Is the interest rate pass-through to retail rates relatively complete for Islamic banks compared to conventional banks in Pakistan? Open Access

Islamic banking has been on the rise for the past three decades. Islamic banking assets grow by an average of roughly 15% every year and are a major factor in global finance (Khan, 2010). With the birth of the Islamic financial setup and its increasing popularity in both Muslim and non-Muslim countries (Majeed and Zainab, 2017), it is important to analyze how the interest channel works for Islamic banks compared with conventional banks. Theoretically, the operations and functions of Islamic banks remain within the domain of Islamic laws, also referred to as Sharia’h. for instance, Islamic laws are embedded in the work of Islamic banks (Yusof and Fahmy, 2008). Even during the global financial crisis of the last part of the 2000s, Islamic bank assets increased when global financial systems collapsed (Zaheer et al., 2013). The Islamic banking system remained immune to the global financial crisis, as Islamic banks do not engage in the sales of conventional debt instruments that involve interest payments (Smolo and Mirakhor, 2010). However, unlike the global financial crisis, Islamic banks remained vulnerable during the COVID-19 pandemic (Chazi et al., 2024).

Pakistan is one of the few countries in which both conventional and Islamic banks have operated in parallel for more than 20 years (Zaheer et al., 2013). The share of the Islamic banking industry in Pakistan in terms of assets of the overall banking industry is roughly 20% as of the end of March 2024 (State Bank of Pakistan, 2024a, 2024b). Currently, Pakistan has five fully fledged Islamic banks. These five banks have 2,671 branches across Pakistan. Some conventional banks have standalone Islamic branches. Pakistan has over 5,000 branches of Islamic banks. However, over 11,000 branches of conventional banks operate in Pakistan (State Bank of Pakistan, 2024a, 2024b). The central bank of Pakistan, the State Bank of Pakistan (SBP), plays a vital role in promoting Islamic banking in Pakistan. SBP has separately developed a regulatory framework based on Sharia’h and has been improving Sharia’h governance and compliance over the years (Rashid et al., 2017). Pakistan launched its first Islamic Sukuk bond in 2005 and was regularly issued in the International Islamic Bond market. Almost half of the companies listed in the equity market of Pakistan (the Pakistan Stock Exchange) adhere to the Shariah principles (Naz and Gulzar, 2023).

It is important to highlight that Islamic banks operate differently from conventional banks. Islamic banking is built on interest-free transactions and aims to avoid immoral and social practices (Butt et al., 2011). What makes it different from a conventional bank is that it relies on tangible assets and real services and implements a profit-and-loss sharing model, and thus, risk-sharing, on both the liability and asset sides (Beck et al., 2013). The profit-sharing model keeps the Islamic banking system less risky than an interest banking system. Nonetheless, similar to commercial banks, Islamic banks also use inter-bank offer rates, such as London inter-bank offer rates (LIBOR) and Karachi inter-bank offer rates (KIBOR), as benchmarks for setting profit rates on their financing products. Hence, changes in these benchmark rates can lead to adjustments in Islamic banks’ profit rates. Some studies reveal that the actual (practical) Islamic banking system may not follow the theory of Islamic laws in the true spirit, and there is room for strengthening the existing system (Hassan and Aliyu, 2018). Nonetheless, market competition is key to having a positive impact on Islamic banks’ social performance of Islamic banks (Meskovic et al., 2024).

The interest rate channel of the monetary policy transmission mechanism is the channel through which changes in the central bank’s policy rate impact the marginal cost of borrowing (for conventional banks) and profit rates for Islamic banks, leading to changes in investment and savings, and consequently, the output. Changes in interest rates affect households (savings), firms (cost structures) and banks (profitability). Higher real returns may encourage households to save money. Similarly, a lower interest rate may encourage a household to take a loan for a car, house, etc. From a firm’s perspective, a change in the interest rate influences investment costs. A higher interest rate increases the overall costs for the firm, and the firm may pass this increase by charging a higher price, which may impact households. Similarly, bank profitability depends on changes in the interest rate (Kohlscheen et al., 2018). If the spreads (the difference between deposit and lending rates charged by banks) increase, then the profitability of a bank increases.

The interest rate channel is relevant for Pakistan, as the operational target of the SBP existing monetary policy framework is to maintain an inter-bank overnight repo rate close to the SBP Policy (target). In one of the relevant studies related to Pakistan, Omer (2019) emphasizes that the interest rate channel is the appropriate channel for Pakistan to analyze the effectiveness of monetary policy transmission. The author takes into consideration the excess liquidity available with Islamic and conventional banks to assess the effectiveness. Few studies estimate the pass-through of the policy rate on the retail rates of overall banking industry in Pakistan. Although different sample periods and methodologies are used, the results are consistent with the pass-through of the policy rate to lending rates being relatively complete compared to the pass-through of the policy rate to deposit rates.

