Trade credit in transition economies: does state ownership matter?

ABSTRACT

This article investigates the effect of residual state ownership on trade credit in Vietnam. The empirical results show that a substantial withdrawal of the state from listed state-owned enterprises (SOEs) does not disturb these firms’ supply of and demand for trade credit, suggesting that trade credit is not a source of soft budget constraint and local privatisation programmes have not been motivated by efficiency. Privatisation seems to be a transitional phenomenon in a state-dominated economy that is inadequately supportive of private sector development and a poor legal system encouraging informal contracts. Interestingly, Vietnamese firms with more accounting conservatism, probably in response to their creditors’ increasing demand for financial reporting quality, are found to provide less trade credit, suggesting that adopting more conservative accounting arising from the government’s privatisation may be associated with firms’ better accessibility to bank credit.

KEYWORDS:

• Trade credit
• state ownership
• accounting conservatism
• privatisation
• transition economies

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. These advantages include the suppliers’ resourceful superiorities in acquiring relevant information (Biais & Gollier, 1997; Burkart & Ellingsen, 2004; Jain, 2001; Smith, 1987), enforcing/monitoring contractual terms (Cunat, 2006; Johnson et al., 2002; Kandori, 1992; McMillan & Woodruff, 1999; Petersen & Rajan, 1997), and repossessing properties from a liquidation event (Frank & Maksimovic, 1998; Giannetti et al., 2011; Longhofer & Santos, 2003; Mian & Smith, 1992; Petersen & Rajan, 1997).

2. Demirgüç-Kunt and Maksimovic (2001) define international determinants of trade credit by using a large sample of manufacturing firms in 40 countries.

3. There are other reasons underlying the use trade credit: The financial motive such as tax deductibility (e.g. Brick & Fung, 1984) and the commercial motive such as demand controlling or business cycle smoothing (e.g. Burkart & Ellingsen, 2004; Emery, 1987; Ferris, 1981; Long et al., 1993; Meltzer, 1960; Nilsen, 2002), product quality guarantee (Deloof & Jegers, 1996; Emery & Nayar, 1998; Lee & Stowe, 1993; Long et al., 1993; Pike et al., 2005; Smith, 1987), and sales support or price discrimination (Brennan et al., 1988; Mian & Smith, 1992; Petersen & Rajan, 1997; Schwartz & Whitcomb, 1978). For an extensive review on relevant theories, see Petersen and Rajan (1997), Fisman and Love (2003), Delannay and Weill (2004), Bastos and Pindado (2007), and R. Chen et al. (2018).

4. Smaller SOEs are often loss-making companies with little contribution to state budget revenues. First privatisation pushing on small SOEs is a central government’s attempt to improve economic performance as well as avoid asset stripping and rent distribution at lower tiers of the government system (Cheshier et al., 2006).

5. The later decline (since 2000) in Vietnam’s SOEs share does not come from an expansion of the private sector but from the growth of foreign investment (Karadjis, 2005).

6. Vietnamese government refers to this process as ‘equitisation’ (‘cổ phần hóa’) rather than ‘privatisation’ (‘tư nhân hóa’). This is partly because ‘the state normally retains a large stake in most of the equitised SOEs, and only a limited amount of their shares are sold to private investors’ (Le, 2017, p. 2) and ‘the percentage of shares transferred to insiders are quite substantial’ (Loc et al., 2006, p. 350).

7. In fact, the period 2011–2015 witnessed an extensive equitisation in Vietnam, which leads to a significant decrease in the number of SOEs (from 1,309 to 958) as well as a result of almost 600 SOEs being equitised (96% of the targeted number in the equitisation plan of the government) and 538 SOEs going to public (IPO) (KPMG, 2017).

8. It also should be noted that the increasing divestment of state assets in this period may be the result of the development of the Vietnamese stock market since 2011 (in terms of size and liquidity), which provides an effective, instrumental condition for such a divestment strategy.

9. Some evidence in other emerging markets also indicates that financial crises exert instantaneous and persistent impacts on trade credit provision (e.g. Love et al., 2007).

10. There has also been plenty of evidence showing the relevance of monetary and macroeconomic conditions to trade credit provision (see Bastos & Pindado, 2013; Baum et al., 2003; Dai & Yang, 2015; Nilsen, 2002) or substitute/complementary relations between trade credit and bank credit (Cull et al., 2009; Delannay & Weill, 2004; Shenoy & Williams, 2017; Uesugi & Yamashiro, 2008). As introduced in Tran and Le (2017) and Le and Tran (2015), the FCI is a composite index reflecting the short-term variations in financial variables, including private credit, stock market indexes, interest rates, and exchange rates. In a bank-centred financial system like in Vietnam, the FCI is related to fluctuations in Vietnamese firms’ accessibility to bank finance.

