Report On

Are Non-Performing Loans Sensitive to Macroeconomic Determinants? An

Empirical Evidence from Banking Sector of SAARC Countries

By

Saom Shawleen
ID: 17104047

A thesis report submitted to the BRAC Business School in partial fulfillment of the
requirements for the degree of
Bachelors of Business Administration

BRAC Business School

BRAC University
September, 2021

© 2021. BRAC University

All rights reserved.
 Declaration

It is hereby declared that

1. The thesis report submitted is my own original work while completing degree at Brac

University.

2. The report does not contain material previously published or written by a third party, except

where this is appropriately cited through full and accurate referencing.

3. The report does not contain material which has been accepted, or submitted, for any other

degree or diploma at a university or other institution.

4. I have acknowledged all main sources of help.

Student’s Full Name & Signature:

_________Saom Shawleen, 17104047___________

Student Full Name
Student ID

Supervisor’s Full Name & Signature:

___________________________________________
Supervisor Full Name
Designation, Department
Institution

ii
 Letter of Transmittal

Mr. Jubairul Islam Shaown

Lecturer,
BRAC Business School
BRAC University
66 Mohakhali, Dhaka-1212

Subject: Submission of the project paper on “Are Non-Performing Loans Sensitive to

Macroeconomic Determinants? An Empirical Evidence from Banking Sector of SAARC
Countries”

Dear Sir,

It is a great pleasure and an honor for me to submit this project report on “Are Non-Performing
Loans Sensitive to Macroeconomic Determinants? An Empirical Evidence from Banking
Sector of SAARC Countries”. This paper is prepared as the BBA program’s partial requirement
under the BRAC Business School, BRAC University. I have put my best effort into preparing
this report by following thoroughly the guidelines of formal report writing. I have attempted
my best to finish the report with the essential data and recommended proposition in a significant
compact and comprehensive manner as possible.

I would like to express my heartiest gratitude to you for providing me with precious advice and
appropriate guidelines that helped me to prepare this study thoroughly. I sincerely supplicate
you to call me if you perceive further study should be conducted on this thesis.

Sincerely yours,

Saom Shawleen
_______________________
Saom Shawleen
ID 17104047
BRAC Business School
BRAC University
Date: 09-11- 2021

iii
 Acknowledgment

First of all, I would like to express my earnest gratitude from the core of my heart to the

Merciful, Beneficent, and Almighty Allah for providing me strength and persistence to

accomplish this study within the programmed time.

Afterward, I show gratitude to my respected parents. They are one of the main reasons that I

am now in BRACU and able to do this kind of project work. They consistently supported me

mentally to accomplish the project.

I take the privilege of extending acknowledgment to my respected instructor, Mr. Jubairul

Islam Shaown, Lecturer, BRAC Business School, BRAC University; who encouraged me to

conduct this rigorous analysis. I would like to express my heartfelt gratitude to him for helping

me to accomplish this report with his skillful support, relentless effort, and extraordinary

gesture. I have attempted to incorporate his knowledgeable opinions and constructive

suggestions in this research paper in the best plausible way to elevate its qualitative worth.

Lastly, I am enormously grateful to him for providing me with relevant materials to accomplish

this report.

iv
 Abstract

The study empirically investigates selected macroeconomic determinants of nonperforming

loans (NPLs) for a panel of 8 SAARC countries (Afghanistan, Bangladesh, Bhutan, India,

Nepal, Maldives, Pakistan, and Sri-Lanka), using annual data for the period 2008-2019. To

examine the association, this study, primarily, conducted OLS model, fixed-effect estimates,

and random-effect estimates and, eventually, applied robust fixed effect estimates to resolve

the problem of heteroscedasticity. The empirical findings confirmed the previous findings,

indicating a significant positive association with the government budget balance and a

significant inverse relationship with GDP, sovereign debt, inflation rate, and money supply.

The findings are useful for formulating macro-prudential along with fiscal policies to avoid the

subsequent NPLs shock in SAARC countries.

Keywords: Non-performing loans; Macroeconomic determinants; SAARC countries; Panel

data; South Asia, Financial crisis

v
 Table of Contents

Declaration................................................................................................................................ii

Letter of Transmittal ............................................................................................................. iii

Acknowledgment ..................................................................................................................... iv

Abstract ..................................................................................................................................... v

Table of Contents .................................................................................................................... vi

List of Tables ........................................................................................................................ viii

List of Figures .......................................................................................................................... ix

List of Acronyms ...................................................................................................................... x

Glossary ................................................................................................................................... xi

Chapter 01 Introduction ......................................................................................................... 1

1.1 Study Background ........................................................................................................ 2

1.2 Objectives .................................................................................................................... 3

1.2.1 Primary Objective ................................................................................................. 3

1.2.2 Specific Objectives ............................................................................................... 3

1.3 Research Gap ............................................................................................................... 3

1.4 Rationale of Study........................................................................................................ 4

1.5 Limitations of Study .................................................................................................... 5

Chapter 02 Review of Literature ............................................................................................ 6

2.1 Background of Non-Performing Loans ....................................................................... 7

2.2 Empirical Literature ..................................................................................................... 7

vi
Chapter 03 Research Framework ........................................................................................ 12

3.1 Samples and Data Collection ..................................................................................... 13

3.2 Variables of Study...................................................................................................... 13

3.2.1 Measured Variable .............................................................................................. 13

3.2.2 Explanatory Variables ......................................................................................... 13

3.3 Conceptual Structure .................................................................................................. 18

3.4 Econometric Framework ............................................................................................ 18

3.5 Analytical Techniques ............................................................................................... 19

Chapter 04 Diagnostic Tests & Findings ............................................................................. 21

4.1 Descriptive Statistics .................................................................................................. 22

4.2 Multi-collinearity Test ............................................................................................... 24

4.2.1 Correlation Matrix .............................................................................................. 24

4.2.2 Variance Inflation Factor (VIF) Test .................................................................. 25

4.3 Regression Analysis: Fixed and Random Effect Model ............................................ 25

4.4 Hausman Test............................................................................................................. 26

4.5 Test of Heteroscedasticity .......................................................................................... 26

4.6 Test of Autocorrelation .............................................................................................. 27

4.7 Test of Cross-Sectional Interdependence................................................................... 27

4.8 Model Specification Result ........................................................................................ 28

4.9 Discussion .................................................................................................................. 29

Chapter 05 Concluding Remarks ......................................................................................... 32

vii
 5.1 Conclusion and Policy Implications .......................................................................... 33

5.2 Future Research Avenues .......................................................................................... 34

References ............................................................................................................................... 35

Appendix A: Pooled OLS Estimation .................................................................................. 44

Appendix B: Fixed Effect Estimation .................................................................................. 45

Appendix C: Random Effect Estimation ............................................................................. 46

Appendix D: Individual Data Source ................................................................................... 47

List of Tables

Table 1 The variables and their anticipated relationships ........................................................ 17

Table 2 Descriptive statistics of selected variables.................................................................. 22

Table 3 Matrix of correlations ................................................................................................. 24

Table 4 VIF Test ...................................................................................................................... 25

Table 5 Hausman (1978) Specification Test ............................................................................ 26

Table 6 Robust Standard Error Fixed Effect Regression Results ............................................ 28

Table 7 Appendix A: Pooled OLS Estimation ......................................................................... 44

Table 8 Appendix B: Fixed Effect Estimation ......................................................................... 45

Table 9 Appendix C: Random Effect Estimation .................................................................... 46

Table 10 Appendix D: Individual Data Source........................................................................ 47

viii
 List of Figures

Figure 1 Conceptual Framework ............................................................................................. 18

Figure 2 Trend of average NPLs rate during the period of 2008-2019 in SAARC countries . 23

ix
 List of Acronyms

CEIP Carnegie Endowment for International Peace

EU European Union

FE Fixed Effect

FSIs Financial Soundness Indicators

GCC Gulf Cooperation Council

GDP Gross Domestic Product

GFC Global Financial Crisis

GMM Generalized Method of Moments

IMF International Monetary Fund

MENA Middle East and North Africa

MS Money Supply

NPLs Non- Preforming Loans

OECD Organization for Economic Co-operation and Development

OLS Ordinary Least Square

RBI Reserve Bank of India

RE Random Effect

SAARC South Asian Association for Regional Cooperation

VAR Vector Auto Regression

VIF Variance Inflation Factor

x
 Glossary

Non-Performing Loans Non-performing loans (NPLs) refers to default loan that the

(NPLs) borrowers are unable to pay interest and principal amount

within a specified period (generally due >90 days) (IMF,

2005)

Money Supply (M2GDP) Money supply refers to the aggregate stock of money in an

economy for a specific period. In general, money supply

(MS) is classified as Reserve Money (M0), Narrow Money

(M1), Broad Money (M2) based on size and type of account

(Badar et al., 2013)

Exchange Rate (EXC) The exchange rate measures the domestic currency’s worth

with another currency (Zameer & Siddiqi, 2010)

Government Budget Government Budget Balance (also called as public fiscal

Balance (FISCAL) balance) refers to the gross difference between federal

revenues and spending that measures the comprehensive

fiscal performance of the government

Gross Domestic Product GDP refers to the aggregate value of finished goods and

(GDP) services that are produced in a country within a specified

period (Badar et al., 2013)

Inflation Rate (INFL) Inflation is referred to as a price spiral for goods and

services for a specified period in a particular economy.