Previous research related to Pakistan has focused only on the magnitude and speed of pass-through in its banking industry. This study compares the pass-throughs of Islamic and conventional banks. An autoregressive distributed lag (ARDL) model is used to estimate the short- and long-run coefficients of pass-through for both Islamic and conventional banks. To test for the existence of a long-run relationship, a “bounds test” was estimated (Pesaran et al., 2001). None of the previous studies conducted in Pakistan included important control variables. This study incorporates important control variables such as real GDP growth, the exchange rate, liquidity positions of the banks, credit risks (non-performing loans), competition within the sector (the Herfindahl–Hirschman Index), the post-covid period and government borrowing.

The remainder of this paper is structured as follows: Following the introduction, the status of Islamic Banks in Pakistan is discussed in Section 2. The theoretical background, literature review, and research gaps are explained in Section 3. In Section 4, we discuss the proposed methodology. Section 5 contains information on the data description and the choice of variables. Section 6 presents our empirical results. Finally, Section 7 concludes the study.

The SBP defines Islamic banking as a system in consonance with the spirit, ethics and value system of Islam. Islamic banking should be administered in line with the principles laid down by the Islamic Shariah. In general, Islamic banking implies avoiding not only interest-based transactions but also unethical and unsocial practices that are against the welfare of society. Unlike conventional banking, Islamic banking transactions are based on tangible assets and real services. The goal of the Islamic banking system is to achieve economic prosperity and ensure people’s welfare.

The need for an Islamic banking system was emphasized as soon as Pakistan became independent in 1947. The founding father of Pakistan, Quaid-e-Azam Muhammad Ali Jinnah, at the inauguration ceremony of the central bank of Pakistan, said, “I shall watch with keenness the work of your Research Organization in evolving banking practices compatible with Islamic ideas of social and economic life […]. The adoption of Western economic theory and practice will not help us to achieve our goal of creating happy and contented people. We must work our destiny in our own way and present an economic system based on the true Islamic concepts of equality of manhood and social justice to the world. We will thereby be fulfilling our mission as Muslims and giving to humanity the message of peace which alone can save and secure the welfare, happiness, and prosperity of mankind.” (Islamic Banking Department, 2024).

One of the visions of the SBP is to make Islamic banking the first choice for providers and users of financial services in Pakistan. As a first step, the SBP established a new department, the Islamic Banking Department (IBD), to develop and strengthen the Islamic banking sector in 2001 (State Bank of Pakistan and Development for International Development, 2014). The Islamic Banking Department of the SBP has been making efforts to promote and develop the Islamic Banking industry in line with best international practices, ensuring Shariah compliance and transparency (Islamic Banking Department, 2008).

In 2001, the SBP-issued policy was exclusively tailored to Islamic banking with detailed criteria for establishing Islamic banks. In 2002, a new clause was added to the Banking Company Ordinance (BCO) 1962, which allowed commercial banks to open subsidiaries for Shari’ a-compliant operations. In the same year (2002), the first fully functional Islamic bank began operations. The SBP continues to play an instrumental role in providing Islamic banks with a platform to explore sectors such as agriculture, small and medium enterprises (SME), housing, and micro finance. The SBP also facilitates Islamic banks in raising awareness and building the capacity of the industry. The SBP, in collaboration with many national and international stakeholders on a regular basis, conducts seminars, conferences and training programs (State Bank of Pakistan and Development for International Development, 2014). In 2007, the SBP introduced guidelines for establishing an Islamic Microfinance Business, setting the stage for the development of this sector (Islamic Banking Department, 2024).

With the emergence of Islamic banks, the unavailability of Shari’a-compliant investment avenues has become a major problem. To address this issue, the government of Pakistan launched the Shariah-compliant Government securities program in 2008–2009. Since then, the government has carried out noticeable issuances through these securities (Islamic bonds), with the objective of increasing the share of Shariah-compliant securities to cater to the excess liquidity available with Islamic banks. As of March 2024, Islamic banks have the option to invest in one, three and five years of Ijara Sukuk instruments (Ministry of Finance, 2024).

