefﬁ- cient of interest is precisely estimated around 0, suggesting that there are no differential pretrends in patenting. Triadic patents. We restrict our sample to triadic patents. These are patents ﬁled at the three main patent ofﬁces: the 24The number of ﬁrms in the sample decreases when we use citations as a measure, since some ﬁrms have portfolios of patents that are never cited. 25The weighted average ¯Tit is now calculated using weights ωin based on a ﬁrm’s patent portfolio until 1980. This ensures that the weights ωin are not themselves determined by the dependent variable ln Ki1988 − ln Ki1980. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 1 0 4 2 2 0 5 1 9 9 6 1 6 5 / r e s t _ a _ 0 0 9 5 1 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 Destination Trends (5) −1.88*** (.42) Input Tariffs −2.87*** (.51) .60 (.76) .02 (.01) Yes Yes Yes No 508 37,980 214 THE REVIEW OF ECONOMICS AND STATISTICS Dep. Variable: (cid:3) ln Kit Change in tariff ((cid:3) ¯Ti) Change in home tariff Change in input tariff TABLE 5.—ROBUSTNESS (2SLS) Triadic Patents (2) −4.65* (2.66) Alternate Dependent Variable (3) −2.18*** (.54) Placebo (1) .004*** (.001) Import Competition (4) −2.87*** (.51) 1.89*** (.48) Home country-industry FE Firm controls Destination market controls ( ˜εi) Destination country trends 1st stage F -statistic Number of ﬁrms Standard errors clustered by home country-industry in parentheses. Firm controls are presample ﬁrm characteristics: the home weight, ωiH ; the number of countries the ﬁrm is patenting in during the preperiod, . Destination market controls controls for industry-speciﬁc innovation trends in a Yes Yes Yes No 2,262 29,320 Yes Yes No Yes 2,027 41,058 Yes Yes Yes No 563 37,980 Yes Yes Yes No 626 21,568 Yes Yes Yes No 833 2,405 ni,Pre; and log knowledge stock in 1985, ln Ki,Pre. The change in tariffs (cid:3) ¯Ti is instrumented with the change in MFN tariffs, (cid:3) ¯T MF N ﬁrm’s destination markets. ∗∗∗ p < 0.01, ∗∗ p < 0.05, and ∗ p < 0.1. i European Patent Ofﬁce (EPO), the Japanese Patent Of- ﬁce (JPO), and the U.S. Patents and Trademark Ofﬁce (USPTO).26 Triadic patents are commonly used in the lit- erature to retain only highly valuable inventions, and they provide a measure of innovation that is robust to administra- tive idiosyncrasies of the various patent ofﬁces. However, by limiting the analysis to triadic patents, the number of obser- vations is reduced by around 94%. The results are shown in in column 2 of table 5. While we observe that the sample size is reduced to around 2,300 observations, our results on the im- pact of trade liberalization on the change in knowledge stock nevertheless remain signiﬁcant, and the magnitude is close to the double as we limit our analysis to these presumably highly valuable inventions. Alternative dependent variable. The change in log knowl- edge stock, (cid:3) ln Ki, was used as the main outcome variable throughout the paper. An alternative measure that is often used in the literature is the log number of patents over the sample period 1992 to 2000, ln (cid:3)Ki. Column 3 of table 5 shows that the results are similar to the baseline estimates. (cid:13) n Import competition. Tariff cuts also heighten import com- petition in ﬁrms’ home markets. The impact of import compe- tition is largely controlled for by the industry-country ﬁxed effect η jm. As an additional robustness check, we include ωin(cid:3)Tjnm, that is, the weighted av- the control variable erage import tariff in ﬁrm i’s home market, using the same weights ωin. The results are shown in column 4 of table 5. We observe that adding the control for import tariffs does not affect our baseline results. The results suggest that increased import competition had the opposite impact relative to im- proved market access: greater import competition reduced innovation.27 26See Dernis and Khan (2004) and Martinez (2010) for additional infor- mation about how triadic patent families are constructed. 27Our result on the impact of import competition is in line with the ﬁndings of Autor et al. (2019), while to some extent contrary to Bloom et al. (2016). k Input tariffs. We also repeat the exercise above using data on the ﬁrms’ input tariff. Speciﬁcally, we calculate the input tariff of each industry and country-pair as T Inp = (cid:13) jnm ξk jTknm, where ξk j is the input cost share of industry j from sector k. The cost shares ξk j are from the World Input- Output Database (WIOD), using U.S. data from the year 2000.28 The input tariff variable for a ﬁrm i in country m and industry j is then calculated as jnm . The results are shown in column 5 of table 5. The input tariff variable is not signiﬁcantly different from 0. We observe that the cor- relation between the import and input tariff variable is high (0.82), which may explain why the variables are imprecisely estimated. ωin(cid:3)T Inp (cid:13) n Destination country trends. The variable ˜εi was included in the regressions to capture patenting trends in destination countries. An alternative empirical strategy is to include des- tination country ﬁxed effects in the regressions and estimate (cid:3) ln Ki = η jm + β(cid:3) ¯Ti + +C(cid:5) i φ + (cid:4) n∈(cid:6)i γn + εi, (11) where γn is a ﬁxed effect for destination n, and we sum over the set of countries (cid:6)i where the ﬁrm has nonzero weights during the preperiod. As an example, if all ﬁrms exposed to the Indian market (but not necessarily with In- dia as home country) have high (cid:3) ln Ki, then this will be controlled for by γIndia. Identiﬁcation of β then only comes from within-country, across-industry variation in tariffs— that among ﬁrms exposed to the Indian market, some ex- perience greater tariff reductions because they belong to an industry getting large tariff cuts in India. Destination country trends will therefore control for the possibility that ﬁrms ex- posed to India may patent more because of unobserved factors speciﬁc to India (e.g., growth in market size or strengthening 28We use a correspondence to convert ISIC industries to two-digit NACE revision 2 codes. Tariffs for nontraded input sectors k are normalized to 0. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 1 0 4 2 2 0 5 1 9 9 6 1 6 5 / r e s t _ a _ 0 0 9 5 1 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 BETTER, FASTER, STRONGER 215 TABLE 6.—MARKET ACCESS AND KNOWLEDGE CREATION BY QUARTILE OF COUNTRIES’ CHARACTERISTICS (2SLS) TABLE 7.—MARKET ACCESS AND KNOWLEDGE CREATION BY QUARTILE OF FIRM INNOVATIVENESS (2SLS) Dep. Variable: (cid:3) ln Ki Change in tariff ((cid:3) ¯Ti) (cid:3) ¯Ti × Q2 h (cid:3) ¯Ti × Q3 h (cid:3) ¯Ti × Q4 h Home country-industry FE Firm controls Destination market controls ( ˜εi) 1st stage F -statistic Number of ﬁrms GDP per capita (2) Trade Intensity Patent Stock (% of GDP) (3) (Km,Pre) (4) 5.02* (2.92) −3.01 (3.01) −7.36** (2.92) −7.96*** (2.96) Yes Yes Yes 85-9819 39,679 −3.28*** (.59) 1.18* (.67) 2.34*** (.81) 2.63*** (.84) Yes Yes Yes 30-647 39,679 52.28*** (.45) −48.53*** (2.20) −54.28*** (.65) −54.97*** (.67) Yes Yes Yes 205-572 40,805 Standard errors clustered by home country-industry in parentheses. Firm controls are presample ﬁrm characteristics: the home weight, ωiH ; the number of countries the ﬁrm is patenting in during the preperiod, ni,Pre; and log knowledge stock in 1985, ln Ki,Pre. The change in tariffs (cid:3) ¯Ti is instrumented with the change in MFN tariffs, (cid:3) ¯T MF N . Destination market controls controls for industry-speciﬁc innovation trends in a ﬁrm’s destination markets. ∗∗∗ p < 0.01, ∗∗ p < 0.05, and ∗ p < 0.1. i of IPR). The estimated coefﬁcient in column 6 in table 5 shows that β is still highly signiﬁcant. D. Heterogeneity Finally, we explore potentially heterogeneous effects across countries and ﬁrms. We start by exploring country heterogeneity by specifying (cid:3) ln Ki = η jm + β(cid:3) ¯Ti + 4(cid:4) (cid:2) βq q=2 (cid:3) (cid:3) ¯Ti × Qq h + C(cid:5) i φ + εi, (12) where Qq n refers to the qth quartile and h to the type of charac- teristic that is investigated. We consider three country char- acteristics: income per capita, trade intensity, and innovative- ness. As measures of these, we use GDP per capita, export plus import in % of GDP, and country-level patent stock, re- spectively. While the latter is based on our own calculations, we use data provided by the World Bank for the measures of GDP per capita and trade intensity.29 We split countries into quartiles based on the relevant char- acteristic. All quartiles are calculated using 1986 data. We estimate equation (12) by 2SLS using the same approach as in the baseline and report the results in table 6. Income. Focusing on GDP per capita, we ﬁnd that the ef- fect of market access is imprecisely estimated in the lowest two income quartiles. The effect is positive and signiﬁcant in the third and fourth quartiles of the income distribution, sug- gesting that market access has a stronger impact on innovation in developed countries compared to developing countries.30 Dep. Variable: Change in tariff ((cid:3) ¯Ti) (cid:3) ¯Ti × Q2 i jn (cid:3) ¯Ti × Q3 i jn (cid:3) ¯Ti × Q4 i jn Home country-industry FE Firm controls Destination market controls ( ˜ε) 1st stage F -statistic Number of ﬁrms Initial Patent Stock Quality Adjusted Initial Patent Stock −2.70*** (.41) .76 (.50) −.67 (.46) .76 (.94) Yes Yes Yes 140-648 40,805 −.83*** (.25) −1.02 (.66) .28 (.62) 2.39*** (.72) Yes Yes Yes 278-449 38,671 Standard errors clustered by home country-industry in parentheses. Firm controls include dummies for the second, third, and fourth quartile of the ﬁrm size distribution and presample ﬁrm characteristics: the home weight, ωiH ; the number of countries the ﬁrm is patenting in during the preperiod, ni,Pre; and log knowledge stock in 1985, ln Ki,Pre. The change in tariffs (cid:3) ¯Ti is instrumented with the change in MFN tariffs, (cid:3) ¯T MF N . Destination market controls for industry-speciﬁc innovation trends in a ﬁrm’s destination markets. ∗∗∗ p < 0.01, ∗∗ p < 0.05, and ∗ p < 0.1. i Trade intensity. The effect of market access on innovation is strongest in the lowest quartile in terms of trade intensity and decreases in higher quartiles (Q1>Q2>Q3>Q4). Das
suggests that market access has a bigger impact on innovation
on countries that are initially relatively closed to trade.

