ARE BANKS USING LOSS LOAN PROVISION TO SMOOTH INCOME? EMPIRICAL EVIDENCE FROM INDIAN COMMERCIAL BANKS PDF

Preview

Summary

ARE BANKS USING LOSS LOAN PROVISION TO SMOOTH INCOME? EMPIRICAL EVIDENCE FROM INDIAN COMMERCIAL BANKS Nikita Chopra Research Scholar University Business School Abstract: Purpose: The banking system being the backbone of the economy is found to indulge in income smoothing. The functioning of financia...

Description

Accelerat ing t he world's research.

ARE BANKS USING LOSS LOAN PROVISION TO SMOOTH INCOME? EMPIRICAL EVIDENCE FROM INDIAN COMMERCIAL BANKS Nikita Chopra Management of Financial Services: Creating Business Value and Sustainability

Cite this paper

Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers

Download a PDF Pack of t he best relat ed papers 

Use of loan loss provisions for capit al, earnings management and signalling by Aust ralian ban… Asokan Anandarajan

T he use of discret ionary loan loss provisions by Islamic banks and convent ional banks in t he Middle E… Hakim Ot hman AN EMPIRICAL ANALYSIS OF PROFIT DIST RIBUT ION MANAGEMENT DECISIONS IN MALAYSIAN ISLAMI… Nurul Syazwani Mohd Noor

ARE BANKS USING LOSS LOAN PROVISION TO SMOOTH INCOME? EMPIRICAL EVIDENCE FROM INDIAN COMMERCIAL BANKS Nikita Chopra Research Scholar University Business School Abstract: Purpose: The banking system being the backbone of the economy is found to indulge in income smoothing. The functioning of financial institutions is different from that of other industries which is why corporate governance of bank a crucial issue. If any other organisation falters, it has limited impact on only its stakeholders, but if a financial institution falters, the repercussions are on the entire economy. Thus the basic research question of the study is “do banks use loss loan provisions to indulge in earning management and what impact do these provisions have on earnings management?’ Prior Literature: Across different countries, it has been found that Loss Loan Provisions is not only used as means of earnings management in the long run but has an effect on the quality of loans and efficiency of the operations (Ma, 1988). As a result, it may lead to failure of the banking system. Research methodology: The data for a period of eleven years from 2007-2017 was collected for a sample of 37 commercial banks (26 PSUs and 11 Private Banks). The panel data has been subject to unit root test to check the stationarity of the data. To understand the long term impact the Toda & Yamamoto 1995 test has been applied with Johnsen Cointegration Test for checking the robustness of the equation. Findings: The estimates of Block ExogeneityWald test shows chi-square distribution with 5 degrees of freedom (Lag Length=8) and the corresponding Probability. In the above-given table, the probability of all the independent variables is more than 0.05. As a result, we accept the null hypothesis that Loss Loan Provisions has a significant positive relation with Earning before Taxes and Provisions, Non-Performing Assets, Total Loan and advance and Capital Adequacy Ratio of the banks. Thus, the banks use LLPs as a tool for earnings management Suggestions: Banks may also smooth the earnings in order to meet the regulatory requirement. While the problem of provision has clearly been expressed in studies conducted both in India and other countries, yet a forward-looking approach that focuses on the borrowing capacity and creditability of the loan payer rather than changing the provisioning policies is required. Original contribution: There have been several studies on banks regarding the credit risk but very few studies have been conducted on earnings management in India. With different time-period taken into consideration, the results also vary. KEYWORDS: Loan loss provisions, income smoothing, banking systems