Currently (at the end of March 2024), the share of Islamic banks in the total assets of Pakistan’s banking industry is just above 20% (State Bank of Pakistan, 2024a, 2024b). Of the 33 banks operating in Pakistan, five are full-fledge Islamic banks. These five banks have 2,671 branches across Pakistan. Some conventional banks have standalone Islamic branches. Pakistan has over 5,000 branches of Islamic banks. However, over 11,000 branches of conventional banks operate in Pakistan, and Islamic banks’ asset quality is also better than that of conventional banks. For instance, gross non-performing financing for overall financing is just over 3% for Islamic banks, whereas it is more than 7% for the banking industry. Similarly, in terms of profitability, the return on asset ratio for Islamic banks is 5.4 compared to 2.9 for the overall banking industry (Islamic Banking Department, 2024).

Our analysis of Islamic banks is restricted to the five full-fledged Islamic banks operating in Pakistan during the sample period. Islamic windows operated by conventional banks are not included in the Islamic bank category due to data unavailability at micro level. It is also important to highlight that in Pakistan, there is a heterogenous ownership structure of banks. Five banks are owned by the federal and provincial governments, over 20 banks are owned by private businesses, and a couple of banks are specialized for lending purposes to the agricultural sector. In contrast, Islamic banks tend to have more private or foreign institutional ownership, which may contribute to differences in managerial autonomy, market discipline and strategic agility. While we do not formally model these effects due to data constraints, we acknowledge that ownership heterogeneity may partly explain variations in rate-setting behavior and responsiveness to policy signals.

Islamic banks in Pakistan predominantly rely on fixed-return, benchmark-linked instruments such as murābaḥah and ijārah, which are commonly indexed to KIBOR (Karachi Interbank Offered Rate). This structural linkage to a market-based benchmark facilitates a quicker and more predictable adjustment in financing rates when the policy rate changes. In contrast, conventional banks may offer a wider variety of lending products, including long-term fixed-rate loans or loans with delayed repricing, which can dampen the speed and magnitude of pass-through.

Islamic banks in Pakistan typically operate on a smaller scale compared to their conventional counterparts. This smaller size may allow them to adjust their balance sheet composition and pricing more swiftly in response to changes in the policy rate. Conventional banks, with more diversified and complex asset-liability structures, may experience institutional inertia that slows their responsiveness.

Islamic banks in Pakistan cater to a niche but rapidly growing market segment. This may subject them to stronger competitive pressures, particularly in deposit mobilization and Shariah-compliant lending. To maintain competitiveness and client loyalty, these banks may be more proactive in adjusting their rates in line with market signals and policy changes.

Given the nature of Shari’a-compliant contracts, where profit-and-loss sharing and asset-backing are common features, Islamic banks may prioritize close alignment with policy benchmarks to manage liquidity and credit risk more effectively. A delayed response to monetary policy changes may expose these banks to higher risks, especially given their lower tolerance for speculative or interest-based instruments.

Studies on the pass-through of interest rates are largely based on the industrial organization theory. However, there are some differences due to the unique nature of banking operations and the market structure. Banks rely largely on short-term deposits to support long-term lending. Changes in interest rates can affect both the cost of funding for banks (e.g. interest paid on deposits) and the interest rate banks charge on loans. Industrial firms may rely on a mix of sources of financing, such as equity, bonds and bank loans, but they typically do not have the same deposit-based funding structure as banks (Berger and Bouwman, 2009). Banks often operate in a more concentrated market with fewer competitors than firms in other industries. Market concentration influences the degree of pass-through of interest rate changes (Wang et al., 2022). In less competitive banking markets, banks may have more discretion in adjusting their lending rates in response to changes in interest rates, leading to less pass-through for consumers. In contrast, firms in industrial organizations may face more competitive pressures in general, which could lead to a more immediate and complete pass-through of interest rate changes to consumers. In addition, banks face extensive regulations and supervision from central banks and financial authorities, which can affect their pricing decisions and the pass-through of interest rate changes. Regulatory requirements such as capital adequacy ratios and liquidity requirements may influence how banks adjust their lending rates in response to changes in interest rates.

The share of Islamic banking not only in Muslim-dominated countries but also around the global is on a rise. The perception of Islamic banking, which is transparent in transactions, operates based on Islamic laws and principles, follows the ethical behaviour of employees and has low service charges, making it popular in Pakistan (Quresh et al., 2012. The main difference between Islamic and conventional banking is that financing in Islamic banking is asset-backed, ensuring that contracts and transactions are associated with identifiable and evident assets (Majeed and Zainab, 2017). The Islamic banking system emphasizes on corporate social responsibility and urges easy access to interest-free loans (Farook, 2007). In the Islamic banking system, profit distribution is based on a pre-agreed upon ratio. Most deals in Islamic banking systems are based on mudarabah profit-sharing and musharakah-joint ventures (Aziz and Afaq, 2018). Some studies argue that there is no major difference between Islamic banking and conventional banking. Islamic banking in terms of its practices is similar to modern conventional banking with certain conditions imposed by Shari’a; hence, the perception that Islamic banks behave differently is not appropriate (Hanif, 2010) . The Islamic banking system does not necessarily follow the profit- and loss-sharing model (Chong and Liu, 2009). Islamic finance practice is not in line with this theory (Siddiqi, 2006). Gulzar and Masih (2015) emphasized that Islamic porift rates are aligned if not determined by conventional rates. They further studied the behavior of banks in Malaysia and concluded that Islamic banks (in Malaysia) would keep the deposit rates close to conventional rates to keep their deposit base intact particularly if banks rely mainly on deposits for funding.

