Use of Interest Rate Swaps by Commercial Banks: An Analysis of Determinants

Contemporary Course Study Submitted To

Prof. Jayadev M. ON August 28, 2007

BY

Chirag Thakral (0611016) Sandeep Kumar (0611042)

INDIAN INSTITUTE OF MANAGEMENT, BANGALORE

ACKNOWLEDGEMENT

We express our sincere thanks to Prof. Jayadev M. (Finance & Control Area) for his guidance and support extended to us in preparation of the study on “USE

OF INTEREST

RATE SWAPS BY COMMERCIAL BANKS: AN ANALYSIS OF DETERMINANTS”. He has been very helpful throughout the project and co-operated and guided us in the best possible way. We have benefited a lot from the depth of his knowledge.

We are also extremely grateful to Mr. Srivatsan Sudersan, Bloomberg, Singapore, for being so helpful and cooperative to us in procuring the data required for the project.

Finally we would like to express our heartfelt thanks to all those who have helped us in this project.

Use of Interest Rate Swaps by Commercial Banks: An Analysis of Determinants Chirag Thakral (0611016)

Sandeep Kumar (0611042) Prof. Jayadev M.

Indian Institute of Management Bangalore August 28, 2007

Abstract Banks like any profit making organisation face a variety of macroeconomic and microeconomic risks and need to manage all these risks using different methods for their efficient functioning. In this study we develop an analysis on hedging of the most significant of financial risks faced by the commercial banks i.e. interest rate risk, using the interest rate swaps. We also look at the revenue generation of banks by trading in interest rate swaps. Banks interest rate exposure associated with a mismatch between assets and liabilities can be measured using traditional GAP and duration GAP analysis. To fill this GAP a commercial bank uses derivatives instruments to adjust the amount of assumed interest rate. Interest rate swaps are the most popular financial derivatives used by the commercial banks. In this study the effect of various bank specific parameters (size, profitability, capitalization, interest rate risk profile, creditworthiness, Stock Price etc.) and economy specific parameters (interest rate volatility, market index etc.) on the interest rate swap positions of the bank are examined. The study used the annual data of about 20 banks (from India, Singapore, Hong Kong and Brazil) for a period of over 5 years from 2002-2006 and tried to find out the pattern of interest rate swap usage for asset liability management by these banks. The various bank specific and economy wide parameters are regressed using Panel Regression, against the notional amount of interest rate swaps obtained from the annual reports of the bank. The results suggest that the size, profitability, creditworthiness and interest rate risk profile of the bank are the bank specific factors that play a major role in the banker’s decision to take interest rate swap position in the market and the extent of the notional amount is dependent on these. The interest rate volatility is a result of the economy of the country in which the bank operates and banks take position in the interest rate swap market accordingly to tackle this factor. The stock market conditions for the bank as well as for the index are not a significant determinant. Keywords: Interest Rate Swaps, Panel Regression, Commercial Banks, Interest Rate Risk

1

Introduction

Bank for International Settlements (BIS)

There are a number of risks banks face that are

publishes statistics on the notional amounts

typical of non financial firms. According to Hem-

outstanding in the OTC (Over the Counter)

pel et al (1994), financial risks are considered to

Derivatives market. The notional amount

be the most critical risks faced by the banks. The

outstanding as of December 2006 in OTC interest

main concern of this study is the analysis of hedg-

rate swaps was $229.8 trillion, up $60.7 trillion

ing used by commercial banks for financial risk,

(35.9%) from December 2005. These contracts

in particular interest rate risk. Interest rate risk is

account for 55.4% of the entire $415 trillion OTC

the risk that the market value of a bank‘s asset

derivative market1. Swaps based on Interest rate are mostly

declines due to a change in interest rates. Looking at financial risks and in particular

used for hedging and speculation. All major

interest rate risks from the management perspec-

banks in world now have an active Derivatives

tive, there are two different approaches managers

structuring desk for corporate wherein they offer

may adopt in controlling interest rate risk, namely

various swap products like Caps, Floors, Digitals,

on-balance sheet adjustments and off-balance

LIBOR Range Accrual, CMS Spread Range

sheet adjustments. Each has particular strengths

Accrual, Knock out swap, CMS Blade Range and

and weaknesses, and should be viewed not as

Snow range runner etc. for hedging and

competing methodologies, but as alternatives

speculation.

which might be appropriate in certain circumstances. On balance sheet adjustment involves the adjustment of some of the bank‘s assets and liabilities in a way that the net effect of an interest rate movement will not harm the market value of these components and as a result the overall profitability of the bank. The off-balance sheet approach involves the use of various non-traditional financial products, widely known as derivative instruments. These instruments appear neither on asset nor on liability side.

