Gia nhập tổ chức WTO và năng suất của các công ty sản xuất Việt Nam: 2006 - 2013
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
WTO ACCESSION AND FIRM PRODUCTIVITY OF VIETNAMESE
MANUFACTURING SECTOR: 2006 – 2013
GIA NHẬP TỔ CHỨC WTO VÀ NĂNG SUẤT CỦA CÁC CÔNG TY SẢN XUẤT VIỆT NAM:
2006 - 2013
Phan Dang My Phuong, Le Dien Tuan
Danang University of Economics
Email: phuong.pdm@due.edu.vn
Abstract
The study in this paper examines the causality relationship between exporting and firm productivity in
an emerging market economy on the path of trade liberalization and market-oriented transition reforms. To
achieve this purpose, we focused on testing whether only the most productive firms can overcome a productivity
threshold to enter export markets (self-selection) or whether firms that enter export markets observe a
subsequent productivity improvement (learning by exporting) using a panel data of firm-level data of
manufacturing firms in Vietnam during the period 2007-2013. This study is conducted based on a random-
effects dynamic probit model for the causality from firm characteristics to export probability and a matching
technique in combination with a difference-in-difference approach for the existence of learning-by-exporting
effect. The econometric results show that not only that productive firms self-select into exporting but also that
they improve their productivity further following their entry. The findings are largely consistent with the
emerging evidence from other developing countries and give support to the export-led growth strategies that
Vietnam is pursuing.
Keywords: WTO, productivity, export
Tóm tắt
Nghiên cứu này xem xét mối quan hệ qua lại giữa xuất khẩu và năng suất trong một nền kinh tế thị
trường hội nhập. Nhóm tác giả tập trung vào kiểm định những công ty với năng suất cao có thể vượt qua rào
cản về năng suất để gia nhập vào thị trường xuất khẩu (tự lựa chọn-SS) hay rằng các công ty đó có những cải
thiện về năng suất (học hỏi thông qua xuất khẩu –LE)) dựa trên dữ liệu mảng về các công ty sản xuất ở Việt
Nam trong giai đoạn đầu của gia nhập WTO 2007-2013. Nghiên cứu này sử dụng mô hình thống kê probit ngẫu
nhiên để xem xét mối quan hệ qua lại giữa các đặc tính của công ty với khả năng xuất khẩu và phương pháp
ước lượng khác biệt trong khác biệt (DID) để xem xét có tồn tại hiệu ứng học hỏi thông qua xuất khẩu. Các kết
quả kinh tế lượng cho thấy rằng những công ty có năng suất cao không chỉ tự lựa chọn vào thị trường xuất khẩu
mà còn cải thiện năng suất sau khi xuất khẩu.
Từ khóa: Tổ chức thương mại thế giới, năng suất, xuất khẩu
1. Introduction
As a transitional economy, Vietnam has been implementing the reform process (Doi Moi) in
the late 1980s, aimed at transforming the centrally planned economy to a market economy. The reform
process has been accelerated with the World Trade Organisation (WTO) accession in 2007. The rapid
removal of non-tariff barriers, follow-up tariff reduction and harmonization of legal systems are among
the significant outcomes from Vietnamese accession to the WTO. It is expected that such changes
would expose domestic firms to foreign competition and promote the efficiency of domestic
firms/exporters in the light of comparative advantage that they hold.
There are two competing theories that are tested to explain the superior performance of
domestic firms/exporters: (i) self-selection (SS) and (ii) learning-by-exporting (LE). The SS theory
proposes that sunk entry costs consisting of transportation, seeking new markets and the modification
of domestic goods are prerequisites to domestic firms entering into export markets. Thus, it is expected
that more productive firms with sufficient funds are able to access the export markets. This proposition
126
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
is also known as trade-induced within-firm productivity improvement (Lach and Tybout, 1998). On
the other hand, the LE hypothesis emphasizes reverse causality, in that productivity improvements of
exporting firms stem from their foreign markets entrance. They obtain incentives on management and
marketing experiences from export trading partners.
This paper explores the impact of WTO accession on firms by using a firm-level dataset
covering the Vietnamese manufacturing sector from 2007 to 2013. It examines differences between
exporter and non-exporter and then tests the two hypotheses–SS using the probit model (to examine
the effects of key factors on firm export participation) and LE using matching techniques. Testing
these two hypotheses allows examination of the causality between Total Factor Productivity (TFP) and
export behaviour right after joining WTO. The entrance of new firms and their learning capacity in the
competitive environment determine their superior performance.
