1
Leverage and investment of SMEs under a predominantly state-owned bank lending
environment: Evidence from Vietnam
By Nina Avramova
n.a.avramova@student.rug.nl
s2506181
Supervisor: Prof. Dr. C. L. M. Hermes
Msc Finance
Faculty of Economics and Business University of Groningen
2 Abstract
This study examines the leverage-investment relation of SMEs in Vietnam. The analysis is unique as it focuses on informationally opaque firms in a predominantly state-owned bank lending environment. The analysis yields three results. First, the relation between leverage and investment in the sample of SMEs is positive. This is consistent with the argument that SMEs rely significantly on bank funding and create long-term relationships with banks to improve their access to capital. Second, the positive relation between bank loans and investment is significantly stronger for low-growth than for high-growth firms. This result is consistent with the argument that state-owned banks in Vietnam provide additional credit to low growth firms or to firms in specific sectors, regions or industries with lower growth rates as part of their mandate to subsidize targeted borrowers. Third, the positive relation between debt and investment is stronger for firms with non-state owned bank borrowing than for firms with state-owned bank borrowing but this result is only robust to exploiting, producing or processing industries and not service and trading industries.
JEL classification: G32
Keywords: Leverage-investment relation; SMEs; Banks; Relationship lending; Vietnam
1. Introduction
An important issue in finance is whether leverage affects firm’s investment decisions. Modigliani and Miller (1958) demonstrate that in a world with frictionless and complete markets, leverage is irrelevant to a firm’s investment policy and its value. However, in a world with incomplete markets and asymmetric information, leverage may have varied and complex impact on investment.
3 This study is unique because in comparison to large firms, SMEs are more financially constrained and bank dependent which creates serious challenges in SME finance (Norden, 2015). Ang et al. (2000) point out that SMEs have unique features that distinguishes them from large firms, among which the reduced value of limited liability, higher bankruptcy costs, and lower agency costs. In comparison to larger firms, SMEs may lack audited financial statements which increases the ‘informational friction’ between borrowers and creditors (Pollard, 2003). SMEs are not required to disclose private information to prospective investors, regulators or creditors and therefore are not able to send credible signals to banks, venture capitalists or trade creditors (Ang et al., 2000). The lack of transparency means that SMEs have difficulty in accessing external capital which limits the accumulation of revenue-generating assets and SMEs‘ potential growth (Vos et al., 2007).
When it comes to access to capital, what matters most to SMEs are personal relationships and social interactions (Vos et. Al, 2007). Cole (1998) shows that prospective lenders are more likely to provide credit to SMEs when pre-existing transactions are present. This is intuitive because previous banking relations help convey private information about the SMEs’ financial history and prospects. The studies of Petersen and Rajan (1994), Berger and Udell (1995), and Cole (1998) also find evidence that firm-creditor relationships generate important information about borrower quality. According to Norden (2015) banks play an extremely important role in financing SMEs and SMEs are more bank-dependent in comparison to large firms that can also access public debt markets and raise equity capital. I believe that this dependency and the relationship that SMEs build with banks have an important implication on the leverage-investment relation for this type of firms.
Various empirical studies exist that investigate the link between leverage and capital investment decisions for public companies in developed countries. Lang et al. (1996), Aivazian et al. (2005a) and Ahn et al. (2006) all report a negative linear relation between leverage and investment although the link is much stronger for firms with low growth. This evidence is consistent with the argument that high leverage in low growth firms discourages managers to undertake negative NPV projects which means that banks and debt-holders have a beneficial role in monitoring low growth firms and limiting the overinvestment bias of managers. Additionally, Lang et al. (1996) argue that a negative relation between investment and leverage can arise because managers of firms with valuable growth opportunities choose to maintain lower leverage in order to be able to take advantage of the investment opportunity when it arises. Several empirical studies - Gaver & Gaver, 1993; Goyal et al., 2002; Jung et al., 1996; Smith & Watts, 1992 - confirm this theory.
4 public companies. I argue that the standard theory of the leverage-investment relation is challenged because of the crucial role of banks in SME financing and the resulting relationship than can have both a bright and a dark side. The strong bank-firm relationship helps reduce the asymmetric information between the bank and the firm so that the bank is better informed about the firm’s future prospects and can extend credit at more favorable terms when the need arises despite existing bank loans. The bank may also provide intertemporal smoothing when the firm experiences adverse business conditions. This is the bright side of relationship lending and can lead to a positive relation between leverage and investment. The dark side of relationship lending arises when the bank extracts additional rents from the firm because it is difficult for it to switch lenders (Rajan, 1992). This situations can create hold-up problems so that the bank extends credit to the firm despite previous loans in order to receive higher rents in the future (Petersen and Rajan, 1995). As a result, the dark side of relationship lending can also lead to a positive relation between leverage and investment.
Another reason this study is unique is because there are only a handful of studies on the leverage-investment relation in the developing nations but only for public companies and they are mainly focused on China. Although China has initiated a transition to a market economy, still most borrowing comes from state-owned banks (Firth et el., 2008). The Vietnamese banking industry, on the other hand, comprises a diverse mix of players, ranging from relatively larger state-owned commercial banks to small privately held banks. (Duxton Asset Management, 2015).
In this paper I argue that the relation between leverage and investment for SMEs in Vietnam operating in a predominantly state-owned bank lending environment is different than the one found for public companies because of SME’s significant reliance on bank financing and therefore formulate the following broad research question: Is there a positive relation between leverage and investment of SMEs in Vietnam? I try to answer the following more detailed questions: Is there a difference in the leverage-investment relation between firms with state-owned and non-state owned bank borrowing? Is there a difference in the leverage-investment relation between firms with high and low growth opportunities?
