4.1 Regression results
Table
3 reports the estimated results using a bivariate probit model based on the joint distribution of two variables (
debt and
equity). This method allows us to theoretically address the simultaneity of debt and equity financing decisions. In all specifications, the correlation (rho) between the error terms (
\({\varepsilon }_{1it}\) and
\({\varepsilon }_{2it})\) of the underlying stochastic utility function associated with debt and equity finance is significant at 1%, indicating that a bivariate probit model is more appropriate than two univariate probit models. We report the marginal effects of intangibles and interactive variables (Intangibles × Age and Intangible × Size) in both debt and equity equations in Table
4. Marginal effects of intangibles were calculated following the bivariate probit estimations (1) and (2) reported in Table
3. The average marginal effect of each variable was predicted on the joint probability distribution of
\({z}_{1it} and {z}_{2it}\) based on the approach of Greene (
2003). For interaction terms, we predict the full interaction effects by computing the cross derivative of the expected value of the dependent variables as suggested by Norton et al. (
2004). According to Ai and Norton (
2003) and Norton et al. (
2004), the magnitude of the interaction effect in nonlinear models does not equal the marginal effect of the interaction term; therefore, the test for the statistical significance of the interaction effect must be based on the estimated cross-partial derivative, not on the coefficient of the interaction term.
Table 3
Intangible assets and firms’ access to debt and equity finance (coefficients)
VARIABLES | Debt | Equity | Debt | Equity | Debt | Equity |
Intangibles | 0.0253*** | 0.0077*** | 0.0538*** | 0.0429*** | 0.0290*** | 0.0133*** |
| (0.0003) | (0.0002) | (0.0041) | (0.0027) | (0.0004) | (0.0004) |
Tangibles | 0.0383*** | 0.0149*** | 0.0380*** | 0.0129** | 0.0391*** | 0.0180*** |
| (0.0037) | (0.0055) | (0.0037) | (0.0054) | (0.0037) | (0.0055) |
Intangibles × Size | | | − 0.0013*** | − 0.0016*** | | |
| | | (0.0002) | (0.0001) | | |
Intangibles × Age | | | | | − 0.0008*** | − 0.0011*** |
| | | | | (0.0001) | (0.0001) |
Size | 0.3872*** | − 0.0699*** | 0.4052*** | − 0.0517*** | 0.3890*** | − 0.0668*** |
| (0.0020) | (0.0016) | (0.0033) | (0.0021) | (0.0020) | (0.0016) |
Age | 0.0088*** | 0.0425*** | 0.0093*** | 0.0434*** | 0.0200*** | 0.0565*** |
| (0.0007) | (0.0006) | (0.0007) | (0.0006) | (0.0012) | (0.0009) |
Cash | − 0.0923*** | 0.0637*** | − 0.0920*** | 0.0631*** | − 0.0916*** | 0.0645*** |
| (0.0015) | (0.0013) | (0.0015) | (0.0013) | (0.0015) | (0.0013) |
Growth | 0.0182*** | − 0.3300*** | 0.0185*** | − 0.3297*** | 0.0183*** | − 0.3301*** |
| (0.0046) | (0.0037) | (0.0046) | (0.0037) | (0.0046) | (0.0037) |
Z-Score | − 0.0048* | 0.0024 | − 0.0048* | 0.0025 | − 0.0047* | 0.0026* |
| (0.0025) | (0.0015) | (0.0025) | (0.0015) | (0.0025) | (0.0015) |
Y2010 | − 0.0860*** | 0.0607*** | − 0.0862*** | 0.0605*** | − 0.0857*** | 0.0616*** |
| (0.0102) | (0.0088) | (0.0102) | (0.0088) | (0.0102) | (0.0088) |
Y2011 | − 0.1498*** | 0.1112*** | − 0.1502*** | 0.1107*** | − 0.1496*** | 0.1120*** |
| (0.0099) | (0.0085) | (0.0099) | (0.0085) | (0.0099) | (0.0085) |
Y2012 | − 0.2612*** | 0.1343*** | − 0.2621*** | 0.1331*** | − 0.2610*** | 0.1352*** |
| (0.0098) | (0.0083) | (0.0098) | (0.0083) | (0.0099) | (0.0083) |
Y2013 | − 0.3392*** | 0.0219*** | − 0.3404*** | 0.0200** | − 0.3381*** | 0.0243*** |
| (0.0098) | (0.0083) | (0.0098) | (0.0083) | (0.0098) | (0.0083) |
Y2014 | − 0.3418*** | − 0.0128 | − 0.3429*** | − 0.0141* | − 0.3409*** | − 0.0101 |
| (0.0098) | (0.0083) | (0.0098) | (0.0083) | (0.0098) | (0.0083) |
Y2015 | − 0.3146*** | − 0.0442*** | − 0.3159*** | − 0.0454*** | − 0.3148*** | − 0.0432*** |
| (0.0097) | (0.0082) | (0.0097) | (0.0082) | (0.0097) | (0.0082) |
Y2016 | − 0.3322*** | − 0.0126 | − 0.3344*** | − 0.0146 | − 0.3344*** | − 0.0148 |
| (0.0127) | (0.0104) | (0.0127) | (0.0104) | (0.0127) | (0.0104) |
Constant | − 7.8634*** | − 0.3183*** | − 8.2640*** | − 0.6947*** | − 7.9642*** | -0.4579*** |
| (0.0402) | (0.0293) | (0.0703) | (0.0414) | (0.0414) | (0.0300) |
Industry | YES | YES | YES |
Rho | 0.0177*** | (0.0026) | 0.0175*** | (0.0026) | 0.0171*** | (0.0026) |
Wald chi2 | 91,642.56 | 91,100.79 | 91,858.51 |
Prob | 0.000 | 0.000 | 0.000 |
Observations | 493,010 | 493,010 | 493,010 |
Table 4
Marginal effects of Intangibles and Tangibles variables
Debt equation |
Intangibles | 0.00547 | 0.000004 | 0.00546 | 0.00548 |
Intangibles × Size | − 0.00028 | 0.000000 | − 0.00028 | − 0.00028 |
TotalME_Intangibles × Size | 0.00337 | 0.000001 | 0.00337 | 0.00337 |