Figure 1.1 classifications of Bank Risks

In recent years, use of interest rate derivative products has increased significantly. The

1

http://en.wikipedia.org/wiki/Interest_rate_swap

2

To illustrate the market conventions, we provide An interest rate swap is a contractual agreement

in fig. 1.2 an example of an intermediated swap

entered into between two counterparties under

transaction. The dealer that makes fixed payments

which each agrees to make periodic payment to

and receives floating payments is said to be on the

the other for an agreed period of time on the basis

bid side of the transaction. Thus, in fig. 1.2, the

of a notional amount of principal. Interest Rate

swap dealer has a bid rate of Treasury plus 15 ba-

Swaps are typically used to create either synthetic

sis points. If the swap has a term of five years, the

fixed or floating rate liabilities or assets in an ef-

five-year Treasury yield will be used as the

fort to hedge against adverse movements in inter-

benchmark. Conversely, the dealer that pays a

est rates. The floating rate index is usually pegged

floating rate and receives a fixed rate is said to be

to short-term interest rates, such as LIBOR or

on the offer side of the transaction. In the example

one-month commercial paper.

in fig. 1.2, the dealer has an offer rate of the five-

Vanilla interest rate swaps are quoted in

year Treasury yield plus 25 basis points. Swap

terms of the fixed rate to be paid against the float-

dealers charge bid-offer spreads for their services:

ing index. The fixed rate is usually quoted as an

they pay less on the bid side and receive more on

absolute rate, so a quote of 3% against 3-month

the offer side. It is intuitive to expect that, all else

Libor would indicate that the fixed rate would be

being equal; the bid-offer spreads depend on the

3% pa. The fixed rates on vanilla swaps are called

swap dealers’ credit reputations2.

swap rates. The swap curve is a yield curve com-

The responsibility of a bank manager is to

prising swap rates for different maturities. In

eliminate risks, but at the same time has to keep in

highly liquid USD swap market, the swap curve

mind that the hedging techniques used by them

has emerged as an alternative to Treasuries as a

eat into the excess returns from risk taking.

benchmark for USD interest rates at maturities exceeding a year. Swap markets serve as a link between government debt, corporate debt and

Literature Review

money markets, across currencies and maturities.

In this section we attempt to look at bank’s motivation to use derivative instruments (especially interest rate swaps) as an exercise to manage the

2

Figure 1.2 Working of interest rate swap

Tong-sheng Sun, Suresh Sundaresan, and Ching Wang (1992), “Interest rate swaps, An empirical investigation”.

3

asset liability mismatch by reviewing work done

business of banking includes a number of impor-

in this area.

tant risk factors (for example, default or credit risk, interest-rate risk, and foreign-exchange risk) that may be related to the use of derivatives for

1) Generation of Income

hedging purposes. An important consideration is

The generation of fee income is a major incentive

the interaction between interest-rate risk and

for banks to provide risk-management services to

credit risk. Schrand and Unal (1998) investigate

corporate clients. Smith (1993) argues that bank-

hedging and coordinated risk management used

ers must recognize the potential benefits of selling

by thrifts to control both credit risk and interest-

risk-management services. The obvious benefits

rate risk. If banks are practicing coordinated risk

come from the generation of fee income and the

management, then the use of derivatives to hedge

opportunities for bankers to create value through

interest-rate risk should also be related to a bank's

cross-selling and enhanced customer relation-

credit exposure. If banks use derivatives (to hedge

ships. Less obviously, since hedging with deriva-

or speculate) in response to credit risk, a positive

tives can reduce the probability of financial dis-

relationship is expected to exist between loan

tress for client firms, banks also benefit by reduc-

losses and derivatives use.

ing their risk exposure to their customers. In addition to above papers we looked at the fol2) Hedging activities

lowing papers:

Most of the big banks use interest rate swaps for

Wall and Pringle (1989)

income generation but the bank at lower end of

Wall and Pringle (1989) surveyed a set of 250

the spectrum use it to hedge against, or speculate

firms which had reported the use of interest rate

on, the movement of economic variables. Smith

swaps. This study tried to find out the different

and Stulz (1985) argue that hedging can reduce

motives for using interest rate swaps. They found

the probability of bankruptcy by reducing the

that no single explanation is adequate in explain-

variance of cash flows. Banks with a higher prob-

ing the behavior of all swap users. They found out

ability of financial distress would be most likely

that the motives for using interest rate swaps dif-

to benefit from a hedging program. This implies

fer between banks and other financial institutions

that banks with greater leverage, and hence a

and the non financial institutions.

greater probability of experiencing bankruptcy, are more likely to use derivatives to hedge. The 4

Song (2005)

ported notional values, is positively related to the

Song (2005) examined the use of interest rate

balance sheet mismatch one to five years forward,

swaps to manage earnings in the banking sector.

the competitiveness of the bank's deposit market,

He finds that banks enter into swap arrangements

the bank's total asset size, primary capitalization

to increase (decrease) earnings when they would

given the bank is a swap dealer, the relative size

have otherwise missed (exceeded) their target. In

of its commercial and industrial loans portfolio,

a follow up analysis, he finds no evidence that

and futures trading. The study by Kim and Kop-

banks enter into offsetting swap arrangements in

penhaver (1992) forms the first attempt to shed

subsequent periods to offset the interest rate risk

some light on the pattern of interest rate swaps

taken on by the initial swap arrangement.

use by commercial banks. It used the method of regression but the factors used were not much.

Kim and Koppenhaver (1992) Kim and Koppenhaver (1992) analysis focused on

L. Shanker (1996)

whether interest rate exposure is a sufficient de-

Shanker (1996) studied the effect that the use of

terminant of a bank to engage in interest rate

interest rate derivatives (futures, options, and

swaps transactions. Thus, the authors tried to re-

swaps) may have in hedging the interest rate risk

late the interest rate swap position of the banks to

of large US commercial banks. Complete data on

their interest rate risk profile. However, as Simons

stock returns and interest rate sensitive assets and

(1995) mentioned in her study the very broad na-

liabilities were collected for 360 banks over the

ture of the data in the banks call reports make it

period 1986-1993. The study suggested that de-

impossible to define such relationships with any

rivatives usage by banks can effectively reduce

degree of precision. According to the findings of

their interest rate risk. However, the study found

the study the authors reported a positive relation-

that when balance-sheet exposure is considered

ship between the long term interest rate exposure

the GAP measure is either the same or greater for

of the bank and the extent of swap market partici-

the group of banks that uses derivatives than for

pation. Thus, the study supported the hedging ar-

the group of banks that does not use derivatives.

gument which requires the bank to increase its

Thus, the above result was only supported when

hedging position along with the increase of its in-

the sensitivity of bank's stock returns to interest

terest rate exposure.

rate changes was considered.

