This paper analyses the role of Sukuk issuance in driving economic growth amid the Global Financial Crisis (2007–2009), considering both short-term fluctuations and long-term outcomes during periods of financial distress.
Using the Pooled Mean Group (PMG) approach of the ARDL model, the study analyses panel data from major Sukuk-issuing countries over the period ranging from 2001 to 2022. The analysis includes four sample periods: Pre-GFC, During-GFC, Post-GFC and the Full sample. Key variables include GDP growth and Sukuk issuance, with controls for Gross Fixed Capital Formation, General Government Final Expenditure, Inflation, Trade Openness, the Financial Development Index and a GFC dummy.
The results show that Sukuk issuance positively impacts long-term economic growth in the Pre-GFC, Post-GFC and full sample periods but has a negative impact during the GFC. Short-term relationships are insignificant. Panel causality analysis reveals a unidirectional link from GDP growth to Sukuk issuance in the Pre-GFC period, supporting Robinson’s “Demand-Following Hypothesis,” and a bidirectional relationship during the GFC, Post-GFC and the full sample, aligning with Patrick’s “Feedback Hypothesis”.
These outcomes yield meaningful guidance for policymakers, financial institutions and investors, highlighting Sukuk’s potential for risk management, long-term growth and portfolio optimization during periods of financial instability.
This study uniquely explores Sukuk’s role in economic growth throughout the Global Financial Crisis (2007–2009), offering insights into its potential as a resilient instrument for risk management and sustained growth.
1. Introduction
Islamic finance presents a distinct alternative to conventional systems, built on principles that prohibit interest (Riba), limit uncertainty (Gharar) and exclude unethical practices. In the aftermath of the 2007–2009 Global Financial Crisis, its emphasis on risk-sharing and asset-backed financing drew increasing global attention (S&P Dow Jones Indices, 2022). Today, the industry spans banking, leasing, capital markets, takaful and microfinance, with Sukuk standing out as a key instrument. The sector’s total value reached $3.38 trillion in 2023, of which Islamic banking accounted for 69.3% and capital markets 29.8%, with Sukuk alone contributing 24.94% (around $0.842 trillion) (International Islamic Financial Market [IIFM], 2010).
Sukuk, Islamic financial certificates that serve as a parallel to conventional bonds, are not a recent innovation. Historically, the notion of “Sakk” dates back to the era of Caliph Umar ibn al-Khattab in the 1st century Hijri (634–644 A.D.). In modern finance, Sukuk are defined as “Certificates of equal value representing undivided shares in ownership of tangible assets, usufructs, services, or the real assets of specific projects” (AAOIFI, 2008). AAOIFI’s Standard 17(2) outlines three key criteria: ownership in real assets, returns derived from post-tax earnings (not interest) and maturity payments based on the market value of underlying assets.
Introduced in the mid-1990s, Sukuk have become a rapidly expanding segment in Islamic finance. They serve as essential tools for infrastructure financing, liquidity management and attracting a wide range of investors. Sukuk offer stable returns and greater transparency, enhancing market stability, especially during times of economic uncertainty (Naz, 2025). As shown in Figure 1, the Sukuk market peaked at USD 49 billion in 2007 before declining to USD 19 billion in 2008 and USD 26 billion in 2009 due to the GFC (IIFM, 2010).
The vertical axis of the vertical bar graph is labeled “Size in Million – U S dollars” and ranges from 0 to 200000 in increments of 20000 units. The horizontal axis is labeled “Year,” and displays years from 2001 to 2010, in increments of 1 year, followed by “Total.” The graph contains 11 vertical bars. The data values from left to right are as follows: 2001: 1,172. 2002: 1,371. 2003: 6,410. 2004: 8,140. 2005: 12,180. 2006: 30,034. 2007: 48,887. 2008: 18,597. 2009: 25,727. 2010: 45,123. Total: 197,642.Cumulative global Sukuk issuance from 2001 to 2010, showing the growth of Islamic bond markets during this period. Source: IIFM Sukuk Reports (2001–2010)
The vertical axis of the vertical bar graph is labeled “Size in Million – U S dollars” and ranges from 0 to 200000 in increments of 20000 units. The horizontal axis is labeled “Year,” and displays years from 2001 to 2010, in increments of 1 year, followed by “Total.” The graph contains 11 vertical bars. The data values from left to right are as follows: 2001: 1,172. 2002: 1,371. 2003: 6,410. 2004: 8,140. 2005: 12,180. 2006: 30,034. 2007: 48,887. 2008: 18,597. 2009: 25,727. 2010: 45,123. Total: 197,642.Cumulative global Sukuk issuance from 2001 to 2010, showing the growth of Islamic bond markets during this period. Source: IIFM Sukuk Reports (2001–2010)
However, the market rebounded, Figure 2 shows Sukuk touched a historic high of USD 188.121 billion in 2021. By December 2022, total Sukuk issuance had grown from USD 1.17 billion in 2001 to a cumulative USD 1,793 billion (IIFM, 2010).
