DIFFERENTIAL EFFECTS OF CREDIT ACCESS ON EMPLOYMENT GROWTH OF FORMAL AND INFORMAL FIRMS: EMPIRICAL EVIDENCE FROM VIETNAM45496

VIETNAM NATIONAL UNIVERSITY - UNIVERSITY OF ECONOMICS AND BUSINESS DIFFERENTIAL EFFECTS OF CREDIT ACCESS ON EMPLOYMENT GROWTH OF FORMAL AND INFORMAL FIRMS: EMPIRICAL EVIDENCE FROM VIETNAM Trinh Q Long1*, Peter Morgan1, Minh B Tran2 Asian Development Bank Institute Central Institute for Economic Management, Vietnam ABSTRACT This paper examines the causal effect of credit access on employment growth, using a unique micro, small, medium-sized (MSMEs) firm-level data collected every two years in Vietnam from 2005 to 2013 The results obtained from fixed-effects (FE) and fixed-effects with instrumental variable (FE-IV) estimators show that firms with credit access experience a higher employment growth than firms without credit access We also find that access to credit is positively associated with employment growth of both formal and informal firms, but the results for formal firms seem to be driven by some high growth firms (and rapidly shrinking firms) The empirical results also indicate that the effect of credit access on employment growth is also heterogenous by firm size and firm age in both types of firms Keyword: credit access; firm growth; employment growth; Vietnam JEL codes: D22, D25 INTRODUCTION It is well documented that the micro-, small-, and medium-sized enterprises (MSMEs) play an important role in economic development in many developing economies (Tybout 2000) While a better understanding of the mechanisms of firm growth is increasingly important for academics and policy makers, studies on the growth of MSMEs and its underlying drivers are few, especially in developing countries (Coad 2009) There are two major perspectives on the firm growth The first approach is the stochastic approach, which was put forward by Gibrat in 1931, argues that firm growth is an idiosyncratic process The deterministic approach, on the other hand, * Corresponding author Email address: ltrinh@adbi.org 47 IN TERNATIONAL CONFERENCE ON - CIFBA 2020 argues that the growth rate of firms differs because of observable industry and firm specific characteristics, of which firm size and age are two major determinants of firm growth (Evans 1987).1 Among the factors that affect firm growth, financing is considered as a crucial factor A firm with limited or no access to external finance may not be able to carry out its optimal investment policy which, in turn, may hamper its growth (Cabral and Mata 2003; Beck and Demirguc-Kunt 2006; Coluzzi, Ferrando and Martinez-Carrascal 2015) Empirically, numerous studies have shown that credit access lifts the burden of financial pressure and thus have a positive effect on firm growth (Coluzzi, Ferrando and Martinez-Carrascal 2015; Rahaman 2011) While the impact of credit access on firm growth is widely studied in developed countries, empirical evidence from developing countries is both limited and inconclusive (Ayyagari et al 2016) Using World Bank Enterprise Surveys, Dinh, Mavridis and Nguyen (2012) and Ayyagari et al (2016) find that firms with credit have employment growth higher than firms without credit access by - 4.2 percentage points Meanwhile, Allen et al (2012) and Beck, Lu and Yang (2015) found no relationship between access to external finance and firm growth in general and employment growth in particular Moreover, except Beck, Lu and Yang (2015) and Raj and Sen (2013), most of current literature using firm-level data from developing countries focus on formal firms We have little understanding about the relationship between access to credit and informal firm growth, despite the preponderance of informal firms in these economies (Farazi 2014).2 However, to our knowledge, the heterogeneity of credit access is rarely studied in the context of developing economies This paper examines the causal effect of access to formal credit3 on the growth of formal and informal firms using a unique longitudinal firm-level data collected every two years from 2005 through 2013 in Vietnam In this study, we use employment growth as the proxy for the firm growth since firm-level employment is usually followed and recorded, thus less subject to measurement errors More specifically, we attempt to answer (i) Is there any difference in the effect of access to credit to employment growth between formal and informal firms; (ii) Is there heterogenous effects of credit access on employment growth across the different classes of firm age and firm size We use fixed-effects estimator to identify the relationship between access to credit and employment growth Because the potential endogeneity of access to credit, to establish the causal effect of access to credit on employment growth, we exploit the variations in financial inclusion, financial depth and financial costs across industries and locations over the year to instrument for access to credit See Hart (2000) and Coad (2009) for excellent reviews on firm growth By 2015, there are about 4,800,000 non-farm informal firms operating in Vietnam, accounted for 90.5% of total number of non-farm firms (MPI 2017) Formal credit is credit from commercial banks and formal financial institutions Throughout the paper, for brevity, we use “access to credit” or “credit access” 48 VIETNAM NATIONAL UNIVERSITY - UNIVERSITY OF ECONOMICS AND BUSINESS We find a strong relationship between