Credit Liability Management in UK banks

Chapter 1: Introduction

Introduction

The purpose of this dissertation is to examine credit liability in UK banks and the surrounding issues in risk taking behavior and credit liability. To analyze these issues, the researcher examines as much as 30 developed and developing countries in Chapter 2. The central theme in examining the first two topics is the risk-taking behavior of banks., and it is one of the core areas of Basel I (signed in 1988) that focuses on the “measurement and management of risk-taking in commercial banks.” This agreement has emerged as one of the most successful international banking accords of the 1990s. Based on the success of Basel I and to remedy some of the limitations of Basel I, the Basel Committee initiated Basel II in 2004; it will soon be implemented by the G10 and several other countries. Basel II specifically outlines the three pillars of modern banking: capital regulations, depositor discipline, and regulatory supervision. The chapters of this dissertation are directly related to two of the most important areas of contemporary banking regulations under Basel II, and the findings herein may contribute to better policy-making, especially in the context of developing countries. “If you are having trouble writing your academic work and are looking for a website  that provides top-quality academic writing services, then look no further and feel free to contact us. At Academic Master, you will get your work not only of top-quality buy also at reasonable price that will not burden your pocket.”

The Concept of Liability Management

Liability management in general refers to the use of purchased funds to meet the investment needs of the bank. This departs from the traditional concept of limiting investment to the bank’s 1evel of exogenously determined deposits. While liability management can be defined in terms of purchased Funds, Kane (1979) discusses the concept in terms of dichotomy. He makes a distinction between the uses of liabi1ity management. Specifically, Kane defines the concept of supplementing asset management with very short borrowing as LM-l. This basically is reserve-position liability management. Kane then defines LM-2 as the management of all liabilities or generalized or loan-position liability management. This study utilizes Kane’s definitions. Following Kane’s discussion, more complete definitions of LM-l and LM-2 can be given. A general goal of bank management is to maintain or increase the level of assets. Liability management provides more flexibility in doing this. Banks have always maintained liquid assets to meet reserve needs and to adjust for deposit inflows and outflows. Of necessity, these liquid investments have not been the most profitable alternatives to the bank. The introduction of LM-2 allows the bank to hold fewer low-yielding liquid assets and to shift to higher earning less liquid assets. LM-l is used to support any additional liquidity needs via purchases of short-term funds. Thus, the primary benefit of LM-l is compositional as the bank adjusts its investment portfolio.

Causes of LM-2

The growth of LM-2 has been of necessity due to increased loan demand and fewer traditional deposits. As with all business firms seeking a profit, commercial banks have reacted with innovation. The conditions leading to innovation have been primarily of a regulatory nature. That is, faced with interest—rate ceilings, banks have sought to develop new sources of funds to offset the effects of disintermediation. The second major regulatory burden is reserve requirements. Reserves are usually non-interest earning and have the effect of raising the true cost of funds to the bank. Liability management banks have attempted to circumvent reserve requirements by shifting funds from high-reserve to lower-reserve sources whenever possible. Quite naturally, banks have the incentive to seek funds from lower- reserve sources.

Long-Term Profits and Customer Relationships

One question that is often considered in liability management is as to why would banks wish to continue making loans and seeking additional funds in constraints. Kane and Malkiel(1965) provide insight into this question. Banks are concerned with their long-term profitability and thus wish to retain certain large customers. As Kane and Malkiel indicate, there is a customer relationship which must be maintained to maximize long-term profits. Specifically, the bank has certain L* customers who have certain desirable banking characteristics such as a long deposit relationship, a Favorable pattern of borrowing, a history of stable deposits, etc- The L* customer is desired by all banks. However, the bank with the account has even more to gain by retaining the account. Due to the relationship, the bank‘s costs of servicing an L* account are less than for new customers. The loss of the L* customer will cause the bank to 7 suffer disutility. Of course, during periods of monetary restraint, granting a loan request to an L* customer will also cause the bank to suffer disutility but not as great as if the account were lost. Basically, the bank can be thought of as experiencing short-term pain for long-term gain. Hendershott and Winder have presented evidence that commercial banks have been more accommodating to customers with the advent of liability management. While liability management has allowed commercial banks to continue to accommodate their best customers and to hold fewer liquid assets, there is a question about the riskiness of the practice. Beebe (1977) addresses the issue of the risk incurred with the move toward liability management. He contends that the decline in asset liquidity, the growth in liability management, and the decline in bank capital have all contributed to greater risk exposure for commercial banks.

Asset Liability Risk Assessment:

Various techniques are used to assess the risk associated with assets and liabilities. Gap analysis and duration analysis are most common of them. Banks use these techniques to manage gaps and match durations of assets and liabilities. These techniques are based on the assumption of fixed cash flows. But the problem of variable cash flows is not addressed by these techniques. Therefore, these techniques are not considered suitable for bonds with call option, floating rate bonds, credits with prepayment options, etc. Although, duration analysis, in theory, address the problem of variable cash flows, but it requires sophistication, and thus becomes problematic in practice. Scenario analysis is a technique that addresses the problem of variable cash flow. Some conditions are assumed under this technique, such as:

• Different interest rate scenarios are assumed for a period such as 5 to 10 years.
• Performance of assets and liabilities are assumed, such as prepayment rate.
• Assumptions about inflation and economic cycle can also be made.

After some assumptions are made, different scenarios of balance sheet performance are projected. If the performance is substandard, then the adjustments are made in assets or liabilities to address the issue. Take an example of loan sanctioning for a corporate client. The bank will take into account various factors like credit history of client, operating and financial efficiency of the project to be undertaken, demand for the product or service, business and economic conditions, etc. After considering all the factors, the bank will grade the case and allow or reject the sanctioning of loan. If allowed, the interest rate charge will include a premium for the risk of lending to this specific client. An important limitation of scenario analysis is that scenarios projected in this technique are dependent on some assumed conditions. The actual conditions may vary significantly from the assumed scenario. With the passage of time banks and other types of firms recognized that there are various kinds of risks associated with assets and liabilities which are ignored by the old asset liability management techniques. For example, interest rate risk, liquidity risk credit risk, equity risk, exchange rate risk, sovereign risk etc. Banks were required to plan more actively for the balance sheet to minimize these risks. Recognition of these types of risks gave new and important dimension to the asset liability management in the financial intermediation. Today banks not only use asset liability management techniques for risk management and compliance purpose, but give also as strategic framework to achieve long term objectives.

Dissertation structure

Chapter 2 of the dissertation presents the liability management in banking sector in general which paves the way to examine risk taking behavior for credit among the banking sector. Chapter 3 examines the popular “credit crunch” literature. Evidence supports the existence of a “credit crunch”; Chiuri et al. (2001), for example, provide evidence of a “credit crunch” in 12 developing countries. Central to the related literature are examinations of whether the adoption and implementation of capital requirements curtails the credit supplies of banks or not. The present study examines the banking sector of UK as well as from the developed and developing countries. Most studies to date have examined this relationship on a country-by-country basis, with a special focus on the 10 OECD countries.