The vertical axis of the vertical bar graph is labeled “Size in Million – U S dollars” and ranges from 0 to 200,000 in increments of 20,000 units. The horizontal axis is labeled “Year” and displays years from 2010 to 2022, in increments of 1 year. The graph contains 13 vertical bars. The data values from left to right are as follows: 2010: 53,125. 2011: 93,173. 2012: 137,599. 2013: 135,557. 2014: 107,833. 2015: 67,818. 2016: 88,318. 2017: 116,717. 2018: 123,150. 2019: 145,702. 2020: 174,641. 2021: 188,121. 2022: 182,715.Annual global Sukuk issuance from 2010 to 2022, showing year-to-year fluctuations in the Islamic bond market. Source: IIFM Sukuk Reports (2010–2022)
The vertical axis of the vertical bar graph is labeled “Size in Million – U S dollars” and ranges from 0 to 200,000 in increments of 20,000 units. The horizontal axis is labeled “Year” and displays years from 2010 to 2022, in increments of 1 year. The graph contains 13 vertical bars. The data values from left to right are as follows: 2010: 53,125. 2011: 93,173. 2012: 137,599. 2013: 135,557. 2014: 107,833. 2015: 67,818. 2016: 88,318. 2017: 116,717. 2018: 123,150. 2019: 145,702. 2020: 174,641. 2021: 188,121. 2022: 182,715.Annual global Sukuk issuance from 2010 to 2022, showing year-to-year fluctuations in the Islamic bond market. Source: IIFM Sukuk Reports (2010–2022)
The Sukuk market has gone global, with 37 countries issuing Sukuk by 2022, including non-Islamic nations such as the UK, Hong Kong and South Africa (S&P Dow Jones Indices, 2022). Asia and the Far East dominate the market with 61.38% share; Malaysia leads with 48.40% (USD 868 billion in issuance since 2001). The GCC and Middle East hold 24.53%, led by Saudi Arabia, Indonesia and the UAE. Sukuk continue to support infrastructure and budget financing across Malaysia, Bahrain, UAE and Pakistan bridging finance and the real economy through asset-backed models and risk-sharing, Sukuk promote sustainable development, inclusive finance and stability (Fitch Ratings, 2024).
The Global Financial Crisis (2007–2009), or “The Great Recession” was triggered by the burst of the U.S. housing bubble and the subprime mortgage crisis, exacerbated by risky instruments such as MBSs and CDOs (The Federal Reserve Board, 2008). The failure of major institutions like Lehman Brothers and Bear Stearns froze global credit markets despite $11.6 trillion in U.S. interventions (UNDP-HDRO, 2010). The crisis caused severe economic contractions: U.S. stock markets lost $8 trillion, unemployment peaked at 10% and household wealth fell by $9.8 trillion. Emerging economies also suffered, with growth dropping from 13.8% in 2007 to 2.1% in 2009 (IMF, 2009). Malaysia’s GDP nearly stalled at 0.1% in Q4 2008, Russia contracted by 10.4% in early 2009 and GCC growth plunged from 7.2% to 0.8%, alongside sharp housing market declines. Overall, developing countries lost an estimated $2.6 trillion in GDP between 2008 and 2010.
Sukuk issuance dropped by over 50% in 2008 as the global credit crunch constrained infrastructure financing. Although early in the crisis Sukuk and conventional bonds moved similarly, Sukuk later showed greater stability amid speculative pressures (Oakley, 2008). From September to December 2008, bond indices fell 6.53% following the collapse of Lehman Brothers and HBOS, while Sukuk registered only modest declines. The East Cameron Gas Sukuk default marked a rare case in Islamic finance. Falling oil prices down 78% since July 2008, further weakened demand in the Middle East. Recovery began in early 2009 with the G20’s $1.1 trillion stimulus, pushing Sukuk prices up by 38.2% by April, though defaults by Dar Al Kuwait, Saad Group and Nakheel raised concerns before Abu Dhabi’s bailout restored confidence (Nanaeva, 2010).
The Global Financial Crisis, characterized by reduced lending, constrained credit and weak GDP growth, underscored the need for alternative financing. Sukuk gained prominence in this context due to their asset-backed and risk-sharing nature. As a resilient and Shariah-compliant instrument, they mobilize long-term capital and support both public and private sector recovery (Naz, 2025). Yet, their role in driving economic growth before, during and after the crisis remains empirically underexplored. This study addresses this critical gap by examining the long-run and short-run relationship between Sukuk issuance and economic growth in the context of the Global Financial Crisis which constitutes the core objective of the study.
The empirical findings confirm that Sukuk issuance has a positive long-run impact on economic growth across the Pre-GFC, Post-GFC and full sample periods, while a negative but significant relationship emerges during the GFC. In contrast, short-term correlations are insignificant across all periods. Causality analysis shows a unidirectional relationship during the Pre-GFC phase, where GDP growth drives Sukuk issuance. Meanwhile, a bidirectional relationship during the GFC, Post-GFC and full sample periods where both Sukuk and GDP growth propels each other.
The findings of this research provide significant insights for market participants, investors and policymakers. The study underlines Sukuk’s importance as a resilient financial tool that promotes sustainable growth during economic uncertainty. Through enabling infrastructure development, enhancing income circulation and supporting inclusive finance, Sukuk contribute to broader developmental goals. The study emphasizes Sukuk’s unique features; risk-sharing and asset-backed structures and positions them as an effective alternative to traditional bonds. It contributes to the fields of Islamic finance, crisis management and economic development, offering practical implications for policy formulation, market development and the promotion of ethical financial practices globally.