5 2. Literature Review and Hypotheses Development
Prior evidence of the leverage – investment relation
A widely debated issue in finance is whether leverage affects firms’ investment policies. On one side are Modigliani and Miller (1958) who maintain that the leverage level is essentially irrelevant. They argue that a firm’s capital structure is irrelevant and a firm with good projects can always find funding no matter its amount of leverage. Those on the other side, however, argue that in a world with incomplete markets and considerable agency costs, leverage may have more complex impact on investment. The existing capital structure literature suggests that a negative relation between leverage and investment arise because high leverage reduces the firm’s ability to finance future investment opportunities (Lang et al., 1996). Consequently, a negative relation exists when managers of firms with good investment prospects choose lower leverage in order to be able to raise capital when the investment opportunity arises (Lang et al., 1996). Hence, management chooses the leverage level according to its private information about the future investment prospects (Lang et al., 1996). Public firms are extremely careful for raising their leverage level above a certain level because of the considerable costs of financial distress that creates a tendency for the firm to engage in actions that are harmful for its debt holders (Grinblatt and Titman, 2008). As a result, financially distressed firms find it difficult to obtain credit when the leverage level is high. Lang et al. (1996), Aivazian et al. (2005a) and Ahn et al. (2006) all report a negative linear relation between leverage and investment based on data from listed firms from the U.S. and Canada. Firth et el. (2008) and Jiang and Zeng (2014) study the leverage-investment link of public firms in China and also find a negative relation.
6 Banks ensure the intermediation between surplus and deficit spending units in an economy (Seward, 1990). They can credibly commit to monitor the return of their investments because they achieve economies of scale and scope in gathering and processing information (Williamson, 1987). In their role of delegated monitors, banks are very careful in financing SMEs. Since reliable information on SMEs is rare and costly, relationship lending is considered the most important lending technology (Norden, 2015). Through relationship lending, the bank relies on soft information gathered through long-term contact with the SME, its owner or manager and the local community (Berger and Udell, 2006). The obtained soft information can include assessment of SME’s future prospects gathered from past communications with SME’s customers, suppliers or competitors (Berger and Udell, 1995). Relationship lending assures access to capital for the firm and access to information for the bank (Berger, 1999). Norden (2015) use a meta-analysis to determine if relationship lending provides benefits for the borrowing firms and find that SMEs obtain more credit and/or lower rates under this lending technology.
In this paper, I argue that the standard leverage-investment theory may not apply to SMEs because of the uniqueness of these firms and their significant bank dependence for raising capital which distinguishes them from public firms (Norden, 2015). Bank financing involves a long-term relationship that can help reduce the asymmetric information problems that are more pronounced in small firms whereas public financing usually does not have this feature (Berger and Udell, 1995). This difference between large, public firms and SMEs can alter the leverage-investment link for small firms because the bank is better informed about SME’s future investment prospects and can adjust its financing when the need arises despite previously obtained loans.
Additionally, borrowing firms may have incentives for moral hazard in both strong and weak bank relationships. If the relationship with the bank is strong, an important borrower in financial distress may have an incentive to count on a “too big to fail” effect and get additional funds from the bank (Norden, 2015). A risky borrower, on the other hand, has an incentive to hide knowledge from the bank in order to benefit from more favorable lending terms and continuous flow of funds (Norden, 2015). Both of these situations entail that the firm has access to additional capital despite previous bank loans so that this can alter the leverage-investment link in case of relationship lending.
7 tangibles in the firm’s assets, the harder it is to provide collateral in order to obtain finance (Freixas and Rochet, 2008). According to Menkhoff et al. (2006), SMEs in less-developed economies have lower probabilities of holding collateralisable assets and as a result, these firms most often have difficulties in obtaining external financing. Beck et al. (2006) examine 12 financing obstacles and find out that collateral requirements are the third most often rated obstacle by 8,000 small and medium-sized firms from 80 developed and developing countries in a survey by the World Bank.
Maintaining a relationship with a bank improves the information dissemination and is beneficial to both parties. (Scholtens, 1999) This way the bank could make a better risk assessment of the project (screening) and can exert control (monitoring) over its outcome. The SME, on the other hand, benefits from more favorable lending conditions which improve over time such as lower cost of capital, improved availability of credit and intertemporal insurance when the firm experiences adverse business conditions (Berlin and Mester, 1999). The bank, however, will only provide insurance if it knows that it will benefit from the continuation of the relationship (Freixas and Rochet, 2008). Moreover, it could be efficient for a bank to provide intertemporal smoothing because the bank’s loan portfolio is better diversified than the firm’s and the risk could be better managed by the bank than the firm (Freixas and Rochet, 2008). I believe that intertemporal insurance also plays an important part in shaping the relationship between leverage and investment in the sample of Vietnamese firms because the period of data collection is from 2009 to 2011, which is immediately after the start of the financial crisis and I believe that it is very likely that state-owned banks supplied additional credit to SMEs hit by adverse business conditions in this period because of their political agenda.
8 Another cost of relationship lending is the soft-budget constraint problem. Boot (2000) describe it as “the potential lack of toughness on the bank’s part in enforcing credit contracts”. This arises when the bank cannot deny extending credit to the firm when it experiences difficulties. The bank might do this in the hope to recover previously provided loan and the problem is that firms that realize this may have the wrong incentives to start with. (Boot, 2000). I argue that the soft-budget constraint problem is also relevant for SMEs in a predominantly state-owned bank lending environment such as Vietnam and can influence the leverage-investment relation for this type of firms. State-owned banks lack market discipline and therefore may not engage in aggressive collection procedure because of its political agenda to subsidize firms in certain industries or regions (Berger and Udell, 2006). Because of this previous loans may not matter for continuing to provide funds to these firms which means the standard theory of the leverage-investment link is distorted.
Based on the arguments about the bright and the dark side of relationship lending including the increased availability of credit, the decrease in asymmetric information about the firm’s future investment prospects, the intertemporal insurance during adverse business conditions, the hold-up situation and its implication for financing projects at early stages and the soft-budget constraint problem and its implication for subsequent funding, I argue that the standard leverage-investment theory does not apply to SMEs. So I formulate my first hypothesis:
H1: There is a positive relation between leverage and investment. The leverage-investment relation for high and low-growth firms
9 overinvestment problem because it “pre-commits firms to pay cash as interest and principal and such commitments can prevent the misuse of free cash flow” (Aivazian et al., 2005).
