SME Lending Y - RBA · risk management department of the bank. Risk management experts evaluate the credit risk of the loan applicant typically using a three-stage lending process

DO NOT COPY

50

RETAIL BANKING II

RETAIL BANKINGACADEMY

203.SME Lending

Course Code 203 - SME Lending

DO NOT COPY

51© Retail Banking Academy, 2014

RETAIL BANKING II


Course Code 203SME Lending

Introduction

It is conventional wisdom that small and medium size enterprises (SMEs) contribute substantially to economic growth in most countries in the sense that they are the principal source of employment generation and output growth. For example, Gerrit de Wit and Jan de Kok (2014)* report that small firms in the European Union as a whole contribute on a greater scale towards job creation than larger firms, a finding that is similar for the US (Neumark et al, 2011)†. An interesting comparison of the role of SMEs in enhancing national economic growth for both developed and developing countries is provided by Pandya (2012)‡.

Not surprisingly there is no consistent definition of SMEs across economies and proposed definitions vary country by country. A firm that is defined as small in a developed country may be regarded as large in one that is emerging. For example, in Egypt, an SME is defined as having between five and 50 employees, whereas in the EU an SME is one that has fewer than 250 employees. The Inter-American Development Bank defines an SME as having no more than 100 employees and less than US$3 million in turnover (i.e., revenue) while in the US an SME has less than 500 employees. Finally, as a general guideline, the World Bank defines an SME as having less than 300 employees, US$15 million in total revenue and US$15 million in total assets.

So SME lending has a direct effect on national economic growth. But SME lending can be quite risky for banks. The main reason is that there is considerable informational asymmetry faced by banks. This is because SMEs typically have highly opaque financial statements so that banks must rely on ‘soft’ data§ to complement their credit evaluation process.

Lending to SMEs typically takes the form of overdrafts, business credit cards, short-term secured and unsecured loans and long-term loans (typically secured). The purpose of SME lending includes working capital financing, trade financing, mortgages for commercial and industrial property, investment in plant and equipment and supply chain financing.

* Refer to “Do small businesses create more jobs? New evidence for Europe”, Small Business Economics, Volume 42, Issue 2, pp 283-295 (February 2014).† Refer to Review of Economics and Statistics, 93(1), page 16-29.‡ Refer to the article presented at the 2012 International Conference on Business and Management 6 – 7 September 2012, Phuket, Thailand.§ Soft data is explained in Chapter 1.


DO NOT COPY

52

RETAIL BANKING II


The lending process begins with the loan officer submitting a completed loan application to the risk management department of the bank. Risk management experts evaluate the credit risk of the loan applicant typically using a three-stage lending process – screening of the prospective borrower followed by the contracting stage, where terms and conditions are set, and finally the monitoring stage, where credit risk migration is tracked and analysed. We adopt the traditional 5Cs model of commercial lending in our consideration of the three-stage lending process.

The remainder of this module is organised as follows: Chapter 1 considers the first stage of the lending to SMEs process: screening. The 5Cs model is the basis for the screening of prospective borrowers. These factors are character, capacity, collateral, capital and conditions. While all five factors are important in establishing the creditworthiness of prospective borrowers, the process defined by the 5Cs model is commonly called ‘collateral-based lending’ so as to emphasise the collateral requirements imposed by the bank. We also consider the credit risk evaluation procedure for unsecured loans in Chapter 2, where this type of lending is called ‘information-based lending’. The bank-customer relationship is examined and the high credit risk associated with new SME borrowers that have been in business less than three years is analysed. If the screening stage results in a decision to proceed with the loan application, then terms and conditions are set (contracting) and the monitoring of credit risk follows. Chapter 2 also deals with these two latter stages and provides a full explanation of the role of the price and non-price terms and conditions. Included are details of how the loan rate is calculated and an extended discussion of the role of funds transfer pricing (FTP) that is presented in Module 206. Chapter 3 deals with the calculation of the expected credit loss that incorporates the realistic assumption that probability of default (PD) and loss given default (LGD) are correlated. This is an input into a model of concentration risk. The potential for the bank’s loan portfolio to exhibit concentration risk can be one of the deciding factors in the risk manager’s decision to lend to an SME. Accordingly, banks place limits on loan concentration.


DO NOT COPY


RETAIL BANKING II


Chapter 1: The Lending Process – The Screening Stage

It is widely accepted by both academics and practitioners that lenders to SMEs are at a disadvantage when it comes to obtaining valid and reliable information on prospective customers. The list of potential problems includes opaque and unaudited financial statements, encumbered collateral that may also be non-marketable and high debt-to-equity ratios. These are so-called ‘hard data’ problems. There are also ‘soft data’ issues that relate to the integrity of the owner(s). While hard data gives the lender a financial picture of the applicant’s ability to repay a loan, soft data provides a more subjective indication to the loan officer of the prospective borrower’s willingness to make repayments in full and on time.

Petersen (2004)* presents a comparison of hard with soft data. He states that “hard information is almost always recorded as numbers. Thus in finance we think of financial statements, the history of which payments were made on time, stock returns, and quantity output numbers as being hard information. Soft information is often communicated in text. It includes opinions, ideas, rumours, economic projections, statement of management’s future plans, and market commentary. The fact that hard information is quantitative means that it can easily be collected, stored, and transmitted electronically”.

In short, hard data are quantifiable while soft data are more subjective. Soft data may take on added importance in the lending process when hard data are not complete or reliable. Indeed, to overcome the problems of opaque financial statements, banks can acquire soft data by sending a loan officer to visit the prospective borrowing firm and speak with key personnel to get an evaluation of the firm’s creditworthiness.

The value of soft data in SME lending is highlighted by numerous researchers in the financial services industry. For example, Chen et al† (2013) show that “the use of soft information significantly improves the power of default prediction models” and hence improves the estimation of expected loan loss.

* Mitchell A. Petersen, Information: Hard and Soft, Kellogg School of Management, Northwestern University and NBER.† Yehning Chen et al, “Soft data and small business lending”, Journal of Financial Services Research (November 2013).


DO NOT COPY

54

RETAIL BANKING II


The Three-stage Lending Process

Lending to SMEs may be characterised by a three-stage process: screening, contracting and monitoring.

The screening phase is the start of the lending process, where the lender evaluates the financial health of the business in terms of the firm’s ability to repay the loan obligation. It is a data collection stage and a preliminary evaluation of key issues. Typically the last three years’ financial statements (including a cash flow statement) are reviewed. In addition, a three-year projection is requested for review of business development and ability to pay. Importantly, because of the possible lack of reliable hard data, there is a risk of adverse selection. Indeed, the academic and professional literature has provided evidence that banks face a higher level of adverse selection risk for SMEs compared to larger corporate firms. This is largely due to information risk arising from opaque financial statements.

Contracting is the second stage of the lending decision process, where the terms and conditions are set by the lender. These terms and conditions include: the cost of the loan to the customer; the loan amount; the term of the loan; and the requirements for collateral. This stage is considered only if the decision is made to proceed beyond the screening stage.

The last stage is the monitoring of the customer’s financial condition and risk-taking activities so as to assess credit risk migration over time. It is also an opportunity for the bank to identify potential sources of moral hazard risk. This risk occurs when the customer invests in projects that are riskier than the lender expected, leading to unexpectedly higher credit risk that was not adequately priced at the contracting stage. The three stages of the lending process are facilitated by the 5Cs of commercial lending. We complete an analysis of the screening stage with a discussion of the purpose of the loan, a practical SME lending scorecard based on the 5Cs with indicative scores and practical advice for the loan officer to achieve an effective screening process.

The 5Cs of Commercial Lending – a Consideration of Collateral-Based Lending

The ‘5Cs of credit’ paradigm* represents a long-established practice utilised by lenders to screen prospective customers so as to establish their creditworthiness. This paradigm comprises the categories of character, capacity, collateral, capital and conditions. Character is identified with integrity, which shows the borrower’s willingness to repay the loan obligation over the term of the loan. Capacity, capital and collateral together form the basis for a quantitative financial analysis. Capacity is a measure of the borrower’s ability to repay the loan while capital or equity represents the company’s stake. This is a commitment of the owner as pre-emption of moral hazard risk for the lender. Collateral is security against the likelihood of the borrower defaulting and reduces the losses due to default. Conditions refer to external variables such as macroeconomic variables – national as well as sector-specific – that are likely to impact the future profitability of the firm and thereby influence the migration of credit risk over time.

In the rest of this chapter, we examine each of the 5Cs in relation to their respective effects on the creditworthiness of the potential borrower – the screening stage.

Character

Character refers to the integrity of the small business owner(s). Integrity means “honouring one’s word”. Indeed, as asserted by Erhard, Jensen and Zaffron† (2013), “in summary, we show that defining integrity as honouring one’s word…and empowers the three virtue phenomena of morality, ethics and legality”.

* George E. Ruth, Commercial Lending, 5th Edition.† See: Werner Erhard, Michael C. Jensen and Steve Zaffron, “Integrity: A positive model that incorporates the normal phenomena of morality, ethics and legality”, Harvard Business School NOM Working Paper No. 06-11 (SSRN id: 2277407)


DO NOT COPY


RETAIL BANKING II


In relation to lending, character has an ethical connotation: a willingness to repay the loan obligation. This is the link to the idea of keeping one’s word.

How can a bank evaluate a potential borrower’s integrity?