Moreover the study found that “the size of interest rate swap positions, as measured by re-

5

Sinkey and Carter (1997)

of hedging the interest rate risk of the banks' port-

Sinkey and Carter (1997) examine cross-sectional

folios. The objective of this study was to deter-

differences in the use of derivatives by banks in

mine which bank's characteristics best explain the

an effort to identify the factors that cause banks to

use of interest rate swaps by commercial banks in

use derivatives. The results of their study suggest

order to hedge the interest rate risk of their portfo-

that many of banks' derivative activities can be

lios? These characteristics may include the bank's

explained by the arguments of contemporary fi-

size, capitalization, assets quality, interest rate

nance theory. Consistent with avoidance of the

risk profile and the firm's revenue performance

costs of financial distress, they find that banks

(these were decided based on previous work done

with greater leverage (and lower capital-to-asset

on these topics). Moreover, the degree to which

ratios) are more likely to use derivatives. Incon-

these characteristics can explain the extent of that

sistent with this argument, however, they find that

use was also examined. The empirical result was

banks with greater on-balance-sheet sources of

derived using hypothesis formulation and per-

risk exposure, as reflected in large maturity gaps,

forming regression on the data available. The re-

are less likely to use derivatives.

sults disclose a positive relationship between the size of banks and the use of interest rate swaps. In

Lyes Boukrami (2002)

addition, the study has found that banks with bet-

The Paper “The Use of Interest Rate Swaps by

ter asset quality tend to use interest rate swaps

Commercial Banks” by Lyes Boukrami, Graduate

more intensively than banks with weaker assets

Business School, and Manchester analyzed and

quality.

illustrated the relationship between one of the most significant financial risks faced by banks i.e. interest rate risk, and the interest rate swap. This study developed a fixed effect model using cross section data from US banks Annual Reports for the Year 2001. The paper defined the Bank’s motives of using derivatives by giving variety of reasons as were mentioned in various papers written before this paper. The paper used annual data for a number of US commercial banks and analyzed the determinants of interest rate swap as a mean

Aim and Scope of the Study Although interest rate swaps have been the fastest growing off-balance sheet product in emerging markets, very less research has been conducted on the analysis of which specific characteristics best describe the use of interest rate swaps by banks and which of these can describe the extent of that use. Other than the bank specific factors like size, profitability, stock price, creditworthiness and interest rate risk profile of the bank, the economy

6

wide factors used are financial market index and

Selection of the variables

interest rate volatility. Although banks don’t dis-

In order to select factors which should be exam-

close much about their strategy of portfolio man-

ined to find out the dependence of interest rate

agement, this study analyses the relationship be-

swap exposure of a bank the model developed by

tween various factors of bank and economy and

Simons (1995) was employed. The study by

the position of bank in an interest rate swap.

Simons was selected since it provides similar re-

Hence, the problem is to “Find out the determi-

search issues with one of the aims of this study.

nants and their extent of influence on the interest

This study deals with the emerging markets as

rate swap positions of a commercial bank as a part

compared to the study by Simons. Moreover, this

of their strategy of asset liability management.”

study deals with a smaller dataset than Simons attributed to the limited period of time available during the face of data collection.

Methodology of Study and Motivation

The limited number of studies conducted in the

In order to solve the above problem our study de-

area of derivatives use by commercial banks sug-

velops a fixed effect model with the cross section

gests that all the authors have more or less used

data for 5 years. The data used in this study was

the same parameters in order to explain the use of

obtained from annual reports of commercial banks

these activities by banks, each one using different

in India, Hong Kong, Singapore and Brazil for the

proxies to capture the effects of these parameters.

year 2002 to 2006. The sample size has been restricted to include only the large banks from a

The most common variables that appear in most

country as per Simons (1995) paper, which

of these studies are (i) the bank's size, (ii) the

proved that interest rate swap use is much more

bank‘s capital position, (iii) the bank‘s credit wor-

widespread among large banks. The resulting

thiness and, (iv) bank‘s interest rate risk profile

sample includes 20 banks. The notional amount of

(v) and the bank revenue performance. The table

the interest rate swaps used for asset-liability pur-

given below provides the most commonly used

poses is regressed against the specific characteris-

parameters by some of the most recent studies on

tics of the banks and economies. The computing

derivatives use by commercial banks

software used for the calculation of the results is EViews 5.0.

7

Variable

Kim and

Simons,

Jagtiani,

Lyes Bou-

Bank specific characteristics

Koppen-

1995

1996

krami,

• Bank’s size

haver,

2002

1992 Size

Total Assets

Total As-

Total Assets

Total Assets

NPA/Asset

S&P Rat-

LLR/Total

LLR/Loans

ings

loans

sets Creditwor-

NA

thiness

The variable Total Assets is used to account for the size of the Bank. It can be seen from the above table that all the papers on this subject have used

IBT/Assets

the same variable and hence there is no ambiguity in selection of this variable.