The rest of the paper is organized as follows. Section 2 analyses the literature survey on export
and firm productivity and methodology. Section 3 presents the empirical results of this research. A
summary and conclusions are discussed in Section 4.
2. Literature review, Theoretical framework and Method
2.1. Literature review
The relation between exports and productivity has been extensively studied by traditional trade
theories from the theoretical comparative advantage for inter-industry trade models (Ricardo and
Hechsker-Ohlin) to the models of intra-industry trade (Helpman and Grossman, 1995). Recent trade
theories on heterogeneous firms tested self-selection (SS) and Learning by Exporting (LE) hypotheses
(Bernard and Jensen 1995).
Self-selection (SS)
The SS mechanism takes place via a link from productivity to exporting. Bernard and Jensen
(1995) examine the fundamental reason for the correlation between export and productivity. If the
productivity cut-off level of new entrants in the domestic market is higher than the minimum cut-off
level, then firms will self-select to export. Bernard and Jensen (1995) demonstrated that the entry sunk
costs are not zero and are based on the firms’ exporting experience and plants’ characteristics such as
size and location. Sunk costs are irrecoverable and create an entry barrier that only more efficient firms
can afford to overcome and self-select to enter the export market.
Exporting firms are generally different from non-exporting firms in key aspects such as more
sophisticated technologies, larger size and higher wages (Lach and Tybout, 1997, Clerides et al., 1999,
Bernard and Vahter, 2007). Firm size is considered as an important factor in influencing the firm
decisions on exports and therefore larger the firms are more likely to export than smaller ones
(Damijan, 2004). Selection effects of most productive firms entering the export markets might be a
conscious process as firms improve their productivity in preparation for exporting (Kneller, 2007). In
the case of Vietnam, it is revealed that exporting firms mainly operating in sectors with low-skilled
labour and domestic raw materials have higher labour productivity than non-exporters (Kokko, 2012).
Foreign firms rather than domestic ones are more likely to export (Sinani, 2010; Girma, 2004).
Foreign firms have already absorbed knowledge as well as experience in overseas markets and
therefore advantaged by lower handling costs and better technology. The presence of foreign firms in
Vietnam’s manufacturing sector has a positive impact on the export decisions of domestic firms
(Thang and Ngoc, 2004; Tran and Bui 2012).
In contrast, a few empirical studies have not had any support in testing the SS hypothesis
(Jensen, 2004, Aw et al., 2002, Castellani, 2003). In the case of South Korea, Aw et al. (2002) argue
that export market entry costs dropped due to export subsidy in the mid-1980s. Thus, the export
127
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
decision of Korean firms was based on their abilities to access financial resources and approach
foreign customers, instead of improving their productivity.
Learning-by-exporting
While the SS mechanism takes places via the link from productivity to exporting, The LE
mechanism considers effects from exporting to productivity (Bernard and Jensen, 1997). Thus, the
firms improve their efficiency as a consequence of a learning mechanism through their participation in
export markets. Experience and knowledge gained from participating in export markets from
competitors and buyers will help firms grow faster than the firms focusing only on domestic markets
(Afonso et al., 2012; Bernard and Wagner, 2007). The effect of learning from exports comes from two-
dimensional knowledge diffusion: vertically as the competition of competitors, horizontally as the
comparison of customers.
Wagner (2012) suggests that competition (vertically) comes directly from competitors
penetrating the international market. Thereby being exposed to more intense competition, exporters
must innovate constantly to survive in an international competitive environment. Exporters must apply
the most modern technologies because, if not applied, they cannot exist in a highly competitive market.
When firms conduct continuous innovation, it leads to technological advancement and ultimately the
increase in productivity (Anderson, 2009).
Together with horizontal competition, foreign customers might push exporting firms to
improve the technological process by providing product designs, technical specifications and technical
support (Rodrik, 1992; Grossman and Helpman, 1995; Clerides et al., 1999). Knowledge accumulated
in the exports activities is primarily the result of these learning activities. Learning through exports
relates to the benefits that exporters derive from the knowledge of buyers. The buyers’ stringent
technical standards help exporting firms to become more aware of new technology and methods
transforming inputs into outputs. Therefore, the products fit the demand of consumers (Bernard and
Jensen, 1999, Bernard and Wagner, 2007, Haidar, 2012). The study of Mertin and Yang (2009)
reaffirms the above point of view by showing that exports can increase firm productivity due to
economies of scale. Enhancing exports allows firms to scale up, reduce costs and as consequence
productivity will be improved.
