Bank lending during the COVID-19 pandemic: A comparison of Islamic and conventional banks (2024)

Journal List
Elsevier - PMC COVID-19 Collection
PMC9916183

As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsem*nt of, or agreement with, the contents by NLM or the National Institutes of Health.
Learn more: PMC Disclaimer | PMC Copyright Notice

Journal of International Financial Markets, Institutions and Money. 2023 Apr; 84: 101743.

Published online 2023 Feb 10. doi:10.1016/j.intfin.2023.101743

PMCID: PMC9916183

Narjess Boubakri,^a Ali Mirzaei,^a and Mohsen Saad^a,^⁎

Author information Article notes Copyright and License information PMC Disclaimer

Associated Data

Data Availability Statement

Abstract

Using a sample of 421 banks from 17 countries, we find that the lending growth of Islamic and conventional banks decreased during the initial phase of the COVID-19 crisis. However, the decrease is significant for conventional banks only. Credit growth for Islamic banks grew around 2.5% faster than that for conventional banks, especially in countries with a macroprudential framework in place in the year leading up to the crisis. Our evidence remains unchanged with alternative empirical methodologies, definitions of bank lending, variations in the pre-crisis period, and proxies for the severity of COVID-19 in different countries.

Keywords: COVID-19, Bank performance, Islamic banks, Macroprudential policies, Bank lending, Credit growth

1. Introduction

To contain the initial spread of the novel coronavirus, COVID-19, governments worldwide resorted to strategies ranging from social distancing to complete lockdowns that put entire countries under national quarantines.¹These measures adversely impacted the economy, including the banking sector (e.g., Altavilla et al., 2020), reviving fears among some about banks' ability to sustain lending activities during challenging times or even of a potential credit crunch such as that witnessed in the 2008 Global Financial Crisis (GCF). Needless to be reminded that the banking panic during the GFC that severely restricted bank credit supply to the corporate sector had a devastating impact on global growth and prosperity (Ivashina and Scharfstein, 2010, Acharya and Naqvi, 2012). To alleviate the impact of the crisis, regulators intervened by imposing new macroprudential requirements to improve bank capital adequacy and the quality of bank assets (Repullo and Saurina, 2011, Stolz and Wedow, 2011, García-Suaza et al., 2012). These protective prudential policy measures led some today to believe that we should observe higher bank resilience during the Covid-19 crisis compared to 2008.

Drawing on this debate, we focus in this paper on bank lending behavior during the first three quarters of the year 2020. One strand of the literature shows that financial distress becomes inevitable in economic downturns unless banks play the critical role of liquidity suppliers by securing funds’ access to borrowers. In particular, prior empirical findings on bank lending during economic downturns highlighted the importance of bank capital (Kosak et al., 2015, Altunbas et al., 2016) and bank ownership (Brei and Schclarek, 2013, Cull and Martinez-Peria, 2013, Chen et al., 2016) in maintaining credit growth.During the COVID-19 crisis, firms' economic activities and cash-flow stream slowed, leading to an increased demand for loans. Hence, banks were expected to pursue counter-cyclical strategies and continue the flow of credit to support the demand for loans (Acharya and Steffen, 2020). As banks’ resilience became more challenging with the persistent COVID-19 conditions, monetary and prudential authorities enacted policy interventions that have been shown in the past to successfully sustain bank lending (Benetton and Fantino, 2018, Rostagno et al., 2019, Boeckx et al., 2020).Under these circ*mstances, we expect banks that operate in countries with macro-prudential policies in place before the crisis to be more resilient.

Another strand of the banking literature focuses on possible performance differences depending on banks’ business orientation, namely Islamic vs conventional. Some researchers highlight the potential superior performance of Islamic banks (IBs) during economic downturns. The main argument is that IB asset-based and risk-sharing intermediation practices dictated by Islamic laws (Shariah) protect IBs from crises' adverse impacts. Indeed, IBs performed quite well in the GFC period based on profitability, credit supply, deposit growth and withdrawals, and returns in the stock market. This contributed to their reputation as being more stable banks than their conventional counterparts and increased their popularity.²Early empirical evidence shows that during the GFC period, credit growth was higher for IBs than conventional banks (CBs) (Hasan and Dridi, 2011, Beck et al., 2013, Ibrahim, 2016, Ibrahim and Rizvi, 2018). Although the existing literature would support higher intermediation capacity for IBs compared to CBs, there is no evidence of whether IBs maintained higher credit growth than their counterparts during the COVID-19 episode.

Building on these strands of the literature on bank lending behavior during crises, and the importance of macro-prudential policies to bank resilience, we zoom in on potential differences between IBs and CBs, and seek to address two essential economic questions in the current study: (i) How did IBs perform in terms of lending activities compared to their peers (CBs) during the outbreak of the COVID-19 crisis? (ii) Do country characteristics, and particularly the utilization of macroprudential measures in the pre-crisis period, affect bank lending behavior during the crisis? These questions are critical given the importance of bank resilience to financial stability and economic growth.

To answer these questions, we use a sample of 421 banks domiciled in 17 countries with dual banking systems. The period of the study extends six months balanced around March 2020 when the World Health Organization declared the coronavirus disease as a pandemic. Around 30 percent of the sample (i.e., 117 banks) is denoted Islamic, while the rest (i.e., 344 banks) is denoted conventional. Inspection of the data shows no difference between the credit growth of both types of banks in the period preceding the crisis. The COVID-19 situation led to lower bank credit growth overall compared to pre-crisis levels. However, the negative impact is only significant for CBs. Growth in loans of IBs is statistically more significant than that of CBs during the crisis.

We estimate regression models that control for a comprehensive set of bank variables to carefully isolate the impact of COVID-19 on credit growth as it varies by bank business orientation. The findings confirm a higher resiliency for IBs compared to their conventional counterparts during the early stage of the COVID-19 crisis. Specifically, our evidence shows that differential lending growth between IBs and CBs during the crisis period was around 2.5% higher for the former. This central finding remains robust against a series of robustness checks, including the estimation of the regression models with alternative empirical approaches, a different definition of loan growth, the extension of the pre-crisis period, the use of lags for control variables to mitigate the potential concern about reverse causality, and additional control variables. Notably, rather than relying on a dummy variable that switches to one for all countries simultaneously, our regression analysis differentiates among the country variations in COVID-19 severity. We employ the John Hopkins University number of COVID-19 cases per 100,000 people in each country, the (Hale et al., 2020) COVID-19 stringency index, and the risk of openness index. The reported results support our main conclusion that IBs sustained their lending activities during the initial phase of the COVID-19 crisis while CBs did not.

We further extend the empirical analysis by investigating the effect of the pre-crisis usage of macroprudential policies on the divergence of loan growth between IBs and CBs during the global health crisis. Implementing macroprudential policies aims to strengthen financial stability and shield economic growth against systematic risks and financial imbalances. Prior research found macroprudential policies to be effective in dampening procyclical bank lending and risk-taking following periods of economic expansions (Claessens et al., 2013, Cerutti et al., 2017, Gómez et al., 2020). Applied to our context, we expect the higher ability of IBs to maintain credit growth during the crisis to be stronger in countries that implemented macroprudential policies in the pre-crisis period. The reported evidence supports our conjecture.

Our analysis advances the existing literature along two fronts: first, we contribute to the literature on Islamic banking by showing that IBs showed higher resilience than CBs during COVID-19. This may have positively impacted economic growth during that period. Second, by showing that this ability to sustain credit growth depends on macroprudential policies, we contribute to the literature on the importance of such policies to financial stability in general.

The paper proceeds as follows. Section 2 reviews the literature on the relative performance of IBs, and further develops our hypotheses. Section 3 describes our data and methodology. Section 4 discusses our empirical results, and Section 5 concludes.

2. Islamic versus conventional banks

2.1. Background and relevant literature review: IBs' credit during crisis times

Like CBs, IBs perform the essential intermediation task by lowering the adverse consequences of information and transaction costs. IBs and CBs engage in activities that reduce the cost of searching for profitable investment opportunities, exercise governance and corporate control, and ultimately allocate resources. However, unlike CBs, IBs adhere to principles that determine Islamically acceptable forms of business transactions. These governing laws, commonly known as Shariah, lead intermediation in IBs to be asset-based and built on a risk-sharing structure. Financial transactions must have 'material finality' by involving the exchange of tangible assets in the real economy, not financial assets. By linking financial transactions to the real economy, lending in IBs must be asset-based and not debt-based.³Besides, Shariah laws establish a risk-sharing framework whereby IBs are prohibited from producing ‘risk-free’ profits such as those made on collateralized loans. Alternatively, profits must be justly earned by taking an equity position in the transactions with a proportional share of risk between the providers of funds and users of funds. Islam's position towards the prohibition of interest and the implication of risk-sharing was concisely put by El Gamal (2000, p. 33), “In Islam, one does not lend to make money, and one does not borrow to finance business.”.

Although comparative studies of IBs and CBs abound, only a few papers have focused on the lending behavior of IBs during crisis times. By comparing the performance of IBs and CBs, Hassan and Dridi (2011) report that IBs' credit growth was twice that of CBs during the GFC and, in general, was less affected by the crisis. The higher solvency of IBs played a crucial role in helping IBs support the demand for loans while the crisis unfolded. Another of their key results was that IBs conducted excessive due diligence and screening, lending loans in sectors that were not affected by the GFC. Beck et al. (2013) reached a similar conclusion using a sample of 510 banks, of which 88 are Islamic, and report that IBs are less likely to disintermediate during crisis times than CBs.

Beck et al. (2013) also reported that IBs observed superior stock market performance than CBs, which they attribute to IBs' higher asset quality and better capitalization. Farooq and Zaheer (2015) investigated how the financial panic affected banks' deposit and lending behavior in Pakistan. During the September – October 2008 financial panic, the authors reported an accelerated pattern of deposit withdrawals that was unique to CBs. IBs experienced fewer deposit withdrawals, and remarkably, some reported higher deposit rates during the financial panic period. Farooq and Zaheer (2015) contend that weaker withdrawals of deposits at IBs may explain their robust credit growth. Ibrahim (2016) provides evidence from Malaysia that lending in CBs tends to be pro-cyclical, as a decline in GDP growth results in lower loan growth. Conversely, the lending behavior of IBs is not influenced by business cycles and can thus be considered counter-cyclical. This finding supports the view that IBs could play a stabilizing role in the economy. Using data on 25 IBs and 114 CBs from 10 dual-banking countries, Ibrahim and Rizvi (2018) find no significant difference in the lending growth of IBs and CBs during normal periods. However, such behavior differs during crisis periods, implying that CBs decrease their lending during the crisis periods while IBs do not.

