Export‐led growth and its determinants: Evidence from Central and Eastern European countries
| Author | Jan Hagemejer,Jakub Mućk |
| Published date | 01 July 2019 |
| DOI | http://doi.org/10.1111/twec.12790 |
| Date | 01 July 2019 |
1994
|
wileyonlinelibrary.com/journal/twec World Econ. 2019;42:1994–2025.
© 2019 John Wiley & Sons Ltd
Received: 12 April 2018
|
Revised: 26 November 2018
|
Accepted: 14 February 2019
DOI: 10.1111/twec.12790
ORIGINAL ARTICLE
Export- led growth and its determinants: Evidence
from Central and Eastern European countries
JanHagemejer1
|
JakubMućk2
1University of Warsaw,Narodowy Bank Polski, Warszawa, Poland
2Warsaw School of Economics,Narodowy Bank Polski, Warszawa, Poland
Funding information
Fundacja na rzecz Nauki Polskiej, Grant/Award Number: START 2018
KEYWORDS
CEEC, common correlated effects estimation, economic growth, global value chains, heterogeneous panels, international trade
1
|
INTRODUCTION
The globalisation processes that occurred over at least the last two decades have changed the pattern of
division of labour and trade around the world. Production of goods has become fragmented, and countries
have become vertically specialised in tasks/stages of production rather than particular products and ser-
vices in the framework of global value chains (GVCs, see, e.g., Baldwin (2013) for a detailed description of
that process). As the supply chains have become difficult to track, the role of trade in driving the economic
growth has also become more complicated, that is, exports require intermediate imports and at the same
time may rely on imported technology, in particular in developing countries. In this paper, we answer the
following questions: What is the direct contribution of exports to economic growth? What is the role of
exports in driving the convergence among countries? What are the main drivers of export performance?
We analyse the European economies with a special focus on Central and Eastern Europe countries
(CEECs)1 who have undergone a great deal of structural change over the past two decades.2 Their
economic transition has involved a gradual removal of trading barriers and barriers to the international
flows of capital. Moreover, they have become the manufacturing backbone of the European economy,
by tight integration with the largely regional global value chains, a high degree of vertical specialisa-
tion on production of intermediate goods as well as reliance on intermediate imports and FDI. At the
beginning of the processes of transition and reintegration with the rest of Europe, the gap in the in-
come levels between the CEECs and the EU- 15 has been substantial and economic convergence has
1The group of CEECs includes Bulgaria, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia
and Slovenia.
2See, for example, Crespo and Fontoura (2007) for evidence on the changes in industrial structure in the period before the EU
accession of the CEECs and review of early literature on CEEC trade structures and foreign direct investment as well as Grela
etal. (2017) for post-accession evidence of the process of convergence in CEECs and its drivers.
|
1995
HAGEMEJER And MUĆK
been a major goal of EU accession. Our objective was to assess the contribution of export expansion
to the economic growth of CEECs over the extended period of time covering a large part of the tran-
sition period (1995–2014), the EU accession and finally the great trade collapse and the global finan-
cial crisis. Moreover, we inquire into the role of exports in the process of economic convergence of the
new EU member states. Lastly, we identify the supply- side determinants of export performance.
Our paper takes a novel approach to growth accounting. We decompose the supply- side aggregate
of GDP into the domestically absorbed and exported components. Traditional national account mea-
sures of net exports do not allow for an accurate assessment of the direct export contribution to growth,
in particular in countries undergoing a significant structural change. While due to increased GVC in-
volvement exports of goods are increasingly dependent on imports of intermediate goods, transition
countries have also been characterised by a sustained upward trend in the import intensity of final
demand components. This is particularly true for growing investment demand that was closely related
to catching- up processes and FDI- driven export expansion. However, due to increased specialisation
consumption demand in countries tightly integrated in GVCs have also increased. These differences
both in levels and in changes of import intensity of different national account components may lead to
a wrong assessment of the contribution of exports to GDP growth with the use of net exports. In order
to circumvent this problem, for example, Kranendonk and Verbr uggen (2008) as well as Cardoso,
Esteves, and Rua (2013) use national input–output tables to identify the import content of exports as
well as other GDP components. We follow a different approach, based on Johnson and Noguera (2012)
who propose a method of identifying sectoral value added generated in a country to a domestically
absorbed component and exports. By the use of annual global input–output tables, we analyse changes
in volumes of those components.3
Our paper is a contribution to the literature on the export–growth nexus, or more importantly, on
the export- led economic growth (ELG). This literature identifies many channels through which trade
can affect GDP growth. It is either direct, that is through the increase in demand and subsequent in-
crease in the physical output, or indirect, that is through the effects of specialisation and related pro-
ductivity boost.4 Moreover, export expansion allows for increased imports of (possibly more advanced)
intermediates and investment goods and they provide additional productivity gains to productivity. A
very useful review of pre- 2000 papers going back to as early as the 1960s is provided by Giles and
Williams (2001) (see also, e.g., Edwards, 1993). The early cross- country papers working on pooled
panel data (long- run averages) document mixed evidence of ELG, with a majority showing at least
some significant and positive association of growth with exports. However, given that there may be
supply effects (output growth driving exports), causality in the export relationship is not obvious.
