Application of financial covenants in credit risk management at commercial banks a case on construction enterprises in vietnam

STATE BANK OF VIETNAM MINISTRY OF EDUCATION AND TRAINING BANKING UNIVERSITY OF HO CHI MINH CITY - BACHELOR THESIS Major: Financial - Banking APPLICATION OF FINANCIAL COVENANTS IN CREDIT RISK MANAGEMENT AT COMMERCIAL BANKS: A CASE ON CONSTRUCTION ENTERPRISES IN VIETNAM Student’s name : Nguyễn Hoàng Thủy Tiên Student’s ID : 030630141644 Guiding teacher : Major ID Nguyễn : 7340201 Ho Chi Minh City, 2018 inh Nh ABSTRACT The research has focused on the application of financial covenants into the Viet Nam market with the objects are construction enterprises The author choose the a non-parametric model called Decision Tree model to measure the thresholds for selected indicators from 100 listed construction companies in the period of time from 2014 to 2017 The detailed process has been demonstrated, from preparing data, selecting variables to building a model and finally obtaining result The performance figures of the obtained model are all reasonable, which proves its capability to apply into covenants Despite providing numerous beneficial contribution, there are still some shortcomings, mainly due to the lack of time and information, have been withdrawn Recommendations for improvement and further studies have also been proposed Based on the results of the study, the paper provides not only the idea about a new mechanism for credit risk management in Viet Nam market but also the suitable thresholds to apply it into construction companies i DECLARATION OF AUTHENTICITY I hereby confirm that I am the sole author of the written these here enclosed and I have compiled it in my own words With my signature, I confirm that this dissertation is my original work, gathered and utilized especially to fulfill the purposes and objectives of this study I have mentioned all people who were significant facilitators of the work I declare that all statements and information contained herein are true, correct and accurate to the best of my knowledge Ho Chi Minh City, December 26th, 2018 Nguyen Hoang Thuy Tien ii ACKNOWLEGEMENTS Firstly, I would like to express my gratitude towards MSc Nguyen Minh Nhat for his dedication, patience and support me with useful recommendations and guidance in the process of this study Secondly, I would like to thank all Banking University's lecturers who have provided me with valuable knowledge as well as experience to help me obtain a solid academic foundation for the future research, career practice and workflow Finally, I would like to thank my friends for the sharing and support throughout my Bachelor program However, due to the limitation of experience and time, the study cannot avoid some certain drawbacks All the comments and advice contributed by members in the committee are gratefully welcomed and appreciated Ho Chi Minh City, December 26th 2018 Nguyen Hoang Thuy Tien iii TABLE OF CONTENTS ABSTRACT i DECLARATION OF AUTHENTICITY _ii ACKNOWLEGEMENTS iii LIST OF ABBREVIATIONS _ vi LIST OF TABLES vii CHAPTER 1: INTRODUCTION vii 1.1 Motivation 1.2 Research goal _ 1.3 Research questions _ 1.4 Research subject and range 1.5 Research method 1.6 Research contribution 1.7 Research Outline CHAPTER 2: LITERATURE REVIEW _ 2.1 Review about Construction Industry in Viet Nam _ 2.2 Credit risk 2.2.1 The importance of credit risk management in commercial banks 2.2.2 Credit risk management models 2.3 Loan covenants _ 10 2.3.1 Covenant types 12 2.3.2 Basic financial covenant structure 16 2.3.3 Prior study on the application of financial covenants _ 22 2.4 Decision Tree Model _ 27 2.4.1 Introduction _ 27 2.4.2 Previous application of Decision tree model 29 CHAPTER 3: RESEARCH METHOD AND DATA _ 33 3.1 Data Sources _ 33 3.2 Data Preparing _ 33 3.3 Variables selection 34 CHAPTER 4: EMPIRICAL RESULTS _ 37 iv 4.1 Obtained Results 37 4.2 Discussion about research results 39 CHAPTER 5: CONCLUSION AND SUGGESTIONS _ 41 5.1 Limitation _ 41 5.2 Recommendations _ 42 5.3 Proposal for further research _ 42 REFERENCES _ viii APPENDIX _ xi v LIST OF ABBREVIATIONS EBITDA Earnings Before Interest, Taxes, Depreciation and Amortization DTs SMEs Decision Tree model Small and Medium Enterprises IS Income Statement BS Balance Sheet vi LIST OF TABLES Name of tables Page Table Types of Loan Covenants Table Input ratios of RiskCalc™ for Japanese private company models Table Selected input ratios by Engelmann and Rauhmeier Table The separation of Default and Non-default firms Table Selection of input variables Table Descriptive Statistics table Table Correlation