The remainder of the paper is organized as follows: Section 2 provides a review of the theoretical and empirical literature. Section 3 outlines the research methodology. Section 4 presents the data analysis and interpretation of results. Section 5 concludes the study and provides implications, recommendations, limitations and directions for future research.
2. Literature review
2.1 Theoretical literature on finance–growth nexus
The theoretical discussion on the finance–growth nexus traces its roots to Schumpeter (1911), who introduced the “Supply-leading hypothesis”, suggesting that financial deepening drives economic development. In contrast Robinson (1952), proposed the “Demand-following hypothesis”, arguing that financial development follows economic growth. Over time Patrick (1966), suggested a shift in causality as economies mature, combining both supply-leading and demand-following perspectives in form of “Feedback hypothesis”. Meanwhile Lucas (1988), posits “Neutrality hypothesis” argued that financial development has no direct effect on growth under specific economic conditions. Further Solow (1956), Neoclassical Growth Theory highlighted the role of labor, capital and technology in driving economic growth.
2.2 Empirical literature on Sukuk and economic growth
Despite the Sukuk market’s significant growth since the 1990s, empirical evidence on Sukuk’s macroeconomic impact, especially on economic growth, is limited. Much of the existing work is descriptive, focusing on structures and compliance issues, with limited empirical depth due to data and market constraints. Recent studies increasingly recognize Sukuk’s long-term developmental potential. For instance Gürbüz et al. (2023), AbdulKareem et al. (2021), present Sukuk as a driver of economic development via capital market deepening, infrastructure investment and financial inclusion, aligning with the supply-leading view of finance-growth dynamics. On contrary, Kartini and Milawati (2020) questioned Sukuk’s short-run effectiveness in Indonesia, reporting negligible or statistically insignificant effects, suggesting that institutional and market readiness may moderate its growth impact. Sukuk’s hybrid structure, combining equity and debt features, also offers financial stability and reduces dependency on conventional borrowing (Gubareva et al., 2024). Its use in infrastructure, energy and social sectors (Smaoui et al., 2021) positions it as a tool for inclusive development. Ali et al. (2024) find that green Sukuk issued in Indonesia (2018–2021) significantly supported economic growth, social development and financial performance, with environmental certification attracting diverse investors. Their short-horizon analysis suggests green Sukuk may be more resilient and appealing to investors during periods of heightened sustainability awareness compared to conventional Sukuk.
The theoretical framing of Sukuk’s impact also varies. Studies like those of Naz and Gulzar (2022b) support the “Supply-leading Hypothesis” arguing that Sukuk issuance propels growth. Conversely Basyariah et al. (2021), suggest a “Demand-following” pattern, where rising GDP stimulates Sukuk development. Some, such as Fatimah and Rahmayanti (2023), adopt Patrick’s “Feedback Hypothesis” showing a two-way causality. Meanwhile, the neutrality view Lucas (1988) finds support in Tan and Shafi (2021), who found no robust link between Sukuk and growth. Kazak et al. (2023) examine conventional and Islamic banks’ impact on Türkiye’s real sector growth using a Fourier-based time-varying causality approach, showing that finance–growth links shift during crises. Their findings highlight the importance of considering contextual factors, such as legal frameworks and macroeconomic stability, when assessing Sukuk’s effectiveness. Overall, the literature points to promising potential, but Sukuk’s growth contribution is neither uniform nor automatic across economies.
2.3 Sukuk and global financial crisis
The Global Financial Crisis (2007–2009) exposed certain vulnerabilities in Islamic finance, including Sukuk, despite their asset-backed and risk-sharing nature. Issuances fell sharply in 2008 amid investor uncertainty, challenging the perception of Sukuk’s inherent stability (Rahim et al., 2021). The subsequent rebound, reflected in the recovery of the Dow Jones Sukuk Index and sovereign issuances such as the UK’s 2014 Sukuk, signaled renewed confidence. Although Sukuk offered diversification and low-correlation benefits during the crisis, their performance during COVID-19 was mixed, as liquidity constraints and oil price shocks affected both conventional and green Sukuk (Gubareva et al., 2024). By contrast, Pakistan’s experience illustrates their potential: during the pandemic, the government employed Sukuk to finance healthcare, relief and infrastructure, highlighting their role as a sustainable instrument for crisis management (Naz, 2025). These outcomes suggest that Sukuk’s resilience is contingent on crisis type, structural robustness and market maturity.