A research by McConnell and Servaes (1995) on listed U.S. companies confirms the overinvestment theory. In another study on Canadian firms, Aivazian et al. (2005) also find evidence that investment is negatively correlated with leverage for firms with low growth opportunities. Firth et al. (2008) examine the relation between leverage and investment among China’s listed firms and confirm the existence of a debt overhang problem even in a bank lending environment where all banks are state-owned. However their finding that the relation is weaker in firms with low growth opportunities and poor operating performance than in firms with high growth opportunities and good operating performance contradicts the existing research. Firth et el. (2008) also find that the negative relation between leverage and investment is weaker in firms with higher level of state shareholding than in firms with lower level of state shareholding. This finding also is opposite to the finding from existing research in developed countries. Firth et el. (2008) argue that the political obligation of state-owned banks to support firms in certain industries and regions as well as their more lenient lending policies restricts them from playing a monitoring and disciplining role in limiting the overinvestment bias of low growth and poorly performing firms. Both of the authors’ findings are consistent with the hypothesis that the state-owned banks in China impose fewer restrictions on the capital spending of low growth and poorly performing listed firms and also for firms with high state ownership (Firth et al., 2008).
In line with the arguments of Firth et al. (2008) that state-owned banks impose fewer restraints on the capital expenditure of low growth firms because of their political agenda to subsidize targeted borrowers and their lack of market discipline to strictly enforce the credit contracts, I argue that the leverage-investment link of low and high growth firms in Vietnam is affected differently than the standard theory suggests. Also it is worth noting that most borrowing in Vietnam comes from state-owned banks (Duxton Asset Management, 2015). Based also on the fact that SMEs are very different from public firms and the implication this may have for the leverage-investment link discussed in the previous section, I formulate my second hypothesis:
H2: The positive relation between leverage and investment is stronger in firms with low growth opportunities than in firms with high growth opportunities.
State-owned banks and Vietnam’s banking sector
state-10 owned institutions generally distribute government support and often are instructed to supply additional credit to SMEs in specific industries, sectors and regions. Moreover, the authors argue that much of this funding may not be directed to creditworthy SMEs because the lending agenda of the government does not always require financing of positive NPV projects or repayment of loans at competitive market rates. Additionally, funds may be dedicated for the political or personal objectives of government officials rather than economic ones (Shleifer & Vishny, 1994; Sapienza, 2004; Dinc, 2005). State-owned banks may not engage in aggressive collection procedures because of their agenda to subsidize targeted firms or because of the lack of market discipline (Berger and Udell, 2006). It is clear that the nature of the bank ownership in Vietnam has implications on the relation between leverage and investment as it affects firm’s cash flows and project evaluation.
Over the past 25 years, the Vietnamese government has initiated many banking reforms to improve the efficiency and competitiveness of the banking system especially through the privatization of state-owned banks (Duxton Asset Management, 2015). In May 2006, the Vietnamese government had announced plans to partially privatize the banks and reduce government ownership to 50% by 2010. However, only two of the SOCBs (VCB & CTG) have successfully sold more than 20% of its shares to private investors thus far. (Duxton Asset Management, 2015). The state still remains the controlling stakeholder, holding at least 65% ownership in them (Duxton Asset Management, 2015). SOCBs have a huge role in Vietnam’s economy (Duxton Asset Management, 2015). There are currently 5 state-owned commercial banks (SOCBs) which had initially been established to fulfill a specialized lending function. Their traditional customer base has been state-owned enterprises (SOEs) but they are increasingly diversifying their customer base to include non-SOEs (Duxton Asset Management, 2015). “State-owned companies use up to 50 percent of state investment, tie up to 60 percent of bank lending and account for more than half of the nation’s bad debt, according to Deputy Finance Minister Truong Chi Trung.” (Bloomberg news, 2013).
11 Figure 1 Credit market share of Vietnamese banks, 2007 - 2013
State-owned banks' policy lending not only leads to high non-performing loan ratios but also distorts firms' investment decisions. (Firth et al., 2008) Through the leverage-investment link, it is argued that the policy lending of state-owned banks would ultimately affect capital allocation efficiency, as well as the economic growth of the economy (Dobson and Kashyap, 2006). Given the large state ownership of Vietnamese banks, the political agenda of state-owned banks to supply additional credit to SMEs in specific industries, sectors and regions and their lack of market discipline to aggressively enforce the credit contracts, I formulate my third hypothesis:
H3: The positive relation between leverage and investment is stronger in firms with state-owned bank borrowing than in firms with non-state state-owned bank borrowing.
3. Data and Methodology
12 Following the methodology of Firth et el. (2008), I first apply the following investment equation on the sample of firms with bank borrowing to explore the impact of debt on firm investment. For consistency, I follow prior studies by adding control variables to the regression to proxy for financial constraints: growth, sales and firm size. Beck et al. (2006) present evidence that financially-constraint firms are the one that are also growth-constraint and have low sales. The authors also find that firm size is an important determinant of the firm’s financing obstacles with smaller firms reporting significantly higher obstacles than larger firms (Beck et al., 2006).
𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡𝑖,𝑡 = 𝛽0+ 𝛽1𝐿𝑜𝑎𝑛𝑠𝑖,𝑡−1+ 𝛽2𝐺𝑟𝑜𝑤𝑡ℎ 𝑖,𝑡+ 𝛽3𝑆𝑎𝑙𝑒𝑠𝑖,𝑡−1+ 𝛽4𝑆𝑖𝑧𝑒𝑖.𝑡+ 𝜀𝑖,𝑡 (I) Where 𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡𝑖,𝑡 is the ratio of gross investment to fixed assets;
𝐿𝑜𝑎𝑛𝑠𝑖,𝑡−1 is the lagged one time period ratio of bank loans to fixed assets;
𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡 is the average sales growth rate between t and t-1; it is a proxy for the firm’s
future growth opportunities;
𝑆𝑎𝑙𝑒𝑠𝑖,𝑡−1 is the lagged one time period ratio of sales to fixed assets; 𝑆𝑖𝑧𝑒𝑖.𝑡−1 𝑖s measured by the natural logarithm of lagged fixed assets; 𝜀𝑖,𝑡 refers to the error term
Most previous studies use panel pooled regressions or firm fixed effects to estimate the investment regression. However, the data I use are cross-sectional and only includes variables for the year 2008 and 2009. Therefore, I will use the OLS estimator1 to explore the impact of leverage on investment.
Furthermore, to test for differences in the roles of leverage in high-growth versus low-growth firms, regression II based on the methodology of Aivazian et al. (2005) and Firth et al. (2008) will be applied. The interaction term accounts for the different impact of debt on investment for firms with low and high-growth prospects. Once again I add control variables to the regression to proxy for financial constraints: growth, sales and firm size.
𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡𝑖,𝑡 = 𝛽0+ 𝛽1𝐿𝑜𝑎𝑛𝑠𝑖,𝑡−1+ 𝛽2𝐺𝑟𝑜𝑤𝑡ℎ 𝑖,𝑡+ 𝛽3𝐷𝐺𝑟𝑜𝑤𝑡ℎ 𝑖,𝑡∗ 𝐿𝑜𝑎𝑛𝑠𝑖,𝑡−1+
𝛽4𝑆𝑎𝑙𝑒𝑠𝑖,𝑡−1+ 𝛽5𝑆𝑖𝑧𝑒𝑖.𝑡+ 𝜀𝑖,𝑡 (II)
where 𝐷𝐺𝑟𝑜𝑤𝑡ℎ 𝑖,𝑡 is a dummy variable that equals 1 if 𝐺𝑟𝑜𝑤𝑡ℎ 𝑖,𝑡 is lower than the mean
value in the sample in that year and 0 otherwise. I expect the coefficient on the interaction term to
1
13 be negative if the relation between debt and investment is stronger for high-growth than for low-growth firms.
I will subsequently apply the following regression specification in order to distinguish between the monitoring effects of state-owned versus non-state owned banks.
𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡𝑖,𝑡 = 𝛽0+ 𝛽1𝐿𝑜𝑎𝑛𝑠𝑖,𝑡−1+ 𝛽2𝐺𝑟𝑜𝑤𝑡ℎ 𝑖,𝑡+ 𝛽3𝐷𝑁𝑎𝑡𝑢𝑟𝑒𝑖,𝑡∗ 𝐿𝑜𝑎𝑛𝑠𝑖,𝑡−1+ 𝛽4𝑆𝑎𝑙𝑒𝑠𝑖,𝑡−1+ 𝛽5𝑆𝑖𝑧𝑒𝑖.𝑡+ 𝜀𝑖,𝑡 (III)
where 𝐷𝑁𝑎𝑡𝑢𝑟𝑒𝑖,𝑡 is a dummy variable that equals 1 if the borrowing source is a state-owned
bank and 0 if it is a non-state owned bank. I expect the coefficient on the interaction term to be positive if the relation between debt and investment is stronger for firms with state-owned bank borrowing compared to firms with non-state bank borrowing.
4. Summary Statistics, Results, and Robustness Checks
Table 2 provides summary statistics of the regression variables for the firms with and without bank loans in 2008. Firms with missing variables in the denominator of the regression equation such as fixed assets in 2009 and sales in 2008 are excluded. The sample of firms without bank loans includes 651 observations out of 1,043 firms that report to have no bank loans in 2008. The sample of firms with bank borrowing includes 161 observations, 93 with state-owned bank borrowing and 68 with non-state owned bank borrowing. The data are winsorized at 5% at each tail to deal with outliers.
Comparing the two samples, it can be seen that the median of the ratio of gross investment to fixed assets for firms with no bank loans is 14 times lower which can be explained by the lack of bank financing for this sample of firms. The median of average sales growth rate between 2009 and 2008 (1.17), however, is close to the one observed for firms with bank loans (1.19). Additionally, the median of lagged sales to fixed assets of firms with no bank loans is considerably lower – 2.66 - compared to 4.40 for firms with bank loans which might explain the inability of those firms to access external financing although their sales growth rate offers some hope to reverse this trend. Finally, the median of natural logarithm of lagged fixed assets for firms with no bank loans is 7.60 compared to 8.60 for firms with bank loans in 2008 which confirms the argument of Menkhoff et al. (2006) that it is harder for SMEs with lower collateralizable assets to obtain external financing.
14 public companies (0.06 in Canada, Aivazian et al. (2005), 0.148 in China, Firth et al. (2008), and 0.0385 in a more recent study in China, Jiang and Zeng, 2014) even though these figures are based on the ratio of net investment by total assets which leads to lower overall values. Moreover, the median of the ratio of lagged bank loans to fixed assets is 1.20 which is much higher compared with the existing studies on public companies (0.50 in Canada, Aivazian et al. (2005), 0.211 in China, Firth et al. (2008), and 0.2375 in a more recent study in China, Jiang and Zeng, 2014). There figures, however, are based on the ratio of total leverage to total assets. The high figure of lagged bank loans to fixed assets for SMEs in Vietnam - suggests that there is a significant reliance on bank financing by these firms.