2.2. Hypotheses development

The comparative literature above argues that the asset-based and risk-sharing features of IBs, as opposed to the debt-based and risk-transfer features of CBs, have shielded IBs from the negative consequences of crises. These unique characteristics could also explain IBs’ ability to extend lending against the cycle. More precisely, according to Shariah principles, IBs raise funds through profit-sharing investment accounts (PSIA) that allow profits to be shared at a pre-determined rate, do not guarantee the nominal value of such deposits, and restrict losses to the account holders. Interest rates are excluded, and the returns on the bank assets determine the depositors' returns. Similarly, the profit- and loss-sharing (PLS) mechanism is also observed on the asset side by financing investments using participation loans such asMudarabahorMusharakah contracts.⁴Equity participation principle provides IBs with the flexibility to adjust to shocks during downturns. Realized losses to the bank asset value are then absorbed by a corresponding reduction in deposits held by account holders. As a result, the assets' and liabilities' values in real terms are constantly aligned with each other. The ability of IBs to engage in pass-through arrangements serves as protection from the asset-liability exposure typically faced by CBs.

The equity-based system in IBs predicts a different lending behavior than that in CBs during economic downturns, such as that associated with the COVID-19 pandemic. In a traditional setting, banks become less willing to maintain the flow of credit to corporate businesses during stressful times when their ability to make interest payments is particularly weakened. However, IBs strictly prohibit lending based on pre-determined interest rates and premises agreements whereby the generated profits are commensurate with the level of risk position or are linked to transactions in the real economy. IBs are expected to be readily available to provide credit to businesses with profitable investment opportunities with much-needed funding. The underlying profit- and loss-sharing mechanism alleviates the borrower's pressure to make interest payments independent of the investment returns. Simultaneously, the PLS mitigates the IBs’ concerns regarding losses since the investments are funded according to agreements requiring the distribution of profits between the borrower and the bank at a pre-determined rate. Evidence consistent with the conjecture that PLS lowers Islamic banks' risk aversion towards extending new loans when economic conditions deteriorate appears in Beck et al., 2013, Ibrahim, 2016.

Based on this discussion and considering the COVID-19 crisis as a natural experiment, we draw our first hypothesis as follows:

Hypothesis 1: IBs sustained a higher lending activity than CBs during the COVID-19 crisis.

Next, we consider the link between the lending behavior of IBs/CBs in the crisis period of COVID-19 to pre-crisis country characteristics, particularly the reliance on macroprudential policies. The devastating effects of the 2008 GFC spurred increasing attention on the role of financial regulations and supervision, particularly macroprudential policies that are meant to help stabilize the financial system. Despite continued efforts of the Bank for International Settlements to advocate macroprudential policies since the 2008 GCF, emerging countries' regulatory and supervisory frameworks started to incorporate macroprudential perspectives only two decades after in the early 2000s.Researchers agree on the success of macroprudential tools in curbing credit growth (e.g., Bruno et al., 2017, Cerutti et al., 2017).⁵Studies that assess the ability of macroprudential policies to influence banking indicators usually find that some devices, such as caps on loan-to-value (LTV) and changes in debt-to-income (DTI) ratios for mortgage loans, are successful in containing credit growth (Claessens et al., 2013, Lim et al., 2011). Alam et al. (2019) found that macroprudential tools such as loan-targeted instruments significantly impacted real credit to households. More recently, Gómez et al. (2020) reported that macro-prudential policies in Colombia reduced the credit cycle and risk-taking from 2006 to 2009.

Although the academic debate on the effectiveness of macroprudential policies in promoting financial stability and mitigating economic shocks is gaining momentum, it remains vastly conducted in conventional banking. Evidence on the consequences of the interplay between Islamic banking and macroprudential tools is scarce and practically nonexistent. Our study addresses this void in the empirical banking literature by comparing how the COVID-19 crisis affected the lending behavior of IBs and CBs in countries that differ in their adoption of macroprudential policies.

3. Methodology and data

3.1. Methodology

The first objective of the current study is to examine whether IBs sustain their lending compared to CBs counterparts during the COVID-19 outbreak. We check this by employing the following model:

$g_{ict} = β_{0} + β_{1} . {Crisis}_{t} \times {IB}_{ic} + δ . X_{ict} + φ_{i} + φ_{ct} + ε_{ict}$

(1)

Above, $i$ , $c$ , and $t$ denote bank $i$ , country $c$ , and quarter $t$ . $g_{ict}$ is the growth (or variation) of bank loans from the end of quarter $t - 1$ to the end of quarter $t$ , estimated as, alternatively, i) growth in total outstanding loans computed as $g (L) = ({Loan}_{t} - {Loan}_{t - 1}) / {Loan}_{t - 1}$ , (ii) change in total outstanding loans to total assets ratio ( $LoanAsset$ ) computed as $Δ (L T A) = ({LoanAsset}_{t} - {LoanAsset}_{t - 1})$ and (iii) change in total outstanding loans as a share of (2018) GDP computed as $Δ (L T G D P) = ({Loan}_{t} - {Loan}_{t - 1}) / {GDP}_{2018}$ . Compared to g(L), which measures growth in the total amount of credit created through new loans made by a bank, $Δ (L T A)$ calculates the change in relation to the bank's loan portfolio to its total assets. The third proxy, $Δ (L T G D P)$ , indicates the increase in a bank's credit scaled by the country's GDP where the bank is domiciled.

${Crisis}_{t}$ is a dummy variable that takes value one during the COVID-19 pandemic (2020Q1-2020Q3) and zero for the three quarters preceding the crisis (2019Q2-2019Q4). Using quarterly series spanning from 2019Q2 to 2020Q3, we select a period of three quarters during the pandemic and three quarters before the crisis. We then examine the relative performance of IBs during the pandemic using an indicator variable, ${IB}_{ic}$ , which is a dummy variable that takes value one for Islamic bank $i$ in country c, and zero otherwise. We treat conventional banks as the base. To examine our first hypothesis, we rely on an interaction term $, {Crisis}_{t} \times {IB}_{ic}$ . The coefficient $β_{1}$ measures the difference between performances in banks (IBs versus CBs) during the COVID-19 pandemic. A positive and significant point estimate of $β_{1}$ indicates that IBs sustained their lending during the pandemic compared to their CBs counterparts (consistent with our first hypothesis).

$X_{i c t}$ is a vector of bank-specific control variables that may explain banks' lending behavior. Specifically, we include a measure of bank size and five proxies of bank healthiness as captured by CAMEL.⁶First, $Size$ represents bank size, measured by the natural logarithm of total assets. Banks of different sizes may have different business models and lending policies. Second, we include a proxy for capital adequacy, the ratio of equity to total assets ( $Equity$ ). Banks with a high level of equity ratio face lower costs of funding. Thus, they can invest in risky assets such as loans. In addition, well-capitalized banks have a lower risk of bankruptcy and consequently have a greater capacity to withstand financial shocks (Demirgüç-Kunt et al., 2013, Kapan and Minoiu, 2018). Third, we consider loan loss provisions to total loan ratio ( $LoanLoss$ ) as a proxy for asset quality. Banks with a high level of impaired loans are assumed to be more fragile and lend less. Fourth, we capture the quality of bank management by including the cost-to-income ratio ( $Cost$ ). We assume that inefficient banks have fewer resources to originate new loans. Fifth, we add return on assets ( $ROA$ ) to consider the earnings dimension, as more profitable banks may be better positioned to lend more. Sixth, we add liquid assets to total deposits and short-term funding ratio ( $Liquidity$ ) to account for a bank's liquidity position, as one may expect that banks holding more liquid assets have a lower potential for growth.

We include bank fixed effects ( $φ_{i}$ ) in all regressions to control for time-invariant bank-specific factors that may affect bank lending policies, such as bank risk culture. This also removes unobserved heterogeneity in bank loan policies (Li et al., 2020). In addition, we capture the effects of country×time variables by including $φ_{ct}$ , which represent the time-variant country fixed effects. They account for time-variant country-specific features that might drive cross-bank differences in loan activities, such as demand for credit, macroeconomic shocks, regulatory differences, and country responses to the crisis (Demirgüç-Kunt et al., 2013). Note that $φ_{i}$ subsumes the level effects of IBs, and $φ_{ct}$ subsumes the level effects of the crisis, and thus they both fully absorb the direct effects of $IB$ and $Crisis$ in Eq. (1). These two types of fixed effects control for time-invariant variations across banks and time-variant factors across countries. By including all potential bank-level control variables that may affect bank lending activities and this rich set of fixed effects, we reduce concerns about omitted variable bias. The disturbance term is $ε_{ict}$ . Regressions are estimated using ordinary least square (OLS) estimates, and the statistical inferences are based on clustered standard errors at the bank level to address potential autocorrelation in the residuals.

The second objective of our study is to assess whether pre-crisis financial conditions affect the relative resilience of IBs to the COVID-19 pandemic (see, for example, Cornett et al., 2011 concerning the 2008 GFC and Li et al., 2020 regarding the COVID-19 pandemic). The existing literature highlights the role of pre-crisis conditions in influencing the performance of banks during the subsequent crisis (see, for example, Demirgüç-Kunt et al., 2013, Brei et al., 2013, Igan and Mirzaei, 2020, Cornett et al., 2011, Ivashina and Scharfstein, 2010, Beltratti and Stulz, 2012, Balvers et al., 2017). Specifically, as previously discussed, we expect that a country's pre-crisis activation of macroprudential policies conditions the resilience of banks during the crisis. To examine the impact of pre-crisis macroprudential policies on the loan sustainability of IBs during the crisis, we conduct a subsample analysis by splitting the dataset into two subsamples at the median proxy of pre-crisis usage of macroprudential policies. We then estimate Eq. (1) for each subsample.

3.2. Data

Bank-level quarterly data come from the ORBIS database by Bureau Van Dijk, which provides financial data for more than 40 million firms (including banks) from more than 100 countries worldwide. Our raw data cover the period 2017 Q4 – 2020 Q3. We, however, utilize 2019 Q2 up to the latest quarter available (at the time of conducting this research), which is 2020 Q3. This enables us to study growth in bank loans around the COVID-19 pandemic and to have a nearly equal number of observations in the sample before and during the crisis. Since not all banks enter the sample every quarter, our final dataset is unbalanced. We select all banks (conventional and Islamic) that belong to countries with dual banking systems. Given our interest in evaluating the resilience of IBs during the COVID-19 crisis, we focus our baseline analysis on banks that are present before and during the crisis.

As a result, 461 banks (117 IBs and 344 CBs) from 17 countries survive the above filtering criteria.⁷The number of banks in our dataset varies by country. On average, each country has about 27 banks with available data. We handle outliers by winsorizing all variables at the 1st and 99th percentiles to reduce the influence of outliers.

3.3. Descriptive evidence

As a preliminary way of exploring the data, we present some descriptive evidence on how the pandemic affected bank lending and whether IBs reacted differently from their CBs counterparts regarding loan growth.