2.4 Sukuk, economic growth and global financial crisis
Although limited, empirical investigations into Sukuk, economic growth and the Global Financial Crisis (GFC) provide meaningful evidence. Said and Grassa (2013) reported a positive correlation between Sukuk and macroeconomic performance but noted that the GFC constrained issuance levels. Ahmad and Radzi (2011) quantified this, with issuances dropping from USD 46.65 billion in 2007 to USD 15.8 billion in 2008 amid a global credit squeeze. Other studies suggest the crisis amplified Sukuk’s role as an alternative to bank financing. Smaoui et al. (2021) argued that Sukuk complemented conventional instruments, serving as a substitute during liquidity constraints. More critically, Metoui and Ghorbel (2023) demonstrated Sukuk’s sustained positive effect on growth across sectors and crisis phases, challenging assumptions that Islamic finance lacks resilience. These insights reflect Sukuk’s evolving role not just as a financing instrument but also as a stabilizing force, though its effectiveness remains context-dependent. Taking a systemic view, Hassan et al. (2023) analyze convergence in Islamic financial development across major jurisdictions using the Fourier panel KPSS test. Their results indicate partial convergence, implying that shock-absorption capacity may be uneven across markets. Situating our cointegration results within this debate, the contribution of Sukuk issuance to cross-country stability may depend on whether national markets are converging toward similar levels of institutional maturity and resilience.
Despite growing interest in Islamic finance, most empirical studies focus on Islamic banking, with limited attention to Sukuk, particularly during financial crises (Sekmen, 2021). While some examine the Sukuk and economic growth link in stable periods (e.g. Yıldırım et al., 2020), there is a notable absence of studies examining this relationship across Pre-, During- and Post-GFC periods, limiting understanding of Sukuk’s resilience and role in financial stability. Existing studies are often country-specific, mainly on Malaysia or Indonesia, while a global analysis of leading Sukuk-issuing countries could yield more generalizable insights. Theoretical support is also inconsistent across supply-leading, demand-following, feedback or neutrality hypotheses. Thus, this study addresses these gaps by empirically investigating the dynamic relationship between Sukuk issuance and economic growth across crisis periods, contributing to a nuanced understanding of Sukuk’s effectiveness as a development and stability tool during financial turmoil. Derived from the literature gap, the current study proposes the following hypotheses:
There exists long-run cointegration among Sukuk issuance and GDP growth before the global financial crisis.
There exists short-run cointegration among Sukuk issuance and GDP growth before the global financial crisis.
There exists long-run cointegration among Sukuk issuance and GDP growth during the global financial crisis.
There exists short-run cointegration among Sukuk issuance and GDP growth during the global financial crisis.
There exists long-run cointegration among Sukuk issuance and GDP growth after the global financial crisis.
There exists short-run cointegration among Sukuk issuance and GDP growth after the global financial crisis.
There exists long-run cointegration among Sukuk issuance and GDP growth across full sample period.
There exists short-run cointegration among Sukuk issuance and GDP growth across full sample period.
3. Methodology
3.1 Research design
This study utilizes panel dataset spanning 22 years from 2001 to 2022, divided into three periods: Pre-GFC (2001 Q1–2006 Q4), During-GFC (2007 Q1–2009 Q4), Post-GFC (2010 Q1–2022 Q4) and Full sample (2001 Q1–2022 Q4) to analyze the long and short-run effects of Sukuk issuance on economic growth. The sample includes eight prominent Sukuk-issuing countries: Bahrain, Indonesia, Malaysia, Pakistan, Qatar, Saudi Arabia, Sudan and the UAE. These countries were chosen for their active participation in Sukuk issuance by value and volume. Their sustainable role since 2001 and issuance activity across pre-crisis, crisis and post-crisis phases further justified their inclusion. The dataset allows a thorough analysis of the Sukuk–economic growth relationship across different stages of the Global Financial Crisis. Table 1 provides a detailed description of the variables used. The annual data are converted to quarterly using the quadratic match-average method to avoid seasonality. Finally, logarithmic transformation was applied to address heterogeneity and outliers.
Description of variables
| Variables | Description (USD-Mil) | Notation | Source |
|---|---|---|---|
| Dependent variable | |||
| GDP growth | Annual % growth rate of GDP at constant 2015 prices, expressed in U.S. dollars | GDPG | World Development Indicator (WDI) |
| Independent variable | |||
| Sukuk | The total value of Sukuk issued | SUK | International Islamic Financial market (IIFM) |
| Control variables | |||
| Gross fixed capital formation | The total value of Capital accumulation/Investment on infrastructure | GFCF | World Development Indicator (WDI) |
| General government final expenditure | The total value of Government expenditure on goods and services | GGFE | |
| Inflation | Annual % change in Consumer price Index | INF | |
| Trade openness | The total value of Import and export as percentage of GDP | TOPN | |
| Financial development index | It is an aggregate of the Financial Institutions Index and the Financial Market Index | FDI | International Monetary Fund (IMF database) |
| Global financial crisis-(Dummy) | Dummy equal to “1” if the observation corresponds to the period 2007Q1–2009Q4, and “0” otherwise | DUMMY-GFC | |
| Variables | Description (USD-Mil) | Notation | Source |
|---|---|---|---|
| Dependent variable | |||
| GDP growth | Annual % growth rate of GDP at constant 2015 prices, expressed in U.S. dollars | GDPG | World Development Indicator (WDI) |
| Independent variable | |||
| Sukuk | The total value of Sukuk issued | SUK | International Islamic Financial market (IIFM) |
| Control variables | |||
| Gross fixed capital formation | The total value of Capital accumulation/Investment on infrastructure | GFCF | World Development Indicator (WDI) |
| General government final expenditure | The total value of Government expenditure on goods and services | GGFE | |
| Inflation | Annual % change in Consumer price Index | INF | |
| Trade openness | The total value of Import and export as percentage of GDP | TOPN | |
| Financial development index | It is an aggregate of the Financial Institutions Index and the Financial Market Index | FDI | International Monetary Fund (IMF database) |
| Global financial crisis-(Dummy) | Dummy equal to “1” if the observation corresponds to the period 2007Q1–2009Q4, and “0” otherwise | DUMMY-GFC | |
Note(s): The table presents the variables employed in the analysis, along with their descriptions, notations and respective data sources
3.2 Description of variables
The study variables, along with their descriptions, abbreviations and data sources, are presented in Table 1.