Table 2 Summary statistics for the firms with and without bank borrowing
INVESTMENT LOANS GROWTH SALES SIZE
Firms with no bank loans
Observations 651 - 651 651 651 Mean 0.77 - 1.27 6.90 7.48 Median 0.06 - 1.17 2.66 7.60 Maximum 6.14 - 2.26 36.67 12.34 Minimum 0.00 - 0.80 0.27 4.25 Std. Dev. 1.54 - 0.35 9.70 2.20
Firms with bank loans
Observations 161 161 161 161 161 Mean 4.3 5.20 1.27 15.50 9.02 Median 0.88 1.20 1.19 4.40 8.60 Maximum 29.8 37.50 2.08 74.95 13.73 Minimum 0.01 0.05 0.90 0.53 5.30 Std. Dev. 7.73 9.64 0.30 21.72 2.38
15 Table 3 t-Test for the mean differences of the sample of firms with and without bank borrowing
INVESTMENT LOANS GROWTH SALES SIZE
t Stat 5,88 - -0,10 4,90 5,69 P(T<=t) one-tail 0,00 - 0,46 0,00 0,00 t Critical one-tail 1,65 - 1,65 1,65 1,65 P(T<=t) two-tail 0,00 - 0,92 0,00 0,00 t Critical two-tail 1,97 - 1,97 1,97 1,97
Table 4 presents the correlation matrix of all the variables used in the regression model. To check for multicollinearity, I check for high correlation between the independent variables. I observe high and statistically significant correlation (0.54) between loans and sales. But because the model coefficients described further have an appropriate sign and are of a plausible magnitude, I can ignore this problem.
Additionally, it can be seen from the table that the variables loans and investment are positively and significantly correlated which provides preliminary evidence that confirms my first hypothesis about the positive leverage-investment relation for SMEs. Moreover, sales is positively and significantly related to investment which is consistent with the prior studies while size is negatively and significantly related to investment which is not consistent with the prior studies. Table 4 Correlation matrix
INVESTMENT LOANS GROWTH SALES SIZE
INVESTMENT 1
LOANS 0.52*** 1
GROWTH -0.02 -0.10 1
SALES 0.40*** 0.54*** -0.09 1
SIZE -0.17** -0.13 -0.14 -0.16** 1
***, **, * correspond to p-values of 1%, 5%, and 10%, respectively.
16 method works by using higher order terms of the fitted values (e.g. 𝑦𝑡2, 𝑦𝑡3, ... ) in an auxiliary
regression of the form (Brooks, 2008):
𝑦𝑡 = 𝛽1+ 𝛽2𝑥2𝑡+ 𝛽3𝑥3𝑡+ ⋯ + 𝛾1𝑦𝑡2+ 𝛾
2𝑦𝑡3 + 𝛾𝑝𝑦𝑡𝑝+ 𝜀𝑡 (IV)
The functional form of the model is correct if 𝛾1= 0, 𝛾2= 0, 𝛾𝑝 = 0. This can be tested using
an F-test. If the null hypothesis that all the coefficients are equal to 0 is rejected, then the functional form is wrong. The result of the test for one fitted term is shown in table 5. Based on the value of the F-statistic (2.25) and the high p-value (0.135), I cannot reject the null hypothesis. Therefore, I can infer that there is no apparent non-linearity in the regression equation and I can conclude that the linear model for the investment-leverage relation is appropriate.
Table 5 Ramsey Reset Test Ramsey RESET Test
Specification: INVESTMENT C LOANS GROWTH SALES SIZE Omitted Variables: Squares of fitted values
Value df Probability
F-statistics 2.252816 (1, 155) 0.1354
Next I test whether the errors have a constant variance using the White’s test (table 6). To test if the var(𝜀𝑡) = 𝜎2, I estimate the model and obtain the residuals, 𝜀𝑡. Then I run an auxiliary
regression of the form (Brooks, 2008):
𝜀𝑡 = 𝛼1+ 𝛼2𝑥2𝑡+ 𝛼3𝑥3𝑡+ ⋯ + 𝛼4𝑥2𝑡2 + 𝛼5𝑥3𝑡2 + 𝛼6𝑥2𝑡𝑥3𝑡+ ⋯ + 𝜗𝑡 (V)
Where 𝜗𝑡 is a normally distributed disturbance term independent of 𝜀𝑡. After obtaining the
𝑅2 from the auxiliary regression and multiplying it by the number of observations, I obtain 𝑇𝑅2~ 𝝌2.
I test the joint null hypothesis that 𝛼2 = 0, 𝛼3= 0, 𝛼3 = 0, etc. Based on the value of the chi-square
17 Table 6 White test for heteroscedasticity
Heteroscedasticity Test: White
F-statistics 5.0187 Prob F. (14,146) 0.0000 Obs*R-Squared 52.3076 Prob. Chi-Square(14) 0.0000 Scaled explained SS 146.8631 Prob. Chi-square(14) 0.0000
Table 7 OLS with White heteroscedasticity-consistent standard errors Dependent Variable: INVESTMENT
Method: Least Squares Included Observations: 161
White heteroscedasticity-consistent standard errors & covariance Variable Coefficient Std. Error t-Statistics Prob.
C 7.0704 3.1860 2.2192 0.0279 LOANS 0.3472 0.1117 3.1061 0.0023 GROWTH -1.7870 1.9094 -0.9358 0.3508 SALES 0.0535 0.0395 1.3519 0.1783 SIZE -0.3395 0.1868 -1.8174 0.0711 R-squared 0.3052
18 Table 8 Jacque-Bera test of normality
After having performed various tests to ensure that the OLS estimator is the best linear unbiased estimator (BLUE), I can infer that debt is positively and significantly related to investment at the 1% level in the sample of Vietnamese SMEs. I therefore confirm my first hypothesis that there is a positive relation between leverage and investment for this type of firms. The result is opposite to what the existing studies on public companies by Aivazian et al., 2005; Firth et al., 2008; Jiang and Zeng, 2014 find. However, given the different nature of the studied firms and the different financial environment that these firms operate in, the results can be explained by their significant dependence on bank financing. Moreover, the improved availability of credit and intertemporal insurance when the firm experiences adverse business conditions, all advantages of relationship lending, can provide another explanation for the positive relationship between leverage and investment in the sample of Vietnamese firms. Additionally, banks might be willing to finance more risky ventures in the early project stages because they can ask for a higher repayment at later stages (Petersen and Rajan, 1995). This can be another possible explanation for the positive relation. Finally, the soft-budget constraint or the decrease in “toughness” on the bank’s part in enforcing the lending contract provides another explanation for the positive leverage-investment relation.