For many countries where a dual system of conventional and Islamic banking exists, studies compare the performance of Islamic banks vis-à-vis conventional banks. Khan (2010) argue that even after three decades, the functions of Islamic banking are similar to that of conventional banking. Kassim et al. (2009) estimate the impact of monetary policy shocks on conventional and Islamic banks and conclude that conventional banks are less sensitive to monetary policy changes compared to Islamic banks. Different impact on conventional vis-à-vis Islamic banks carries implication for the risk management. A study on Malaysia by Gulzar and Masih (2015) concluded that even after four decades, retail rates offered by Islamic banks are aligned with rates of conventional banks. They further argued in their study that even most of the products offered by Islamic banks in Malaysia are similar to that of conventional banks. Nonethless, the authors acknowledged that the central bank of Malaysia is a pioneer in promoting the growth of Islamic finance industry and is reinforcing the practice of Profit and Loss (PLS) contracts in true spirits. Farook et al. (2012) concluded in their study that most Islamic banks do not follow asset returns in spirit and distribute their profits in line with market-based interest rates. Cevik and Charap (2015) analyzed and compared the pattern of deposit rates offered by conventional banks and the rate of return contracted by Islamic banks for Malaysia and Turkey. Their main findings on the basis of cointegration analysis and Granger causality confirmed a long-run relationship between the conventional bank deposit rates and the PLS returns. Similarly, one study explored the dynamics of liquidity holdings in Islamic and conventional banks (Abdo et al., 2023). In Indonesia, Islamic financial development helped promote economic growth and capital accumulation (Abduh and Omar, 2012).

In Pakistan, a limited number of studies estimate the pass-through of the policy rate to the retail rates of banks. However, previous research related to Pakistan, for instance, Hussain and Khan (2016) and Khan and Hanif (2012), focus only on the magnitude and speed of the pass-through of the overall banking system or in terms of ownership of the bank, that is, private, public or foreign banks; very little or no importance has been given to estimating the pass-through for conventional and Islamic banks separately. Most studies use the ARDL framework to estimate the pass-through of the policy rate to retail rates in Pakistan. Khan and Hanif (2012) and Hussain and Khan (2016) use an ARDL framework to analyze cointegration and estimate the speed and magnitude of the pass-through of the SBP policy rate to market interest rates.

Previous research related to Pakistan has focused only on the magnitude and speed of pass-through in its banking industry. This study compares the pass-throughs of Islamic and conventional banks. An autoregressive distributed lag (ARDL) model is used to estimate the short- and long-run coefficients of pass-through for both Islamic and conventional banks. To test for the existence of a long-run relationship, a “bounds test” suggested by Pesaran et al. (2001) is estimated. None of the previous studies conducted in Pakistan included important control variables. This study incorporates important control variables, described in Section 5.

The ARDL model is a widely used framework in most studies in this field. These models have a wide range of possible applications (Kripfganz and Schneider, 2023). In this model, the dependent variable is a function of its lagged values, and the current and lagged values of the explanatory variables. The ARDL model can accommodate flexible lag structures, is applicable to small samples and is useful for panel data. ARDL can estimate the relationships between both levels and differences. In addition, this technique provides estimates of short- and long-run cointegrating relationships even if the variables are not integrated in the same order. It can accommodate a mixture of stationary and nonstationary variables I(0) and I(1), even without the need to pretest the order of integration. Furthermore, both the short- and long-run coefficients can be consistently estimated in a single step, and the estimator’s asymptotic normality eases the statistical inference (Kripfganz and Schneider, 2023). To test for the existence of a long-run relationship, Pesaran et al. (2001) suggest a ‘bounds test” that compares the values of conventional F- and t- statistics to pairs of critical values. Within these bounds, the test is inconclusive; outside these bounds, the test does not reject the null hypothesis.

