![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
ABSTRACT
Using data of Chinese listed firms, we find that firms that purchase financial products from banks are more likely to obtain loans from these banks than firms that do not buy. The probability is 52% and 58% higher for the three-year and one-year look-back window, respectively. This phenomenon is more pronounced for borrowers who are non-SOE or low-risk ones. Our findings reveal that by purchasing financial products from banks, firms can build relationship with them and consequently obtain loans more easily. Our paper contributes to the extant literature on the channels and mechanisms of relationship building between banks and firms.
Disclosure statement
No potential conflict of interest was reported by the authors.
This work was supported by the National Natural Science Foundation of China [71672188,71872010,71902009]; The Fundamental Research Funds for the Central Universities, and the Research Funds of Renmin University of China [19XNL007].
Notes
1. The costs a bank incurs when doing a credit check on a firm.
2. Purchasing financial products from banks can generate revenues for firms. Yields from financial products are higher than for deposits for most of the times.
3. However, the research of Hens and Rieger (Citation2011, Citation2014) finds that purchasing financial products may not bring expected profits for investors.
4. Chorafas (Citation2006) believes that wealth management requires unique skills that are not widespread and that firms feel that the complex technology used in banking can get them a better deal than managing their wealth by themselves.
5. Investors with incorrect beliefs arising from probability weighting or probability misestimation.
6. Since firms may tend to borrow from their deposit banks, the observed Purchase and Borrow Association (‘PBA’ hereafter) between banks and firms may be induced by the deposit behavior of the firms. However, after checking the bank account management regime in China as well as the financial product purchasing data and firm borrowing data, we excluded this concern. Specifically, according to the Bank Account Management Regulation of China, each company should hold only one basic bank account in its principle bank where the majority cash flows of the company reside. However, according to our statistical data, each company on average obtains bank loans from 3.67 different banks every year. During our sample period 2007–2015, each company obtains loans from an average of 12.03 banks. As for the financial products, each company on average purchase financial products from 3.46 different banks every year. The number of stock-years that purchase financial products from more than one bank in each year accounts for 80.60% of all the stock-years. When we examine financial product purchasing behaviors during our whole sample period 2007–2015, each company on average purchase financial products from 7.56 different banks. The number of stock-years that purchase financial products from more than one bank during the sample period accounts for 90.17% of all the stock-years. As can be concluded, most of the firms purchase financial products from more than one bank and obtain loans from various banks. Therefore, these numeric facts contradict the one principle bank account institution in China, making the deposit-driven endogeneity unreliable.
7. It cannot be excluded that some firms may choose to purchase financial products from banks solely for the purpose of pleasing and building relationships with these banks with the expectation of potentially obtaining more far-reaching financial support from them.
8. Because relationship lenders gain an information advantage over a non-relationship lender.
9. Another possibility is that even if SOEs have purchased financial products from a bank and have established a relationship with that bank, they do not necessarily apply for support from that bank when they need capital because they have numerous alternative channels from which to raise money, such as the capital market, government or even other banks that require fewer debt covenants. In addition, SOEs are less efficient in their investments (Chen et al. Citation2011) and wealth management, so they are less likely to endeavor to invest their spare money in financial products to raise cash management efficiency.
10. Factors such as the rapid growth of the credit derivative market, rise in the bankruptcy and developing credit risk literature led to shift emphasis on modeling and evaluation of credit risk at the start of 21st century which is marked by the devastating financial crisis (Singh and Mishra Citation2016).
11. This is based on the assumption that the banking market is a lenders’ market.
12. When deciding to extend credit to a PBA borrower, banks face the opportunity costs of their reduced ability to lend to non-PBA borrowers. Given the value of the financial products purchased by PBA borrowers, only when the financial risks of the non-PBA borrowers are sufficiently low, will banks lend to non-PBA borrowers rather than PBA borrowers. In summary, all other conditions being equal, banks will choose lower-risk borrowers when deciding whether or not to build PBAs with firms and with what kind of firms to build these PBAs.