19 between leverage and investment is stronger in firms with low growth opportunities than in firms with high growth opportunities. The result is consistent with the argument that state-owned banks impose fewer restraints on the capital expenditure of low growth firms in their goal to subsidize targeted borrowers and because of the lack of market discipline to strictly enforce the credit contracts.
Table 9 OLS with White heteroscedasticity-consistent standard errors for low and high-growth firms Dependent Variable: INVESTMENT
Method: Least Squares Included Observations: 161
White heteroscedasticity-consistent standard errors & covariance
Variable Coefficient Std. Error t-Statistics Prob.
C 1.9171 2.6574 0.7214 0.4717 LOANS 0.1916 0.1058 1.8098 0.0723 GROWTH 1.5993 1.5073 1.0610 0.2903 DGROWTH*LOANS 0.5037 0.1306 3.8561 0.0002 SALES 0.0470 0.0347 1.3558 0.1771 SIZE -0.2731 0.1913 -1.4276 0.1554 R-squared 0.4004
Furthermore, to test for differences in the monitoring of state-owned versus non-state owned banks, I add a dummy variable 𝐷𝑁𝑎𝑡𝑢𝑟𝑒𝑖,𝑡 to the regression equation. As can be seen in table
20 Table 10 OLS with White heteroscedasticity-consistent standard errors for firms with state-owned and non-state owned bank borrowing
Dependent Variable: INVESTMENT Method: Least Squares
Included Observations: 161
White heteroscedasticity-consistent standard errors & covariance
Variable Coefficient Std. Error t-Statistics Prob.
C 4.6163 1.7408 2.6517 0.0088 LOANS 0.4721 0.1374 3.4343 0.0008 GROWTH -0.0155 0.0785 -0.1974 0.8438 NATURE*LOANS -0.2804 0.1669 -1.6800 0.0950 SALES 0.0657 0.0381 1.7233 0.0868 SIZE -0.3257 0.1738 -1.8736 0.0629 R-squared 0.3400
To test the robustness of my results, I will first restrict the sample to exploiting, producing or processing firms. These firms have more tangible assets compared to service or trading firms and therefore might be less dependent on relationship lending for raising bank capital. The results presented in tables 10, 11 and 12 are similar to the previous ones. None of the coefficients I am interested in changes sign. So I confirm the previous results that loans are positively related to investment, that the positive relation between bank loans and investment is significantly stronger for low-growth than for high-growth firms, and that the positive relation between debt and investment is stronger for non-state owned banks than for state-owned banks, this time being confident at the 5% level.
Table 11 OLS with White heteroscedasticity-consistent standard errors for firms from exploiting, producing or processing industries
Dependent Variable: INVESTMENT Method: Least Squares
Included Observations: 99
White heteroscedasticity-consistent standard errors & covariance
21 C 9.1189 4.2486 2.1463 0.0344 LOANS 0.1553 0.0748 2.0763 0.0406 GROWTH -1.6000 2.2462 -0.7123 0.4780 SALES 0.0799 0.0324 2.4639 0.0156 SIZE -0.5540 0.2826 -1.9599 0.0530 R-squared 0.2700
Table 12 OLS with White heteroscedasticity-consistent standard errors for low and high-growth firms from exploiting, producing or processing industries
Dependent Variable: INVESTMENT Method: Least Squares
Included Observations: 99
White heteroscedasticity-consistent standard errors & covariance
Variable Coefficient Std. Error t-Statistics Prob.
C 2.9660 3.8583 0.7687 0.4440 LOANS 0.1712 0.0720 1.0988 0.0812 GROWTH 2.5599 2.0921 1.2235 0.2242 DGROWTH*LOANS 0.5768 0.1034 5.5785 0.0001 SALES 0.0679 0.0283 2.3979 0.0185 SIZE -0.4961 0.2462 -2.0149 0.0468 R-squared 0.4531
Table 13 OLS with White heteroscedasticity-consistent standard errors for firms with state-owned and non-state owned bank borrowing from exploiting, producing or processing industries
Dependent Variable: INVESTMENT Method: Least Squares
Included Observations: 99
White heteroscedasticity-consistent standard errors & covariance
22 C 8.2200 4.2026 1.9559 0.0535 LOANS 0.2482 0.0866 2.8659 0.0051 GROWTH -1.2002 2.2183 -0.5410 0.5898 NATURE*LOANS -0.2179 0.1071 -2.0345 0.0447 SALES 0.1003 0.0334 2.9987 0.0035 SIZE -0.5360 0.2782 -1.9267 0.0571 R-squared 0.3011
I further restrict the sample to firms in the service and trading industries to check the robustness of my previous results. Firms in these industries have less tangible assets compared to exploiting, producing or processing firms and therefore should be more dependent on relationship lending for raising bank capital. The results are presented in tables 13, 14 and 15. Only the coefficient for the nature of bank borrowing changes sign, all other coefficients remain the same. So I confirm the previous results that loans are positively related to investment however this result is significant at the 10% level. I also confirm the finding that the positive relation between bank loans and investment is stronger for low-growth than for high-growth firms, also at the 10% confidence level. However, this time the positive relation between debt and investment is stronger for state-owned banks compared to non-state owned banks which is opposite to my previous results. So when service and trading industries are concerned, state-owned banks supply additional credit to these industries. Table 14 OLS with White heteroscedasticity-consistent standard errors for firms from service and trading industries
Dependent Variable: INVESTMENT Method: Least Squares
Included Observations: 62
White heteroscedasticity-consistent standard errors & covariance
Variable Coefficient Std. Error t-Statistics Prob.