The ARDL model also provides estimates of long-run parameters even in the presence of weak endogeneity (Cho et al., 2021) ^[1]. Conversely, if the data contains a unit root and there is no co-integration, then the ARDL estimates fail to identify the correct relationship between the variables. In this case, variables with unit roots must first be differenced, and modelled in this form to avoid spurious regressions.

The ARDL (p,q) model used to estimate the long-run relationship between the dependent and explanatory variables is as follows:

where: rr = retail rate offered by the bank (dependent variable), pr = policy rate announced by the central bank (explanatory variable), z_t = set of L control variables, μ_t is is an error term.

Akaike Information Criterion (AIC) was used to obtain the optimal lag orders for p and q. A maximum lag length of 13 was used for all specifications. AIC is preferred over Bayesian Information Criterion (BIC) because it tends to select less restrictive models (Kripfganz and Schneider, 2023). It is important to highlight that lag selection under AIC, BIC or 13 lags (selected arbitrarily) yields similar results.

Two retail rates are used in this study.

1. weighted Average Lending Rate on Gross Disbursements; and

2. weighted Average Lending Rate on New (fresh) deposits.

Two proxies for the policy rate were used in this study. Six-Month Karachi Inter-bank Offer Rate (KIBOR); and Market Treasury Bills’ weighted average Yield for three months.

The final equation for estimation is as follows:

In (2), the ARDL short-run coefficients are $ψyi ψxi; -π1$ = α, which is the error correction coefficient (speed-of-adjustment coefficient), and the long-run coefficient is -$π2/π1$⁠. The error correction coefficient shows the speed of adjustment of the retail rate in response to a change in policy rate.

Model (2) is run separately for the five Islamic banks and the 28 conventional banks currently operating in Pakistan.

Monthly time series data on the six months Treasury Bill rate, six-months Karachi Interbank Offer Rate used as a proxy of policy rate and weighted average lending rate (WALR), and weighted average deposit rate (WADR) of individual conventional and Islamic banks from January 2005 to December 2023 were used. The main data source is the SBP.

• Weighted average lending rate (WALR)

This is the rate of loans and advances recoverable by borrowers at the end of a month. Loans and advances include zero markup and interbank placements. The data are available for four different types of banks: public, private, foreigns and specialized.

Weighted Average Lending Rate = Sum (rate charged × amount disbursed)/sum (amount disbursed):

• Weighted average deposit rate (WADR)

This is the rate offered for fresh deposits. Fresh deposits include outstanding positions of fresh deposits (new accounts) mobilized during the month and deposits repriced or rolled over during the month. Deposits also include zero markups and interbank placements.

Weighted Average Deposit Rate = Sum (rate of deposit × amount of deposit)/sum (amount of all deposits).

The six-months Karachi Inter Bank Offer Rate (KIBOR) is used as a proxy for the policy rate. Six-months KIBOR is the interest rate benchmark at which a conventional bank is ready to lend Rupees to another bank for six months without any collateral. The data are available on a daily basis. The monthly averages were used in this study lending rates of banks are linked to a benchmark rate known as the Karachi Inter-bank Offered Rate (KIBOR), which was published by the Financial Market Association of Pakistan. The SBP-mandated banks benchmark their lending to the corporate sector to the KIBOR, so that the pricing mechanism is transparent.

• Real GDP growth – Year on Year (YoY) growth rate (interpolated) ^[2].

• Monthly Exchange Rate – (YoY) growth rate.

• Monthly Government Borrowing from the commercial banks – YoY growth rate.

• YoY growth rate of stock market.

• Post-Covid period.

• Liquidity position (available funds) – monthly net injections.

• Herfindahl-Hirschman Index (HHI) (to capture competition).

• Non-performing loans (NPLs) as percent of total loans (credit risk).

Before going for estimations, augmented Dickey–Fuller (ADF) and GLS-transformed Dickey–Fuller (DF-GLS) tests were applied to check for the presence of a unit root in the series of dependent and explanatory variables. Testing for a unit root is crucial for avoiding spurious regressions. The null hypothesis is that the variable contains a unit root, in contrast to the hypothesis that the variable is generated by a stationary process (possibly a trend stationary). This trend is included in the unit root tests at levels; however, at the first difference, the time trend is not included. All variables in the two data sets (pre- and post-pandemic periods) are I (1), integrated of order 1 based on the ADF and DF-GLS tests. There is no evidence of I (2); therefore, the ARDL model can be run.

As mentioned in the Data Description section, there are two types of deposit and lending rates, fresh and outstanding. Both rates were used in this study for the analysis and robustness. However, in the main text, weighted average rates of fresh loans and deposits mobilized during a specified period (i.e. a month) are used. Using fresh rates makes more sense because the objective of this study is to assess the current impact of the change in the policy rate on bank retail rates. Similar to retail rates, this study used different proxies for policy rates. In the main text, the six-month Karachi Inter Bank Offer Rate (KIBOR) is used; however, other proxies for the policy rate are used to check the robustness of the main results.