Table 1, Panels A and B provide definitions and descriptive statistics of the main variables used in the present study for the entire sample and the subsamples of IBs and CBs. According to Panel A, the average loan growth of IBs during pre-crisis was almost equal to that of CBs but decreased for both types of banks during the crisis. Nonetheless, IBs allocated more loans than CBs during the COVID-19 pandemic. The differences between pre and during the crisis are statistically insignificant for IBs but significant for CBs. IBs are, on average, smaller, better capitalized, but less efficient than conventional banks. In addition, they are more liquid but less profitable than their rival conventional ones. These results align with Beck et al., 2013, Ibrahim and Rizvi, 2018.

Table 1

Definition and summary statistics of main variables: Islamic versus conventional banks. Panel A reports a detailed definition and a comparison of mean values between IBs and CBs for all variables in our analysis. Panel B reports a univariate analysis of bank lending. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Our sample includes 461 banks (out of which 117 are IBs) in 17 countries over 2019Q2-2020Q3 (pre-crisis: 2019Q2-2019Q4 vs crisis: 2020Q1-2020Q3). Panel A: Descriptive statistics.

Table 1A: Definition and summary statistics of all variables
		All banks			IBs			CBs
		Pre-crisis	Crisis	Diff.	Pre-crisis	Crisis	Diff.	Pre-crisis	Crisis	Diff.
Variable	Definition	(1)	(2)	(3)= (2)-(1)	(4)	(5)	(6)= (5)-(4)	(7)	(8)	(9)= (8)-(7)

Dependent variables
g(L)	The growth in total outstanding loans, calculated as g(L)=(Loan_t-Loan_t-1) /Loan_t-1.	0.024	−4E-04	−0.024***	0.022	0.014	−0.008	0.025	−0.005	−0.030***

Δ(LTA)	The chnage in total outstanding loans to asset ratio (LoanAsset), calculated as Δ(LTA)=(LoanAsset_t-LoanAsset_t-1).	−0.001	−0.006	−0.005***	−0.003	−0.002	0.001	−0.001	−0.008	−0.007***
Δ(LTGDP)	The change in total outstanding loans as share of (2018) GDP, calculated as Δ(LTGDP)=(Loan_t-Loan_t-1)/GDP₂₀₁₈.	0.422	0.259	−0.163**	0.546	0.603	0.057	0.378	0.143	−0.236***
Bseline controls
Size	Natural logarithm of a bank total assets.	14.561	14.591	0.030	14.483	14.510	0.027	14.588	14.619	0.031
Equity	The ratio of equity to total assets of a bank.	0.202	0.203	0.001	0.241	0.241	0.000	0.188	0.190	0.002
LoanLoss	Loan loss provisions to total loan ratio. A loan loss provision is an expense set aside as an allowance for bad loans.	0.008	0.008	0.000	0.008	0.008	0.000	0.008	0.008	0.000
Cost	Bank cost-to-income ratio, as calculated by dividing the operating expenses by the operating income.	0.616	0.605	−0.010	0.714	0.697	−0.017	0.584	0.576	−0.008
ROA	Return on assets, which is defined as profit before tax as a share of average assets of a bank.	0.011	0.012	0.001	0.004	0.006	0.002	0.013	0.014	0.001
Liquidity	The bank liquid assets to total deposits and short-term funding ratio.	0.439	0.462	0.023	0.564	0.540	−0.024	0.397	0.436	0.038
Other controls
ZSCORE	Bank Z-score, as a proxy for individual bank overall risk. It is computed as sum of return on asset and capital to asset ratio divided by return volatility. Return volatility is measured based on a 5-quarter window basis of volatility of the return on assets of the bank.	8.826	5.685	−3.140***	8.520	5.909	−2.611***	8.931	5.608	−3.323***
AssetDiversification	Asset diversification as measured by 1-\|(Net loans – Other earning assets)/(Total earning assets)\|. Asset diversity takes values between zero and one with higher values indicating greater diversification.	0.516	0.525	0.008	0.444	0.444	0.000	0.539	0.549	0.010
FeeIncome	Bank ratio of fee and other operating income to total assets.	0.016	0.015	−0.001	0.017	0.013	−0.004**	0.016	0.016	0.000
WholesaleFunding	Bank short-term funding to total assets ratio.	0.077	0.072	−0.006	0.074	0.071	−0.003	0.078	0.072	−0.007

Open in a separate window

Panel B: Univariate analysis of bank lending: Islamic versus conventional banks.
Table 1B: Univariate analysis of bank lending
	N	g(L)	Δ(LTA)	Δ(LTGDP)
Differences

IB_Pre - CB_Pre	1,251	−0.003	−0.002	0.168
IB_Crisis - CB_Crisis	1,205	0.019***	0.007**	0.460***

Open in a separate window

Furthermore, Table 1, Panel B shows a univariate analysis of the differences in bank lending between IBs and CBs in pre and during the COVID-19 pandemic. We find no difference between the credit growth of both types of banks in pre-crisis. However, when it comes to the COVID-19 crisis, the growth in loans by IBs is statistically more significant than that of CBs.

Table 2 presents the number of IBs and CBs, average loan growth (or variation), and the severity of the crisis by country. Out of 117 IBs in our sample, 18 banks come from Malaysia, followed by 16 from Iraq. As measured by growth in gross loans, the mean of credit growth is the highest in Syria (8%) and lowest in Pakistan (-2%) during the whole sample period. Focusing on the Oxford stringency index to measure the severity of the COVID-19 pandemic, the most affected countries are Oman and Palestine. In contrast, the least affected are Indonesia and Malaysia. Fig. 1plots the change in the average bank loan growth (measured by the growth in gross loans) between the COVID-19 crisis and pre-COVID-19 for both IBs and CBs to complement this country-level analysis. The change in credit growth is negative for eight countries and positive for the remaining countries when considering IBs. However, the change in bank loan growth for CBs is negative for most (more than 80% of) countries.

Table 2

Number of banks, loan growth, and severity of the crisis by country.

		Number of banks						Loan growth			Severity of the crisis
ID	country	Total	IBs	%	CBs	%	g(L)	Δ(LTA)	Δ(LTGDP)	Case per 100,000 population	Oxford stringency index	Opening risk index
1	Bahrain	13	8	61.5	5	38.5	−0.008	−0.002	1.078	6.481	72.530	0.769
2	Bangladesh	36	8	22.2	28	77.8	0.017	−0.007	0.294	2.155	80.557	0.448
3	Egypt	21	3	14.3	18	85.7	0.034	0.002	0.314	2.855	72.840	0.555
4	Indonesia	104	11	10.6	93	89.4	0.003	−0.007	0.013	2.387	59.263	0.643
5	Iraq	30	16	53.3	14	46.7	0.039	0.000	−0.004	4.097	83.333	0.647
6	Jordan	19	5	26.3	14	73.7	0.015	0.001	1.102	2.740	70.680	0.389
7	Kenya	30	3	10.0	27	90.0	0.021	−0.002	0.232	1.556	80.867	0.474
8	Kuwait	15	10	66.7	5	33.3	0.021	0.001	1.416	5.601	77.470	0.673
9	Malaysia	42	18	42.9	24	57.1	0.010	0.001	0.389	3.018	59.570	0.418
10	Oman	9	2	22.2	7	77.8	0.002	−0.004	0.421	5.241	86.420	0.667
11	Pakistan	31	9	29.0	22	71.0	−0.019	−0.012	−0.140	3.201	65.280	0.613
12	Palestine	5	2	40.0	3	60.0	0.035	0.001	2.064	3.901	86.423	0.603
13	Qatar	10	5	50.0	5	50.0	0.012	0.001	1.383	6.633	77.160	0.693
14	Saudi Arabia	13	5	38.5	8	61.5	0.034	0.001	1.531	4.925	72.840	0.562
15	Syria	14	3	21.4	11	78.6	0.083	−0.013	0.319	0.295	67.130	0.565
16	Turkey	47	3	6.4	44	93.6	−0.012	−0.002	−0.001	4.750	69.910	0.585
17	UAE	22	6	27.3	16	72.7	0.015	−0.002	0.560	5.014	69.137	0.536

All		461	117	25.4	344	74.6	0.012	−0.004	0.342	3.333	69.778	0.570

Open in a separate window

Fig. 1

Change in bank loan growth (as measured by $g (L)$ ) between crisis (average 2020Q1-2020Q3) and pre-crisis (average 2019Q1-2019Q4) periods: Islamic vs conventional banks.

To investigate further the impact of the COVID-19 crisis on IBs and CBs, we compare the resilience of banks based on bank distance to default (DTD) and probability of default (PD). The former reveals how far away a bank is from default, with higher figures denoting lower risk. The latter reflects the default risk of publicly-listed firms by quantitatively analyzing numerous covariates. The data are for 43 IBs, and 145 CBs obtained from the Credit Research Initiative, National University of Singapore. Appendix Table A1 summarizes the effects of the COVID-19 crisis on these bank risk indicators. It is worth highlighting that the impact on both types of banks is negative, but the adverse impact is more pronounced for CBs.

Overall, we find that: (i) the COVID-19 pandemic adversely affected the lending behavior of both IBs and CBs; (ii) there is no significant difference between loan growth for IBs and CBs in the pre-crisis period; (iii) IBs performed better with regards to lending during the crisis, as compared to their conventional counterparts, and (iii) the adverse impact of the COVID-19 on banks is also evident if considering other performance indicators, such as firm survival. Next, we turn to formal regression analyses to understand the relative performance of IBs during the crisis based on such an essential heterogeneity in bank lending observed across both types of banks.

4. Results

4.1. Baseline results

We start our analysis by examining how IBs behave differently from CBs during the COVID-19 pandemic concerning loan growth. Specifically, our main research question is whether IBs maintain their lending during the pandemic more than CBs. Table 3reports the results from estimating Eq. (1), using growth in total outstanding loans ( $g (L)$ ), as well as change in total outstanding loans to total assets ratio ( $Δ (L T A)$ ) and change in total outstanding loans as a share of GDP ( $Δ (L T G D P)$ ) as the response variables. Our interest variable is the interaction term between Crisis and IB dummy indicators ( $Crisis \times I B$ ).