It is suggested that one way to evaluate integrity is to consider the potential borrower’s history of repayment of previous obligations as well as the length of the credit history. In short, integrity must be validated over a minimum period of time. Indeed, as asserted by Ruth (2004, See footnote on previous page), “persons lacking character assign a low priority to repaying debts and are quick to default on any loan commitment at the first sign of financial trouble”.

It is interesting that the FICO credit score (owned by the Fair Isaac Corporation in the United States), which is used to evaluate the creditworthiness of potential borrowers, comprises five factors: payment history; amount owed; length of credit history; credit mix; and new credit. The highest weight of 35 percent is applied to payment history, followed by 30 percent for length of credit history. While many economies (especially, but not only emerging ones) do not have access to a credit score for SMEs, it is recommended that internal credit models should give added significance to these two factors of the FICO score.

The value of the customer’s integrity is seen to be very important in the practice of SME lending. For example, Amrei Botha, head of SME banking at Standard Bank Africa, stated in an interview (see below) that, in establishing the creditworthiness of loan applicants, added emphasis is placed on the customer’s willingness to repay his/her obligations.

“Prospective clients are asked to fill in a business questionnaire, which gives us an indication of their willingness to repay debt. We also assess their ability to repay by evaluating sales turnover, gleaned from banking records. Unbanked entrepreneurs are asked to transact with us for a month, to make this information available.

“So essentially what we are trying to assess where an informal business applies for a loan is first of all whether they have the willingness to repay the loan, then second of all, whether they have got the ability to repay the loan,” said Botha.

Source: The Africa Report, 2 October 2012

In summary, it is recommended that two indicators of character should be applied in the screening stage of the lending process. These are:

• number of missed payments on all loan obligations;• length of repayment history, with a minimum of three years.

These indicators give a degree of measurability to character which is normally an example of soft data.

While character is a key component of the 5Cs paradigm, it is not sufficient to grant a loan. The following lesson is always important to heed:

While good character is not sufficient to grant a loan, bad character (lack of integrity) is sufficient to deny a loan.

Good character can be tested by negative economic conditions. This is when individuals of integrity seek to restructure their loan obligations rather than quickly seek to default.

The second component of the 5Cs paradigm is ‘Capacity’, which measures the customer’s ability to prepay the loan obligation.


DO NOT COPY

56

RETAIL BANKING II


Capacity*

It is understood that net profits do not repay a loan. This section will present a rationale for this statement and recommend that lenders switch from the analysis of SME profitability to producing cash flow projections. The ability of a loan applicant to repay a loan is based on cash flow and not just profitability.

Indeed, in the Financial Handbook for SMEs† it is noted that ”most SMEs keep track of their revenue and profits by looking at their profit and loss statement. As long as it shows a profit, entrepreneurs usually assume that profit is the same as cash, but that may not be true.”

An article in the Daily Telegraph (9 December 2013) makes a key point in relation to the importance of cash flow. There, it was reported that more than 50 percent of SMEs polled in the UK admitted having cash flow concerns. The report implied that this may be the reason why ”recent funding and lending figures show that access to credit for SMEs remains subdued. Of the one in five small businesses that applied for a loan in the past two years, more than half had their application rejected or are still waiting for feedback.”

We now examine some cash flow ratios that are typically used by lenders to assess the borrower’s ability to repay its loan obligations. Firstly, we consider a measure of cash flow – earnings before interest, taxes, depreciation and amortisation (EBITDA).

EBITDA is defined simply as total revenues less operating expenses excluding depreciation and amortisation and before interest and tax payments. Hence, it is a measure of operating cash flow.

Here is a simple income statement of an SME that illustrates the calculation of EBITDA.

Income Statement for the year ending 31 December 2013(All monetary values are stated in thousands of dollars)

Revenues 5,060

Total Operating Expenses (OPEX) 3,130

Interest Expense 250

Depreciation & Amortisation 370

Taxes (20 %) 336

Net Earnings 1,344

In this case, EBITDA = (5,060 – 3,130) + 370 = 2,300. Note that OPEX includes depreciation and amortisation which is added back to obtain a value for EBITDA.

A common ratio that measures an SME’s operating cash flow relative to the interest it pays on total outstanding debt is Interest Coverage Ratio (ICR). ICR is calculated as follows:

ICR = PaymentsInterestTotal

EBITDA

In our example, ICR = 2,300 / 250 = 9.2. This means that the current level of EBITDA covers 9.2 times the current level of interest payments. While this ratio is relatively simple and easy to calculate, one of its major drawbacks is that an SME’s EBITDA is likely to fluctuate much more than its interest payments. Hence, projections of ICR are likely to be unreliable. In addition,

* As a first step, the loan officer must ensure that the borrower has legal capacity to act on behalf of the SME.† Financial Handbook for SMEs (Action Community for Entrepreneurship, The Association of Banks in Singapore, 2007)


DO NOT COPY


RETAIL BANKING II


interest expense may not represent all of the SME’s total debt obligations. For example, it does not account for mandatory principal repayments.

It should be noted that ‘capacity’ summarises the SME’s ability to repay its loan obligations and hence ratios should be based on cash flows. EBITDA does not include interest, taxes or required asset investments. All of these are real cash expenses and should be considered when evaluating a company’s ability to repay its total debt obligations.

Accordingly, banks consider total debt service ratio (TDSR) as a relatively more reliable indicator of the SME’s ability to repay its total debt obligations. The formula for TDSR is defined as follows:

paymentsincipalInterestTotalsInvestmentTaxesEBITDATDSR

&&−

=Pr Re

A value for TDSR that is below 1.2 is viewed as high risk; a value above 1.5 indicates low risk while a value within this range is considered as medium risk.

Evaluation of a New Loan Obligation

One approach for the evaluation of a new loan obligation is to consider the SME’s debt service ratio (DSR), which compares the SME’s free cash flow (FCF) with the terms of the new loan obligation.

First we present the procedure for the calculation of FCF:

FCF = EBITDA – Existing (Interest Expense and Mandatory Principal Repayments)

• Projected tax payments

• Change in Working Capital Requirements*

• Required Replacement Investment Expenditure

A positive FCF reflects the SME’s ability to finance new loans. Simply put, the SME’s FCF is what is available to finance new loans. We now present the formula for DSR as:

)( paymentsincipalMandatorypenseInterestExNewFCFDSR

+=

Pr Re

This ratio typically ranges between two and four and the choice of a minimum value by the lender depends on several factors – including its risk appetite, age of the business, portfolio concentration and sector risk.

Although we have presented the procedure to calculate the value of financial ratios for one period, it is important to examine the trend in the values of this ratio over time. For example, a relatively stable and high ratio for DSR that is calculated on a quarterly basis over a three-year period provides a degree of assurance that that trend is likely to continue.

One of the principles of commercial lending is commonly described as the ‘Two Ways Out’, which means that the borrower must have two distinct and independent ways of repaying the loan. Adequate free cash flow is one of these two ways out. The most common second way is the provision of collateral that may be liquidated to recover the bank’s outstanding exposure.

We now consider the role of ‘Collateral’ in the screening of potential SME borrowers.

* Note that a negative change will increase free cash flow while a positive change will result in a decrease. To fully understand this point, consider Question 3 of the Multiple Choice Questions at the end of this module.


DO NOT COPY

58

RETAIL BANKING II


Collateral

Collateral is an asset or set of assets that are pledged by the borrower to secure a loan. These assets typically can include accounts receivable, inventory, equipment and real estate. They constitute a secondary source of loan repayment should the borrower be unable or unwilling to repay the full amount. But non-interest earning assets, in particular, which are pledged as collateral may themselves become impaired and lose value over time. Hence, in converting collateral to cash at the time of default, the lender may find the recovery (market) value to be lower than expected. This is termed a ‘collateral gap’ when asset values decline in the markets.

As reported in the Wall Street Journal (22 July 2010), ”many small businesses, thwarted in efforts to get loans, are saying it takes money to get money. That’s because property and equipment assets have fallen in value, so businesses seeking loans are being asked for alternative collateral, often in the form of cash so that the loan is backed in case the borrower defaults. According to Kathie Sowa, a commercial banking executive at Bank of America Corp., one of the nation’s largest small-business lenders, basic underwriting standards haven’t changed: Cash flow must be sufficient to support the loan, and there must be a secondary source of repayment. Collateral was typically a combination of accounts receivables, inventory, real estate, equipment, and other business or personal assets. But since real-estate and equipment values have plummeted, she says, business owners who may have landed loans in the past are now falling short of having sufficient assets.”

We now ask the key question – why do banks (sometimes) require collateral?

As already stated above, one can reason that collateral provides a reduction in the risk of a loss to the bank should the borrower default. It is a second source of repayment. Loss given default (LGD) is reduced and hence the recovery rate is increased. But there are other reasons why a bank might require collateral from an SME. Three such reasons are proposed in the professional literature. We consider each in turn.

a) It is well-accepted that banks as lenders have an informational disadvantage compared to the borrower, who typically has quite opaque financial statements. One implication is that borrowers, after the loan is granted, may take on projects with risk levels that are higher than the bank expected. This can put the bank’s loan repayment in jeopardy should such a project fail. This is the classic moral hazard problem. By being required to provide collateral, the borrower has an interest in curtailing high-risk activities. This occurs at the monitoring stage of the lending process and is fully discussed in Chapter 2.

b) The informational disadvantage referred to in a) above also creates a problem for the bank at the screening stage of the lending process – before the loan is granted. This is the classic adverse selection problem in lending where some information is hidden from the bank. As stated by Bester (1985)*, if the bank cannot discern borrowers’ riskiness because of hidden information, then collateral may serve as a screening device to distinguish between borrowers and to mitigate the adverse selection problem. Specifically, this literature states that lower-risk customers are likely to willingly provide collateral while high-risk customers are less likely to respond positively to the bank’s request for collateral. This follows from the observation that a lower-risk borrower has a greater incentive to pledge collateral than a risky borrower, because of his lower probability of failure and loss of collateral. The bank then lowers its risk of adverse selection. This is called the adverse-selection hypothesis, and it occurs at the screening stage of the lending process.

c) The last rationale for the bank requiring collateral is based on what is commonly called the observed-risk hypothesis. This states that banks charge riskier borrowers higher loan rates and require higher collateral from these borrowers. Hence, there is a positive relationship between risky customers and the requirement for collateral. This also occurs at the screening stage of the lending process.