The codes used in the subsequent results are listed in Table 1. The results of the autoregressive distributed lag model for the overall banking sector are provided in Tables 3 and 4. In all specifications, the optimal lag selection is based on the AIC. For the robustness check, 13 lags were selected arbitrarily in each specification as monthly data were employed in this study. It is important to highlight that lag selection under AIC, BIC or 13 lags (selected arbitrarily) yields similar results.

The pass-through of the six-month KIBOR (proxy of the policy rate) to the weighted average lending rate on fresh loans in the overall banking industry is close to one in all specifications (Table 2). The inclusion of a constant and trend did not significantly change the results of the main coefficient. However, inclusion of trend and a constant impact the speed of adjustment and have implications for some of the control variables. The speed of adjustment improves from −0.28 to −0.41, and the exchange rate and government borrowing variables become statistically significant in the extended specification.

A 100-basis points (bps) change in the 6-month KIBOR leads to a 93-bps change in the lending rate in the long run (specification 4). The lower and upper bounds of the 95% confidence interval are 0.80 and 0.97, respectively. The coefficient of the speed of adjustment (-$π1$⁠) is statistically significant and shows that 39% of the pass-through adjustment is completed within the first month of change in the six-month KIBOR rate. This implies that interest rate adjustment is completed within a quarter. The coefficient of real GDP turns out to be statistically insignificant, reflecting that fluctuations in economic activity do not impede or provide impetus for pass-through adjustment. The coefficients of the exchange rate, liquidity conditions, credit risk and government borrowing are significant at the 5% level. The coefficient of the HHI is also insignificant as the competition (the market share) within banks has not changed significantly over the last two decades. Although the share of Islamic banks are on a consistent rise, the coefficient suggests that the impact on the interest rate pass-through is not significant for the overall banking industry.

Several post-estimation tests were conducted for the main model to test for structural breaks, serial correlation, heteroscedasticity and coefficient stability. Breusch and Pagan’s (1980) LM test for autocorrelation showed no serial correlations. To check for coefficient stability, a cumulative sum (CUMSUM) test was conducted ^[3]. The result of the decomposition of the IM test (Cameron and Trivedi, 1990) indicates that heteroscedasticity is not present, and the main estimation procedure does not require adjustments. Structural break testing using the “estat sbsingle” STATA command was also implemented to test for structural breaks. These results do not indicate any structural breaks.

Table 3 provides the results for the pass-through of the KIBOR to the deposit rates. A 100-bps change in the six-month KIBOR leads to a 58-bps change in the deposit rate in the long run, indicating only a partial pass-through. The lower and upper bounds of the 95% confidence interval are 0.43 and 0.63, respectively. The coefficient of speed of adjustment is statistically significant and shows that 50% of the adjustment is completed within the first month of the change in the six-month KIBOR rate. As in the case of lending rates, the coefficient of the exchange rate is significant, but the size of the coefficient estimate indicates that its impact on pass-through is negligible. Most of the control variables are insignificant in the case of deposit rates. The coefficient of post-covid period is statistically significant implying that the pass-through to deposit rates has improved post-Covid. There is no evidence of a structural break as even dropping this variable (Post-covid), the coefficient of the main independent variables does not change significantly. For robustness purpose, the Chow test (a statistical method) is also employed to detect whether there’s a significant change in the coefficients of a regression model across two different periods (pre and post Covid). The result does not validate any structural break point.

Before separately providing the results of the ARDL model for Islamic and conventional banks, it is important to present the results of the bounds test. If the Bounds test reveals no long-run relationship, the results of the ARDL model are not valid. Tables 4 and 5 provide the results of the bounds test. The Bounds Test involves comparing the values of F- and t-statistics to pairs of critical values (details are in the methodology section). The test results were inconclusive within these bounds. If the values are outside these bounds, the test either rejects or does not reject the null hypothesis.

Table 6 shows that a 100 bps rise in KIBOR leads to an increase of 100 bps in the lending rate of the Islamic banks reflecting complete pass-through in all specifications. The lower and upper bounds of the 95% confidence interval are 0.76 and 1.30, respectively. The coefficient of speed of adjustment shows that around 28% of the adjustment is completed within a month. The profit rates (lending rates) are adjusted completely by Islamic banks compared to relative incomplete adjustment of the overall banking sector.