Sovereign Debt (DEBT) Sovereign debt refers to the central government

consolidated debt to finance the trade deficit and/or budget

deficit.

xi
Chapter 01 Introduction
1.1 Study Background

Non-performing loans (NPLs) are those financial assets that remain unpaid and therefore, no

installment and/or interest payments are received by banks as planned. According to IMF

(2005), the loans are considered as NPLs when it does not generate principal or interest rate

for at least 90 days. Moreover, Joseph et al. (2012) denoted that when the loans are past due,

generally more than ninety days and unable to ‘perform’/ receive interest anymore. However,

when the banks are unable to collect the total principal or interest on the due date and there is

no possibility of repayment in foreseeable future, this is called non-performing loans (Alton &

Hazen, 2001) and this study is based on their NPLs concept. Besides, the high inflation rate,

vulnerable fiscal and monetary policy, and weak economic activities are primarily responsible

for increasing the bank’s exposure to credit risk and, consequently, threaten the financial

stability.

Increasing NPLs are considered as a major proxy of credit risk since the entire banking system

is directly impacted by NPLs. The financial crisis in Asia during 1997 and the financial crisis

of 2007-2008 are the most vivid paradigm of how non-performing loans lead to an unstable

banking system. The NPLs ratio is one of the vigorous indicators of the onset of the banking

crisis as it reduces the credit growth (Reinhart & Rogoff, 2010) and, hence, overall economic

stability significantly disrupts (Ivanović, 2016). Rising NPLs ratio is the indicator of a

susceptible financial system, while a lower rate of NPLs is a signal of financial soundness.

High NPLs further affect any country’s respective commercial banks and ultimately

commercial banks ’ significant exposure to credit risk jeopardizes entire financial system and,

thereby, state’s economy (Feijó, 2011). Certainly, a consistent upward shift of NPLs adversely

affect the financial efficiency and thus chance of the banking crisis are introduced (Louzis et

al., 2012; Nkusu, 2011). More precisely, NPLs reduce the investment opportunities, restraints

interest revenues, and boost the liquidity crisis that is initially responsible for bankruptcy in a

2
financial system. Hence, the prerequisite of maintaining financial and economic stability is to

detect the determinants that affect NPLs.

The remainder structure of the study is as follows. Sect. 02 reviews empirical pieces of

literature on macro-economic determinants of non-performing loans. Sect. 03 explains data

sources and study variables, formulates hypotheses, illustrates the econometric framework, and

discusses analytical techniques. Sect. 04 depicts the diagnostic tests and analyzes the research

findings while section 05 includes concluding remarks.

1.2 Objectives

1.2.1 Primary Objective

The study purpose is to explore the macroeconomic determinants that have a noteworthy

influence on non-performing loans (NPLs) in the banking sector of SAARC countries.

1.2.2 Specific Objectives

 To gain a deep insight regarding NPLs.

 To identify the influence of macro-economic determinants on NPLs.

 To consider the analytical finding’s implications from practical perspectives.

1.3 Research Gap

Although various studies examined the determinants of non-performing loans in worldwide,

according to the best knowledge of the author, no studies have been conducted previously on

the impact of macro-economic determinants on NPLs based on the banking system of the

SAARC countries. Here, the paper attempted to evaluate the non-performing loans’ sensitivity

in SAARC countries, using six macroeconomic determinants (broad money supply, exchange

rate, government budget balance, GDP, inflation rate, and sovereign debt) under the panel

dataset and static panel estimation technique.

3
1.4 Rationale of Study

Although in 2007-2008, the world economy experienced Global Financial Crisis (GFC), the

great recession mainly affected America, EU countries (especially Ukraine), Latin America

(especially Argentina), and African Countries (especially Jamaica). Conversely, according to

Carnegie Endowment for International Peace (2009), the Asian countries i.e. China, India, and

Japan were listed as the ‘least affected countries’. The GFC has more or less affected all of the

SAARC countries significantly.

Afghanistan is overburdened with strategic and political problems and the country is highly

reliant on foreign aid. According to World Bank (2012) almost 90% of Afghanistan’s budget

is dependent on foreign aid. During the financial crisis, the big donors were severely affected,

and therefore, in 2011 their donation fell by 3% (OECD, 2011). Hence, the financial crisis in

2008 highly affected Afghanistan’s crisis-stricken banks. Pakistan’s banking system also

experienced an increasing NPLs volume by almost two-fold from 2009-20011 (Badar et al.,

2013). The status of NPLs of Bhutanese and Maldives banks are somewhat satisfactory as in

recent years, the non-performing loans ratio are declining from 13.3% in 2008 to 8.45% in

2019 and 16.40% in 2009 to 9.39% in 2019 respectively (IMF, 2021). Sri Lankan banking

system also witnessed NPLs shock in recent decades and according to the study of Fonseka

(2009) Bangladeshi banking system experienced highest and Sri Lankan banking system

witnessed second highest NPLs among Bangladesh, Malaysia, Sri Lanka, Indonesia, Thailand,

and Philippines.

Despite Bangladesh, India and Nepal are currently holding the position of the strongest

economy (Investopedia, 2020), the scenario of the Nepalese banking system is alarming as the

aggregate NPLs are undoubtedly in an upward trend from 2010 (Koju et al., 2018). Moreover,

Bangladesh is highly burdened with non-performing loans in the last few decades. Based on

the international standard, the NPLs rate of 2% or below is considered as standard but
4
unfortunately, the NPLs rate of Bangladesh is 5 to 6 fold higher in recent years (Kumar et al.,

2020). Furthermore, the Reserve Bank of India (2017) stated that the financial health of the

Indian banking system deteriorated the asset quality because of the minimum capital

requirement (as per the Basel norm). Hence, the banking groups witnessed a crucial surge in

NPLs in recent years.

The continuous upward trend refers that the banks are heading towards a liquidity crisis and an

unstable financial stability will prevail in the recent future. Hence, maintenance of long-term

economic and financial stability significantly requires assessing the macroeconomic variables

of NPLs in SAARC economy.

1.5 Limitations of Study

One of the main limitations of this paper is time constrain. The report might have provided

better analysis if the period was extended. This report is conducted based on only static panel

estimation, whereas dynamic panel estimation could have been used to avoid biased and

inconsistent results (if any). Lastly, the sample size of this paper is <100, and therefore, further

inference of results is somewhat questionable.

5
Chapter 02 Review of Literature

6
2.1 Background of Non-Performing Loans

According to the published data at the country level and overall banking system level, the

amount of nonperforming loans and its factors are increasing significantly in recent years. The

surge in the credit risk during and after the global crisis took the researchers’ attention and,

therefore, factors that negatively impact the bank’s portfolio has been severely investigated.

The result of several studies revealed that a surge of irresistible problematic loans along with

the banking sector’s fragility and financial vulnerability undermined the banking crisis in ’90s.

Due to a negative shock in social welfare and economic growth, González-Hermosillo (1999)

Barseghyan (2010), and Zeng (2012) concluded the Non-Performing Loans as ‘financial

pollution’.

Keeton (1999) studied 50 US banks between 1982 and 1996 and showed that the lax credit

standards are one of the pivotal reasons for a sudden surge in NPLs. In line with the previous

study, McGoven (1998) also found that unsecured loans, low credit standards, and borrowers’

attitudes have a crucial impact on raising loan loss in the US banking system. Moreover, a

study on banking system of Argentina between 1993 and 1996 was conducted by Bercoff et al.

(2002) where he applied an Accelerated Failure Time (AFT) method and that revealed both

bank-specific1 and macroeconomic determinants had equal influence on NPLs.

2.2 Empirical Literature

This paper reflects several researchers’ investigative and analytical studies regarding the

factors that impact the non-performing loans for both individuals and a panel of countries.

Some of the researchers investigated both macroeconomic and bank specific variables, while

others investigated macro-economic or bank-specific factors separately.

1
Bank specific variables are solvency rate, operating efficiency, quality of asset, leverage and deposits, bank size,
liquidness etc.

7
Using macroeconomic determinants, Babouček and Jančar (2005) explored the NPLs in the

Czech Republic for 11 years and used the VAR method to examine the macroeconomic

determinants’ (real GDP, unemployment percentage, inflation, , exports, imports, interest

percentage, exchange rate and aggregate bank loans) influence. Their findings demonstrated

that GDP growth rate reduces the NPLs ratio while raising inflation and exchange rate

deteriorate the bank’s loan portfolio quality.

The correspondence between NPLs and macroeconomic along with bank-specific determinants

was explored by Espinoza and Prasad (2010) while considering a panel of 80 banks from the

GCC zone. The study highlighted that macroeconomic variables, particularly interest rate and

non-oil GDP had remarkable effect on the credit risk. Moreover, the study also depicted that

some distinct bank-specific factors (credit growth, capital sizes, and efficiency) had an impact

on non-performing loans. The study results suggested a short, yet, strong feedback effect from

the banking industry to the economy.