* H. Bester, “Screening vs. Rationing in Credit Markets with Imperfect Information”, The American Economic Review 75(4), 850-855 (1985).


DO NOT COPY


RETAIL BANKING II


So, we see that there are three reasons why a bank requires collateral – reduction of loan loss in the event of default, mitigating the risk of adverse selection and moral hazard.

In a study of 5, 843 loans made to borrowers located in 43 countries, Laurent Weill and Christophe Godlewski* provided evidence showing ”support for the adverse selection and observed-risk hypotheses, as both hypotheses may be empirically validated depending on the degree of information asymmetries in the country.”

What is the lesson from this study?

When there is a high level of information asymmetry, as may be the case in emerging economies (so that the bank has a significant information disadvantage), banks require more collateral as part of their screening process in an effort to compensate for hidden information that could lead the bank to make adverse selection (i.e., wrong choice of a borrower).

We now consider another factor in the 5Cs paradigm – ‘Capital’.

Capital

Capital, also known as shareholder equity, is the difference between the accounting value of assets and liabilities and represents the owner’s equity stake in the company. It represents the owners’ stake and lenders look to a high value of shareholder equity relative to the amount of long-term debt as an indication of the total risk of the company.

An important accounting ratio in this regard is the debt-to-equity ratio† which is a measure of the long-term solvency of the company. It is defined as the accounting value of long-term debt divided by owners’ equity.

There are some important considerations of this key ratio.

a) A high value of debt-to-equity implies that the company already has a high level of debt relative to the owners’ stake. This could mean that total current interest expense is quite high, which may negatively affect the free cash flow of the company. The company’s ability to repay additional loan obligations may be negatively affected – i.e., capacity suffers.

b) While we asserted above that net profits do not repay loans but cash flow does, there is an important implication when a company is unable to create positive profits over time. This arises from the fact that net profits (positive or negative) become retained earnings (after possible dividend payments) in the next accounting period. A loss would mean negative retained earnings and hence, all else being equal, a reduction in owners’ equity leading to a higher debt-to-equity ratio. It can also lead to moral hazard problems for the bank if the owners’ equity is reduced sufficiently and owners’ seek high-risk investments to earn a higher level of profits.

c) An erosion of accounting value of the company’s assets due to, for example, increased depreciation and amortisation or impairments will reduce asset values and lower owners’ equity. Since there is no change in the amount of long-term debt, the debt-to-equity ratio will increase and the company’s risk of insolvency will increase.

Accordingly, a thorough analysis of the company’s business plan, balance sheet, profit and loss statement and cash flow statement is required to obtain a full and transparent picture of the company’s long-term viability. This is evidenced in part by the debt-to-equity ratio. While there is no hard and fast rule, a common level for the maximum value of the debt-to-equity ratio is two to one. This means that a bank may not lend to a company where the debt-to-equity ratio is already at two to one. But lending practices vary in different market conditions and depend on the borrower’s Capacity (ability to repay) as well as Character (willingness to repay).

* University of Strasbourg (November 2006).† There are several variants of the debt-to-equity ratio. One version considers total debt to equity where total debt is the sum of short-term and long-term debt.


DO NOT COPY

60

RETAIL BANKING II


Similar to the recommendation made for the ICR and TDSR, the bank requires three-year projections of the debt-to-equity ratio on an annual basis.

The final component is ‘Conditions’, which refers to current and anticipated economic and market conditions facing the prospective borrower.

Conditions

In screening prospective borrowers, the bank must assess the probability of default over the proposed term of the loan. Clearly, credit risk varies over time and hence it is important to assess future macroeconomic conditions – international, national and local – in order to assess potential sources of future credit risk. For example, a forecast of a downturn in the national economy might increase unemployment, which could adversely affect the business of the prospective borrower. This would reduce the borrower’s ability to repay the loan obligations (i.e., Capacity) and may lead to a higher probability of default (PD). Further, asset values may decline in capital markets so that collateral values may be reduced, leading to a higher loss given default (LGD). We see that market conditions can have a dual impact on PD as well as on LGD. Negative economic events can increase both PD and LGD leading to higher credit risk than existed at the beginning of the screening process.

Hence, the loan officer must seek the expertise of the bank’s economists in order to assess the prospective borrower’s vulnerability to the forecasted economic conditions. The increased risk may lead to a denial of the loan or the requirement for a higher level of good quality collateral.

The above analysis of the 5Cs of lending to SMEs shows that the five factors are inter-related and hence should not be considered in the screening as if they are independent. Here are some inter-relationships that are noteworthy:

• Capital – a high debt-to-equity ratio can lead to higher principal repayments and interest expense and a lower free cash flow leading to lower ability to repay loan obligations (Capacity);

• Conditions – forecasts of an economic downturn may increase credit risk over time leading to a denial of the loan request or a requirement for a higher level of good quality collateral.

We now consider another key factor that a bank typically considers in its screening process – the ‘purpose’ of the loan. Generally, SMEs seek bank loan financing for three main reasons:

a) Improve short-term liquidity by seeking working capital loans or an increase in the limits of lines of credit;

b) Replace inefficient or depreciated assets such as equipment and,

c) Invest in projects for higher profitability, market share or increased net cash flow.

Banks typically prefer to invest in option c) since these projects are expected to generate cash flows to finance the new loan. Banks do not typically look favourably at working capital loans intended to enhance short-term liquidity since there may be operational issues in the SME that should first be addressed before seeking a loan. For example, there may be problems with collecting receivables, unduly high inventory levels and low cash and cash equivalents. As a general rule, a positive value of TDSR implies that the SME has a positive level of free cash flow implying that a) and b) are covered. This leaves c) as a bank’s preferred reason to lend to an SME.

Finally, in this section, we provide a summary of practical advice for loan officers in their screening process.


DO NOT COPY


RETAIL BANKING II


Advice for Loan Officers – Screening Stage

1. Have a second source of repayment – e.g. collateral or personal guarantees for family-owned businesses.

2. Know the principals of the SME as well as the type of business the SME is in.

3. Seek the advice of senior business and risk managers when you are dealing with a new loan applicant or an SME that is new in business.

4. Have effective documentation to support the screening stage.

5. Place a higher weighting on cash flow than on profitably.

6. Be aware of the value of soft data.

We close this chapter with a recommended SME Lending Scorecard that is based on the 5Cs paradigm. We propose an indicator attribute for each of the 5Cs with a suggested scoring scheme.*

5Cs LOW-RISK MEDIUM-RISK HIGH-RISK

CHARACTER

Number of Missed Payments

(Payment History ≥ 3 years)

0 1 > 1

CAPACITY

Total Debt Service Ratio (TDSR) > 1.5 Between 1.2 and 1.5 inclusive

< 1.2

Debt Service Ratio (DSR) > 4 Between 2 and 4 inclusive < 2

COLLATERAL ASSET

Class of Committed Asset* Level 1 Level 2 Level 3

CAPITAL

Debt-Equity Ratio (D/E) < 1 Between 1 and 2 inclusive

> 2

CONDITIONS

Forecasted Business Cycle Upturn Stable Downturn

An overall scoring scheme is proposed as follows:

• An SME loan applicant is scored as LOW-RISK if all indicator attributes are scored as low risk;

• An SME loan applicant is scored as HIGH-RISK if at least one indicator attribute is scored as high risk;

• If the SME loan applicant is scored as neither low-risk nor high-risk then the defaulting score is MEDIUM-RISK.

Comment

If the loan officer cannot take mitigating actions (e.g. seek better quality collateral or personal guarantees), then SME loan applicants that are scored as high-risk are to be eliminated from further consideration. Medium-risk loans may only be considered if sufficient mitigating

* These levels refer to IFRS classifications of availability of market prices to value assets.


DO NOT COPY

62

RETAIL BANKING II


measures can be taken.

If the decision is to proceed, then the contracting and monitoring stages follow. Chapter 2 deals with these two remaining stages of the lending process for both collateral-based and information-based lending.


DO NOT COPY


RETAIL BANKING II


Chapter 2: The Lending Process – Contracting and Monitoring

Contracting comes into effect after the risk manager has decided that screening of the creditworthiness of the prospective borrower has revealed positive results. This stage specifies a set of terms and conditions that are intended to:

• mitigate the borrower’s credit risk for the bank and, • increase the bank’s ability to monitor credit risk migration over the term of the loan.

It is important that the terms and conditions set by the loan officer must also facilitate effective monitoring after the loan is granted.

Terms and conditions fall into two main categories:

a) price of the loan and,b) non-price aspects that can include collateral requirement, loan size, loan maturity and loan covenants.

There are generally two approaches utilised by a bank to set the price of a loan. Firstly, the price is based on the combination of the bank’s cost of funding through a fund transfer pricing procedure and a target profit margin where the latter reflects competitive pressures. This is sometimes called uniform pricing.

