The Effects of Ownership Structure on Borrowing Rate: The Case of Listed Iranian Companies

Abstract

The Iranian banking system engages in financial repression, and there are legal restrictions on interest rate. However, micro-evidence shows that the interest rates paid by Iranian firms in this condition are different across firms. This article aims to investigate the roots of the difference in borrowing rates by exploring the effects of the borrowers’ ownership structure and bank–firm relationship features. Using data collected from companies listed in the Tehran Stock Exchange (TSE) for the period 2007–2016, empirical models are estimated through applying the dynamic panel data regression method. According to the estimation results, the presence of an institutional stockholder, and particularly banks, among a firm’s shareholders can generally reduce its borrowing rate. Moreover, the results show that the financing rate is significantly lower in the firms with more than 20 per cent of their shares owned by the government. In addition, the findings suggest that borrowers of unhealthy banks, in terms of non-performing loans (NPL), bear a higher finance rate.

Keywords

Borrowing rate bank–firm relationship shareholder structure non-performing loan (NPL)stock market companies

I. Introduction

The effects of ownership structure on a firm’s performance have been studied in the agency theory framework developed by studies like Jensen and Meckling (1976). In this context, a considerable amount of research has examined the impact of ownership structure through the manager–shareholder agency problem framework (see Ang et al., 2000; McConnell & Servaes, 1990; Morck et al., 1988; Shleifer & Vishny, 1997), but the ownership structure effect has been less studied in the lender–shareholder framework. Anderson et al. (2003) examine the family ownership structure and show that founding-family ownership is related to a lower cost of debt financing. Some earlier studies indicate that a firm’s ownership structure strongly influences its agency cost of debt. Kim and Sorensen (1986) find a higher percentage of inside-ownership influence on the agency cost of debt. Also, Moh’d et al. (1998) show that ownership concentration affects the agency cost. Since there is a tight relationship between the agency cost of debt and the borrowing rate across borrowers, the direct relationship between the borrowing rate and ownership structure can be examined as a hypothesis. Although the effect of ownership structure on the borrowing rate has been investigated in a few aspects, the effect of different types of institutional ownership on the cost of debt has not been studied extensively. Thus, in this article, we examine whether the different existing types of institutional shareholders reduce borrowing rates. In the studies devoted to identifying borrowing-rate determinants, the borrower’s financial situation plays the central role (Altman, 1968; Dimitras et al., 1999; Emel et al., 2003; Ohlson, 1980). However, some studies, such as Altman and Saunders (1997), suggest that the only way to evaluate borrower risk is not to take advantage of financial information and ratios. Thus, besides the financial characteristics, many studies have aimed to identify which factors can influence borrowing rates in different ways. For instance, Diamond (1989) and Pittman and Fortin (2004) show that a firm’s age is a relevant factor in determining its borrowing rate. Lehman and Neuberger (2001), Petersen and Rajan (1994) and Gambacorta (2008) study the effects of the bank–firm relationship characteristics on the rate. Lin et al. (2011) and Chava et al. (2009) observe the impact of governance structure, and some studies like Park and Woo (2009) and Graham et al. (2008) address the effects of the information transparency of borrowers. However, little work extensively examines the impact of ownership structure on the borrowing rate. This article investigates the determinants of the borrowing rate at the firm level, with more emphasis on ownership structure and institutional ownership. Also, using bank–firm data enables us to observe the bank–firm relationship’s effect as a determinant factor of changes in the rate. Because of information asymmetry and financial frictions in financial markets, creditors are not able to completely assess the borrowers’ risk using their accessible financial data. In the undeveloped and bank-based financial systems, these problems become more difficult, because the assessment process of potential borrowers is not concentrated in the specialized credit scoring agencies and each bank independently assesses its own borrowers’ risk and sets its rates. Thus, in an undeveloped and bank-based system like the Iranian credit market, institutional variables like ownership structure and bank–firm relationship have more power to explain the variation in the rate across borrowers and time.

In this research, we examine the impact of institutional factors, such as ownership structure, bank–firm relationship and financial health of the lenders, on the financing rate. The main hypothesis is that the companies (stock market companies) that are owned by banks, governments and large holding companies as major stockholders can get loans with lower financing rates. Also, we investigate whether the financial health of lenders and other features related to the bank–firm relationship (features of lenders such as bank type [private or public], relationship length and numbers of the firm’s lenders) can affect the borrowing rate. To do so, we use bank–firm matched data. The matched data allow us to test which characteristics of borrowers and lenders influence the borrowing rate. The main contributions of this article are threefold. First, the existing studies on determinants of borrowing rate ignore the effect of different types of institutional ownership. In this article, we analyse the impact of the presence of institutional owners on borrowing rates using bank–firm data. Also, we examine whether bank–firm factors and bank-specific characteristics drive borrowing rates. Second, using detailed data on the ownership structure of firms allows us to investigate the impact of shareholder structure on financial cost, in terms of new aspects, and to address questions that have not been answered yet. For instance, does the indirect ownership of borrowing firms by the major financial and political institutions affect the borrowing rate similar to direct ownership? Is there a threshold in the relationship between the share of institutional owners and the financing rates? Are the influencing factors of borrowing rates different across firms? Finally, unlike the majority of studies in this area, these relationships are examined in a developing country with an undeveloped, bank-based financial market that suffers from financial repression. World Bank (2001) finds that information asymmetry of financial markets is stronger in developing countries than in advanced economies. Thus, it is expected that the role of the institutional factors such as ownership structure and bank–firm relationship in allocating loans and determining loan rates becomes more important. According to the findings, the borrowers’ ownership structure drives the borrowing rate, that is, in a market with information asymmetry, among borrower stockholders, the presence of banks having a more centralized ownership structure causes a decrease in the borrowing rate. Among the bank-specific variables, a higher non-performing loan (NPL) ratio in the creditor bank significantly increases the financing rate of borrowing firms.

In Section II, a literature review of the theoretical and empirical studies describing the factors influencing the borrowing rate. In Section III, we look at the statistics on the borrowing rate and its determinants and explain the methodology of statistical analysis. In Section IV, major findings of estimations are presented and interpreted under several subsections, and finally, the last section concludes the article.

II. Review of Literature

The most significant factor explaining the difference in financial cost between firms is their profitability and financial risks. In this regard, different studies elaborate the difference in borrowing rates of firms by considering various indicators, such as the firm’s size (Booth, 1992; Houston et al., 2014), age (Diamond, 1989; Pittman & Fortin, 2004), cash flow fluctuation (Lin et al., 2011) and liquidity status (Pittman & Fortin, 2004), as a proxy for risks of a borrower firm. In addition to the aforementioned factors, a wide range of structural and institutional features related to the firm and the business environment have been examined as factors affecting the financial cost of loans. Therefore, the literature review focuses on three strands of research grouped according to these factors.

The first strand of research in the literature focuses on the effect of different dimensions of the bank–firm relationship on a firm’s financial cost. The key purpose of studies in this area is to analyse asymmetric information in the relationship between the creditor and the borrower. Diamond (1989) argued that since long-term relationships between banks and borrowers mitigate the problem of asymmetry of information, a long-term relationship is beneficial to borrowers and reduces the financial cost to borrow. Using data for US firms, Petersen and Rajan (1994) showed that the type of relationship created between the bank and the firm has a significant impact on the costs and the availability of loans for firms; a close relationship of the firm with the bank decreases the loan cost and increases loan availability. They also found that if firms get loans from several banks instead of relying on a single bank, the financial cost will increase. Lehman and Neuberger (2001) studied the small and medium enterprises (SMEs) of the German economy as a bank-based economy and concluded that the relationship and interaction between firms and banks have a significant effect on the loan cost, required collateral for loans and access to loans by firms. Gambacorta (2008) approved this relationship using a different method. He demonstrated that in order for banks to determine the loan rate, they consider some factors, such as the firm’s liquidity and capital and the depth of relationship with the borrowers. Hubbard et al. (2002) illustrated that the financial health indicators, particularly the capital ratio of creditor banks, affect the loan rate for the borrowing firms, in such a way that getting loans from banks with weak financial soundness can increase the financial cost. Assessing the financial crisis in 2007, Santos (2011) showed that higher financial costs were imposed on firms that borrowed from risky banks during the crisis. In this article, regarding the background and effect of information asymmetry on the financing rate, we examine whether the variables that illustrate the quality of the relationship between the creditor banks and borrowing firms (such as the number of creditor banks or the bank–firm relationship length) can influence a firm’s borrowing rates in the Iranian economy with an undeveloped financial and banking system.