Kastrati (2011) analyzed the impact of NPLs ratio during the period of 1994-2009 on 15

transition countries (Azerbaijan, Albania, Armenia, Bosnia and Herzegovina, Belarus,

Bulgaria, Moldova, Macedonia, Kosovo, Romania, Serbia, Montenegro, Croatia, Georgia, and

Ukraine) by using dynamic panel data method. The report demonstrated the nonperforming

loans were highly persisting from one year to another and inflation rate, real economic growth

rate and competition had a noteworthy impact on the NPL ratio.

Using dynamic GMM and fixed effect model, Ghosh (2015) analyzed the bank specific and

economic variables of aggregate NPLs by taking 50 banks in both Columbia and USA between

1984 and 2013. The results implied that increasing GDP, housing price index, and personal

income growth rate declines NPLs, while sovereign debt and rate of unemployment increase

8
the NPLs significantly. Afterward, Konstantakis et al. (2016) confirmed the impact while

conducting a study on the Greek economy from 2001 to 2015.

Boudriga et al. (2010) conducted an empirical investigation based on 12 selected MENA

countries considering a sample of 46 banks and analyzed the impact of nonperforming loans

on the bank-specific, institutional, and business environment factors for the 2002-2006 year

timespan. Their result revealed foreign participation from developed countries, institutional

environment, loan-loss provision, and credit growth possess a significant impact on bad debt.

The factors of nonperforming loans in the Greek Bank during 2003Q1-2009Q3 were

scrutinized by Louzis et al. (2012) where they used dynamic panel data estimation techniques

and considered various types of loans (mortgage, consumer, and business). The study findings

exhibited that the NPLs ratio of business loans were, primarily, highly sensitive to the

unemployment rate change, the NPLs ratio of consumer loans were highly sensitive to real

growth rates changes, while NPLs ratio of mortgage was comparatively less sensitive to change

in macroeconomic environment. Moreover, they presumed the hypothesis of ‘sovereign debt’

and asserted that higher sovereign debt leads to rising NPLs.

De Bock and Demyanets (2012) investigated the macroeconomic variables of NPLs spanning

from 1996-2000 by considering a group of 25 emerging markets. Their study highlighted that

economic expansion, trade growth (goods), exchange rate, capital flows had a significant

impact on NPLs.

Messai and Jouini (2013) conducted an empirical study on 85 banks in Spain, Italy, and Greece

where they evaluated both macroeconomic and bank specific variables of nonperforming loans

on for a period of 2004 to 2008 and found a significant relationship between financial and

macroeconomic variables (i.e. rate of unemployment, rate of GDP growth, loan loss reserves,

and return on assets), real interest rate and nonperforming loans.

9
Caporale et al. (2014) investigated the impact of financial and macroeconomic determinants

on the quality of bank portfolio based on the Italian banking sector spanning from 2008-2012

and they found that economic recession leads to a high volume of NPLs since the banks grant

high volume of credits during economic boom. The findings revealed that the deterioration of

economic condition caused a record in NPLs surplus in the recessionary period and, during the

pre-crisis years, the Italian banks promoted the lending policy.

Reinhart and Rogoff (2011) investigated a group of 70 developed and developing countries

consisting of 209 sovereign default2 and 290 banking crises episodes for a prolonged span from

1800 to 2009. In the seminal paper, they revealed that the sovereign default/ government default

affected the quality of bank portfolio, irrespective of commitment size. In line with the findings,

a strong linkage between sovereign debt and nonperforming loans rate was explored by Makri

et al. (2014) while investigating 14 EU countries from 2000 to 2008.

The effect of financial crisis 2007-2008 on financial soundness indicators of banks were

investigated by Kasselaki and Tagkalakis (2014) on 20 industrialized OECD countries

spanning the period from 1997 to 2009. The study demonstrated that the global financial crisis

caused a significant increase in nonperforming loans, real interest rates (both short and long

term). Authors suggested the policymakers to develop a prior warning system so that the

stability of the banking sector is fragile/threatened or not will be known beforehand.

Chaibi and Ftiti (2015) examined the macroeconomic impact of nonperforming loans on

quality of bank’s portfolio where the laisse-faire economy (France) was compared with the

credit-based economy (Germany) for the span of 2005-2011. The authors found that, in both

of the economic contexts, all the chosen macroeconomic determinants (excluding the rate of

2
Sovereign default is the failure of a government to repay some or all of the country’s outstanding debt
(Investopedia, 2021).

10
inflation) had a noteworthy influence on NPLs. The findings suggested that a laisse-faire

economy possesses a greater credit risk than a credit-based economy.

The empirical study of Castro (2013) demonstrated that macroeconomic environment and

nonperforming loans have noteworthy correlations in the euro-zone countries. Moreover, the

study considered the financial variable’s impact on the NPL ratio.

Individual empirical analysis of Jakubík and Reininger (2013), Škarica (2014), Klein (2013)

confirmed the prior findings that the real economy affects significantly to nonperforming loans

in the observed countries, while they examined the factors of NPLs for Eastern and South-

Eastern, and Central European countries and their findings.

Bofondi and Ropele (2011) explored the macroeconomic variables of NPLs on the banking

system of Italy between 1990Q1 and 2010Q2. The study highlighted that aggregate money

supply, lending rates, and rate of unemployment are directly associated with NPLs and GDP is

negatively interrelated with NPLs. Contrarily, Ahmad (2003) conducted a study on NPLs as a

proxy of credit risk based on 65 Malaysian deposit-taking institutions, and Kalirai and

Scheicher (2002) analyzed the NPLs impact on the Austrian banking system, and both of them

found a significant negative association between credit risk and money supply. Fofack (2005),

however, found no impact of money supply on nonperforming loans.

Lastly, Khemraj and Pasha (2009) scrutinized the determinants of nonperforming loans in

Guyana spanning from 1992 to 2004. The findings showed that the real effective exchange

rate, the high lending rate had a positive relationship with nonperforming loans while GDP had

a negative relationship with NPLs. In line with the previous study, Beck et al. (2013) explored

the NPLs’ determinants for 75 developed and emerging economies for 2000-2010 and found a

significant impact on lending rate, share price, nominal effective exchange rate (based on local

currency), and GDP growth rate.

11
Chapter 03 Research Framework

12
3.1 Samples and Data Collection

The study examines the impact of various macroeconomic determinants on NPLs for the eight

SAARC countries. The paper included an extended and the most recent period of 12 years

(2008-2019) to reflect a realistic scenario of the influence of macroeconomic determinants on

the bank’s loan portfolio during the financial crisis of 2007-2008 and post-crisis condition. The

analysis of study was carried out based on secondary data and these were extracted from the

IMF, World Bank and the Annual Report/Bulletin of some of the country’s respective Central

Bank.

3.2 Variables of Study

3.2.1 Measured Variable

3.2.1.1 Non-Performing Loans (NPLs)

According to IMF (2005), non-performing loans (NPLs) refers to the default loan that the

borrowers are unable to pay interest and principal amount within a specified period (generally

due >90 days). In other words, when the scheduled payment is overdue and no longer likely to

be paid in the foreseeable future, then the loans and advances become non-performing (Alton

& Hazen, 2001). Hence, NPLs are the pivotal unit for measuring the loan loss. Since the non-

performing loans ratio measures the bank’s financial soundness of credit portfolio and asset

quality, therefore, this study considers NPLs ratio as a measured variable and this has been

taken as a proxy of credit risk.

3.2.2 Explanatory Variables

3.2.2.1 Money Supply (M2GDP)

Money supply denotes to the aggregate stock of money in a financial system for a specific

period. In general, Money Supply (MS) is classified as Reserve Fund (M0), Narrow Money

(M1), Broad Money (M2) based on size and account types. As a proxy of money supply, this

13
study considered Broad Money (M2) since it includes both M0 and M1. Moreover, reserve

money, also known as central bank money is a central bank’s obligation that comprises

currency and depository accounts of the central bank. Narrow money (M1) comprises M0 and

all scheduled bank’s time and demand deposits, while broad money consists of M1 and all

foreign currency deposits. Moreover, the money supply has a significant impact on

nonperforming loans since the money supply influences borrowers’ behavior. During the

expansionary monetary policy, the required reserve rate and discount rate get reduced and

therefore, productivity and profitability increase (N. H. Ahmad & Ariff, 2007). Increased

money supply, thereby, revitalizes the investment and consumption pattern which consequently

increases income. Additionally, increasing the money supply decreases the cost of funds that

results in cheaper funds. These consequences escalate the borrower’s ability to pay the

outstanding on due time and thus bank’s credit risk exposure declines. Hence, the study

formulates:

H1. Broad Money Supply is inversely related to NPLs

3.2.2.2 Exchange Rate (EXC)

In developing economies, exchange rate volatility creates economic instability. The exchange

rate measures the domestic currency’s worth with another currency (Zameer & Siddiqi, 2010).