matrix vii CHAPTER 1: INTRODUCTION 1.1 Motivation In a project financed transaction, the lenders will want to ensure that the revenue stream is protected and the project performs as it is supposed to with no default on the loan However, credit risk is an inevitable but controllable factor that every bank has to consider when accepting a company loan With traditional methods of risk management such as credit risk measurement models, banks will only capable of controlling risk after the loan is disbursed but not before Loan covenant is a method with practical control mechanism which will give banks the right to create limitations on what the project company can without bank‟s approval and the ability to step into management of the investment decisions made by the company Specifically, this forecast mechanism will help the bank control the company‟s developing direction by keeping financial ratios in thresholds which will benefit the bank more than the company Banks will also seek to have trigger events, which allow them additional rights and powers in the event of their occurrence (for instance, if certain ratios such as debt to equity ratios or debt service cover rations are breached by the company) Given the priority of banks to ensuring security of the project revenue stream, a number of financial ratios will be key to the analysis of a project financed transaction Financial ratios can quantify different aspects of the project company‟s business and operations and are an integral part of analyzing its financial position During due diligence, before financial close, banks will run these ratios using various sensitivities, for example testing the financial ratios in the event construction costs increase by 20%, or revenues fall by 10% After financial close, the banks will use these ratios as part of the project monitoring and control functions Where ratios not achieve the levels required, they will have a series of possible interventions including blocking dividend distribution, sweeping cash from existing accounts, applying reserve account money to debt service, taking control of additional rights of the borrower or its shareholder If these breaches persist, eventually, such breaches will amount to events of default permitting the lenders to accelerate, cancel outstanding loan amounts or suspended existing loans It may also permit them to increase the interest margin, require compensation of the lenders for additional investigation costs and other fees and fines Jing Wang (2017) stated that loan covenants are designed to assist creditors‟ preferred mechanisms to control and protect loan values in contracting environments with corporations They can give banks a more active role in corporate governance through either the experience or the forecast mechanisms When the covenant is designed with the forecast mechanisms, they either make sure shareholder‟s actions are consistent with the bank‟s preference or incorporate positive prospect of the borrowers, so that future creditor intervention is not necessary; when the covenant is designed with the experience mechanisms, they can reduce moral hazard problems through creditor intervention after the lending relationship is underway A proper set of loan covenants can reduce the probability of default, the loss given default and the net exposure at default by a significant amount for commercial borrowers The net result is an expected loss rate of almost 20% less with a set of strong covenants than without A large part of this expected loss reduction is the fact that covenants give the lender an average of seven or more months of early warning between the time the covenants are tripped and the time a delinquency occurs The construction market in 2018 may continue to be optimistic, thereby enabling businesses in the industry to expand their business and consolidate solid resources This results in the raising need of loans from construction companies, but it also leads to a bigger problem in bank lending – risk management The issue that creditors are most concerned about will be categorizing the kind of risk they are facing and choose the corresponding types of covenants In Viet Nam, loan covenants haven‟t been considered a strategy for risk management in banks Therefore, this thesis will help creditors – commercial banks in this situation, build a suitable covenant model that can apply to Viet Nam‟s financial market, typically with construction companies 1.2 Research goal The purpose of this study is to focus on how to use loan covenants in credit construction companies in three most recent years are gathered and