The second group of studies in this area examines the impact of the ownership structure of borrowers on the loan rates applied to them. These studies show that the shareholder structure and the board composition of the borrowing firms influence the financing rate through affecting the firm’s performance and risk. In this regard, some studies, such as Lin et al. (2011) and Chava et al. (2009), focus on the governance structure of borrowers and state that the board composition has a significant impact on the financing rate. They arrive at the result that a less focused and decentralized decision-making mechanism in companies impose a higher financing cost than a centralized control system that allows a small group to make decisions. Moreover, Houston et al. (2014) showed that firms with political connections benefit from low-interest-rate loans. The authors explained the relationship between access to lower loan rates and the existence of political connections with the board through two channels: borrower channel and bank channel. In the borrower channel, political connections reduce financial risk, and this leads to decreasing the financial cost. In the bank channel, bankers themselves seek to establish political relationships and tend to lend to firms and individuals with political connections. Thus, low-interest-rate loans are more accessible for borrowers with political connections. Anderson et al. (2004) demonstrated that the independence and size of boards have an inverse relationship with the loan rate. Some studies in this area examine whether the composition of stockholders affects the borrowing rate. Robert and Yuan (2010) examined the influence of institutional shareholders on the financial cost of a firm and found that the institutional shareholders cause a decrease in the financial cost. Also, they showed that the relationship is stronger in markets with more information asymmetry. Also, Sanchez-Ballesta and García-Meca (2011) studied the impact of the appearance of banks and the government as stockholders on reducing the borrowing rate. Borisova et al. (2015) studied the effect of ownership structure on financial costs in different situations. Through examining firms of 41 countries for the period 1991–2010, they illustrated that the presence of the government as a stockholder in a borrowing firm during non-financial-crisis periods increases the financial cost. Conversely, in financial-crisis periods, the existence of governmental stockholders can decrease the financial cost.

Generally, the effects of ownership structure on the financing rate can be analysed through two channels. First, since there is asymmetric information in the credit markets, information about the ownership structure of borrowers can be useful in the loan risk assessment process. Thus, the existence of a powerful shareholder can contribute to lessening the estimated default risk and reduce the loan rates. Second, according to empirical studies, the ownership and governance structure can influence the performance of a firm and its operational risk and therefore affect the loan rate. Thus, the existence of institutional shareholders can potentially reduce the borrowing rate through these channels. This article aims to answer the question: does the presence of banks, large holding companies or the government as stockholder(s) in a firm reduce its borrowing rates?

The third strand of research in the literature investigates the effects of the information transparency of firms. Park and Wu (2009) and Graham et al. (2008) showed that frequent adjustments of financial statements lead to an increase in the financial cost. Graham et al. (2008) also suggested that the loan rates, required collateral, loan maturity and loan contract terms for firms altering their financial statements (relative to their predicted budget) become worse than before the adjustment of the information. Some studies in this area address the role of audition in the loan rate. Anderson et al. (2004) found that the independence and size of the audit committees are inversely related to the cost of debt.

In general, according to the literature, it can be stated that the financing rates vary across firms based on their operational risk. Thus, some indicators and ratios that reveal a firm’s financial status in terms of profitability, liquidity and operational risks are able to affect its financial cost. However, given the asymmetric information in the credit markets, creditor banks use all financial and non-financial information in the loan risk assessment process. Institutional variables, such as the ownership structure of firms or the background of bank–firm relationships, can lessen the risk of asymmetric information between banks and firms. Thus, these factors can influence the loan rates and loan accessibility. More specifically, the presence of institutional shareholders (banks, large holding companies and government sectors), long-lasting relationships with banks (i.e., long-term or short-term relation), number of a firm’s lenders and NPL ratio of creditor banks are the most important factors that can make a difference in borrowing rates. In this research, we empirically analyse the difference in the loan rate across firms and investigate the effect of the mentioned factors using data obtained from Iranian manufacturing companies.

III. Data and Statistics

To investigate the determinants of financial cost, we need detailed financial statements at the firm level. Thus, we use the financial information of 313 companies listed in the Tehran Stock Exchange (TSE). The studied data are organized in an unbalanced panel for the period 2007–2016. To remove outliers, we drop the observations below the 1st percentile and above the 99th percentile.¹

The income statement and balance sheet of the TSE-listed companies are used to measure their borrowing rate. The main indicator for the borrowing rate is calculated by dividing the financial cost (an item selected from the income statement) by the average balance of total loans (the balance sheet item) for the current and previous period.² Since the inflation rate in Iran has been considerably high during the studied period, there is a significant difference between the nominal rates and real rates. Hence, to eliminate the inflation bias in the estimations and analysis, we use real borrowing rate (r_t) as the main dependent variable throughout the article. Then, after calculating the nominal borrowing rate in each firm-year observation, we subtract the calculated nominal borrowing rate from the inflation rate in that year. This research investigates the roots of difference in the loan rates across firms. In this regard, Figure 1 illustrates the histogram distribution of the nominal borrowing rate among the listed companies in 2016. It maps the observed dispersion of the rate among companies. Finding the reason for the different rates and this dispersion is the main purpose of the present study. Figure 2 demonstrates the average nominal rate and the real rate of loans, along with the producer price index. The average nominal rate has increased over the years, while its fluctuations compared to inflationary changes are not significant.

Figure 1.

Source: The author.

Figure 2.

Source: The author.

The variables related to the ownership structure of the borrowers theoretically play an important role in the difference in borrowing rates across the firms. We measure several indicators related to the ownership structure of companies by using the published list of shareholders who own more than 1 per cent of a company’s outstanding shares. The first indicator measures the shares of banks and credit institutions in the shareholder structure (sh_bnk). Stockholding by the government and public companies (sh_gov) and shareholding by the large holding companies³ (sh_hld) are two ownership structure variables. In these three indicators, in addition to direct stockholding, owning indirectly through other companies has also been considered.⁴ We also measure the shares of the retail shareholders (sh_oth) as the fourth indicator. It is expected that the larger share of institutions (bank, the government and holding companies) in a borrower firm decrease information asymmetry and also positively affect the firm’s performance. Thus, this reduces the borrowing rate of the firm. It seems that stockholding by these institutions does not have the same effect on the rate. We expect that the loan rates of firms owned by banks will be lower than those of firms owned by large holding companies and the government. Also, a larger share of retail shareholders decentralizes the power of shareholders, which induces a higher level of risks and, subsequently, a higher loan rate. In the present study, besides examining the mentioned hypothesis, we test whether the share of the institutional stockholder can impact the financing rate or not. We also examine whether the relationship between the loan rate and share of the institutions in the borrowing firms is not linear and after a threshold for the percentage of ownership the relationship becomes stronger.

Characteristics of the bank–firm relationship is the next important factor that can explain the difference in borrowing rates across firms. Previous studies have shown that the quality and background of the bank–firm relationship can affect the financing rate. For measuring the bank–firms variables, we extract data from ‘Notes to the financial statements’. The data published in ‘Notes to the financial statements’ contain the firms’ outstanding loan data in detail, including their lender banks. By collecting and processing this data set, the bank–firm relationship can be measured using some indexes, such as the number of creditors banks (bnum), the type of creditor banks in terms of creditor ownership (such as private, privatized or public bank) (bpri) and the creditor’s financial health indicators⁵ (like NPL [bNPL] and the capital-to-assets ratio [bEqu]). A hypothesis examined in this article is that the firms that borrow from various banks, instead of relying on a single main bank, bear a higher financial cost. Another hypothesis is that the imposed loan rates for firms that borrow from healthier banks (fewer NPLs and high capital-to-assets ratio) are lower.