But uniform pricing does not consider sources of risk that are specific to the loan applicant.

To deal with these additional sources of risk, the loan officer will make adjustments to non-price terms and conditions such as loan size, loan maturity, loan covenants and collateral requirement.

The second approach to pricing of loans to SMEs is based on the level of credit risk incurred by the bank by lending to the particular SME. This procedure is called ‘risk-based pricing’*. This methodology results in differential loan prices according to the riskiness of the prospective borrowers. An added benefit of this approach is that it is possible to link the price of the loan to a Risk Adjusted Return on Capital (RAROC) benchmark set by the bank. However, there is evidence that banks more often adopt the uniform pricing method to set SME loan interest rates. Since risk-based pricing is fully explained in Module 209 and is not commonly used by banks to obtain

* Details of risk-based pricing are presented in module 209, which deals with risk management.


DO NOT COPY

64

RETAIL BANKING II


customer loan prices, we only consider the details of the uniform pricing approach in this chapter.

The remainder of this chapter deals with price and non-price terms and conditions that are applied to the contracting and monitoring stages of the lending process to SMEs.

Pricing of Loans to SMEs

We now deal with the details of uniform pricing, where all SMEs are charged the same price for a same loan duration, although small variations are possible, if, for example, collateral commitments comprise mostly liquid assets. As an illustration, ICBC (Asia) currently offers the prime rate plus one percent to all SME customers for unsecured loans over a three-year term. All qualified customers will receive this price, although small discounts resulting in a price such as prime rate + 50 basis points may be offered to some SMEs which have a long-term relationship with the bank or which have additional business with the bank.

But the key guideline remains:

Terms and Conditions for a Loan to a Particular SME =

Identical Loan Rate for Borrowers + Non-Price Terms and Conditions that Incorporate the SPECIFIC risks of the SME

Uniform Loan Pricing Model

There are different approaches to obtaining loan prices. We provide a review of a popular model that is supported by the Global Association of Risk Professionals (GARP) in its professional examination for Financial Risk Managers (FRM) and then consider our preferred approach that is based on funds transfer pricing (FTP), a procedure that is fully explained in Module 207. The description of the FTP is enhanced by a consideration of the markup for credit risk, liquidity risk and optionality risk.

Saunders* presents a formula which may be stated, using the symbols in Module 207, as follows:

L = f + m + FTP where:

L = customer loan rate;

f = average acquisition and servicing costs over the duration of the loan;

m = target profit that includes expected loss on the loan;

FTP = reference (interbank) swap rate + markup for liquidity risk + markup for optionality risk

We rewrite the loan pricing model as follows:

Since FTP = reference (interbank) swap rate + liquidity markup + optionality markup, we have:

L= f+m+ interbank swap rate + liquidity markup +optionality markup

We now consider non-price terms and condition of a commercial loan to prospective SME borrowers and show how the markups on credit risk, liquidity risk and optionality are managed by risk managers.

* See: Anthony Saunders, Credit Risk Measurement, (New York: John Wiley & Sons, 1999).


DO NOT COPY


RETAIL BANKING II


Non-price Terms and Conditions

We stated in Chapter 1 that collateral is traditionally viewed as a risk-reducing contractual feature; it serves to compensate moral hazard risk for the bank and is potentially an effective screening device for prospective borrowers with hidden information and thereby reduces adverse selection risk for the lender. Banks are at an informational disadvantage when dealing with SMEs, which typically have unaudited and highly opaque financial statements and are not normally evaluated by ratings agencies nor followed by the financial press.

In this section we:

• consider some practical issues such as guidelines for setting the amount and type of collateral;

• advise caution to loan officers who may inadvertently succumb to the so-called ‘lazy bank hypothesis’ and finally;

• consider the role of the duration of the bank-customer relationship in determining collateral requirements by the bank.

Collateral, Haircut and Maximum Loan Amount

The (monetary) amount of collateral required for a loan is impacted by a haircut applied by the bank to the stated collateral value. Specifically, an estimate of the market value of the collateral is determined by the bank’s specialists and then a haircut (i.e., a discount percentage factor) is applied to determine the maximum loan amount. For example, if the market value of the asset that will serve to secure the loan is $1 million, and the bank applies a 30 percent haircut, then the maximum loan amount is 70 percent of the value of the collateral or $700,000.

Formally, Maximum Loan Amount = Collateral Value x (100% - Haircut (%))

This formula is related to the traditional loan-to-value (LTV) ratio which measures loan value to market value. In the example above, the max LTV is equal to 70 percent since the haircut is 30 percent.

Two issues are immediately evident. These are:

a) How to estimate the market value of the collateral asset?b) What is the appropriate haircut percentage?

We consider each of these questions in turn.

The accounting profession* has classified assets into three categories – the so-called three-level hierarchy. The first is Level 1 where these assets are traded on an active and liquid market and market quotes are readily available. This is sometimes called mark-to-market. Examples of Level 1 assets are shares of corporations that are traded on stock markets and also debt securities such as bonds that are traded in a market. Of course, cash and cash equivalents fall into this category.

The main point is that if these assets are offered to secure a bank loan, then their market values (as proxies for fair values) are easily observed. From a bank’s perspective, Level 1 assets are the best form of collateral commitment by an SME.

Level 2 assets are transacted between parties quite infrequently and so do not have consensus market prices as is the case for Level 1 assets. The value of the asset is typically estimated or appraised using a model. Simply put, Level 2 assets do not have observable prices, but data is available as input to a model to calculate a price. This is sometimes called mark-to-model. A classic example of a Level 2 asset is real estate.

Level 2 assets are not typically traded in an open market on a regular basis and are normally estimated through a recognised model using observable input data.* See Financial Accounting Standards Board (FASB) 157.


DO NOT COPY

66

RETAIL BANKING II


Level 3 assets are extremely illiquid and the market price is practically based on in-house opinion or estimates. Examples include inventory, accounts receivable etc.

The typical collateral assets of an SME are inventory, accounts receivable and real estate, which are Level 2 and Level 3 assets for which estimates of their respective market values are quite difficult to obtain with a high degree of confidence. There is a lot of estimation work so reliable values are not assured.

Banks prefer liquid collateral (i.e. Level 1 assets) since, should it become necessary to reclaim the funds borrowed, they can sell it relatively easily in an open market. It is no wonder that in times of severe economic distress, banks seek Level 1assets as collateral. The difficulty for borrowers is that this is precisely when they have less Level 1 assets and a preponderance of Level 2 and 3 assets.

The Role of Prudent Haircuts

While liquid collateral gives the bank an opportunity to sell the asset without undue delay, haircuts lower credit risk to the bank. Indeed, haircuts are determined by the expected future price volatility of the collateral asset and hence the degree of uncertainty about the true value of the asset. For this reason, Level 3 assets will typically be subject to a significantly higher haircut relative to Level 1 assets.

We summarise some key points on non-price terms and conditions in the box below:*

The size of the haircut on collateral assets for SMEs is typically high since collateral assets are normally in the category of Level 2 and Level 3.The actual value imposed by a bank will depend on local market conditions since only Level 1 assets are typically traded on liquid markets. Simply put, the size of a haircut to collateral is determined individually and not nationally, except for Level 1 assets.

The literature on SME lending has found some interesting relationships on non-price terms and conditions.

a) A borrower is less likely to default when the collateral value far exceeds the loan amount. The borrower is likely to repay the loan and free-up the collateral.

This assertion is quite interesting in relation to the calculation of the markup of credit risk. A high haircut implies a low loan-to-value (LTV) and hence less credit risk for the bank. This also implies a lower markup for credit risk.

b) Conventional wisdom states that loans backed mainly by Level 1 assets are less risky and tend to earn lower loan rates.

c) Loan size and collateral requirement are directly related. For example, Jackson and Kronman* conclude that larger loans should be more frequently secured. It is clear from the equation above, that the size of the loan is positively related to the collateral commitment.

d) Some authors also assert that loan size is linked to the probability of default by the borrower, since a firm that receives a large amount of debt attains a higher leverage level which increases the risk of non-payment. This point was raised by Avery et al.,† amongst others.

This assertion implies that the SME’s level of new debt may increase the customer’s level of leverage, which can increase the financial risk of the customer. Hence, the risk manager can manage the markup for credit risk by restricting the loan size so that the customers’ debt-to-equity ratio is not inordinately high.

* Thomas H. Jackson and Anthony T. Kronman, ”Secured Financing and Priorities among Creditors”, The Yale Law Journal, 88, 1143-1182, (1979).


DO NOT COPY


RETAIL BANKING II


*Lazy Bank Hypothesis

Manove, Padilla and Pagano† define a bank (specifically the loan officer) as “lazy” if it voluntarily chooses prospective borrowers who commit to a high level of collateral but without conducting effective customer screening – as prescribed in Chapter 1. By this action, the provision of high quality collateral is viewed by the loan officer as a substitute for effective screening. Accordingly, loan officers who accept loans on the basis of collateral requirements increase the risk of adverse selection – the risk of choosing the wrong loan applicant.

The lesson is clear – the screening stage of the lending process is an absolutely necessary step in that process. Collateral commitments alone are not substitutes for an efficient three-stage lending process. An assessment of the customer’s willingness and ability to repay the loan obligations is crucial.

