# Foreign Exchange Risk Management in Commercial Banks in Pakistan

Model Summary

Model R R Square Adjusted R Square

Std. Error of the

Estimate

1 .167^{a }.028 .000 .39342807

a. Predictors: (Constant), Size, Exchange Rate Volatility, Ownership Status

ANOVA

b

Model Sum of Squares df Mean Square F Sig.

1 Regression .462 3 .154 .996 .398^{a}

Residual 16.098 104 .155

Total 16.560 107

a. Predictors: (Constant), Size, Exchange Rate Volatility, Ownership Status

b. Dependent Variable: NFXNA

Coefficients^{a}

Unstandardized Coefficients

Standardized

Coefficients

Model B Std. Error Beta t Sig.

1 (Constant) .784 .125 6.266 .000

Ownership Status -.054 .121 -.046 -.447 .656

Exchange Rate

Volatility

.336 .238 .139 1.414 .160

Size 9.561E-10 .000 .060 .580 .563

Table iv: Multiple Linear Regression Output of Relationship Between "NFXNA & "Size, OS & ERV"

Above mentioned tables show the output from the SPSS Statistics Processor. To

check whether there is any significant relationship between the dependent variable

and three independent variables is significant, F-Test is used and to check whether

partial regression coefficients are significant or not t-test is used. From the output

tables, the table showing the coefficients, it is evident that the three different

values of t-statistic, for three partial regression coefficients are not significant at

the desired level of significance. Further value of R Square is very small, which

shows that very small variation is explained because of linear relationship. The

value of F, calculated using ANOVA, is also not significant at the desired level of

Page | 41

significance. Since, the value of F-statistic is not significant, the multiple linear

regression is not overall significant.

Since, the regression is not overall significant, there is no relationship between

dependent variable & independent variables, the null hypothesis 1 is

substantiated. Now, based on our results, it can be said that “Net Foreign

Currency Exposure Relative to Net Assets” of commercial banks in Pakistan does

not depend on Size of Bank, Ownership Status of Bank & Exchange Rate

Volatility.

3. Relationship Between NFX & Net Assets

The below Model 2 is formed to check the relationship between Net Foreign

Currency Exposure and Net Assets of commercial banks in Pakistan.

��� � � � �

��^{��}

A scatter diagram of both variables is below:

Figure 1: Scatter Diagram of NFX & Net Assets (Model 2)

Simple linear regression is used in this model and the output from SPSS Statistics

processor is below:

Model Summary

Model R R Square

Adjusted R

Square

Std. Error of the

Estimate

Page | 42

1 .960^{a }.922 .921 7.06015E6

ANOVA

b

Model Sum of Squares df Mean Square F Sig.

1 Regression 6.264E16 1 6.264E16 1256.630 .000^{a}

Residual 5.333E15 107 4.985E13

Total 6.797E16 108

a. Predictors: (Constant), Net Assets

b. Dependent Variable: Net Foreign Currency Exposure

Table v: Output for Regression: NFX & NA

Coefficients^{a}

Unstandardized Coefficients

Standardized

Coefficients

Model B Std. Error Beta t Sig.

1 (Constant) -2420679.816 851025.994 -2.844 .005

Net Assets .973 .027 .960 35.449 .000

a. Dependent Variable: Net Foreign Currency Exposure

As per the above mentioned output of SPSS for model 2, it is evident that the

value of regression coefficient B is 0.973 with a standard error of 0.027. The

calculated value of t-statistic used to check its significance shows that it is

significant at the desired level of significance rather highly significant. Value of

R-square is very high which shows that there is a very strong relationship between

dependent and independent variables. To check the overall significance of model,

calculated value of F-Statistic is also significant.

In the light of the above results, null hypothesis 2 is rejected and the alternative

hypothesis 2 is accepted which states that there is a relationship between Net

Foreign Currency Exposure and Net Assets of Commercial Banks in Pakistan.

As there is a direct relationship between these two variables, if the Net assets of

the bank are changed, Net FX Exposure will also be changed. Model 2 differs

from the model 1 in that Model 1 takes into account NFXNA and studies its

relationship with Size of Bank which is measured by Net Assets of bank whereas

in model 2, Net Foreign Currency Exposure is related to Net Assets of bank.