Intangibles × Age | − 0.00016 | 0.000000 | − 0.00016 | − 0.00016 |
TotalME_ Intangibles × Age | − 0.00006 | 0.000000 | − 0.00006 | − 0.00006 |
Equity equation |
Intangibles | 0.00167 | 0.000001 | 0.00166 | 0.00167 |
Intangibles × Size | − 0.00035 | 0.000000 | − 0.00035 | − 0.00035 |
TotalME_Intangibles × Size | − 0.00054 | 0.000000 | − 0.00054 | − 0.00053 |
Intangibles × Age | − 0.00025 | 0.000000 | − 0.00025 | − 0.00025 |
TotalME_Intangibles × Age | − 0.00013 | 0.000000 | − 0.00013 | − 0.00013 |
Marginal effects of Intangibles in Debt and Equity equations were calculated following bivariate probit estimations (1) and (2) reported in Table
3. Average marginal effect of each variable was predicted on the joint probability distribution of
\({z}_{1it} \mathrm{and} {z}_{2it}\) based on the approach of Green (
2003). As we estimated bivariate probit via multivariate probit suggested by Cappellari and Jenkins (
2003) using Stata 17, predictions following mvprobit are obtained using the command mvppred. For interaction terms, we calculate the marginal effects of interaction terms (Intangibles × Size; Intangibles × Age) following bivariate probit estimations (3)–(6). Then we estimate the full interaction effects (TotalME) of interaction terms which equal the cross-partial derivative of the expected value of dependent variables
\({z}_{1it} and {z}_{2it}\) based on the approach suggested by Norton et al. (
2004)
As observable from Table
3, the coefficients associated with the variable
Intangibles are positive and statistically significant in both debt and equity equations, suggesting that intangible-rich firms have better access to external finance than firms with fewer or no intangible assets. This finding thus fully supports hypothesis H1.
Hypothesis H2 theorises the relative attractiveness of intangible assets to debt financers and equity investors. To test this hypothesis, we compare the marginal effects of intangibles in debt and equity equations: \(54.7\times {10}^{-4}\) and \(16.7\times {10}^{-4}\), respectively. We can interpret these figures as follows. On average, for any 100% increase in the value of intangible assets (i.e. 1 unit increase in Intangibles variable), the probability for firm to use debt finance increases by 0.547 percentage points whereas the probability for firm to use equity finance increases by only 0.167 percentage points, holding all other variables constant. Although the magnitude of the variation in the two effects is economically humble, they are statistically significantly different from each other. Although this finding runs contrary to the extant literature, it is sensible in the context of Vietnam’s less-developed equity market and supports the study’s hypothesis H2.
To explain the result, we need to take into account the nature of the equity market for SMEs in Vietnam. Unlike large firms, Vietnamese SMEs are not listed in stock markets which, according to the efficient market hypothesis, could reflect intrinsic value of firms (Malkiel,
2005). Moreover, the alternative equity markets are insufficiently developed to be accessible to the majority of Vietnamese SMEs (Scheela & Dinh,
2004; Scheela et al.,
2018). Prior research indicates that informal finance in the form of money borrowed from family, relatives and friends is the most important source of equity financing for SMEs in Vietnam (Nguyen & Canh,
2020). These unprofessional and network-based investors may not fully appreciate the value of intangible assets (Mateus & Hoang,
2021; Thornhill & Gellatly,
2005), which they treat as a complement to rather than a substitute for asset tangibility. However, the banks in Vietnam will consider the values of intangibles when making their lending decisions. There are probably two reasons for this. First, the intangible assets in our study are separately identifiable, valuable and potentially collateralisable, and are instrumental in generating cash flows (i.e. priced, audited and reported in balance sheets). Hence, banks can assess their value at low to no cost. As such, they may be treated like tangible assets in the lending decision-making processes (Lim et al.,
2020). Second, the secondary markets for trading intangible assets (e.g. trademarks, franchises, copyrights) is starting to develop in Vietnam (Quan et al.,
2020). Therefore, banks may find it feasible to liquidate the intangible assets that firms have pledged as collateral in the event of payment default. For these very specific characteristics of a developing financial market, the effect of intangibles is stronger on debt finance than equity finance in the case of Vietnamese SMEs.