Research Method

In order to examine risk-taking behavior and competitiveness, we use the two most widely accepted models to measure changes in bank risk-taking (competitiveness) before and after 1997, namely the Shaffer (1993) model and the Panzer and Rosse (1982, 1987) methodology. This combination is rarely used in the existing literature. As a final step in this study, we examine the link between depositor discipline and bank risk-taking (competitiveness), in light of Gruben et al. (2003).

CHAPTER II: CREDIT LIABILITY MANAGEMENT

The Theory of Liability Management

Mason (1979) provides an excellent presentation of the theory of liability management. His analysis focuses upon a two-asset two- liability situation. The assets are elastically-supplied investments in-elastically-supplied loans. Mason dichotomizes liabilities into what he refers to as Type-I and Type-II deposits. Type I deposits are a combination of demand deposits and time and passbook savings accounts, with the supply assumed to be quite interest inelastic. Type-II deposits are CDs or other liabilities which are responsive to interest rates. Hence, the marginal opportunity cost of not investing in loans is given by the marginal revenue of government securities. The optimal portfolio of assets will then be at the point where the marginal revenue from investing in loans is equal to the marginal opportunity cost of investing in government securities.

Instruments of Liability Management

Negotiable Certificates of Deposits (CDs) are nothing more than receipts for funds that have been deposited at the issuing bank. The certificates specify the details of the deposit such as the maturity date, the rate of interest, and the amount of the deposit. The rates on CDs are market determined with the yield on newly issued CDs called the primary CD rate and the rate on outstanding CDs termed the secondary rate, while domestic CDs may be issued in either bearer or registered Form, most negotiable CDs are issued in bearer form to facilitate transactions in the secondary market. Domestic CDs are typically paid for in immediately available funds and redeemed for immediately available Funds. This section reviews the literature on bank liability management. While some of the papers examined do not deal with liability management per se, they are important for the information they provide concerning bank behavior. The literature dealing with liability management is quite limited; the most relevant studies are couched within the broader area of bank balance-sheet behavior, where much has been done in the past twenty years. The early balance-sheet studies usually involve the selection of a single asset or the selection of the set of assets to hold. They typically assume that the liability side of the balance sheet is beyond the control of bank management and is taken as given. Of course, prior to the l960s, this was generally the case as banks relied almost entirely on traditional deposits to support their asset investment. Despite the development of liability management in the 1960s, it was not really until the early l970s that explicit consideration was made of the portfolio of liabilities to hold. The literature on balance-sheet behavior can be broken into two categories. The first category of studies involves theoretical considerations involved in the selection of assets and liabilities. Most of these normative works involve the development of mathematical models to determine optimal quantities of individual assets and liabilities to hold, and models to determine optimal balance-sheet positions. The second major category involves actual bank behavior. Again, most of the positive research considers only a limited number of assets and liabilities. The general failure of these efforts lies with two implicit assumptions. First, most of the studies assume that there is no relationship between quantities of assets and quantities of liabilities. Second, they consider only average behavior patterns over time. That is, no attention is focused on balance-sheet behavior during phases of the economic cycle. This section reviews the major normative works on liability management. Emphasis is placed upon the research most directly related to the subject at hand.

Normative Studies

Chambers and Charnes (1961) set the stage for the development of computerized models of balance-sheet management. Their development of a linear-programming (LP) model is based on maximization of a profit function subject to certain constraints. It is assumed that the levels of future demand and time deposits, rates of interest, and the bank’s net worth are known with certainty. The bank is faced with determining the optimal mix of specified assets-cash, government securities of various maturities, and other securities and loans which will maximize profit over five planning periods. The Chambers- Charnes model specifies two constraints. The first constraint insures that reserve requirements are met while the second focuses upon Federal Reserve examiners’ measures of capital adequacy and liquidity. Although the model of Chambers and Charnes appears simplistic by today’s standards, it is regarded as breaking the ground for future computerized optimization models. However, since the model considers only a single objective function, a small number of assets, and only five planning periods, it is limited in application. Cohen and Hammer (1972) seek to further the practical use of linear- programming techniques in bank funds management. Their study discusses the development of an LP model and its use in a New York City bank for asset management. In contrast to the Chambers and Charnes profit-maximization objective, the Cohen and Hammer model seeks to maximize the present value of net income. Their model also considers a wider range of constraints. Specifically, it incorporates constraints to account for risk, availability of funds, policy considerations, and market conditions. Input requirements are management forecasts of future interest rates, future loan demand, future deposit levels, and future levels of other exogenous variables. Uncertainty in these variables is not considered. This framework represents an improvement over Chambers and Charnes in two ways. First, a wider range of constraints is considered and second, the model was put into actual use. While the l960s and early l970s saw many studies involving variations of the LP method, drawbacks to the approach remain. Mckinnev (1977) points out some of the pitfalls in using mathematical- programming models for financial planning. First, LP requires the identification of an objective function to be maximized- while this would appear to be simple, a bank often has multiple objectives. Second, the technique requires the explicit specification of constraints. Some inter-temporal models do not consider the fact that constraints may change over time. Also, some constraints are difficult to state in precise terms. Often, there are qualitative values to be incorporated into the model. Even overcoming these problems, it is often difficult for bank management to translate the information from the model into practice. In addition, McKinney contends that since the risk-return trade off is not considered, a bank may actually prefer a lower risk-lower return position while an LP solution will give a single optimal position in terms of maximizing the objective function. McKinney concludes by recognizing the value of mathematical models, but warning that care must be taken in their application. Fortson and Dince (1977), recognizing the need to incorporate multiple objectives, present a model which explicitly accommodates multiple goals. Goal programming (GP) is a mathematical-programming method which considers multiple and even conflicting goals in arriving at an optimal specification for a set of choice variables. Specifically, the approach requires the specification of management desires (goals) and environmental conditions (constraints). Penalties are attached to deviations from the stated goals with the objective function being minimization of total penalties. Fortson and Dince consider a four-quarter planning horizon in the application of the model to a north Georgia bank. While the results confirm the viability of the model to balance-sheet planning, discrepancies are apparent. These were attributed to forecast errors pointing out the need to consider different planning horizons. The authors conclude that the model is workable and suggest that use of the model for simulation will allow management to consider various outcomes.   Eatman and Sealey (1979) advance further the incorporation of multiple objectives in mathematical-programming models. Their model specifically optimizes a set of objective functions subject to a set of constraints. Instead of yielding one optimal solution, their procedure yields a set of efficient solutions. From this set of efficient solutions, bank managers select the desired solution according to their implicit utility function. The authors point out three advantages of their approach. First, all necessary information for a feasible solution is included in the model. Second, decision- makers explicitly select the solution from the efficient set. Finally, the method can be accommodated in existing LP models with only minor modifications. Hagdon and wicks (1966) present a nonnative model incorporating several policy variables to determining the earning assets for a profit-maximizing commercial bank. They impose several restrictive assumptions. First, the bank is assumed to have a given amount of funds to lend. Second, the authors assume that the particular use of the funds has no bearing whatsoever on the availability or cost of Funds. The solution of their model involves selection of alternative earning assets which provide the highest value of Ex (defined to be the net rate of return in excess of the net rate on treasury bills), subject to regulatory restrictions. To account for risk, the value E_x is divided by A_f (the loss rate expected on the particular alternative). The selection procedure is simply a matter of ranking the earning assets and selecting from them until available funds are exhausted. Typical of the early studies, however, the bank is assumed to have no control over the liability side of the balance sheet. This study’s contribution is that it provides an alternative to the mathematical-programming models which dominate the normative literature. Francis (1978) demonstrates the application of the Markowitz portfolio Framework to the balance-sheet portfolio decision. Francis treats the choice set of variables for the balance sheet as being all assets and liabilities, with liabilities indexed as negative assets. Associated with each asset and liability is a rate of interest representing either return or cost. The risk is determined as the variability of the rate of interest, treating each of the choice variables just as alternative investments; Francis determines an efficient frontier by minimizing risk for each level of return. Each point on the frontier represents a portfolio of balance-sheet variables with the return on the portfolio defined to be the return on equity. He defines return on equity as a weighted average of rates of return on the five assets and three liabilities under consideration. The portfolio risk is defined as the variance of the return on equity. To demonstrate the use of the portfolio selection, Francis analyzes the portfolio management of small, medium, and large-size banks. He uses data on 58 large banks (over $200 million in total deposits), 198 medium banks (S50 to S200 million in total deposits), and 603 small banks (less than $50 million in total deposits). The sample period is 1966 to 1977. Efficient frontiers for each size category are developed with the actual portfolios then compared to the efficient portfolios. Francis determines an unconstrained efficient frontier and then an efficient frontier constrained to allow for reserve requirements. Of interest is the fact that the efficient frontier for small banks dominates that of large and medium banks for both the constrained and unconstrained portfolios. Francis explains that this is due to the higher rate large banks must pay to secure funds. While they do earn a higher return, the return on equity is below that of small banks. When comparing the actual portfolios to the efficient portfolios, he found that large banks were closer to their theoretical portfolios than were small banks. He concludes that large banks are more efficient in their portfolio management. To summarize, the normative studies have largely concentrated on the maximization of some objective. Such studies have failed to capture the relationship between the asset and liability sides of the balance sheet. They have been informative to academics and practitioners alike as work continues toward improving balance-sheet management.