11. With all regressions, we include foreign ownership as an indicator of the effect of Vietnam’s internationalisation process. All estimates of foreign ownership are insignificant. Vietnamese government has set up limitations and restrictions on foreign involvement in local enterprises. Also, SOEs which are often underperforming and inefficient compared to private firms may not be the potential targets of foreign investors. We do not report these results for the sake of brevity but are available from the authors upon request.

12. Vietnam has not so far completely adopted the International Financial Reporting Standards (IFRS).

13. For simplicity, superscripts are omitted.

14. Trade receivables and trade payables in this study refer to accounts receivable and accounts payables, respectively.

15. Corporate governance may also affect trade credit (Dai & Yang, 2015). Ownership concentration – combined ownership of the largest shareholders – is an effective mechanism of corporate governance in Vietnam characterised by a weak institutional environment (Nguyen et al., 2015; Tran and Le, 2020). However, our sample shows a high correlation between state ownership and ownership concentration. Therefore, our specification does not include ownership concentration to avoid the multicollinearity problem (Boubaker et al., 2017). Even so, our examination indicates that when included, these variables do have a statistical significant effect on trade credit. Our remaining results remain qualitatively the same.

16. For instance, industry-specific differences in trade credit are documented by Fisman and Love (2003). Firm size and profitability may affect the degree of adverse selection costs, which should impact firm policy on trade credit extension (Bastos & Pindado, 2007).

17. All right-hand-side variables in Equaiton (1) are lagged by one year to alleviate simultaneity issues. For 2SLS regressions, we use predicted state ownership to instrument the potentially endogenous state ownership in the first stage. The intuition of this external instrument, as in Becker and Milbourn (2011), is that a faster decrease in state ownership is predicted in firms that start out with a higher state ownership early in the sample. Accordingly, the value of state ownership predicted for firm i in t^th year after the earliest year t₀ (observable for firm i in our unbalanced data sample) is $STATE\mathrm{\_}OW{N}_{i,t}=STATE\mathrm{\_}OW{N}_{i,t_0}+t\times \left[\left(\overline{STATE\mathrm{\_}OWN}-STATE\mathrm{\_}OW{N}_{i,t_0}\right)/\left(2015-t_0\right)\right]$, where $1\le t\le 5$, $2010\le t_0\le 2015$, and $\overline{STATE\mathrm{\_}OWN}$ is the sample average of STATE_OWN.

18. 2SLS and Heckman approaches are also implemented for comparison.

19. The relative level of high/low conservatism is determined by the median of the sample.

20. To mitigate the influence of outliers, we winsorise all the variables in the model at conventional levels.

21. It should be noted that the differences in state ownership for the two cases (22% lower right after and 20% lower at one year after privatisation) are based on the sample of 109 NPFs and a reduced sample of 75 NPFs, respectively.

22. Splitting the full sample into groups of large and small firms also shows a considerable dependence of trade credit levels on firm size. Trade credit terms are significantly higher in smaller firms (see Table C2 of Appendix C).

23. For 2SLS regressions, the results for the first stages are not reported for the sake of brevity. It also should be noted that all estimates of predicted state ownership (the instrument) in the first stages are strongly significant.

24. The results for estimating C_SCORE as a proxy for accounting conservatism are shown in Appendix B.

25. It should be noted that our out-of-sample evidence on the determinants of trade credit is robust to other regression approaches, including robust-cluster OLS, 2SLS, and Heckman selection bias correction. In other words, we re-estimate specification (1) using all these approaches on the sample of 480 firms. For brevity, these results are untabluated and available upon request.

26. The estimated coefficients of C_SCORE on TC_PAY in Tables 5 and 6 are positive. It is plausible that more conservative firms (higher C_SCORE) tend to prefer providing less (lower TC_REV) and receiving more trade credit (higher TC_PAY). However, the positive effect of C_SCORE is only significant at the 5% level in the model of SCE in Table 6 (i.e. Column (6)). It remains statistically insignificant in the models of NPF and SOE (columns (4) and (5), respectively). This implies the absence of robust results regarding TC_PAY. In fact, out-of-sample evidence in Table 7 shows insignificantly negative estimates of C_SCORE in models of payables.

27. Our main results in this study remain the same when we use alternative measurements of the dependent variables such as logarithmic transformations, asset-scaled ratios, and alternative variables for firm size (logarithm of total assets, instead of SALES_LN) and firm profitability (return on assets, instead of ROS), and alternative specifications without lagging the right-hand-side variables or with additional variables such as foreign ownership or corporate governance (namely, ownership concentration), or/and non-winsorised data.

28. In this perspective, our result that a higher level of cash flow (higher information asymmetry) is related to a lower level of trade credit provided is surprising.