Table 3

Bank lending during the COVID-19 pandemic: Islamic banks vs conventional banks. Baseline results This table reports the results estimating $g_{ict} = β_{0} + β_{1} . {Crisis}_{t} \times {IB}_{ic} + δ . X_{ict} + φ_{i} + φ_{ct} + ε_{ict}$ where $i$ , $c,$ and $t$ denote bank $i$ , country $c$ and quarter $t$ . $g_{ict}$ is, alternatively, i) growth in total outstanding loans: $g (L)$ , (ii) change in total outstanding loans to total assets ratio: $Δ (L T A)$ , and (iii) change in total outstanding loans as a share of GDP: $Δ (L T G D P)$ , from quarter $t - 1$ to quarter $t$ . ${Crisis}_{t}$ is a dummy variable that takes value 1 during the COVID-19 pandemic (2020Q1-2020Q3) and zero before the pandemic (2019Q2-2019Q4). ${IB}_{ic}$ is a dummy variable that takes value 1 if bank $i$ domiciled in country $c$ is an Islamic bank, and zero otherwise. $X_{ict}$ is a vector of bank-specific variables that may explain the lending behavior of banks. We include bank fixed effects ( $φ_{i}$ ) and country-year/quarter fixed effects ( $φ_{ct}$ ) in all regressions. See Table 1 for a detailed definition of variables. Regressions are estimated using OLS. The statistical inferences are based on clustered standard errors at the bank level (associated t-values reported in parentheses). ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Our sample includes 461 banks (out of which 117 are IBs) in 17 countries over the period 2019Q2-2020Q3.

	g(L)	Δ(LTA)	Δ(LTGDP)
	(1)	(2)	(3)
Crisis_t×IB_ic	0.025**	0.013***	0.557**
	(2.037)	(2.883)	(2.567)
Size_ict	0.105***	−0.135***	0.771*
	(2.744)	(-7.063)	(1.854)
C: Equity_ict	0.197	0.172**	0.805
	(1.581)	(2.184)	(1.059)
A: LoanLoss_ict	−1.004	−0.223	−1.585
	(-1.369)	(-1.219)	(-0.439)
M: Cost_ict	−0.030	−0.010	0.011
	(-1.030)	(-0.918)	(0.098)
E: ROA_ict	−0.139	−0.041	1.627
	(-0.385)	(-0.282)	(0.431)
L: Liquidity_ict	−0.014	−0.017**	−0.162
	(-0.959)	(-2.018)	(-1.194)
Constant	−1.829***	2.341***	−10.355
	(-2.733)	(6.939)	(-1.376)
Bank FEs	Y	Y	Y
Country*Year-Quarter FEs	Y	Y	Y
# Countries	17	17	17
# Banks	461	461	461
N	2,104	2,104	2,104
Adj. R²	0.331	0.060	0.269

Open in a separate window

4.2. Sensitivity tests

So far, we have observed the resilience of IBs during the COVID-19 outbreak for loan growth, which is in line with our first hypothesis. We do several sensitivity tests to ensure our primary finding is robust in this section.⁸Results of six robustness tests are reported in Table 4Panels A, B, and C when the dependent variable is $g (L)$ , $Δ (L T A)$ , and $Δ (L T G D P)$ , respectively.

Table 4

Sensitivity tests.

4A: $g (L)$
	Cluster at country	WLS	Net loans	Since 2017Q4	Lag controls	Other controls
	(1)	(2)	(3)	(4)	(5)	(6)
Crisis_t×IB_ic	0.025*	0.038***	0.024**	0.017*	0.036***	0.020*
	(1.897)	(2.620)	(2.055)	(1.865)	(3.499)	(1.731)
Size_ict	0.105*	0.056	0.132***	0.050	−0.153***	0.107***
	(1.938)	(0.689)	(3.276)	(1.273)	(-2.711)	(2.678)
C: Equity_ict	0.197	−0.943	0.175	0.368***	0.498**	0.235*
	(1.154)	(-1.176)	(1.391)	(3.372)	(2.253)	(1.682)
A: LoanLoss_ict	−1.004	−0.582	−1.597**	−0.873*	−1.179*	−0.533
	(-1.297)	(-0.387)	(-2.207)	(-1.764)	(-1.687)	(-0.946)
M: Cost_ict	−0.030	−0.023	−0.026	−0.028**	−0.054*	−0.041
	(-0.960)	(-0.543)	(-0.865)	(-2.434)	(-1.773)	(-1.382)
E: ROA_ict	−0.139	1.542*	−0.162	−0.280	−0.380	0.177
	(-0.333)	(1.847)	(-0.440)	(-0.907)	(-0.871)	(0.463)
L: Liquidity_ict	−0.014	−0.041	−0.013	−0.020*	0.047	−0.037**
	(-0.966)	(-1.368)	(-0.812)	(-1.905)	(1.473)	(-2.072)
Other controls_ict
(ZSCORE, AssetDiversification, FeeIncome, WholesaleFunding)	–	–	–	–	–	√
Constant	−1.829*	−0.843	−2.315***	−0.851	2.641***	−1.843***
	(-1.933)	(-0.572)	(-3.277)	(-1.238)	(2.648)	(-2.656)
Bank FEs	Y	Y	Y	Y	Y	Y
Country*Year-Quarter FEs	Y	Y	Y	Y	Y	Y
# Countries	17	17	17	17	17	17
# Banks	461	461	461	461	461	461
N	2,104	2,104	2,104	3,848	2,114	2,074
Adj. R²	0.331	0.322	0.339	0.277	0.369	0.359

Open in a separate window

4B: $Δ (L T A)$
Table 4B: Sensitivity tests
	Cluster at country	WLS	Net loans	Since 2017Q4	Lag controls	Other controls
	(1)	(2)	(3)	(4)	(5)	(6)
Crisis_t x IB_ic	0.013***	0.015**	0.035***	0.010***	0.007	0.012**
	(4.676)	(2.563)	(3.013)	(2.784)	(1.359)	(2.560)
Size_ict	−0.135***	−0.211***	−0.268***	−0.080***	0.132***	−0.123***
	(-14.442)	(-5.312)	(-5.923)	(-6.872)	(5.384)	(-6.658)
C: Equity_ict	0.172***	0.378*	0.481***	0.165***	−0.079	0.186***
	(3.949)	(1.792)	(3.080)	(3.503)	(-1.193)	(2.600)
A: LoanLoss_ict	−0.223*	0.264	−1.533**	−0.108	−0.067	−0.326
	(-1.771)	(0.954)	(-2.087)	(-0.780)	(-0.397)	(-1.552)
M: Cost_ict	−0.010**	−0.009	−0.031	−0.007	−0.004	−0.010
	(-2.658)	(-0.756)	(-1.217)	(-1.301)	(-0.630)	(-0.871)
E: ROA_ict	−0.041	0.001	−0.444	−0.109	−0.048	−0.152
	(-0.596)	(0.005)	(-1.054)	(-1.063)	(-0.383)	(-0.951)
L: Liquidity_ict	−0.017	−0.081***	−0.026	−0.023***	0.033**	−0.024***
	(-1.676)	(-4.086)	(-1.430)	(-3.192)	(2.230)	(-2.705)
Other controls_ict
(ZSCORE, AssetDiversification, FeeIncome, WholesaleFunding)	–	–	–	–	–	√
Constant	2.341***	3.652***	4.631***	1.395***	−2.299***	2.205***
	(14.533)	(5.165)	(5.849)	(6.781)	(-5.372)	(6.748)
Bank FEs	Y	Y	Y	Y	Y	Y
Country*Year-Quarter FEs	Y	Y	Y	Y	Y	Y
# Countries	17	17	17	17	17	17
# Banks	461	461	461	461	461	461
N	2,104	2,104	2,104	3,848	2,114	2,074
Adj. R²	0.060	0.203	0.173	0.069	0.070	0.087

Open in a separate window

4C: $Δ (L T G D P)$
Table 4C: Sensitivity tests
	Cluster at country	WLS	Net loans	Since 2017Q4	Lag controls	Other controls
	(1)	(2)	(3)	(4)	(5)	(6)
Crisis_t x IB_ic	0.557*	1.240*	0.495***	0.447**	0.537**	0.511**
	(2.052)	(1.745)	(2.738)	(2.540)	(2.491)	(2.293)
Size_ict	0.771	3.291	0.812**	0.879***	−1.728***	1.102**
	(1.485)	(0.831)	(2.265)	(3.485)	(-2.609)	(2.534)
C: Equity_ict	0.805	10.356	0.810	1.370*	0.275	0.723
	(1.239)	(0.528)	(1.168)	(1.705)	(0.220)	(0.899)
A: LoanLoss_ict	−1.585	23.595	−4.727	−1.177	−2.397	−2.572
	(-0.697)	(0.825)	(-1.445)	(-0.417)	(-0.749)	(-0.719)
M: Cost_ict	0.011	−1.085	0.078	−0.027	0.023	0.038
	(0.090)	(-1.174)	(0.808)	(-0.444)	(0.218)	(0.279)
E: ROA_ict	1.627	50.329**	3.320	−0.462	1.115	3.404
	(0.388)	(2.024)	(1.151)	(-0.191)	(0.308)	(0.831)
L: Liquidity_ict	−0.162	−1.682	−0.161	−0.212*	0.264	−0.318*
	(-1.088)	(-1.265)	(-1.170)	(-1.775)	(1.161)	(-1.708)
Other controls_ict
(ZSCORE, AssetDiversification, FeeIncome, WholesaleFunding)	–	–	–	–	–	√
Constant	−10.355	−55.849	−12.628**	−10.575**	32.966***	−15.111*
	(-1.143)	(-0.796)	(-1.985)	(-2.314)	(2.814)	(-1.945)
Bank FEs	Y	Y	Y	Y	Y	Y
Country*Year-Quarter FEs	Y	Y	Y	Y	Y	Y
# Countries	17	17	17	17	17	17
# Banks	461	461	461	461	461	461
N	2,104	2,104	2,104	3,848	2,114	2,074
Adj. R²	0.269	0.417	0.307	0.238	0.265	0.277

Open in a separate window

Our baseline specification includes standard errors clustered at the bank level. We employ-two alternative empirical approaches to this clustering. First, a general preference is a cluster at the higher level of aggregation at the country level. Second, we use Weighted Least Square (WLS) regression as an alternative to clustering with the number of banks in each country as weights. Columns 1 and 2 in Table 4 report both results. The estimation results with the country-level clustering and WLS are consistent with the main results. IBs fared better during the COVID-19 crisis period in terms of credit growth.

Third, a reasonable concern is that our choice of measure drives the results for loan growth. We use the data for gross loans in the baseline to construct our dependent variables. We now use net loans, defined as gross loans minus allowances for loan losses, to formulate bank loan growth. While a bank's liquidity can be viewed as the value of its gross loans, net loans represent the actual performing loans. Therefore, we check whether this alternative measure corroborates our findings. The results are reported in Column 3. We again find that IBs sustain their lending during the current health crisis compared to CBs.

Fourth, we consider an alternative time horizon for the dependent variables. We used three periods for pre-crisis and three periods during the crisis in the baseline. We aimed to rule out any misrepresentation from possible survival bias by extending the pre-crisis to 2017 Q4 (the past available year data). The reported results in Column 4 confirm the primary finding that Islamic banks are more resilient.

Fifth, our findings remain mostly robust to using the lag of control variables in Column 5, Panels A and C. Using the lagged variable can mitigate the potential concern about the reverse causality between bank balance sheet variables and bank credit growth.