NPL/Assets

The bank’s size was taken into account as

Capitaliza-

Capital to

Equity to

KBIND

Equity to

tion

Asset Ratio

assets ratio

Capital

assets ratio

research has shown the importance of size in

NII/Total

banks' use of futures (Koppenhaver, 1990), swaps

Income

(Kim & Koppenhaver, 1993), and interest-rate

Ratio Interest Rate

GAP meas-

GAP

Risk

ure

measure

Return Per-

ROA

NA

formance

NA

NA

IBT/Total Assets Ratio

Where IBT= Income before Taxes; NPL= Non-Performing Loans; NPA= Non-Performing assets; KBIND= The GAP between the Capital level and the Binding level. NII= Net interest income (the difference between interest income and interest expenses)

Table 1: Usage of variables by different papers

derivatives (Carter & Sinkey, 1998; Gunther & Siems, 1996). Large banks due to their large size may exploit economies of scales using swaps as an asset-liability tool because of more expertise and credit worthiness in the a\swap markets. According to Kim and Koppenhaver (1992. p.63-

The variables that may be important in explaining the use of interest rate swaps for asset liability management purposes of a commercial bank, for the purpose of this study, are presented below. Note that all the data that forms the explanatory variables of this study were gathered from the consolidated balance sheets and the income statements of the banks under examination by searching through the annual reports of each bank or looking up on Bloomberg Terminal. The list of banks and their websites was retrieved from Bloomberg Terminal.

64) large banks are in a position to utilize specialized management skills to manage effectively an off balance-sheet activity as swap market participation. Therefore, size seems to be an important variable in explaining the characteristics of bank using swap and thus, total assets is included as an independent variable for the model. If the analysis reveals that this variable is significant, it can be deduced that there are barriers to entry in the swaps market, and that size gives an advantage to large entities that use interest rate swaps for asset liability management activities.

8

Hypothesis

seen in the table given above. As the capital of a

H10: There is no relationship between the usage

bank increases, the bank’s risk handling ability

of interest rate swaps as an asset liability man-

increases and hence the need for use of interest

agement tool and the size of the bank.

rate swap (for hedging purposes) decreases.

H11: A significant relationship exists between the

Hence to investigate this relationship the variable

usage of interest rate swaps as an asset liability

equity to assets ratio has been included in the

management tool and the size of the bank.

model. Moreover, Kim and Koppenhaver (1992) also point out bank capital is an important dimen-

• Bank’s Creditworthiness

The ratio of the Non Performing Assets (NPA)

sion that has to be considered when choosing to participate in swaps market.

to the total assets, accounts for the creditworthi-

Hypothesis

ness or asset quality of a given bank. This ratio

H30: There is no relationship between the usage

gives the percentage of bad loans with respect to

of interest rate swaps as an asset liability man-

the total loans given. Banks with relatively poor

agement tool and the equity to assets ratio of the

asset quality (as measured by high levels of Non

bank.

Performing Assets to total assets), will need to

H31: There is a significant relationship between

conserve capital and may find interest rate swaps

the usage of interest rate swaps an asset liability

as a more desirable, capital efficient way to man-

management tool and the equity to assets ratio of

age their balance sheet.

the bank.

Hypothesis H20: There is no relationship between the usage of interest rate swaps as an asset liability man-

• Bank’s Interest rate risk exposure The ratio of net interest income (the difference

agement tool and the NPA ratio of the bank.

between interest income and interest expenses) to

H21: There is a significant relationship between

total income (appears in the income statement as

the usage of interest rate swaps an asset liability

net income: the sum of net interest income and net

management tool and the NPA Ratio of the bank.

non-interest income) is used to measure the interest rate risk exposure of the bank without taking

• Bank’s Capitalization

the interest rate swap into consideration. This

The equity to assets ratio is used to find out the

measure is included unlike the GAP measures

capitalization for a bank. This is in accordance

used by other studies (Kim and Koppenhaver,

with most of the papers on this subject as can be

1992; Simons, 1995) in order to avoid a major

9

limitation of the GAP as a proxy of interest rate

exposure (Net Interest Income to Total Income

risk when used in this kind of studies. In order to

ratio) of the bank.

perform a GAP analysis for a particular portfolio, one prerequisite is to classify the assets and li-

• Bank’s Profitability

abilities according to their maturity characteris-

The ratio of profit before tax to total assets is

tics. That is, all the assets that mature within a

used as a measure for the profitability of the bank.

given period (e.g. zero to three months) of time

The higher the ratio the more profitable is the

must be pooled in one aggregate measure. The

bank. The bank’s profit increases cash flow and

same is also true for the liabilities. Now, in the

hence the credibility of the bank. This helps the

case of derivatives instruments or more exotic

bank to have a greater position in interest rate

products, since these products are usually re-

swaps. Hence we expect a positive co-relation be-

priced in multiple periods of time (e.g. swaps) it is

tween the bank’s profitability and its interest rate

rather difficult to classify them according to their

swap exposure.

re-pricing characteristics. Therefore, when using

Hypothesis

GAP measures, one fails to account for the por-

H50: There is no relationship between the usage

tion of risk that these instruments contribute to the

of interest rate swaps as an asset liability man-

overall risk of the portfolio. Our measure of the

agement tool and the Income Before Taxes to To-

interest rate risk takes into consideration all these

tal Assets ratio of the bank.

limitations and hence is a better estimate of the

H51: A significant relationship exists between the

interest rate risk. The larger the value of this ratio

usage of interest rate swaps as an asset liability

the greater is the interest rate risk exposure of a

management tool and the Income Before Taxes to

bank.