Bank-Customer Relationship

There is evidence that the quality and duration of the bank-customer relationship is an important determinant of collateral requirement by the bank. Specifically, customers with a longer relationship with their bank incur a lower incidence of collateral (i.e. may even obtain collateral-free loans). This evidence is provided by, for example, Harhoff and Korting.‡ The rationale is quite clear. An SME with a longer relationship with the bank provides the latter with important information regarding the SME’s willingness and ability to repay loan obligations. The degree of information asymmetry (where the bank is at an informational disadvantage) is reduced and so the role of collateral in limiting credit risk and moral hazard is of less importance. Accordingly, the bank may require a reduced level of collateral or may even offer collateral-free loans.

The duration of the bank-customer relationship is negatively related to the level of collateral required by the bank. The longer the duration of the bank-customer relationship, the lower the level of collateral required by the bank since the credit risk is better judged from the increased transparency.

At this point, we see that loan size and collateral requirement are positively related with the maximum loan amount determined after applying a haircut to the market value of the collateral asset.

We now consider the role of the duration of the loan in controlling and monitoring credit risk for the bank.

Duration of Loan (or Time to Maturity)

It is clear that loans with a long maturity require that loan officers make a similarly long-term judgment of the creditworthiness of the loan applicant. Obviously, an SME that is deemed to be creditworthy at the time the credit decision is made will not necessarily be creditworthy in the future. And the likelihood of this being the case increases with a longer term for the loan. Indeed, banks will be less willing to grant loans with a long maturity to riskier SMEs or to offer unsecured lending. As we discussed in Chapter 1, the chance of occurrence of an adverse economic event (Conditions in the 5Cs paradigm) looms larger when the duration of the loan is longer. In this case, the provision of collateral by the loan applicant has the power to decrease this source of credit risk and so reduce the markup for credit risk.

For this reason, bank loans to SMEs are likely to have a relatively shorter time to maturity. This feature of loans to SMEs coupled with loan covenants are likely to address the issues arising from information asymmetry.

* R.B. Avery, R.W. Bostic, and K.A. Samolyk , “The role of personal wealth in small business governance” Journal of Banking and Finance, 22, 1019-1061 (1998).† M. Manove, A. J. Padilla and M. Pagano, “Collateral versus Project Screening: A Model of Lazy Banks”, Rand Journal of Economics, 32(4), 726-744 (2001).‡ Dietmar Harhoff and Tim Korting, “Lending Relationships in Germany: Empirical Results from Survey Data“, CEPR Discussion Papers 1917 (1998).


DO NOT COPY

68

RETAIL BANKING II


Loan Covenants

Covenants are financial or operating conditions imposed by the bank in the loan contract that serve to protect the interests of the lender. Paglia* indicates that covenants place requirements and/or restrictions to constrain the SME’s ability to undertake activities that are potentially harmful to the interests of the bank. Hence, covenants provide control and flexibility in the monitoring stage of the lending process and, if the borrower violates these covenants, lenders may require that the outstanding principal become due immediately or it may trigger an increase in interest rates on the loan.

A key rationale for the inclusion of loan covenants is that if they are violated, banks have the opportunity to react before the borrowing firm actually defaults on its loan. This is an important point since actual default may lead to a significant credit loss for the bank. In fact, violation of loan covenants by the SME gives the bank an opportunity to restructure the loan.

Loan covenants are generally classified as financial and non-financial. Financial covenants place restrictions on the SME’s accounting ratios such as debt-to-equity, total debt service ratio (TDSR) and interest coverage ratio (ICR). As pointed out by Mather and Peirson,† financial covenants are defined as ”…covenants that use accounting data in their formulation either as an absolute amount or as a ratio”’. Non-financial covenants can include restrictions on asset sales, forbidding additional encumbrances on committed collateral, description of uses of the loan, agreed to business activities or a requirement for periodic financial reporting.

Clearly, as asserted by Berger and Udell‡, financial covenants are likely difficult to enforce for SMEs with unaudited and highly opaque financial statements. This places added emphasis on the bank’s monitoring of non-financial covenants after the loan is granted.

What is the main lesson?

In order to facilitate effective monitoring, it is advisable that lenders in the contracting stage,

• seek collateral commitments with conservative haircuts especially for Level 2 and 3 assets;

• issue loans with relatively short maturity to minimise the effects of potential adverse economic events; and

• include in the loan contract effective non-financial covenants which provide restrictions and early warning signals for the bank.

These all have the effect of reducing the markup for credit risk

We now deal with determinants of the markup for liquidity risk and markup for optionality.

Markup for Liquidity Risk

The Basel Committee on Banking Supervision (BCBS)§, reported that many of the fundamental principles of liquidity risk management were neglected by banks during the recent financial crisis. Banks must now include a markup for liquidity risk especially for long-term loans. Table 3 of a research paper published by the Bank for International Settlements¶ reveals a typical relationship between term liquidity premium** and loan maturity in years. Here is a copy of this

* John K. Paglia, “An Overview of Financial Covenants in Commercial Bank Loans,” The Risk Management Association (RMA) Journal, (July 2007).† Paul Mather, and Graham Peirson, “Financial covenants in the markets for public and private debt”, Accounting and Finance, Vol.46 Issue 2 (2006).‡ A.N. Berger and G.F. Udell, “The economics of small business finance: The roles of private equity and debt markets in the financial growth cycle”, Journal of Banking & Finance, 22(6-8): 613-673 (1998).§ Refer to Liquidity Risk: Management and Supervisory Challenges, BCBS, (February 2008).¶ See Joel Grant, “Occasional Paper No.10, Liquidity transfer pricing: a guide to better practice” BIS. Bank for International Settlements. 2011. Web. < www.bis.org/fsi/fsipapers10.pdf>.** The method for the calculation of the term liquidity premium is relatively complex. Details of this method


DO NOT COPY


RETAIL BANKING II


table in part:

Maturity of Loan 1 2 3 4 5

Term Liquidity Premium (basis points) 5 10 18 28 40

Clearly, the markup for liquidity risk increases non-linearly with the time to maturity of the loan and the precise size of the markup depends on the outstanding maturity of the loan and financial market circumstances. Hence, by issuing loans with a shorter time to maturity, the risk manager can lower the markup for liquidity risk. It is noteworthy that as the term to maturity of the loan increases not only does the liquidity markup increase but also, all else being equal, credit risk increases since the likelihood of default increases for long-term loans. This shows an inherent link between liquidity risk and credit risk with the term to maturity of loans.

Markup for Optionality

Prepayment risk is an important consideration in the pricing of commercial mortgage loans. This risk arises from an unexpected lower return of the outstanding loan obligations to the bank when the SME borrower refinances the loan or prepays. The literature on prepayment optionality on loans shows that ”prepayment risk may significantly impact the return on loans, return on equity and real estate loans to total loans ratios of various commercial banks”.*

One of the reasons for this result is that a significant reduction in market interest rates relative to the contractually fixed rate mortgage rate increases the likelihood that the SME borrower will exercise the option to prepay. In this case, banks cannot earn the interest income they intended to earn when the original mortgages were created. Accordingly, banks face reinvestment rate risk.

It is conventional wisdom that the higher the loan-to-value (LTV) ratio (i.e. the higher the leverage), the greater is the debt service requirement, which likely negatively affects the borrower’s ability to repay the loan obligation. Indeed LTV and debt service ratios (DSR) affect default risk of mortgages to SMEs. But there should be caution in making a strong conclusion. An SME may have a high LTV but if there are other favourable conditions such as a strong positive cash flow, the link between LTV and default risk may be less pronounced.

Lesson

Banks may manage the markup for optionality risk by limiting loan maturity (so that interest rate volatility is expected to be lower). In addition, a charge at the time of prepayment may act as a deterrent for the borrower to prepay.

In summary, there are actions that the bank can take to lower the respective markups for credit risk, liquidity risk and optionality risk. In this way, the bank can charge a loan rate that is competitive and not deviate substantially from the loan standard conditions.

So far we have described the three-stage lending process for collateralised loans. In practice, this applies to both short-term and long-term lending to SMEs by banks. But banks also issue unsecured loans to SMEs. Hence, we complete this chapter with a consideration of unsecured (i.e. collateral-free) loans.

Unsecured Loans – a Consideration of Information-Based Lending

Banks have become more willing to provide unsecured loans to SMEs. Here are some recent examples:

are found in Chapter 7 of Leonard Matz and Peter Neu, Liquidity Risk, Measurement and Management: A practitioner’s guide to global best practices (Wiley Finance, 2007).* Alex Fayman and Ling T. He, “Prepayment risk and bank performance”, The Journal of Risk Finance, Vol. 12 Issue 1, pp.26 - 40 (2011).


DO NOT COPY

70

RETAIL BANKING II


• “Diamond Bank (Nigeria) to push unsecured loans to SMEs” (Business Daily, 4 August 2013) – Individual SMEs can access a maximum of Naira three million in loans from Diamond bank without collateral

• “ICBC (Asia) offers unsecured loans to SMEs” – Fast approval can be finished within 48 hours; No collateral required; Interest rate as low as Prime +1%; Repayment period up to 36 months (ICBC website)

• “HSBC proposes to launch unsecured lending to small and medium enterprises (SMEs) by July this year” – Business Standard, 22 February 2014.

The process of establishing the creditworthiness of unsecured loan applications is commonly called ‘information-based lending’. It is fundamentally based on four of the Cs described in Chapter 1 (Collateral is excluded), the business experience of the owner(s) of the SME as well as on the duration of the bank-customer relationship. Before we provide details of information-based lending, it is important to consider new loan applicants separately from applicants who have an established relationship with the lender. For new loan applicants we further separate this category into two classes – SMEs that have been in business for less than three years and those that have been in business longer than three years.