divided into six categories as a preparation Table Selection of input variables Variables Description EBITDA to Interest Expense EBITDA to Interest Expense (X5) Debt to Capitalization (X6) Total Debt to Net worth Debt to EBITDA (X7) Total Debt to EBITDA Current Ratio (X8) Current Asset/Current Liabilities Quick Ratio (X9) (Current Asset – Inventory)/Current Liabilities Source: Author The reasons why author pick these variables are explained as following: EBITDA to Interrest Expense The EBITDA-to-interest coverage ratio is a ratio that is used to assess a company's financial durability by examining whether it is at least profitable enough to pay off its interest expenses The EBITDA-to-interest coverage ratio is also known as EBITDA coverage A ratio greater than indicates that the company has more than enough interest coverage to pay off its interest expenses Because EBITDA does not account for depreciation-related expenses, an interest coverage ratio of 1.5 might actually be sufficient Debt to Net Worth The Debt/Net Worth ratio gives analysts and investors a better idea of a company's financial structure and whether or not the company is a suitable investment All else being equal the higher the debt-to-capital ratio, the riskier the company Firms with higher level of this ratio are believed to be default more frequently 35 Debt to EBITDA Debt/EBITDA is a measure of a company's ability to pay off its incurred debt The ratio gives the investor the approximate amount of time that would be needed to pay off all debt, ignoring the factors of interest, taxes, depreciation and amortization This ratio is commonly used by credit rating agencies to assess a company's probability of defaulting on issued debt, a high Debt/EBITDA ratio suggests that a firm may not be able to service its debt in an appropriate manner and warrants a lowered credit rating Current Ratio The current ratio is a liquidity ratio that measures a company's ability to pay short-term and long-term obligations To calculate the ratio, we compare current assets to current liabilities The current ratio is called “current” because, unlike some other liquidity ratios, it incorporates all current assets and liabilities A current ratio that is in line with the industry average or slightly higher is generally considered acceptable A current ratio that is lower than the industry average may indicate a higher risk of distress or default Similarly, if a company has a very high current ratio compared to their peer group, it indicates that management may not be using their assets efficiently This ratio is believed to be negatively associated with credit risk Quick Ratio The quick ratio is an indicator of a company‟s short-term liquidity position, and measures a company‟s ability to meet its short-term obligations with its most liquid assets Since it indicates the company‟s financial position to instantly use its near cash assets (that is, liquid assets) to get rid of its current liabilities, it is also called as the acid test ratio Liquid assets are the assets that can be quickly converted into cash with minimal impact to the price received in the open market, while current liabilities are a company's debts or obligations that are due to be paid to creditors within one year 36 CHAPTER 4: EMPIRICAL RESULTS The output of obtained model will be presented in this chapter And as introduced in Chapter 3, Decision Tree model will be exploited 4.1 Obtained Results All of them comprise variables chosen above with the Median, Maximum and Minimum value as well as Standard deviation as below Table Descriptive Statistics table Mean SD Min Max 163.5291 2396.001 -17.67726 46822.25 Debt to Net Worth -0.5991805 82.24245 -1634.473 110.2076 Debt to EBITDA 6.453728 28.54285 -279.6604 287.4631 Current Ratio 1.553662 0.9403967 0.3354524 11.67916 Quick Ratio 1.017466 0.797584 0.1542044 10.15117 EBITDA to Interest Expense Source: Author As it can be seen, Min value of EBITDA to Interest Expense is negative because there are some companies having negative EBITDA in this period of time Similarity, Min values of accounts Debt to Net Worth and Debt to EBITDA are negative because of the same reason Moreover, Current Ratio is 1.55 which shows that the majority of construction companies have an effective operating cycle and able to pay off short term debt Although a few companies have a rate of less than 1, there is a possibility of not getting a good financial situation, but that does not mean that the company will go bankrupt because there are many ways to raise more capital Quick Ratio is 1.01 much more smaller than Current Ratio (1.55) shows that the liquidity of short-term assets is relatively low To see the associations or predictive relationship between variables, we can call the correlation matrix: 37 Table Correlation matrix X1 X2 X3 X4 X1 X2 0.0016 X3 -0.0061 0.0136 X4 0.0111 0.0238 -0.0517 X5 0.0372 0.0059 -0.0435 0.8629 X5 Source: Author It can be seen from Table that the connection among these variables is quite weak As a result, it could be said that all variables not depend on each other strongly, and that is very positive to the model The output of obtained model will be presented below Demonstration of Decision Tree Source: Author 38 A quick look at the output of the Decision Tree model can help categorize the firms that considered “good” and the ones considered “bad” to the banks The thresholds which is classified as non-default For example, it can be explained that companies whose current ratio is under 0.995 and the ratio quick ratio is under 0.59 will be classified as default The more specific explanation will be presented in the next section 4.2 Discussion about research results This section will present regression results of the model based on Decision Tree data, the thresholds suitable for the borrowers, in this case are construction companies, are: Current Ratio In this result from the regression, the threshold of Current Ratio at 0.995 might be a safe value, it shows that the company has the ability to pay short-term and long-term obligations A current ratio that is less than one may seem alarming but the current ratio indicates the company can‟t cover all its current debts, doesn‟t mean it won‟t be able to once the payments are due In theory, the higher the current ratio, the more capable the company is of paying its obligations because it has a larger proportion of short-term asset value relative to the value of its shortterm liabilities However, a high ratio (over 3) could indicate the company is not using its current assets efficiently, is not securing financing very well, or is not managing its working capital Although there is some ambiguity, the current ratio can still be a useful measure of a company‟s short-term solvency when it is placed in context of what has been historically normal for the company and its peer group Quick Ratio A figure of is considered to be the normal quick ratio, as it indicates that the company is fully equipped with sufficient assets that can be instantly liquidated to pay off its current liabilities As from the obtained results, the suggested Quick Ratio threshold for covenants is higher than 0.59, however, in some circumstances, having a healthy Quick Ratio doesn‟t mean the company is able to meet up its short term liabilities On the other hand, if the company negotiates rapid receipts of payments from its customers and secures longer terms of payments from its 39 suppliers, it may have a very low quick ratio but may be fully equipped to pay off its current liabilities EBITDA to Interest Expense As from the obtained results, the suggested threshold of EBITDA to Interest Expense ratio for covenants is 0.145, even though this is a low number, a wellestablished utility will likely have very consistent production and revenue, particularly due to government regulations, so even with a relatively low interest coverage ratio it may be able to reliably cover its interest payments Debt to EBITDA As from the results, the suggested threshold of Debt to EBITDA ratio for covenants is 0.715, which is relatively ideal for banks and investors Because there is no absolute number showing how much Debt / EBITDA ratios is acceptable, commercial banks often give certain limit that require borrowers to keep, such as a Debt to EBITDA ratio < based on the financial situation of the company and the agreement between them and banks Debt to Net Worth As from the obtained results, the suggested threshold of Debt to Net Worth ratio for covenants is -0.99, however, in some situations, this ratio can be too generalized and need to be investigated more Because if a lot of debt is used to finance growth, a company could potentially generate more earnings than it would have without that financing If leverage increases earnings by a greater amount than the debt‟s cost (interest), then shareholders should expect to benefit However, if the cost of debt financing outweighs the increased income generated, share values may decline The cost of debt can vary with market conditions, thus unprofitable borrowing may not be apparent at first In conclusion, the thresholds found above can be supportive to a covenant design but need more consideration, for example the