Total Assets ratio of the bank.

Hypothesis H40: There is no relationship between the usage of interest rate swaps as an asset liability man-

• Bank’s market performance The bank’s closing stock price on the last day of

agement tool and the bank’s interest rate risk exposure (Net Interest Income to Total Income ratio) of the bank. H41: There is a significant relationship between the usage of interest rate swaps an asset liability management tool and the bank’s interest rate risk

10

match a bank may open a position in the interest

• Interest rate volatility

rate swaps.

The ratio of standard deviation to mean of the

Hypothesis

interest rate of an economy is used as a measure

H60: There is no relationship between the usage

of the market performance of the country. The

of interest rate swaps as an asset liability man-

measure of the deviation of the interest rate with

agement tool and the Stock Price of the bank.

respect to the mean interest rate is a safe measure

H61: A significant relationship exists between the

for the volatility of the markets (or interest rates).

usage of interest rate swaps as an asset liability

Frequent changes in the interest rate of an econ-

management tool and the Stock Price of the bank.

omy will affect the positions of various banks in the interest rate swaps.

Economy specific characteristics

Hypothesis

• Stock Market index performance

H80: There is no relationship between the usage

The market index of the most prominent stock

of interest rate swaps as an asset liability man-

exchange of a country is used as a measure of the

agement tool for a bank and the interest rate vola-

market performance of the country. With the

tility of the country.

change in the market performance of the econ-

H81: A significant relationship exists between the

omy, the assets side of the bank’s balance sheet

usage of interest rate swaps as an asset liability

may change. This may cause an asset liability

management tool for a bank and the interest rate

mismatch in the bank’s balance sheet. To care of

volatility of the stock exchange of the country.

this mismatch a bank may open a position in the interest rate swaps.

The summarization of all the above selection of

Hypothesis

variables is done in the table given below:

H70: There is no relationship between the usage

Variable Used

Representative for

of interest rate swaps as an asset liability man-

Total assets NPA/Total Assets

Size Creditworthiness/Asset Quality Capitalization Interest Rate Risk exposure

agement tool for a bank and the market index of the stock exchange of the country. H71: A significant relationship exists between the usage of interest rate swaps as an asset liability management tool for a bank and the market index of the stock exchange of the country.

Equity-to-Assets Ratio Net Interest Income / Total Income Profit before Tax / Total Assets Stock Price of Bank Market Index of stock exchange of country Standard deviation/mean of the interest rate of country

Profitability Bank’s Market Performance Market’s performance Market Volatility

Table 2: Variables used for various parameters

11

Results Descriptive statistics All the data is collected and converted into USD for consistent results. Table given below provides some preliminary descriptive statistics for the variables used in the empirical analysis. A total of 80 observations of the banks have been used. The notional principal of swaps used for asset liability management averaged $505,209.69 million and ranged from $911 million to $6,749,847 million. The Bank size as measured by the assets size had an average of $127,974 million, but most of the banks in the sample have a lower amount of assets. The largest bank in the sample is HSBC with a total amount of assets of $1,860,758 million, whereas the smallest bank was UTI Bank with total assets of $4,049 million. The Asset quality on an average is 1.22% with a variation of 1.27%, representing the variety in the types of the banks in the sample. The prof-

Variable Total assets (Million USD) NPA/Total Assets Equity-toAssets Ratio Net Interest Income / Total Income Profit before Tax / Total Assets Stock Price of Bank Market Index of stock exchange of country Standard deviation/mean of the interest rate of country

N

Min

Max

Mean

Std. Deviation

80

4,049

1,860,758

127,974

294508.9684

80

0

0.05284

0.01223

0.012729715

80

0.00093

0.11928

0.04913

0.045089438

80

0

0.75758

0.39461

0.192685056

80

0.00439

0.04875

0.01563

0.007468489

80

0.26264

36.23170

7.07589

6.86613013

80

70.3967

20,818.6

2,708.278

4949.453003

80

0.01870

0.21045

0.14132

0.056806605

Table 3: Descriptive summary of Bank Data

itability of BANCO ITAU bank is highest at 4.875% with the average profitability of the sample of the banks at only 1.56%. In terms of Market index performance of each country, the sample is very much varied with a standard deviation of 4949.45 and with the best valued index of Brazil at 20,818.6 in 2006 and worst valued index of India at 70.39 in 2002.

12

Regression Results

of the changes in the interest rate swaps are due to

The data we have used to do regression is a time-

the variations of the independent variable used in

series data available for around 5 years for differ-

the model. This supports the good selection of the

ent banks, which makes the sample data as cross-

proxies.

sectional data. So we have used the Panel Regression to regress the interest rate swap notional values of the banks against the bank and economy specific factors. Following Results were obtained:

Analysis of Hypothesis

Dependent Variable: INTEREST RATE SWAP

Size of the Bank

Method: Least Squares

The regression results in table shows that there is

White Heteroskedasticity - Consistent Standard Errors & Covariance

a positive relationship between the use of interest

Included observations: 80

rate swaps and size. The variable's total asset coVariable

Coefficient

Std. Error

t-Statistic

Prob.

efficient has a positive sign and equals to 0.4325

TOTALASS

0.432594

0.055559

7.786160

0.0000

in its regression coefficient and is a significant

CAPITAIS

1557281.