To identify the effects of the ownership structure and bank–firm characteristics, it is required to control the effects of some financial ratios and indexes that introduce a firm’s current and potential risks. For this reason, in the light of research in the literature, other variables that are taken into consideration include the return on assets (ROA) as the profitability index, the current-capital-to-assets ratio (wk) and standard deviation of cash flow (sdch) as indicators of liquidity risk, the size (size) and fixed-assets-to-total-assets ratio (fasset) as indicators of the power of providing collateral and the beta coefficient (beta) as a measure of systematic risk, as well as the production growth in two-digit International Standard Industrial Classification (ISIC) codes (igrow). Moreover, since our main dependent variable is driven by the cost of debt, to identify the effect of the targeted variables, we should control for debt maturity and the combination of debt. Following Lin et al. (2011) and Roberts and Yuan (2010), we control for loan maturity by introducing the short-term-loan-to-total-loan ratio (dur) as an explanatory variable. Also, because each debt item has its own rate and the debt combination affects the cost of debt, we should control for the effect of the debt combination, and thus we consider the effect of the ratio of total loan to total debt (loanr). Since the better financial status and higher profitability of potential borrowers lower the future risk, a borrower’s financing rate is negatively correlated with its financial-soundness indicators. Moreover, larger companies and companies with a higher fixed-assets-to-total-assets ratio pose a lower default risk to creditors, and their rate is expected to be at a relatively lower level. The beta coefficient, too, is treated as a systematic risk indicator and can be taken into account for measuring the effectiveness of risk stemming from stock price fluctuations. Additionally, increasing the output growth in an industry also reduces the risk of stock market companies. The definitions of all applied variables are provided in Appendix 1.

The means and standard deviations of the calculated borrowing rates and the effective factors for the studied periods are reported in Table 1. Because the inflation rates in the studied periods are often higher than the nominal loan rates, the average real borrowing rate is negative. In the Iranian capital market, banks, the government and large holding companies are the major institutional owners. Thus, we measure the institutional ownership in the companies by stacking up the proportion of shares held by these institutions. According to Table 1, on average, 14.6 per cent of the sample companies’ shares directly belong to the institutions. Also, these institutions, directly and indirectly, own about 18.58 per cent of the sample companies. Banks, the government and holding companies, directly and indirectly, own 5.56 per cent, 3.61 per cent and 9.46 per cent of the companies’ shares, respectively. Also, the direct ownership proportions of these institutions are 3.64 per cent, 3 per cent and 7.95 per cent, respectively.

Table 1.

The Statistics of Affective Variables in the Period 2006–2016

Variables	Symbol	Obs.	Mean	Std. Dev.
Real borrowing rate (%)	r_i,t	2,894	−2.48	15.78
Nominal borrowing rate (%)	i_i,t	2,894	15.85	10.39
Institutional shareholder (direct + indirect) (%)	sh_inst_i,t	3,787	14.62	24.56
Institutional shareholder (direct) (%)	shd_inst_i,t	3,787	18.58	25.41
Banks share (direct + indirect) (%)	sh_bnk_i,t	3,787	5.56	13.55
Government sector share (direct + indirect) (%)	sh_gov_i,t	3,787	3.61	12.21
Holding companies’ share (direct + indirect) (%)	sh_hld_i,t	3,787	9.45	20.49
Banks’ share (direct) (%)	shd_bnk_i,t	3,787	3.64	12.17
Holding companies’ share (direct) (%)	shd_hld_i,t	3,787	7.95	19.45
Government sector share (direct) (%)	shd_gov_i,t	3,787	3.03	11.90
Retail shareholders (%)	sh_oth_i,t	3,787	26.09	21.50
Creditor banks’ NPL (%)	bNPL_i,t	1,339	13.80	5.44
Creditor banks’ capital ratio (%)	bEqu_i,t	1,339	9.23	6.89
Share of private banks in firm’s loans (%)	bpri_i,t	1,339	14.62	25.04
Number of banks (creditors)	bnum_i,t	1,884	3.18	2.41
Length of credit relationship	blon_i	253	3.20	1.12
ROA (%)	prof_i,t	3,445	12.33	440.3
Short-term loan to total loan	dur^i,t	3,447	0.83	0.27
Loan to debt	loanr_i,t	3,804	0.35	0.26
Working capital to assets	wk_i,t	4,362	0.06	0.40
Beta coefficient	beta_i,t	4,613	1.12	28.30
Fixed asset to total assets	fasset^i,t	3,937	0.38	0.23
Net cash flow standard deviation (billion rials)	sdch_i	3,911	740.4
Industrial growth (%)	igrow_j,t	3,460	0.04	0.18
Debt to total assets	debtr_i,t	3,804	0.65	0.42
Logarithm of total assets value	size_i,t	3,804	13.84	1.67

Source: The author.

Model Specification

According to previous studies, three categories of variables can be considered as explanatory variables for the borrowing rate: ownership structure variables, variables related to bank–firm relationship and other variables controlling the borrower’s financial and operational risks. In the model specification process, using the borrowing rate (in the level form) or its changes as a dependent variable results in a considerable difference in the interpretation of the results and also changes the model variables. A lag of the dependent variable should be introduced to model as the explanatory variables if we seek to study the borrowing rate changes while to evaluate the varied levels of borrowing rate, the model does not contain the lag of the dependent variables. Thus, the models are expressed as follows:

r_{i, t} = α + β r_{i, t - 1} + γ B_{i, t - 1} + θ W_{i, t - 1} + μ C_{i, t - 1} + δ_{j} + ω_{t} + ε_{i t}

(1)

r_{i, t} = α + γ B_{i, t - 1} + θ W_{i, t - 1} + μ C_{i, t - 1} + δ_{j} + ω_{t} + ε_{i t}

(2)

B_{i, t} {=[bNPL}_{i, t} {bEqu}_{i, t} {bpri}_{i, t} {bnum}_{i, t} {blon}_{i, t}]

W_{i, t} = [s h_b n k_{i, t} s h_h l d_{i, t} s h_g o v_{i, t} s h_o t h_{i, t}]

C_{i, t} = [d u r_{i, t} l o a n r_{i, t} d e b t r_{i, t} f a s s e t_{i, t} s i z e_{i, t} w k_{i, t} p r o f_{i, t} s d c h_{i} b e t a_{i, t} i g r o w_{i, t}]

where r_it indicates the real borrowing rate of manufacturing company i in period t, B_it₋₁ is the vector of variables related to bank relation of company i in time t − 1, W_it₋₁ is the vector of variables of the ownership structure of firm i in period t and C_it₋₁ indicates the control variables of firm i in period t − 1. Moreover, δ_j and ω_t consider the fixed effects at the industry level and the time fixed effects. All the explanatory variables have been modelled with a lag in order to reduce the endogeneity problem.⁶

The vector B includes the bank health indicators of the creditors, such as the capital ratio (bEqu), the NPL ratio, the share of private banks in firm’s loans (bpri), the number of firm’s lender banks (bnum) and the length of relationship between bank and firm (bloan). The vector W_i indicates the ownership structure features, which include the variables of shareholding by banks (sh_bank), shareholding by the government (sh_gove), shareholding by holding companies (sh_hld) and the retail stockholders’ share (sh_oth). The C vector, too, consists of control variables at the firm level, such as the ability to provide collateral (fasset), size (size), working capital ratio (wk), industry growth rate (igrow), beta coefficient (beta), debt ratio (debtr), loan ratio (loan r), short-term loan ratio (dur) and ROA (prof).

