The impact of loans to small and medium enterprises the case study of vietnam

UNIVERSITY OF ECONOMICS HO CHI MINH CITY VIETNAM ERASMUS UNVERSITY ROTTERDAM INSTITUTE OF SOCIAL STUDIES THE NETHERLANDS VIETNAM – NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS THE IMPACT OF LOANS TO SMALL AND MEDIUM ENTERPRISES: THE CASE STUDY OF VIET NAM By BUI THI HONG CHINH MASTER OF ARTS IN DEVELOPMENT ECONOMICS Ho Chi Minh City, January 2018 i UNIVERSITY OF ECONOMICS HO CHI MINH CITY VIETNAM ERASMUS UNVERSITY ROTTERDAM INSTITUTE OF SOCIAL STUDIES THE NETHERLANDS VIETNAM – NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS THE IMPACT OF LOANS TO SMALL AND MEDIUM ENTERPRISES: THE CASE STUDY OF VIET NAM By BUI THI HONG CHINH Supervisor Dr NGUYEN THI THUY LINH Ho Chi Minh City, January 2018 ii ACKNOWLEDGEMENT This process of writing a thesis is a collaborative experience involving the support and helps from many people I want to express my gratitude to those who give me the tremendous support to complete this thesis First of all, I would like to thank gratefully to my supervisor Dr Nguyen Thi Thuy Linh I cannot finish my thesis if not get the support and the advice from her Beside that, I want to thank to Dr Pham Khanh Nam who give me many useful and important advice that help me very much through the time I the thesis Moreover, I want to thank Dr Truong Dang Thuy who helps me answer questions when I need it And I want to thank the professors and the teacher staff of Viet Nam – Netherlands Program that gave many supports to me to have the knowledge, to solve the difficult problems in my studying process Last, I want to thank my closet friends and my family iii ABBREVIATIONS SME Small and Medium Enterprise RD Regression Discontinuity Design PSM Propensity Score Matching Method DD Difference in difference VIF Variance inflation factor IV Instrumental variable HET Heteroscedasticity SOE State Owned Enterprise OECD Organization for Economic Co-operation and Development DNNN State enterprises NHNN State Bank NHTM Commercial Bank iv ABSTRACT After a period of growth and affected by the crisis, Vietnam's economy has decreased Hundreds of thousands of small and medium-sized enterprises went bankrupt and shut down Loan is a solution for business to expand scale, increase sales and profits, but it can create jobs, increase salaries to improve social welfare or not? To verify that argument, the author uses the PSM method and combines with DD on the SME data set from 2009 to 2013 to more accurately assess the impact of the loans The results show that loans not have the effect of improving employee incomes, as well as creating more jobs In addition, the loans from informal sources with low cost not help enterprises to expand their operations because of the small scale Loans from official sources are large scale, but the high costs overwhelm profits Moreover, the impact of formal loans also causes businesses to reduce their jobs The topic also shows other factors such as export, type of ownership, scale, production technique, entrepreneurial qualification that affects to employment and wage v TABLE OF CONTENTS ACKNOWLEDGEMENT III ABBREVIATIONS .IV ABSTRACT V CHAPTER 1: INTRODUCTION 1.1 PROBLEM STATEMENT 1.2.1 Research objectives 1.2.2 Main research question CHAPTER 2: LITERATURE REVIEW 2.1 REVIEW OF THEORY 2.1.1 Definition of small and medium enterprises and types of credits 2.1.2 The impact of loans to employee from producer theory 2.1.3 Factors affecting the operation of the business 2.2 REVIEW OF EMPIRICAL STUDIES 10 2.2.1 Impact of loan to SMEs in Viet Nam 10 2.2.2 Previous researches 12 2.3 SUMMARY 13 CHAPTER 3: RESEARCH METHODOLOGY 14 3.1 ANALYTICAL FRAMEWORK 14 3.2 ECONOMETRICS MODELS 15 3.2.1 Impact assessment methodology 15 3.2.2 Research proposal and select model 18 3.2.3 Dependent variables 21 3.2.4 Independent variables 21 3.3 DATA 24 CHAPTER 4: RESEARCH RESULTS 25 4.1 OVERVIEW OF THE RESEARCH TOPIC 25 4.2 DESCRIPTIVE STATISTICS 28 vi 