Innovativeness. The results on innovativeness mirror the
results on GDP per capita; for the two ﬁrst quartiles, im-
proved market access has a zero or negative impact on inno-
vation, while for the third and fourth quartiles, the impact is
positive.

Nächste, we consider the role of ﬁrm-level heterogeneity fo-
cusing on ﬁrm innovativeness. We split ﬁrms into quartiles
of ﬁrm innovativeness, measured relative to industry coun-
try average innovativeness, letting Qq
n refer to the qth quar-
tile and h to the type of measure. We use two measures of
innovativeness: ﬁrms’ initial (1985) patent stock relative to
industry-country average and ﬁrms’ quality-adjusted initial
patent stock relative to industry-country average.31 We pro-
ceed by estimating

(cid:3) ln Ki = η jm + β (cid:3) ¯Ti +

(cid:14)

(cid:15)

βq

(cid:3) ¯Ti × Qq
i jn

4(cid:4)

q=2

4(cid:4)

q=2

δq × Qq
i jn

+ C(cid:5)
ich

Phi + εi,

(13)

based on the same IV approach as in the baseline and report
the results in table 7. The effect of greater market access is
positive and signiﬁcant for ﬁrms in all quartiles. The impact

29GDP per capita (constant 2010 US$) with indicator ID: NY.GDP.
PCAP.KD, and Trade (% of GDP) with indicator ID: NE.TRD.GNFS.ZS.
30The results in the literature on developed versus developing countries

are mixed; see Steinwender and Shu (2019).

31We calculate the quality-adjusted patent stock for each ﬁrm using cita-
tions as a measure of patent quality as follows: Ci jnPre/C jnPre, where Ci jnPre
is ﬁrms’ citation-weighted patent stock in the preperiod and C jnPre is the
average citation-weighted patent stock in ﬁrms’ industry and country.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

216

THE REVIEW OF ECONOMICS AND STATISTICS

appear to be stronger in the ﬁrst and third quartiles; Jedoch,
the standard errors for the interaction terms are relatively
groß.

When we measure innovativeness in terms of quality-
adjusted patent stock, the effect of improved market access
appears to have the strongest impact on innovation in the
lowest two quartiles.