Table 7 shows that a 100-bps rise in KIBOR leads to an increase of 90 bps in the deposit rate of Islamic Banks. The lower and upper bounds of the 95% confidence interval are 0.40 and 1.28, respectively. This is a relatively high pass-through compared to conventional banks. As Islamic banks aim for the welfare of society, apparently adjustment of the deposit rate is in line with the changes in the policy rate.

A 100-bps increase in KIBOR leads to an increase of 90 bps in the lending rate of conventional banks, reflecting an incomplete pass-through compared with Islamic banks. The coefficient of speed of adjustment shows that approximately 27% of the adjustments were completed within a month (Table 8).

Table 9 shows that a 100-bps increase in KIBOR leads to an increase of 58 bps in the deposit rate of conventional banks. This is again relatively incomplete compared to Islamic banks. To check the robustness of the main results (KIBOR and fresh retail rates), different retail rates and proxies for policy rates were used. Different retail rates and different proxies for the policy rate yield similar results.

The above results show that the pass-through of policy rates to retail rates varies between conventional and Islamic banks. However, the pass-through to lending rates are higher and more complete relative to deposit rates across both bank types (Figures 1 and 2). One possible reason why pass-through to lending rates is higher than deposit rates is that corporate loans are pegged to the KIBOR. The pegging of lending rates to the KIBOR (quoting the rate as KIBOR plus) implies that the lending rate for conventional banks and profit rates for Islamic banks should increase or decrease in line with the changes in the KIBOR. However, the deposit rates are not pegged at any rate. Most deposits held by commercial banks are current account deposits (i.e. zero rate of return). This also enables banks to avoid completely adjusting the deposit rate because most of their deposit bases are insensitive to interest rate changes. In an economy like Pakistan, pass-through to lending rates are incomplete, mainly because banks prefer to earn higher returns by placing their funds in (almost) no-risk sovereign bonds and securities. These results are consistent with those of Gregor et al. (2021). Their meta-analysis of interest rate pass-through was based on 50 studies and more than thousand estimates. They conclude that the average pass-through of a 100-bps change in the policy rate to lending rates is approximately 80 bps.

A Wald test was conducted on the estimated long-run coefficients of KIBOR to determine whether the pass-through of policy rates to retail rates was complete. The null hypothesis suggests that a coefficient of 1 indicates complete pass-through (Fazal and Salam, 2013). The test results, presented in Table 10, reveal that the pass-through-to-deposit rates are incomplete for the overall banking sector and conventional banks at 5% level of significance. However, for Islamic banks, the test does not reject the possibility of a complete pass-through for both lending and deposit rates, indicating that the coefficient is statistically equal to one, signifying a complete pass-through at 5% level of significance. However, given the wide confidence interval and a lower R-squared, the point estimate for Islamic banks maybe a little imprecise.

One important limitation of our analysis is the potential aggregation bias arising from treating Islamic banks as a homogeneous group. While our results show that Islamic banks, as a sector, exhibit a stronger and faster pass-through of policy rate changes compared to conventional banks, this finding is based on aggregated data from the five fully fledged Islamic banks operating in Pakistan. In practice, these institutions may differ in important respects – including Shariah board interpretations, product offerings (e.g. degree of reliance on murābaḥah vs mushārakah), customer segments and institutional maturity – which could influence their individual responsiveness to monetary policy.

Unfortunately, bank-level disaggregated data on Islamic banks’ lending and deposit rates are not publicly available on consistent basis, which constrains our ability to model within-group heterogeneity. As a result, our estimates should be interpreted as reflecting average behavior across the Islamic banking sector, rather than as representative of all individual institutions.

We suggest that future research, using more granular data – ideally at the bank or even contract level – could provide deeper insights into how heterogeneity among Islamic banks affects the monetary policy transmission mechanism. This would help identify whether specific bank characteristics (e.g. size, risk management practices, funding structure or adherence to different Shariah schools) systematically influence their pass-through behavior.

In a hybrid banking system, different levels of pass-through of policy rates to retail rates and varying speed of adjustment towards the equilibrium for Islamic and conventional banks highlight important monetary policy, regulatory, and financial stability implications. The findings that Islamic banks exhibit relatively different pass-through to retail lending and deposit rates than their conventional counterparts suggest the following key implications:

• The asymmetry in rate pass-through implies that an identical change in the central bank’s policy rate may not transmit equally across the banking system. Since Islamic banks adjust differently (more completely) to policy changes, they could react more strongly, causing uneven credit supply and savings incentives across bank types. To tackle this, the SBP may need to differentiate or supplement its monetary policy instruments to ensure effective transmission across the entire financial system. This could include further and proactive work dual liquidity windows or differentiated reserve requirements that account for structural differences between Islamic and conventional banks.