Positive Studies

Brechling and Claxton (1965) consider three classes of bank assets in their presentation of the factors affecting the choice of assets by British commercial banks. They explicitly assume that the level of funds available for investment in assets is exogenous. Specifically, three assets are considered. To account for size, the values are divided by total deposits. The three ratios are liquid assets to deposits, investments to deposits, and advances to deposits. These are each regressed on four exogenous variables: the rate of interest on investments and advances, the level of economic activity, first differences in the rate of interest, and the level of liquid assets. Obviously, this early empirical work is a very simplified approach to the balance-sheet portfolio decision. Even the authors acknowledged that their effort represents nothing more than a first attempt to apply portfolio techniques to British commercial banks. Anderson and Burger (1969), in a l969 study, test two hypotheses regarding bank portfolio behavior. First, they examine the “accommodation principle” which implies that bank portfolio behavior is determined by the demand for bank loans. This is in contrast to the profit-maximization principle which implies that market forces determine bank portfolio behavior. The second hypothesis is that a significant change in portfolio behavior has occurred in the years most recent to their study. From the empirical model they develop and test, they conclude that banks attempt to manage their asset portfolios according to the profit-maximization principle. Of course, in light of the Kane and Malkiel (1965) concept of the L* customer, there may be no real distinction between the accommodation principle and the profit-maximization principle. That is, the demand for loans does affect bank behavior because their long-run profit- ability is enhanced by meeting the loan requests of their valued customers. Recognizing that the nature of banking was changing, the authors posit that banks had changed their behavior patterns in the early 1960s. The authors define this change as a move to a more “profit-oriented behavior,” They explain that banks appeared to base decisions more on costs and yields than in earlier times. To test for a change, three demand equations are estimated for the time periods from 1953 to 1960 and from 1951 to 1967, Chow tests for equality of coefficients indicate a significant difference. The hypothesis of a difference in behavior is supported as would be expected from what is now known. While the ‘study is limited in scope, it does provide evidence that liability management has introduced a new element into bank balance-sheet decisions. Hester and Pierce (1975), in a l975 work, provide a micro-econometric analysis of portfolio behavior by commercial and mutual savings banks. This comprehensive work develops a theoretical basis for bank behavior based on the institution’s objective function, set of activities, environmental restrictions, institutional frictions, and uncertainty. Although this work was completed in the mid-seventies, it was seven years in the works. The major difficulty with the study remains the decision to treat available funds as exogenous. While no empirical models in the study treat them, as endogenous, the authors do discuss the likely effects of such treatment. They posit that banks will adjust to deposit flows by adjusting purchased funds. This contrasts with their model where adjustments occur in securities and loans or at the discount window- with endogenous funds, asset selection will be less affected by traditional deposit histories. Further, they conclude that the decision to use purchased funds or liquid assets as adjustment factors will depend on future trends in a variety of interest rates. The major finding of their work is that a deposit shock has a measurable effect on an individual bank’s portfolio for about one year. For an aggregate of banks, Hester and Pierce hypothesize the lag to be greater than one year. Hester proposes a model of portfolio behavior by large commercial banks that is tested on a panel of 320 weekly reporting banks from 1969 to 1974. The model is a banking sector model only- Exogenous variables are demand deposits, time deposits (excluding negotiable certificates of deposit), bank equity, miscellaneous illiquid assets and liabilities, and all interest rates. To account for changes in the regulatory environment, a dummy variable also is included. Endogenous variables include twelve asset categories, three liability categories, and two interest rates. Each endogenous variable is treated as a function of the same set of exogenous variables, The basic hypothesis of the study is that a change in an exogenous variable will affect assets and liabilities in predictably different ways. The adjustments are expected to be determined by regulation, adjustment costs (implicit and explicit), and net rates of return on the various assets and liabilities. For actual estimation, all quantities are deflated by the bank‘s total assets as a means of correcting for heteroscedasticity. From the estimation, Hester reports only on ten assets and liabilities. During the period under study, banks were acquiring net Fed Funds and CDs which were intermediated into cash assets, commercial and industrial loans, financial loans, and securities. The major “shock absorber” during the period appears to be Fed Funds as opposed to short-term government securities. The one-year lag incorporated into the model does not appear to be long enough to completely adjust the balance sheet to exogenous demand-deposit changes. The lag structure includes 17 demand-deposit variables, 17 time-deposit variables, and 10 CD variables. Hester concludes that this is not sufficient to capture the adjustment process. The results also indicate that the one-year adjustment period for time deposits is too short. However, time deposits do appear to be close substitutes for new acquisitions of CDs. Lagged changes in CDs cause an initial reduction in net Fed Funds but have no significant effect over time. In equilibrium, 42% of CDs are in business loans, 23% in mortgage loans, and l6% in other securities. The estimated interest rate coefficients tend to be statistically insignificant. Hester attributes this lack of explanatory power to three potential shortcomings: the absence of transactions costs, misspecification. The Hester and Pierce study