To remove the effects of the different scales of the variables, the variables are transformed to the z-score form (i.e., subtracting the mean and dividing by standard deviation for each variable).

The main focus of this article is on estimating Equation (1) that includes the first lag of the dependent variable as an explanatory variable. Nevertheless, the estimation results of Equation (2) are presented in Online Appendix⁷ (which confirm the main findings of the estimations of Equation [1]). Since the panel data specification of Equation (1) contains the lag of the dependent variable, the consistent estimations for the coefficients require using the generalized method of moments (GMM) dynamic panel data method developed by Blundell and Bond (1998). In this article, we use the Blundell and Bond (1998) method that uses instrument variables as proxy for the first lag of the dependent variable. As suggested in Blundell and Bond (1998) and Arellano and Bond (1991), the instruments are chosen among both levels and differences of deeper lags of the dependent variable and lags of explanatory variables. In addition, Sargan’s J test is used for testing over-identifying restrictions.

IV. Major Findings

The regression analysis framework allows examining the effects of potentially effective factors on the borrowing rate. In this section, to provide robust results, besides the main estimation results, the sensitivity of the estimation results is checked using different dimensions, as well as different definitions of variables.

The estimation results of Equation (1) are reported in Table 2 (column [1]). The first lag of the dependent variable has a significantly positive effect. Thus, the lag of real borrowing rates has a significant explanatory power on change in the explanatory power. Because of considering the effect of the lagged dependent variables, the coefficients of other explanatory variables measure the influence on changes in the borrowing rate (not level). The estimated coefficients of banking variables indicate that a higher level of NPL ratio of the firm’s creditor positively influences the financing rate of the firm, and a higher share of private banks in a firm’s loans is correlated with lower rates. The estimated model does not support any significant effect of the number of creditor banks and the capital ratio of the creditors.

Among the ownership structure variables, shareholding by banks (indirectly and directly) in a firm has a negative and significant effect on the borrowing rate, whereas shareholding by the government and large holding companies does not have a significant impact on the borrowing rate. Furthermore, a larger share of retail stockholders increases the loan rates. Hence, firms with banks as stockholders and lower shares for retail stockholders have lower borrowing rates. The most significant factor among the control variables is the standard deviation of cash flow, which has a significant and positive impact on the rate, which means more fluctuations in a firm’s cash flow through increasing the firm’s future risk lead to an increase in the financing cost. Since the financial structure of large companies is significantly different from that of SMEs, the weighted GMM estimation method is applied, and the results are presented in column (2) of Table 2. In this regard, to highlight the observations for larger firms, the weight in the method is the logarithm of the total value of assets. The results obtained through the weighted GMM estimation show that considering a firm’s size does not have a significant effect on the estimated coefficients and that the coefficients are numerically close to the estimated coefficient in column (1). Only the effect of shareholding by banks is weaker in the case of the weighted method. In other words, the decrease in borrowing rate stemming from an increase in the shares held by banks is higher in smaller firms than in larger ones. Cash flow fluctuations, too, have a stronger effect on large firms. To ensure the results are robust, industry fixed effects in column (1) are replaced by firm fixed effects in column (3). Comparison of the estimated coefficients in these columns shows that the main results remain unchanged.

Table 2.
The Effects of Bank-Specific and Ownership Variables on the Real Borrowing Rates

Dep. Var.: r_t (1) (2) (3)

r_i _,t−1 0.42* 0.43* 0.23

(4.28) (4.32) (2.38)

Bank–firm var. bNPL_i,t−1 0.11 0.12** 0.12*

(2.23) (2.47) (1.89)

bEqu_i,t−1 −0.065 −0.078 −0.04

(−0.86) (−1.04) (−0.52)

bpri_i,t−1 −0.16 −0.15 0.037

(−2.32) (−2.37) (0.47)

bnum_i,t−1 −0.083 −0.084 −0.06

(−1.00) (−1.03) (−0.23)

blon_i 0.21* 0.21* 0.135

(2.81) (2.79) (1.11)

Ownership var. sh_bnk_i,t−1 −0.095 −0.089 −0.08**

(−2.15) (−2.02) (−2.01)

sh_hld_i,t−1 0.066 0.046 0.059

(1.39) (0.96) (0.55)

sh_gov_i,t−1 −0.11 −0.10 −0.26*

(−1.46) (−1.40) (−1.66)

sh_oth_i,t−1 0.053 0.054 −0.06

(2.07) (2.11) (−1.39)

Control var. dur_i,t−1 0.038 0.054 0.24***

(0.47) (0.68) (2.82)

loanr_i,t−1 0.0091 0.0066 −0.015

(0.15) (0.11) (−0.2)

debtr_i,t−1 −0.14 −0.063 0.35

(−0.58) (−0.25) (1.26)

fasset_i,t−1 −0.20 −0.16 −0.03

(−1.29) (−1.02) (−0.18)

size_i,t−1 −0.0097 −0.037 0.29

(−0.077) (−0.30) (1.08)

wk_i,t−1 −0.28 −0.20 0.21

(−1.10) (−0.77) (0.84)

prof_i,t−1 −8.08 −12.0 −11.2

(−0.22) (−0.33) (−0.3)

sdch_i 0.36* 0.41** 0.01

(1.75) (2.17) (0.02)

beta_i,t−1 −0.097 −0.11 −0.01

(−0.49) (−0.55) (0.05)

igrow_i,t−1 −0.025 −0.023 0.003

(−0.66) (−0.62) (0.11)

Constant −0.74 −0.76 −0.77

(−0.74) (−0.73) (1.25)

Sargan J test (prob.) 76.9 (0.48) 76.6 (0.49) 79(0.4)

Time fixed effects Yes Yes Yes

Firm fixed effects No No Yes

Industry fixed effects Yes Yes No

Number of firms 149 149 149

Observations 692 692 692

Dep. Var.: r_t	(1)	(2)	(3)
	r_i _,t−1	0.42***	0.43***	0.23**
(4.28)	(4.32)	(2.38)
Bank–firm var.	bNPL_i,t−1	0.11**	0.12**	0.12*
	(2.23)	(2.47)	(1.89)
bEqu_i,t−1	−0.065	−0.078	−0.04
	(−0.86)	(−1.04)	(−0.52)
bpri_i,t−1	−0.16**	−0.15**	0.037
	(−2.32)	(−2.37)	(0.47)
bnum_i,t−1	−0.083	−0.084	−0.06
	(−1.00)	(−1.03)	(−0.23)
blon_i	0.21***	0.21***	0.135
	(2.81)	(2.79)	(1.11)
Ownership var.	sh_bnk_i,t−1	−0.095**	−0.089**	−0.08**
	(−2.15)	(−2.02)	(−2.01)
sh_hld_i,t−1	0.066	0.046	0.059
	(1.39)	(0.96)	(0.55)
sh_gov_i,t−1	−0.11	−0.10	−0.26*
	(−1.46)	(−1.40)	(−1.66)
sh_oth_i,t−1	0.053**	0.054**	−0.06
	(2.07)	(2.11)	(−1.39)
Control var.	dur_i,t−1	0.038	0.054	0.24***
	(0.47)	(0.68)	(2.82)
loanr_i,t−1	0.0091	0.0066	−0.015
	(0.15)	(0.11)	(−0.2)
debtr_i,t−1	−0.14	−0.063	0.35
	(−0.58)	(−0.25)	(1.26)
fasset_i,t−1	−0.20	−0.16	−0.03
	(−1.29)	(−1.02)	(−0.18)
size_i,t−1	−0.0097	−0.037	0.29
	(−0.077)	(−0.30)	(1.08)
wk_i,t−1	−0.28	−0.20	0.21
	(−1.10)	(−0.77)	(0.84)
prof_i,t−1	−8.08	−12.0	−11.2
	(−0.22)	(−0.33)	(−0.3)
sdch_i	0.36*	0.41**	0.01
	(1.75)	(2.17)	(0.02)
beta_i,t−1	−0.097	−0.11	−0.01
	(−0.49)	(−0.55)	(0.05)
igrow_i,t−1	−0.025	−0.023	0.003
	(−0.66)	(−0.62)	(0.11)
Constant	−0.74	−0.76	−0.77
(−0.74)	(−0.73)	(1.25)
Sargan J test (prob.)	76.9 (0.48)	76.6 (0.49)	79(0.4)
Time fixed effects	Yes	Yes	Yes
Firm fixed effects	No	No	Yes
Industry fixed effects	Yes	Yes	No
Number of firms	149	149	149
Observations	692	692	692

Source: The author.