access to credit and employment growth Overall, firms with access to credit exhibit a growth of about 20 percentage points higher than firms without access to credit Access to credit is positively associated with employment growth of both formal and informal firms However, we find that while the results for informal firms are robust, that for formal firms are not robust and may be driven by some firms that experienced very high growth or very rapidly shrinking growth Moreover, the empirical results also show heterogeneity in the effect of credit access on the employment growth of both formal and informal firms by classes of firm age and firm size Access to credit have positively effects on informal firms aged from 10 to 20 years old and on formal firms aged less than 10 years old For both formal and informal firms, while micro ones with credit access experienced a sharp decline in growth rate, medium-sized ones enjoy much higher growth rate than their counterparts For small sized firms, access to credit has a positive and statistically significant effects on the employment growth of informal firms but not on that of formal firms Our paper contributes to the literature in several aspects First, our sample comprises of both formal and informal firms This helps our paper deviates from Dinh, Mavridis and Nguyen (2012) and Ayyagari et al (2016) which focuses on only formal firms and from Beck, Lu and Yang (2015) and Raj and Sen (2013), which focus only on informal firms Having both formal firms and informal firms in our sample allows us to see the differences in access to credit by type of firm formality Second, our study provides further evidence in the heterogenous effect of credit access and employment growth in the developing country context While literature focusing on credit access and employment growth in developed countries have shown a heterogeneity in the effects of credit access by firm size and firm age, there are quite few evidences for developing countries Ayyagari et al (2016) find a difference in the effect of finance on job growth between large firms and MSMEs firms but not see whether such difference exists among MSMEs ones Third, we examine the effect of credit access on employment growth in a single developing country to minimize the cross-country differences that cannot capture by-country dummy variables Our paper is proceeded as follows Section provides a brief literature review on the effect of credit access and firm growth and employment growth We introduce MSMEs development and credit access in Vietnam in section The empirical approach is presented in section 4, followed by explanation of data and some descriptive analysis in section We report our empirical results in section Section provides concluding remarks ACCESS TO CREDIT AND EMPLOYMENT GROWTH: LITERATURE REVIEW The impact of access to credit on the real activity of firms has been studied extensively since 1980s (see, e.g Schiantarelli 1996; Hubbard 1998; Stein 2003) The main argument is that, due to the imperfections in capital markets, the costs of internal finance and external finance are different Thus, firms will depend more on internal finance for their operation This may not be optimal for firm performance However, so far, both theory and empirical evidence focus on the effect of financing 49 IN TERNATIONAL CONFERENCE ON - CIFBA 2020 constraints on firms’ investment decisions, while fewer papers concentrate directly on firm growth (Ayyagari et al 2016, Coluzzi, Ferrando and Martinez-Carrascal 2015) Existing theories suggest a number of channels through which access to finance can impact the firms’ growth (Nickell and Nicolitsas, 1999; Spaliara, 2009; Campello, 2003) through firms’ choice of production input factors such as labor, capital and capital labor ratio Under the assumption that the capital is not fully substitutable for labor, being unable to access credit to invest in capital goods may curtail any plans they have to foster firms’ growth However, if capital and labor inputs are partially substitutable, firms that cannot access credit may increase their employees as an alternative thus altering the capital labor ratio Given this ambiguity, estimating the direction of the relationship is an empirical question In their seminal work, Rajan and Zingales (1998) show that financial development facilitates economic growth by reducing the cost of external financing Moreover, in a developed financial market, industries that are more dependent on external financing grow faster than industries that are less dependent on external finance Levin (2005) further explains that the development of financial market helps to channel credit to the most productive firms and thus, ultimately boost the growth of the whole economy Rajan and Zingales (1998) and Levin (2005)’ arguments have been extensively examined in the literature at macro level At micro-level, a growing empirical evidence have shown a positive effect of financial development on firm growth (Demirguc-Kunt and Maksimovic, 1998) Earlier studies linked financial development indicators at the macro level with firm growth for a cross section of countries or country-specific studies (see, e.g Demirguc-Kunt and Maksimovic, 1998; Beck et al 2008; Butler and Cornaggia 2007) The availability of the firm-level data has raised the interest of examining the direct effect of access to credit and firm growth in developing