4.3 REGRESSION RESULTS 30 4.3.1 OLS regression results 30 4.3.2 PSM combined with DD results 32 4.3.3 Verification of model stability 41 4.4 DICUSSIONS 42 CHAPTER 5: CONCLUSIONS AND POLICY IMPLICATIONS 47 5.1 CONCLUSIONS 47 5.2 POLICY IMPLICATIONS 48 5.3 LIMITS OF THE STUDY 49 REFERENCES 50 APPENDICES 55 APPENDICES 1: DIVIDE THE SIZE OF THE BUSINESS 55 APPENDICES INFLATION AND PRICE INDEX VND (1994=1) 56 APPENDICES INFLATION AND PRICE INDEX VND (1994=1) (CONT) 56 APPENDICES IMPACT ASSESSMENT BY MATHEMATICAL METHOD 57 APPENDICES GROUPS WERE DEVIDED BY PSM METHOD 58 APPENDICES REGRESSION DISCONTINUITY DESIGN - RD 59 APPENDICES INSTRUMENTAL VARIABLE - IV 60 APPENDICES DEFINITION OF SOME VARIABLES 61 APPENDICES DESCRIPTIVE STATISTICS 62 APPENDICES 10 ANALYSIS OF CORRELATION BETWEEN QUANTITATIVE VARIABLES 64 vii LIST OF TABLE TABLE 3.1 DESCRIPTION AND MEASUREMENT VARIABLES 22 TABLE 4.1 DATA STATISTICS 25 TABLE 4.2 STATISTICS DESCRIBE THE PARTICIPANTS AND THE CONTROL GROUPS BEFORE THE LOAN 28 TABLE 4.3 IMPACT OF LOAN TO SMALL AND MEDIUM ENTERPRISES- BASIC MODEL 31 TABLE 4.4 REGRESSION MODEL OF THE LOANS TO SME 32 TABLE 4.5 TREND POINT OF GENERAL SUPPORT AREA 34 TABLE 4.6 IMPACT OF LOANS TO SME ON THE LABOUR COSTS 35 TABLE 4.7 IMPACT OF LOANS TO SME ON THE NUMBER OF EMPLOYEES 37 TABLE 4.8 IMPACT OF EACH TYPE TO SME ON THE LABOUR COSTS 38 TABLE 4.9 IMPACT OF EACH TYPE TO SME ON THE NUMBER OF EMPLOYEES 40 TABLE 4.10 INVESTMENT AND LABOUR 42 TABLE 4.11 THE SCALE OF THE MOST IMPORTANT LOAN 43 viii LIST OF FIGURE Figure 2.1 Illustration of the impact of loan to SMEs………………………………5 Figure 2.2 Optimal coordination of production factors when loan increases…….…6 Figure 3.1 Impact of loans to SMEs when enterprises participate and not join in loans………………………………………………………………………….…….15 Figure 3.2 Impact assessment by DD method……………………………… … 18 Figure 3.3 Illustrate the general support area and the observation area discarded with PSM…………………………………………………………………… ……19 Figure 4.1 The supply of formal credit………………………………………….…26 Figure 4.2 The supply of informal credit………………………………………….27 Figure 4.3 The biggest difficulties prevent the development of SMEs……………44 ix REFERENCES Acevedo, G L., & Tan, H W (Eds.) (2011), “Impact evaluation of small and medium enterprise programs in Latin America and the Caribbean”, World Bank Publications ADB (2014), “Asia Small and Medium-sized Enterprise (SME) Finance Monitor 2014” Adorno, V., Bernini, C., & Pellegrini, G (2007), “The impact of capital subsidies: new estimations under continuous treatment” Giornale degli economisti e annali di economia, page 65-95 Asdrubali (1991), ‘Monitoring and Reputation: The Choice between Bank Loans and Directly Placed Debt’, Journal of Political Economy, 99(4), page 685-723 Bentolila, S., Jansen, M., Jiménez, G., & Ruano, S (2013) “When credit dries up: Job losses in the great recession?” Bentolila, S (2004), “Basic Econometrics”, New York: McGraw Hill Berger, N (1989), “Giving Women Credit: The Strengths and Limitations of Credit as a tool for Alleviating Poverty” World Development Vol 17, No.7, page 1015-1033 CIEM (2010), Business Environment in Vietnam: Survey of Small and Medium Enterprises in 2009, Labor and Social Affairs Publish CIEM (2012), Characteristics of Business Environment in Vietnam, Survey Results of Small and Medium Enterprises in 2011, Labor Export - Social Institute CIEM (2014), Business Environment in Vietnam: Survey of Small and Medium Enterprises in 2013, Financial Publish Claessens, S., S Djankov and L.C Xu (2000), “Corporate Performance in the East Asian Financial Crisis,” World Bank Research Observer, 15(1), page 22-47 50 De Mel, Suresh, D McKenzie, and C Woodruff (2008a), “Are Women More Credit Constrained? Experimental Evidence on Gender and Microenterprise returns” American Economic Journal: Applied Economics 1(3): 1-35 De Mel, S., McKenzie, D., and Woodruff, C (2008b), “Returns to Capital: Results from a Randomized Experiment.” Quarterly Journal of Economics 