Notes: *, ** and *** Indicate significance at the 10%, 5% and 1% levels, respectively.

Table 2 reports the estimation results of Equation (1). In column (1), we use the generalized method of moments (GMM) method suggested by Blundell and Bond (1998). In column (2), Equation (1) is estimated using the weighted dynamic GMM estimation in which the size (asset value logarithm) is used as a weighted variable. The estimation of column (3) is similar to the estimation of column (1), using firm fixed effects instead of industry fixed effects.

The ownership variables in the results reported in Table 2 include both direct and indirect ownership. In Table 3, we separate direct and indirect ownership into two variables and examine their effects through different estimations. Except the definitions of the ownership variables, all other options of the estimations reported in Table 3 are similar to the estimations of Table 2. The results presented in columns (1) and (2) of Table 3 show the direct shareholding effects, and columns (3) and (4) show the indirect shareholding effects. The results show that the negative effect of shareholding by banks on the borrowing rate is significant only if the banks own firms directly, and indirect ownership linkages between the firms and banks do not decrease the borrowing rate. Other variables have similar effects to those reported in Table 2, and the variables such as the NPL ratio, private banks’ shares, length of the bank–firm relationship, retail shareholders’ share and cash flow fluctuations have a significant effect on these estimates.

Nominal Borrowing Rate

To obtain an accurate picture of the credit rate determinants in an economy with high rates of inflation, we use the real borrowing rate series in the main model’s estimations (reported in Tables 2 and 3). The question arises whether the influence of factors on the nominal borrowing rate and that on the real rate are different from each other. Table 4 reports the estimations in which the dependent variable is the nominal borrowing rate. The setting of the estimation in columns (1) and (2) in Table 4 is the same as that of the estimation in columns (1) and (2) in Table 2 (estimation method in columns [1] and [2]is GMM and weighted GMM, respectively). Comparison of the results with the main specification results (estimation column [1] in Table 2) shows that the effects of NPL ratio of creditor banks and shareholding by the banks remain significant, without any sign change. Thus, these variables impact both the nominal and real borrowing rates in the same way. Moreover, both variables of the firm’s ability to provide collateral and profitability of the firm have significant effects only on the nominal rate. Also, the results do not confirm a significant linear relationship between the stockholding of large holding companies and the government and both the nominal and real borrowing rates.

Table 3.

Impact of Direct and Indirect Institutional Ownership on the Real Borrowing Rate

Dep. Var.: r_t		(1)	(2)	(3)	(4)
	r_t ₋₁	0.42***	0.41***	0.46***	0.46***
		(4.34)	(4.16)	(4.65)	(4.54)
Bank–firm var.	bNPL_i,t−1	0.097**	0.088*	0.11**	0.12**
		(2.04)	(1.81)	(2.23)	(2.46)
	bEqu_i,t−1	−0.052	−0.072	−0.057	−0.070
		(−0.69)	(−0.90)	(−0.74)	(−0.92)
	bpri_i,t−1	−0.15**	−0.11*	−0.18***	−0.18***
		(−2.31)	(−1.69)	(−2.72)	(−2.69)
	bnum_i,t−1	−0.058	−0.12	−0.080	−0.085
		(−0.70)	(−1.31)	(−0.94)	(−1.02)
	Blon	0.20***	0.18**	0.20***	0.21***
		(2.67)	(2.43)	(2.72)	(2.78)
Ownership var.	shd_bnk_i,t−1	−0.074*	−0.082*
		(−1.68)	(−1.89)
	shi_bnk_i,t−1			−0.049	−0.052
				(−1.21)	(−1.27)
	shd_hld_i,t−1	0.058	0.062
		(1.33)	(1.27)
	shi_hld_i,t−1			−0.022	−0.020
				(−0.33)	(−0.31)
	shd_gov_i,t−1	−0.13	−0.025
		(−1.47)	(−0.24)
	shi_gov_i,t−1			−0.028	−0.026
				(−0.61)	(−0.57)
	sh_oth_i,t−1	0.050**	0.054**	0.053*	0.056**
		(1.97)	(2.13)	(1.95)	(2.08)
Sargan J test (prob.)		79.2 (0.41)	89.2 (0.1)	78.2 (0.44)	77.2 (0.47)
Control var. and cons.		Yes	Yes	Yes	Yes
Time fixed effects		Yes	Yes	Yes	Yes
Industry fixed effects		Yes	Yes	Yes	Yes
Observations		692	692	692	692

Source: The author.

Notes: *, ** and *** Indicate significance at the 10%, 5% and 1% levels, respectively.

The reported estimations capture the effects of direct and indirect ownership variables separately. In the estimation columns (1) and (2), the direct ownership of the mentioned institutions is investigated using the GMM and weighted GMM methods, respectively. In the estimation columns (3) and (4), we use the GMM and weighted GMM methods to examine the effect of the indirect variables, respectively.

Table 4.

The Effect of Ownership and Bank-Specific Variables on the Nominal Borrowing Rate

Dep. Var.: r_t		(1)	(2)
	r_i _,t−1	0.36***	0.37***
	r_i _,t−1	(2.70)	(2.71)
Bank–firm var.	bNPL_i,t−1	0.34**	0.36**
		(2.15)	(2.26)
	bEqu_i,t−1	−0.11	−0.14
		(−0.56)	(−0.71)
	bpri_i,t−1	−0.22	−0.23
		(−1.24)	(−1.28)
	bnum_i,t−1	−0.095	−0.032
		(−0.43)	(−0.14)
	Blon	0.0053	−0.044
		(0.024)	(−0.20)
Ownership var.	sh_bnk_i,t−1	−0.21**	−0.19*
		(−2.01)	(−1.86)
	sh_hld_i,t−1	0.14	0.10
		(0.84)	(0.65)
	sh_gov_i,t−1	−0.30	−0.33
		(−1.08)	(−1.23)
	sh_oth_i,t−1	−0.024	−0.040
		(−0.24)	(−0.39)
		(0.0051)	(0.19)
Constant		0.0	7.79**
Constant		(0.0)	(2.51)
Sargan J test (prob.)		51.5 (0.34)	55.0 (0.22)
Control var. and cons.		Yes	Yes
Time fixed effects		Yes	Yes
Industry fixed effects		Yes	Yes
Observations		700	700

Source: The author.

Notes: *, ** and *** Indicate significance at the 10%, 5% and 1% levels, respectively.

Table 4 reports the estimation results of Equation (1) with the nominal borrowing rate as the dependent variable. The control variables that are not reported in the table are similar to the main specifications in Table 2. In the estimation columns (1) and (2), we use the GMM and weighted GMM methods, respectively.

A Threshold for the Institutional Ownership Impact

The impact of institutional shareholders on a firm’s performance (and consequently on its borrowing rate) depends on the different investment strategies that are adopted by the shareholders. Institutional shareholders that adopt the buy-and-hold strategy and pursue long-term and managerial goals behave differently from the investors who invest based on short-term vision and seek short-term returns. Therefore, a hypothesis can be proposed that the presence of institutional shareholders has a real effect on a firm’s performance and reduces its financing cost only if the institutional shareholders can contribute to the firm’s decision-making, for which the proportion of shares held by the institutions should be above a certain threshold. To examine this hypothesis, dummy variables (equal to 1 if institutions’ shares are above a threshold, and 0 otherwise) are introduced into the model (Equation [1]). For assessing the sensitivity of the threshold, we consider four different values of 5 per cent, 10 per cent, 20 per cent and 30 per cent. In the estimations of columns (1)–(5) in Table 5, the threshold levels are defined as 0 per cent, 5 per cent, 10 per cent, 20 per cent and 30 per cent, respectively.