3.3 Econometric model
This study examines the Sukuk–economic growth relationship during the Global Financial Crisis, using the Neo-Classical Growth Model (growth as a function of labor, capital and technology). Following Setianingsih and Widyastuti (2020), Sukuk is viewed as a driver in Islamic finance, mobilizing funds for infrastructure and capital, promoting efficient resource allocation through profit-and-loss sharing and supporting sustainable investments. It boosts job creation, labor productivity and human capital, while green and social Sukuk advance technological progress by financing renewable energy and smart city projects.
3.4 Model estimation techniques
The analysis followed a structured sequence, starting with descriptive statistics, followed by diagnostic procedures including cross-sectional dependency and stationarity tests. Subsequently, Johansen panel cointegration tests (Fisher and Kao) were employed. The PMG/ARDL approach was then applied, and the process concluded with panel causality tests to identify directional relationships. All estimations were carried out using EViews 12 software.
3.4.1 Pooled Mean Group estimation of ARDL
The ARDL (Autoregressive Distributed Lag) model analyses lagged values, while the PMG estimator (Pesaran et al., 1999) extends ARDL to panel data, allowing short-run coefficients to vary across cross-sections while keeping long-run coefficients consistent. Pesaran et al. (1999) highlight ARDL’s error correction form as a more advanced cointegration method compared to Johansen and Engle & Granger.
Earlier studies, such as Naz and Gulzar (2022a), employed the PMG/ARDL technique to examine long- and short-run dynamics in panel data. The present research has selected, PMG/ARDL due to its strong alignment with the dataset for several reasons. First, the dataset (N < T) with eight cross-sections and varying time series lengths (Pre-GFC: T = 24, During-GFC: T = 12, Post-GFC: T = 52, Full sample: T = 88) suits PMG’s capabilities. Second, PMG/ARDL estimates both long- and short-term coefficients simultaneously, making it ideal for the study’s objectives. Third, PMG/ARDL handles mixed stationarity I(0) and I(1) effectively. Fourth, it is robust to outliers, using lagged terms and error correction to stabilize estimates. Fifth, it addresses issues of endogeneity, heteroscedasticity, autocorrelation and multicollinearity. Lastly, PMG/ARDL provides higher estimation efficiency, better degrees of freedom and low co-linearity compared to other models.
Equation (1) specifies the long-run PMG-ARDL(p, q) model employed to investigate the dynamic relationship between Sukuk issuance and economic growth.
∆LGDPG is the change in Logarithm of GDP growth and is a dependent variable and LSUK Logarithm of Sukuk issuance is an independent variable, remaining LGFCF, LGGFE, LINF, LTOPN, LFDI are the Logarithm of Gross fixed capital formation, General government final expenditure, Inflation, Trade openness, Financial development index, respectively, DUMMY-GFC are control variables. All exogenous variables are with l = 1,2,3,4,5,6,7, εit is an error term and Δ is the first difference operator.α1 is an intercept, γ1 … …γ8 are coefficient of lagged dependent variable called scalars. δ1 … … δ8 represents the change in exploratory variables would bring change in GDP growth. δ1, δ2, δ3, δ6, δ7, >0 states that increase in value of GDP growth, Sukuk issuance, Gross fixed capital formation, Trade openness, Financial development index, leads to an increase in the GDP growth. However, δ4, δ5, δ8 < 0 states that increase in value of General government final expenditure and Inflation, DUMMY-GFC leads to decrease in the GDP growth. The “t” specifies the dimension of a time period, t = 1,2 …., T and “i” represents the cross-section dimension, i = 1, 2 …. N. p, q are optimal lag orders.
Notably, Eq. (1) applies to the full sample period (2001Q1–2022Q4), including the GFC dummy variable (DUMMY-GFC) to capture the crisis’s impact (2007–2009). The dummy is excluded from Pre-GFC and Post-GFC models, as these periods fall outside the crisis, and During-GFC directly measures the crisis phase. The specification of short-run dynamics using the PMG/ARDL approach involves constructing an error correction model, which can be formulated as follows in equation (2):
∆LGDPG the change in Logarithm of GDP growth is a dependent variable and ∆LSUK, change in Logarithm of Sukuk issuance is an independent variable, remaining ∆LGFCF, ∆LGGFE, ∆LINF, ∆LTOPN, ∆LFDI are the change in Logarithm of Gross fixed capital formation, General government final expenditure, Inflation, Trade openness, Financial development index, respectively ∆DUMMY-GFC is the change in Global Financial Crisis are control variables, with l = 1,2,3,4,5,6,7,8 and Δ is the first difference operator. The residuals ε lit (l = {1, 2, 3, 4,5,6,7, 8}) are independent and control variables. ECT l, it-1 (l = {1, 2, 3, 4, 5,6, 7, 8 }) is the error correction term defined by the long-term relationship. The parameter μ li indicates the speed of adjustment to the equilibrium level.