The problem that occurs frequently is the rising exchange rate for the home country meaning

foreign currency appreciation against the home currency and, therefore, home country’s cost

of imported goods increase. When the exchange rate increases, the local currency value

depreciated and hence the local goods and commodities become cheaper. Consequently,

domestic country’s export increase, and imports become costly. Therefore, local customer

retention becomes challenging as the price of the final product (which is sold locally) becomes

expensive. As the value of local currency depreciates, costly imported inputs create pressure

on the financial letter of credit which is issued by a commercial bank to the traders, and

14
therefore, the bank’s default risk increases (Sirpal, 2009). In other words, currency depreciation

has an expansionary impact that leads to an increased operating profit in terms of export as the

export becomes cheaper but causes shrinkage in import due to the opposite rationale (Nucci &

Pozzolo, 2001). Additionally, significant local currency depreciation deteriorates the net worth

of a firm, primarily, via the ‘balance sheet effect3 (Pratap & Urrutia, 2004). Hence, the study

proposes:

H2. Exchange rate fluctuation is positively related to NPLs

3.2.2.3 Gross Domestic Product (GDP)

Gross Domestic Product (GDP) is a significant economic element in the economic cycle that

measures economic development. The relationship between credit risk exposure of banks and

the economic cycle is dialectical. During the economic stagnation/recession the credit risk of

financial intermediaries generally increases since the economy suffers to maintain the

employment, prices, and outputs at the desired level. On the flip side, during the economic

boom, increasing economic activities lead to an increase in cash volume for both households

and businesses. Again, confidence among the lenders and borrowers boosts up for new

investment and increasing borrower’s income level strengthen their capacity to repay the

financial outstanding (Poudel, 2013). Therefore, the sequential impacts lead to a reduction of

credit risk for commercial banks. Therefore, the study originates:

H3. GDP growth rate has a negatively related to NPLs.

3.2.2.4 Government Budget Balance (FISCAL)

Government Budget Balance (also called public fiscal balance) refers to the gross difference

between federal revenues and spending that measures the comprehensive fiscal performance

3
Balance sheet-effect refers to the fact that financial net worth declines due to real depreciation and therefore
increases the burden of dollar-dominated outstanding debt for any firm.

15
of the government. During the expansionary fiscal policy, the government uses budgetary

instruments (e.g. raise government spending/decrease tax) that increase the money supply

(Dimitrios et al., 2016) and boosts productivity. Hence, the expansionary fiscal policy (which

leads to the budget deficit) alleviates/mitigates the credit risk. Hence, the study originates the

following hypothesis:

H4. Government Budget Balance is positively associated with NPLs.

3.2.2.5 Inflation Rate (INFL)

Inflation is referred to as a price spiral for goods and services for a specified period in a

particular economy. Since inflation depreciates the original value of money, hence, high

inflation leads to the high cost of borrowing/ loan interest, and therefore, the borrower’s

obligation increases and that results in an increased default risk (Poudel, 2013). Based on the

Price Stability Indicator, high inflation degrades the borrower’s real income and their loan

repayment capability decrease, whereas, low inflation leads to economic growth. Conversely,

increasing inflation rate decreases the aggregate value of loans and, thereby, the borrower’s

ability to timely pay their financial obligations increase that ultimately reduce the default risk

(Koju et al., 2018). Hence, correlation between inflation rate and nonperforming loans might

be ambiguous (Nkusu, 2011). Therefore, the study formulates the following hypothesis:

H5. Inflation has a positive/negative relationship with NPLs.

3.2.2.6 Sovereign Debt (DEBT)

Sovereign debt refers to the central government consolidated debt for financing the trade deficit

and/or budget deficit. This is also termed as public debt/general government gross debt/per

capita outstanding debt. Sovereign debt is, primarily, caused due to trade deficit or budget

deficit (Koju et al., 2018). When sovereign debt increases, the government might take some

16
fiscal measures i.e. curtail the public social spending4 (Louzis et al., 2012). These

consequences render a negative shock in household income and, therefore, borrowers become

unable to pay their outstanding on due time. Hence, the aggregate NPLs increase drastically.

Hence, study generates:

H6. Sovereign debt is positively related to NPLs.

Table 1 The variables and their anticipated relationships

Notations Variable Explanation Source(s) Estimated Effect

NPLs Ratio of Non-performing Loans to Total Bank World Bank, IMF & * (dep. v)
Loans Annual Report
M2GDP Broad Money Supply as % of GDP World Bank (-)
EXC Exchange Rate (average) IMF (+)
GDP GDP Growth Rate as annual percent World Bank (-)
FISCAL Government Net Lending/Borrowing as % of GDP World Bank (+)
INFL Inflation on Consumer Prices as annual % World Bank (+)/(-)
DEBT Sovereign Debt, Consolidated as % of GDP IMF (+)
Source: Author’s Compilation

4
Public social spending refers to the general government expenditures where the government offers tax-breaks,
cash benefits and other direct in-kind support of goods and services especially to the low-income/ underprivileged
households, unemployed, sick, disabled or elderly individuals.

17
3.3 Conceptual Structure

The conceptual structure for this paper is illustrated as follows:

Broad Money
(M2GDP)

Exchange Rate
(EXC)

Government
Budget Balance
(FISCAL)
Non-Performing
Loans (NPLs)
Gross Domestic
Product (GDP)

Inflation Rate
(INFL)

Sovereign Debt
(DEBT)

Figure 1 Conceptual Structure

3.4 Econometric Framework

The following equation has been evolved to estimate the determinants of NPLs:

NPLit= β0 + β1M2GDPit + β2EXCit + β3GDPit + β4FISCALit + β5INFLit + β6DEBTit + ɛit

This study further converted the equation into a logarithm model to make the dataset concise

since a large sample is unable to anticipate a relevant result. In special cases, log transformation

demonstrates results with significant accuracy. The small dataset displays a small variance and

hence it shows a coherent and accurate outcome. Therefore, the study has taken the natural log

of NPLs, M2GDP, EXC, and DEBT.

where,

18
 NPLs = Non-Performing Loans

M2GDP = Broad Money (% of GDP)

EXC = Exchange Rate

GDP = Gross Domestic Product

FISCAL = Government Budget Balance

INFL = Inflation Rate

DEBT = Sovereign Debt

β0 = The Intercept

β1 - β6 = Respective Coefficient Terms

i (Number of Countries) = 1, 2, … ,8

t (Time-Interval) = 1,2, … ,12

ɛit = Random Error

3.5 Analytical Techniques

This study conducted an empirical investigation and analysis by applying panel data regression

analysis. This study is by nature, exploratory, the dataset is strongly balanced and certain

similar entities in a specific timeframe will be observed. Hence, panel data regression analysis

is preferred. In the panel data regression, a combination of time series (year) and cross-sectional

data (individual countries) will ostensibly correspond to error/residuals (Zulfikar & STp, 2019).

Hence, a conventional regression model that is Pooled OLS (Ordinary Least Square) might

produce biased estimation.

OLS model was applied to check for multi-collinearity by conducting correlation matrix and

Variable Inflation Factor (VIF) test. The multi-collinearity problem occurs due to a significant

correlation between the independent variables. Hence, the increasing p-value decreases the t-

statistics that create a significant variable distinctly insignificant (F. Ahmad & Bashir, 2013).

Therefore, highly correlated variables were dropped to resolve this issue.

19
Furthermore, Fixed Effect (FE) & Random Effect (RE) Estimates were applied for checking

significance of explanatory variables. In general, Fixed Effect estimates are considered better

since it assumes that residuals (ɛ) are not correlated with regressors (βs). As a result, Hausman

Test was employed to selected the appropriate estimation between FE and RE estimation.

Based on the outcome, Fixed Effect estimation was carried out and Modified Wald Test was

conducted afterward, to check whether the residuals are heteroskedastic or homoscedastic. To

determine the first-order serial correlation/autocorrelation and cross-sectional independence in

FE Model, Wooldridge Test and Pesaran Test were employed respectively. Finally, the strongly

balanced dataset turned out as heteroscedastic, and consequently, Fixed Effect Robust Standard

Error was conducted to remove heteroscedasticity.

20
Chapter 04 Diagnostic Tests & Findings

21
4.1 Descriptive Statistics

The descriptive analysis of both dependent and explanatory determinants of SAARC countries

for the period of 12 years (2008-2019) with a total observation of 96 illustrate in Table 02. The

result of descriptive statistic in this study indicated that the mean value of non-performing loans

(NPLs) of SAARC countries recorded as 7.93%, while the standard deviation (SD) was 5.95%

and the range of significant disparity among the countries are between .5% (Afghanistan, in

2010) and 48.4% (Afghanistan, in 2011). The average broad money supply (M2GDP) was

58.51% with a minimum value of 29.512% (Afghanistan, in 2008) and a maximum value of

97.17% (Nepal, in 2019). Moreover, minimum value of Exchange Rate (EXC) was 12.8 (the

Maldives, in 2008) with a SD of 35.99, while the maximum was 178.745 (Sri Lanka, in 2019)

with a mean value of 73.82.