Page | 43

4. Comparison of Net FX Exposure of Commercial banks in Private

Sector & Public Sector

While studying the Net Foreign Currency Exposure of Commercial Banks in

Pakistan, next objective is to compare the net foreign currency exposure of Public

Sector Commercial Banks (PSCB) and Local Private Banks (LPB). Results

(Output, using SPSS Statistics Processors) of the independent sample t-test are

below:

Group Statistics

Ownersh

ip Status N Mean Std. Deviation Std. Error Mean

NFXNA PSCB 14 .8557330 .36257606 .09690253

LPB 94 .7904280 .39891236 .04114467

Independent Samples Test

Levene's Test for Equality of

Variances t-test for Equality of Means

F Sig. t df

NFXNA Equal variances assumed .874 .352 .578 106

Equal variances not

assumed

.620 18.028

Independent Samples Test

t-test for Equality of Means

Sig. (2-tailed) Mean Difference

Std. Error

Difference

Page | 44

NFXNA Equal variances assumed .565 .06530500 .11305264

Equal variances not assumed .543 .06530500 .10527575

Table vi: Results of independent sample t-test to compare NFX of PSCR & LPB

Levene’s Test is used to check the Equality of variances, since, the value of F is

not significant, therefore, variances of both groups is not significantly different

and hence equal variances are assumed.

Assuming equal variances, the value of t=0.578 with degree of freedom 106 is not

significant at the desired level of significance. There is no difference between

means NFXNA of Public Sector Commercial Banks & NFXNA of Local Private

Banks. Hence our Null Hypothesis 3 is substantiated.

5. Comparison of Net Foreign Currency Exposure of Islamic Vs

Conventional Banks

Independent sample t-test is used to check if there is a significant difference

between Islamic & Conventional banks as far as there Net Foreign Currency

Exposure is concerned. Below are the results (output from SPSS Statistics

Processor):

Group Statistics

Type N Mean Std. Deviation Std. Error Mean

NFXNA Conventional 92 .7677321 .40961668 .04270549

Islamic 16 .9780714 .21426364 .05356591

Independent Samples Test

Levene's Test for Equality of

Variances t-test for Equality of Means

Page | 45

F Sig. t df

NFXNA Equal variances assumed 12.572 .001 -2.001 106

Equal variances not

assumed

-3.070 37.623

Table vii: Results of independent sample t-test to compare NFXNA for Islamic & Conventional Banks

Independent Samples Test

t-test for Equality of Means

Sig. (2-tailed) Mean Difference

Std. Error

Difference

NFXNA Equal variances assumed .048 -.21033936 .10509514

Equal variances not assumed .004 -.21033936 .06850595

Table viii: Results of independent sample t-test to compare NFXNA for Islamic & Conventional Banks

Here Levene’s test gives the value of F which is significant at p<0.01, therefore

equality (Homogeneity) of variances is not assumed. Using Independent Sample

t-test, value of t is -3.070 with degree of freedom of 37.623. This value of t is

significant as p-value is less than 0.01, therefore, it is established that there is a

significant difference between means of these two groups. Hence, null hypothesis

4 is rejected and alternative hypothesis 5 is accepted stating that there is a

significant difference NFXNA of Islamic Banks and Conventional Banks.

Page | 46

Findings on Usage of Different Tools for Foreign Exchange Risk

Management

Foreign Exchange Risk is managed by all the commercial banks in Pakistan and

the findings related to research objectives as listed in the previous sections are

given below in the same order as they appear previously:

5. Foreign Exchange Risk Management: Tools & Practices

Foreign exchange risk management practices as adopted by commercial banks in

Pakistan can be broadly listed as below:

•

•

•

Foreign Currency Portfolio Diversification

Foreign Currency Assets & Liabilities Matches

Use of Currency Derivatives

Foreign Currency Portfolio Diversification

Almost all commercial banks hold a portfolio of different foreign currencies i.e.

they deal in multiple currencies. Major currencies, in which they deal in and hold

open positions at the same time, are US Dollar (US$), Great British Pound (GBP),

Japanese Yen, Euro, UAE Dirham & Others.

Foreign Currency Assets & Liabilities Matches

Foreign currency assets and liabilities matching is a very common practice by the

commercial banks in Pakistan to hedge against foreign exchange risk. However,

such matches are strictly done within the limits. These limits are set internally by

the banks themselves, mostly by the Asset Liability Committee, as advised by the

State Bank of Pakistan. These limits control foreign currency exposure through

dealer limits, open foreign currency position limits & counterparty exposure

limits. Foreign currency assets and liabilities are also managed within the strict

limits as prescribed by the State bank of Pakistan.

Use of Currency Derivatives

The most important practice of managing foreign exchange risk involves currency

derivatives. Currency Derivatives are tools employed by every commercial bank,

however, selection of derivatives from the available ones vary from bank to bank.

In Pakistan, a descriptive analysis of such tools (currency derivatives), as used by

the commercial banks included in our sample, is given below:

Page | 47

^{Category }Frequency Percent Cumulative Percent

Only Forward Exchange Contracts 63 57.3 57.3

Forward Exchange Contracts & Swaps 33 30.0 87.3

Forward Exchange Contracts, Swaps & Options 14 12.7 100.0

Total 110 100.0

Table ix: Descriptives: Currency Derivatives Usage

In the above table, 110 commercial banks included in sample, use different mix of

currency derivatives.

•

•

•

•

Majority of banks use forward exchange contracts only.