Particularly, Eq. (2) applies to the full sample period (2001Q1–2022Q4), including the GFC dummy variable (DUMMY-GFC) to capture the crisis’s impact (2007–2009). The dummy is excluded from Pre-GFC and Post-GFC models, as these periods fall outside the crisis, and During-GFC directly measures the crisis phase.
4. Results and analysis
4.1 Diagnostic tests
Descriptive statistics was carried out across the four sample periods: Pre-GFC, During-GFC, Post-GFC and the full sample. These results verified the consistency and reliability of the data, allowing further diagnostic tests and the subsequent main analysis to proceed.
4.1.1 Cross-section dependence test
Cross-sectional dependence tests are essential in panel data models to ensure valid coefficient estimates. This study utilized the Breusch-Pagan LM, Pesaran scaled LM, Bias-corrected scaled LM and Pesaran CD tests to detect dependence.
Cross-sectional dependence tests as shown in Table 2, yielded mixed signals across different methods; however, the majority of test statistics confirmed independence in the Pre-, During-, Post-GFC and Full sample periods. While some tests (e.g. Pesaran CD and scaled LM) indicated mild dependence in certain periods, the Breusch–Pagan LM test consistently supported independence. Overall, the results validate the assumption of cross-sectional independence, allowing for reliable application of first-generation panel unit root tests in subsequent analysis.
Cross-section dependence test
| CDS tests | Breusch-Pagan LM | Pesaran scaled LM | Pesaran CD | |||
|---|---|---|---|---|---|---|
| T-statistic | P-value | T-statistic | P-value | T-statistic | P-value | |
| Pre-GFC | 25.745 | 0.5871 | −0.3013 | 0.7632 | 2.0290 | 0.0424 |
| During-GFC | 12.051 | 0.9962 | −2.1312 | 0.0331 | −0.4980 | 0.6185 |
| Post-GFC | 9.5593 | 0.9995 | −2.4642 | 0.0137 | 0.0099 | 0.9921 |
| Full-sample | 39.334 | 0.0757 | 1.5145 | 0.1299 | 0.5059 | 0.6129 |
| CDS tests | Breusch-Pagan LM | Pesaran scaled LM | Pesaran CD | |||
|---|---|---|---|---|---|---|
| T-statistic | P-value | T-statistic | P-value | T-statistic | P-value | |
| Pre-GFC | 25.745 | 0.5871 | −0.3013 | 0.7632 | 2.0290 | 0.0424 |
| During-GFC | 12.051 | 0.9962 | −2.1312 | 0.0331 | −0.4980 | 0.6185 |
| Post-GFC | 9.5593 | 0.9995 | −2.4642 | 0.0137 | 0.0099 | 0.9921 |
| Full-sample | 39.334 | 0.0757 | 1.5145 | 0.1299 | 0.5059 | 0.6129 |
Note(s): The table presents cross-sectional dependence test results (Breusch–Pagan LM, Pesaran scaled LM and Pesaran CD) with respective p-values across Pre-, During-, Post-GFC and full-sample periods
4.1.2 Panel unit root test
Ensuring data stationarity is crucial for valid econometric analysis, as non-stationary variables can affect cointegration and causality tests. Panel unit root tests (LLC, IPS, ADF and PP) were applied to examine the statistical properties of the predictor and outcome variables. In the flexible PMG/ARDL framework, variables maybe I(0) or I(1), consistent with Pesaran et al. (1999), and no variable should exceed I(1).
Table 3 presents the panel unit root test results across the four samples (Pre-, During-, Post-GFC and full sample), confirming a mix of I(0) and I(1) variables, with none integrated at I(2). These findings validate the use of the PMG/ARDL model and confirm the data’s suitability for long-run estimation.