found short-term governments to be the buffer. Hester explains the difference as due to differences in the sample and increased Fed Funds activity of the hypothesis, and lack of correspondence between rates and quantities. Using aggregate data, Hester reestimates the model with better results. He finds greater significance of interest rates and an overall better fit. In summary, he concludes that portfolio variations can best be explained by deposit histories. His findings also show that banks seek to maximize the present value of the customer relationship. A drawback to Hester’s study, however, is the failure to consider behavior over particular economic cycles. His results reflect average behavior over a period of time as opposed to specific responses. Brainard and Tobin (1979), in what has become the definitive article on financial model-building, present theoretical arguments to support explicit consideration of the interdependencies of markets in both theoretical and empirical models of financial behavior. To support their arguments, a fictitious economy is con- structured with various behavior patterns determined through simulation. Specifically, Brainard and Tobin argue that balance-sheet models should follow several rules of consistency. First, the set of interest rates associates with a particular asset-demand or liability-supply equation should appear in all asset-demand or liability-supply equations. This is due to the fact that the net effect of an interest-rate change, over the entire balance sheet, must be zero- Failure to include a rate across all equations implicitly assumes that the offsetting effect of a change comes in some unspecified equation. Likewise, the same holds true for exogenous balance-sheet items- Here, there is an effect across all equations of unity. In summary, Brainerd and Tobin contend that the same set of regressors must appear in all asset-demand and liability-supply equations. Equations for all choice variables must be specified. All exogenous balance-sheet items must be included in all equations. All column-sum coefficients on the interest rates must be zero. The contribution of Brainard and Tobin has been in providing a theoretical underpinning for financial models. Most of the studies since their paper have incorporated their recommendations. Hendershott and winder (1979) examine the impact of liability management on the availability of funds to meet the loan demand of bank customers. They attempt to determine whether the growth of markets for purchased funds has allowed banks to become more accommodating to their customers. They consider five endogenous assets categories: business loans, consumer credit, home mortgages, other mortgages, and a net-liquidity variable. The net-liquidity variable is defined as the difference between U. S, Government securities and reserve-adjusted purchased funds (large CDs, net Fed Funds, repurchase agreements, open market paper, and borrowings from the Federal Reserve). These endogenous variables are hypothesized to be a function of current and lagged exogenous net sources of funds and extraordinary demand for funds by businesses and households. Extraordinary demand is defined as the difference between a sector’; current demand for funds and its long-run normal demand. Using quarterly aggregate data for the period From 1961 to 1977, the mode1 is estimated. To test for the impact of 1iabi1ity management, two subsets are uti1ized. The first, from 1961 to 1970, captures the period when the markets for purchased funds were deve1oping. The second period, from 1971 to 1977, ref1ects the period when the markets were well deve1oped. A two-stage process is fo11owed in estimation. First, seeming1y- unre1ated regression (SUR) is used to determine unconstrained co-efficient. These estimates are then examined as to sign, size, and significance to determine  inc1usion in the Final model. The Final model is then estimated using SUR. For the 1961 to 1970 period, the authors conc1ude that banks are accommodative to the short-term needs of business firms and to the extraordinary demand for funds by households. Further, about four-fifths of demand deposits are channeled into business loans, with about 15% going into commercial and multi-family mortgages. Small time and savings deposits are spread over all uses with little substitution found among assets. The results of the 1971 to 1977 estimates support the hypothesis that banks are more accommodating with liability-management alternatives. Further, households are accommodated to a greater degree than businesses. To summarize, Hendershott and Winder find evidence of a shift in the asset portfolio of commercial banks from 1iquid assets to higher earning 1ess liquid assets. They conc1ude that banks are better ab1e to meet the needs of both business and consumers. Moreover, during their second sub-period, there is no evidence of credit rationing at the expense of households relative to the needs of business. Over the long run, the authors conclude that households are the greatest beneficiaries of liability management. The Aigner (1973) study estimates a system of reduced-form demand equations for the short-term assets and liabilities of an individual bank. The major contribution of the study is in the econometric specification which generalizes an earlier substitution model to incorporate complementary results. That is, the system explicitly considers complementary relationships among the variables. Berndt, McCurdy, and Rose (1980) in a study, reexamine the Parkin model. Specifically, they express concern about autocorrelation which was implicitly assumed not to exist in the original Parkin paper. The authors conclude that their adjusted model provides more plausible results leading them to question Parkin’s interest-rate response estimates. The Humphrey (1977) study examines commercial bank liability management in a mu1ti·asset/multi-liability framework for the period 1970 to 1975. It is among the first to simultaneously consider various liability categories and explicitly consider complement- substitute relationships. Specifically, three separate hypotheses are tested:

1. the aggregation of Fed Funds, demand deposits, time deposits, and CDs into a composite debt measure affects statistical analyses of behavior.
2. each liability is intermediated into liquid assets just as into non-liquid assets.
3. the own-price elasticities are zero for liability demand and asset supply.