C 4.9552 3.2990 1.5020 0.1386
LOANS 0.2653 0.1623 1.6339 0.1078
GROWTH 1.5003 1.6903 0.8875 0.3785
SALES -0.0277 0.0266 -1.0413 0.3021
23
R-squared 0.2097
Table 15 OLS with White heteroscedasticity-consistent standard errors for low and high-growth firms from service and trading industries
Dependent Variable: INVESTMENT Method: Least Squares
Included Observations: 62
White heteroscedasticity-consistent standard errors & covariance
Variable Coefficient Std. Error t-Statistics Prob.
C 2.3998 3.5043 0.6848 0.4963 LOANS 0.2122 0.1073 1.9771 0.0530 GROWTH 2.6878 1.6664 1.6129 0.1124 DGROWTH*LOANS 0.4240 0.2427 1.7467 0.0862 SALES -0.0302 0.0252 -1.1993 0.2355 SIZE -0.4673 0.2689 -1.7377 0.0878 R-squared 0.2506
Table 16 OLS with White heteroscedasticity-consistent standard errors for firms with state-owned and non-state owned bank borrowing from service and trading industries
Dependent Variable: INVESTMENT Method: Least Squares
Included Observations: 62
White heteroscedasticity-consistent standard errors & covariance
Variable Coefficient Std. Error t-Statistics Prob.
C 4.5701 3.1365 1.4570 0.1507
LOANS 0.2229 0.1029 2.1651 0.0347
GROWTH 1.4726 1.4964 0.9840 0.3293
24
SALES -0.0339 0.0249 -1.3599 0.1793
SIZE -0.5275 0.2602 -2.0271 0.0474
R-squared 0.2751
5. Conclusion
My results offer three main findings. First, there is a positive relation between leverage and investment in the sample of SMEs. I argue that the relation between leverage and investment is different for SMEs than the one documented for public companies in the developed countries of US and Canada and the developing country of China and my results support this hypothesis. As a reasoning, I point to the fact that the SMEs in Vietnam operate in an extremely uncertain environment and because SMEs are so informationally opaque, they cannot send credible signals to banks and other creditors. This is the reason that they have difficulty in raising external capital and most often borrow from banks with whom they establish long-term relationship. The positive relation between bank loans and investment could be explained by various factors such as the improved availability of credit, the decrease in asymmetric information and intertemporal insurance, the bright side, but also the dark side of relationship lending which often creates a hold-up situation for the firm but makes possible the financing of projects at early stages, as well as the soft-budget constraint that prevents the bank to ‘toughly’ enforce the lending contract and leads to increased funding.
25 My results suggest that relationship lending is effective. However, banks should strive to diminish the firms’ incentives for moral hazard and try to incentivize them by adjusting the lending terms throughout the relationship. Additionally, banks should strive to monitor borrowers to prevent the inefficient allocation of capital because this can have an adverse effect on the competitiveness and the growth of the economy. I recommend that the Vietnam’s government should continue its reform in the Vietnam’s banking sector in order to improve its efficiency and competitiveness. This should greatly benefit the business environment in the country, and assist in the development process. According to Beck and Kunt (2006) small firms benefit greater than large firms from improved financial environment so more competitive and efficient banking system should become a goal of the Vietnamese government since SMEs can contribute enormously to production and growth.
26 Appendix A
OLS is a good estimator provided that the so-called Gauss-Markov assumptions (table 1.A) are satisfied (Brooks, 2008). If these assumptions hold, then the estimators 𝛽0, 𝛽1, 𝛽2, etc.
determined by the regression equation (I) will have some desirable properties which are known as the Best Linear Unbiased Estimators (BLUE) (Brooks, 2008). ‘Best’ means that the OLS estimator has minimum variance of all linear unbiased estimators (Brooks, 2008). ‘Linear’ means that 𝛽0 and 𝛽1 are
linear estimators (Brooks, 2008). ‘Unbiased’ means that on average, the actual values of 𝛽0 and
𝛽1 will be equal to their true values (Brooks, 2008). ‘Estimator’ means that 𝛽0 and 𝛽1 are estimators
of the true values of 𝛽0 and 𝛽1 for the population (Brooks, 2008). If even one of the assumptions is
not satisfied, then the OLS estimator is no longer BLUE.
Table 1.A Assumptions concerning disturbance terms and their interpretation Technical notation Interpretation
(1) E(ut) = 0 The errors have zero mean
(2) var(ut) = σ2 The errors have constant and finite variance.
(3) cov(ut, uj) = 0 The errors are uncorrelated between observations.
(4) cov(ut, xt) = 0 The errors and independent variables are uncorrelated.
(5) 𝑢𝑡 ~ 𝑁(0, 𝜎2) The disturbances are normally distributed
As can be seen from table 1.A, all assumptions involve the error term so all diagnostic tests are performed on the residuals. The first assumption is never violated if there is a constant in the model (Brooks, 2008). The second assumption assumes that the errors are homoscedastic which can be checked by using White’s (1980) general test for heteroscedasticity. If the assumption of homoscedasticity is violated, one way to deal with it is by using White’s heteroscedasticity-consistent standard errors (Brooks, 2008). The effect of using this correction is that the standard errors for the slope coefficients are increased relative to the usual OLS standard errors. This will make hypothesis testing more ‘conservative’, so that more evidence will be required to be able to reject the null hypothesis (Brooks, 2008). The third assumption of no autocorrelation of the residuals is less likely to occur when dealing with cross-sectional data so I will skip it.
27 Furthermore, the fifth assumption states that the disturbance terms are normally distributed. Normality in the residuals is required to test the hypotheses stated earlier. I use the Jacque-Berra (JB) test to check if the coefficient of skewness and kurtosis are jointly zero. The JB test statistics is zero in the case of normal distribution. However, even if the test statistics significantly deviates from this value, the normality assumption can be ignored if the sample size is sufficiently large (Brooks, 2008). Moreover, an implicit assumption of the OLS estimator is that the appropriate ‘functional’ form is linear (Brooks, 2008). This means the relationship between x and y can be represented by a straight line. I will use Ramsey’s (1969) RESET test to determine if the model is linear.