123 (4): 1328–1373 Dinh Tuan Minh, To Trung Thanh, Edmund Malesky, Nguyen Duc Thanh (2010), "Impact of interest rate support policy on enterprises' operations." VEPR Policy Discussion Gertler, M., & Gilchrist, S (1991), “Monetary policy, business cycles and the behavior of small manufacturing firms”, (No w3892), National Bureau of Economic Research Government (2001), Decree No 90/2001/ ND-CP on supporting the development of small and medium enterprises Government (2009), Decree 56/2009 / ND-CP on supporting the development of small and medium enterprises Hansen, H., Rand, J., & Tarp, F (2009), “Enterprise growth and survival in Vietnam: does government support matter?” The Journal of Development Studies, 45(7), 1045-1070 Harvie, C., Narjoko, D., & Oum, S (2010), “Firm characteristic determinants of SME participation in production networks”, ERIA Discussion paper series, 11 Hulme, D (2000), “Impact assessment methodologies for microfinance: Theory, experience and better practice” World development, 28(1), 77-99 Kokko, A and Sjoholm, F (2004), “The Internationalization of Vietnamese SMEs, Stockholm School of Economics”, Asian Economic Papers, Vol.4, No.1 51 Ministry of Planning and Investment (2015), Circular No 13/2015/TT-BKHDT promulgating List of priority areas for support and criteria for selection of priority beneficiaries of the Enterprise Development Fund Small and medium Nguyen Kim Anh (2011), Microfinance for Poverty Reduction in Viet Nam Accreditation and Comparison, Statistical Publishing House, Hanoi Nguyen Xuan Thanh (2016), "Should money support business?", Cafef, accessed on 08/01/2016,at:http://cafef.vn/business/Should-money-support-business201601081114- 32353.chn OECD (2006), “The African Economic Outlook 2005-2006”: Kampala Uganda Oh, Inha., Lee, J D., Heshmati, A., & Choi, G G (2009), “Evaluation of credit guarantee policy using propensity score matching” Small Business Economics, 33(3), 333-354 Petersen, MA and RG Rajan (1994), 'The Benefits of Firm-creditor Relationship: Evidence from Small Business Data', Journal of Finance, 49(1), page 1-38 Pindyck, Robert S and Daniel L Rubinfeld (2013), “Microeconomics”, 8th edition, Upper Saddle River New Jersey: Pearson Prentice Hall Ramanathan, R (2002), “Introductory Economictrics with Applications” 5th edition, Harcourt College Publisher Pham, T T T., & Lensink, R (2008), “Is microfinance an important instrument for poverty alleviation” Pham Chi Lan (2016), "How is the private sector being encroached?", Cafef, accessed 29/01/2016, at: http://cafef.vn/business/ pham-chi-lan-how-is-the-privatesector-being-encroached-?-20160129071020078.chn Phan Thi Linh (2015), "Solutions to access bank capital for small and medium enterprises", Financial Journal, No 11, page 15-20 52 Rubin, T., Demirguc-Kunt, A and Levine, R (2005), "SMEs, Growth, And Poverty: Cross-Country Evidence," Journal of Economic Growth, v 10(3, Sep), 195-230 Ruiz, C., Love, I (2012), “Impact assessment framework: SME finance Washington DC”; World Bank Signore, S., Pierfederico, A (2015), “The Economic Impact of EU Guarantees on Credit to SMEs Evidence from CESEE Countries” EIF Research & Market Analysis Working Paper 2015/29 Tran Hoang Ngan (2015), "Credit support, low interest rates for SMEs", Cafef, accessed 29/01/2016, at: http://cafef.vn/micro-investment/tran-hoang-ngan-supportcredit-low-interest-rate-for-business-20151110131214835.chn The World Bank (2007), "Vietnam: Building a Comprehensive Strategy to Improve Access to Microfinance Services (of the Poor)," World Bank The Chamber of Commerce and Industry of Vietnam (VCCI) (2015), "Why are more and more private enterprises going bankrupt?", NDH, accessed 29/01/2016, http://ndh.vn/why-are-more-and-more-private-enterprises-going-bankrupt2015111903424- 125p4c147.news The Prime Minister (2013), Decision No 601/QD-TTg on the establishment of the Small and Medium Enterprise Development Fund Tran Dinh Thien et al (2015), "Chapter 7: Developing and liberalizing