countries (Beck et al 2005, Ayyagari et al 2008, Dinh et al 2012, Aterido et al 2011) Most of the studies find a positive relationship between access to credit and firm growth, although some evidence shows no or weak linkage between access to credit and firm growth (Allen et al 2012; Beck, Lu and Yang 2015) Recent evidence also suggested that the effects of financial conditions, thus the ease of credit access, are heterogeneous across firms For example, Fort et al (2013) analyzed the responsiveness of firm growth to business cycles in the US for the period 1981-2010 They find that young firms were associated with a large decline in net employment growth and job creation as well as a large increase in job destruction during the 2008 global financial crisis Moreover, young firms often face difficulties in obtaining external finance because of larger costs incurred by asymmetric information; more difficult to monitor and inexperienced (HernándezCánovas and Martínez-Solano, 2010; Akoten et al 2006) The effects of credit access may also be differential across the firm size In a seminal work, Gertler and Gilchrist (1994) assess the role of credit market frictions in propagating business cycles They find that large and small firms have similar responses to easing credit conditions; however, small firms exhibit much sharper declines in sales and inventories during periods of credit market tightening relative to 50 VIETNAM NATIONAL UNIVERSITY - UNIVERSITY OF ECONOMICS AND BUSINESS large firms Similarly, Chari et al (2007) also find that small firms are more sensitive to credit shock than larger firms In fact, various studies have shown that firm size is one of the most important indicators that affect the credit access There are various reasons, both at demand and supply sides, explaining why firm size matter for firms access to finance For example, the amount that banks are asked to lend is too small to offset transaction and screening cost The cost of asymmetric information tends to be higher for smaller firms (Hernandez-Canovas and Martinez-Solano 2010) Firm size also is a signal of repayment ability (Drakos and Giannakopoulous 2011) Various empirical evidence has shown that access to credit is more difficult for smaller firms than for larger firms (Hernández-Cánovas and Martínez-Solano 2010; Drakos and Giannakopoulos 2011; Bigsten et al 2003; Malesky and Taussig, 2009; Rand, 2007) In terms of firm formality, formal firms tend to be able to access to formal credit than informal firms (Beck et al 2008) They are also face fewer problem with collateral requirement and paperwork bureaucracy (Demirgỹỗ-Kunt and Levine, 2005) Meanwhile, informal firms are more likely to be credit rationed and financial constrained due to collateral requirement as well institutional factors (Drakos and Giannakopoulos, 2011) MSME DEVELOPMENT IN VIETNAM Vietnam’s economic achievement has been remarkable as a result of its Renovation (Doi Moi) launched in 1986, particularly towards economic prosperity and human development Economic has been accelerated by rapid annual growth, averaging from 5% to 6% during 1991-2016, creating fundamental background for transforming the country to a lower middle-income country Along with the positively economic growth, MSMEs have emerged as a dynamic and central force in the development of Vietnamese economy According to the official definition of firm class stipulated in the Decree No 56/2009/ ND-CP, 95% of Vietnamese enterprises are MSMEs According to a recent report of the Ministry of Planning and Investment, formally registered MSMEs contributed around 32% of GDP in 2017, with a total investment of 36-38% of total investment into the economy MSMEs have seen exponential growth over the last decade (MPI 2018) The General Statistics Office of Vietnam reported a total of 517.9 thousand enterprises in the economy, of which 507.86 thousand firms are categorized as micro, small and medium enterprises by 2017 (GSO 2018) During 2006-2015, MSMEs sector in Vietnam has witnessed a remarkable increase in the number of enterprises, from around 45,000 enterprises in 2006, the number of MSMEs has increased to over 120 thousand in 2015, roughly 2.6 times compared to 2006 The average growth rate of this period reached the level of approximately 14 percent per year In addition, MSMEs sectors has been considered as major contributor for creating more employment opportunities Out of nearly 13 million jobs in the economy, MSMEs sectors has created 60%, or 7.8 million employments The sorting trend of enterprises has been seen more clearly after the global financial crisis in 2011, with a slow down during 2011-2014, both in terms of number of enterprises and capital The recovery of MSMEs development showed more positive 51 IN TERNATIONAL CONFERENCE ON - CIFBA 2020 trend during 2015-2017, with quite rapid movement, especially in 2016, Vietnam witnessed a new record of newly register enterprises of more than 110 thousand firms and another 126,895 firms in 2017 This positive trend can be attribute to the promulgation of new Enterprises Law and Investment Law in 2014 and various policies for improving business environment, supporting start-up of firms and facilitating the MSMEs sector Another typical feature of Vietnam’s economy has been the existence of informal sector Data from the latest Economic Establishment Census showed that, by July 2017, there were more than 