Finally, while we acknowledge that we control for a range of bank-level variables, some unobservable factors could explain our primary finding of less vulnerability of IBs during the COVID-19 pandemic. We now attempt to control for additional bank-level control variables to address the concern that our findings may be biased due to omitted variables. We add to the model the following four bank-specific variables. (i) Z-score ( $ZSCORE$ ), computed as $Zscore = \frac{ROA + E T A}{sigROA}$ , where $ROA$ is the return on assets, $ETA$ is the equity to asset ratio, and $SigROA$ is the standard deviation of $ROA$ over a five-quarter window. When facing adverse shocks, solvent banks are in a better position to absorb such shocks, and thus, it is natural that solvent banks maintain their lending during a crisis. Following Boubakri et al. (2017), we include two diversification variables: (ii) asset diversification ( $AssetDiversification$ ) as measured by 1-|(Net loans – Other earning assets)/(Total earning assets)| and (iii) fee income ( $FeeIncome$ ) measured as the ratio of fee and other operating income to total assets. It is argued that diversified banks (both in terms of assets and income) generate more profits. Hence, these banks are more resilient to financial instability and more willing to allocate funds (Elsas et al., 2010). (iv) Wholesale funding ( $WholesaleFunding$ ), defined as bank short-term funding to total assets ratio. We capture the bank's liabilities structure, which may affect its stability and credit growth (Craig and Dinger, 2013, Zheng, 2020). Ippolito et al. (2016) find that banks reliant on wholesale funds reduce their loans more than banks reliant on core deposits during a crisis. The estimation results with these additional control variables are reported in Column 6. They show that the baseline results that loan growth in IBs is higher than CBs during the COVID-19 crisis remain unchanged.

We next employ an alternative model by excluding the $Crisis$ dummy from the model and splitting the sample into the pre-crisis versus crisis periods. Table 5shows the results where Columns 1–3 report results when running regression for the pre-crisis period, and Columns 4–6 report results during the crisis period. The dummy $IB$ is statistically significant during the crisis but insignificant in pre-crisis. These results also reinforce our main finding that IBs were more resilient during the early stage of the COVID-19 pandemic.

Table 5

Robust to splitting the sample to pre-crisis and crisis periods. This table reports the results estimating $g_{ict} = β_{0} + β_{1} . {IB}_{ic} + δ . X_{ict} + φ_{ct} + ε_{ict}$ where $i$ , $c,$ and $t$ denote bank $i$ , country $c$ and quarter $t$ . $g_{ict}$ is, alternatively, i) growth in total outstanding loans: $g (L)$ , (ii) change in total outstanding loans to total assets ratio: $Δ (L T A)$ , and (iii) change in total outstanding loans as a share of GDP: $Δ (L T G D P)$ , from quarter $t - 1$ to quarter $t$ . ${IB}_{ic}$ is a dummy variable that takes value 1 if bank $i$ domiciled in country $c$ is an Islamic bank, and zero otherwise. $X_{ict}$ is a vector of bank-specific variables that may explain the lending behavior of banks. We include country-year/quarter fixed effects ( $φ_{ct}$ ) in all regressions. See Table 1 for a detailed definition of variables. Regressions are estimated using OLS. The statistical inferences are based on clustered standard errors at the bank level (associated t-values reported in parentheses). ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Our sample includes 461 banks (out of which 117 are IBs) in 17 countries over 2019Q2-2020Q3.

	Splitting sample to
	Pre-crisis (2019Q2-2019Q4)			Crisis (2020Q1-2020Q3)
	g(L)	Δ(LTA)	Δ(LTGDP)	g(L)	Δ(LTA)	Δ(LTGDP)
	(1)	(2)	(3)	(4)	(5)	(6)
IB_ic	−0.002	−0.004	−0.033	0.027***	0.005*	0.482***
	(-0.222)	(-1.518)	(-0.244)	(3.634)	(1.934)	(2.684)
Size_ict	−0.000	−0.002**	0.252***	0.003	0.000	0.099**
	(-0.115)	(-2.077)	(5.570)	(1.211)	(0.543)	(2.337)
C: Equity_ict	0.030	−0.021*	0.435	−0.058	0.017	−0.471
	(0.594)	(-1.688)	(0.928)	(-1.071)	(1.064)	(-0.924)
A: LoanLoss_ict	−0.222	−0.208**	−6.029	−0.758**	−0.107	−5.472
	(-0.459)	(-2.271)	(-1.307)	(-2.312)	(-0.988)	(-1.620)
M: Cost_ict	−0.015	−0.003	−0.027	0.010	0.006	−0.032
	(-1.049)	(-0.692)	(-0.215)	(0.576)	(1.176)	(-0.209)
E: ROA_ict	0.140	0.130	1.759	0.033	−0.097	−1.300
	(0.432)	(1.614)	(0.527)	(0.150)	(-1.432)	(-0.341)
L: Liquidity_ict	−0.012	−0.003	0.008	0.008	−0.004	0.069
	(-0.786)	(-0.855)	(0.103)	(0.866)	(-1.145)	(0.890)
Constant	0.031	0.036*	−1.888	−0.054	−0.020	−1.661**
	(0.472)	(1.663)	(-1.245)	(-1.110)	(-1.145)	(-2.072)
Bank FEs	N	N	N	N	N	N
Country*Year-Quarter FEs	Y	Y	Y	Y	Y	Y
# Countries	17	17	17	17	17	17
# Banks	461	461	461	461	461	461
N	1,093	1,093	1,093	1,011	1,011	1,011
Adj. R²	0.095	0.046	0.243	0.425	0.083	0.217

Open in a separate window

Another robustness test is conducted to ascertain the baseline results in Table 3. Not all countries were affected in the same manner by the COVID-19 pandemic. For example, while Bahrain was affected more severely by the crisis, Kenya was affected slightly. We show that the resilience of IBs during the COVID-19 outbreak for bank lending is more pronounced in countries critically affected by the pandemic. We assert that IBs’ resilience in maintaining their credit growth during the health crisis is evident in countries more severely affected by the crisis. We apply a form of Eq. (1) where our interest variable is an interaction term between the severity of the COVID-19 crisis and Islamic bank dummy ( $S e v e r i t y_c r i s i s \times I B$ ). We use three proxies for the severity of the COVID-19 pandemic. (i) We employ the number of COVID-19 cases per 100,000 people in each country. The data are from the John Hopkins University dataset. (ii) The second proxy is the stringency of COVID-19, a country-level severity of the lockdown measures in response to the pandemic. This composite measure is based on nine response indicators, including school closures, workplace closures, and travel bans (Hale et al., 2020). (iii) The third proxy is the risk of openness index. It calculates a country's risk from adopting an 'open' policy stance. The data for the last two proxies are from the Oxford University dataset. Columns 1–3, 4–6, and 7–9 in Table 6report the results on bank lending of IBs in countries that were affected severely by the COVID-19 crisis where the severity is the number of cases, stringency index, or risk of openness index, respectively. We find that IBs in countries significantly hit by the COVID-19 crisis performed relatively better than CBs, only when severity is measured by the stringency index or the risk of openness index.

Table 6

Robust to the severity of the crisis. This table reports the results estimating $g_{ict} = β_{0} + β_{1} . {S e v e r i t y_c r i s i s}_{ct} \times {IB}_{ic} + δ . X_{ict} + φ_{i} + φ_{ct} + ε_{ict}$ where $i$ , $c,$ and $t$ denote bank $i$ , country $c$ and quarter $t$ . $g_{ict}$ is, alternatively, i) growth in total outstanding loans: $g (L)$ , (ii) change in total outstanding loans to total assets ratio: $Δ (L T A)$ , and (iii) change in total outstanding loans as a share of GDP: $Δ (L T G D P)$ , from quarter $t - 1$ to quarter $t$ . ${S e v e r i t y_c r i s i s}_{ct}$ is a proxy for severity of the COVID-19 pandemic in country $c$ in quarter $t$ . ${IB}_{ic}$ is a dummy variable that takes value 1 if bank $i$ domiciled in country $c$ is an Islamic bank, and zero otherwise. $X_{ict}$ is a vector of bank-specific variables that may explain the lending behavior of banks. We include bank fixed effects ( $φ_{i}$ ) and country-year/quarter fixed effects ( $φ_{ct}$ ) in all regressions. See Table 1 for a detailed definition of variables. Regressions are estimated using OLS. The statistical inferences are based on clustered standard errors at the bank level (associated t-values reported in parentheses). ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Our sample includes 461 banks (out of which 117 are IBs) in 17 countries over 2019Q2-2020Q3.

	Cases per 100,000 population			Oxford stringency index			Risk of openness index
	g(L)	Δ(LTA)	Δ(LTGDP)	g(L)	Δ(LTA)	Δ(LTGDP)	g(L)	Δ(LTA)	Δ(LTGDP)
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
Severity_crisis_ct×IB_ic	0.003	0.001	0.053	0.0003*	0.0002***	0.009**	0.037*	0.018**	0.722*
	(1.242)	(0.731)	(1.125)	(1.747)	(2.785)	(2.564)	(1.682)	(2.485)	(1.887)
Size_ict	0.109***	−0.132***	0.868**	0.106***	−0.135***	0.750*	0.106***	−0.135***	0.805*
	(2.870)	(-7.015)	(2.045)	(2.760)	(-7.068)	(1.820)	(2.771)	(-7.068)	(1.915)
C: Equity_ict	0.193	0.170**	0.700	0.196	0.172**	0.796	0.197	0.172**	0.781
	(1.545)	(2.167)	(0.905)	(1.570)	(2.170)	(1.062)	(1.577)	(2.178)	(1.019)
A: LoanLoss_ict	−1.019	−0.232	−1.943	−1.020	−0.231	−1.915	−1.002	−0.222	−1.597
	(-1.385)	(-1.272)	(-0.528)	(-1.392)	(-1.268)	(-0.535)	(-1.372)	(-1.216)	(-0.448)
M: Cost_ict	−0.030	−0.010	0.027	−0.030	−0.010	0.008	−0.030	−0.010	0.028
	(-1.018)	(-0.888)	(0.237)	(-1.031)	(-0.921)	(0.071)	(-1.008)	(-0.882)	(0.241)
E: ROA_ict	−0.179	−0.064	0.694	−0.150	−0.046	1.562	−0.146	−0.046	1.352
	(-0.494)	(-0.444)	(0.176)	(-0.415)	(-0.319)	(0.421)	(-0.407)	(-0.316)	(0.355)
L: Liquidity_ict	−0.015	−0.018**	−0.196	−0.015	−0.018**	−0.168	−0.014	−0.018**	−0.175
	(-1.030)	(-2.100)	(-1.403)	(-0.991)	(-2.037)	(-1.219)	(-0.973)	(-2.029)	(-1.268)
Constant	−1.889***	2.285***	−12.009	−1.839***	2.337***	−9.986	−1.837***	2.332***	−10.921
	(-2.857)	(6.888)	(-1.568)	(-2.749)	(6.942)	(-1.338)	(-2.757)	(6.946)	(-1.438)
Bank FEs	Y	Y	Y	Y	Y	Y	Y	Y	Y
Country*Year-Quarter FEs	Y	Y	Y	Y	Y	Y	Y	Y	Y
# Countries	17	17	17	17	17	17	17	17	17
# Banks	461	461	461	461	461	461	461	461	461
N	2,104	2,104	2,104	2,104	2,104	2,104	2,104	2,104	2,104
Adj. R²	0.329	0.055	0.266	0.330	0.060	0.270	0.330	0.059	0.268

Open in a separate window

4.3. Role of pre-crisis usage of macroprudential measures

We finally check our section hypothesis that using macroprudential tools in pre-pandemic may affect the performance of banks during the pandemic. Previous studies (e.g., Abedifar et al., 2013, Bilgin et al., 2021) find that country-level variables may shape the performance of IBs versus CBs.