Turning to the moderating effect of firm age and size, it may be seen from Table
3 that the coefficients associated with the interaction terms between intangibles and firm size/age are negative and statistically significant in all specifications. This finding indicates that the effect of intangible assets on access to both debt and equity is stronger for smaller and younger firms. In a dynamic and technology-intensive economic landscape, nascent and small firms are more flexible and willing to rely on intellectual capital when becoming involved in innovation activities, whereas larger and older firms with established track records and fixed assets are less likely to do so (Petruzzelli et al.,
2018). As a result, older and large firms are unlikely to leverage their intellectual assets to gain access to external finance. Conversely, smaller and younger firms, which are relatively free of bureaucracy and which have the autonomy to explore innovative solutions, are preferred by lenders and investors when they have more intangibles (Petruzzelli et al.,
2018). As such, hypotheses H3 and H4 are supported.
Regarding firm-specific factors, the estimated results show significant positive coefficients for firms’ size and age to debt financing, implying that access to bank loans is easier for bigger and older firms as they are perceived by lenders to be less risky. This is highly consistent with the extensive literature on SMEs’ financial constraints (Ullah,
2020). Firm age also positively affects equity financing, but firm size surprisingly shows a negative effect on access to equity. One possible explanation for this finding is that smaller firms are keen to grow by acquiring and investing more equity, whereas bigger entrepreneurs are more likely to generate profit, preferring to use their internally generated funds via retained earnings rather than external sources of equity finance. Interestingly, the cash variable has a positive effect on equity financing but negatively impacts firm accessibility to debt finance. The level of cash holding contains information about a firm’s investment opportunities (Ferrando & Mulier,
2015), hence sending a good signal to equity investors and incentivising them to provide funding. Although this finding runs contrary to the literature, it lends support to Lim et al. (
2020) who argue that firms that generate substantial cash may prefer to be debt-free so that its managers are not subject to scrutinisation by lenders. Growth in profit boosts the probability of accessing debt finance because in Vietnam’s conservative credit environment, profit is considered to be the factor that most affects borrowing capacity and the likelihood of repaying loans. However, it seems that entrepreneurs with increasing profits are more likely to use internally generated funds (i.e. retained earnings), which is consistent with our argument about the effect of firm size on equity finance. Lastly, the effect of the
Z-score is significant for debt finance because this represents the credit risk assessment employed by banks and financial institutions in their lending decisions. However, its effect on equity financing is negligible.
4.2 Robustness check
We conducted two robustness checks. First, we split the sample into two periods 2008–2011 and 2012–2016 to test whether the findings are different during and after the GFC. This also enables us to examine the relative importance of intangible assets in the economic conditions of an era that is becoming increasingly digitalised and knowledge-based. During 2010/2011, the Vietnamese government introduced several measures aimed at enabling firms to cope with the uncertainty generated by the GFC and at improving the competitiveness of the domestic private sector. These included financial support schemes for firms to engage in R&D activities, and training schemes geared to transforming firms’ business models from labour intensive to knowledge intensive (Haaker et al.,
2021; Vu & Tran,
2020). The changing economic and institutional conditions inevitably led to the increasing importance of intangible assets. For example, Quan et al. (
2020) evidently show that since 2010, intangibles have made up a large share of the total assets of Vietnamese firms. Based on their arguments, we use a split-sample method to examine the increasing importance of intangibles to Vietnamese SMEs.
The regression results are presented in
Appendix, Table
5. It can be seen that intangibles helped SMEs obtain debt and equity finance both during and after the GFC. This finding reveals that banks in Vietnam began to accept intangible assets as collateral some time ago. In terms of the moderating effects of firm age and size, the findings are quite consistent with the full sample analysis. Specifically, intangibles are essential to smaller and younger firms when they seek debt and equity finance.
Second, we employ a multinominal logit model to estimate the effect of intangibles on three groups of firms: (i) those with access to debt finance only, (ii) those with access to equity finance only and (iii) those with access to both sources of external finance. Even though this approach fails to take into account the simultaneously determined process of debt and equity financing decisions, it provides us with a nuanced understanding of the effects of intangibles on access to debt and equity finance when they are examined individually and collectively. The results are presented in Table
6 in the
Appendix. They show that intangibles have a significant positive effect on both debt and equity finance, with higher coefficients being observed in debt than in equity equations. In terms of the moderating effects, the findings are consistent with the main results; that is, smaller and younger firms gain more debt and equity from intangibles than their larger and older counterparts.