INTRODUCTION The financial system is the backbone of a country. The smooth functioning of the banks depicts sound economic conditions. The functioning of financial institutions is different from that of other industries which is why corporate governance of bank a crucial issue. If any other organisation falters, it has limited impact on only its stakeholders, but if a financial institution falters, the repercussions are on the entire economy. Thus, regulations of banking industries and corporate governance complement each other. One issue of corporate governance in banks and financial institution is of earnings management. The banking institutions have been excluded from earning management from other type of organisations because the fundamental characteristics of this industry are different from the others (Peasnell, Pope and Young, 2000). There have been previous studies which suggest the use of loan loss provisions (LLPs) as a means for earnings management. These loan loss reserves have been created by banks in case of expected future losses on their loan. However, it has been found that bank managers have an incentive to manage earnings through these provisions. Despite the strict regulation and prudential norms, managers are still in a position to take advantage because, in certain decisions such as loan impairment, they have to exercise subjective judgement. Thus, it not only leads to earnings management of loss loan provisions but also has a signalling effect about the future of the banks (Ahmed, Takeda and Thomas, 1999, p. 2). But the degree of earnings management varies from country to country due to different regulatory policies. The foundation of the Indian Banking system has been credited to two committees, the Report of the Committee on Financial System (Reserve Bank of India, 1991) and the Report of the Committee on Banking Sector Reforms (Government of India, 1998), both presided by then RBI Governor, M Narasimham. These reports enhanced the functioning of commercial banks through the recommendation by policies such as a reduction in Cash Reserve Ratio, Statutory Liquidity Ratio and dismissal of administered interest rates. Narasimham Committee recommendations led to the modernisation of the financial sector, more autonomy to public sector bank, implementation of prudential norms and change in the ownership structure of banks in India. These reforms were adopted to create an effective, prolific and profitable financial sector. As per the reports published by Reserve Bank of India, the assets and liabilities continue to grow at 7.7%, there is a growing percentage of

banks’ delinquent loans, especially of public sector banks (PSBs) and a subsequent rise in provisions created for non-performing assets (NPAs) to lower the risk and ease the financial position. Thus the basic research question of the study is ‘do banks use loss loan provisions to indulge in earning management and what impact do these provisions have on earnings management?’. Furthermore the paper aims to contribute to the literature by the formulating an empirical model with the support of panel data techniques to understand the long-run impact of these provisions using Toda & Yamamoto 1995 test. REVIEW OF LITERATURE Earnings management is a subjective term. It has been expressed by different formula and definitions. Ma (1988) has defined income smoothing as the deliberate decrease in earnings fluctuations as compared to a given level of normality. Many researchers have highlighted it as a manager's discretionary decisions. Thus earnings management exists where a manager uses biased judgement in reporting the financial of company, either to misinform stakeholders about the performance of the company or to benefit from contracts that derive rewards from reported accounting numbers (Wahlen, 1994). Although the term income smoothing and the activities classified as income smoothing were not new terms, the earliest credited literature contributed on banks using loss loan provisions as an earnings management tool can be found in studies conducted by Greenawalt and Sinkey (1988) and Ma (1988). Greenawalt and Sinkey (1988), using a sample of 106 large bank holding companies from 1976 to 1984 focused on how the loss loan provisions behave to affect the quality of the loans. Using pooled time-series, they found that in comparison to the money-centre banks it was the regional banking companies which engaged in earning management at a larger scale, thus affecting the quality of loans. Ma (1988) while analysing the reason for a large number of U.S bank failures during the late 1970s tried to understand whether the U.S banks indulged in income smoothing. It was found that Loss Loan Provisions do not affect the quality of loans but the LLPs are increased when the operating income is high. The previous researchers have linked the earning management incentives to the Transaction cost theory which indicates that reporting decrease in income would lead to higher transaction cost and the Prospect theory which highlights that fall in earnings would send a negative signal to credit rating agencies and stakeholders. According to

Kanagaretnam, Lobo & Mathieu (2004) earnings management also depends on bank-specific factors such as the cost of borrowing may reflect the perceived risk associated with the bank. Using the bank by bank approach on the sample of U.S. banks, Collins, Shackelford and Wahlen (1995) found that banks do indeed use LLPs as an instrument of income smoothing. Similar results were found by Bhat (1996) who concluded that the sample of US banks used in his study depicted a strong relationship between LLPs and earnings. He generalised that banks with low growth, the book to asset high loan to deposit ratios are more likely to indulge in earnings management. A research conducted on a sample of U.S banks by Cornett, McNutt and Tehranian (2006) further confirms the significant relation of LLP with earnings management. They further suggest that LLPs are also significantly related to the percentage of shares the CEO and directors own where the CEO/Chair duality exists. In a research conducted by Kim, Liu and Rhee (2003) to study the relationship between earnings management and size of the firm for 18 years from 1983 to 2000 including banks conclude that small firms do earnings management to avoid reporting losses while the large firm does earnings management to avoid reporting a decrease in earnings. The evidence of banks involving earnings management through LLPs is not only restricted to U.S. Hasan and Hunter (1999) have used a sample of Spanish banks to examine the efficiency of bank managers to take a decision regarding the LLPs and understand the factors affecting the efficiency of such decisions. Thus, their conclusion highlights the high degree of inefficiency in LLP decision making by the bankers. Chipalkatti and Rishi (2007) in an attempt to examine the use of Loss Loan Provisions to meet the regulatory requirements of non-performing assets and capital adequacy norms, has identified that banks with weak financial position and performance do not engage in earning management, but understate their NPA. Reverte (2008) while studying the relational of institutional structure and earnings management suggested that where there are high-quality reporting practices the income smoothing is significantly lower as in the case of the European Union institutions. Pinho (1997) conducted a study on the determinants of LLP in Portuguese banks which suggested that with a high-interest margin, low loan to asset ratio had an effect on the provisions for loss loan. One such technique is to keep “renewing” defaulted loans and use of extraordinary items. A new contribution made to the literature of LLPs is by Burgstahler and Dichev (1997)