threshold for current ratio is at 0.995, but a bank can limit it at more than or more than 1.2 depends on the subject company Another example to quick ratio, the threshold for “safe zone” is 0.59, but the bank can use a limit of more than 0.6 for composing a suitable covenant Similarly, other ratios can be adjusted before put in a financial covenant 40 CHAPTER 5: CONCLUSION AND SUGGESTIONS A variety of theoretical and empirical studies has been reviewed to build a strong and trustworthy fundamental for this paper Despite being a contribution to the field of credit risk assessment and modeling, there are some limits existing in this paper, which open direction for further research 5.1 Limitation The most striking drawback of this study is small set of data Due to the limit of time, only 100 enterprises are investigated and restricted to construction field Number of input variables is 5, which is suitable to the amount of observations However, this sample is considered small, resulting in the lack of significant outcomes Another limit of the study is the use of technical default, most lenders would rather not be confronted with minor technical defaults that borrowers easily could have avoided by, for example, providing the lender with a required notice of material event But since technical defaults are so prevalent, loan managers must spend considerable time figuring out whether the borrower is in breach, whether such breach affects the risk profile of the loan and whether the breach demands some sort of formal response such as default notice, consent, waiver or loan amendment Thus, the final result of this study is openly suggested, it is not a specific and correct in all cases It means that the result could be variable because it depends on the purpose of every individual or company who apply different values or different variables for different purposes Decision tree based model required an input classification that is just a technical judgment, not truly one This judgment varies depending on the factors/ratios that are decided by authors, like in this research, which sometimes lacks of objectiveness Different input classifications lead to different result Hence, technical methods decided by authors may negatively affect to the accuracy of the model In addition, because of the limit of time and Vietnam‟s market features, the number of selected financial ratios may not be as much as expected, so this study 41 could not cover every single aspect of 100 construction companies in the dataset 5.2 Recommendations Because the input of this study is mainly from financial statements, banks need to guarantee the accuracy as well as the authenticity of the data Banks also should collect information about their clients through different sources such as from CIC, The State Bank of Vietnam and the company‟s partners Future researchers should pick up more financial ratios and more companies in a longer duration for more precise results 5.3 Proposal for further research It is recommended that further research investigated in the following areas: Widening set of data and period of time: In order to obtain more trustworthy outcome, it is required that the number of investigated enterprises be extended to 500 or even above And so is the amount of input variables The sample should cover enterprises in various fields so that the flexibility and adoptability of Decision tree model can be tested More comparison between Decision tree model and other different models: Further research should focus on finding out the difference as well as the advantages and disadvantage of those models When the reasons that lead to one outperforming the other is found, they may become beneficial to the application in reality Finally, other researchers could reuse the framework of this study to expand the scale with a much bigger dataset in a longer duration of time 42 REFERENCES Domestic Research Lê Văn Tư (2005), Quản trị rủi ro ngân hàng thương mại, NXB Hà Nội Trần Thị Xuân Hương 2002, „Phân tích tài xếp loại doanh nghiệp cơng tác thẩm định tín dụng ngân hàng‟, Tạp chí phát triển kinh tế Foreign Research Anthony, Sasch, Felix and Bjorn (2012), Covenant Violations, Loan Contracting, and Default Risk of Bank Borrowers, NYU Working Paper No 2451/31417 Creation Achleitner et al (2012) Private Equity Lemons? Evidence on Value in Secondary Buyouts, SSRN Electronic Journal, available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1756901 Allen, D.E and Powell, R.J 2011, „Credit risk measurement methodology‟, Proceeding of 19th International Congress on Modelling and Simulation, Perth, Australia, Available