28 References
Ahn, S., Denis, D., Denis, D., 2006. Leverage and investment in diversified firms. Journal of Financial Economics 79, 317-337.
Aivazian, V., Ge, Y. Qiu, J., 2005. The impact of leverage on firm investment: Canadian Evidence. Journal of Corporate Finance 11, 277-291.
Ang, J., Cole, R., Lin, J., 2000. Agency costs and ownership structure. The Journal of Finance 55. 81-106
Beck, T., Demirguc-Kunt, A., Laeven, L., Maksimovic, V., 2006. The determinants of financing obstacles. Journal of International Money and Finance 25, 932–952.
Beck, T., Demirguc-Kunt, A., 2006. The determinants of financing obstacles. Journal of International Money and Finance 25, 932–952.
Berger, A., Udell, F., 1995. Small and medium-size enterprises: Access to finance as a growth constraint, Journal of Banking and Finance 30, 2931-2943.
Berger, A., Udell, G., 2006. A more complete conceptual framework for SME finance. Journal of Banking and Finance 30, 2945-2966.
Berlin, M., Mester, L., 1992. Debt Covenants and Renegotiation. Journal of Financial Intermediation 2, 95-133.
Bloomberg News. 2013. Vietnam Gets Tough on State Firms in Economic Growth Push. Available at
http://www.bloomberg.com/news/articles/2013-07-24/vietnam-gets-tough-on-state-firms-in-economic-growth-push, last accessed on 17/12/2015.
Brooks, C., 2008. Introductory Econometrics for Finance. Cambridge University Press, Cambridge. Boot, A., 2008. Relationship banking: What do we know? Journal of Financial Intermediation 9, 7-25. Cole, R., 1998. The importance of relationship to the availability of credit. Journal of Banking and Finance 22, 959 - 977
Cull, R., Xu, L.C., 2003. Who gets credits? the behavior of bureaucrats and state banks in allocating credit to Chinese state-owned enterprises. Journal of Development Economics 71, 533–559.
29 Dinc, I.S., 2005. Politicians and banks: political influences on government-owned banks in emerging markets. Journal of Financial Economics 77, 453–479.
Dobson, W., Kashyap, A.K., 2006. The contradiction in China's gradualist banking reforms. Brookings Papers on Economic Activity 2, 103–162.
Duxton Asset Management. (2015). Vietnam Banking Industry Report. Available at
http://www.duxtonam.com/vietnam-banking-industry-report-23-jan-2015/, last accessed on 17/12/2015.
Fama, E., 1985. What is different about banks? Journal of Monetary Economics 15, 29-39.
Firth, M., Lin, C., Wong, S., 2008. Leverage and investment under a state-owned bank lending environment: Evidence from China. Journal of Corporate Finance 14, 642-653.
Freixas, X., Rochet, J., 2008. Microeconomics of Banking. MIT Press, London.
Grinblatt, M., Titman, S., 2008. Financial markets and corporate strategy. Mc-Graw-Hill. London Gaver, J., & Gaver, K. (1993). Additional evidence on the association between the investment opportunity set and corporate financing, dividend, and compensation policies. Journal of Accounting and Economics, 16, 125–160.
Gertler, M., Gilchrist, S., 1994. Monetary policy, business cycles, and the behavior of small manufacturing firms. Quarterly Journal of Economics 109, 309-340.
Goyal, V., Lehn, K., & Racic, S. (2002). Growth opportunities and corporate debt policy: The case of the U.S. defense industry. Journal of Financial Economics, 64, 35–59.
Jensen, M.C., 1986. Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review 76, 323–329.
Jensen, M.C., Meckling, W.H., 1976. Theory of the firm: managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3, 305–360.
Jiang, W., Zeng, Y., 2014. State ownership, bank loans, and corporate investment. International Review of Economics and Finance 32, 92-116.
30 Lang, L., Ofek, E., Stulz, R., 1996. Leverage, investment, and firm growth. Journal of Financial Economics 40, 3-29.
McConnell, J.J., Servaes, H., 1995. Equity ownership and the two faces of debt. Journal of Financial Economics 39, 131–157.
Menkhoff, L., Neuberger, D., Suwanaporn, C., 2006. Collateral-based lending in emerging markets: evidence from Thailand. Journal of Banking and Finance 30, 1–21.
Modigliani, F., Miller, M.H., 1958. The cost of capital, corporation finance, and the theory of investment. American Economic Review 53, 433–443.
Myers, S., 1977. Determinants of corporate borrowing. Journal of Financial Economics 5, 147–175. Norden, L., 2015. The role of banks in SME finance. Journal of Economic Literature. Available at http://repub.eur.nl/pub/77636, last accessed on 17/12/2015.
Petersen, M., Rajan, R., 2002. Does distance still matter? The information revolution in small business lending. Journal of Finance 57, 2533-2570.
Pollard, J. S. (2003). Small firm finance and economic geography. Journal of Economic Geography, 3(4), 429-452.
Rajan, R., 1992. Insiders and outsiders: the choice between informed and arm’s-length debt. Journal of Finance 47, 1367-1406.
Sapienza, P., 2004. The effects of government ownership on bank lending. Journal of Financial Economics 72, 357–384.
Scholtens, B., 1999. Analytical issues in external financing alternatives for SBEs. Small Business Economics 12, 137-148.
Seward, J., 1990, Corporate financial policy and the theory of financial intermediation. Journal of Finance 45, 351-377.
Sharpe, S., 1990. Asymmetric information, bank lending and implicit contracts: a stylized model of customer relationships. Journal of Finance 45, 1069-1087.
31 Stulz, R.M., 1990. Managerial discretion and optimal financing policies. Journal of Financial Economics 26, 3–27.
Vos, Ed., Yeh, A., Carter, S., Tagg, S., 2007. The happy story of small business financing. Journal of Banking & Finance 31, 2648 - 2672
