The Effect of Exports on Growth of Small and Medium-Sized Enterprises: Evidence from Vietnamese Manufacturing Firms

Abstract

The paper aims to examine the impact of exports on the growth of Vietnamese manufacturing small and medium-sized enterprises (SMEs) by exploring the information of 36,053 enterprises across 24 manufacturing sectors from the Vietnam Annual Enterprise Survey (VAES) in the period 2014-2019. To deal with the problem of variable variance, autocorrelation, and endogeneity of the model, the paper uses the OLS regression method with a strong standard error method and system GMM. Export participation by SMEs is positively associated with business growth in terms of sales and total assets, according to the findings. The GMM estimate shows that the rate of sales growth among exporters is 36.5 percent greater than that of non-exporting enterprises in the case of the sales growth measure. Exporters' average total asset growth rate is 19% greater than the rate estimated for non-exporting businesses. The study's findings indicate the need of adopting policies that promote SMEs in transition economies like Vietnam to engage in exporting activities. Furthermore, the findings show that financial assistance and suitable ownership would enable SMEs to take advantage of export opportunities to increase sales and total assets.

1. Introduction

SMEs have traditionally played a crucial role in the economy, serving as the engine of growth, providing jobs, and boosting the economy’s innovation and competitiveness (Maggioni, 2012). SMEs need to grow to survive. At the firm level, a company that continues to grow has a better chance of surviving in the market. As a result, research has been conducted to determine the key factors of SME growth. Many studies look at data from SMEs all around the world to see what qualities of businesses and owners contribute to firm growth. When it comes to business features, access to overseas markets is frequently cited as a key component in enhancing corporate growth (Nam & Bao Tram, 2021). Because of its low risk, cheap entry cost, high flexibility, and low resource utilization, exporting is the most typical strategy for SMEs to join overseas markets. As a result, the entrepreneur’s financial cost and risk are lower than with other foreign entry modes (Manolova et al., 2010).

In the literature on international economics and growth, the economic relationship between export and firm growth has long been a disputed topic. Despite universal agreement that only high-performing organizations can overcome the sunk cost of beginning exports (Self Selection), there is much disagreement concerning the benefits of exports to corporate performance (Learning-by-Exporting). Participating in international trade, namely exporting, has long been thought to be a viable means of supporting businesses in reaching their growth potential (Manolova et al., 2010). Firms gain experience by exporting, gaining access to new technology and new products, and improving their products to better serve the home and foreign markets. However, the benefits of exporting on company growth have been disputed in some circumstances (Liu et al., 1999) or shown to be dependent on a variety of factors, including industry, firm ownership, and degree of internationalization (Ngo & Nguyen, 2020).

Previous research, on the other hand, has paid little attention to the impact of export gains on the expansion of SMEs, which are characterized by a lack of innovation, low productivity, and a low level of internationalization. Furthermore, some of the research mentioned above contain methodological flaws. The majority of the variables in the literature-based growth equations are likely to be endo- genous. Although some writers address the endogenous nature of the export factors in their models (Ngo & Nguyen, 2020), the majority of research utilizes simple fixed or random effect estimates that ignore endogeneity. As a result, their findings should be regarded with caution. My equations are estimated using a generalized method of moments (GMM) system estimators that account for the endogenous nature of all regressors, which is an improvement over previous work.

According to a review of the literature, there hasn’t been much research in Vietnam on the effects of export on SMEs’ growth. The Vietnamese context is appealing because, since ‘Doi Moi’ (renovation) in 1986, this Southeast Asian country has experienced a significant restructuring process. Vietnam has a reputation for being one of the world’s most open economies, with active participation in a number of international trade accords (Nguyen et al., 2021).

Vietnam is following an export-oriented growth model, with manufacturing sectors accounting for a large portion of the country’s total export value (Nguyen, 2020). The effects of exporting on business growth determine whether the Vietnamese government should concentrate the majority of its resources on promoting exports. As a result, this is a critical issue for a developing economy like Vietnam. This research examines the impact of exports on the growth of Vietnamese manufacturing SMEs using a dataset of Vietnamese manufacturing SMEs and the GMM estimator, which adjusts for both endogeneity and unobserved, time invariant factors.

2. Literature Review

There is a growing body of empirical research on whether selling in international markets boosts a company‘s performance (Castellani, 2002; Wagner, 2007). Selling in international markets improves performance by allowing businesses to take advantage of the expanded scale economies associated with a combined domestic and international market, or by providing firms with enhanced learning capabilities through access to new quality standards, new technology, and knowledge spillovers (Manolova et al., 2010).

According to Wagner (2007), the influence of export on firm growth is greater in developing nations, where the percentage of exporters and international integration is low.