1048973.

1.484577

0.1424

variable. So we must reject the null hypothesis

PROFITAB

7096655.

3075197.

2.307708

0.0242

and accept that ‘a significant relationship exists

INTERSTR

219241.0

94472.64

2.320683

0.0234

CREDITWO

-5576091.

1414314.

-3.942612

0.0002

STOCKPRI

-1704.937

2134.464

-0.798766

0.4273

asset liability management tool and the size of

MARKETIN

-56.81285

54.22311

-1.047761

0.2986

the bank’. It can be deduced that the size and the

INTERE_A

-799486.7

363000.8

-2.202438

0.0311

between the usage of interest rate swaps as an

expertise, as well as skill, that may be implicitly

R-squared

0.742724

Mean dependent var

76707.22

available to large banks, give a comparative ad-

Adjusted R-squared

0.715437

S.D. dependent var

176841.9

vantage to these entities in using interest rate

S.E. of regression

94335.31

Akaike info criterion

25.84890

Sum squared resid

5.87E+11

Schwarz criterion

26.09799

Log likelihood

-948.4095

Durbin-Watson stat

0.922159

swaps for hedging and trading applications. Creditworthiness By using the NPA to total assets ratio as the vari-

Table 4: Regression Results Summary Table

able of the asset quality of the bank, the regression results in table shows that there is a negative

The R-square of the model equals 74.3% and the

relationship between the use of interest rate swaps

R-square adjusted of the model equals to 71.5%,

and asset quality of the banks assets. The vari-

which are both consistent. This means that 71.5%

able's coefficient has a negative sign and equals to

13

-5576091 in its regression coefficient and is a sig-

the view that Bank capital significantly enhances

nificant variable with p-value equal to 0.0002. So

the swaps participation is refuted in our sample.

we must reject the null hypothesis and accept that

These findings highlight the importance of distin-

‘There is a significant relationship between the

guishing the determinants of swaps participation

usage of interest rate swaps an asset liability

from the factors influencing the extent of swaps

management tool and the NPA Ratio of the

participation.

bank’. This suggests that banks with stronger asset qual-

Interest rate risk exposure

ity tend to be a bigger user of interest rate swaps than banks with relatively weaker asset quality. Therefore, our study is in line with earlier studies to conclude that swaps are perceived by regulators as risky and banks with weak assets quality are subject to more scrutiny by regulators when they attempt to use swaps.

As per Kim and Koppenhaver, 1992 there is expected to be a positive relationship that would satisfy the hedging argument that higher the interest rate risk profile of the bank, the higher would be its interest rate swap hedging position. The regression results in table shows that there is a positive relationship between the use of interest rate

Capitalisation

swaps and the interest rate risk profile. The variable's coefficient has a positive sign and equals to

The regression results in table shows that there is no relationship between the use of interest rate swaps and capitalisation. The variable's total asset coefficient has a positive sign and equals to 1557281 in its regression coefficient, but the pvalue of 0.14 suggests that it is not a significant variable to determine the interest rate swap position. So we must accept the null hypothesis and

219241.0 in its regression coefficient and is a significant variable with p-value of 0.0234. So we must reject the null hypothesis and accept that ‘There is a significant relationship between the usage of interest rate swaps an asset liability management tool and the bank’s interest rate risk exposure (Net Interest Income to Total Income ratio) of the bank.’

accept that ‘There is no relationship between the usage of interest rate swaps as an asset liability management tool and the equity to assets ratio of

Profitability

the bank’.

The regression results in table shows that there is

Inconsistent with the findings from the

a positive relationship between the use of interest

Lyes Boukrami (2002) study for the US banks,

rate swaps and profitability of the bank. The vari-

14

able's coefficient has a positive sign and equals to

to determine the interest rate swap position. So

7096655 in its regression coefficient. Profitability

we must accept the null hypothesis and accept that

of a bank comes out to be a significant variable in

‘There is no relationship between the usage of

the usage of interest rate swap with p-value

interest rate swaps as an asset liability manage-

0.0242. So we must reject the null hypothesis and

ment tool and the market index of the stock ex-

accept that ‘A significant relationship exists be-

change of the country’.

tween the usage of interest rate swaps as an asset liability management tool and the Income before Taxes to Total Assets ratio of the bank’.

Interest rate volatility The regression results in table shows that there is

Bank’s Stock market performance The regression results in table shows that there is no relationship between the use of interest rate swaps and Stock Price of the bank. The variable's coefficient has a negative sign and equals to 1704.937 in its regression coefficient, but the pvalue of 0.4276 suggests that it is not a significant variable to determine the interest rate swap position. So we must accept the null hypothesis and accept that ‘There is no relationship between the usage of interest rate swaps as an asset liability management tool and the Stock Price of the

a negative relationship between the use of interest rate swaps and interest rate volatility of the country. The variable's coefficient has a negative sign and equals to 799486.7 in its regression coefficient. Interest rate volatility of a country comes out to be a significant variable in the usage of interest rate swap with p-value 0.0311. So we must reject the null hypothesis and accept that ‘A significant relationship exists between the usage of interest rate swaps as an asset liability management tool for a bank and the interest rate volatility of the stock exchange of the country.’