capital markets", Vietnam Development Market Report, Tri Thuc Publish Tran Hoang Nhi (2016),"Why Small and Medium Enterprises Can’t Grow?",saigonbusinessman,accessed06/04/2016,at:http://www.saigonbusinessman vn/problem/micro-star-business-industry-king-warehouse/1096384/ Truong Tan Sang (2013), "Five years after the global financial crisis in Vietnam", Vnexpress, dated29/01/2016, at http://business.vnexpress.net/photo/microeconomics/ Five-years-after-the-global-financial-crisis-in-Vietnam-2877946.html 53 Shahidur R.Khandker, Gayatri B.Koolwal, Hussain A.Samad (2010), Impact Assessment Handbook - Quantitative Methods and Practices, The World Bank Wang, X (2013), “The Impact of Microfinance on the Development of Small and Medium Enterprises: The Case of Taizhou, China” Asian journal of business and management sciences, 2(9) 54 APPENDICES Appendices 1: Divide the size of the business Size Too small business Small business Medium business Number of Total capi- Number of employees tal employees Total capital Number of employees 10 people or less Under 20 billion VND From 10 to 200 people From 20 billion to 100 billion VND From 200 to 300 people 10 people or less Under 20 billion VND From 10 to 200 people From 20 billion to 100 billion VND From 200 to 300 people 10 people or less Under 20 billion VND From 10 to 50 people From 10 billion to 50 billion VND From 50 to 100 people Area I Agriculture, forestry and fishery II Industry and construction III Trade and services Source: Article 1, part 3, Decree 56/2009 / ND-CP dated 30/6/2009 55 Appendices Inflation and price index VND (1994=1) Year 1994 Inflation Price index 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 17.00% 8.72% 6.60% 8.85% 5.76% 3.40% 1.92% 3.95% 6.67% 8.22% 1.000 1.170 1.272 1.356 1.476 1.561 1.614 1.645 1.710 1.824 1.974 Source: Calculated from the Ministry of Finance and CIEM Appendices Inflation and price index VND (1994=1) (cont) Year 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Inflation 8.16% 7.26% 8.25% 21.70% 5.70% 10.00% 18.68% 9.09% 6.59% 4.09% Price index 2.135 2.290 2.479 3.017 3.189 3.508 Source: Calculated from the Ministry of Finance and CIEM 56 4.163 4.542 4.841 5.039 Appendices Impact assessment by mathematical method The basic mathematical presentation is as follows: Yi = αXi + βTi + εi Signed for businesses i: Ti: is the dummy variable If the ith join, it will get the value of 1; opposite, it will get the value of Yi|Ti : the variable that has the result of business activities under condition T Average treatment effect – ATE for all business is: ATE = E(Yi|Ti=1) – E(Yi|Ti=0) = E(Yi|Ti=1) – E(Yi0|Ti=1) + E(Yi0|Ti=1) – E(Yi|Ti=0) = ATT + Selection Bias In practice it is impossible to observe the counter-facto situation Yi0|Ti=1, so it is impossible to estimate the exact ATE 57 Appendices Groups were devided by PSM method With attnd command in nearest – neighbor matching method, we select the nearest n observations for each value range Each unit that participates in the policy, will be compared to a opposite unit with the nearest trend point (n = commonly used) or uses a non-participating object to collate with different participants With attr command in radius matching method, we choose to observe the given value range In the attnd statement there are very high trend difference points (the participants are not very close together) This problem makes poor quality comparisons and should be avoided by setting thresholds or tolerances above the maximum trend point (within range) We only compare there alternatives among trend points within a certain range However, if the number of participating companies is high, it will likely increase the sampling error With atts command in stratification matching method, we compare each given value range It divides the support capital into several levels and calculates the impact of the program in each space For example, in each space, the impact of the program would be the median difference in outcomes between