Table 2 Descriptive statistics

Variable Obs Mean Std. Dev. Min Max

NPLs 96 7.928 5.947 .5 48.4
M2GDP 96 58.506 16.347 29.512 97.17
EXC 96 73.816 35.988 12.8 178.745
GDP 96 5.57 3.291 -7.229 21.391
FISCAL 96 -4.368 3.769 -17.9 7.9
INFL 96 6.733 4.608 -6.811 26.419
DEBT 96 53.624 25.784 6.769 114.2
Source: Author’s Calculation

The average GDP growth rate (GDP) recorded as 5.57% with a lower standard deviation of

3.29% and demonstrated a substantial-high variability since the minimum and maximum

values were -7.23% (Maldives, in 2009) and 21.39% (Afghanistan, in 2009) respectively. The

minimum value of GDP is, however, negative and the mean value is very small which refers

that some of the SAARC countries experienced a negative growth during 2008-2019.

Concerning the public finance determinants, the mean value indicated a budget deficit of 4.37%

of GDP under the government budget balance (FISCAL) variable with a SD of 3.77%, while

individual country’s highest budget deficit was -17.9% (Maldives, in 2009) of GDP and a much

22
smaller budget surplus of 7.9% (Bhutan, in 2010) of GDP. The inflation rate (INFL) average

recorded as 6.73% and this ranges between a minimum value of -6.81% (Afghanistan, in 2009)

and the maximum value of 26.42% (Afghanistan, 2008) with a standard deviation of 4.61%.

Finally, the average sovereign debt (DEBT) recorded as 53.624% with 25.784% disparity,

while the minimum value was 6.769% (Afghanistan, in 2012) and the maximum value was

114.2% (Bhutan, in 2016) and this indicated that from 2008 to 2019, many SAARC countries

registered high debt with an unsustainable level.

AVERAGE NPL RATE OVER THE PERIOD

60
OF 2008-2019
40

20

0
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Afghanistan Bangladesh Bhutan India Maldives
Nepal Pakistan Sri Lanka Average

Source: Author’s Complication

Figure 2 Trend of average NPLs rate during the period of 2008-2019 in SAARC countries

Figure 2 illustrates the dynamics of the average NPLs ratio of SAARC countries from 2008 to

2019. During the great recession of 2008, individual country’s loan portfolio quality was

significantly deteriorated. When the global financial crisis was at its onset, the unemployment

rate surged, corporate and household incomes declined, some of the national currencies

depreciated, and asset values, especially, real estate value decreased. The post-crisis impact

exhibits that among the other SAARC countries, Maldives, Pakistan, Afghanistan, and

Bangladesh were recorded a dramatic increase of NPLs during the period of 2009-2013.

23
4.2 Multi-collinearity Test

In the equation of multiple regression, multi-collinearity exists when there is a significant

correlation between two or more explanatory variables. When a multi-collinearity problem

undermines the significance of the independent variable as a higher standard error leads to a

lesser significance in the regression coefficient (Allen, 1997). Additionally, significant

collinearity of independent variables is unacceptable since a dataset with a multi-collinearity

problem is unable to measure the role of explanatory variables accurately while explaining the

difference of the dependent variable’s value (Haitovsky, 1969). To test the multi-collinearity

among the determinants, the following tests were run after conducting the Pooled OLS

regression (Appendix A).

4.2.1 Correlation Matrix

Table 3 Matrix of correlations

Variables L_NPLs L_M2GDP L_EXC GDP FISCAL INFL L_DEBT

L_NPLs 1.000
L_M2GDP -0.259 1.000
L_EXC -0.275 0.242 1.000
GDP -0.355 -0.013 -0.112 1.000
FISCAL -0.187 0.051 0.217 0.204 1.000
INFL -0.092 0.043 0.088 -0.242 -0.061 1.000
L_DEBT 0.072 0.446 0.042 -0.034 -0.440 0.073 1.000
Note: Natural logarithm of Non-performing Loans (NPLs), Natural logarithm of Broad Money Supply as % of
GDP (M2GDP), Natural logarithm of Average Exchange Rate (EXC), GDP Growth Rate annual percentage
(GDP), Government Budget Surplus/Deficit (FISCAL), Inflation Rate (INFL), Natural logarithm of Sovereign
Debt (DEBT)
Source: Author’s Calculation

Table 03 explains correlation among the independent variables. The correlation matrix explains

whether the multi-collinearity problems are present in the dataset. If the value of the correlation

coefficient is greater than .80 or 80%, then there exists a multicollinearity problem between

two independent variables (Gujarati, 2003; Kennedy, 2008). The outcome of the dataset

illustrates that the independent variables are not significantly correlated as all of the correlation

coefficients are less than .50. Therefore, the multi-collinearity problem is not present in this

strongly balanced panel dataset.

24
4.2.2 Variance Inflation Factor (VIF) Test

Marquaridt (1970), Neter, Wasserman, and Kutner (1989), and Kennedy (2008) recommended

the maximum acceptable VIF value is 10 (VIF≤ 10), and hence, it correspondent to the

tolerance of .10, since 1/.10 =10. Hence, the rule of thumb is VIF≤10 is the highest critical

value to decide the existence of multi-collinearity. However, Richter et al. (2016) and Ahmed

et al. (2021) considered the benchmark value of VIF was ≤5 based on their research criteria.

The result of VIF and its tolerance level is illustrated in Table 04. The outcome demonstrates

the VIF values of all the independent variables are within the cut-off VIF as the values are

below 5 and within cut-off tolerance statistics as the tolerance values are greater than .05.

Hence, the dependent variables indicate that there is no existence of multi-collinearity as this

panel dataset demonstrates the VIF’s critical value of 0.05<VIF<5.

Table 4 VIF Test

VIF 1/VIF
L_M2GDP 1.428 .7
L_EXC 1.149 .87
GDP 1.141 .876
FISCAL 1.524 .656
INFL 1.071 .934
L_DEBT 1.728 .579
Mean VIF 1.34 .
Source: Author’s Calculation

4.3 Regression Analysis: Fixed and Random Effect Model

Although pooled OLS model was conducted earlier for testing the multi-collinearity problem,

one of the significant shortcomings of this model is it ignores the individuality and

heterogeneity in the data. Hence, fixed and random effect models were carried out (Appendix

B & Appendix C). In the fixed-effect model, EXC and FISCAL demonstrate a positive

relationship with the NPLs rate, whereas, M2GDP, GDP, INFL, along with DEBT indicate

negative relation with NPLs. The outcome suggests that M2GDP, GDP, INFL, and DEBT have

a significant relationship with the NPLs.

25
However, the random effect estimates illustrate that only FISCAL has a positive impact and

the rest of the variables – M2GDP, EXC, GDP, INFL, DEBT have a negative relationship with

NPLs. The result exhibits that M2GDP, GDP, and INFL have a significant relationship with

the NPLs which is slightly different from the outcome of fixed effect estimates. Moreover, both

of the models are statistically significant as the F statistic value indicates that both of the models

are below the .05 significance level.

4.4 Hausman Test

To identify the suitable test between FE and RE estimates, Hausman test was carried out.

Hausman test refers if the p-value is less than 5% then the null hypothesis (H0: Difference in

coefficients not systematic) is rejected.

Table 5 Hausman (1978) Specification Test

Coef.
Chi-square test value 18.694
P-value .005
Source: Author’s Calculation

Table 05 shows that the p-value is at the .05 (significance) level. Hence, H0 is rejected, and a

fixed (systematic) effect model is accepted.

4.5 Test of Heteroscedasticity

To check the heteroscedasticity in fixed effect estimates, the Modified Wald test was carried

out. If heteroscedasticity presents in the FE model, then the standard error for coefficients and

respective t-values most likely provide the wrong outcome. In that case, the outcome suggests

to reject null hypothesis since p-value <.05 and, hence, it can be determined that the residuals

are failed to comply with the assumption of homoscedasticity.

Modified Wald Test

chi2 (8) = 731.55
Prob>chi2 = 0.0000

26
The result demonstrates that the value of chi2 is less than at the .05 (significance) level and

hence, the study rejects the null hypothesis. Therefore, there is a presence of heteroscedasticity

in the fixed effect regression estimates.

4.6 Test of Autocorrelation

Autocorrelation or serial correlation leads to a smaller standard error of coefficient than its

actual values and therefore, the R2 gets Inflated (Mehmood et al., 2013). This report deals with

time elements where t=12>20 and, hence, it indicates a micro panel dataset. In general, the

autocorrelation does not exist in the dataset that is dealing with less than 20 years of the time

variable. However, the Wooldridge test for autocorrelation was carried out to test the exactness.

Wooldridge Test
F (1, 7) = 3.538
Prob > F = 0.1020

The outcome it indicates that the p-value is 10.20% which is more than .05 and, hence, the

study fails to reject null hypothesis. Therefore, it is above the significance level and there is

no presence of autocorrelation in this strongly balanced dataset.

4.7 Test of Cross-Sectional Interdependence

To check whether the residuals have contemporaneous correlation/ cross-sectional dependence,

the Pasaran CD test was carried out. If the residuals are correspondent to across the variables,

then the outcome will be biased.