The set of variables includes five classes of liabilities and two classes of assets. Liability categories are Fed Funds, domestic demand deposits, domestic time and savings deposits, negotiable CDs, and bank equity capital. The study assumes all returns on assets and rates on liabilities to be endogenous with the exception of Fed Funds. The process of liability management (LM) can be viewed as originating with an increase in loan demand measured by non-liquid assets A_n, which leads to an increase in purchased funds, Hence, the intermediation elasticities 9_CD’D and 9_F,N are positive. To some extent Humphrey concludes that the traditional-adjustment sequence and the LM-adjustment sequence are effective substitutes. While the evidence is weak, he further suggests that the LM- adjustment process may be more important, at least during the period of his study. Humphrey experiments further by estimating the model for individual banks but there are no changes in his conclusions. To summarize, Humphrey draws three conclusions: Cl) the use of an aggregate debt measure is inappropriate for statistical analysis of bank behavior, (2) liabilities are intermediated into liquid and nonliquid assets in different ways, and (3) the hypothesis of zero own—price elasticities is rejected. Spindt and Tarhan (1980) provide the most recent study of LM behavior. They statistically evaluate the dynamic effects of variation in externally—determined conditions on the liquidity structure of large money—market banks. Their empirical model considers the extent to which New York City banks utilize traditional liquid-asset adjustment as opposed to LM adjustment. They identify U.S. Treasury bills, loans to brokers and securities dealers, and other short-term securities as liquidity assets. Liquidity liabilities are identified as net fed funds, borrowings from Federal Reserve Banks, Eurodollars, and large negotiable CDs, These items constitute the set of endogenous variables. The bank is assumed to be a price-taker with respect to all liquidity instruments, Their original specification considers specific rates with each of the liquidity instruments- However, due to multicollinearity problems, the authors confess to “mining” the data to arrive at their final model, which includes only four money—market rates: the Fed Funds rate, the 9l-day Treasury bill rate, the 2-3 month Eurodollar rate, and the 3 month CD rate. Spindt and Tarhan discuss several possible explanations, but conclude that the role of interest rates in bank portfolio decisions requires further research. Results on the forecasted stock of exogenous balance—sheet items are much more clear with signs generally conforming to expectations. That is, an increase in the expected stock of exogenous asset items leads to a decrease in ex ante liquid asset stocks and increases in ex ante purchased liability stocks. Expected increases in exogenous liabilities induce ex ante decreases in stocks of purchased funds and ex ante increases in liquid assets. The authors conclude that the initial impact of changes in expected exogenous stocks is on net Fed Funds and loans to securities dealers. They further suggest that these items now serve as a buffer to deposit stocks replacing the traditional use of excess reserves. The estimated systems dynamics matrix suggests that the set of liability instruments are substitutes while the liquid assets appear complementary. Examining the off- diagonal row elements relating to net Fed Funds and loans to securities dealers, they have the largest coefficients- However, the column values corresponding to lagged quantities of the two instruments are smaller suggesting that while changes in the lagged liquidity structure impact on net Fed Funds and loans to securities dealers, changes in the two categories themselves have little effect on liquidity structure. The authors conclude that there is no fixed adjustment ordering among liquidity instruments and that the greatest impact of shifts in loan demand and deposit flows initially falls on the Fed Funds market. This section has presented a survey of the early work on bank balance-sheet behavior, beginning with normative studies and ending with empirical ones. The articles were presented in order of their relevance to the work at hand. The recent works have focused upon the balance sheet in a simultaneous framework considering both the asset and liability sides. Although these studies represent a significant advancement over the studies of the l960s, they still fail to capture the nature of bank behavior- One drawback is the use of a single time period for estimation. Such a procedure yields only average coefficients representing average responses over a time period. Since liability management is used as a tool to stabilize funds variability and to expand the level of investible funds, it would seem that liability-management behavior might differ according to the stage of the business cycle. The main distinctive features of loan commitments which attract economists’ interest are option-like exercise and multiple fee structure. Thus, many studies focus on explaining why loan commitments exist and how they are priced. Morgan (1993) uses a model based on contract theory to see how incentive problems between lender and borrower affect the fee structure of loan commitments. His model predicts that the spread over the safe rate will increase with the size of commitment loan and the least-known firms with high monitoring cost benefit the most from using band loan commitments. However, the reality is that a fixed spread for all sized loans is charged and relatively well-known firms use loan commitments more heavily. He leaves an open question about why the small firms which would benefit most from loan commitments use less of it. Shockley and Thakor (1997) scrutinize a micro data to make testable predictions that pricing structure of loan commitment is more complicated and delicate to firms whose assets are harder to value and whose credit quality is poor. Also, they predict the announcement of a loan commitment purchase should be greeted with an abnormal positive price reaction since obtaining a stable source of funds is regarded as good news in the markets. Both predictions are confirmed in their empirical test. As to credit risk from using loan commitments, Boot and Thakor (1991) formally establish that loan commitments may reduce banks’ asset risk. And Avery and Berger (1991) show that commitment loans tend to have slightly better than average performance, suggesting that either commitment generate little risk or that this risk is offset by the selection of safer borrowers to receive commitments. Some may point out that loan commitment is rather discretionary due to MAC (Material Advance Change) clauses, hence adding no credit or liquidity risk. Banks can get out of their obligation under certain circumstances, attesting MAC clauses. Nevertheless, in reality, it appears that courts have often obstructed banks, right to invoke MAC clauses and deny credit to a loan commitment owner, arguing that the banks had not acted in good faith (Edelstein, 1991). Additionally, if a bank concerns about its reputation by exercising MAC clauses and its long-term effects, invoking the clauses is costly. Therefore, being with MAC clauses does not entitle banks to full freedom of leaving obligations from loan commitments at any time. Notwithstanding, it cannot be said that loan commitment is an unconditional obligation to firms all the time. An empirical study by Sufi (2005) shows that banks, recognizing that the line of credit can be prone to abusive use by corporate borrowers, closely monitor and extend the line of credit through covenants on firms’ profitability. In this regard, it seems obnoxious to think that loan commitment lies between unconditional obligation of funding and term loans without commitment. In terms of pricing loan commitments, a recent study by Chava (2004) presents interesting results. Using pricing model and Dealscan database, he shows that banks offer loan commitments (on a stand-alone basis) to firms below the fair price and the extent of underpricing is significantly and positively related to the length of the relationship between the bank and the borrower. The empirical fact that loan commitments are widely used despite its negative price on a stand-alone basis implies the existence of informational value accruing from mutual relationships between lenders and borrowers. At macro level, Morgan (1998) reports that bank loans not made under commitment show after tight policy, while loans under commitment accelerate or are not changed. In a similar context, Sofianos et al. (1990) use VAR and find an evidence of a smaller impact of monetary policy on loans made under commitment agreements than on loans not made under commitment. They conclude loan commitments effectively insulate borrowers from the effect of credit rationing and force monetary policy to work exclusively through the interest channel. Kashyap et al. (2002) ask why the dominant proportions of loan commitments are issued by commercial banks, not by other types of financial intermediaries. Their key insight is that deposit withdrawal and loan commitment take-down are essentially the same in terms of liquidity provision. Hence, if those two liquidity demands are not perfectly correlated, dealing with two seemingly separate functions under one roof produces the synergy effect, allowing banks to hold less amount of buffer stock liquidity. Recent studies by Gatev and Strahan (2005) and Gatev at al. (2005) confirm this by showing that, during 1998 crisis, banks experiencing a large inflow of transaction deposits provide liquidity to firms using commercial back up lines. Also they present evidence that this deposit-lending synergy is more powerful during crisis, as nervous investors move their funds into banking sector.

Liability model

This section introduces a simple one-period model to highlight the determinants of using loan commitments. The simple case will show how the optimal amount of loan commitments is affected by the degree of adverse selection in capital  markets, the degree of uncertainty in borrows’ liquidity demand and other parameters. Next, an extended version allows a bank to use an option of recalling term loan. When the amount of liquidity held inside falls short of the realized take-down from loan commitments, a bank needs to make up for the shortfall. In this situation, the options open to a bank are (1) to get uninsured funds through external financing and (2) to reduce the amount of term loans to be issued. The relative importance of these two options depends on their relative marginal costs. If a bank’s external financing cost is higher compared to that of others, it will cit down more of its term loan if other things are being equal. This extended version will show that the ratio of  loan commitments to term loans will increase in bad times and this effect is stronger for small or stand-alone banks, which are more liquidity constrained compared to large banks.

