The only domestic markets. The term “exceptional export performance”

The relationship between international trade and
productivity growth has been extensively investigated. It is always at the core
of intense debates amongst academic researchers and policymakers over the past
decades since the pioneering work of A. B. Bernard, Jensen, and Lawrence (1995). Since then, many empirical
and theoretical studies in this field have flourished. These contributed great
insights into the literature, with an intensive
focus on the investigation of the relationship between characteristics of
firms, especially productivity, and exporting behavior of firms. Firms that
export are found in empirical studies to be better than firms that serve only
domestic markets. The term “exceptional export performance” initially used by A. B. Bernard and Jensen (1999a) to describe their findings of
the superiority of exporters in the U.S. manufacturing sector. Currently, this
is broadly confirmed by many other researchers in different countries, implying
the fact that exporters are superior to non-exporters almost everywhere. Two
alternatives which are not mutually exclusive1
hypothesis have been proposed to explain the superior performance of exporters,
but empirical evidence is not definitive as each of them has distinctive policy
implications. The first is self-selection hypothesis which considers the
post-exporting productivity effects. This implies that only the more productive
ones are able to recoup the sunk costs2
of entry into foreign markets and survive
in the tough foreign competition (Roberts & Tybout, 1997). An alternative explanation
is learning by exporting in which exporting makes firms more productive. This
is due to the fact that foreign competition and exposure can also speed up
technological acquisition through disembodied technology and knowledge diffusion.
This helps to achieve economies of scale and thereby improve the manufacturing
process, reduce production costs and improve product quality (Almeida & Fernandes, 2008; De Loecker, 2007; Van Biesebroeck, 2005).


Both channels stated above are plausible and it is
widely believed that exporters are superior to non-exporters. This can be
attributed to the self-selection effect or the learning effect from exporting
or both, though they vary significantly with respect to empirical methodology
and measurement of firm productivity. Thus, the literature discussing the
causality of an exporting-productivity relationship shows mixed findings. The
former is well supported by most of the existing empirical studies for both developed
and developing countries. For instance, (AB Bernard & Jansen, 1999) for the US; (J. R. Baldwin & Gu, 2003) for Canada; (Arnold & Hussinger, 2005; A. B. Bernard & Wagner, 1997, 2001)
for Germany; (Imbruno, 2008) for Italy; (Delgado, Farinas, & Ruano, 2002) for Spain; (Aw, Chung, & Roberts, 2000; Liu, Tsou, & Hammitt, 1999; Roberts
& Tybout, 1997)
for Taiwan; (Clerides, Lach, & Tybout, 1998) for Colombia, Mexico and
Morocco; (Isgut, 2001; Roberts & Tybout, 1997) for Columbia; (Poddar, 2004)
for India; (V. H. Vu,
2012) for Vietnam; (Sinani & Hobdari, 2010) for Estonia found corroborate
evidence of self-selection, but failed to find any evidence supporting learning
by exporting. The above empirical studies confirm that exporting firms are more
productive than non-exporting firms and revealed that higher productivity of
firms occurs before entry into export markets. While some other studies have
found no significant effect regarding the causality from firm productivity to
the decision to export ((A. B. Bernard & Jensen, 2004) for the U.S.; and (Aw et al., 2000) in Korea).

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now


Similarly, a mixed picture also appears regarding
empirical findings of the learning by exporting hypothesis in both developed
and developing countries, albeit, feeble and less in number (see (Wagner, 2007a)). For example, using UK data, (Crespi, Criscuolo, & Haskel, 2008; Girma, Greenaway, & Kneller,
2003; Greenaway & Kneller, 2004, 2007)
have found that firms boost their productivity advantage after being exporters.
Similar results also found from Canada and Slovenia manufacturing plants by J. R. Baldwin and Gu (2003) and De Loecker (2007) respectively. In contrary to developed
countries, in which evidence for learning effect is rare, in the developing
countries learning by exporting effects is more popular. For instance, (Kraay, 1999; Park, Yang, Shi, & Jiang, 2010) for Chinese firms; (Blalock & Gertler, 2004) for Indonesian firms; (Yasar, Nelson, & Rejesus, 2006) for Turkey and (Bigsten et al., 2004) for Sub-Saharan African
countries have found evidence of a post-exporting productivity gain. On the
other hand, a number of studies have found no proof of the learning by
exporting effect, even for major exporting countries (for instance: (Hailu & Tanaka, 2015) (A. B. Bernard & Jensen, 1999c; Hung, Salomon, & Sowerby, 2004) for the USA; (Fu, 2005)
for China; (V. H. Vu, 2012) for Vietnam; (Arnold & Hussinger, 2005; Wagner, 2002) for Germany).