Evidences from both SS and LE
There are empirical evidences supporting for both SS and LE mechanisms (Boermans et al.
2013; Garcia, 2012; Kazem, 2007; Blalock and Gertler, 2004; Castellani, 2002; Alvarez, 2002). Many
empirical studies have got only support for SS mechanism (Vu, 2012; Alborno et al. 2007; Greenaway
and Yu, 2004; Bernard and Jensen, 1999). However, there are some studies that found no evidence of
both mechanisms (Greenaway et al., 2005; Sharma and Mishra, 2015).
Table 1: Exports and productivity of firms in different countries
Authors
Aw and Hwang (1995)
Country
Taiwan
Methodology
Translog production
function
SS
LE
Yes
Yes
No
Clerides et al. (1998)
Colombia,
Mexico, Morocco function
Translog production
Morrocco: Yes
Columbia & Mexico: No
No
Bernard and Jensen (1999) USA
Linear probability with Yes
fixed effect
Kraay (1999)
Castellani (2002)
Delgado et al. (2002 )
China
Italy
Spain
Dynamic panel
Cross-section
Yes
Yes
No
Yes
No
Panel data, matching yes
Panel data, matching yes
Panel data, matching Yes
Bernard and Vahter (2002) Germany
Greenaway (2002)
Alvarez (2002)
Yes
No
UK
Chile
Cross-section
Yes
Yes
128
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
Greenaway and Yu (2004) UK
DID
Yes
yes
not
No, firms exiting because
of the market share lost
Yes
Blalock ang Gertler (2004) Indonesia
Translog production
function
DID
Fernandez et al. (2005)
Alborno et al. (2007)
Colombia
Argentina
Yes for young firms and
examined industries targeted to
high-income countries
Panel data, GMM
Yes
No, higher productivity
for exporting to more
developed countries
Yes if exporting to more
developed countries
Yes, technological
capacity affects to
knowledge diffusion
Yes, improvement in
labour recruitment
No
Kazem (2007)
Garcia (2012)
Egypt
Spain
Panel data, matching Yes
Translog production
function
Yes
Boermans et al. (2013)
Vu (2012)
Ghana and
Nigeria
Vietnam
Cost function
Yes
Yes
Fixed effect
Source: Author’s summary from previous studies
A few empirical studies tend to support the SS mechanism rather than LE mechanism. This is
mainly because the LE mechanism appears only when firms have the capacity to absorb new
knowledge to improve and upgrade their current production processes or may decide to fully utilize the
new production process (Salomon and Shaver, 2005). However, there are differences between firms in
terms of technology capability (Wernerfelt, 1984, Barney, 1991, Peteraf, 1993). In addition, absorption
capacity influences the development of technological capacity (Cohen and Levinthal 1990). Firms with
better absorption capacity will be appropriate to learn from foreign partners and absorb the knowledge
via the production process (Garcia et al., 2012). Few empirical studies confirm the evidence of LE
mechanism by establishing the positive association between exports and productivity (García et al.
2012; Aw et al. 2008; Castellani 2007). For this purpose, empirical studies used investment in R&D as
a proxy. Investment in R&D not only improves the ability of exporters to apply foreign knowledge
into innovation activities but also helps firms improve productivity by learning through exports
(Salomon and Jin 2010).
Some argue that The LE mechanism works based on the place of exports. The reason is that
customers in more developed countries expect a higher quality of products and help to transfer
technology so they will motivate exporters to improve production. Based on this hypothesis, one can
observe positive effects of exports in developing countries as importers in developed countries require
a certain level of standard from exporters in developing countries (Love, 2013; Boermans, 2010; Eaton
et al., 2008; Vu, 2012; Lopez, 2005).
2.2. Methodology
This paper intends to examine the causal relationship between productivity and exporting of the
Vietnamese manufacturing firms during post-WTO accession and tests following three hypotheses:
H1. Exporting firms are more productive than non-exporting firms.
H2. Firms with higher productivity are more likely to self-select into export markets.