bank’. Country’s Stock Market index performance The regression results in table shows that there is no relationship between the use of interest rate swaps and Stock Market index. The variable's coefficient has a negative sign and equals to -56.81 in its regression coefficient, but the p-value of 0.2986 suggests that it is not a significant variable

15

Key Interpretations & conclusions

has a consistent relationship with the use of inter-

After analyzing the data of 20 top commercial

est rate swaps came as per our expectations.

banks of some of the emerging markets, our study

From the economic factors considered to

sheds some light on the factors which influence

play an important role in the interest rate swap

the use of interest rate swaps for asset liability

positions of the banks, only the interest rate vola-

purpose by the banks. The study has found that

tility proved out to be a significant factor as fre-

larger the bank, higher will be their position in the

quent changes in the interest rate of an economy

interest rate swap market. This can also be attrib-

will have to be managed by banks, which affect

uted to the fact that larger the banks, better are the

the positions of various banks in the interest rate

technical expertise and better the customer rela-

swaps. The stock price or stock market index are

tionship and hence easier for them to get involved

not a factor in the banker’s decision to take posi-

in swaps and derivative activities.

tions in the interest rate swap market.

In addition, the study has found that banks with

Irrespective of whether swaps are for

better asset quality tend to be more intensive users

hedging or speculating, the positive relationship

of interest rate swaps, than those with weaker as-

between interest rate exposure and swap activity,

set quality. There is very little publicly available

if it exists, may not imply that the market does not

information on the swap positions of banks. This

care about the bank's risk and allow risky banks to

means that it is very difficult to estimate the expo-

be more active as guarantors in the swap market.

sure of interest rate movements.

Depending on the risk propensity of the institu-

As per the earlier studies as the capital of a

tion, risk can be controlled using a wide variety of

bank increases, the bank’s risk handling ability

techniques that fall into two broad categories,

increases and hence the need for use of interest

namely direct and synthetic methods. Direct re-

rate swap (for hedging purposes) decreases. But

structuring of the balance sheet involves the

our study concludes that capitalization of the bank

changing of the contractual characteristics of the

is not much related to the interest rate swap posi-

bank‘s assets and liabilities in a way that a par-

tion in the market. The Study also concludes that

ticular duration or maturity can be achieved. Since

banks with higher profitability manage their risk

direct restructuring may not be always possible or

better than other banks and hence take bigger po-

desired, the synthetic method comes to comple-

sitions in the interest rate swap market. The con-

ment the drawbacks of the direct method and to

clusion that interest rate risk profile of the bank

add a certain degree of flexibility to the asset liability management process. The synthetic

16

method relies on the use of capital markets in-

little more difficult as the data disclosure is not

struments; such as interest rate swaps, interest rate

very high outside US banks. The language in

futures and interest rate options; and customized

which banks in different countries disclose their

agreements to alter the balance sheet risk expo-

annual reports also limited the scope of our study.

sure. Among these instruments, interest rate

For Example the Chinese banks couldn’t be stud-

swaps are considered to be the most widely

ied because of annual reports being published in

used instrument for hedging against interest rate

Chinese language. We believe that in future more

risk.

disclosure information will be available in public domain for interest rate swaps, which will make

Limitations One problem that we encountered during the

the empirical study of the interest rate swap issues by researchers more interesting and reliable.

study is the limited disclosure of interest rate swap information by the banks. Most of the banks

About the Authors

only disclose the notional amount of swaps in

Prof. Jayadev M is a professor in the Finance

their annual reports, which doesn’t provide com-

and Control at the Indian Institute of Management

plete information about the risks of interest rate

Bangalore.

swap transactions. Russian banks don’t disclose their interest rate swap positions in their annual

Chirag Thakral is a second-year student at the Indian Institute of Management, Bangalore.

reports. So there is a need of a more detailed disclosure by the banks. Another limitation was the difficulty in

Sandeep Kumar is a second-year student at the Indian Institute of Management, Bangalore

identifying the interest rate swap from other swaps mentioned in the annual reports. In many cases, interest rate swaps include plain vanilla fixed for floating swaps, basis swaps, indexamortizing swaps, and other, more exotic types of swaps contracts. Different countries have different disclosure norms and hence different level of interest rate swap disclosure. Studying the commercial banks of emerging markets made the information extraction a

17

References 1. Ahmed . A S Beatty. A, Takeda . C , (1997), "Evidence on Interest risk management and derivatives usage by commercial banks", working paper, University of Rochester, NY. 2. Arnold, T.S. (1984). “How To Do Interest Rate Swaps”, Harvard Business Review. September/October, pp. 96-101. 3. Aspinwall, R. (1982). “Managing Interest Rate Risk.“ In: Proceedings of a Conference on bank structure and Competition. Chicago: Federal Reserve Bank of Chicago, pp.463470. 4. Badger, P. (1995), “Financial insurance, Interest Rate Risk, and Derivatives”, Bankers Magazine, volume 178, number 3, May/June. p.32-37. 5. Barton, J. “Doe the Use of Financial Derivatives Affect Earnings Management Decisions?" The Accounting Review 76 (2005): 126. 6. Bicksler, J. and Chen, A.H. (1986). “An Economic Analysis of Interest Rate Swaps“. Journal of Finance, volume 41, July, pp. 645-655. 7. Carter.D.A & Sinkey J.F , (1996); " An empirical investigation into the use of interest rate derivatives by US commercial banks. Working paper. Athens, GA: University of Georgia. 8. Carter.D.A & Sinkey J.F , (1998); " The use of interest rate derivative by end users: the case of large community banks, Journal of Financial Services Research 14 pp17-34. 9. Colquitt. L.L & Hoyt. R.E, (1997), "Determinants of corporate hedging behaviour," Journal of Risk and Insurance, Vol 64