Pesaran's Test

Cross-sectional interdependence = -1.779,

Pr = 0.0752

The null hypothesis is there is no cross-sectional interdependence across the entities. The

outcome indicates that the p-value is 7.52% which is above .05. Hence, the study fails to reject

the null hypothesis.

27
4.8 Model Specification Result

The above-mentioned diagnostic tests illustrate that there is no existence of multi-collinearity,

autocorrelation, and cross-sectional dependence within FE estimates. However, there is a

presence of heteroscedasticity in the fixed effect regression estimates. To deal with the

heteroscedasticity problem, the robust standard error for the FE model was carried out since

robust standard error estimates automatically corrects heteroscedasticity.

Table 6 Robust Standard Error Fixed Effect Regression Results

L_NPLs Coef. St.Err. t-value p-value [95% Conf Interval] Sig

L_M2GDP -1.868 .679 -2.75 .029 -3.474 -.262 **
L_EXC .655 .761 0.86 .418 -1.146 2.455
GDP -.094 .025 -3.82 .007 -.152 -.036 ***
FISCAL .033 .011 3.00 .02 .007 .059 **
INFL -.027 .013 -2.12 .072 -.057 .003 *
L_DEBT -1.048 .533 -1.97 .09 -2.308 .212 *
Constant 11.464 2.557 4.48 .003 5.417 17.512 ***

Mean dependent var 1.825 SD dependent var 0.757

R-squared 0.418 Number of obs 96
F-test 14.406 Prob > F 0.001
Akaike crit. (AIC) 123.776 Bayesian crit. (BIC) 139.162
*** refers p<.01, ** refers p<.05, and * refers p<.1
Source: Author’s Calculation
According to the findings, the robust standard error in the FE model (Table 06) indicates,

NPLs = 11.464 - 1.868M2GDP + .655EXC - .094GDP + .033FISCAL - .027INFL -

1.048DEBT

The findings suggest that the entire robust standard error of the Fixed Effect Model’s F-statistic

value of 14.406 renders .1% which is at the .05 (significance) level. As these values indicate

the fitness of the model, hence the model is fit for analysis. Additionally, the R-square is 41.8%

meaning that the macroeconomic determinants explain nearly 42% variance in the dependent

variable (NPLs). The empirical result also indicates that at the .05 (significance) level, M2GDP

and GDP have negative and FISCAL is positively associated with NPLs. While at .10 level,

inflation and sovereign debt have significant negative association with NPLs. However, the

average exchange rate illustrates an insignificant positive relationship with NPLs.

28
4.9 Discussion

The broad money as a percentage of GDP (M2GDP) exhibits p-value is statistically significant

(.029) at 95% level and reveals an inverse relationship with non-performing loans. The

outcome implies that the interest rate reduces significantly if the growth of money supply

increases; as the borrowers tend to receive funds with a comparatively cheaper rate and hence,

this increases their ability to pay back their financial obligations. The findings are consistent

with Kalirai and Scheicher (2002), Vogiazas and Nikolaidou (2011), and Poudel (2013).

Ahmad (2003) also found a significant negative relationship between these factors while

assuming M3 as a proxy of the money supply. Additionally, Adusei (2018) considered M1,

M2, and M2+, while Badar et al. (2013) took M2 as a proxy of money supply and found similar

findings. However, at the .05 level, Akinlo and Emmanuel (2014), Leka, Bajrami, and Duci

(2019) found an insignificant but positive relationship with NPLs and argued that increased

money supply deteriorates the bank’s portfolio due to inaccurate credit analysis.

Although the exchange rate (EXC) reveals an anticipated relationship with NPLs, the value is

not statistically significant. The direct relation of EXC to NPLs signifies that rising exchange

rates undermine the performance of import-oriented sectors due to trade deterioration in the

entire economy and therefore, the crisis in the banking system exacerbates. The findings are

similar to the results of Fofack (2005), Khemraj and Pasha (2009), and Akinlo and Emmanuel

(2014), but in contrast with the result of Ahmad and Bashir (2013), since they stated that the

international competitiveness inversely impacts the NPLs.

The outcome exhibits that the beta coefficient of annual GDP growth rate (GDP) have an

inverse relationship at the .01 (significance) level which implies a noteworthy impact on NPL.

The result indicates that the economic growth improves the business performance and increase

their payment capacity and that leads to a decrease in NPL. Additionally, during the economic

downturn, the borrower’s income and collaterals’ value goes down, and hence, the borrower’s
29
ability to pay decreases. The findings are clearly in line with the predetermined hypothesis and

in line with the findings of Salas and Saurina (2002), Rajan and Dhal (2003) Fofack (2005),

Jesus and Gabriel (2006), Espinoza and Prasad (2010), Dash and Kabra (2010), Louzis et al.

(2012), Castro (2013), Messai and Jouini (2013), Jakubík and Reininger (2013), Klein (2013),

Makri et al. (2014), Škarica (2014), Erdinc and Abazi (2014), Kasselaki and Tagkalakis (2014),

Chaibi and Ftiti (2015), Dimitrios et al. (2016).

The result of the government budget surplus/deficit (FISCAL) positive and statistically

significant (.02) at 95% level. This outcome is in line with the study’s hypothesis. The

government budget balance has a positive relationship with NPLs due to the measures of

budgetary consolidation. The budgetary consolidation occurs either due to insignificant

budgetary expenditure (e.g. defense services, grants to state, pensions, public employee’s

remunerations, health services, etc.) and/or significant budgetary revenues (through raising

prevailing taxes/ adding new taxes, excise customs, and other duties, etc.). These measures

could yield a lower income that reduces the borrower’s ability to repay the outstanding debt

and hence, the bank’s NPLs increase. The findings are similar to that of Roman and Bilan

(2015) and Dimitrios et al. (2016) where they argued that budgetary consolidation results in a

low budget deficit of high budget surplus and, thereby, deteriorates the bank portfolios. Our

study is, however, opposite to the study of Makri, Tsagkanos, and Bellas (2014), since they

argued that by nature, the government budget surplus/ deficit has a negative influence on

sovereign debt, and hence, government budget balance is adversely correlated with NPLs.

The inflation rate (INFL) depicts an adverse relationship with loan portfolio quality and the p-

value <.10 which is at the .10 (significance) level. This refers that if inflation rate upsurges by

1%, then NPLs will be decreased by .027%. The finding is confirmed by the study of Shu

(2002), Zribi and Boujelbegrave (2011), Vogiazas and Nikolaidou (2011), Zribi and

Boujelbegrave (2011), Klein (2013), Erdinc and Abazi (2014), Škarica (2014), Chaibi and Ftiti

30
(2015), Ekanayake and Azeez (2015), Anjom and Karim (2016), Koju et al. (2018).

Nonetheless, Makri et al. (2014) and Kasselaki and Tagkalakis (2014) found an insignificant

association between inflation and credit risk. In contrast with these findings, Rinaldi and

Sanchis-Arellano (2006) and Nkusu (2011) found significant and direct association between

inflation and credit risk, while an insignificant positive relationship was found by Castro

(2013). He argued that high inflationary pressure decreases the real income of borrowers and

the real value of outstanding debt. Hence, the final result of inflation can be neutralized as one

effect is compensated through another impact.

The expected hypothesis regarding sovereign debt (DEBT) is completely dismissed. The

findings suggest that the beta coefficient of sovereign debt of SAARC countries has a strong

but negative association with NPLs at the .10 level. The rationale behind the outcome is that

the remodeling, innovation, and development in the financial sector may reduce credit risk.

Since innovation and development require massive government borrowing, it ultimately

reduces nonperforming loans. This result is consistent with Garr (2013), Anjom and Karim

(2016) and Dimitrios et al. (2016), and contrast to that of Louzis et al. (2012), Makri et al.

(2014), Roman and Bilan (2015), and Koju et al. (2018). They asserted that higher amount of

sovereign debt decreases the loans in the financial market and then interest rates on loans

increase. Since increasing interest rates increase the cost of loans, hence NPL increases due to

poor loan repayment capability of borrowers in a timely manner.

31
Chapter 05 Concluding Remarks

32
5.1 Conclusion and Policy Implications

Non-performing loans are one of the significant proxies of the financial and economic stability

of any country. The paper empirically evaluates the relationship between the macroeconomic

performance and banking system across SAARC countries for the period of 2008-2019 and

analyzes the shock of financial crisis 2007-2008. The study findings indicate slower economic

growth, low level of inflation, sluggish money supply growth, currency depreciation are driving

macroeconomic factors that are responsible for high NPLs in the SAARC economy. Moreover,

this paper uses two public finance variables and the findings suggest that high government

budget balance and lower level of sovereign debt are primarily responsible for a high level of

NPLs. These findings are useful for formulating macro-prudential along with fiscal policies to

avoid the subsequent NPLs shock in SAARC economics.

This is the first study that empirically examined the impact of selected macroeconomic

variables of NPLs in SAARC countries. The study pinpoints that the bank’s quality of loan

portfolio is adversely affected by a wide range of macroeconomic determinants. To reduce the

aggregate NPLs in the SAARC economy, the respective country’s government should identify

the financial sector’s vulnerabilities and, thereby, emphasize on boosting the economic growth,

appreciating local currency, ensuring a moderate level of money supply along with inflation

rate. This empirical analysis enables the respective country government and regulatory

authorities of banks to forecast the NPLs trends in the upcoming years for both the macro-level

and individual commercial banks in SAARC countries. Hence, the banks will stay alerted

during the adverse economic situation and be able to avoid credit risk.

