Document Text (Pages 41-50) Back to Document

Foreign Exchange Risk Management in Commercial Banks in Pakistan

by Sabri, Maroof Hussain, MS


Page 41

Model Summary

Model R R Square Adjusted R Square
Std. Error of the

Estimate
1 .167a .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 .398a
Residual 16.098 104 .155

Total 16.560 107
a. Predictors: (Constant), Size, Exchange Rate Volatility, Ownership Status
b. Dependent Variable: NFXNA
Coefficientsa

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


Page 42

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


Page 43

1 .960a .922 .921 7.06015E6

ANOVA
b
Model Sum of Squares df Mean Square F Sig.
1 Regression 6.264E16 1 6.264E16 1256.630 .000a
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

Coefficientsa

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


Page 44

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


Page 45

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


Page 46

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


Page 47

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


Page 48

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


Page 49

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


Page 50

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

© 2009 OpenThesis.org. All Rights Reserved.