13. The report was issued by the central clearing company’s national banking information registration system.
14. Some companies purchased dozens of these financial products in a very short time frame, four companies bought over 100 financial products in 2015 and six companies spent over 10 billion RMB (about 1.5 billion USD) on these financial products (Wind Financial Database).
15. We exclude financial industry firms because their purpose(s) for applying for bank loans may be different from those of firms in other industries.
16. Our sample period begins in 2007 as this was when data for financial products in CSMAR was first collected.
17. For each loan record that includes several bank lenders, the record is split for several loan contracts between the borrower and each lending bank, and the amount of the loan is averaged for each bank.
18. The names of both listed and unlisted banks in China, downloaded from the Bankscope Database, are used as the reference when identifying and numbering banks by their names.
19. Bharath et al. (Citation2007) economize on the size of the data set by keeping only transactions by lead banks ranking in the top 40 banks by market share in the prior year. Our sample differs from Bharath et al. (Citation2007) in that we do not exclude banks that rank low by market share because of the particular feature of financial industry in China. In China, the number of banks is limited, and the size is considerably large for all banks. Thus, we prefer to keep all banks in our sample. For robustness purpose, we condense our sample based on market share of banks following Bharath et al. (Citation2007).
20. As a robustness test, for each financial product in a specific year we check on whether the company has obtained loans from the bank during the following one-year and three-year window. Our conclusions are unaltered. We acknowledge that we cannot exclude the reverse causality for sure, but our one-year look-back window and three-year look-back window design can help to alleviate this concern by providing time sequence evidence.
21. This sample selection process is largely based on a one-year look-back window. The process based on the three-year look-back window is very similar to the one-year window, except for the trivial difference in the number of observations left due to sacrifice in different look-back periods.
22. Detailed data processing procedures are illustrated in Appendix B.
23. These results are equally strong if we use the other four continuous (,
,
and
) measures that take into account both the existence and intensity of past purchasing experiences.
24. Our additional test can to some extent help to alleviate the endogeneity concern considering the data characteristics in our sample. Because unlike the data structure used in our baseline model, the two kinds of designs in the additional test section use firm-year structure which is quite common and avoids the redundant data concern.
25. We provide some evidence in Appendix C.
26. Robustness tests are conducted on Models (1), (1a) and (1b). Most of the results for robustness tests are consistent with those of the three models, supporting our three hypotheses. For brevity and space saving, we don’t tabulate the results for the robustness tests of Models (1a) and (1b).
27. PROB is estimated by using Column (1) in Table C.1. The results using other columns of Model (C.1) to estimate the predicted probability of purchasing financial products from banks for each firm-year are similar to the results here. For brevity, we do not tabulate the results using the other columns of Model (C.1).
28. While all of the results are supportive, for brevity we only present the results for the 30th and 90th percentiles in , Panels C and D, respectively.
29. We also conduct this robustness test using different percentiles of bank retention (including 50%, 30%, 20%, 10% and 5%) by the loan market shares. Most of the results are consistent with our three hypotheses.
30. We argue that since extending loans is one of the main businesses of banks, PASTLOAN, a dummy variable that equals one if a particular firm borrowed from a particular bank in the past three years, can also be an ideal indicator reflecting past business connections between a specific bank/firm pair (Bharath et al. Citation2007).
31. We decompose CONNECTION into three indicators showing different connections between firms and banks, and control for each of them one by one in the model. Specifically, the three indicators are: (1) FIRMHOLDBANK, a dummy variable taking the value of one if a firm is one of the shareholders of a particular bank, zero otherwise; (2) BANKHOLDFIRM, a dummy variable taking the value of one if a bank is one of the shareholders of a particular firm, zero otherwise; (3) COMMONEXECUTIVE, an indicator variable that equals one if a firm and a bank have at least one person as their common executives, zero otherwise. The results remain the same. For brevity the results are not tabulated.
32. A detailed definition of all the variables in the model is listed in Appendix A.
33. We ask researchers to interpret these results with caution because the exact purchase and benefit periods are uncertain.
34. Number of firms that purchased financial products from banks.
35. Number of firms that received bank loans.