Chapter III: Impact of Basel I-Like Bank Capital Requirements on Bank Credit Risk

In 1988, all the OECD countries, as well as Switzerland and Luxemburg, signed the Basel I Accord (capital adequacy regulations). It was implemented during a three-year period, from 1990 to 1993. During the 1990s and 2000s, the Accord emerged as the landmark document for bank capital regulations and supervision in about 100 countries, both developed and developing. Evaluation studies have shown that the regulation did increase the capital ratios of all banks in developed countries, including undercapitalized banks. These studies also found a weak negative relationship between capital ratios and the risk-taking of banks, indicating that the improvement of capital ratios of undercapitalized banks was not accompanied by a concomitant increase in risk. At present, there is a dearth of literature on the impact of such regulations in developing countries. Most of the previous studies are of a single-country nature and are not cross-sectional. Critics caution that developing countries have very different banking histories, structures and environments; those countries also differ from developed countries in their levels of government intervention in banking, degrees of governmental bank ownership, degrees of bank liberalization, levels of privatization vis-à-vis state-owned banks, degrees of regulatory forbearance, and in their institutional set-ups and other socio-economic factors that affect banking operations. Therefore, in the present study, we examine the reaction of banks in developing countries to Basel I-like capital regulations. Recent increases in the number of bank crises around the world have renewed the interest of professional policy makers and academia in identifying the determinants of such crises. Several studies on the determinants of bank crises have already identified regulatory changes as a major factor. It is instructive to assess the impact of capital regulations on bank risk-taking behavior and on the relationship between capital ratios and bank risk-taking. This is the primary focus of the present study. The literature begins, in earnest, with Shrieves and Dahl (1992), who use several periods of cross-sectional data on commercial banks in the United States, under the two equation (capital equation and risk equation) simultaneous equations framework. They found that the effectiveness of risk-based capital regulations depends on how well the regulations reflect the true risk exposure of banks. The results of studies by Aggarwal and Jacques (1997, 2001) on U.S. banks are not easy to interpret. Samples selected for these studies are from 1991-1993 and 19911996 respectively. Both of these samples coincide with the passage and implementation of the Federal Deposit Insurance Corporation Improvement Act (FDICIA) in December 1991. Prompt Corrective Action (PCA), contained in Section 131 of FDICIA, went one step further than the Basel I Accord by defining three regulatory ratios (the Basel capital standards plus an advantage requirement) and five categories in which banks are classified according to their compliance with the three ratios. The impacts of Basel I and PCA under FDICIA therefore overlapped for U.S. banks, and it is thus difficult to ascribe the findings of the two papers by Aggarwal and Jacques (1997, 2001) to the Basel I Accord alone. However, if one were interested in implementing such regulations with a concomitant hardening of the central bank’s behavior, then the results remain valid. Thus the authors find that undercapitalized banks increased their capital target ratios more quickly than banks that had higher initial capital. The study by Jacques and Nigro (1997) deals exclusively with the consequences of the Basel I Accord, as it concentrates on the years 1990-91, which is the period before the FDICIA was passed. Their finding was in line with those of Aggarwal and Jacques (1997, 2001); however, the problem inherent in this study is the very low number of undercapitalized institutions in Jacques and Nigro’s sample – less than 2 percent of the total number of banks, which may reduce the reliability of some of their estimates. Following studies present non-U.S. evidence regarding the relationship between capital ratios and credit risk. Ediz, Michael, and Perraudin (1998), employ confidential U.K. data, including detailed information about the balance sheets and profit and loss accounts of all British banks, from 1989 to 1995. It uses a limited-information technique that differs from that of the Shrieves and Dahl (1992) framework mentioned earlier, and they found evidence that capital regulations were effective in increasing the capital to meet the minimum standard. Unfortunately, Ediz et al.’s model leads to a puzzling conclusion: banks are adjusting their capital levels each year by more than the difference between the current level and the target they have in mind, which means that banks are overshooting their targets (and by more and more each year). The study by Rime (2001) is interesting, because it provides the first application of the Shrieves and Dahl (1992) model to non-U.S. banks during 1989-1996. His results indicate that Swiss banks reacted to capital regulations by increasing their capital, but that this did not change banks’ risk-taking behavior. One of the possible problems with this study is that Rime adopted the PCA regulatory classification to measure regulatory pressure on Swiss banks; this may be inappropriate, given that the additional requirements set out by PCA have not been formally adopted by any country other than the United States. Patrick Van Roy (2003) studies banks in seven G10 countries (Canada, France, Italy, Japan, Sweden, United Kingdom, and the United States) in a panel data set for the 1988-95 period, and achieved similar results. Godlewski (2004) studies the response of banks in 30 emerging market countries in Central and Eastern Europe, Asia, and South America to such regulations; his results corroborate the existing findings for banks in developed countries, and also show that regulatory, environmental, and legal milieu play important roles in bank capitalization and credit risk-taking behavior in emerging market economies. To assess the impact of deposit insurance and regulatory restrictiveness on the effectiveness of capital regulations, studies by Demirguc-Kunt et al. (1999, 2001) and Barth et al. (2000) are examined. Demirguc-Kunt et al. provide evidence that explicit deposit insurance tends to be detrimental to bank stability, especially when bank interest rates are deregulated and the institutional environment is weak. Evidence provided by Barth et al. (2000) indicates that a positive relationship exists between the degree of regulatory restrictiveness and banking sector fragility. We summarize the finding of the articles discussed in the review section, as follows: these articles generally support the idea that undercapitalized banks increased their capital adequacy ratios in the first half of the 1990s; a similar trend was observed for well-capitalized banks, but to a lesser extent. However, there is little consensus among the reviewed literature that banks, whether adequately capitalized or not, engaged in riskier activities because of changes to capital regulations. Based on the background presented above, in the present study, we analyze the following: (i) the change in actual capital ratios in response to regulatory minimum capital ratios of banks; (ii) the relationship between change in risk and change in capital ratios to examine if the increase in capital ratios came at the cost of higher risk-taking or not; (iii) impact of different important elements on the relationship between change in risk and change in capital, which include: liberalization ( foreign direct investment as a percentage of Gross Domestic Product(GDP); bank activities restrictiveness index ; different characteristics of deposit insurance schemes; and domestic credit as percentage of GDP.