33
5.2 Future Research Avenues

This paper solely focused on external (macroeconomic) factors for examining the NPLs of the

SAARC economy. Hence, future research avenues should be directed towards comparison

among the SAARC countries, aiming at examining both macroeconomic and bank-specific

variables of NPLs in SAARC countries.

34
 References

Adusei, C. (2018). Determinants of non-performing loans in the banking sector of Ghana

between 1998 and 2013. Asian Development Policy Review, 6(3), 142–154.

Ahmad, F., & Bashir, T. (2013). Explanatory power of macroeconomic variables as

determinants of non-performing loans: Evidence form Pakistan. World Applied Sciences

Journal, 22(2), 243–255.

Ahmad, N. H. (2003). Formation of credit risk, price Effect of Regulatory changes and the path

Linking credit Risk and total Risk. Unpublished Ph. D. Thesis, Universiti Utara Malaysia.

Ahmad, N. H., & Ariff, M. (2007). Multi-country study of bank credit risk determinants.

International Journal of Banking and Finance (IJBF), 5(1), 35–52.

Ahmed, S., Majeed, M. E., Thalassinos, E., & Thalassinos, Y. (2021). The Impact of Bank

Specific and Macro-Economic Factors on Non-Performing Loans in the Banking Sector:

Evidence from an Emerging Economy. Journal of Risk and Financial Management, 14(5),

217.

Akinlo, O., & Emmanuel, M. (2014). Determinants of non-performing loans in Nigeria.

Accounting & Taxation, 6(2), 21–28.

Allen, M. P. (1997). The problem of multicollinearity. Understanding Regression Analysis,

176–180.

Alton, R. G., & Hazen, J. H. (2001). As economy flounders, do we see a rise in problem loans.

Federal Reserve Bank of St. Louis, 11(4), 45–65.

Anjom, W., & Karim, A. M. (2016). Relationship between non-performing loans and

macroeconomic factors with bank specific factors: a case study on loan portfolios–

SAARC countries perspective. ELK Asia Pacific Journal of Finance and Risk

35
 Management, 7(2), 1–29.

Babouček, I., & Jančar, M. (2005). Effects of macroeconomic shocks to the quality of the

aggregate loan portfolio (Vol. 22). Czech National Bank.

Badar, M., Javid, A. Y., & Zulfiquar, S. (2013). Impact of macroeconomic forces on

nonperforming loans: An empirical study of commercial banks in Pakistan. Wseas

Transactions on Business and Economics, 10(1), 40–48.

Barseghyan, L. (2010). Non-performing loans, prospective bailouts, and Japan’s slowdown.

Journal of Monetary Economics, 57(7), 873–890.

Beck, R., Jakubik, P., & Piloiu, A. (2013). Non-performing loans: What matters in addition to

the economic cycle?

Bercoff, J., Giovanni, J., & Grimard, F. (2002). Argentinean banks, credit growth and the

tequila crisis: A duration analysis. Unpublished Paper.

Bofondi, M., & Ropele, T. (2011). Macroeconomic determinants of bad loans: evidence from

Italian banks. Bank of Italy Occasional Paper, 89.

Borio, C., Drehmann, M., & Tsatsaronis, K. (2014). Stress-testing macro stress testing: does it

live up to expectations? Journal of Financial Stability, 12, 3–15.

Boudriga, A., Taktak, N. B., & Jellouli, S. (2010). Bank specific, business and institutional

environment determinants of banks nonperforming loans: evidence from mena countries.

Economic Research Forum, Working Paper, 547, 1–28.

Caporale, G. M., Di Colli, S., & Lopez, J. S. (2014). Bank lending procyclicality and credit

quality during financial crises. Economic Modelling, 43, 142–157.

Castro, V. (2013). Macroeconomic determinants of the credit risk in the banking system: The

36
 case of the GIPSI. Economic Modelling, 31, 672–683.

Chaibi, H., & Ftiti, Z. (2015). Credit risk determinants: Evidence from a cross-country study.

Research in International Business and Finance, 33, 1–16.

Dash, M. K., & Kabra, G. (2010). The determinants of non-performing assets in Indian

commercial bank: An econometric study. Middle Eastern Finance and Economics, 7(2),

94–106.

De Bock, M. R., & Demyanets, M. A. (2012). Bank asset quality in emerging markets:

Determinants and spillovers. International Monetary Fund.

Dimitrios, A., Helen, L., & Mike, T. (2016). Determinants of non-performing loans: Evidence

from Euro-area countries. Finance Research Letters, 18, 116–119.

Ekanayake, E., & Azeez, A. A. (2015). Determinants of non-performing loans in licensed

commercial banks: Evidence from Sri Lanka. Asian Economic and Financial Review,

5(6), 868.

Erdinc, D., & Abazi, E. (2014). The determinants of NPLs in emerging Europe, 2000-2011.

Journal of Economics and Political Economy, 1(2), 112–125.

Espinoza, M. R. A., & Prasad, A. (2010). Nonperforming loans in the GCC banking system

and their macroeconomic effects. International Monetary Fund.

Feijó, C. A. (2011). Credit risk and macroeconomic interactions: Empirical evidence from the

Brazilian banking system. Modern Economy, 2(05), 910.

Fofack, H. (2005). Nonperforming loans in Sub-Saharan Africa: causal analysis and

macroeconomic implications (Vol. 3769). World Bank Publications.

Fonseka, N. (2009). Bouncing back–the opportunities and constraints. 20th Anniversary

37
 Convention of Association of Professional Bankers in Sri Lanka.

Garr, D. K. (2013). Determinants of credit risk in the banking industry of Ghana. Developing

Country Studies, 3(11), 64–77.

Ghosh, A. (2015). Banking-industry specific and regional economic determinants of non-

performing loans: Evidence from US states. Journal of Financial Stability, 20, 93–104.

González-Hermosillo, B. (1999). Developing indicators to provide early warnings of banking

crises. Finance & Development, 36(002).

Gujarati, D. N. (2003). Basic Econometrics. Forth Edition. Singapura: McGraw-Hill.

Haitovsky, Y. (1969). Multicollinearity in regression analysis: Comment. The Review of

Economics and Statistics, 486–489.

IMF. (2005). The Treatment of Nonperforming Loans BOPCOM-05/29.

IMF. (2021). Financial Soundness Indicators (FSIs). https://data.imf.org/?sk=51B096FA-

2CD2-40C2-8D09-0699CC1764DA&sId=1390030341854

Investopedia. (2020). South Asian Countries: The New Face of Emerging Economies.

https://www.investopedia.com/articles/investing/022316/south-asia-new-face-emerging-

economies.asp

Investopedia. (2021). INVESTOPEDIA. https://www.investopedia.com/terms/s/sovereign-

default.asp

Ivanović, M. (2016). Determinants of credit growth: The case of Montenegro. Journal of

Central Banking Theory and Practice, 5(2), 101–118.

Jakubík, P., & Reininger, T. (2013). Determinants of nonperforming loans in Central, Eastern

and Southeastern Europe. Focus on European Economic Integration, 3, 48–66.

38
Jesus, S., & Gabriel, J. (2006). Credit cycles, credit risk, and prudential regulation.

Joseph, M. T., Edson, G., Manuere, F., Clifford, M., & Michael, K. (2012). Non performing

loans in commercial banks: a case of CBZ Bank Limited in Zimbabwe. Interdisciplinary

Journal of Contemporary Research in Business, 4(7), 467–488.

Kalirai, H., & Scheicher, M. (2002). Macroeconomic stress testing: preliminary evidence for

Austria. Financial Stability Report, 3, 58–74.

Kasselaki, M. T., & Tagkalakis, A. O. (2014). Financial soundness indicators and financial

crisis episodes. Annals of Finance, 10(4), 623–669.

Kastrati, A. (2011). The Determinants of Non-Performing Loans in Transition Countries.

Financial Stability Report, Central Bank of the Republic of Kosovo (June), 97–102.

Keeton, W. R. (1999). Does faster loan growth lead to higher loan losses? Economic Review-

Federal Reserve Bank of Kansas City, 84(2), 57.

Kennedy, P. (2008). A guide to econometrics. John Wiley & Sons.

Khemraj, T., & Pasha, S. (2009). The determinants of non-performing loans: an econometric

case study of Guyana.

Klein, N. (2013). Non-performing loans in CESEE: Determinants and impact on

macroeconomic performance. International Monetary Fund.

Koju, L., Koju, R., & Wang, S. (2018). Macroeconomic and bank-specific determinants of non-

performing loans: evidence from Nepalese banking system. Journal of Central Banking

Theory and Practice, 7(3), 111–138.

Konstantakis, K. N., Michaelides, P. G., & Vouldis, A. T. (2016). Non performing loans

(NPLs) in a crisis economy: Long-run equilibrium analysis with a real time VEC model

39
 for Greece (2001–2015). Physica A: Statistical Mechanics and Its Applications, 451, 149–

161.