from , 12 December 2011 Beaven W H (1966), Financial Ratios as Predictors of Failure, Journal of Accounting Research, vol 4, page 71-111 Basel Committee on Banking Supervision (BCBS) (2004), Basel II: International Convergence of Capital Measurement and Capital Standards: A Revised Framework Available from http://www.bis.org/publ/bcbs107.htm Cem Demiroglu & Christopher James (2006) The Information Content of Bank Loan Covenants Chris Nichols (2014), What loan covenants matter for banks, November Felix Freudenberg et al (2012) Covenant Violations, Loan Contracting, 26 and Default Risk of Bank Borrowers Hans B Christensen and Valeri v Nikolaev (2011) Capital Versus Performance Covenants in Debt Contracts, available at https://doi.org/10.1111/j.1475679X.2011.00432.x 10 Jing Wang (2014) Debt covenant renegotiations and creditor control rights, Journal of Financial Economics 113 (3) viii 11 Jing Wang (2017) Debt covenant design and creditor control rights: Evidence from the tightest covenant - Journal of Corporate Finance 44, page 331 – 352; available at https://www.journals.elsevier.com/journal-of-corporate-finance 12 John K Paglia and Donald J Mullineaux (2006) “An Empirical Exploration of Financial Covenants in Large Bank Loans”, Banks and Bank Systems, Vol 1, pages 103-122; available at www.businessperspectives.org/journals_free/bbs/BBS_2006_02_Paglia.pdf 13 Lance Moir and Sudi Sudaranam (2007) Determinants of financial covenants and pricing of debt in private debt contracts: the UK evidence, Accounting and Business Research, 37:2, pages 151-166, available at http://dx.doi.org/10.1080/00014788.2007.9730066 14 Michel Crouchy et al (2001) Prototype risk rating system, Journal of banking & Finance 25, page 47 – 95 15 Michael R Roberts (2015) The role of dynamic renegotiation and asymmetric information in financial contracting, Journal of Financial Economics volume 116, pages 61-81, available at https://www.sciencedirect.com/science/article/pii/S0304405X14002591 16 Ohlson, J A 1980 „Financial Ratios and the Probabilistic Prediction of Bankruptcy‟, Journal of Accounting Research, Spring, 109-131 17 Ong, M 2005, Internal Credit Risk Models Capital Allocation and Performance Measurement, Risk Books 18 Peter Demerjian (2010) Financial Covenants, Credit Risk and the Resolution of Uncertainty 19 Redouane Elkamhi et al (2015) Accounting Quality and Financial Covenants in Loan Contracts, Journal of Accounting Research 20 Servigny, N D., & Renault, O (2004), Measuring and managing credit 21 Sinkey J 1992, „Commercial Bank Financial Management‟, risk Mcmillan 22 Song, Yan 2015, „Decision tree methods: applications for classification and prediction‟, Shanghai Arch Psychiatry Apr 25; Available from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4466856 , Apr 25, pg 130–135 ix 23 Tysk D (2010) Forecasting defaults and analysing the interaction between defaults and the real economy, Central Bank of Iceland, June 16 24 Zopounidis C, Dimitras AI 1998, Multicriteria Decision Aid Methods for the Prediction of Business Failure, Kluwer, Boston x APPENDIX 100 construction companies chosen Trading code Data Base AME HNX B82 UPCoM BCE HoSE BHT UPCoM BII HNX C32 HoSE C47 HoSE C92 HNX CDC HoSE CDO UPCoM CSC HNX CTD HoSE CTX HNX CX8 HNX DC4 HNX DIG HoSE DIH HNX FCN HoSE HAS HoSE HBC HoSE HDG HoSE HU1 HoSE HU3 HoSE HUT HNX ICG HNX IJC HoSE ITA HoSE xi L10 HoSE L18 HNX L35 HNX L43 HNX L61 HNX LCG HoSE LCS HNX LGL HoSE LHC HNX LIG HNX LM7 HNX LM8 HoSE LO5 HNX LUT HNX MCG HoSE MCO HNX MDG HoSE MEC HNX NDN HNX NHA HNX PHC HoSE PPE HNX PPI HoSE PPS HNX PTC HoSE PVA UPCoM PVV HNX PXA UPCoM PXI HoSE PXL UPCoM QTC HNX xii S55 HNX S74 HNX S99 HNX SC5 HoSE SD2 HNX SD5 HNX SD6 HNX SD9 HNX SDC HNX SDD HNX SDH UPCoM SDT HNX SDU HNX SIC HNX SJE HNX SSM HNX THG HoSE TV1 UPCoM TV2 HNX TV3 HNX TV4 HNX UDC HoSE V12 HNX V21 HNX VAT HNX VC1 HNX VC2 HNX VC3 HNX VC5 UPCoM VC6 HNX VC7 HNX xiii VC9 HNX VCC HNX VCG HNX VE2 HNX VE3 HNX VE9 HNX VMC HNX VMI HNX VNE HoSE VRC HoSE Correlations Test xiv ... not provide adequate financial information to give an exact forecast Financial ratios have also been considered by a vast number of rating agencies when evaluating credit quality Standard & Poor‟s... performs against the financial projections provided by the business owner, CFO, or management Financial covenants mostly base on financial numbers (financial ratios) In many papers authors indicate... focused more on credit risk management in business and gradually approaching risk management standards under the Basel II international treaty on risk management Credit risk measurement was fully