5.1 million individual establishments in the economy, increasing by 11.2 as compared to that of 2012 (GSO 2018) The mentioned sector attracts 8.7 million employees (32.3% of the total employment) Of which, only 25.9% has been formally registered as household businesses while the remaining has not been registered However, to the other side, Japan External Trade Organization (JETRO 2017) found out that MSMEs in Vietnam have been encountering different barriers, the most three main obstacles include lack of financial accessibility, the ineffectiveness of support from the government, and limited business capacity as of their small scale, low level of technological advancement Such obstacles have deterred the development of MSMEs in Vietnam, preventing them from having stable position in the market and less ability in integrating into international market and global value chains Recognizing the importance of MSMEs and their growth constraints, the Vietnamese government has implemented several policies to facilitate access to finance for these firms For example, in 2017, the Law on supporting small and medium enterprises was promogulated This Law emphasizes on the necessity of financial support, including of access to credit and guarantees for SMEs Following this Law, a Decree on establishing SME credit guarantee funds at the provincial level was issued in order to facilitate better access to credit In 2013, the SME Development Fund (SMEDF) was established This fund provides funds for firms that produce high quality products, use new materials and energy, innovation of equipment and technology, create more jobs for women, and invest in environmental-friendly technology development Despite these positive developments, financing for SMEs and sustainable production remains limited There still exists a big gap between demand and supply of formal credit Nearly 80% of MSMEs in Vietnam finance their investment projects by internal sources, rather than external sources (CIEM, DoE, ILSSA, 2014; Yoshino and Wignaraja, 2015) Some 26% of MSMEs applied for a formal loan, and 24% of them had problems getting the loan Access to credit has worsened for micro and small, and rural enterprises; however, it has increased for medium enterprises (CIEM, DoE, ILSSA, 2014) There are several reasons behind the undersupply of credit Firstly, SMEs often not approach banks because they lack information on financing opportunities or because they not have the capacity to comply with application procedures Collateral requirements also hinder MSMEs to approach financial institutions Secondly, banks are reluctant to lend to SMEs because of high costs for screening and monitoring credits and administration procedures Many financial institutions still prefer to cater to large enterprises and have tightened 52 VIETNAM NATIONAL UNIVERSITY - UNIVERSITY OF ECONOMICS AND BUSINESS credit conditions for SMEs since the global financial crisis in 2008 Thirdly, the underdevelopment of credit rating and guarantee frameworks also plays a significant role in limiting credit access for MSMEs EMPIRICAL STRATEGY Following Rahaman (2011), Coluzzi, Ferrando and Martinez-Carrascal (2015) and other literature and using the fix-effects (FE) estimators, we estimate our benchmark growth equation as follows: (1) Where: captures time fixed-effects; captures firm fixed-effects; is error terms The dependent variable, , is the employment growth rate of firm at time and measured as difference in logarithm of employment at time t and time t-1.4 is a dummy variable, which takes the value of one if firm accessed to formal credit, i.e., access to credit from commercial banks or other formal financial institutions, during period and zero otherwise.5 is a vector of control variables at time to mitigate the possible endogeneity issue Since the data is collected at the end of period, the values of the variables at time could be assumed to be the initial conditions for the firm operating at time We follow literature to control for the following variables: - Firm size: Firm size is predominantly identified by the extant industrial economics literature as one of the sources of heterogeneity in employment growth Smaller firms may have more potential for growth (Rahaman 2011; Coad, Segarra and Teruel 2016) In our study, firm size is represented by the number of employed at the end of time t-1 - Firm age: While some argues that younger firms may experience more turbulent growth trajectories and have a higher chance of failure than older firms (Rahaman 2011), some evidence shows that younger firms grow more quickly (Jovanovic 1982; Coad, Segarra and Teruel 2016) - Growth opportunities: The annual growth rate of sales is introduced as a proxy for growth opportunities as in Coluzzi, Ferrando and Martinez-Carrascal (2015) Sales (and wages) are deflated using the sectoral producer price index, realeased by the General Statistics Office The base year of this index is 2010 - Other control variables: We control for owners’ technical education, industry For robustness checks, we use an alternative definition of employment growth Please refer to the Appendix for definition and estimation results Informal firms could access to credit from formal financial institution by using their physical asset or the owners’ access (such as house/land0 as collaterals 53 IN TERNATIONAL CONFERENCE ON - CIFBA 2020 dummies, province dummies, year