H3. Firms increase their productivity by learning- by- exporting
In order to test hypotheses, we first estimate export premium which is defined as the percentage
difference between export statuses in the mean level of firm characteristics such as revenue, TFP,
employment, wage, capital intensiveness and age (Bernard and Jensen, 1999). The export status will
129
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
then be used to establish the link between productivity and exports. The positive link indicates that the
Vietnamese firms gained confidence in post –WTO accession and vice versa.
Export Premium
ꢀꢁꢂꢃ∗ꢄ = ꢅꢆꢇꢃꢄ + ꢈꢃꢄ
(1)
where ꢉ indexes firms and ꢊ is the time index
ꢂꢃ∗ꢄ is the characteristic in consideration
ꢇ is the export status (0,1)
ꢃꢄ
ꢅꢆ is a parameter to be estimated
ꢈꢃꢄ is the error term assumed to be ꢉꢉꢋ.
The premium is calculated as the mean in the data set pooled across 2007–2013. This simple
export premium does not control for differences in firms’ characteristics, like firms’ location and
industry type. Incorporating these characteristics gives the conditional export premium. It is defined as
the difference in the mean level with other characteristics, location and industry type being controlled.
This is done by estimating the multivariate regression for the whole seven-year period controlling for
industry and location.
ꢀꢁꢂꢃ∗ꢄ = ꢌꢆꢇꢃꢄ + ꢌꢍꢂꢃꢄ + ꢌꢎꢏ + ꢌꢐꢑ + ꢒꢃꢄ
ꢂꢃ∗ꢄ is the vector of firm characteristics above
ꢏ is a vector of time dummies
(2)
ꢑ is a vector of industry and location dummies
ꢌꢆ, ꢌꢍ, ꢌꢎ and ꢌꢐ are vectors of parameters to be estimated
ꢒꢃꢄ is the iid error term.
The export premium is calculated as follows:
∗ꢔꢕꢖꢗꢘꢄꢔꢘ
ꢃꢄ
ꢙꢗꢙꢚꢔꢕꢖꢗꢘꢄꢔꢘ
ꢃꢄ
ꢙꢗꢙꢚꢔꢕꢖꢗꢘꢄꢔꢘ
ꢃꢄ
ꢓ(ꢂ
− ꢂ
)/ꢂ
ꢛ ∗ 100
(3)
After estimating all the parameters, the simple exporter premium is calculated as
(ꢜꢝ − 1) ∗ 100
(4)
ꢞ
The conditional exporter premium is calculated as
(ꢜꢟ − 1) ∗ 100
ꢞ
The two parameters ꢅꢆ and ꢌꢆ describe the difference between exporters and non-exporters.
Table 2: Summary of expected signs of dependent variables for export premium model
Variable
Exporter
Revenue
Employment
TFP
Variable descriptions
Expected signs
Having exporting activity (1=yes,0=no)
Total sales
+
+
+
+
-
Total employees
Total Factor Productivity
Wage
Total labour payment divided by total employees
Capital size per total employees
Number of years in business
Capital intensity
Age
+
Source: Author’s construction
130
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
Self–selection
A firm will decide to export if the net expected profit from this decision is positive. Foreign
market participation may incur some costs in collecting market information and adjusting production
processes and products to satisfy foreign customers. Most of these costs are due to their sunk nature.
With the assumption that entry sunk costs are not occurring if they did not export last period, a firm
will look beyond the current stage in the exporting decision (Tybout, 1997).
The export variable can be formulated in the presence of sunk costs:
(
)
+ ꢈꢃꢄ > 0
ꢥꢊℎꢜꢦꢧꢉꢨꢜ
1 ꢉꢡ ꢌꢢꢃꢄꢚꢣ − ꢤ 1 − ꢇ
ꢃꢄꢚꢣ
ꢇ = ꢠ
ꢃꢄ
(5)
0
where N is the sunk entry costs of exporting
ꢇ is a binary variable indicating exporting status
ꢃꢄ
ꢢꢃꢄꢚꢣ is the firm-specific factors affecting exporting decision
ꢈꢃꢄ is the iid error term.
The variable of interest is TFP, considered as a key indicator that determines the trend of being
an exporter. The Levinsohn and Petrin (LP) approach is used to construct TFP values. Other
observable firm characteristics are capital intensity, wage, firm size, age, and firm foreign ownership.
Capital size per total employees represent capital intensity and one would expect a negative association
between capital intensity and export status. Labour-intensive firms in a developing country tend to
export and so workforce quality and wage rates will be expected to be positive with export status.