VI. Abschluss

We set out to analyze the impact of improved market access
facilitated by trade agreements on worldwide innovation. To
do so, we use the decline in tariffs during the 1990s in the
aftermath of the GATT Uruguay Round and a comprehensive
global data set of patenting. Our results show that the Great
Liberalization of the 1990s had a large, positive net impact
on innovation. Our results indicate that a 1 percentage point
tariff cut in export markets leads to 2% Zu 3% growth in
ﬁrms’ knowledge stock, suggesting that trade policy was an
important factor driving global innovation in the 1990s. Unser
ﬁndings underscore the importance of trade liberalization for
ﬁrms’ long-term performance and for aggregate economic
Wachstum. They point to the large, dynamic gains from trade—
gains that are typically not observed and therefore neglected
in empirical analyses.

Our estimates are robust to a set of econometric issues, Und
in particular we provide evidence in support of patents as a
useful measure of innovation. While the results are directly
relevant for the analysis of trade policy, they also add to the
broader literature on economic factors that govern innovation
und Wachstum.

VERWEISE
Abramovsky, L., R. Grifﬁth, G. Macartney, and H. Müller, “The Location of
Innovative Activity in Europe,” Institute for Fiscal Studies working
Papier 08/10 (2008).

Acemoglu, D., and J. Linn, “Market Size in Innovation: Theory and Evi-
dence from the Pharmaceutical Industry,” Quarterly Journal of Eco-
nomics 119:3 (2004), 1049–1090. 10.1162/0033553041502144

Aghion, P., A. Bergeaud, M. Lequien, and M. Melitz, “The Heterogeneous
Impact of Market Size on Innovation: Evidence from French Firm-
Level Exports,” NBER working paper 24600 (2018).

Aghion, P., N. Bloom, R. Blundell, R. Grifﬁth, and P. Howitt, “Competition
and Innovation: An Inverted U Relationship,” Quarterly Journal of
Economics 102:2 (2005), 701–728.

Aghion, P., A. Dechezlepretre, D. Hemous, R. Martin, and J. V. Reenen,
“Carbon Taxes, Path Dependency and Directed Technical Change:
Evidence from the Auto Industry,” Journal of Political Economy
214:1 (2016).

Aghion, P., C. Harris, and J. Vickers, “Competition and Growth with
Step-by-Step Innovation: An Example,” European Economic Re-
view 41:3–5 (1997), 771–782. 10.1016/S0014-2921(97)00036-6

Amiti, M., and J. Konings, “Trade Liberalization, Intermediate Inputs, Und
Productivity: Evidence from Indonesia,” American Economic Re-
view 7:5 (2007), 1611–1638. 10.1257/aer.97.5.1611

Autor, D., D. Dorn, G. H. Hanson, G. Pisano, and P. Shu, “Foreign Com-
petition and Domestic Innovation: Evidence from U.S. Patents,”
American Economic Review: Insights 2 (2019), 357–374. 10.1257/
aeri.20180481

Bertrand, M., E. Duﬂo, and S. Mullainathan, “How Much Should We Trust
Differences-in-Differences Estimates?” Quarterly Journal of Eco-
nomics 119:1 (2004), 249–275. 10.1162/003355304772839588

Bloom, N., M. Draca, and J. Van Reenen, “Trade Induced Technical
Change? The Impact of Chinese Imports on Innovation, IT and
Productivity,” Review of Economic Studies 83:1 (2016), 87–117.
10.1093/restud/rdv039

Boler, E. A., A. Moxnes, and K. H. Ulltveit-Moe, “R&D, International
Sourcing and the Joint Impact on Firm Performance,” American Eco-
nomic Review 105:12 (2015), 3704–3739. 10.1257/aer.20121530

Bustos, P., “Trade Liberalization, Exports and Technology Upgrading:
Evidence on the Impact of MERCOSUR on Argentinean Firms,”
American Economic Review 101:1 (2011), 304–340. 10.1257/
aer.101.1.304

Calel, R., and A. Dechezlepretre, “Environmental Policy and Directed Tech-
nological Change: Evidence from the European Carbon Market,” this
REZENSION 98:1 (2016).

Dekle, R., J. Eaton, and S. Kortum, “Global Rebalancing with Gravity:
Measuring the Burden of Adjustment,” IMF Economic Review 55:3
(2008), 511–540.