dummies, interactions between industry and year dummies and between province and year dummies To examine the differential effect of credit access by classes of firm size and firm age, we estimate the following equations: 𝑗𝑗 𝑗𝑗 (2) 𝑗𝑗 𝑗𝑗 (3) ′ 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔ℎ𝑖𝑖𝑖𝑖 = 𝛼𝛼0 + ∑ 𝛼𝛼1 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑖𝑖𝑖𝑖 ∗ 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑖𝑖𝑖𝑖−1 + 𝑋𝑋𝑖𝑖𝑖𝑖−1 𝛼𝛼2𝑠𝑠 + 𝛿𝛿𝑡𝑡 + 𝜇𝜇𝑖𝑖 + 𝜖𝜖𝑖𝑖𝑖𝑖 𝑗𝑗 ′ 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔ℎ𝑖𝑖𝑖𝑖 = 𝛼𝛼0 + ∑ 𝛼𝛼1 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑖𝑖𝑖𝑖 ∗ 𝐴𝐴𝐴𝐴𝐴𝐴𝑖𝑖 + 𝑋𝑋𝑖𝑖𝑖𝑖−1 𝛼𝛼2𝑠𝑠 + 𝛿𝛿𝑡𝑡 + 𝜇𝜇𝑖𝑖 + 𝜖𝜖𝑖𝑖𝑖𝑖 Where: 𝑗𝑗 Dependent variable (growthit) and control variables ( ) are similar to equation (1) Since we are interested in the differential effects of credit access on employment growth by classes of firm size, we divided our sample into three groups of firms We follow Vietnam’s official definition of firm size6 to divide firms into three groups: micro firms, which consists of firms with number of employees less than 10 people, small firms (consisting firms with employees from 10 to 49) and medium firms (with more than 50 employees)7 In equation (2), indicates firm i belonging to firm size class j (j = 1.3) at time t-1 For firm age, we categorized firm into three classes, those with less than 10 years of operation, those from 10 to 19 years of operation and those with more than 20 years of operation Our empirical strategy requires that only firms that participated in three consecutive surveys to be eligible in the estimation sample Since we are also interested in examining the differential effects of access to credit by firm type (i.e formal and informal firms)9, we would estimate equations (1), (2) and (3) separately for informal firms and formal firms In addition, we also estimate heterogeneous effects of credit access by firm type using the following equation: 𝑓𝑓 𝑖𝑖𝑖𝑖𝑖𝑖 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔ℎ𝑖𝑖𝑖𝑖 = 𝛼𝛼0 + 𝛼𝛼1 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑖𝑖𝑖𝑖 ∗ 𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖𝑖𝑖−1 + 𝛼𝛼1 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑖𝑖𝑖𝑖 ∗ 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑖𝑖𝑖𝑖−1 Where: ′ +𝑋𝑋𝑖𝑖𝑖𝑖−1 𝛼𝛼2 (4) + 𝛿𝛿𝑡𝑡 + 𝜇𝜇𝑖𝑖 + 𝜖𝜖𝑖𝑖𝑖𝑖 𝐹𝐹𝐹𝐹𝑟𝑟𝑖𝑖𝑖𝑖−1 (𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑖𝑖𝑖𝑖−1) is a dummy variable which takes value of one if firm is a formal (an informal) firm at t-1 and zero otherwise Credit access may be potentially endogenous, either because of reverse causality or existence of unobservable factors that affect both credit access and employment growth To deal with this, we use the fixed-effects with instrumental variable (FE-IV) estimator We exploit local variation in financial development Micro firm is defined in the Vietnamese Government’s Decree No 56/2009/ND-CP None of firms in the sample for this study have more than 300 employees Since we control for sales growth in the last period (i.e the sale growth from t-2 to t-1, only firms that participated at least three consecutive surveys are included in our estimations According Vietnamese Government’s Decree No 56/2009/ND-CP, formal firms include limited liability companies and joint stock companies (i.e all firms operating under the Enterprise Law) while informal firms include household business, cooperatives and private business entities 54 VIETNAM NATIONAL UNIVERSITY - UNIVERSITY OF ECONOMICS AND BUSINESS as instrument for access to credit Moreover, we also take into account the fact that external finance dependence is different from industry to industry (Rajan and Zingales 1998) in construction of our measurement of financial development Three variables are used to measure the financial development: (i) Financial inclusion, which is measured as the percentage of firms in a given industry located in a province being able to access to credit; (ii) Financial depth, which is calculated as the average of the ratio of formal credit to total sales in a given industry-province; (iii) Financial cost, which is measured as the average ratio of interest payments to total sales in a given industry-province level.1011 These three variables are usually used in the literature as indicators for financial development (share of firms that can access to formal credit; the depth of financial development and the cost of formal credit) Various studies have shown that higher financial development will allow firms easier to access to finance and reduce the cost of borrowing (O’Toole and Newman 2017; Fafchamps and Schündeln 2013), which in turn affect firm performance We assume that firm’s credit access is determined by two components: financial development at the industry-province level ad firm-specific components The first component could be viewed as a function of specific features that shape that industry-province, including the common industry-specific production techniques, common industry-specific and province-specific values and traditions, and common labor pools (and compositions) These features may determine the demand and supply of credit different from industry to industry and from province to province Eventually, they determine the extent to which the opportunity to access to credit would be common to all firms in the same industry and province in each year If the industryprovince component is uncorrelated with the