Larger firms can benefit from their size via economies of scale in production due to greater
demand. Accessing remote markets requires more resources that only firms of a certain size can afford
(Majumdar, 1997; Taymaz, 2005; Alvarez and Crespi, 2003). Foreign-owned enterprises are more
likely to become exporters with the advantages of their experience and knowledge of overseas markets
(Bartelsman et al., 2000; Demsetz and Villalonga, 2001; Criscuolo, 2005). Firms with long trading
history are more likely to seek foreign markets for further development; older firms tend to pursue
exporting activities (Majumdar, 1997, De Kok et al., 2006, Cucculelli et al., 2014). In addition, firm
age sometimes has a close relation to experience, performance and firm size.
The self-selection model of export decision is therefore specified as:
ꢩ(ꢇ = 1) = ꢪ(ꢏꢫꢩꢃꢄꢚꢣ, ꢇꢃꢄꢚꢣ, ꢨꢉꢬꢜꢃꢄꢚꢣ, ꢀꢁꢭꢀꢃꢄꢚꢣ, ꢮꢯꢜꢃꢄꢚꢣ, ꢫꢰꢱꢃꢄꢚꢣ, ꢋꢈꢲꢲꢉꢜꢨ)
(6)
ꢃꢄ
where ꢪ(. ) is a normal cumulative density function.
It can be estimated by probit with panel data. Using random and fixed-effects a model often
properly treats unobserved firm heterogeneity. Two specifications are used: first, using probit with
unobserved effects to identify the determinants of firms’ exporting decision; second, Heckman’s
random-effects dynamic probit is applied to control unobserved effects and the dynamic process.
Dummy variable are included, one each for industry, region and time to proxy for spatial, industrial
and regional characteristics.
Learning-by-exporting
LE hypothesis suggests that firm productivity increases after entering export markets and
investigates the effects of exporting on firm productivity by employing a matching technique to
overcome the problem of self-selection. A propensity score matching (PSM) technique makes
consistent comparisons between exporters and non-exporters (Girma et al. 2004; Vahter, 2002; Becker
131
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
and Ichino, 2002).1 The matching method resolves all the systematic differences related to both the
exporting decision and firm productivity.
Following the methodology of Heckman et al. (1997), the average treatment effect on the
treated (ATT) of exporting firms is calculated. The ATT evaluates the effect of exporting activities on
firm performance. The ATT for exporters is defined as:
ꢣ
ꢗ
ꢣ
ꢗ
(
)
(
)
(
|
)
ꢇ = 1
ꢱ ∆ꢳꢃꢄꢴꢵ − ꢳꢃꢄꢴꢵꢶꢇ = 1 = ꢱ ∆ꢳ ꢶꢇ = 1 − ꢱ ꢳ
(7)
ꢃꢄ
ꢃꢄ
ꢃꢄ
ꢃꢄꢴꢵ
ꢃꢄꢴꢵ
Where ∆ꢳꢃꢄꢴꢵ is firm ꢉ’s outcome at period ꢊ + ꢨ (ꢨ > 0)
ꢇ is a binary variable indicating exporting status of firm ꢉ at period ꢊ
ꢃꢄ
∆ꢳꢣꢃꢄꢴꢵ is the outcome at time ꢊ + ꢨ of firm ꢉ that exported at ꢊ (ꢇ = 1)
ꢃꢄ
ꢳꢗꢃꢄꢴꢵ is the outcome of firm ꢉ that has not exported.
The matching technique is used to estimate the counterfactual outcome by finding a valid
control group among non-exporting firms. The counterfactual is estimated by the corresponding
average value of the outcome of non-exporting firms. The basic principle of matching is to select for
the non-exporters group that does not have any difference with the exporter group.
After building the control group through matching, causal effects of exporting on productivity
are estimated by the new sample consisting of the exporter group and matched control group. The
impacts of both exporting participation and exporting involvement will be investigated.
The equation explaining TFP of firm ꢉ at time ꢨ > 0 is
ln ꢳꢃꢵ = ꢌꢷ + ꢌꢷꢆꢇꢃꢷ + ꢌꢷꢸꢂꢃꢷ + ꢌꢹꢨ + ꢌꢆꢹꢇꢃꢷꢨ + ꢌꢍꢹꢂꢃꢷꢨ + ꢺꢃ + ꢻꢃ
(8)
where ꢇ is dummy variables indicating different exporter statuses
ꢃꢷ
ꢂꢃꢷ is the observed firm-specific factor in the base year
ꢺꢃ is unobserved firm effect.