Developments in the time- series methodology in the 1980s have lead to another strand of the literature
focused on (Granger) causality. While a bulk of papers in this group of works are individual country
papers, there are many exceptions that look at multiple countries' samples providing comparable re-
sults (among others Greenaway & Sapsford, 1994 for multiple developed and developing countries
that have undergone trade liberalisation, Dutt & Ghosh, 1994, 1996 for several developing countries
3Additionally, we use the method proposed by Wang, Wei, and Zhu (2013) to assess the role of vertical exports of intermediate
goods in the growth of exported value added of analysed countries.
4The microdata evidence on trade- related productivity gains is ample. Many of the studies are based on the seminal paper by
Bernard and Jensen (1999) who investigate the productivity- based selection into exporting as well as export- led productivity
improvements (learning by exporting). While the former process is confirmed by the data, the evidence for the latter is rather
scarce. Wagner (2007) surveys more recent evidence and reaches similar conclusions, that is, the evidence of learning by ex-
porting is restricted to selected countries. However, given the fact that exporters are more productive and tend to grow larger
and attract more resources, this reallocation alone is enough to see trade- induced productivity growth.
1996
|
HAGEMEJER And MUĆK
from Africa, South America and Asia, Amoateng & Amoako- Adu, 1996 for Africa; and Riezman,
Whiteman, & Summers, 1996 for 126 countries). However, the evidence on ELG and the direction of
causality are mixed, that is, both the significance of the relationship and the direction of causality vary
across and within papers.
Doubts on the direction of causality are related to possible endogeneity of exports. The seminal
paper by Frankel and Romer (1999) provides a solution to the problem, that is instrumental variables
approach. Their idea is to use exogenous geographical variables describe the size of the countries and
the proximity of the countries (in a similar way, the gravity model of trade defines those characteris-
tics5). Their results point to the upward bias of earlier cross- sectional estimates of ELG, but they do
find a robust positive impact of exports on growth. Newer papers have improved upon the identifica-
tion strategy by additionally exploiting the exogenous variation in trade costs. Examples include the
development of the steam engine in Pascali (2017), the fall of air transport cost in Feyrer (2009) and,
more recently, the improvements in container transport technology in Altomonte, Bonacorsi, and
Colantone (2018).
Our approach lets us distil the direct part of aggregate value added that is related to exports, that
is, we know exactly how much exports contribute to the growth of value added. We do not need to
analyse the growth–exports relationship per se as we are able to compute the exact measure of the
export contribution to growth through the supply- side analysis, and we can remain outside the debate
on the choice of instrumental variables as there are none required. However, this approach has its
disadvantages, that is, we do not capture all possible effects of exports on growth that were mentioned
above. As far as direct effects are concerned, we capture them all, that is, we can observe the external
demand contribution to the growth of value added after it was cleared of the imports of intermediates.
This calculation also includes the productivity effects that are related to exporting both in sectors that
export directly and in sectors that export only indirectly (i.e., provide intermediates to importing sec-
tors). However, we do not capture some additional spillovers, for example the gains from productivity
improvements that translate to the growth of domestically absorbed value added (e.g. sectors that gain
from overall productivity improvements but are completely focused on the domestic markets).
Our decompositions show that exports were a sizable component of economic growth of the
CEECs' economies and that they were a main factor behind economic convergence of the region, in
particular, after the EU accession. Export performance of the CEECs has been better than the one in
most of the comparator EU- 15 countr ies and remained to be important growth factor even after the
global economic crisis. We show that the rate of convergence within the CEECs due to exports was
twice as large as the one due to supply to the domestic market.
We proceed to analyse the determinants of export- related growth components, focusing on the
supply side. We apply standard panel methods and, due to significant substantial cross- sectional de-
pendence and possible heterogeneity, the common correlated effect estimator (Pesaran, 2006). We
find that export- related growth is associated mainly with capital deepening. While previous growth-
accounting exercises show that capital accumulation was a predominant source of growth in Central
and Eastern Europe (see, e.g., Dombi, 2013 or Havlik, 2005), we document that it has led to an ex-
pansion of export capacity rather than that of the domestic supply. This is in line with the findings by
Timmer, Erumban, Los, Stehrer, and de Vries (2014) who show that over a similar period, emerging
economies increased their export specialisation in capital- intensive goods that generated high value
added per unit of output, while the substantial capital accumulation was fuelled by its high returns and
competitive wages. In this context, our results are in line with the relatively new literature on the factor
content of trade. For example, Chor (2010) shows using an empirical application of a heterogeneous
5Rodriguez and Rodrik (2001) suggest that the geographical variables affect other variables than trade that also affect growth.
To continue reading
Request your trial