The estimation results show that the relationship between shareholding by banks and a firm’s loan rate has two threshold levels in which the relationship has breaks. First, on average, companies with any percentage of their shares being owned by banks (more than 0%) have a lower borrowing rate. Second, companies with more than 10 per cent of their shares being owned by banks pay a lower loan rate than others. Moreover, the estimated coefficient of the bank shareholding dummy variable is higher in estimation column (3) than in estimation column (1). This means that the companies having at least one bank in their shareholder composition have lower borrowing rates than other companies, and among the companies with shares owned by banks, those with more than 10 per cent of their shares being held by banks have lower borrowing rates.

According to the results of Tables 2–4, ownership of a firm’s shares by large holding companies and the government does not have a significant linear relationship with changes in the corporate’s borrowing rates. However, when we convert the government ownership variable into a dummy variable, as shown in estimation column (4) of Table 5, this relationship becomes significant. Thus, the companies with more than 20 per cent of their shares being owned by the government have a lower financial cost than those with under 20 per cent of their shares being owned by the government. The coefficient of government ownership dummies in estimation columns (4) and (5) are, respectively, −0.75 and −1.10. Comparison of the estimated coefficients shows that the companies with more than 30 per cent of their shares being owned by the government have lower rates than those with 20–30 per cent of their shares being owned by the government. This method (converting the ownership variables into dummy variables) is also applied for the direct ownership variables, and the results are presented in the Online Appendix.

Table 5.

Threshold Effect of the Ownership Variables on Changes in the Real Borrowing Rate

Dep. Var.: r_t		(1)	(2)	(3)	(4)	(5)
	r_i _,t−1	0.32***	0.091	0.059	0.19	0.25
		(2.59)	(0.52)	(0.32)	(1.14)	(1.44)
Bank–firm var.	bNPL_i,t−1	0.15**	0.26***	0.24***	0.30***	0.34***
		(2.01)	(2.92)	(2.69)	(3.19)	(3.13)
	bEqu_i,t−1	−0.14	−0.24**	−0.27**	−0.20*	−0.20
		(−1.55)	(−2.14)	(−2.42)	(−1.70)	(−1.55)
	bpri_i,t−1	−0.088	−0.24**	−0.23**	−0.24**	−0.27**
		(−1.03)	(−2.07)	(−2.11)	(−2.22)	(−2.35)
	bnum_i,t−1	−0.16	0.10	0.11	0.10	0.14
		(−1.20)	(0.71)	(0.79)	(0.69)	(0.88)
	blon_i,t−1	−0.012	0.077	0.073	0.11	0.097
		(−0.11)	(0.85)	(0.85)	(1.17)	(0.94)
Ownership var.	sh_bnk_i,t−1 ≠ 0	−0.27**
		(−2.45)
	sh_hld_i,t−1 ≠ 0	0.18
		(1.44)
	sh_gov_i,t−1 ≠ 0	0.28
		(1.51)
	sh_oth_i,t−1 ≠ 0	0.0018
		(1.32)
	sh_bnk_i,t−1 > 5%		−0.18
			(−1.20)
	sh_hld_i,t−1 > 5%		0.31*
			(1.73)
	sh_gov_i,t−1 > 5%		−0.49
			(−1.41)
	sh_oth_i,t−1 > 5%		−0.12
			(−0.58)
	sh_bnk_i,t−1 > 10%			−0.39**
				(−2.35)
	sh_hld_i,t−1 > 10%			0.32
				(1.51)
	sh_gov_i,t−1 > 10%			−0.33
				(−0.91)
	sh_oth_i,t−1 > 10%			0.0055
				(0.035)
	sh_bnk_i,t−1 > 20%				−0.22
					(−1.09)
	sh_hld_i,t−1 > 20%				0.26
					(1.06)
	sh_gov_i,t−1 > 20%				−0.75*
					(−1.77)
	sh_oth_i,t−1 > 20%				0.092
					(0.68)
	sh_bnk_i,t−1 > 30%					−0.31
						(−0.89)
	sh_hld_i,t−1 > 30%					0.48
						(1.20)
	sh_gov_i,t−1 > 30%					−1.10**
						(−2.13)
	sh_oth_i,t−1 > 30%					0.13
						(0.93)
Sargan J test (prob.)		46.1 (0.2)	29.3 (0.45)	29.6 (0.43)	28.8 (0.47)	23.4 (0.76)
Control var. and cons.		Yes	Yes	Yes	Yes	Yes
Time fixed effects		Yes	Yes	Yes	Yes	Yes
Industry fixed effects		Yes	Yes	Yes	Yes	Yes
Observations		692	732	732	732	732

Source: The author.

Notes: *, ** and *** Indicate significance at the 10%, 5% and 1% levels, respectively.

The control variables in these estimations are similar to the main specifications in Table 2.

In the estimation column (1), the stockholding of banks is introduced to the model in a dummy variable format (the firm, directly or indirectly owned by the institutions equal 1, otherwise 0). In the estimations (2), the dummy variable is defined as follows: the variable equals one if the proportion of the firm is owned by the institutions is more than 5 percent, and equals zero otherwise In the estimations of (3), (4), and (5), the threshold of the dummy variable increase to 10, 20, and 30 percent, respectively.

Effective Factors on Firms with Different Ownership Structures

Based on the estimation results reported in the previous sections, the ownership structure variables have a significant effect on a firm’s borrowing rate. Moreover, it can be examined how the existence of institutional shareholders in a firm can affect the strength of the relationships between the model variables. In other words, the direction and strength of the relationship between the loan rate and explanatory factors can be different across firms with varying ownership structures. To examine this, we divide the sample of companies into two groups based on their having or not having at least one bank in their shareholder composition, and then we estimate the effect of the explanatory variables on their borrowing rate. Estimation column (1) of Table 6 includes the firms that do not have a bank in the shareholder composition, and estimation column (2) includes the firms with at least one bank as a shareholder. Also, estimation column (3) includes firms that have more than 5 per cent of their shares being owned by banks.⁸ Comparison of estimation columns (1)–(3) shows that among bank–firm variables, the length of the bank–firm relationship (blon) has a significant and negative impact only on the firms with more than 5 per cent of their shares being owned by banks. The effects of the other bank–firm variables are the same across all the sub-samples. Also, among control variables, cash flow fluctuations only affect the borrowing rate of the firms in which no bank is a shareholder.

Table 6.

Effect of Bank–Firm Variables on Borrowing Rates in the Sub-samples Divided by the Banks’ Shareholding

Dep. Var.: r_t		(1)	(2)	(3)
	r_i _,t−1	0.43***	0.50***	0.19
		(3.45)	(2.93)	(1.27)
Bank–firm var.	bNPL_i,t−1	0.0049	−0.065	0.084
		(0.084)	(−0.55)	(1.10)
	bEqu_i,t−1	−0.085	0.21	−0.036
		(−0.92)	(1.25)	(−0.36)
	bpri_i,t−1	−0.026	0.0060	0.031
		(−0.36)	(0.046)	(0.33)
	bnum_i,t−1	0.014	0.16	0.11
		(0.16)	(0.94)	(1.01)
	blon_i	−0.069	0.41	−0.36**
		(−0.77)	(1.55)	(−2.20)
Sargan J test (prob.)		73.1 (0.05)	21.8 (0.11)	45.8 (0.44)
Control var. and cons.		Yes	Yes	Yes
Time fixed effects		Yes	Yes	Yes
Industry fixed effects		Yes	Yes	Yes
Observations		312	420	174

Source: The author.

Notes: *, ** and *** Indicate significance at 10%, 5% and 1% levels, respectively.