Taking annual average differences between ꢊ= 0 and ꢨ > 0, yielding
ꢣ
(
)
ꢀꢁꢳꢃꢵ − ꢀꢁꢳꢃꢷ = ꢌꢹ + ꢌꢆꢹꢇꢃꢷ + ꢌꢍꢹꢂꢃꢷ + ꢼꢃꢵ (8)
(9)
ꢵ
With ꢼꢃꢵ = ꢻꢃꢵ − ꢻꢃꢷ
In a difference-in-differences approach, time trend, unobserved firm effects and observed firm
characteristics are controlled.
ꢌꢆꢹ represents ꢳꢏꢏ in terms of log of growth rates of TFP.
In this study, the matching process will be implemented by STATA algorithm by Becker and
Ichino (2002). Specifically, the sample is divided into ꢭ intervals regarding the propensity score ꢽꢃ,
and checked whether the average propensity score between the treated and controlled units is not
different in each interval. If the test fails over a period of time, the interval is divided into half and the
test of differences is repeated until it stays in each time period. Then, the necessary condition of the
equilibrium hypothesis is examined. If in each interval, the means of each characteristic are not
different between the treated and controlled units, this condition is considered to be satisfactory.
3. Results and discussion
Export premium
From equation 2, the estimated coefficient ꢌ of the export variable denotes the superiority of
1
This has become a very popular approach for estimating causal treatment effects, especially when evaluating labour
market policies but it is also used in diverse fields of study.
132
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
exporters over non-exporters in the same industry. Table 3 presents the results.
Table 3: Export premium
Variables
Revenue
Simple export
premium (%)
Standard errors Conditional export premium
(%)
Standard
errors
258.27
27.55
312.80
3.27
0.0678***
0.0624***
0.0587***
0.0275
50.52
16.42
158.57
4.5
0.0375***
TFP
0.0408***
0.0532***
0.0269
Employment
Average wage
Capital intensity -20.01
0.0556***
0.0745***
0.0469***
-42.79
150.49
14.64
0.0412***
0.0735***
0.0532**
Capital
Age
226.48
32.29
Note: ***, **, * denote 1 per cent, 5 per cent and 10 per cent significance level respectively.
Source: Author’s calculation from VES
At the unconditional mean, employment and revenue level in exporting firm are about nearly
300 per cent higher than those in non-exporters. The positive export premium remains the same with
smaller magnitudes. In terms of the conditional export premium, the largest premium is in employment
and capital, then revenue later TFP and age. All these differences are statistically significant at 1 per
cent level, except for wages, indicating that exporters are significantly larger than non-exporters.
Exporters also have high premium in TFP in comparison to non-exporters. In terms of
experience years, exporters are better than non-exporters in terms of TFP, size and age. In particular,
exporters are more productive and indicate support for our hypothesis 1.
Self-selection hypothesis
What is the possibility of a non-exporter becoming an exporter as defined by SS (Bernard and
Jensen, 1999)?. The estimated coefficients in Table 7 are representing the marginal effects of each
explanatory variable on the link function in probit models. The model (1) is probit in pooled data, the
model (2) is Heckman’s random-effects dynamic probit model.
Table 4: Effects of TFP on exporting probability
Variables
(1)
(2)
0.432***
(0.0466)
0.666***
(0.0120)
ln(ꢏꢫꢩ
)
ꢃꢄꢚꢣ
-0.0907***
(0.0227)
-0.0844***
(0.0491)
ꢀꢁꢾꢿꢃꢄꢚꢣ
0.0414
0.00370
(0.0953)
ꢀꢁꣀꢮꢯꢜꢃꢄꢚꢣ
ꢀꢁꢨꢉꢬꢜꢃꢄꢚꢣ
ꢀꢁꢮꢯꢜꢃꢄꢚꢣ
ꢫꢰꢱꢨ
(0.0550)
0.320***
(0.0162)
0.578***
(0.0427)
0.0315
0.0861
(0.0316)
(0.0431)
0.676***
(0.0492)
1.337***
(0.187)
Industry FEs
Year Fes
Yes
Yes
Yes
Yes
Note: ***, **, * denote 1 per cent, 5 per cent and 10 per cent significance level respectively.