We consider the impact of macro-prudential policies on the lending behavior of IBs versus CBs during the COVID-19 pandemic. Some studies investigate the mitigating role of macro-prudential policies in the aftermath of the current health crisis (Igan et al., 2022). Following the classification proposed by Cerutti et al. (2017), we account for macro-prudential policies using a total index measure (MPI_Total), a financial institution-based index (MPI_Finance), aimed at improving the liquidity position of banks, and a borrower-based index (MPI_Borrower), aimed at controlling the borrowers’ leverage and financial positions. $M P I_B o r r o w e r$ covers i) loan-to-value ratio (LTV), and ii) debt-to-income ratio (DTI). $M P I_F i n a n c e$ covers: i) limits on foreign currency loans, ii) limits on domestic currency loans, iii) reserve requirement ratio, iv) limits on interbank exposures, v) countercyclical capital-buffer requirement, vi) dynamic loan loss provisioning, vii) leverage ratio for the bank, viii) capital surcharges on systematically important financial institutions, ix) concentration limits, and x) tax on financial institutions. For a given country, the value of the $M P I_B o r r o w e r$ variable is between 0 and 2. Similarly, the value of the $M P I_F i n a n c e$ variable ranges from 0 to 10, and the value of the total variable ( $M P I_T o t a l$ ) from 0 to 12. A yearly dummy variable is designated a value of unity if the tool was activated (or was in place) and zero otherwise (Cerutti et al., 2017).

Using the data for macro-prudential tools in the latest available year, 2017, we split the sample countries into a high or low category based on their macro-prudential usage. We assign countries to the high macro-prudential usage category if their macro-prudential indicator ranks above the cross-country median. Similarly, countries with macro-prudential indicators falling below the cross-country median are grouped in the low macro-prudential usage category. Our cross-country median of $M P I_T o t a l$ is 4.73, with a minimum of 1 (Kenya) and a maximum of 8 (Turkey). Regarding $M P I_B o r r o w e r$ , the sample median is 1.33, with a range of 0 to 2. Finally, the sample median of $M P I_F i n a n c e$ is 3.4, with the lowest score at 0 and the highest at 6.

The results are presented in Table 7. We find that IBs extended more loans than CBs during the COVID-19 crisis only in countries that activated macroprudential policies, i.e., countries with high macro-prudential usage (either in terms of MPI_Total, MPI_Finance, or MPI_Borrower) in the year approaching the health crisis. There is no significant difference in the lending behavior between Islamic and conventional banks during the crisis period in countries with low macro-prudential usage. This finding is consistent across all three macro-prudential indicators, except for the results in Column 7, as reported in Columns 1–3, 8–9, and 13–15, respectively. The reported results highlight the importance of macro-prudential policies in supporting the differential ability of IBs to sustain credit growth during bad times over their counterparts.

Table 7

Role of the pre-crisis usage of macroprudential measures This table reports the results estimating ${Δ L o a n}_{ict} = β_{0} + β_{1} . {Crisis}_{t} \times {IB}_{ic} + δ . X_{ict} + φ_{i} + φ_{ct} + ε_{ict}$ where $i$ , $c,$ and $t$ denote bank $i$ , country $c$ and quarter $t$ . Each panel displays the results obtained by running the regression in a subsample determined by the median value of pre-crisis macroprudential measures proxy. $g_{ict}$ is, alternatively, i) growth in total outstanding loans: $g (L)$ , (ii) change in total outstanding loans to total assets ratio: $Δ (L T A)$ , and (iii) change in total outstanding loans as a share of GDP: $Δ (L T G D P)$ , from quarter $t - 1$ to quarter $t$ . ${Crisis}_{t}$ is a dummy variable that takes value 1 during the COVID-19 pandemic (2020Q1-2020Q3) and zero before the pandemic (2019Q2-2019Q4). ${IB}_{ic}$ is a dummy variable that takes value 1 if bank $i$ domiciled in country $c$ is an Islamic bank, and zero otherwise. $X_{ict}$ is a vector of bank-specific variables that may explain the lending behavior of banks. We include bank fixed effects ( $φ_{i}$ ) and country-year/quarter fixed effects ( $φ_{ct}$ ) in all regressions. See Table 1 for a detailed definition of variables. Regressions are estimated using OLS. The statistical inferences are based on clustered standard errors at the bank level (associated t-values reported in parentheses). ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Our sample includes 461 banks (out of which 117 are IBs) in 17 countries over 2019Q2-2020Q3.

	MPI_Total						MPI_Finance						MPI_Borrower
	Countries with low usage			Countries with high usage			Countries with low usage			Countries with high usage			Countries with low usage			Countries with high usage
	g(L)	Δ(LTA)	Δ(LTGDP)	g(L)	Δ(LTA)	Δ(LTGDP)	g(L)	Δ(LTA)	Δ(LTGDP)	g(L)	Δ(LTA)	Δ(LTGDP)	g(L)	Δ(LTA)	Δ(LTGDP)	g(L)	Δ(LTA)	Δ(LTGDP)
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)
Crisis_t×IB_ic	0.025	0.009	−0.069	0.022*	0.014***	0.745***	0.035*	0.009	0.549	0.016	0.015***	0.541**	−0.001	0.003	−0.125	0.022**	0.013***	0.582***
	(0.913)	(0.980)	(-0.154)	(1.761)	(2.800)	(2.878)	(1.665)	(1.077)	(1.241)	(1.219)	(2.626)	(2.139)	(-0.018)	(0.297)	(-0.079)	(2.035)	(2.743)	(2.599)
Size_ict	−0.097	−0.141***	0.580	0.139***	−0.146***	0.745	−0.076	−0.137***	0.965	0.142***	−0.145***	0.601	0.364	0.038	15.542	0.104***	−0.140***	0.502
	(-0.961)	(-3.819)	(0.515)	(3.532)	(-7.164)	(1.597)	(-0.884)	(-4.013)	(0.804)	(3.462)	(-6.912)	(1.424)	(1.692)	(0.290)	(1.630)	(2.669)	(-7.183)	(1.359)
C: Equity_ict	0.013	0.058	−0.184	0.196	0.198**	0.822	−0.052	0.073	−0.992	0.221*	0.196**	1.031	0.225	0.144	33.005	0.192	0.180**	1.011
	(0.034)	(0.640)	(-0.086)	(1.553)	(2.329)	(0.915)	(-0.152)	(0.821)	(-0.450)	(1.722)	(2.303)	(1.165)	(0.162)	(0.276)	(1.334)	(1.550)	(2.301)	(1.399)
A: LoanLoss_ict	2.225	0.562	5.414	−1.516*	−0.381**	−0.508	2.014	0.397	5.476	−1.579*	−0.339*	−1.520	9.827***	1.444**	6.979	−1.397**	−0.271	−1.405
	(1.388)	(1.372)	(0.331)	(-1.932)	(-1.972)	(-0.143)	(1.384)	(1.039)	(0.370)	(-1.963)	(-1.728)	(-0.434)	(3.937)	(2.363)	(0.228)	(-2.019)	(-1.494)	(-0.390)
M: Cost_ict	−0.017	−0.008	1.110	−0.029	−0.010	−0.004	0.000	0.013	0.893	−0.030	−0.011	−0.034	−0.188	−0.023	3.975	−0.029	−0.010	−0.027
	(-0.238)	(-0.243)	(0.892)	(-0.928)	(-0.885)	(-0.037)	(0.005)	(0.438)	(0.943)	(-0.977)	(-1.002)	(-0.303)	(-0.938)	(-0.349)	(0.824)	(-0.978)	(-0.905)	(-0.232)
E: ROA_ict	2.718	0.711	36.384*	−0.304	−0.081	0.671	2.451	0.765*	29.682	−0.306	−0.083	−0.156	0.194	−0.288	26.370	−0.199	−0.037	0.461
	(1.445)	(1.327)	(1.754)	(-0.814)	(-0.556)	(0.184)	(1.611)	(1.730)	(1.505)	(-0.814)	(-0.580)	(-0.042)	(0.109)	(-0.586)	(0.493)	(-0.532)	(-0.249)	(0.129)
L: Liquidity_ict	−0.043	−0.036***	−0.357	−0.008	−0.012	−0.111	−0.039	−0.037***	−0.250	−0.009	−0.011	−0.123	−0.091	−0.042	−1.428	−0.010	−0.016*	−0.142
	(-1.482)	(-5.615)	(-0.922)	(-0.543)	(-1.360)	(-0.914)	(-1.302)	(-5.188)	(-0.653)	(-0.619)	(-1.341)	(-1.012)	(-0.727)	(-1.510)	(-1.109)	(-0.762)	(-1.864)	(-1.053)
Constant	1.683	2.441***	−7.933	−2.104***	2.208***	−11.185	1.314	2.368***	−14.295	−2.154***	2.196***	−8.973	−6.318*	−0.648	−272.633	−1.806***	2.423***	−5.639
	(0.954)	(3.783)	(-0.399)	(-3.454)	(7.050)	(-1.563)	(0.886)	(3.967)	(-0.675)	(-3.389)	(6.805)	(-1.388)	(-1.701)	(-0.280)	(-1.653)	(-2.659)	(7.067)	(-0.836)
Bank FEs	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
Country*Year-Quarter FEs	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
# Countries	17	17	17	17	17	17	17	17	17	17	17	17	17	17	17	17	17	17
# Banks
N	541	541	541	1,563	1,563	1,563	643	643	643	1,461	1,461	1,461	135	135	135	1,969	1,969	1,969
Adj. R²	0.240	0.047	0.160	0.378	0.070	0.307	0.226	0.041	0.132	0.389	0.071	0.334	0.607	0.051	0.149	0.336	0.068	0.274

Open in a separate window

5. Conclusion

Adherence to Islamic principles makes IBs blend concepts of moral and social values with banking transactions. Enshrined in Shariah, Islamic values condemn interest rates or excessive risk as unethical and advocate fair banking, as exemplified in profit-loss sharing contracts. The salient feature of aligning business transactions with Shariah principles is that it acts as an additional layer of governance on top of any other regulations and has proven beneficial to IBs, especially in times of crisis. Supported by empirical findings, previous studies that compare IBs to their counterparts acknowledge that the IBs are more stable during stressful periods such as the recent GFC. Our findings lend further support to the resilience of IBs during challenging times. The reported evidence shows that credit grew 2.5% faster for IBs compared to CBs during the initial phase of the COVID-19 crisis period.