• The existing regulatory frameworks should explicitly acknowledge that the structure of Islamic banks’ risk profiles and income is different from those of conventional banks. Their reliance on the PLS method implies Islamic banks are structurally more risk-averse and less interest rate-sensitive in theory, though market-linked benchmarks (like KIBOR) are commonly used. In this context, the needs to fine-tune supervisory tools to reflect these operational realities. This could mean developing more sophisticated separate stress-testing frameworks, capital adequacy guidelines and liquidity measures for Islamic banks.

• The better pass-through of rates in Islamic banks can enhance public trust in their responsiveness and transparency, potentially boosting deposit mobilization and lending activities, particularly among religiously motivated populations. To further encourage financial inclusion, especially in underserved and religiously observant segments, SBP and policymakers should enhance awareness and financial literacy campaigns emphasizing the stability and efficiency of Islamic banks.

• Despite operating under Shariah principles, Islamic banks in Pakistan still often peg pricing to conventional benchmarks like KIBOR. The more effective pass-through indicates they are not insulated from interest rate movements, raising conceptual and operational questions about authenticity and consistency. Countries like Pakistan can take the lead in developing Shariah-compliant Islamic benchmarks (e.g. Islamic Interbank Benchmark Rate) to reduce reliance on interest-based indicators, thereby aligning operations more closely with Islamic principles.

• As both a regulator and a promoter of Islamic finance, SBP must play a balancing role – ensuring monetary policy effectiveness while also fostering the unique attributes of Islamic finance.

• This study underscores the lack of empirical research on monetary policy transmission in Islamic banking. Evidence-based policymaking requires regular analysis of pass-through behavior, bank-level responses and consumer sensitivity.

Islamic banking has witnessed significant growth over the past three decades, with assets expanding by approximately 15% annually, making it a key player in global finance. The resilience of Islamic banks, especially during the global financial crisis of the late 2000s, highlights the distinct nature of their operations based on Islamic law. In contrast to conventional banks, Islamic banks avoid interest-based transactions and emphasize profit-and-loss sharing, which mitigate risk. In Pakistan, Islamic banking has been operating alongside conventional banking for over two decades, with Islamic banks now holding a 20% share of the total banking assets as of March 2024. The SBP plays an important role in promoting and regulating the Islamic banking sector. Despite the growing prominence of Islamic banks, there is limited research on how the interest rate channel, a critical aspect of monetary policy transmission, functions differently for Islamic banks than conventional banks.

This study addresses this gap by comparing the pass-through of the policy rate with the retail rates of Islamic and conventional banks in Pakistan. Using an ARDL model, this study estimates both short- and long-run effects by incorporating important control variables, such as real GDP growth, exchange rate, and government borrowing. The findings reveal that Islamic banks pass-through to lending rates are near-complete, whereas they are incomplete for conventional banks. Additionally, the pass-through to deposit rates of Islamic banks are nearly double those of conventional banks, indicating that Islamic banks are more effective in adjusting the pass-through and therefore exhibit a more efficient monetary policy transmission mechanism. The findings of this study contribute to the literature by highlighting the differences between how Islamic and conventional banks in Pakistan respond to changes in monetary policy, offering insights for policymakers and stakeholders in the financial sector.

Islamic banks demonstrate a more complete and responsive adjustment to policy changes, which could lead to divergent credit supply and savings behavior. To ensure effective monetary transmission, the SBP may need to adapt its policy tools – such as dual liquidity windows or differentiated reserve requirements – while also refining regulatory frameworks to account for the unique risk structures and operational models of Islamic banks. The stronger pass-through observed in Islamic banks can enhance public trust and financial inclusion, particularly among religiously observant populations, but also raises concerns about the continued reliance on conventional benchmarks like KIBOR. This points to the need for developing Shariah-compliant reference rates. As both a regulator and promoter of Islamic finance, the SBP must strike a balance between ensuring macroeconomic stability and nurturing the distinct principles of Islamic banking. Finally, the study underscores the urgent need for more empirical research to support evidence-based policymaking in this dual banking landscape.

The author would like to thank two anonymous referees whose constructive comments have significantly improved the quality of this draft. The author is also indebted to Prof. Alfred Haug and Prof. Dorian Owen for their valuable feedback on various sections of this study.

This study did not require ethical approval because it did not involve human participants and was based on a secondary data analysis.

I hereby declare that the ideas, concepts and research presented in this work are entirely my own. I used AI tools solely for purposes of enhancing articulation, grammar and rephrasing, ensuring clarity and coherence. The original intellectual content, analysis and findings remain the product of my independent effort.