Source: Author’s calculation
133
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
The positive and statistically significant coefficients of TFP imply firms with higher
productivity in Vietnam tend to self-select into foreign markets and gain support for our hypothesis 2.
As far as capital intensity is concerned, both models with negative sign show that firms with a smaller
ratio of capital to labour tend to export. It indicates that firms with labour-intensive status reflecting
comparative advantages are more likely to serve foreign markets. The effect of average wage is
positive and statistically insignificant. It could be explained that the majority of Vietnamese exporters
operate in labour-intensive sectors and offer a lower average wage.
As for other observable variables, firm size and firm age are also determinants of exporting
probability. The significant impact of firm size on serving foreign markets is also large. Firms with
larger size are more likely to pursue exporting activities. Firms with many years of doing business
might be more efficient when accessing exporting transactions. Moreover, Foreign-owned enterprise is
also a significant determinant, having up to twice the effect as TFP.
The estimates indicate that the Vietnamese manufacturing firms with higher productivity tend
to select into serving foreign markets because of their capacity to cover the entry sunk costs considered
as the entry barrier of engagement in international trade.
Learning-by-exporting
In this section, the study examines whether exporting participation could enhance firms’
productivity. Propensity score matching (PSM) is used to estimate the impacts of the determinants of
exporting participation. PSM is undertaken for the entry year of serving a foreign market and that of
one year later. The estimated coefficients of the determinants of exporting participation are presented
in Table 8. All the variables with one-year lag are statistically significant. Using a one-year lag
prevents all the reserve impacts when calculating exporting probability.
Table 5: Determinants of the manufacturing firm entering into export activities
Variables
Lag TFP
Coefficients
Entering year
Coefficients
After one year entering
0.2132***
(0.0863)
0.4307***
(0.0586)
Lag Size
Lag FIEs
0.1292***
(0.0384)
0.0840***
(0.0274)
0.5087***
(0.1765)
0.3674***
(0.1348)
Industry effects
Year effects
Yes
Yes
Yes
Yes
Note: ***, **, * denote 1 per cent, 5 per cent and 10 per cent significance level respectively.
Standard errors presented parentheses.
Source: Author’s calculation
Firms need time to acquire and then adjust their capacity to exploit the benefits of export.
Specifically, there may be a gap between the time of export participation and the time for the benefit of
this participation to be made. Differences in exporting experiences seem to be a considerable factor in
making firms different.
After finding the probability of the firms’ entry into exporting activities, firms are matched by
the one-to-one nearest neighbour matching method for matched controls. The PSMATCH (Aianesi,
2004) is employed for this estimation. After applying the PSM technique, the estimated ATT is shown
in Table 5.
134
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
Table 6: Effects of exporting behaviour on firm performance
TFP
Entering After one year Entering After one year Entering After one year entering
year entering year entering year
0.1532 0.1789 0.0708
(0.0854)
Employment
Revenue
ATT
0.0950* 0.1526
(0.0502) (0.0765)
0.0837
(0.0662)
296
(0.0324)
381
(0.0475)
Export Starters 381
Non-exporter 2,367
381
296
296
1,320
2,367
1,320
2,367
1,320
Note: Standard errors in parentheses
Source: Author’s calculation
With the evidence in Table 6, it is reasonable to state that the learning effects from exports on TFP
are significant. Exports have a positive correlation with firms’ TFP with 15.3 per cent of the TFP growth
increase when starting exporting. In addition, it is significant that the exports of the previous period have
impacts on firm size and profit. After participating in exporting activities, firms have opportunities to
increase their employment in later years. Exporting induces a higher growth rate of revenue.
In terms of the balancing check, the quality of matching is efficient through the t-test results
(Table 7). In the matched sample, there are no significant differences in the characteristics of the
entrants serve the foreign market and the matched non-exporters.