who provided evidence of earnings management through cashflow from operations and working capital management. Evidence has also been found where the banks do not use LLPs as an earnings management tool. These results differ due to a difference in time period. Taktak et al., (2010) conducted a study of 66 Islamic banks for a period of 6 years from 2001-2006 using Beidleman and Eckel coefficients in regression analysis to test whether Islamic banks use LLP for income smoothing, which resulted in a negative result. Beatty et al. (1995) with a sample of 707 public banks and 1160 private banks from 1988-1998 found that there exists only a small numerical relation between LLPs and earnings. Thus, no significant evidence could link loss loan provisions to earnings management.

Wall and Koch (2000) state

although the evidence suggests that bank manager have an incentive to manage earnings there may be a difference in results due to different periods. Ahmed et al. (1999), in an attempt to understand the impact of changing policies did not find any evidence that suggested the use of LLPs for earnings management. Anandarajan(1999) analysed the use of LLPs for earnings management and signalling effect across Australian banks and found results to be inconsistent with the previous researchers. All the studies suggest a mixed response pertaining to the difference across country, time and type of regulatory structure. Thus, on the basis of these research works, I conclude that there is strong evidence that LLPs are used as a tool for earnings management by banks, as it is supported by recent studies. Also, the incentives for bank managers to smooth income using LLPs are well documented. RESEARCH DESIGN In India, by the end of the financial year 2017, the banking system comprised of 27 Public Sector banks, 20 private banks and 43 foreign banks. The study has considered eleven years period from 2007 to 2017. The foreign banks have been excluded from the study since their banking operations and accounting format differ on account of multiple trade transaction. Secondly, data for foreign banks is unavailable. Thus, on the basis of capital market capitalisation, 26 public sector banks and 11 private banks have been selected (Annexure A). The data has been collected from CMIE PROWESS and the annual reports of the selected banks. From the reviewed literature, the variables have been selected and a model has been formulated which closely follows the models used to test the income smoothing hypotheses

as used by Greenwalt and Sinkey (1988) aa well as Kanagaretnam, Lobo, and Mathieu (2003) in their study. The model is as follows:

Where:

LLP = α1 + β1LNTA + β2TL + β3NPA+ β4NPL + β5EBTP + β6CAR + ε

Variables LLPit TA TL NPA EBTP CAR

Table 1: Definitions of the Variables Empirical Definition Specific and general LLP of Bank i in year t normalised by total assets Natural logarithm of Total asset bank i in year t Total loans and advances of the bank i in year t normalised by total assets Non-performing assets Earnings before tax and provision Capital Adequacy Ratio( Tier-I+ Tier II)

Source: Researchers’ Compilation

In the study, Loss Loan Provisions have been taken as the dependent variable to determine income smoothing as a result of the behaviour of the explanatory variables. The literature consists of various variables that explain the loss loan provisions of the banks. The variable NPA (Non-Performing assets) and TL (total loans normalised by total asset) have been taken as a measure to depict a bank’s risk profile. The variable EBTP, on the basis of prior research, has been taken as a proxy for income smoothing because the precondition of income smoothing is that the managers have incentives to smoothen income. Thus it is expected to be positive. Since the banks vary in size and have economies of scale in operations due to the resources the natural logarithm of total assets (TA) has been taken to neutralise the effect of size of banks. The following hypotheses have been formulated: H1: There is a significant positive relationship between Loss Loan Provisions and EBTP. H2: There is a significant positive relationship between Loss Loan Provisions and NPA H3: There is a significant positive relationship between Loss Loan Provisions and Loans and Advance of the banks. H4: There is a significant positive relationship between Loss Loan Provisions and Capital Adequacy Ratio. H5: There is a significant positive relationship between Loss Loan Provisions and Total assets RESULTS AND DISCUSSIONS Prior to using the regression, to verify the compatibility of test with data (there are no irregularities in the data set), tests were conducted. To check the stationarity (or non-

stationarity) of the independent variables, panel unit root test was used. The following part of the research paper discusses the results and findings of the study with the above-mentioned variables for 37 banks from 2007-2017. The first step is to understand the normality of the time series.