Panel unit root test
| Variable | Individual intercept @ I(0) | Individual intercept @ I(I) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| LLC | IPS | ADF | PP | LLC | IPS | ADF | PP | Status | |
| Pre-GFC | |||||||||
| LGDPG | 0.0000** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LSUK | 0.0012** | 0.3847 | 0.5477 | 0.7278 | 0.0000** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LGFCF | 0.5808 | 1.0000 | 0.9999 | 0.10000 | 0.0262** | 0.1376 | 0.1053 | 0.0049** | I(1) |
| LGGFE | 0.0071** | 0.7474 | 0.8729 | 0.0005** | 0.0179 | 0.0653 | 0.1257 | 0.0097** | I(0) |
| LINF | 0.0031** | 0.0056** | 0.0007** | 0.0911 | 0.0000** | 0.0000** | 0.0000** | 0.0310** | I(0) |
| LTOPN | 0.2421 | 0.9918 | 0.9959 | 0.9549 | 0.0848 | 0.0366** | 0.0939 | 0.0106** | I(I) |
| LFDI | 0.5223 | 0.5715 | 0.5476 | 0.9416 | 0.7480 | 0.1703 | 0.3551 | 0.0376** | I(1) |
| During-GFC | |||||||||
| LGDPG | 0.7480 | 0.9885 | 0.9938 | 0.9979 | 0.0013** | 0.0281** | 0.0763 | 0.0745 | I(I) |
| LSUK | 0.0000** | 0.0003** | 0.0014** | 0.0136** | 0.1071 | 0.3852 | 0.6230 | 0.5573 | I(0) |
| LGFCF | 0.0192** | 0.3727 | 0.4767 | 0.5157 | 0.0510** | 0.2097 | 0.3436 | 0.3206 | I(0) |
| LGGFE | 0.0088** | 0.9416 | 0.9035 | 0.2442 | 0.2977 | 0.3789 | 0.4701 | 0.4808 | I(0) |
| LINF | 0.7783 | 0.8501 | 0.8723 | 0.9904 | 0.0156** | 0.3050 | 0.4890 | 0.4075 | I(I) |
| LTOPN | 0.0559** | 0.5121 | 0.7273 | 0.8202 | 0.4730 | 0.6894 | 0.9029 | 0.4899 | I(0) |
| LFDI | 0.4254 | 0.4915 | 0.4686 | 0.3219 | 0.0401** | 0.2529 | 0.1578 | 0.0003** | I(I) |
| Post-GFC | |||||||||
| LGDPG | 0.0109** | 0.0000** | 0.0000** | 0.0165** | 0.2396 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LSUK | 0.0142** | 0.0000** | 0.0000** | 0.0067** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LGFCF | 0.0008** | 0.2088 | 0.3851 | 0.2807 | 0.6902 | 0.0004** | 0.0020** | 0.0000** | I(0) |
| LGGFE | 0.0000** | 0.0144** | 0.0238** | 0.0100** | 0.9230 | 0.0002** | 0.0004** | 0.0000** | I(0) |
| LINF | 0.4969 | 0.0121** | 0.0231** | 0.2741 | 0.9963 | 0.0023** | 0.0000** | 0.0000** | I(0) |
| LTOPN | 0.6813 | 0.5218 | 0.1846 | 0.0446** | 0.7325 | 0.0063** | 0.0153** | 0.0000** | I(0) |
| LFDI | 0.0024** | 0.0017** | 0.0036** | 0.2909 | 0.4899 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| Full sample | |||||||||
| LGDPG | 0.0193** | 0.0000** | 0.0000** | 0.0000** | 0.0769 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LSUK | 0.0000** | 0.0002** | 0.0011** | 0.0107** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LGFCF | 0.0009** | 0.3198 | 0.4430 | 0.1665 | 0.1586 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LGGFE | 0.0000** | 0.1127 | 0.1728 | 0.0010** | 0.1025 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LINF | 0.1517 | 0.0000** | 0.0000** | 0.0004** | 0.7388 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LTOPN | 0.0078** | 0.5371 | 0.6148 | 0.9329 | 0.7377 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LFDI | 0.0635 | 0.0007** | 0.0014** | 0.0752 | 0.0381** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| DUMMY GFC | 0.5251 | 0.0283** | 0.1211 | 0.0878 | – | – | – | – | I(0) |
| Variable | Individual intercept @ I(0) | Individual intercept @ I(I) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| LLC | IPS | ADF | PP | LLC | IPS | ADF | PP | Status | |
| Pre-GFC | |||||||||
| LGDPG | 0.0000** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LSUK | 0.0012** | 0.3847 | 0.5477 | 0.7278 | 0.0000** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LGFCF | 0.5808 | 1.0000 | 0.9999 | 0.10000 | 0.0262** | 0.1376 | 0.1053 | 0.0049** | I(1) |
| LGGFE | 0.0071** | 0.7474 | 0.8729 | 0.0005** | 0.0179 | 0.0653 | 0.1257 | 0.0097** | I(0) |
| LINF | 0.0031** | 0.0056** | 0.0007** | 0.0911 | 0.0000** | 0.0000** | 0.0000** | 0.0310** | I(0) |
| LTOPN | 0.2421 | 0.9918 | 0.9959 | 0.9549 | 0.0848 | 0.0366** | 0.0939 | 0.0106** | I(I) |
| LFDI | 0.5223 | 0.5715 | 0.5476 | 0.9416 | 0.7480 | 0.1703 | 0.3551 | 0.0376** | I(1) |
| During-GFC | |||||||||
| LGDPG | 0.7480 | 0.9885 | 0.9938 | 0.9979 | 0.0013** | 0.0281** | 0.0763 | 0.0745 | I(I) |
| LSUK | 0.0000** | 0.0003** | 0.0014** | 0.0136** | 0.1071 | 0.3852 | 0.6230 | 0.5573 | I(0) |
| LGFCF | 0.0192** | 0.3727 | 0.4767 | 0.5157 | 0.0510** | 0.2097 | 0.3436 | 0.3206 | I(0) |
| LGGFE | 0.0088** | 0.9416 | 0.9035 | 0.2442 | 0.2977 | 0.3789 | 0.4701 | 0.4808 | I(0) |
| LINF | 0.7783 | 0.8501 | 0.8723 | 0.9904 | 0.0156** | 0.3050 | 0.4890 | 0.4075 | I(I) |