We further show that our primary finding of loan growth sustainability for IBs during the COVID-19 pandemic is more robust in countries where regulators were more active in utilizing macroprudential policies in the pre-COVID-19 pandemic. These observations imply that the lending behavior of IBs during downturns was shaped by bank regulations in the years preceding the crisis.

Our results on the determinants of bank resilience during downturns, especially in dual banking systems where IBs compete with CBs, are essential given the central role of bank resilience in financial stability and economic growth. Our evidence also sheds light on how banking principles affect bank resilience. We add to the literature on Islamic banking by showing that IBs were more resilient than CBs during the COVID-19 period. In addition, by documenting that this ability to sustain credit growth depends on macroprudential policies, we contribute to the literature on the importance of such policies to financial stability in general.

CRediT authorship contribution statement

Narjess Boubakri: Writing – original draft, Writing – review & editing. Ali Mirzaei: Conceptualization, Methodology, Software, Data curation, Investigation. Mohsen Saad: Writing – original draft, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Footnotes

¹By April 2020, the lockdown had already been established in as many as 80 countries.

²By the end of 2018, the Islamic financial services industry's Size had already surpassed the 2 trillion dollars to reach $2.19 trillion (Islamic Financial Services Board, 2019).

³CBs aim to make loans with low credit risk at a pre-determined interest rate. In comparison, IBs link lending to the purchase and subsequent selling of an underlying tangible asset to borrowers.

⁴Mudharabah is a participation contract between the bank and the borrower, such that profits are shared at a pre-determined ratio and losses are exclusively suffered by the bank, except for situations that involve misconduct, negligence, or breach of contract by the investor. Whereas, Musharakah requires the Islamic bank and the customer to contribute capital to a business adventure such that profits are shared according to the agreement and losses in proportion to capital contribution.

⁵Please refer to Galati and Moessner (2018) for a survey on macroprudential policies and their effectiveness.

⁶CAMELis anacronymfor capital adequacy, assets, management capability, earnings, and liquidity.

⁷Following the literature on IBs, we include only fully-fledged IBs, neglecting CBs with an Islamic window.

⁸Arguably, the financial sectors of Iraq, Palestine, and Syria, which experienced war in recent years, are unstable and do not function properly. Including bank data from these countries in our empirical analysis could bias our findings. To address this concern, we rerun the regressions in baseline Table 3 after dropping bank data from each country separately, any pairwise combination, and all three countries. Our preliminary results remain unchanged. For brevity, we only present the results in Appendix Table A3 after excluding all three countries.

Appendix A.

See Table A1, Table A2, Table A3

Table A1

Change in bank distance to default (DTD) and probability of default (PD) from pre-crisis to the crisis: Islamic vs conventional banks.

	IBs=43			CBs=145
	Pre-crisis 2019Q4	Crisis 2020Q3	Diff.	Pre-crisis 2019Q4	Crisis 2020Q3	Diff.
Risk indicator	(1)	(2)	(3)= (2)-(1)	(4)	(5)	(6)= (5)-(4)
DTD	2.391	1.896	−0.495	2.137	1.529	−0.608
PD (1month)	0.00025	0.00033	0.00008	0.00031	0.00043	0.00012
PD (3months)	0.00080	0.00102	0.00022	0.00100	0.00136	0.00036
PD (12months)	0.00371	0.00445	0.00074	0.00469	0.00593	0.00124

Open in a separate window

Table A2

Bank lending during the COVID-19 pandemic: Islamic banks vs conventional banks.

	g(L)	Δ(LTA)	Δ(LTGDP)	g(L)	Δ(LTA)	Δ(LTGDP)
	(1)	(2)	(3)	(4)	(5)	(6)
Crisis_t				−0.031***	−0.007***	−0.275***
				(-7.526)	(-3.880)	(-3.549)
IB_ic	−0.003	−0.004*	−0.074	−0.006	−0.005**	0.035
	(-0.346)	(-1.742)	(-0.544)	(-0.839)	(-2.318)	(0.279)
Crisis_t×IB_ic	0.031***	0.010***	0.604***	0.041***	0.012***	0.433***
	(2.790)	(2.697)	(3.208)	(3.443)	(3.267)	(2.652)
Size_ict	0.001	−0.000	0.180***	0.002	−0.000	0.184***
	(0.591)	(-0.870)	(5.510)	(0.854)	(-0.448)	(5.794)
C: Equity_ict	−0.007	0.004	0.067	−0.008	0.005	0.041
	(-0.261)	(0.414)	(0.179)	(-0.302)	(0.613)	(0.118)
A: LoanLoss_ict	−0.475*	−0.172**	−6.324*	−0.331	−0.241***	−2.549
	(-1.818)	(-2.143)	(-1.867)	(-1.295)	(-3.220)	(-1.102)
M: Cost_ict	−0.005	0.001	−0.057	0.004	−0.003	0.044
	(-0.493)	(0.178)	(-0.526)	(0.351)	(-0.727)	(0.474)
E: ROA_ict	0.090	−0.016	−0.315	0.206	−0.173***	1.276
	(0.558)	(-0.290)	(-0.119)	(1.232)	(-3.504)	(0.990)
L: Liquidity_ict	−0.003	−0.003	0.049	−0.004	−0.002	0.060
	(-0.386)	(-1.196)	(0.859)	(-0.503)	(-0.945)	(1.233)
Constant	0.005	0.012	−0.628	−0.018	0.011	−1.582*
	(0.128)	(0.676)	(-0.438)	(-0.481)	(0.849)	(-1.836)
Country*Year-Quarter FEs	Y	Y	Y	N	N	N
Country FEs	N	N	N	Y	Y	Y
# Countries	17	17	17	17	17	17
# Banks	461	461	461	461	461	461
N	2,104	2,104	2,104	2,104	2,104	2,104
Adj. R²	0.281	0.064	0.230	0.057	0.026	0.110

Open in a separate window

Table A3

Bank lending during the COVID-19 pandemic: Islamic banks vs conventional banks. The sample set excludes Iraq, Palestine, and Syria.

	g(L)	Δ(LTA)	Δ(LTGDP)
	(1)	(2)	(3)
Crisis_t×IB_ic	0.021*	0.013***	0.449**
	(1.915)	(2.809)	(2.098)
Size_ict	0.114***	−0.136***	0.715*
	(2.918)	(-7.002)	(1.656)
C: Equity_ict	0.029	0.178**	0.762
	(0.215)	(2.183)	(0.945)
A: LoanLoss_ict	−1.924*	−0.293	−2.384
	(-1.831)	(-1.008)	(-0.403)
M: Cost_ict	−0.004	−0.008	0.021
	(-0.148)	(-0.638)	(0.142)
E: ROA_ict	−0.466	−0.078	1.353
	(-0.774)	(-0.391)	(0.280)
L: Liquidity_ict	−0.009	−0.016*	−0.174
	(-0.663)	(-1.901)	(-1.210)
Constant	−1.963***	2.357***	−9.375
	(-2.889)	(6.897)	(-1.205)
Bank FEs	Y	Y	Y
Country*Year-Quarter FEs	N	N	N
# Countries	14	14	14
# Banks	412	412	412
N	1,968	1,968	1,968
Adj. R²	0.304	0.040	0.235

Open in a separate window

Data availability

Data will be made available on request.