However, some studies have found evidence of
coexistence of both hypotheses: for instance; (Aw, Roberts, & Winston, 2007) for Taiwan; (Alvarez & Lopez, 2005) for Chile; (Kimura & Kiyota, 2006) for Japan; (Girma, Greenaway, & Kneller, 2004; Greenaway & Kneller, 2004) for the UK; (Hahn, 2005)
for Korea; (Fernandes & Isgut, 2005) for Colombia; (Bigsten et al., 2004) as well as (Van Biesebroeck, 2005) for SSA Countries; and (Bigsten & Gebreeyesus, 2009) for Ethiopia. Some other
studies, while, have failed to find any evidence for either hypothesis and
conclude that the performance characteristics of exporters and non-exporters
are remarkably similar (e.g., (Girma et al., 2003); (Kim, Gopinath, & Kim, 2009) and (Sharma & Mishra, 2011) using data from Swedish,
Korean, and Indian firms, respectively).


While trying to explain the reason behind the
observed blended results across countries and time of the export and
productivity nexus, the empirical literature has recently moved toward other
aspects of firm heterogeneity. This includes international trade associated
with the macroeconomic environment; the degree of competition and entry costs
in the export markets that firms are likely to face. Some other studies also
consider particular behavior of firms involved in international activities for
the existence of a mixed result, for example, product and country
diversification ((Andersson, 2001) for Sweden and (Wagner, 2007b) for Germany); import behavior ((Castellani, Serti, & Tomasi, 2010) for Italy); Geo-economic
orientation ((Damijan, Polanec, & Prašnikar, 2004) for Slovenia) and FDI
behavior ((Helpman, Melitz, & Stephen, 2004) for USA). According to Blalock and Gertler (2004), firms in countries with poor
technology and low productivity can gain a greater marginal benefit from
exposure to exporting and show that the level of economic development may be
the main reason for differing results. It also asserts that the variance in
characteristics of geographical and economic conditions of countries may be the
reason for the nexus (Wagner, 2007b). Lastly, Sharma and Mishra (2011) also indicate the diverse
conclusions may originate from using a wide variety of econometric
methodologies and approaches for testing these two hypotheses.