Variable TA TL NPA EBTP CAR LLP

Table 2: Unit Root Test Statistic Value Levin, Lin & Chu t* -29.3220 Im, Pesaran and Shin W-stat -3.66476 Levin, Lin & Chu t* -10.9012 Im, Pesaran and Shin W-stat -2.28569 Levin, Lin & Chu t* -5.42882 Im, Pesaran and Shin W-stat -2.93721 Levin, Lin & Chu t* -13.7591 Im, Pesaran and Shin W-stat -5.51759 Levin, Lin & Chu t* -5.18768 Im, Pesaran and Shin W-stat -5.18768 Levin, Lin & Chu t* -16.0819 Im, Pesaran and Shin W-stat -4.31376

Source: Researchers’ Compilation

Sig. 0.0000 0.0001 0.0000 0.0111 0.0000 0.0017 0.0000 0.0000 0.0000 0.0027 0.0000 0.0000

Conclusion I(2) I(2) I(2) I(2) I(2) I(2) I(2) I(2) I(0) I(0) I(2) I(2)

Table 2 articulates the result of unit root tests. Unit root test is applied to check whether the data is stationary or not to avoid spurious results. The result of the unit root test suggests that variables such as capital adequacy ratio are stationary at level difference while EBTP is stationary at the second level difference. The remaining variables such as Total loans, total assets and Non-Performing assets are also stationary at the second level difference. Thus, the null hypothesis is rejected because there is no trend in the analysed data. The alternate hypothesis is accepted that data is stationary. Hence, the results of the test determine the test to be applied to study the relationship between loss loan provisions and the influencing variables. Now, we can move forward to explore the relationship between loss loan provisions and other variables. Table 3 reports the lag selection criterion where lag eight has been selected based on the Final prediction error, Akaike information criterion and sequential likelihood-ratio (LR) test all chose eight lags, as indicated by the “*” in the output. Toda and Yamamoto (1995) proposed a simple procedure to estimate an augmented VAR using asymptotic distribution of Wald statistics (an asymptotic χ2 –distribution). Since the results of the unit root test clearly show the stationarity of the variable at level, first difference and second difference, the model for regression analysis chosen is the Toda & Yamamoto Test. This test is an extension of

Granger Causality test which neither binds the variables to a specific order of difference I(0), I(1), I(2), nor to the cointegration of uninformed order (Wolde- Rufael, 2005). Table 3: VAR Lag Order Selection Criteria VAR Lag Order Selection Criteria Endogenous variables: LLP NPA TOTAL_ASSET TOTAL_LOANS CAR Exogenous variables: C Sample: 2007 2017 Included observations: 111 Lag

LogL

LR

FPE

AIC

SC

HQ

0 1 2 3 4 5 6 7 8

-7278.306 -7053.040 -7023.423 -7006.603 -6952.946 -6918.545 -6887.416 -6809.537 -6754.030

NA 426.1789 53.36437 28.78977 87.01198 52.68586 44.87168 105.2417 70.00784*

6.77e+50 1.83e+49 1.69e+49 1.98e+49 1.19e+49 1.03e+49 9.52e+48 3.84e+48 2.36e+48*

131.2307 127.6223 127.5391 127.6865 127.1702 127.0008 126.8904 125.9376 125.3879*

131.3528 128.3546* 128.8817 129.6394 129.7333 130.1741 130.6739 130.3314 130.3920

131.2802 127.9194 128.0838 128.4787 128.2100 128.2881 128.4253 127.7200 127.4179*

* indicates lag order selected by the criterion LR: sequential modified LR test statistic (each test at 5% level) FPE: Final prediction error AIC: Akaike information criterion SC: Schwarz information criterion

The empirical results of Toda and Yamamoto (1995) have been reported in Table 4. The estimates of Block ExogeneityWald test shows chi-square distribution with 8 degrees of freedom (Lag Length=8) and the corresponding Probability. In the above-given table, the pvalue of the earnings before...