New Loan Applicants

As indicated in the introduction to this module, banks grant both short-term and long-term collateralised loans but typically grant only short-term unsecured loans. We will discuss below why limiting the maturity of unsecured loans tends to reduce credit risk especially for new SMEs.

The screening of the creditworthiness of an SME loan to an applicant who is not a bank customer is typically based on a great degree of subjectivity. A critical factor is the length of time the SME has been in business. If it is the case that the SME is not just a new loan applicant but is also a relatively new business, then the bank may be subject to a very high level of credit risk. We define an SME to be a new business if its time in business is less than three years. Why?

Statistics show that for start-ups in the UK more than 33 percent of businesses fail within the first three years. This percentage can be much higher for high-tech startups.*

For new SMEs (i.e. in business life for less than three years), there is usually a lack of information on the financial performance of the firm so that projections of expected future profitability and cash flow (i.e. a measure of ability to repay) will not give a high degree of confidence. In addition, this may not be enough time to assess the business acumen of the principals of the SME nor their character (i.e. willingness to repay).

It may be concluded that high business risk in a new business in the early years will lead to high credit risk for the lender.

Lesson

It is recommended that a lender refrain from granting unsecured loans to new SMEs during the first three years of being in business unless there are exceptional circumstances. Even in these exceptional cases, it is advisable that banks offer unsecured loans for a short-term (see, for example the case of ICBC [Asia] where loan maturity is 36 months) and with loan sizes that are limited to a strict maximum loan amount so as to control the bank’s credit exposure.

For the second category of new SMEs that have been in business for more than three years, but not with the lending bank, there is likely to be more financial information on the firm’s profitability and cash flow. There is also more opportunity to assess the business experience and expertise

* The Huffington Post reported on 12 September 2012 that about 75 percent of venture capital-funded new businesses in the United States fail.


DO NOT COPY


RETAIL BANKING II


of the principals with a higher degree of reliability. As asserted by Minniti and Bygrave*, over a longer time period, owners of new SMEs gain accumulated knowledge that has a self-reinforcing capacity. It is this specific industry knowledge that contributes to the SME’s business survival after three years and contributes to future profitability.†

We now consider the case of an unsecured loan application by an SME which already has an established relationship with the bank. We have already mentioned the role of a durable bank-customer relationship on the decision of the bank to require collateral. Indeed, it can be similarly asserted that a bank-customer relationship provides the risk manager with both hard (financial) and soft (willingness to repay) data that are crucial for the decision to grant an unsecured loan. There is greater information transparency, which permits a better analysis of the SME’s credit risk. It is interesting that a robust bank-customer relationship can be mutually beneficial. Not only do banks obtain crucial information about the principals of the business and important financial data, Bharath, Dahiya, Saunders and Srinivasan‡ show that banks with established relationships with SMEs have a higher probability of securing future loan contracts (42 percent) than non-relationship banks (three percent). Accordingly, banks should consider only making unsecured loans to SMEs in business for at least three years that are customers of the bank.

New SMEs with < 3 years in business

New to the bank and new in the industry

New SMEs with ≥ 3 years in business

New to the bank but relatively experienced in the industry

SMEs with > 3 years in business and an established relationship with the bank

Not new to the bank and quite experienced in the industry

The credit risk for lending in this category of new

SMEs is very high.

Banks should normally refrain from making unsecured loans

to this category of SMEs.

The credit risk for lending to this category ranges from medium to high. There is more reliable data on the

expected future profitability of the SME and on the

industry knowledge and experience of the principals.

Banks might consider making unsecured loans to this category of new SMEs

The credit risk for this category of established

SMEs is medium. The bank-customer relationship provides more reliable

soft and hard data.

Banks may consider making unsecured loans only to

this category of SMEs.

We end these chapters with some advice§ for credit managers:

* M. Minniti and W. Bygrave , ”A Dynamic Model of Entrepreneurial Learning”, Entrepreneurship Theory and Practice, 23(3) pages 5-16 (2001).† A. Cooper, F. Gimeno-Gascon, and C. Woo, “Initial Human and Financial Capital as Predictors of New Venture Performance.” Journal of Business Venturing, vol. 9 (5), 371-395 (1994). ‡ S. Bharath et al, ”So What Do I Get? The Bank’s View of Lending Relationships”, Journal of Financial Economics, Vol. 85, No.2, pages 368-419 (2007).§ As reported in: Ken Brown and Peter Moles, Credit Risk Management (Edinburgh School of Business, 2008, 2011).


DO NOT COPY

72

RETAIL BANKING II


Rouse, Bankers’ Lending Techniques, Second Edition, (Chartered Institute of Bankers: Financial World Publishing, p.26, 2002) advises the risk manager as follows:

1. Take time to reach a decision.

2. Do not be too proud to ask for a second opinion.

3. Get full information from the customer and do not make unnecessary assumptions, i.e. do not lend to a business you do not fully understand.

4. Ensure that you obtain a detailed business plan with a projection of quarterly cashflow over a three-year period.

5. Do not take a customer’s statements and representations at face value; ask for evidence to support the statements.

6. Distinguish between facts, estimates and opinions when forming a judgment.

7. Think again when the ‘gut reaction’ suggests caution, even though the factual assessment looks satisfactory.

The final chapter in this module deals with two issues in credit risk related to commercial lending – the calculation of expected loss on a portfolio of lending under the realistic assumption that LGD and PD are correlated and, a consideration of concentration risk in the bank’s loan portfolio.


DO NOT COPY


RETAIL BANKING II


Chapter 3: Credit Risk Management - Expected Loan Loss and Concentration Risk in Lending to SMEs

It is probably instructive to review an article published in the Economist in 1998 that claims that:

“banks’ credit-risk models are mind-bogglingly complex. But the question they try to answer is actually quite simple: how much of a bank’s lending might plausibly turn bad? Armed with the answer, banks can set aside enough capital to make sure they stay solvent should the worst happen.

“No model, of course, can take account of every possibility. Credit-risk models try to put a value on how much a bank should realistically expect to lose in the 99 percent, or so, of the time that passes for normality. This requires estimating three different things: the likelihood that any given borrower will default; the amount at risk and the amount that might be recoverable [in case of default]”*.

This article identifies the key factors that determine ”how much a bank should realistically expect to lose”’ – these three factors are Probability of Default (PD), Loss Given Default (LGD) and Exposure at Default (EAD). The well-known formula† that links these three factors is as follows:

Expected Loss (EL) = PD x LGD x EAD where PD and LGD are expressed as percentages and EAD is a monetary amount. This formula is very simple but it makes a very strong assumption – that is, PD and LGD are uncorrelated. This assumption is not very realistic as we have explained in Chapter 1. Several researchers have demonstrated that PD and LGD are linked to macroeconomic events.

For example, during an economic downturn with high periods of unemployment and lower consumer demand, SMEs will typically experience lower sales revenue and, all else being equal, lower profitability and lower ability to repay their respective loan obligations. In addition, asset values usually erode leading to a decline in collateral values. So, in cases of both secured and unsecured loans LGD will rise. Furthermore, LTV values will rise and cause an increase in the probability of default, PD. This shows that LGD and PD are positively correlated.

Failure to recognise that PD and LGD are positively correlated will underestimate the value * “Model behavior”, The Economist, 28 February 1998.† This formula is presented in Retail Banking I.


DO NOT COPY

74

RETAIL BANKING II


of expected loss on a loan.

We illustrate this case in the table below. Statistical formulae required for this calculation are included in the footnote* below.

In the table we assume that the Exposure at Default (EAD) is $5 million with PD = 1% and LGD = 80%, then Expected Loss (EL) = 0.01 x 0.80 x $5,000,000 = $40,000

So the traditional formula where it is assumed that PD and LGD are uncorrelated gives a value for EL to be $40,000.

Now we consider the case where PD and LGD are correlated:

PD and LGD are correlated

We assume the following inputs for the three variables:

Exposure at Default (EAD) $5,000,000

Expected probability of default (PD) 1%

Standard deviation (i.e. volatility) of PD 9.95%

Expected loss given default (LGD) 80%

Standard deviation (i.e. volatility) of LGD 20%

Correlation between PD and LGD 50%

Based on these values, the expected loss, EL = $89,750, which is more than twice the value when it is assumed that PD and LGD are not correlated*. If the correlation between PD and LGD is increased (e.g. 60%), then the resulting value for EL is increased substantially to $99,700.

There is a substantial underestimation of EL when using the formula that assumes that LGD and PD are not correlated. There is an immediate implication for the provision of loan losses and on RAROC.

Lesson

The risk manager must determine the extent of the potential correlation between probability of default (PD) and loss given default (LGD) by analysing macroeconomic conditions (i.e. business cycles).

In addition, while the volatility of PD is calculated from the average historical values of PD, it is harder to estimate the volatility of LGD. Hence, it is necessary to evaluate the sensitivity of expected loss (EL) to a range of values for the volatility of LGD.

The value of EL is an important input into a simple model of concentration risk in the bank loan portfolio. In relation to concentration risk, it is important for the risk manager to assess whether the proposed loan to the SME contributes meaningfully to the concentration risk of the loan portfolio.

* The formula for EL is based on well-known formulae in elementary statistics. In this case, EL = EAD x [(Correlation between PD and LGD x Standard deviation of PD x Standard deviation of LGD) + (Expected PD x Expected LGD)]. Note that the standard deviation of PD = square root of (PD x (1-PD)) and the standard deviation of LGD is assumed, as well as the correlation between D and LGD.