time-varying unobservable factors, then we can use the industry-province component as the instrumental variable for the credit access of a specific firm There are, however, time-varying factors that may be correlated with the industry-province component To account for such potential time-varying unobservable factors, following Fafchamps and Schundeln (2013), in our estimations, we control for growth opportunities at the industry-province level The growth opportunities at time t are calculated as the average of the sales and employment growth at the given industry-province level at time.12 We categorize firms into seven industries: food-processing, garment and textiles, wood and furniture, chemicals, non-metal materials, machinery and other industries based on the 2-digit code Vietnam Industrial Code 1993, which corresponds with ISIC revision On average, each industry-province cell has about 48.9 firms (sd: 38.2) 11 To calculate the average, for given year, we include all (available) firms, including those that are not satisfied to be included in our estimation sample (i.e firms not participated in three consecutive surveys) 12 For robustness check, the growth opportunities are measured as the highest growth rate of firms in a given industry-province level at time The results are upon request 10 55 IN TERNATIONAL CONFERENCE ON - CIFBA 2020 DATA AND DESCRIPTIVE ANALYSIS 5.1 Data sources The data is jointly collected by the University of Copenhagen (Denmark), the Central Institute for Economic Management (Vietnam), and the Institute for Labor Studies and Social Affairs (Vietnam) every two years from 2005 through 2013 in 10 provinces in Northern Vietnam (Ha Noi, Vinh Phuc, Ha Tay, Hai Phong), Central Vietnam (Nghe An, Quang Nam, Khanh Hoa), and Southern Vietnam (Lam Dong, Ho Chi Minh City, Long An and Dong Nai).13 These 10 provinces cover about 30 percent of manufacturing firms in Vietnam In each province, the selection of the sample was stratified by ownership types to ensure an adequate number of enterprises with different ownership types, including private, collectives, partnerships, private limited enterprises, and joint stock enterprises After each survey round, the sample firms that had exited or stopped cooperating with the survey were replaced by randomly selected firms from a firm list compiled by the General Statistics Office of Vietnam (for formal firms) and firm list provided by local authorities (for informal firms) The sample, therefore, included a wide range of firms operating in Vietnam: private, collectives, partnerships, private limited enterprises, and joint stock enterprises The inclusion of domestic private firms (both formal and informal firms) made this survey unique compared with other firm-level datasets in Vietnam such as GSO’s annual Enterprises Survey or World Bank’s Enterprise Survey, which cover only formal firms (i.e firms that are registered and have a tax code) In each of ten selected provinces, the selection of the sample was stratified by ownership types to ensure an adequate number of enterprises in each province with different ownership forms There were about 2500 firms selected in 2005 based upon the above-mentioned sample selection approach These firms continued to be surveyed in the subsequent rounds of survey (i.e 2007, 2009, 2011 and 2013) The survey instrument (i.e questionnaire) are also consistent across the rounds of survey Information collected included the firm’s and owner/managers’ production, sales and markets, and some other characteristics The questionnaires also contained questions about internationalization activities the firms have undertaken 5.2 Descriptive analysis Table presents descriptive analysis In this study, our sample includes 5,696 firmyear observations (1,862 firms) Of which 3,908 observations (1,286 firms) are informal firms and 1,658 observations (568 firms) are formal firms Most of firms in our sample are aged from 10 to 20 years (account for about 35-45%) The share of informal firm aged more than 20 years old is much larger than share of formal firms aged more than 20% The descriptive data also shows that for informal firms, most of informal firms are micro firms (account for about 80%) Only 1-2% of informal firms could be categorized as medium firms Meanwhile, more than half of formal firms are small firms (with number of employees ranged from 10 to 50) About one 13 For details, see CIEM, DoE and ILLSA (2014) 56 VIETNAM NATIONAL UNIVERSITY - UNIVERSITY OF ECONOMICS AND BUSINESS whole sample while that in column and are from estimation with two subsamples of informal firms and formal firms respectively Kleibergen-Paap statistic and CraggDonald Wald F-statistic show that our instrumental variable does not suffer from underidentification, weak instrument problems Meanwhile Hansen J statistics show that our set of instrumental variables satisfied the overidentification test The estimation results still show that credit access has a positive effect on employment growth, but the magnitude of estimates is larger On average, access to credit raises the employment growth by 20.6 percentage points The figure is slightly lower for informal firms (column 3) Access to credit leads to employment growth by 19.0 percentage points for the informal firms and about 21.8 percentage points for the formal firms This result is consistent with Coluzzi, Ferrando and Martinez-Carrascal (2015), which