10. Flannery. M.J & James. C.M, (1984), "The effects of interest rate changes on the common stock returns of financial institutions," Journal of Finance, Vol 39, pp 1141-1153. 11. Felgran, S.D. (1987). “Interest Rate Swaps: Use, Risk. and Prices”, New England Economic Review. November/December, pp. 2232. 12. Geczy. C, Minton. B.A, and Schrand. C, (1997), "Why firms use currency derivatives," Journal of Finance, Vol 52, pp1323-1354. 13. Goodman, Laurie S., (1990) “The use of interest rate swaps in managing corporate liabilities”, Journal of Applied Corporate Finance, 2 (4), 35-47. 14. Gunther. J.W, & Siems. T.F, (1996), "the likelihood and extent of bank participation in derivatives activities," Financial Industry Studies, Federal Reserve Bank of Dallas, Dallas, TX. 15. Haight, T.G. and Kehly, J.C.Jr. (1995). “The Role of Derivatives as an Asset/Liability Management Tool”. Bankers Magazine, volume 178, number 3, May/June, p.22-26. 16. Jagtiani, J. (1996). “Characteristics of Banks that are More Active in the Swap Market”. Journal of Financial Services Research, volume 10, pp. 131-141. 17. Jong and Driessen (1999) "LIBOR versus Swap Market Models: An Empirical Comparison", European Finance Review, 5(3), 201237. 18. Kim, S. and Koppenhaver, G.D. (1992). ‘An Empirical Analysis of Bank Interest Rate Swaps”. Journal of Financial Services Research. volume 7. number 1. pp. 57-72 19. Kangis. P, Elgar. J and Haar L, (1998), "The use of Derivatives Instruments for Hedging

18

Purposes," Journal of Financial Services Marketing, Vol 3, No 1, p 81. 20. Kuprianov, Anatoli. (Summer 1994) The role of interest rate swaps in corporate finance, Economic Quarterly, 80, 49-68. 21. Kwan. S.H, (1991), "Re-examination of interest rate sensitivity of commercial banks stock returns using of random-coefficient model," Journal of Financial Services Research, Vol 5, pp 61-76.

cies", Journal of Financial and Quantitative Analysis 20 , No . 4, pp. 583-601. 30. Shanker, L. (1996). “Derivatives Usage and Interest Rate Risk of Large Banking Firms”. The Journal of Futures Markets, volume 16, number 4, pp. 459-474. 31. Simons, K. (1993). “Interest Rate Structure and the Credit Risk of Swaps”. New England Economic Review, July/August, pp. 23-34.

22. Litzenberger, R.H. (1992). “Swaps: Plain and Fanciful”. Journal of Finance, volume 47, number 3, pp. 831-850.

32. Simons, K. (1995). “Interest Rate Derivatives and Asset-Liability Management by Commercial Banks”. New England Economic Review, January, February, p.l7-28.

23. Loeys, G.J. (1985). “Interest Rate Swaps: A New Tool for Managing Risk“. Federal Reserve Bank of Philadelphia, Business Review, May/June, pp. 17-25.

33. Smith, C. W., Smithson, C.W., Wakeman, L.M. (1988). “The market for interest rate swaps”. Financial Management 17, (4), 34 44, Winter.

24. Lyes, Boukrami (2002), “The Use of Interest Rate Swaps by Commercial Banks”, Manchester Metropolitan University

34. Song, C. J. “Are Interest Rate Swaps Used to Manage Bank Earnings?" Working paper, Michigan State University, 2005.

25. Remolona E.M. and P.D. Wooldridge (2003) ‘The Euro Interest Rate Swap Market’, BIS Quarterly Review, March.

35. Sun, T.S., Sundaresan, S., and Wang, C. (1993). “Interest Rate Swaps: An empirical Investigation”. Journal of Financial Economics, volume 34, number 1, pp. 77-99.

26. Ronn.E.L, Verma. A.K, (1986)," Pricing the risk adjusted: An option based model," Journal of Finance, Vol 41, pp 871-896. 27. Rudnick, L.T. (1987). “Discussion of Practical Aspects of Interest Rate Swaps”. In Proceedings of a Conference on Bank Structure and Competition. Chicago: Federal Reserve Bank of Chicago, pp. 206-213. 28. Sinkey. J.F and Carter. D.A, (1997), " Evidence on Financial Characteristics of banks that do and do not use derivatives," Working Paper, University of Georgia, Athens. 29. Smith Clifford W and Rene Stulz, (1985), " The Determinants of Firms' Hedging Poli-

36. Venkatachalam M. “Value-Relevance of Banks’ Derivatives Disclosures.” Journal of Accounting and Economics 22 (1996) 37. Wright, W.M. and Houpt, J.V. (1996). “An Analysis of Commercial Bank Exposure to interest Rate Risk”. Federal Reserve Bulletin, volume 82, number 2, February, p.l15-128. 38. Yu, W.T. and Pang, W.K. and Li, L.K. (2002) “Borrowing cost reduction by interest rate swaps––an option pricing analysis,” European Journal of Operational Research 154 (2004) 764–778. 39. Bloomberg Database

19