Other research, on the other hand, showed only shaky evidence of the beneficial effects of exporting on exporter growth rates (Castellani, 2002) or no evidence of firm growth enhancement following international market entry. Leichenko (2000) used regional data for the United States from 1980 to 1991 to illustrate the impact of exports on the company in a study on Mexican plants. According to the author, rising international market sales will have a detrimental impact on the employment expansion of the company.

Exporting operations did not have a substantial impact on business growth in Taiwan, according to Liu et al. (1999), who observed no significant performance differences between exporters and non-exporters. Export effects on company growth, according to Ngo and Nguyen (2020), are highly dependent on individual sectors. Given the conflicting and ambiguous results of previous research on the effects of exporting on business growth, it is worthwhile to look into the case of SMEs in manufacturing in Vietnam, a developing country that has been increasingly exposed to export and overseas markets over the last three decades.

3. Conceptual Model and Data

3.1. Conceptual Model

In the present study, the author’s empirical model to evaluate the role of the export on the SMEs’ growth follows Castellani (2002) and has the following form:

\(\text { Growth }_{i t}=\beta_{0}+\beta_{1} \mathrm{EXP}_{i t-1}+\beta_{2} \text { Growth }_{i t-1}+\beta_{3} \text { Firm }\\ \text { Control variables }{ }_{i t-1}+\beta_{4} \text { Owner Control } \text { variables }_{i t}+\varepsilon_{i}\)       (1)

In the model, the dependant variable refers to the growth rate of sales and total assets. The coefficients β₁, β₂, β₃, β₄ and are parameters to be estimated.\(\varepsilon_{1}\) is the stochastic error term. t and i denote year and firm, respectively, in the model. The author includes firm-specific characteristics measured in year t –1. Firm control variables include firm size, firm age, leverage, firm ownership. Regarding owner control variables, factors related to the entrepreneur (gender, age, and academic level) are measured in year t. Dummies for years and two-digit ISIC industries are also included to control for yearly and industry fixed effects. The advantage of the model is that: the cross-sectional nature of the data, together with the use of lagged regressors, reduces the endogeneity problems between export and sales/total asset growth.

The model is used to test the hypothesis on the relationship between export and firm growth. The export premium, β₁, shows the average percentage difference between exporters and non-exporters in the same industry. It is expected that ​​​​​​​β₁ > 0 to prove that exporting positively impacts the sales/total asset growth of SMEs.

Dependent variable:

We used two measures to capture the growth of Vietnamese manufacturing SMEs in different dimensions. We computed the annual growth rate of net sales and total assets for each firm.

GRSALE: Firm sales growth - the difference in the natural logarithm of net sales between two consecutive years.

GRASSET: Firm total asset growth - the difference in the natural logarithm of total assets between two consecutive years.

Independent variable:

LAG_EXP: To test the effect of exporting on SME growth, the study uses a dummy variable LAG_EXP, which takes the value of 1 if a firm export in year t – 1, otherwise, receives the value of 0.

LAG_GRSALE: Lagged Sales Growth.

LAG_GRASSET: Lagged total asset growth

Firm Control variables:

For causal conclusions, the author lagged all the firm control variables, as in previous research (Castellani, 2002). The author includes the following firm control variables in the model: The size of the SME is determined by the log of net sales and the log of total assets, respectively. LIMITED is a dummy variable that represents whether the company is a limited company or a joint-stock company. LEV is the financial leverage of the enterprise; FAGE is the firm age, which is calculated using the natural logarithm of the firm’s age; and LIMITED is the dummy variable that represents whether the company is a limited company or a joint-stock company.

Owner control variables:

In SMEs, the most important decisions are made by one or a few people. Therefore, owner/manager traits can influence the decision-making process and firm growth. Some owner characteristics are included in the model as follows: MIDDLE AGE (equals 1 if owner/manager is more than 40 years old); GENDER (equals 1 if the owner/ manager is male); EDUCATION (One if the owner/manager has a college, undergraduate or postgraduate degree).

Apart from those above variables, the author also adds dummies for years and two-digit ISIC industries to control for yearly and industry fixed effects.

Two hypotheses are set up based on regression results:

H1: Exporting activities has a positive impact on the SMEs sales growth.

H2: Exporting activities have a positive impact on the SMEs total asset growth.

3.2. Data

The Vietnam Annual Enterprise Survey (VAES), which is conducted annually by the General Statistical Office (GSO) of Vietnam, is the data source for this study (Table 1). The VAES was created to offer annual data on the performance of businesses in a variety of industries. At the firm level, the surveys collected data on the activities of the enterprises, including ownership, location, industry, labor force, and wages, assets and liabilities, exports and imports of goods, and business outcomes (revenue, costs of goods, administrative costs, and net profit).