Table 7: The t-test results for exports and non-exporters in the matched sample
Variable
Lag TFP
Lag Size
Lag FIEs
Mean (treated group)
1.5123
Mean (control group)
1.4878
t-test p value
0.965
4.0275
4.0561
0.786
0.2385
0.1638
0.668
Source: Author’s calculation
In summary, exporting participation has been found to have expected, significant and robust
impacts on firm performance, supporting the theoretical LE mechanism. Serving foreign markets is a
good way to enhance TFP for the Vietnamese manufacturing firms. It is likely that the increase in TFP
stemming from exports comes through in a later period rather than in the entry year. After participating
in the export market, under the impact of foreign markets, firms’ capital and size tend to be larger,
contributing to the increase in revenue. Firms must take advantage of their capacity and knowledge
absorbed from overseas partners for better performance in the near future
4. Conclusion
This paper intends to show the impact of WTO accession on firm productivity and exports. A
decomposition analysis shows that aggregate productivity growth is driven by the substantial
contributions of new entrant firms and the survivors are still holding a vital stabilizing role in the
economy.
We do present significant empirical evidence to show that Vietnamese exporting firms have
had higher productivity than non-exporters during 2007 -2013. In order to perform this, we tested SS
hypothesis and LE hypothesis using data from VES, Vietnam. The results show support for both SS
and LE hypotheses and this reflects the theoretical prediction of that WTO accession have had a
positive impact on firm performance.
135
Kỷ yếu Hội thảo quốc tế “Thương mại và phân phối” lần 2 năm 2020
The advantage of scale will help businesses to cover the cost of goods manufactured, creating
more competitive prices and accessing foreign markets. Therefore, firms with better performance in
comparison to others are able to participate in export markets and grow stronger. After participating in
exports, under the impact of foreign markets, firms’ capital and size increase significantly and
meanwhile profit has also improved.
For the analysis of export effects on productivity, research shows that exports have a major
impact on productivity changes and increase firms’ competitiveness. Therefore, the Vietnamese
Government should implement programs to help domestic firms intensively engage in exporting
activities. Policies of government should focus on the following three core issues: (1) focusing on
supporting micro-enterprises to participate in the internationalization process by linking micro-sized
enterprises together, linking SMEs with SOEs and FIEs, (2) enhancing absorption capacity for firms and
(3) supporting firms to seek markets and legal consulting when participating in international markets.
REFERENCE
1. Bernard, AB & Jensen, JB 1999, ‘Exceptional exporter performance: cause, effect, or both?’, Journal of
international economics, vol. 47, no. 1, pp. 1-25.
2. Bernard, A & Jensen, J 1999, Exporting and productivity, National bureau of economic research
3. Bernard, AB & Jensen, JB 1995, ‘Exporters, Jobs, and Wages in US Manufacturing: 1976–1987’,
Brookings Papers on Economic Activity: Microeconomics 1995, vol. 1995, pp. 67–119.
4. Lach, S & Tybout, J 1998, ‘Is learning by exporting important? Micro-dynamic evidence from Colombia,
Mexico, and Morocco’, The Quarterly journal of economics, vol. 113, no. 3, pp. 903-47.
5. Kneller, R 2007, ‘Industry differences in the effect of export market entry: learning by exporting?’, Review
of World Economics, vol. 143, no. 3, pp. 416-32.
6. Damijan, J, Polanec, S & Prašnikar, J 2004, Self-selection, export market heterogeneity and productivity
improvements: Firm-level evidence from Slovenia, LICOS Discussion paper.
7. Sinani, E 2010, ‘Export market participation with sunk costs and firm heterogeneity’, Applied Economics,
vol. 42, no. 25, pp. 3195-207.
8. Girma, S, Greenaway, A & Kneller, R 2004, ‘Does exporting increase productivity? A microeconometric
analysis of matched firms’, Review of International Economics, vol. 12, no.
9. Kokko, A 2012, ‘The internationalization of Vietnamese small and medium-sized enterprises’, Asian
Economic Papers, vol. 4, no. 1, pp. 152-77.
10. Jensen, 2004, Aw et al., 2002, Castellani, 2003
11. Jenkins, R 2004, ‘Vietnam in the global economy: trade, employment and poverty’, Journal of
International Development, vol. 16, no. 1, pp. 13-28.
12. Wagner, J 2012, ‘International trade and firm performance: a survey of empirical studies since 2006’,
Review of World Economics, vol. 148, no. 2, pp. 235-67.
136
11 trang
18440
Free
Bạn đang xem tài liệu "Gia nhập tổ chức WTO và năng suất của các công ty sản xuất Việt Nam: 2006 - 2013", để tải tài liệu gốc về máy hãy click vào nút Download ở trên
File đính kèm:
- gia_nhap_to_chuc_wto_va_nang_suat_cua_cac_cong_ty_san_xuat_v.pdf