CHAPTER IV: CREDIT EXPENSION

The G10 countries adopted and implemented the Basel I Accord in 1988 and 1990-1993, and central bank authorities in different countries around the world started to implement national versions of the Accord throughout the 1990s and early 2000s. The primary objective behind such a worldwide acceptance was to promote the soundness of the national financial system, to “catch up” with the international banking standards in the wake of financial liberalization and integration, and to foster economic growth with the help of a better-functioning financial system. During the 1990s and early 2000s, several concerns were raised by academia and financial practitioners around the world about several adverse impacts of the Basel I Accord, like capital regulations on the banking system of host countries in particular, and consequences in terms of the economy as a whole. Over the years, this debate took the form of two key questions (IMF, Survey): What is the reaction of banks’ behavior to such an Accord, which entails a capital charge of the risks they take? What is the impact of such changes in their behavior with respect to lending? The second question is central to what is now popularly called the “credit crunch” literature. The previous chapter of this dissertation primarily examined the impact of regulatory capital ratios on actual capital ratios and risk taking of banks. But in the present study, we examine a different and related field, namely, the impact of capital regulations (minimum capital requirements) from different point of views: (i) the lending behaviour of banks; (ii) several variables representing equity; loan; and profitability; and (iii) risk sensitivity of loans. If banks face a constraint of capital for the loans they want to make, they must raise new capital. Nonetheless, asymmetric information and “lemons” problems may prevent them from issuing new capital. Myers and Majluf (1984) point out that banks may prefer to shrink rather than issue new equities, due to asymmetric information and “lemons” problems. Holmtrom and Tirole (1997) provide evidence revealing the importance of capital as determinants of investment, monitoring, and interest rates, and the importance of capital in terms of macro-economic implications, especially with regard to banks. As a result of such difficulties associated with raising new capital and the concomitant importance of the role of capital in the decision-making of banks, banks may decrease lending in response to capital regulation. This is the central argument of the “credit crunch” literature. Notwithstanding the above argument and evidence on “credit crunch,” banks may respond positively to capital requirements and increase – rather than decrease – capital and bank lending. To date, a substantial body of literature has dealt with this issue. Recent work by Chami and Cosimano (2001) reveals that banks are more likely to expand its lending to meet pent-up demands for credit, and risk punishments from other banks. Using the data from the states and form New Jersey during 1991-1992, Bernanke and Lown (1991) demonstrated that loan growth at individual banks during the 19911992 was positively linked to initial capital ratios. They also find evidence that declines in bank capital have contributed to the slowdown in lending, which is consistent with ‘capital crunch’ hypothesis. However, they cautioned that the magnitude of the effect is not insignificant but also not extremely large either. Peek and Rosengern (1995) find the similar evidence of “credit crunch” for similar period of time by focusing on bank deposits. In addition, the results of several studies on Japanese banks (Ito & Sasaki, 1998; Kim & Moreno, 1994; Who, 1999; Honda, 2002) show that the result of Basel I on Japan was similar to that on the United States. Berger and Udell (1994) investigate the impact of risk-based capital adequacy on credit crunch in the United States as well as some alternative explanations of credit decline of early 1990s. They compared how bank portfolios changed in the early 1990s from the 1980s and how these changes were related to key variables related with risk based capital and other variables (large banks, banks with weaker capital ratios, and banks supervised by the OCC). Their findings indicate that the risk based capital related credit crunch hypothesis fares the worst of all the alternative explanations of the bank credit reallocation of the 1990s. On the other hand, macro-regional effects dominate the other factors behind the credit decline. With respect to emerging market countries, Chiuri et al. (2002) argue that the introduction of higher minimum bank capital requirements may well induce a contraction of bank credit, and aggregate slowdown. Their sample includes 16 emerging countries, 10 of which experienced both regulatory change and financial crises; another five were non-crisis countries. This study confirms the existence of “credit crunch.” Nonetheless, Barajas et al. (2005) did not find strong evidence of a Basel-induced credit crunch in Latin America. Watanabe (2004) analyzes the impact of prudential regulation in slowing down credit expansion, countering the effectiveness of monetary policy in stimulating economic conditions in Japan. Stagnation in Japan persisted during the last decade despite monetary easing, as evident by “zero interest rate policy” since February 1999, which proved to be ineffective. Over the years, several scholars have studied the impact of capital regulations intensively. Aggarwal and Jacques (1997, 2001) and Jacques and Nigro (1997) deal exclusively with the impact of such regulations on the risk-taking behavior of commercial banks and changes in capital ratios. One of the major objectives of Basel was to increase the capital ratios of undercapitalized banks; these studies on developed countries find that the basic relationship between risk-taking and capital ratios were negative and that such regulations did, in fact, increase the capital ratios of undercapitalized banks. Van Roy (2003) later found evidence that risk-taking by commercial banks and capital ratios were negatively related in developing countries, confirming earlier findings of the literature. Hussain and Hassan (2004) find similar results, and also find that the impact of Basel I fell short of expectations (i.e., did not increase the capital ratios of banks). As already pointed out, many developing and emerging market countries have adopted and implemented (or are at different phase of implementing) a national capital regulation regime. During the 1990s and early 2000s, the principles of Basel I have emerged as the core principles of capital regulation regimes around the world, but data on implementation years and stages throughout the world was relatively scarce and remains scattered. The objective of the present study is to examine the impact of the implementation of Basel I-like capital requirements on bank lending in emerging market countries.