Moreover, the nonexistence of a consistent
measurement of productivity can be responsible for the uncertain and mixed
result of export productivity linkage. Some previous studies often use the
conventional technique for estimating Total Factor Productivity (TFP) such as
the Solow residual method which defines TFP growth as the residual of output
growth after the contribution of labor and capital inputs have been subtracted from
total output growth. This approach depends on an established assumption which
includes the form of the production function is known; all firms are working
effectively with no space for any inefficiency; neutral technical change and
have a constant return to scale, which means that TFP growth equal to technical
progress growth. If these assumptions do not hold, TFP measurements will be
biased (Arcelus & Arocena, 2000; T. Coelli, 1998). Some others also use labor
productivity to stand for productivity, yet this index just represents a part
of the picture of productivity and should be considered as one of the
attributes of exporting manufacturing firms in the Ethiopian context. In a study in Ethiopia, while
considering the relationship between export status and firm productivity, Bigsten and Gebreeyesus (2009) used three different measures
of firm performance – TFP, labor productivity (Q/L) and unit labor cost (ULC)
of different industries from 1996 to 2005 and found a support for both
hypotheses. Nonetheless, the above study and these methodologies don’t permit
the decomposition of TFP change into its components such as technical progress
change, technical change and scale efficiency change ((S. C. Kumbhakar & Lovell, 2003)). Rather, most studies often
consider productivity under a single umbrella of investigation that does not
give due consideration regarding the different parts of productivity and the
importance of their influence even if it helps to understand whether gains in
productivity levels are achieved through the efficient use of inputs or through
technological progress. This will constrain further investigation concerning
the relationship between export participation and productivity with its
decompositions just when an aggregated index for productivity is considered.
The only exception in this regard is the study by Fu (2005)
for China who utilized a random effects panel data model to test the effect of
export status on productivity growth and its parts. Its decomposition into
technical efficiency and technical progress is also made by using a frontier
approach which is examined by employing the Malmquist index. But it still
overlooks the contribution of export intensity on scale efficiency and used
industry not firm-level data. Furthermore, V. H. Vu (2012) also examines the causality of exporting and
firm productivity using a different sample retrieved from a survey of
Vietnamese SME firms by decomposing TFP change into technical progress change,
technical efficiency change, and scale efficiency. The study failed to find
evidence in support of export participation on any of TFP components. However,
their study based on the data surveyed, just for private small and medium firms
and a short panel dataset which may not give a full picture of this


Despite there are numerous empirical studies using
datasets from different countries to test the hypothesis so far, it would seem
at very much informative stage and were no dominant explanation exists (Sharma & Mishra, 2011). Besides, the above issues
bring up a question about whether the measurement of productivity can offer an
alternative explanation for the mixed results in the relationship between
productivity and export. Thus, the present study motivated by the existing
empirical research crevice and the need to revisit the validity of the two
hypotheses within manufacturing firms in Ethiopian context for the period
2000-2011. We measure productivity change by using Stochastic Frontier Approach
(SFA) to release the assumption of a full efficiency of firms and decompose into
its components, including technical change, scale change and technological
progress change by following S. Kumbhakar and Lovell (1998). Although other approaches
like data envelopment analysis (DEA) may divide productivity change, the SFA
has been preferred in this study. This is due to its advantages with regard to
controlling with inefficiencies resulted from omitted variables, measurement
errors, outliers and stochastic noise, which may result in a possible upward
bias of inefficiency scores (Del Gatto, Di Liberto, & Petraglia, 2011). Besides, we used two
additional productivity measures which are Levinsohn and Petrin (2003) methodology and labor
productivity calculated by output per total employees, to test the robustness
of the results. We then used different econometric methods to deal with the
causality between export participation and productivity change with its


The paper has two main novelties vis?à?vis previous
literature in the area of heterogeneous-firm trade empirical literature: first,
in relation to decomposing productivity, to the best of my knowledge, it is the
first investigation to consider the impact of export participation on each
component of TFP in the context of African firms. Second, decomposing TFP can
provide another way to explain the mixed findings in empirical studies as well
as provides additional insights into understanding the recent debate on TFP
growth. In sum, it will broaden
our empirical insight into what policies and strategies should be pursued to improve
productivity and thus their competitiveness in the global market in line with
each component. Our argument is that export participation can impact negatively
on productivity change, but it may create positive effects on each component of
productivity. Therefore, considering TFP as an aggregated index will hide such
interesting points. Furthermore, we complement the evidence by studying the
impact of other firm-specific characteristics as determinants of its export
performance. Finally, Ethiopia makes for a particularly interesting study in
light of two main reasons: its’ relatively strong export-oriented policy and
the absence of such empirical analysis.


The remainder of the paper is organized as follows.
Section 2 presents some background information about Ethiopian economy. Data
sources are presented in section 3 while section 4 specifies the empirical
models. Finally, section 5 and 6 consecutively discusses the results and
concluding remarks.

1 Meaning that both effects can
sequentially play a role, before and after firms start exporting

2 such as transport costs, modification
costs to meet foreign tastes and regulations, distribution or marketing costs
and setup costs to establish distribution network which prevent less productive
firms from entering foreign markets.