DO NOT COPY


RETAIL BANKING II


Concentration Risk in Lending to SMEs

BCBS* states that “concentration of credit risk in asset portfolios has been one of the major causes of bank distress”. Indeed, Basel II proposed a model to estimate loan portfolio credit risk which assumes that the number of borrowers as well as the size of individual loans in the bank loan portfolio are high enough to ensure that no borrower’s behaviour can have an unduly harmful effect on the loan portfolio value as a whole – that is, banks must ensure proper diversification.†

Concentration risk in a bank’s loan portfolio refers to the additional risk borne by a bank by having too many loans and also overly large loan sizes in one sector, industry (i.e. sectoral concentration) or geographical area, or even in one firm (i.e. single name concentration, also called granularity). Credit concentration in the bank’s loan portfolio exists when there is a large aggregation of loans with common characteristics. Indeed, section 773 of Basel II states, “concentrations include:

• Significant exposures to an individual counterparty or group of related counterparties;

• Credit exposures to counterparties in the same economic sector or geographic region;

• Credit exposures to counterparties whose financial performance is dependent on the same activity or commodity; and

• Indirect credit exposures arising from a bank’s Credit Risk Management activities (e.g. exposure to a single collateral type or to credit protection provided by a single counterparty).”

The implication of an undiversified (i.e. concentrated) loan portfolio is clear – the bank may experience highly correlated default rates among SME borrowers. That is, by having a high degree of credit concentrated in a single industry, there is a risk that a single negative event can lead to a high percentage of firms in that industry experiencing financial distress – a type of contagion that generates very high credit risk for the bank.

A new loan to an SME after being subjected to effective screening with appropriate price and non-price terms and conditions is viewed favourably if it adds to the diversification of the bank’s loan portfolio. But the bank must have expertise in lending in this sector.

Limiting Concentration Risk

Altman and Saunders‡ state that the ‘early approaches’ to concentration risk analysis were based either on: (1) subjective analysis where the experts decide on a maximum percentage of loans to allocate to an economic sector or geographic location and (2) on limiting exposure in an area to a certain percentage of capital (e.g. 10 percent). This latter approach is based on the ‘exposure to capital’ ratio.

How does the bank set a limit on the maximum loan amount to a particular SME?

A model that sets concentration limits is proposed by Saunders and Cornett§ and is based on the exposure to capital ratio. The steps in implementing this model are as follows:

(a) Set a maximum tolerable loan loss expressed as a ratio of capital. This is an expression of management’s risk appetite which sets a limit on the maximum exposure for a single SME loan in the sector.

* BCBS, “Studies on credit risk concentration”, Bank for International Settlements, Basel Committee on Banking Supervision, WP No.15 (2006).† The definition of a bank’s significant exposures is provided in Module 208.‡ Edward Altman and Anthony Saunders, ”Credit risk measurement: Developments over the last 20 years“ Journal of Banking and Finance, 21, pages 1721-1742 (1998).§ Anthony Saunders and Marda Cornett, Financial Institutions Management: A Risk Management Approach, 7th Edition, Chapter 12 (McGraw Hill, 2011).


DO NOT COPY

76

RETAIL BANKING II


(b) Obtain the historical average loan loss amount expressed as a ratio of loans issued (LGD) to this sector.

(c) Concentration limit, which is the maximum loan exposure for a particular SME as a percentage of capital, is obtained by the relationship:

CapitalAppetiteRisk

CapitalLGDSectorAverage

CapitalLimitionConcentrat

��

We illustrate by an example:

Senior management of a retail bank place a maximum loss of one percent on single loan losses expressed as a ratio of capital. Historically, this sector incurred an average loan loss for this sector expressed as a ratio of the loan amount of 40 percent. From the formula we obtain that:

(Concentration Limit for one borrower as a percentage of capital) x 40% ≤ 1%.

Equivalently, the concentration limit ≤ 1% / 40% = 2.5%. This means that the maximum exposure for a single SME is 2.5 percent of capital.

The approach described above shows that bank management can set concentration limits for each borrower in such a manner that establishes alignment with the bank’s risk appetite. But this does not account for possible sectoral concentration. Limiting the loan amount for each borrower does not necessarily increase the diversification of the loan portfolio. There can still be a large number of borrowers in the same sector. Hence, it is advisable that the bank set maximum limits on loan amounts for each sector of the economy.

While the models for sectoral diversification can be quite complex and beyond the scope of this module, it is important to highlight two important conclusions on sectoral concentration risk that have emerged from academic research. These are:

a) Sector concentration risk is the main contributor to economic capital for loan portfolios of all sizes. (Burton et al*). This has important implications for RAROC of a loan portfolio in that, all else being equal, a higher allocation of economic capital reduces RAROC.

b) Sector concentration can generate significant default contagion with unexpectedly high loss rates for the lender. This is very important for measures of bank loan profitability.

Lesson

Some sectors in the economy are more sensitive to changes in the overall macroeconomic environment. An overly high concentration of loans in these sectors can create a high variability in loan loss rates with a potential for higher capital allocation for credit risk.

We now summarise the key results of this module on SME lending.

Summary

The module utilised the 5Cs model augmented with the specific purpose of the loan as the basis for establishing the creditworthiness of prospective SME borrowers. This process is commonly called ‘collateral-based lending’ so as to emphasise the collateral requirements imposed by the bank. We presented an SME lending risk scorecard that can be utilised by the loan officer during the screening process.

* S. Burton, A. Chomsisengphet, and E. Hadfield, “The Effects of Name and Sector Concentrations on the distribution of Losses for Portfolios of Large Wholesale Credit Exposures”, Presented at BCBS/Deutsche Bundesbank Journal of Credit Risk Conference (Elville: 18-19 November 2005).


DO NOT COPY


RETAIL BANKING II


This stage is typically followed by two sequential stages – contracting and monitoring. The contracting stage provides the mechanics of loan pricing. The price of the loan is adjusted by non-price terms and conditions so as to account for the specific risks of the borrower. These non-price terms and conditions such as collateral requirement, loan amount, loan maturity and loan covenants also help the bank in the monitoring of credit risk after the loan is approved.

This module also considered the credit risk evaluation procedure for unsecured loans. The bank-customer relationship was examined as was hard financial data, together with soft data that is obtained through this relationship. The special situation of SMEs that are new borrowers and have been in business for a short period of time was analysed. If the screening stage results in a decision to proceed with the loan application, then terms and conditions (contracting) are set and the monitoring of credit risk follows.

Chapter 3 dealt with the calculation of the expected loss with the realistic assumption that probability of default (PD) and loss given default (LGD) are correlated. When this correlation is not recognised, the expected loss is significantly under-estimated with the resultant over-estimation of the bank’s profitability.

We also considered the implications of potential concentration risk in the bank’s loan portfolio. This lack of diversification arises from over-lending to a particular obligor or sector or geographical region and can lead to correlated default rates and unexpectedly high credit risk for the bank.


DO NOT COPY

78

RETAIL BANKING II


Multiple Choice Questions

1. Some banks place a relatively high weight on whether a prospective borrower has shown a willingness to repay its loan obligations. In relation to the 5Cs model, this behaviour is mainly associated with which of the following?

a) Collateralb) Capacityc) Characterd) Capital

2. The SME borrower may be required to commit assets which serve as collateral for the loan. The maximum amount lent is typically reduced to a percentage of the estimated value of the collateral asset. Which one of the following collateral assets is likely to be subject to the greatest reduction?

a) Cash and cash equivalentsb) Inventoryc) Commercial real estated) SME investment in short-term government bonds

3. The current financial statements of an SME show the following information (all expressed in thousands of dollars)

EBITDA = 10,050Existing Interest Expense and Principal Repayments = 2,000Change in Working Capital Requirements = - 3,080Required Replacement Investment Expenditures = 6,210Projected Tax Payments = 100

The Free Cash Flow is equal to:

a) -1,340b) 4,820c) 6,210d) 2,840

4. According to the ‘Two Ways Out’ principle, loan officers should seek at least two distinct and independent ways by which the borrower may repay the loan. The second most common way out is for the prospective borrower to commit collateral assets. The first way out is directly based on which of the following?

a) Adequate personal guaranteesb) A high level of EBITDAc) An interest coverage ratio (ICR) that is higher than twod) Adequate and positive free cash flow


DO NOT COPY


RETAIL BANKING II


5. A key factor that a bank typically considers in its screening process is the ‘purpose’ of the loan. The following are reasons why an SME may seek a bank loan. Which one of these reasons would a bank find least compelling to grant a loan?

a) Improve short-term liquidity by seeking working capital loans or an increase in the limits of lines of credit.b) Replace inefficient or depreciated assets such as equipment.c) Invest in new projects for higher profitability, market share or increased net cash flow.d) Pay dividends to the principals of the SME.

6. A rationale for the bank requiring collateral is based on what is commonly called the observed-risk hypothesis. Which of the following is not covered by this hypothesis?

a) Banks charge riskier borrowers higher loan rates.b) Banks require higher collateral from riskier borrowers.c) The provision of collateral by the riskier borrower reduces adverse selection risk for the bank.d) There is a positive relationship between risky borrowers and the requirement for collateral.