Dernis, H., and M. Khan, “Triadic Patent Families Methodology,” OECD

Wissenschaft, Technology and Industry working paper 2004/2 (2004).

Estevadeordal, A., and A. M. Taylor, “Is the Washington Consensus Dead?
Growth, Openness, and the Great Liberalization, 1970s–2000s,” this
REZENSION 95:5 (2013), 1669–1690.

Goldberg, P. K., A. K. Khandelwal, N. Pavcnik, and P. Topalova, “Imported
Intermediate Inputs and Domestic Product Growth: Evidence from
Indien,” Quarterly Journal of Economics 125:4 (2010), 1727–1767.
10.1162/qjec.2010.125.4.1727

Gopinath, G., and B. Neiman, “Trade Adjustment and Productivity in
Large Crises,” American Economic Review 104:3 (2013), 793–831.
10.1257/aer.104.3.793

Grifﬁth, R., R. Harrison, and H. Simpson, “Product Market Reform and
Innovation in the EU,” Scandinavian Journal of Economics 112:2
(2010), 389–415. 10.1111/j.1467-9442.2010.01602.x

Griliches, Z., “Patent Statistics as Economic Indicators: A Survey,” Journal

of Economic Literature 28:4 (1990), 1661–1707.

Halpern, L., M. Koren, and A. Szeidl, “Imported Inputs and Productivity,”
American Economic Review 105:12 (2015), 3660–3703. 10.1257/
aer.20150443

Harhoff, D., F. Narin, F. M. Scherer, and K. Vopel, “Citation Frequency
and the Value of Patented Inventions,” this REVIEW 81:3 (1999),
511–515.

Kopf, K., and T. Mayer, “Gravity Equations: Workhorse, Toolkit, and Cook-
book” (S. 131–195), in G. Gopinath, E. Helpman, and K. Rogoff,
Hrsg., Handbook of International Economics. Amsterdam: Sonst,
2014.

Kopf, K., T. Mayer, and J. Ries, “The Erosion of Colonial Trade Link-
ages after Independence,” Journal of International Economics 81:1
(2010), 1–14. 10.1016/j.jinteco.2010.01.002

Khandelwal, A. K., and P. Topalov, “Trade Liberalization and Firm Pro-
ductivity: The Case of India,” this REVIEW 93:3 (2011), 995–1009.
Lerner, J., “The Importance of Patent Scope: An Empirical Analysis,”
RAND Journal of Economics 25 (1994), 319–333. 10.2307/2555833
Lileeva, A., and D. Treﬂer, “Improved Access to Foreign Markets
. for Some Plants,” Quar-
Raises Plant-Level Productivity .
terly Journal of Economics 125:3 (2010), 1051–1099. 10.1162/
qjec.2010.125.3.1051

Loecker, J. D., P. K. Goldberg, P. A. Khandelwal, and N. Pavcnik, “Prices,
Markups and Trade Reform,” Econometrica 84:2 (2016), 445–510.
10.3982/ECTA11042

Martinez, C., “Insight into Different Types of Patent Families,” OECD Sci-
enz, Technology and Industry working papers 2010/2 (2010).
Nagaoka, S., K. Motohashi, and A. Goto, “Patent Statistics as an In-
novation Indicator” (S. 1083–1127), in Handbook of the Eco-
nomics of Innovation, Bd. 2. Amsterdam: Sonst (2010). 10.1016/
S0169-7218(10)02009-5

OECD, Patent Statistics Manual. Paris: OECD Publishing (2009).
Schmookler, J., “Invention and Economic Growth. Cambridge, MA: Har-

vard University Press (1966).

Steinwender, C., “The Roles of Import Competition and Export Opportu-
nities for Technical Change,” CEP discussion paper 26483 (2015).
Steinwender, C., and P. Shu, “The Impact of Trade Liberalization on Firm
Productivity and Innovation” (S. 39–68), in Josh Lerner, and Scott
Stern, Hrsg., Innovation Policy and the Economy. Chicago: Universität
of Chicago Press (2019).

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
/
R
e
S
T
/

A
R
T
ich
C
e
–
P
D

F
/

1
0
4
2
2
0
5
1
9
9
6
1
6
5
/
R
e
S
T
_
A
_
0
0
9
5
1
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3
PDF Herunterladen