the intervention and control observations The weighted average of the estimates which affect this space will show the overall program impact with the proportion of participants in each space that is weighted With attk command in kernel matching method, we use an internal regression with non-parametric The risk is that there is a small group of non-participants that satisfies the criteria within the overall support and results in counter-measures Without the number of parameters such as the internal kernel and linear comparison, the weighted average of all non-participants is used as a counter-argument for each participant In contrast, linear regression calculates region-weighted regression with no quantitative number on the results of the comparison group near to each intervential observation 58 Appendices Regression Discontinuity Design - RD This method selects program participants based on certain criteria, at which point is the cut point For example, if you want to get credit for the poor, the participants must first earn a poor income Then, select the above and below cut points for comparison Subjects that are close to the cut point are often assumed to have similar characteristics If the policy is effective, the difference will be very clear around the cut point Impact assessment using RD Y= Labour productivity + + + ++ + + ++ Control group Participating group ++ + + + ++ + + + + 20 Labours X= Number of employees Source: Ruiz, Claudia; Love, Inessa (2012) Figure 7, page 27 59 Appendices Instrumental variable - IV This method use to overcome the endogenous state, the offense assumption, the goal is to clean up the correlation between Ti and ei in equation (3.1), meaning cov (T, ε) ≠ The reason is that the selection error from the unobserved characteristics will be included in the estimated surplus The remainder will contain variables that have correlations with the dummy variables This will make the normal OLS estimate to be biased (Ramu Ramanathan, 2002) This way will be done in two steps, called Two-Stage least squares - 2SLS: Step 1: Regression Ti = αZi + ui to find the probability of participating in the program of the business i This variable Zi must be correlated with T (cov (Z, T) ≠ 0) and not correlated with εi (cov (Z, ε) = 0) Step 2: Using Ti from step to regression Yi = βTi + εi And β will measure the effect of the program Choice of tool variables is a difficult problem, if poor quality variables can increase the error, or higher than standard OLS calculations So, we need to prepare well before using method IV 60 Appendices Definition of some variables Number of regular employees: full time employees, contracted to work or work more than months, average working 20 days per month or 20 hours per week, or 183 days per year Regular and sufficient time reflects the size of the business Regular employees are often self employed or family members Number of employees paid in the enterprise (PaidLabour): are labor chosen from society This shows the ability of the enterprise to create jobs for employment, contributing to solving the problem of creating jobs for laborers 61 Appendices Descriptive statistics Labour (person) Salary (million dong) Total assets (million VND) Business features AGE8 MICRO8 HOUSEHOLD8 OWNLAND8 HAND8 EXPORT8 SOUTH8 LowTech8 Characteristics of the business owner HAge8 HGen8 Kinh8 Edu8 ProEdu8 Economic characteristics NSLD8 (million VND/person/year) ShareDept8 Source: SME 2009 Control group Number of observations = 1042 mean sd max 11,58 22,43 270 71,32 181,3 1942 Participating groups Number of observations = 574 Mean sd max 25,30 40,11 350 182,2 387,5 5206 989,5 2848 2,221 39148 1773 4103 3,977 42562 15,99 0,798 0,752 0,666 0,0873 0,0422 0,309 0,601 12,03 0,402 0,432 0,472 0,282 0,201 0,462 0,490 0 0 0 55 1 1 1 13,10 0,589 0,592 0,585 0,0383 0,0801 0,277 0,523 9,751 0,492 0,492 0,493 0,192 0,272 0,448 0,500 0 0 0 55 1 1 1 46,71 0,670 0,917 0,530 0,146 10,52 0,470 0,277 0,499 0,353 20 0 0 90 1 1 