shows that firm without financial obstacles may have enjoyed a higher growth than firms with financial obstacles by 13-47 percentage points, depending on the prevalence of SMEs in the economy In fact, the SMEs account for more than 95% of the firms (MPI, 2017) The results also suggest that there is differential effect of access to credit by firm formality Formal firms with credit access seem to create more jobs than their informal counterparts Results in column based on the average of all firms in the sample While there is a large different between informal and formal firms, we estimate two subsamples separately (column and 5j we see that both formal firms and informal firms with access to credit still enjoyed a higher growth rate than their counterparts and formal firms still have larger effects on employment growth However, we could see that the coefficient is weaker, only significant at 10% level On average, the effects of credit access are larger for formal firms than for informal firms and this differential is statistically significant at 1% level With regards other control variables, we not see any significant change in estimates in comparison to the results from the FE estimation However, we find some different among informal firms and informal firms Columns and show that firm size has negative associated with employment growth for both types of firms, but this magnitude of estimates is larger among informal firms than formal firms Similarly, we see that firms that pay higher average wage sees to employed more It is also interested to note that older formal firms seem to experience higher employment growth than younger one while firm age have no effects on employment growth among informal firms We also find that owners who have university is associated with higher growth among formal firms but have no effect on the employment growth of informal firms Meanwhile, those with only 3-year college degree is correlated with employment growth among the informal firms Coad (2009) and Coad, Segarra and Teruel (2016) show that firm growth may not normally distributed, but tent shaped with long tail at both sides of the distribution curve, suggesting that some firms may enjoy very high growth or shrink very rapidly while most of firms not grow or grow slowly Existence of some very high growth firms (and very rapidly shrinking firms) in the sample may drive our results presented in Table To tackle with this issue, we dropped the 5% of firms with lowest growth and 5% of firms with highest growth rate from our sample and reestimate table 59 IN TERNATIONAL CONFERENCE ON - CIFBA 2020 Table Effects of credit access on employment growth   Credit access (1) (2) (3) 0.206*** (4) Informal firms, FE-IV 0.153** (5) Formal firms, FE-IV 0.161* All firm FE All firm, FE-IV All firms, FE-IV 0.073*** [0.019] [0.063] [0.064] [0.095] 0.190** Credit access * Informal firms [0.077] 0.218** Credit access * Formal firms Firm size (lagged) Being a formal firm (lagged) Firm age Lagged average wage Lagged revenue growth Owner attended vocation school Owner attended 3-year college Owner attended university Located in urban area [0.089] -1.066*** -1.070*** -1.068*** -1.129*** -0.970*** [0.022] [0.025] [0.024] [0.030] [0.042] 0.089* 0.090* 0.078 0.043 0.015 [0.050] [0.048] [0.066] [0.109] [0.058] 0.249** 0.242** 0.242** 0.095 0.469*** [0.101] [0.095] [0.096] [0.127] [0.163] 0.069*** 0.069*** 0.069*** 0.050*** 0.132*** [0.016] [0.016] [0.016] [0.019] [0.033] 0.065*** 0.057** 0.058** 0.019 0.106*** [0.024] [0.026] [0.026] [0.033] [0.040] 0.045 0.049* 0.048* 0.038 0.063 [0.028] [0.029] [0.029] [0.031] [0.085] 0.074** 0.080** 0.079** 0.110** -0.019 [0.037] [0.037] [0.037] [0.044] [0.070] 0.099** 0.107*** 0.106** 0.029 0.120* [0.042] [0.041] [0.041] [0.068] [0.066] -0.015 -0.056 -0.061 -0.090 0.054 [0.199] [0.175] [0.173] [0.237] [0.198] 0.131 0.120 0.140 0.256 [0.110] [0.111] [0.136] [0.232] Sales growth at industry-province level Labor growth at industry-province level   0.135 0.160 0.082 0.160 [0.135] [0.289] [0.149] [0.289] Kleibergen-Paap rk LM statistic 93.097 87.085 52.28 43.07 Cragg-Donald Wald F statistic 39.718 28.575 22.464 17.092 Hansen J Statistics (p-value) 0.2774 0.276 0.2244 0.7209 Numbers of firms 1862 1862 1862 1286 568 Observations 5696 5696 5696 3908 1658 Note: Robust standard errors in brackets; statistical significance: * p < 0.1; ** p < 0.05; *** p < 0.01 Column 1, the standard FE is adopted, while columns -5 present results from FEIV estimations In all specifications, we control for industry dummies, province dummies, year dummies Two variables, average sales growth and employment growth, are used as measures of growth potential in the region Source Authors’ calculation 60 VIETNAM NATIONAL UNIVERSITY - UNIVERSITY OF ECONOMICS AND BUSINESS Table Effects of credit access on employment growth: Robustness check with trimmed sample   Credit access (1) (2) (3) 0.158*** (4) Informal firms, FEIV 0.208** (5) Formal firms, FEIV 0.091 All firm FE All firm, FE-IV All firms, FE-IV 0.043*** [0.015] [0.053] [0.091] [0.089] 0.178** Credit access * Informal firms [0.080] 0.132* Credit access * Formal firms Firm size (lagged) Being a formal firm (lagged) Firm age Lagged Average wage Lagged revenue growth Owner attended vocation school Owner attended 3-year college [0.074] -0.797*** -0.800*** -0.799*** -0.863*** -0.697*** [0.026] [0.024] [0.024] [0.033] [0.037] 0.028 0.027 