The control variables in these estimations are similar to the main specifications in Table 2. Estimation column (1) shows the effect of the bank–firm variables and control variables for the firms that are not owned by a bank. The sample used in estimation column (2) includes the firms with at least one bank as a shareholder. The sample used in estimation column (3) includes the firms with more than 5 per cent of their stocks being owned by banks.

In Table 7, the firms of the sample are divided based on whether they have holding company and state ownership or not (the analysis method is the same as in Table 6, for which the benchmark was having or not having banks in the shareholder composition). Estimation column (1) shows the effect of bank-specific and control variables among firms that do not have any holding company as a shareholder. Estimation column (2) includes the firms with part of their shares being owned by holding companies. Comparison of these two estimations suggests that the influential factor on the borrowing rate among the firms owned by holding companies is the NPL of the creditor, and the influential factor on the borrowing rate among the firms that do not have an ownership relationship with any holding company is the number of creditor banks. Also, the sample used in estimation column (3) includes non-state-owned firms, and the sample in column (4) includes state-owned firms. Based on the comparison of these two estimations, among the non-state-owned firms, the borrowing rate is related to the capital of the creditor banks, loan-to-debt ratio and fixed assets ratio (bonding power). However, none of the considered variables affects the borrowing rate of firms with a state shareholder.

Table 7.

Effect of bank- Fim variables on Borrowing Rates in the subsamples that divided by the by the holding and government shareholding shareholding

Dep. Var.: r_t		(1)	(2)	(3)	(4)
	r_i _,t−1	0.26**	0.37**	0.34***	0.16
		(2.09)	(2.20)	(3.10)	(0.75)
Bank–firm var.	bNPL_i,t−1	−0.085	0.23*	0.054	0.016
		(−1.09)	(1.91)	(0.91)	(0.12)
	bEqu_i,t−1	−0.065	−0.057	−0.15**	−0.14
		(−0.86)	(−0.41)	(−2.10)	(−1.16)
	bpri_i,t−1	−0.030	−0.076	0.030	0.085
		(−0.35)	(−0.70)	(0.35)	(0.53)
	bnum_i,t−1	−0.28***	−0.14	−0.071	−0.067
		(−2.90)	(−0.97)	(−0.79)	(−0.58)
	blon_i	0.083	0.27	0.16	0.13
		(0.68)	(1.57)	(1.44)	(0.64)
Sargan J test (prob.)		46.5 (0.06)	29.3 (0.08)	71.7 (0.03)	34 (0.11)
Control var. and cons.		Yes	Yes	Yes	Yes
Time fixed effects		Yes	Yes	Yes	Yes
Industry fixed effects		Yes	Yes	Yes	Yes
Observations		334	398	585	147

Source: The author.

Notes: *, ** and *** Indicate significance at 10%, 5% and 1% levels, respectively.

The sample of estimation (1) includes firms that do not have relationship with the holding companies as shareholders. Estimation (2) includes firms that partially owned by the holdings companies. Estimation (3) represent the relationship among firms that government did not own them and sample of estimation (4), in other hands, includes firms that partially owned by the government.

V. Conclusion

Borrowing rates vary across firms. The root of this variance lies in different firms’ risk exposures. However, according to empirical observations and studies, the difference in the financial and operational risk indexes, mainly extracted from the financial statement, cannot completely explain the difference in the rates. Due to credit market imperfections, the factors such as characteristics of the bank–firm relationship and the ownership structure of firms can mitigate information asymmetry and affect the firm’s borrowing rate. This article investigated the effect of shareholding by influential institutions, such as banks, the public sector and large holdings, on the borrowing rate paid by firms. Also, we investigated the effect of creditors’ (banks’) soundness indicators (e.g., NPL and the capital adequacy ratio) and the characteristics of the bank–firm relationship (the number of creditor banks, length of the bank–firm relationship) on borrowing rates. According to the estimation results, an increase in the proportion of shares held by banks in a firm reduces the borrowing rate paid by the firm. Also, a reduction of the proportion shares held by retail shareholders (interpreted as a more centralized ownership structure) is effective in lowering the borrowing rate. Among the bank variables, the NPL ratio has a significant and sustained effect on the borrowing rate. Thus, firms that borrow from banks with a higher NPL ratio pay higher borrowing rates. Among other variables, an increase in cash flow fluctuations (as an indicator of liquidity risk) will increase the borrowing rate. According to the results, there are break points in the relationships between the borrowing rate and the proportion of shares held by banks and the government. Thus, the borrowing rate for companies that are partially owned by banks is significantly lower than that for others. Also, among firms partially owned by banks, those that have more than 10 per cent of their shares being owned by banks have lower borrowing rates. The results do not support the existence of a linear relationship between government shareholding and borrowing rates. However, the findings indicate that the loan rate of firms with more than 30 per cent of their shares being owned by the government is lower than that of firms with 20–30 per cent of their shares being owned by government. Also, firms with more than 20 per cent of their shares being owned by the government have a lower rate than other state-owned firms. In summary, according to the estimation results, the existence of institutional shareholders in a firm’s shareholder composition and a centralized structure of ownership in the firm reduce its borrowing rates. Therefore, to mitigate information asymmetry, lenders assess the ownership structure of firms. Existence of powerful institutions in the shareholder composition of a borrower reduces the expected default risk and, consequently, its cost of fund.

The Iranian credit market is an undeveloped credit market where information asymmetry is more severe, and it also suffers from financial repression (legal restrictions on interest rates). In this circumstance, according to the results, firms without ownership relationship with powerful institutions (e.g., banks, the government and large holding companies), which statistically are SMEs, pay higher rates on their loans. The higher financial cost for SMEs can be a challenge for them in competing with their larger counterparts. Thus, under these conditions, government policies should support SMEs and firms without institutional owners to reduce their cost of funds.

Footnotes

Acknowledgements

The author would like to thank S. Ali Madanizadeh and Amineh Mahmoudzadeh for their comments that greatly improved this research.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author received no financial support for the research, authorship and/or publication of this article.

Appendix

Appendix 1.