7. An SME is a seeking a loan from a bank for the first time and the company has been in business for less than three years. Which of the following recommendations to the bank loan officer, in relation to unsecured loans, would be most appropriate for him/her to take into consideration so as to reduce unexpected credit loss?

a) Banks should normally refrain from making unsecured loans to this category of new SME.b) Banks should consider making unsecured loans to this category of new SME.c) Banks may consider focusing on making unsecured loans only to this category of new SME.d) Banks should never make unsecured loans to new SME loan applicants.

8.Which statement is incorrect?

a) Banks may manage the markup for optionality risk by limiting loan maturity so that interest rate volatility is expected to be lower.b) Banks can issue loans with a shorter time to maturity to reduce the markup for liquidity risk.c) Banks can include in the loan contract effective non-financial covenants which provide restrictions and early warning signals for the bank and thereby reduce the markup for credit risk.d) Banks can issue loans with relatively long maturity to minimise the effects of potential adverse economic events and thereby reduce the markup for credit risk.

9. Consider the following example:

Senior management of a retail bank places a maximum loss of two percent on loan losses expressed as a ratio of capital in the consumer services sector. Historically, this sector incurred an average loan loss expressed as a ratio of total loans of 40 percent. Which of the following statements is incorrect?

a) The maximum exposure for a single SME as a percentage of total loan capital for the sector is five percent.b) The bank’s risk appetite expressed as percentage of capital is five percent.c) Limiting the loan amount for each borrower does not necessarily increase the diversification of the loan portfolio.d) The bank may have all loans respecting the single exposure limit but still have an unbalanced consumer services sector concentration.


DO NOT COPY

80

RETAIL BANKING II


10. Which of the following statements is incorrect?

a) A high value of Total Debt Service Ratio (TDSR) for an SME loan applicant indicates a relatively high level of capacity to repay the loan.b) Level 3 assets are the best form of collateral commitment by an SME.c) An example of concentration in the bank’s loan portfolio is when there are credit exposures to counterparties in the same economic sector or geographic region.d) Loan covenants in loan contracts provide warning signals to the bank prior to actual default by an obligor.

Answers

1 2 3 4 5 6 7 8 9 10

c b b d d c a d b b

CASE STUDY

Juan Perez is the sole owner of a restaurant specialising in Spanish cuisine, La Capilla, which has been in business for the last eight years. The restaurant is located in the central business district of Madrid and is rated very highly by customers on social networks as well as on travel websites and magazines. Last year the restaurant received its first Michelin star for superior quality and customer service. Perez recently hired a business consultant to advise him on his strategy to open a new restaurant in Barcelona using a similar business model to the one in Madrid. The current loan, with a principal of €600,000, has a fixed interest rate of 10 percent and matures in 15 years when the full principal is due. It was made by Banco Provincial, which financed the opening of La Capilla five years ago. The business plan requires an investment of €300,000 that is expected to be financed by a new fixed rate loan for a period of 10 years. The structure of repayments is expected to be similar to the existing loan – that is, quarterly interest payments with the full principal payable at the maturity date.

Perez met with his loan officer, Pablo Morales, and presented the following financial statements for La Capilla.

Income Statement for La Capilla (Full Year ending 31 December 2013; monetary values are stated in thousands of euros)

Revenues 3,030

Operating Expenses 1,810

Interest Expense 60

Depreciation and Amortisation 80

Taxes (20%) 232

Net Earnings 928


DO NOT COPY


RETAIL BANKING II


Balance Sheet for La Capilla (Full Year ending 31 December 2013; monetary values are stated in thousands of euros)

Assets Liabilities

Cash and Cash Equivalents 487 Accounts Payable 420

Restaurant Equipment (net) 1,242 Taxes payable 80

Inventory (Food, Wine & Alcohol and Supplies)

500 Current Liabilities 500

Current Assets 2,229 Bank Loan 600

Intangible Assets 350 Total Liabilities 1,100

Total Assets 2,579 Shareholder Equity 1,479

Morales notes that the current working capital (i.e., current assets – current liabilities) is €1,629,000, which declined from last year by a value of €65,000. Investment made in 2013 to replace equipment and furnishings in the restaurant was €30,000.

Pablo Morales is also presented with quarterly forecasts of La Capilla’s EBITDA over the next three years. Morales assumes that, for the next three years, La Capilla will fund the €300,000 loan and will also provide for the increased operating expenses as the new business seeks to create financial viability. Morales knows that Perez did not miss any interest payment on his existing loan over the last five years and that his periodic review of the company’s financial statements shows a profitable company with comparatively healthy balance sheet. But expanding in current macroeconomic conditions in Spain requires caution and for this reason his bank has adopted a strictly low tolerance for risk in the restaurant sector. Banco Provincial has a lending policy that requires a minimum interest coverage ratio (ICR) of 8.0 for SMEs in the restaurant sector.

Quarterly Forecasts of EBITDA over a 3-Year Period from 1 January 2014 (Consolidation of financial statements of both restaurants; all monetary values are stated in thousands of euros)

End of Quarter 1 2 3 4 5 6 7 8 9 10 11 12

Revenues 760 768 775 783 814 847 881 916 962 1000 1060 1113

Operating Expenses 480 528 580 639 652 665 678 692 706 720 734 749

Depreciation & Amortisation

40 40 40 40 40 40 40 40 40 40 40 40

EBITDA 320 280 235 184 202 222 243 264 296 320 366 405

Max TOTAL Interest Expense for ICR ≥8

40 35 29 23 25 28 30 33 37 40 46 51

Interest Expense on Current Loan

15 15 15 15 15 15 15 15 15 15 15 15

Max Interest Expense for New Loan

25 20 14 8 10 13 15 18 22 25 31 36

Note: It is assumed that a new restaurant will take one year to build up market share, thereby contributing marginally to revenues but adding a significant level of operating costs to the consolidated income statement. The following assumptions are made in forecasting EBITDA:


DO NOT COPY

82

RETAIL BANKING II


a) Revenues will increase by 1% each quarter in year one, 4% each quarter in year two and 5% in each quarter in year three.

b) Operating expenses will increase by 10% each quarter in year one, 2% in each quarter in year two and year three.

c) Depreciation and Amortisation will double in value and remain constant over the next three years.

Banco Provincial offers unsecured loans at a fixed annual rate of 12% to loan applicants in the high-end restaurant sector for a term of 10 years and for a maximum loan size of €500,000. For a loan amount of €300,000, the quarterly interest expense is €9,000.

Multiple Choice Questions

1. The EBITDA (stated in thousands of euros) for the full year ending 31 December 2013 is

a) 1,118b) 894.40c) 1,300d) 1,056

2. With respect to the bank’s lending policy, which one of the following statements is incorrect?

a) The current value of the interest coverage ratio (ICR) violates the bank’s lending policy for SMEs in the restaurant business.b) The current value of ICR is greater than the minimum value required by Banco Provincial.c) Constraints set in the bank’s lending policy are likely to mitigate concentration risk in the bank’s SME loan portfolio.d) The minimum value for ICR is likely a reflection of the bank’s tolerance for credit risk in the restaurant sector.

3. With respect to CAPITAL in the 5C model, Banco Provincial deems a debt-to-equity higher than 2 to be high risk while a value lower than 1 is low risk. Values between 1 and 2 inclusive would be in the range described as medium risk. Which one of the following statements is correct?

a) La Capilla’s current debt-to equity ratio would be placed in the low risk category.b) The size of the outstanding bank loan is less than the current value of shareholder equity and is classified as medium risk.c) Large loans to SMEs are typically high risk.d) The current value of the debt-to-equity ratio shows that Perez has a relatively low equity stake in the company.

4. Which one of the following affects net earnings but not cash flow for La Capilla?

a) Interest expense on the bank loanb) Depreciation of restaurant equipment c) Taxes payable d) Revenues

5. Which one of the following statements is correct in relation to free cash flow at the end of 2013?

a) The change in working capital at the end of 2013 relative to the previous year was €65,000.b) The required replacement investment of €30,000 in 2013 will increase La Capilla’s free cash flow.


DO NOT COPY


RETAIL BANKING II


c) The value of the free cash flow at the end of 2013 was €913,000.d) Free cash flow is always higher than net earnings in the restaurant sector

6. Which one of the following statements is correct?

a) Quarterly forecasts of EBITDA show a constant decline over the next three years.b) Quarterly growth rates of revenues are higher than the quarterly growth rates in operating expenses over the three-year period.c) In creating the quarterly forecasts of EBITDA, it is assumed that depreciation and amortisation expense varies over time.d) Quarterly forecasts of EBITDA decline in the first year since La Capilla funds the increased operating expenses of the new restaurant.

7. Which statement is correct?

a) If the new loan is granted according to the terms of the bank’s lending policy for unsecured loans, the minimum value of ICR will be breached in three of the four quarters in the first year.b) If Morales increases the term of the loan, the bank will incur less risk to the repayment of interest.c) The ICR constraint will be breached in quarter 3 of year 1 only.d) This loan will require a repayment of interest and a portion of the principal.

8. Morales is considering the following actions to reduce credit risk associated with repayment of the principal of each loan at the maturity date:

a) Seek personal guarantees from Perez.b) Reduce the new loan size, which will also reduce the interest payment.c) Require collateral from La Capilla in the form of cash and cash equivalents.d) Extend the term of the new loan to match that of the existing loan.

Which actions will not meet his objective?

I: a), b) and c) onlyII: a) only III: d) only IV: c) and d) only

Answers

1 2 3 4 5 6 7 8

c b a b a d c III


SME Lending Y - RBA · risk management department of the bank. Risk management experts evaluate the credit risk of the loan applicant typically using a three-stage lending process

Documents