0.045 -0.124 0.022 [0.038] [0.037] [0.054] [0.092] [0.044] 0.104 0.095 0.090 0.018 0.289** [0.076] [0.074] [0.074] [0.096] [0.129] 0.037*** 0.038*** 0.038*** 0.025* 0.083*** [0.013] [0.013] [0.013] [0.015] [0.025] 0.070*** 0.067*** 0.066*** 0.042 0.092*** [0.017] [0.018] [0.018] [0.028] [0.025] 0.006 0.008 0.008 -0.004 -0.000 [0.023] [0.023] [0.023] [0.025] [0.062] 0.013 0.016 0.016 0.009 -0.021 [0.028] [0.028] [0.028] [0.031] [0.055] 0.059* 0.065** 0.064** 0.025 0.093* [0.033] [0.032] [0.032] [0.048] [0.051] -0.015 -0.056 -0.061 -0.090 0.243** [0.199] [0.175] [0.173] [0.237] [0.120] 0.131 0.120 0.140 0.256 [0.110] [0.111] [0.136] [0.232] 0.135 0.160 0.082 0.160 [0.135] [0.289] [0.149] [0.289] Kleibergen-Paap rk LM statistic 173.32 120.611 83.716 58.079 Cragg-Donald Wald F statistic 108.01 89.454 78.757 30.822 Owner attended university Located in urban area Sales growth at industry-province level Labor growth at industry-province level    Hansen J Statistics (p-value) 0.1076 0.1408 0.1689 0.4422 Numbers of firms 1697 1697 1697 1163 520 Observations 4979 4979 4979 3397 1452 Note: Robust standard errors in brackets; statistical significance: * p < 0.1; ** p < 0.05; *** p < 0.01 Column 1, the standard FE is adopted, while columns -5 present results from FE-IV estimations In all specifications, we control for industry dummies, province dummies, year dummies Two variables, average sales growth and employment growth, are used as measures of growth potential in the region Source Authors’ calculation 61 IN TERNATIONAL CONFERENCE ON - CIFBA 2020 The estimation results with trimmed sample are reported in Table We find that the effects of access to credit has reduced for both FE and FE- IV estimators This indicates that our previous result may be driven by firms with extreme performance Interestingly, results from trimmed sample shows that the effects of access to credit is larger for informal firms than for formal firms This result is reversed with previous result (column in Table 2), suggesting that formal firms tend to be extreme cases than informal firms This is confirmed when we estimate two subsamples separately (columns and 5) We find that access to credit is still have a positive and statistically significant effects on informal growth, it loses its effect for formal firms Coefficient on access to credit variable is still positive but not statistically significant Meanwhile, other regressors are qualitatively similar to the previous results For further robustness check our results, we replicate tables and with employment growth defined as Davis et al (2002) Tables A1 and A2 in the Appendix present our estimation results Qualitatively, we find the similar results as presented above, i.e access to credit facilities employment growth and formal firms that access to credit finance tends to raise employment growth among informal firms at the higher rate than that of formal firms Effects of credit access by class of firm size and firm age Tables and report differential effect of credit access on employment growth by classes of firm age and firm size The whole sample is used in column of two tables while subsample of informal firms and formal firms are used in columns and 3, respectively As presented above, we categorize firms into three different classes based on their age and their size Interaction between variable credit access and three classes of firm age and firm size will create three more potentially endogenous variables To deal with this, we interact our main instrumental variable (financial inclusion) with three dummy variables corresponding to the class that a firm belongs to Other instrumental variables are financial depth and financial cost As before, Kleibergen-Paap statistic and Cragg-Donald Wald F-statistic suggest that our instrumental variable does not suffer from underidentification, weak instrument problems Our instrumental variables also satisfied the overidentification test The result in column 1, Table shows that access to credit have differential effects across firm age classes Access to credit have stronger effect on employment growth among younger firms For example, credit access may help firms aged less than 20 years to grow by 23.5 percentage points than counterpart firms without credit, while the figure for firms aged more than 20 years are only 15.5 percentage points This difference is statistically significant at 1% level This result suggests that firms that operated (and survived) very long may have accumulated a certain amount of wealth and therefore access to credit is not as sensitive to their growth as younger firms The results, however, are different when we examine informal and formal firms separately (columns and 3, Table 4) The effect of credit access is only statistically significant among informal firms aged from 10 to 20 years old and among formal firms aged less than 10 years old Access to credit may help informal firms aged from 10 to 20 years old to grow by 19.9 percentage points The figure for formal firms aged 62 VIETNAM NATIONAL UNIVERSITY - UNIVERSITY OF ECONOMICS AND BUSINESS less than 10 years old is 22.2% The effect of credit access on formal firms aged less than 10 years old is statistically significant at 10% level For other firms, access to credit not have any statistically significant effects on employment growth Table Credit access and employment growth, by class of firm age   Credit access * Age