The Variables’ Definitions

Variable	Symbol	Short Definition
Nominal rate (%)	i_i,t	From the dividing of the financial cost (a write-off of the profit and loss account) of firm i at time t, the average of loan total balance (from the balance sheet items) was obtained in periods t and t−1. Considering that the loan balance only represents an image of the company’s situation at the end of fiscal year t and does not consider changes during the year, the average loan balance of periods t and t−1 have been used.
Real rate (%)	r_i,t	The nominal rate of financing (ii,t) of firm i at time t of the index growth of the producer price in year t.
Institutional shareholder (%) (direct + indirect)	sh_inst_i,t	The percentage of firm is shares that, directly or indirectly, are owned by government and state corporations, banks and large holding companies in year t.
Institutional shareholder (%) (direct)	shd_inst_i,t	The percentage of firm is shares that are directly owned by the government and related institutions, banks and large holding companies in year t.
Banks’ share (%) (direct + indirect)	sh_bnk_i,t	The percentage of firm is shares that are directly and indirectly owned by banks in year t.
Government sector share (%) (direct + indirect)	sh_gov_i,t	The percentage of firm is shares that are directly and indirectly owned by the government and related institutions in year t.
Holding companies’ share (%) (direct + indirect)	sh_hld_i,t	The percentage of firm is shares that are directly and indirectly owned by large holding companies in year t. The large holding companies in the Iranian economy include the holding company related to Iran’s social security agent, Bonyad Mostazafan Holding, Astan Ghods Razavi Holding and the Execution of Imam Khomeini’s Order (EIKO).
Banks’ share (direct) (%)	shd_bnk_i,t	The percentage of firm is shares that are directly owned by banks in year t.
Holding companies’ share (direct) (%)	shd_gov_i,t	The percentage of firm is shares that are directly owned by large holding companies in year t.
Government sector share (direct) (%)	shd_hld_i,t	The percentage of firm is shares that are directly owned by the government and related institutions in year t.
Banks’ share (indirect) (%)	shi_bnk_i,t	The percentage of firm is shares that are indirectly owned by banks in year t.
Holding companies’ share (indirect) (%)	shi_hld_i,t	The percentage of firm is shares that are indirectly owned by large holding companies in year t.
Government sector share (indirect) (%)	shi_gov_i,t	The percentage of firm is shares that are indirectly owned by the government and related institutions in year t.
Retail shareholders (%)	sh_oth_i,t	Shareholding (ownership) percentage of minority shareholders from firm i at time t.
Creditor banks’ NPL (%)	bNPL_i,t	The weighted average of non-performing loans (NPLs) ratio of the creditor banks to firm i at time t, where the weight of each creditor bank is determined by the amount of loan given to firm i in year t.
Creditor banks’ capital ratio (%)	bEqu_i,t	The ratio of weighted average capital to the assets of the creditor banks to firm i at time t, where the weight of each creditor bank is determined by the amount of credit given to firm i in year t.
Private banks’ participation (%)	bpri_i,t	Percentage of loans taken by firm i at time t from private-owned banks.
Number of banks (creditors)	bnum_i,t	This indicator shows the number of loans a firm has received each year.
Length of credit relationship	blon_i	This variable indicates the number of years firm i continuously borrowed from its main bank (without interruption).
ROA (%)	prof_i,t	Return on total assets (net profit as percentage of total assets) in year t for company i.
Short-term loan share	dur_i,t	Loans with a short-term nature (less-than-1-year maturity) to the total value of the loans in year t for company i.
Loan to debt	loanr_i,t	Total loan to total debt value in year t for firm i.
Working capital to assets	wk_i,t	The difference of current assets and current liabilities (working capital) to total assets value in year t for company i.
Beta coefficient	beta_i,t	Beta is the intensity of the return changes on company is share to the market or index, including its share over the year t, using daily data.
Fixed assets to total assets	fasset_i,t	The value of fixed assets to the total value of assets for company i in year t.
Net cash flow standard deviation (million rials)	sdch_i	The standard deviation of net cash flow for company i in year t.
Industrial growth (%)	igrow_j,t	The true growth of production in industry j (two-digit ISIC code) related to company i in year t.
Debt to total assets	debtr_i,t	Ratio of debt to total assets value (leverage ratio) for company i in year t.
Logarithm of the total assets value	size_i,t	Natural logarithm of total assets value (in million rials) in company i in year t.

Source: The author.

Notes

References

Altman

E. I.

(1968). Financial ratios, discriminant analysis and the prediction of corporate bankruptcy. The Journal of Finance, 23(4), 589–609.

Altman

E. I.

, & Saunders

(1997). Credit risk measurement: Developments over the last 20 years. Journal of Banking & Finance, 21(11–12), 1721–1742.

Anderson

R. C.

, Mansi

S. A.

, & Reeb

D. M.

(2003). Founding family ownership and the agency cost of debt. Journal of Financial Economics, 68(2), 263–285.

Anderson

R. C.

, Mansi

S. A.

, & Reeb

D. M.

(2004). Board characteristics, accounting report integrity, and the cost of debt. Journal of Accounting and Economics, 37(3), 315–342.

Ang

J. S.

, Cole

R. A.

, & Lin

J. W.

(2000). Agency costs and ownership structure. The Journal of Finance, 55(1), 81–106.

Arellano

, & Bond

(1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The Review of Economic Studies, 58(2), 277–297.

Blundell

, & Bond

(1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87(1), 115–143.

Booth

J. R.

(1992). Contract costs, bank loans, and the cross-monitoring hypothesis. Journal of Financial Economics, 31, 25–41.

Borisova

, Fotak

, Holland

, & Megginson

W. L.

(2015). Government ownership and the cost of debt: Evidence from government investments in publicly traded firms. Journal of Financial Economics, 118(1), 168–191.

10.

Chava

, Livdan

, & Purnanandam

(2009). Do shareholder rights affect the cost of bank loans? Review of Financial Studies, 22(8), 2973–3004.

11.

Diamond

D. W.

(1989). Reputation acquisition in debt markets. Journal of Political Economy, 97(4), 828–862.

12.

Dimitras

A. I.

, Slowinski

, Susmaga

, & Zopounidis

(1999). Business failure prediction using rough sets. European Journal of Operational Research, 114(2), 263–280.

13.

Emel

A. B.

, Oral

, Reisman

, & Yolalan

(2003). A credit scoring approach for the commercial banking sector. Socio-Economic Planning Sciences, 37(2), 103–123.

14.

Gambacorta

(2008). How do banks set interest rates? European Economic Review, 52(5), 792–819.

15.

Graham

J. R.

, Li

, & Qiu

(2008). Corporate misreporting and bank loan contracting. Journal of Financial Economics, 89(1), 44–61.

16.

Houston

J. F.

, Jiang

, Lin

, & Ma

(2014). Political connections and the cost of bank loans. Journal of Accounting Research, 52(1), 193–243.

17.

Hubbard

R. G.

, Kuttner

K. N.

, & Palia

D. N.

(2002). Are there bank effects in borrowers’ costs of funds? Evidence from a matched sample of borrowers and banks. The Journal of Business, 75(4), 559–581.

18.

Jensen

M. C.

, & Meckling

W. H.

(1976). Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics, 3(4), 305–360.

19.

Kim

W. S.

, & Sorensen

E. H.

(1986). Evidence on the impact of the agency costs of debt on corporate debt policy. Journal of Financial and Quantitative Analysis, 21(2), 131–144.

20.

Lehmann

, & Neuberger

(2001). Do lending relationships matter? Evidence from bank survey data in Germany. Journal of Economic Behavior & Organization, 45(4), 339–359.

21.

Lin

, Ma

, Malatesta

, & Xuan

(2011). Ownership structure and the cost of corporate borrowing. Journal of Financial Economics, 100(1), 1–23.

22.

McConnell

J. J.

, & Servaes

(1990). Additional evidence on equity ownership and corporate value. Journal of Financial Economics, 27(2), 595–612.

23.

Moh’d

M. A.

, Perry

L. G.

, & Rimbey

J. N.

(1998). The impact of ownership structure on corporate debt policy: A time-series cross-sectional analysis. Financial Review, 33(3), 85–98.

24.

Morck

, Shleifer

, & Vishny

R. W.

(1988). Management ownership and market valuation: An empirical analysis. Journal of Financial Economics, 20, 293–315.

25.

Ohlson

J. A.

(1980). Financial ratios and the probabilistic prediction of bankruptcy. Journal of Accounting Research, 18(1), 109–131.

26.

Park

J. C.

, & Wu

(2009). Financial restatements, cost of debt and information spillover: Evidence from the secondary loan market. Journal of Business Finance & Accounting, 36, 1117–1147.

27.

Petersen

M. A.

, & Rajan

R. G.

(1994). The benefits of lending relationships: Evidence from small business data. The Journal of Finance, 49(1), 3–37.

28.

Pittman

J. A.

, & Fortin

(2004). Auditor choice and the cost of debt capital for newly public firms. Journal of Accounting and Economics, 37(1), 113–136.

29.

Roberts

, & Yuan

L. E.

(2010). Does institutional ownership affect the cost of bank borrowing? Journal of Economics and Business, 62(6), 604–626.

30.

Sanchez-Ballesta

J. P.

, & Garcia-Meca

(2011). Ownership structure and the cost of debt. European Accounting Review, 20(2), 389–416.

31.

Santos

J. A.

(2011). Bank corporate loan pricing following the subprime crisis. Review of Financial Studies, 24(6), 1916–1943.

32.

Shleifer

, & Vishny

R. W.

(1997). A survey of corporate governance. The Journal of Finance, 52(2), 737–783.

33.

World Bank (2001). Finance for growth: Policy choices in a volatile world. A World Bank Policy Research Report. World Bank.