43,96 0,676 0,956 0,636 0,226 9,916 0,468 0,204 0,482 0,419 22 0 0 72 1 1 55,810 94,202 1243 0,0689 0,168 2,1e+06 79,426 128,762 6125 1,7e+06 2,396 0,163 0,346 Labor: the control group has an average of 11 - 12 laborers, while the participation group is 25 - 26 people The control group ranged from to 270 with a standard deviation of 22 The participants were to 350 and 40, respectively, who were more than 13 to 14 participants in the control group and statistical meaning This shows that enterprises with loans have a higher number of employees 62 6,092 Total Assets: The average asset of the participating group was 1.7 billion and 0.9 billion of the control group Group assets of the participating group are also higher from million to 42 billion The difference in mean total assets was a control group of 783 million compared to the control group in a statistically significant Salaries: Since the participants have more assets and more labor, the average salary for labor to participant group is more than 110 million Average attendance of the group is VND 182 million between and 5.2 billion; the control group was 71 million VND between and 1.9 billion VND AGE: the mean of participation group operates 13-14 years, while the control group is 15-16 years The control group operated more years than loan group about years MICRO: the participants were significantly larger than the other group The control group was over 79% of the control group, while the participants were 58% HOUSEHOLD: 75% of the control group was household enterprises, while the participating group had a lower rate, only 59% HAND, EXPORT and LowTech: The participants in the loan had more points than the control group: more exports (more than 3.8%), using hand tools (less than 4.9%), low technology (less than 7.8%) Edu: The participation group graduated high school at 63% compared to 53%, and college qualifications (ProEdu) was at 22% compared with 14% HAge: The mean age of participating group was 43-44 years, compared with 46-47 of the control group NSLD: The participating group was 79 million VND/person/year compared with 55 million VND/person/year of the control group, an average of 24 million/person/year in a statistically significant ShareDept: the participation group was 16% while the control group was 6% 63 Appendices 10 Analysis of correlation between quantitative variables LNWAGE lnLABOR LNASSET lnNSLD AGE HAge LNWAGE 1.0000 lnLABOR 0.7383* 1.0000 LNASSET 0.6593* 0.7181* 1.0000 lnNSLD 0.2926* 0.2136* 0.3289* 1.0000 AGE -0.1948* -0.1832* -0.1213* -0.1288* 1.0000 HAge -0.1441* -0.1504* -0.0659* -0.1014* 0.2664* 1.0000 ShareDept 0.1065* 0.1371* 0.0608* 0.0528* -0.0382* -0.0597* ShareDept 1.0000 * Statistically significant, meaning level 5% Quantitative variables have no correlation The highest correlation coefficient was only 0.73 between the lnLABOR and LNWAGE variables, and the between lnLABOR and LNASSET was 0.71 Less more 0.7 for the correlation of the remaining variables The quantitative independent variables are not strongly correlated with each other, so the model will have less multi-collinearity The two independent variables (AGE and HAge) are negative with the dependent variable (LNWAGE, lnLABOR), the remainder is the positive relationship The LNASSET variable has the highest positive correlation for two dependent variables compared to the other independent variables 64 ... operation of the business The subject of the study is the impact of loan to small and medium enterprise in Viet Nam, in other words, the causal relationship of the loan to the salary and employment of. .. the loan to SMEs impact on the labour costs? Does the loan to SMEs impact on the number of employees? Does the each type of credit to small and medium enterprises impact on the labour costs and. .. and the number of employees? 1.3 Scope of study This thesis investigates the impact of loan to employee on the number of employees and the labour costs by using the dataset of Survey of Small and