Table 1: Data Description

The definition of SMEs used in this study is as follows: stand-alone businesses with fewer than 300 employees, as defined by the Vietnam SME Law. From 2014 to 2019, we collected data on all Vietnamese SMEs in the manufacturing sector, as defined by Vietnamese law. Furthermore, the data is collected on a yearly basis, allowing for the development of a longitudinal profile of an SME’s internationalization process. The VSIC 2007 and VSIC 18 industry classification systems are used here, and they closely correlate to the 4^th Revision of the International Standard Industrial Classification of All Economic Activities (ISIC Rev. 4).

Table 1 shows the descriptive data for our important variables. The GRSALE has a standard deviation of 0.5096 and a mean value of 0.0549. There are businesses that are quite excellent when the revenue growth rate is high, 1.7999 times over the previous year, and there are also businesses that are inefficient when the sales growth rate is negative, the smallest being –1.4498 times, among the SMEs participating in the survey.

Overall, sales grew at a 5.49 percent annual rate. The mean value of the GRASSET is 0.0980, with a standard deviation of 0.3773. A relatively positive picture of the growth rate in Vietnamese manufacturing SMEs may be obtained using two growth metrics. The average value of the LAG EXP variable is 0.0642, indicating that 6.42 percent of manufacturing SMEs participate in international commerce via exporting. LNSALE’s average firm size is 8.4650 million VND, which is equal to the natural logarithm of 15655.12 million VND. LNASSET has an average of 8.8432, which is the natural logarithm of 15025.41 million VND. With a standard deviation of 0.6143 and a mean value of 1.9928, the FAGE variable indicates that the average age of sampled firms is 8.76 years. Leverage is on average 49.57 percent. The majority of the companies in the sample (88.47 percent) are limited or joint-stock companies. In terms of owner characteristics, 66.67 percent of the sample’s owners/ managers are over 40 years old, 59 percent of the managers are male, and 78.44 percent have a college, undergraduate, or postgraduate degree.

The correlation matrix between the independent variables is shown in Table 2. The association between LNSALE and LNASSET is substantial, according to the findings. However, in two distinct models to predict sales growth rate and asset growth rate, LNSALE and LNASSET are independent variables of company size. As a result, multicollinearity does not occur in any of the models. The model’s flaws have also been tested by the authors. The authors used the Breusch-Pagan test to determine the variance and found the following results: Chi2(1) = 3222.14 with Prob > chi2 = 0 for the model to estimate sales growth. Chi2(1) = 2095.76 with Prob > chi2 = 0 for the model to estimate total asset growth. As a result, Prob > chi2 = 0.0000 (statistical significance at 1%) indicates that both models exhibit variable variance. The study employs the robust OLS approach to overcome these infractions (strong standard error method).

Table 2: Correlation Matrix Among Variables

Note: GRSALE: the natural logarithm of revenue between two consecutive years; GRASSET: the difference in the natural logarithm of a total asset between two consecutive years. LAG_EXP = 1 if firm export in year t – 1, otherwise equals to 0; LNSALE is the size of the SME by the natural logarithm of net sales); LNASSET is the size of the SME by the natural logarithm of the total asset); LEV: Total liabilities/Total assets; FAGE: the natural logarithm of the age of the firm, LIMITED = 1 if the company is a limited company or a joint-stock company, otherwise equal to 0; MIDDLE AGE = 1 if owner/manager is more than 40 years old, otherwise equals to 0; GENDER = 1 if the owner/manager is male, otherwise equals to 0; EDUCATION = 1 if the owner/manager has a college, undergraduate or postgraduate degree, otherwise equals to 0.

Moreover, while estimating the model, we realize that they are dynamic panel data models including a lagged variable with small T (time-year) and large N (number of panel firms). The potential association between the lagged variable and the past or present realization of such “small T and big N” model errors should be explored. The author uses the system GMM estimator proposed by Arellano and Bond (1991) to overcome the endogeneity problem.

In the process of estimating the model by system GMM, the author considers yearly and industry dummies, firm lagged control characteristics and owner characteristics as exogenous variables and the rest as endogenous variables. For three reasons, the GMM estimator outperforms the fixed-effects estimator. For starters, it aids the author in accounting for time-invariant effects that are not detected. Second, it can deal with any endogeneity that the lagged dependent variable may introduce. Finally, system GMM works nicely with my data, which includes “small T and huge N.”