Kumar, D., Hossain, M. Z., & Islam, M. S. (2020). Non-performing loans in banking sector of

Bangladesh: An evaluation. International Journal of Applied Economics, Finance and

Accounting, 6(1), 22–29.

Leka, B., Bajrami, E., & Duci, E. (2019). Key Macroeconomic Drivers on Reducing Non

Performing Loans in Albania. Academic Journal of Interdisciplinary Studies, 8(2), 88.

Louzis, D. P., Vouldis, A. T., & Metaxas, V. L. (2012). Macroeconomic and bank-specific

determinants of non-performing loans in Greece: A comparative study of mortgage,

business and consumer loan portfolios. Journal of Banking & Finance, 36(4), 1012–1027.

Makri, V., Tsagkanos, A., & Bellas, A. (2014). Determinants of non-performing loans: The

case of Eurozone. Panoeconomicus, 61(2), 193–206.

Marquaridt, D. W. (1970). Generalized inverses, ridge regression, biased linear estimation, and

nonlinear estimation. Technometrics, 12(3), 591–612.

McGoven, J. (1998). Why bad loans happen to good banks. The Journal of Commercial

Lending, 75(6), 44–51.

Mehmood, B., Younas, Z. I., & Ahmed, N. (2013). Macroeconomic and bank specific

Covariates of non-performing loans (NPLs) in Pakistani commercial banks: Panel data

evidence. Journal of Emerging Economies and Islamic Research, 1(3), 34–48.

Messai, A. S., & Jouini, F. (2013). Micro and macro determinants of non-performing loans.

International Journal of Economics and Financial Issues, 3(4), 852.

Neter, J., Wasserman, W., & Kutner, M. H. (1989). Applied linear regression models.

40
Nkusu, M. M. (2011). Nonperforming loans and macrofinancial vulnerabilities in advanced

economies. International Monetary Fund.

OECD. (2011). Development: Aid to Developing Countries Falls Because of Global Recession.

https://www.oecd.org/newsroom/developmentaidtodevelopingcountriesfallsbecau-seof-

globalrecession.htm

Poudel, R. P. S. (2013). Macroeconomic Determinants of credit risk in nepalese banking

industry. Proceedings of 21st International Business Research Conference, 10–11.

Rajan, R., & Dhal, S. C. (2003). Non-performing loans and terms of credit of public sector

banks in India: An empirical assessment. Reserve Bank of India Occasional Papers, 24(3),

81–121.

Reinhart, C. M., & Rogoff, K. S. (2010). Growth in a Time of Debt. American Economic

Review, 100(2), 573–578.

Reinhart, C. M., & Rogoff, K. S. (2011). From financial crash to debt crisis. American

Economic Review, 101(5), 1676–1706.

Reserve Bank of India. (2017). Financial Stability Report. Paper Knowledge . Toward a Media

History of Documents, 15.

Richter, N. F., Cepeda-Carrion, G., Roldán Salgueiro, J. L., & Ringle, C. M. (2016). European

management research using partial least squares structural equation modeling (PLS-

SEM). European Management Journal, 34 (6), 589-597.

Rinaldi, L., & Sanchis-Arellano, A. (2006). Household debt sustainability: What explains

household non-performing loans? An empirical analysis.

Roman, A., & Bilan, I. (2015). An empirical analysis of the macroeconomic determinants of

non-performing loans in EU28 banking sector. Revista Economică, 67(2), 108–127.

41
Salas, V., & Saurina, J. (2002). Credit risk in two institutional regimes: Spanish commercial

and savings banks. Journal of Financial Services Research, 22(3), 203–224.

SHIMELSE ALI, URI DADUSH, L. F. (2009). Financial transmission of the crisis: What’s

the lesson. International Economic Bulletin.

https://carnegieendowment.org/2009/06/17/financial-transmission-of-crisis-what-s-

lesson-pub-23284

Shu, C. (2002). The impact of macroeconomic environment on the asset quality of Hong

Kong’s banking sector. Hong Kong Monetary Authority Research Memorandums, 2002,

1–26.

Škarica, B. (2014). Determinants of non-performing loans in Central and Eastern European

countries. Financial Theory and Practice, 38(1), 37–59.

Vogiazas, S. D., & Nikolaidou, E. (2011). Investigating the determinants of nonperforming

loans in the Romanian banking system: An empirical study with reference to the Greek

crisis. Economics Research International, 2011.

World Bank. (2012). Migration and Development Brief 19. http://sitere-

sources.worldbank.org/INTPROSPECTS/Resources/334934-1288990760745/Mi-

grationDevelopmentBrief19.pdf

Zameer, S., & Siddiqi, M. W. (2010). The impact of exports, FDI and external debt on exchange

rate volatility in Pakistan. International Journal of Contemporary Research in Business,

2(7), 337–354.

Zeng, S. (2012). Bank non-performing loans (NPLS): A dynamic model and analysis in China.

Zribi, N., & Boujelbegrave, Y. (2011). The factors influencing bank credit risk: The case of

Tunisia. Journal of Accounting and Taxation, 3(4), 70–78.

42
Zulfikar, R., & STp, M. M. (2019). Estimation model and selection method of panel data

regression: an overview of common effect, fixed effect, and random effect model.

43
 Appendix A: Pooled OLS Estimation
Table 7 Appendix A: Pooled OLS Estimation

L_NPLs Coef. St.Err. t-value p-value [95% Conf Interval] Sig

L_M3 -.804 .269 -2.99 .004 -1.337 -.27 ***
L_EXC -.308 .11 -2.80 .006 -.526 -.089 ***
GDP -.101 .021 -4.71 0 -.143 -.058 ***
BUG .015 .022 0.70 .485 -.028 .058
INF -.029 .015 -1.96 .054 -.058 0 *
L_DEBT .265 .12 2.21 .03 .027 .502 **
Constant 6.153 .987 6.23 0 4.191 8.115 ***

Mean dependent var 1.825 SD dependent var 0.757

R-squared 0.328 Number of obs 96
F-test 7.240 Prob > F 0.000
Akaike crit. (AIC) 193.775 Bayesian crit. (BIC) 211.726
*** p<.01, ** p<.05, * p<.1
Source: Author’s Calculation

44
 Appendix B: Fixed Effect Estimation
Table 8 Appendix B: Fixed Effect Estimation

L_NPLs Coef. St.Err. t-value p-value [95% Conf Interval] Sig

L_M3 -1.868 .543 -3.44 .001 -2.949 -.787 ***
L_EXC .655 .412 1.59 .116 -.166 1.475
GDP -.094 .017 -5.44 0 -.128 -.059 ***
BUG .033 .023 1.45 .152 -.012 .078
INF -.027 .013 -2.03 .046 -.054 -.001 **
L_DEBT -1.048 .314 -3.33 .001 -1.674 -.423 ***
Constant 11.464 2.539 4.52 0 6.413 16.515 ***

Mean dependent var 1.825 SD dependent var 0.757

R-squared 0.418 Number of obs 96
F-test 9.831 Prob > F 0.000
Akaike crit. (AIC) 125.776 Bayesian crit. (BIC) 143.726
*** p<.01, ** p<.05, * p<.1
Source: Author’s Calculation

45
 Appendix C: Random Effect Estimation
Table 9 Appendix C: Random Effect Estimation

L_NPLs Coef. St.Err. t-value p-value [95% Conf Interval] Sig

L_M3 -.907 .43 -2.11 .035 -1.751 -.064 **
L_EXC -.169 .232 -0.73 .465 -.623 .285
GDP -.104 .018 -5.72 0 -.139 -.068 ***
BUG .035 .024 1.49 .136 -.011 .081
INF -.033 .013 -2.60 .009 -.059 -.008 ***
L_DEBT -.054 .207 -0.26 .793 -.46 .351
Constant 7.343 1.773 4.14 0 3.867 10.819 ***

Mean dependent var 1.825 SD dependent var 0.757

Overall r-squared 0.199 Number of obs 96
Chi-square 38.261 Prob > chi2 0.000
R-squared within 0.327 R-squared between 0.107
*** p<.01, ** p<.05, * p<.1
Source: Author’s Calculation

46
 Appendix D: Individual Data Source
Table 10 Appendix D: Individual Data Source

Variables Sources
The Ratio of Non-performing Loans to Total World Bank. World Development Indicators, IMF.
Bank Loans Financial Soundness Indicators Annual Report/Bulletin of
Da Afghanistan Bank, Bangladesh Bank, Nepal Rastra
Bank, National Bank of Moldova, Royal Monetary
Authority of Bhutan, and Central Bank of Sri Lanka
Broad Money Supply (% of GDP) World Bank. World Development Indicators
Exchange Rate (average) IMF. International Financial Statistics
GDP Growth Rate (annual %) World Bank. World Development Indicators
General Government Net Lending/Borrowing World Bank. World Development Indicators
(% of GDP)
Inflation, Consumer Prices (annual %) World Bank. World Development Indicators
General Government Gross Debt, Total (% of IMF. World Economic Outlook
GDP)

47