Some use a mix of forward exchange contracts and currency swaps.

Few banks use a mix of forward exchange contracts, currency swaps &

foreign exchange options.

None of the bank uses Futures for the Foreign exchange risk management

The main reason behind the no usage of Futures is lack of futures exchange in

Pakistan. Same is the reason behind the lesser usage of foreign currency options.

FX options are both exchange traded and over the counter. As there is no

exchange for trading of FX options therefore the FX options being traded here by

the commercial banks are usually over-the-counter. Forward exchange contracts

& currency swaps are also over the counter. Forward exchange contracts is the

most common and most important tool used by the commercial banks in Pakistan

as every bank use it whereas the currency swaps are the second popular tool used.

Swaps are usually long term foreign exchange risk management tool. Few banks

also use options, only 12.7% of our sample i.e. listed commercial banks of

Pakistan. As there is no specific exchange for the exchange traded options in

Pakistan, therefore, the option used by few banks are over the counter ones.

6. Currency Derivatives Usage

Currency derivatives are crucial for the management of foreign exchange risk. In

Pakistan, all the commercial banks use currency derivatives. Even if their policy

is not to use Derivatives, they still use forward exchange contracts.

Results show that all the banks whether public sector commercial banks or local

private banks use forward exchange contracts. 110 banks out of sample of 110

banks use forward exchange contracts. However, 33 out of 110 banks use swaps

in addition to forwards and 14 out of a total of 110 banks use foreign currency

options along with forwards and swaps as well. This is evident from the results

Page | 48

mentioned in the above table that public sector commercial banks do not use

foreign currency options at all.

Currency Derivatives Usage by Ownership Status

Currency derivatives are used by both Public Sector Commercial Banks and Local

Private Banks. To what extent these banks use currency derivatives, below are the

results:

Ownership Frequency Percent Cumulative Percent

Public Sector Forwards 9 64.3 64.3

Forwards & Swaps 5 35.7 100.0

Total 14 100.0

Local Private Forwards 54 56.3 56.3

Forwards & Swaps 28 29.2 85.4

Forwards, Swaps & Options 14 14.6 100.0

Total 96 100.0

Table x: Currency Derivatives Usage by Ownership Status

Descriptive study, to check whether Islamic banks or conventional banks use

currency derivatives or not, is conducted. Findings in this context are mentioned

in a table below:

Currency Derivative Usage by Type of Bank

Type Frequency Percent

Cumulative

Percent

Conventional Forwards 52 55.3 55.3

Forwards & Swaps 28 29.8 85.1

Forwards, Swaps &

Options

14 14.9 100.0

Total 94 100.0

Islamic Valid Forwards 11 68.8 68.8

Forwards, Swaps 5 31.3 100.0

Total 16 100.0

Table xi: Currency Derivative Usage by Type of Bank

Page | 49

Sample contains 110 banks in total, including 16 Islamic and 94 conventional

banks. All the banks, both Islamic and conventional banks, use forward exchange

contracts. Only 5 out of 16 Islamic banks use swaps along with forwards and none

of Islamic bank use foreign exchange options. However commercial banks use all

the three types of currency derivatives i.e. forwards, swaps and options. It is

evident from the above findings that only 14 out of 94 conventional banks use

foreign currency options. Therefore, the most popular tool is forward exchange

contracts and the least popular and used tool is foreign currency options.

Findings on Factors that affect Currency Derivative Usage

Using Binary Logistic Regression it is tried to establish what are the different

variables that affect the selection of tools for a commercial bank. A set of five

independent variables (Size of bank, Ownership Status of Bank, Type of Bank,

Net Foreign Currency Exposure and Exchange rate Volatility) are included in

model to check their relationship with tools selection as a dependent variable. The

objective is to find out the most significant variables which influence dependent

variable. For this purpose, Stepwise Backward method based on the Likelihood

ratio test is used with the help of SPSS.

Backward Stepwise Binary Logistic Regression took place in the following steps,

in order to reach the significant independent variables:

1. In the beginning the model is fitted with only constant and no independent

variable. After this model is again fitted using all of the five independent

variables. Therefore, step 1 involves the model with all the independent

variables in it.

2. In step 2, model is refitted removing independent variable “Type of Bank”

which proved to be insignificant on the basis of Likelihood Ratio test.

3. In step 3, another independent variable “Current year exchange rate

volatility” is removed from the model and the number of independent

variables in the model is reduced to three. These independent variables

include Ownership Status of Bank, Size of Bank (as measured by Net

Assets) and Net foreign currency exposure relative to net assets.

Both Type & Current ERV are removed from model based on the Likelihood ratio

test. The summary of results of likelihood ratio test is given below:

Model if Term Removed

Variable

Model Log

Likelihood

Change in -2 Log

Likelihood df

Sig. of the

Change

Page | 50