References

Abedifar P., Molyneux P., Tarazi A. Risk in Islamic banking. Eur. Finan. Rev. 2013;17:2035–2096. [Google Scholar]
Acharya, V., and Steffen, S., 2020. ‘Stress Test’ for Banks as Liquidity Insurers in a time of COVID”, mimeo.
Acharya V., Naqvi H. The seeds of a crisis: a theory of bank liquidity and risk taking over the business cycle. J. Finan. Econ. 2012;106:349–366. [Google Scholar]
Alam Z., Alter A., Eiseman J., Gelos G., Kang H., Narita M., Nier E., Wang N. Digging deeper - evidence on the effects of macroprudential policies from a new database. Int. Monetary Fund Working Paper. 2019;19/66 [Google Scholar]
Altavilla C., Barbiero F., Boucinha M., Burlon L. The great lockdown: pandemic response policies and bank lending conditions. Eur. Central Bank Working Paper Series. 2020;No 2465 [Google Scholar]
Altunbas Y., Tommaso C.D., Thornton J. Do better-capitalized banks lend less? Evidence from European banks. Financ. Res. Lett. 2016;17:246–250. [Google Scholar]
Balvers R.J., Gu L., Huang D. Profitability, Value, and Stock Returns in Production- Based Asset Pricing without Frictions. J. Money Credit Bank. 2017;49:1621–1651. [Google Scholar]
Beck T., Demirgüç-Kunt A., Merrouche O. Islamic vs. conventional banking: business model, efficiency and stability. J. Bank. Financ. 2013;37:433–447. [Google Scholar]
Beltratti A., Stulz R.M. The credit crisis around the globe: why did some banks perform better? J. Financ. Econ. 2012;105:1–17. [Google Scholar]
Benetton M., Fantino D. Competition and the pass-through of unconventional monetary policy: evidence from TLTROs. Banca d’Italia Working Papers. 2018;No 1187 [Google Scholar]
Bilgin M.H., Danisman G.O., Demir E., Tarazi A. Bank credit in uncertain times: Islamic vs. conventional banks. Financ. Res. Lett. 2021;39:101563. [Google Scholar]
Boeckx J., de Sola Perea M., Peersman G. The transmission mechanism of credit support policies in the euro area. Eur. Econ. Rev. 2020;124(C) [Google Scholar]
Boubakri N., Mirzaei A., Samet A. National culture and bank performance: evidence from the recent financial crisis. J. Financ. Stab. 2017;29:36–56. [Google Scholar]
Brei M., Gambacorta L., von Peter G. Rescue packages and bank lending. J. Bank. Finan. 2013;37:490–505. [Google Scholar]
Brei M., Schclarek A. Public bank lending in times of crisis. J. Financ. Stab. 2013;9:820–830. [Google Scholar]
Bruno V., Shim I., Shin H.S. Comparative assessment of macroprudential policies. J. Financ. Stab. 2017;28:183–202. [Google Scholar]
Cerutti E., Claessens S., Laeven L. The use and effectiveness of macroprudential policies: new evidence. J. Financ. Stab. 2017;28:203–224. [Google Scholar]
Chen Y.-S., Chen Y., Lin C.-Y., Sharma Z. Is there a bright side to government banks? Evidence from the global financial crisis. J. Financ. Stab. 2016;26:128–143. [Google Scholar]
Claessens S., Ghosh S.R., Mihet R. Macro-prudential policies to mitigate financial system vulnerabilities. J. Int. Money Financ. 2013;39(C):153–185. [Google Scholar]
Cornett M.M., Mcnu*tt J.J., Strahan P.E., Tehranian H. Liquidity risk management and credit supply in the financial crisis. J. Financ. Econ. 2011;101:297–312. [Google Scholar]
Craig B.R., Dinger V. Deposit market competition, wholesale funding, and bank risk. J. Bank. Financ. 2013;37 3605–362. [Google Scholar]
Cull R., Martinez-Peria M.S. Bank ownership and lending patterns during the 2008–2009 financial crisis: evidence from Latin America and Eastern Europe. J. Bank. Financ. 2013;37:4861–4878. [Google Scholar]
Demirgüç-Kunt A., Detragiache E., Merrouche O. Bank capital: Lessons from the financial crisis. J. Money Credit Bank. 2013;45:1147–1164. [Google Scholar]
El Gamal M.A. Islamic Society of North America; Plainfield: 2000. A Basic Guide to Contemporary Islamic Banking and Finance. [Google Scholar]
Elsas R., Hackethal A., Holzhauser M. The Anatomy of Bank Diversification. J. Bank. Financ. 2010;34:1274–1287. [Google Scholar]
Farooq M., Zaheer S. Are Islamic banks more resilient during financial panics? Pac. Econ. Rev. 2015;20:101–124. [Google Scholar]
Galati G., Moessner R. What do we know about the effects of macroprudential policy? Economica. 2018;85:735–770. [Google Scholar]
García-Suaza A.F., Gómez-González J.E., Pabón A.M., Tenjo-Galarza F. The cyclical behavior of bank capital buffers in an emerging economy: size does matter? Econ. Model. 2012;29:1612–1617. [Google Scholar]
Gómez E., Murcia A., Lizarazo A., Mendoza J. Evaluating the impact of macroprudential policies on credit growth in Colombia. J. Financ. Intermed. 2020;42:10084. [Google Scholar]
Hale T., Webster S., Petherick A., Phillips T., Kira B. Oxford Covid-19 government response tracker. Blavatnik School of Government. 2020 https://www.bsg.ox.ac.uk/research/research-projects/oxford-covid-19-government-response-tracker [PubMed] [Google Scholar]
Hasan M., Dridi J. The effects of the global crisis on Islamic and Conventional Banks: a comparative study. J. Int. Econ. Commer. Policy. 2011;2:163–200. [Google Scholar]
Ibrahim M.H. Business cycle and bank lending procyclicality in a dual banking system. Econ. Model. 2016;55:127–134. [Google Scholar]
Ibrahim M.H., Rizvi S.A.R. Bank lending, deposits and risk-taking in times of crisis: a panel analysis of Islamic and conventional banks. Emerg. Mark. Rev. 2018;35:31–47. [Google Scholar]
Igan D., Mirzaei A. Does going tough on banks make the going get tough? Bank liquidity regulations, capital requirements, and sectoral activity. J. Econ. Behav. Organ. 2020;177:688–726. [Google Scholar]
Igan D., Mirzaei A., Tomoe M. Does macroprudential policy alleviate the adverse impact of COVID-19 on the resilience of banks? J. Bank. Financ. 2022 In press. [PMC free article] [PubMed] [Google Scholar]
Ippolito F., Peydrò J., Polo A., Sette E. Double bank runs and liquidity risk management. J. Financ. Econ. 2016;122:135. [Google Scholar]
Islamic Financial Services Board (IFSB), 2019. Islamic Financial Services Industry Stability Report, Kuala Lumpur, Malaysia: Islamic Financial Services Board.
Ivashina V., Scharfstein D. Bank lending during the financial crisis of 2008. J. Financ. Econ. 2010;97(3):319–338. [Google Scholar]
Kapan T., Minoiu C. Balance sheet strength and bank lending: Evidence from the global financial crisis. J. Bank. Financ. 2018;92:35–50. [Google Scholar]
Kosak M., Li S., Loncarski I., Marinc M. Quality of bank capital and bank lending behaviour during the global financial crisis. Int. Rev. Financ. Anal. 2015;37:168–183. [Google Scholar]
Li L., Strahan P.E., Zhang S. Banks as Lenders of First Resort: Evidence from the COVID-19 Crisis. Rev. Corp. Financ. Stud. 2020;9(3):472–500. [Google Scholar]
Lim C., Columba F., Costa A., Kongsamut P., Otani A., Saiyid M., Wezel T., Wu X. Macroprudential policy: what instruments and how to use them? Lessons from country experiences. Int. Monetary Fund Working Paper. 2011;11/238 [Google Scholar]
Mirzaei A., Saad M., Emrouznejad A. In Press; Annals of Operations Research: 2022. Bank Stock Performance during the COVID-19 Crisis: Does efficiency explain why Islamic banks fared relatively better? [PMC free article] [PubMed] [Google Scholar]
Repullo R., Saurina J. CEPR Discussion Papers; 2011. The countercyclical capital buffer of Basel III: A critical assessment. CEPR Discussion Papers 8304. [Google Scholar]
Rostagno M., Altavilla C., Carboni G., Lemke W., Motto R., Saint Guilhem A., Yiangou J. A tale of two decades: the ECB’s monetary policy at 20. Eur. Central Bank Working Papers. 2019;No 2346 [Google Scholar]
Stolz S., Wedow M. Banks’ regulatory capital buffer and the business cycle: evidence for Germany? J. Financ. Stability. 2011;7:98–110. [Google Scholar]
Zheng Y. Does bank opacity affect lending? J. Bank. Financ. 2020;119:105900. [Google Scholar]

Articles from Journal of International Financial Markets, Institutions and Money are provided here courtesy of Elsevier

FAQs

Bank lending during the COVID-19 pandemic: A comparison of Islamic and conventional banks? ›

Using a sample of 421 banks from 17 countries, we find that the lending growth of Islamic and conventional banks decreased during the initial phase of the COVID-19 crisis. However, the decrease is significant for conventional banks only.

Read On ›

What impact did COVID have on banks? ›

Issues the banking and capital markets industry might face:

We could also see a rise in payment delays due to cash-flow constraints on businesses arising from a temporary drop in demand for services. The hard part: telling the difference between a routine payment delay and significant deterioration in credit quality.

Discover More Details ›

Why did small business FinTech lending dry up during the COVID 19 crisis? ›

Why Did Small Business FinTech Lending Dry Up During the COVID-19 Crisis? FinTech small business lenders fund loans mostly through credit facilities and securitizations. This business model could make them financially constrained when a shock reduces the value of existing loans.

Why might the bank lending channel be less effective today than it once was? ›

The federal has decreased its influence over the banks in the form of the shift in legislation in response to the policy shift, making the banker routes less functional. The bank lending system has become powerless as a result of the international fall in the sector of bank loans.

See Details ›

Why did public banks lend more during the global financial crisis? ›

Using a novel bank-level dataset covering 25 emerging market economies, we show that public banks lent relatively more during the GFC because they pursued an objective of helping to stabilize the economy, rather than because they had superior fundamentals or access to public or depositors' funding.

Find Out More ›

How has the pandemic changed people's way of banking? ›

The pandemic accelerated a trend toward more digital banking and less reliance on in-person banking, though branches remain important for certain segments of the population, including many small-business owners.

Tell Me More ›

What has been one of the biggest changes in the banking industry due to COVID-19? ›

Summary. COVID-19 has accelerated digitization of banking as customer expectations change during the pandemic. This potential digital change, brought forward by COVID-19, will also help banks cope with the tougher operating environment the pandemic has brought.

Show Me More ›

Why do banks not lend to the poor? ›

Collateral is an asset that the borrower owns such as livestock, buildings, vehicles, and deposits with banks) and uses this as a guarantee to a lender until the loan is repaid. Usually, poor people or farmers may not have sufficient collateral to get loans from the banks.

Explore More ›

Why have banks been so reluctant to loan funds to businesses in recent years? ›

Banks are often cautious about lending to small businesses because they see them as risky or not profitable enough. They have strict rules and regulations to follow, and small businesses may not meet their criteria. Private credit investors are more willing to take risks and can be more flexible in who they lend to.

Why do banks have lending limits? ›

Lending limits protect the safety and soundness of national banks, promote diversification of loans, and help ensure equitable access to banking services. These limits prevent excessive loans to one person, or loans to related persons who are financially dependent.

Show Me More ›

Will there be a global recession in 2024? ›

Data for 2024 is a forecast. UN Trade and Development (UNCTAD) forecasts global economic growth to slow to 2.6% in 2024, just above the 2.5% threshold commonly associated with a recession. This marks the third consecutive year of growth below the pre-pandemic rate, which averaged 3.2% between 2015 and 2019.

Read The Full Story ›

Why did banks fail during the financial crisis? ›

The primary driver of commercial bank failures during the Great Recession was exposure to the real estate sector, not aggregate funding strains. The main “toxic” exposure was credit to non-household real estate borrowers, not traditional home mortgages or agency-issued MBS.

See Details ›

What caused many banks to run out of money when the market crashed? ›

Many of the small banks had lent large portions of their assets for stock market speculation and were virtually put out of business overnight when the market crashed.

Get More Info Here ›

How did COVID cause financial problems? ›

Many households and firms in emerging economies were already burdened with unsustainable debt levels prior to the crisis and struggled to service this debt once the pandemic and associated public health measures led to a sharp decline in income and business revenue.

How did COVID affect savings? ›

The COVID-19 pandemic has generated a sense of financial insecurity—even among the well-off. While the pandemic is a health crisis, the biggest motivator for saving more is “in case I have large, unexpected costs,” reinforcing the evidence that individuals now want to be prepared for unwelcome contingencies.

Have banks caught the corona effects of COVID on lending in the US? ›

Exploiting spatial and time variation, we find that banks geographically more exposed to lockdown measures experience an increase in loss provisions and non-performing loans. Exposures to the pandemic itself have a similar, but slightly weaker effect.

View Details ›

What has been the effect of the COVID-19 pandemic on capital spending? ›

In 2022, of those companies that were impacted by the coronavirus pandemic but had returned to normal level of operations in 2020, 2021 or 2022, 4.1 percent of companies canceled, 12.45 percent postponed, 11.65 percent decreased, and 2.8 increased some of their budgeted capital expenditures during the coronavirus ...