4. Results

To tackle the phenomena of variable variance, auto-correlation, and endogeneity, Table 3 shows the regression results utilizing the OLS regression method with a strong standard error approach and a system GMM. The p-values of the Hansen and Sargan tests in all GMM models show that the hypotheses that all instruments are valid cannot be rejected. There is a significant first-order autocorrelation and no evidence of significant second order autocorrelation, according to the p-values for AR (1) and AR (2). It signifies that the test data show that all models are properly specified.

Table 3: Regression Results

Note: GRSALE: the natural logarithm of revenue between two consecutive years; GRASSET: the difference in the natural logarithm of a total asset between two consecutive years. LAG_EXP = 1 if firm export in year t–1, otherwise equals to 0; LNSALE is the size of the SME by the natural logarithm of net sales); LNASSET is the size of the SME by the natural logarithm of the total asset); LEV: Total liabilities/Total assets; FAGE: the natural logarithm of the age of the firm, LIMITED = 1 if the company is a limited company or a joint-stock company, otherwise equal to 0; MIDDLE AGE = 1 if owner/manager is more than 40 years old, otherwise equals to 0; GENDER = 1 if the owner/manager is male, otherwise equals to 0; EDUCATION = 1 if the owner/manager has a college, undergraduate or postgraduate degree, otherwise equals to 0. Note: Robust standard errors in parentheses ***p < 0.01; **p < 0.05; *p < 0.1. Significant at the 0.05 level.

In terms of the study‘s primary findings, Table 3 shows that, across annual horizons, exporters show significantly faster growth than non-exporters in terms of total assets and particularly, in sales (1 percent significance level). All regression tests show that this is true for the two growth measures studied. The GMM estimate shows that the rate of sales growth among exporters is 36.5 percent greater than that of non-exporting enterprises in the case of the sales growth measure.

When looking at the total asset growth rate, a similar pattern emerges: the average total asset growth rate of exporters is 19% greater than the rate projected for non exporting enterprises. As a result, exporting operations in Vietnamese manufacturing SMEs had a positive effect on SME sales and total asset growth rate, confirming the two hypotheses. Companies can grow their consumer base and enhance sales by exporting to other markets. Furthermore, if exporters serve more demanding markets, they can only benefit from exports through learning and competitive consequences. As a result, they may be able to better serve the domestic market. To satisfy market demand, the production volume and capacity may expand as the sales volume increases. Exporting to international markets is viewed as a crucial way to grow the business because it expands the market and creates the potential for expansion. The findings of this study are comparable to those of previous studies (Golovko & Valentini, 2011).

Regarding firm control variables in the model, the significant and negative coefficient of firm size (by either LNSALE or LNASSET) in all models indicates that small firms have a higher growth rate than large firms. This supports Gibrat’s law (suggested that the growth of the firm is independent of the firm size) in the sample of manufacturing SMEs in Vietnam. Moreover, the firm age variable (FAGE) has a significance level of 1% in the OLS but insignificant in system GMM. The coefficient of Leverage (LEV) is positive and significant in all models, suggesting that the more external capital Vietnamese SMEs can raise, the more sales and total assets they can expand. Similarly, in all models, the coefficient of company ownership (LIMITED) is positive and significant at 1%. It means that a firm with a limited or joint-stock ownership structure can grow more quickly than a company with a different ownership structure. Similarly, in all models, the coefficient of company ownership (LIMITED) is positive and significant at 1%. It means that a firm with a limited or joint-stock ownership structure can grow more quickly than a company with a different ownership structure.

Owner Age (MIDDLE AGE) is significantly negative in all models for the owner control variables, whereas Owner Education (EDUCATION) is significantly positive. The sales and total assets growth of companies with middle-aged owners is lower than that of companies with young owners. It‘s also been proven that the more education executives have, the faster their companies grow. Furthermore, in the study, the gender of the management (GENDER) had no significant impact on the firm‘s growth.

5. Conclusion

High-growth SMEs have been proved to be important contributors to national economic development. As a result, research on SMEs’ growth factors is gaining a lot of traction. The impact of exporting operations on the expansion of SMEs in the manufacturing sector in Vietnam is investigated in this article. This sector is recognized as the driving engine of Vietnam’s economic growth due to its prominent role in industrialization. The major findings of our research show that exporting will result in a faster rate of growth for SMEs in terms of sales and total assets. Scholars, practitioners, and policymakers will all benefit from this research.

From a research perspective, the findings suggest that export is important for sales growth while also highlighting the importance of export for total asset growth in SMEs. In keeping with earlier arguments, the study’s findings are positive for practitioners, implying that enterprises, particularly small and medium-sized businesses, might benefit significantly from export efforts. Exporting may be viewed as a viable growth option by Vietnamese SMEs. Finally, the findings of this research have important policy consequences. Export promotion policies should be implemented to encourage enterprises to engage in foreign trade to boost overall economic growth.