Chapter V: Depositor Discipline and Bank Risk-Taking Behavior

In the wake of successful regulatory changes in the developed countries, many developing countries started to liberalize their financial sectors in the 1990s. In spite of their initial success, the liberalization policies now face a backlash in several developing countries. Critics present the cases of international financial crises (the Mexican Crisis in 1995 and the Asian Crisis of 1997) and domestic crises (Japan in the late 1990s and Turkey during the early 2000s). The argument centers on the timing, pace and the ways in which such policies were implemented. These have renewed the profession’s interest in the issues of competition and stability in the banking sector. One consequence of liberalization and privatization in the banking sector of developing countries was that it increased competition in the banking sector of the host countries, as documented by IMF (1998) and several other studies. Liberalization opened up the banking sector to open competition by the private, foreign and even government banks. This paved the way for increased risk taking. Malaysia, Korea, Thailand, Indonesia, and Philippines faced such situations since the starting of the 1990s. The focus of the present study is to follow that episode with focus on 1997 Asian crisis. More specifically we indent to examine the changes in risk taking behavior of banks before and after 1997. In addition to risk taking behavior of banks we will examine evidence of depositor discipline in the after math of 1997 crisis. If depositors punish banks when quality of assets of commercial banks decline, then we claim that there is evidence of depositor discipline. But if this is not the case when we claim otherwise. Until recently, most of the literature in this field has concentrated on the determinants of financial/bank crisis (i.e., Calomiris, 1990; de la Caudra & Valdes, 1992; Kaminksy & Rienhart, 1996; McKinnon & Pill, 1996; Hagen & Ho, 2003; Demirguc-Kunt & Detragiache, 1998, 2005), where incentives that lead banks to take on more risk have taken a back stage. However, our approach to studying the risk taking problems of banks is the change in risk incentives directly and it is relatively new. Following Gruben et al. (1997, 1998, 2003), we examine the shifts in bank risk and the factors that directly make such activities more attractive. Gruben et al. (1999) point out that the two major factors that cause banking panic are lack of market (depositor) discipline and financial liberalization. Demirguc-Kunt and Detragiache (1998a, 1998b, 2005) and Kaminsky and Reinhart (1996) find evidence indicating that the risk-taking activities of banks increase in the wake of liberalization, especially in countries where financial institutions are underdeveloped, law enforcement is weak and regulatory supervision is inadequate, which is more likely in developing countries. In developing countries, such liberalization often results in increased opportunities for excessive risk-taking and fraud. Demirguc-Kunt and Detragiache (1998a, 1998b, and 2000a) find evidence that risk-taking activities on the part of banks also increase due to the moral hazards created by deposit insurance. This shows that explicit deposit insurance reduces depositor discipline, which increases moral hazard. Thus two factors directly related to the stability of banks are market discipline and financial liberalization. Apart from the recent increase in the number of bank crises and the resultant academic interest, in June 2004, the Basel Committee and the 10 OECD countries finalized Basel II. It is expected that before long, this will become the internationally accepted standard for bank supervision and regulation, like its predecessor Basel I, which is at different phases of implementation in over 100 countries. The primary focus of Basel II is risk measurement and risk management, making the changes in banks’ risk-taking in response to policy changes more important than before. At the same time, the Committee has also decided that “market discipline” be made one of the three pillars on which future financial regulation should be based, because such discipline imposes strong incentives on banks to conduct their business in a safe, sound, and efficient manner, and also to hold adequate capital with respect to the regulatory minimum. These steps are expected to reduce the risk of bank portfolios (Ghosh & Das, 2004), but as it has been pointed out, there is little empirical evidence supporting this hypothesis. This paper looks to fill that gap. Despite strong interest of and rapid progress in research into the issues (not to mention the urgency of such studies), research with respect to developing and emerging market countries remain incomplete, with the exception of some jurisdictions (i.e., Argentina (1995), Canada (1984-86), Mexico (1995), Singapore (1997-99), Norway (1987-89), and Texas Savings and Loan Associations (1984-90)). Kaminsky and Reinhart (1999) refer to the South American financial problem of 1995 and to the Asian Crisis of 1997 as the “twin crisis”; the countries involved in these crises face concurrent exchange rate and banking problems. The present study expands upon the existing literature for South Asian countries, to include South Korea, Philippines, Thailand, Malaysia, and Indonesia. One of the relevant studies for the present study is Gruben and McComb (2003), who apply the Breshnahan methodology to the Mexican banking sector of the mid-1990s and find that the Mexican banking system was super-competitive – that is, marginal prices were set below marginal costs. This was called “super-competition,” where banks take risks now to capture a larger share of the market, and in the future hopefully reap the benefits of such a hostile expansion of market share. The two most relevant studies for the current paper are Gruben et al. (1998) and Gruben et al. (2003). The study of Argentina, Mexico and Canada by Gruben et al. (1998) finds that lending risk (measured as a level of “super-competition”) increases significantly in the aftermath of liberalization, in countries where market discipline is weak. However, where depositors discipline banks by withdrawing deposits when asset quality falls , banks do not behave in a risky fashion. Again, another study by Gruben et al. (2003) on six jurisdictions (Argentina, Canada, Mexico, Norway, Singapore, and Texas) provides evidence that bank risk increases significantly in the aftermath of liberalization, but only where depositors fail to discipline banks and where market discipline and bank risk were persistently and inversely related. As a result, current paper attempts to find answers to the following three important questions: (i) does depositor discipline exist in the selected South East Asian countries; (ii) does risky lending occur prior to bank crises, and (iii) what is the relationship between depositor discipline and bank risk-taking behavior? We follow a joint Breshnan (1982)/Shaffer (1989)/Gruben (1999) framework. The review of the literature discussed above provides evidence on the behavior of banks with respect to risk taking and strength of depositor discipline within the context of South East Asian Crises of 1997. The results are in line with that of Gurben (1998, 2003) and other studies. Even in the after-math of Asian Crises, the strength of depositor discipline (tested as a negative relationship between depositor discipline and assets quality) was absent. Depositors as a group did not withdraw deposits from banks in reaction to a decline in their asset quality. This also highlights the importance of the need to improve measures that strengthen depositor discipline, such as, the measures put forward in Basel II. Timely and accurate dissemination of information can go a long way to solve these problems. Overall, any measure that enhances transparency and accountability of the decision making of banks is welcome. These are some of the objectives of Basel II. For the test of a break in the competitive behavior of banks before and after 1996, we find no change in banks’ risk-taking behavior before and after the crisis. Also the index has a value of zero, implying competitive behavior of banks in these countries. This to is in line with Gruben et al. (1998, 2003). However, use of PR-methodology to measure bank competitive behavior indicates the monopolistic competition and not perfect competition exists in these markets. However, we did not find a consistent link between depositor discipline and bank risk taking, which can be explained by the weak state of depositor discipline. At least until today depositor discipline has not grown strong enough to affect banks risk-taking behavior. More needs to be done to make depositor discipline effective. Caprio et al (2004) also arrived at the same conclusion. They pointed out that “it seems that greater emphasis on the third pillar in the Basel II Accord that on the refinements of the risk-weighting system of the first pillar may be warranted for most developing countries.

Overall Conclusion

We present the overall conclusion of the three essays in the present dissertation in the following couple of paragraphs.  In contrast to previous studies that deal with one developed countries with a few exceptions, evidence of these chapters come from multi-country panel data. It is often pointed out that the two most important impacts of Basel accord are: one; undercapitalized banks did increase their actual capital ratios in response to regulatory minimum capital ratio; and two; there is limited evidence that such regulations were associated with a decline in bank lending activities due to pressure on banks to hold more capital. The discussion presented in previous chapter provides evidence that the increase in actual capital ratio of undercapitalized banks did not come at the cost of higher risk. Indeed, evidence reveals that the relationship between these two variables is negative. In the second step, we examine the relationship between capital ratios and risk under different conditions. Evidence for the proxy of financial development remains ambiguous. There is evidence that banks operating in more restrictive environment, abstain from taking higher risk. Chapter 3 also provides long-run evidence that the average capital ratios and equity of banks increased in the aftermath of implementation of capital regulations. It also provides evidence that such regulations have strong impact on the important decision making variable of banks, for example, measures of banks capital and bank loans. Chapter 4 provides evidence that the state of depositor discipline is weak in five selected countries of South East Asia, which include Malaysia, South Korea, Philippines, Thailand and Indonesia. Depositors did not punish commercial banks in response to a decline in the quality of assets. Banking industry in these countries remained perfectly competitive during this time period and there was no shift of bank competition. But we could not establish the link between depositor discipline and bank capital.