| LTOPN | 0.0559** | 0.5121 | 0.7273 | 0.8202 | 0.4730 | 0.6894 | 0.9029 | 0.4899 | I(0) |
| LFDI | 0.4254 | 0.4915 | 0.4686 | 0.3219 | 0.0401** | 0.2529 | 0.1578 | 0.0003** | I(I) |
| Post-GFC | |||||||||
| LGDPG | 0.0109** | 0.0000** | 0.0000** | 0.0165** | 0.2396 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LSUK | 0.0142** | 0.0000** | 0.0000** | 0.0067** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LGFCF | 0.0008** | 0.2088 | 0.3851 | 0.2807 | 0.6902 | 0.0004** | 0.0020** | 0.0000** | I(0) |
| LGGFE | 0.0000** | 0.0144** | 0.0238** | 0.0100** | 0.9230 | 0.0002** | 0.0004** | 0.0000** | I(0) |
| LINF | 0.4969 | 0.0121** | 0.0231** | 0.2741 | 0.9963 | 0.0023** | 0.0000** | 0.0000** | I(0) |
| LTOPN | 0.6813 | 0.5218 | 0.1846 | 0.0446** | 0.7325 | 0.0063** | 0.0153** | 0.0000** | I(0) |
| LFDI | 0.0024** | 0.0017** | 0.0036** | 0.2909 | 0.4899 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| Full sample | |||||||||
| LGDPG | 0.0193** | 0.0000** | 0.0000** | 0.0000** | 0.0769 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LSUK | 0.0000** | 0.0002** | 0.0011** | 0.0107** | 0.0000** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LGFCF | 0.0009** | 0.3198 | 0.4430 | 0.1665 | 0.1586 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LGGFE | 0.0000** | 0.1127 | 0.1728 | 0.0010** | 0.1025 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LINF | 0.1517 | 0.0000** | 0.0000** | 0.0004** | 0.7388 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LTOPN | 0.0078** | 0.5371 | 0.6148 | 0.9329 | 0.7377 | 0.0000** | 0.0000** | 0.0000** | I(0) |
| LFDI | 0.0635 | 0.0007** | 0.0014** | 0.0752 | 0.0381** | 0.0000** | 0.0000** | 0.0000** | I(0) |
| DUMMY GFC | 0.5251 | 0.0283** | 0.1211 | 0.0878 | – | – | – | – | I(0) |
Note(s): The table reports panel unit root test outcomes (LLC, IPS, ADF, PP) for all variables across pre-, during- and post-GFC periods, as well as the full sample, at both levels I(0) and first differences I(1) at the 5% level of significance, with the Status column showing the inferred integration order
4.2 Panel cointegration test
In the next step, multivariate panel cointegration analysis was conducted using Kao (1999) and Fisher-type tests. These tests are essential for identifying long-term relationships among variables and avoiding misleading results.
4.2.1 Johansen–Fisher panel cointegration test
The Johansen–Fisher combined panel test was applied to test the null hypothesis of “no cointegration” for the Pre-GFC, Post-GFC and Full sample datasets. However, it could not be computed for the During-GFC period due to insufficient observations. Results for the applicable periods are presented in Table 4.
Table 4 presents the Johansen–Fisher combined test results. For the Pre-GFC period, both the Trace and Max–Eigen tests at the 5% significance level confirm a strong long-term relationship between Sukuk issuance and economic growth. Throughout the Post-GFC period, the Trace test affirms a robust long-run correlation, while the Max–Eigen test highlights the significance of 5 out of 8 cointegrating equations at the 5% level, reinforcing the evidence of a substantial relationship. For the Full sample period, the Trace test indicates significance in 6 out of 8 equations, and the Max–Eigen test supports 3 out of 8 at the 5% level, demonstrating a consistent and enduring long-term association since the inception of Sukuk issuance. The results provide consistent evidence against the null hypothesis of “no cointegration” confirming strong long-term relationships across all periods.
Johansen Fisher panel cointegration
| Unrestricted cointegration rank test (trace and maximum eigenvalue) | ||||
|---|---|---|---|---|
| Hypothesized no. of CE(s) | Fisher stat* (trace test) | Prob | Fisher stat* (max-eigen test) | Prob |
| Pre-GFC | ||||
| None | 177.1 | 0.0000** | 679.7 | 0.0000** |
| At most 1 | 380.0 | 0.0000** | 204.7 | 0.0000** |
| At most 2 | 362.6 | 0.0000** | 258.2 | 0.0000** |
| At most 3 | 224.9 | 0.0000** | 126.7 | 0.0000** |
| At most 4 | 127.0 | 0.0000** | 68.73 | 0.0000** |
| At most 5 | 80.31 | 0.0000** | 60.21 | 0.0000** |
| At most 6 | 52.55 | 0.0000** | 52.55 | 0.0000** |
| Post-GFC | ||||
| None | 151.7 | 0.0000** | 70.96 | 0.0000** |
| At most 1 | 91.02 | 0.0000** | 28.38 | 0.0285** |
| At most 2 | 69.73 | 0.0000** | 34.41 | 0.0048** |
| At most 3 | 46.86 | 0.0001** | 19.23 | 0.2569 |
| At most 4 | 39.09 | 0.0011** | 22.35 | 0.1322 |
| At most 5 | 31.82 | 0.0105** | 26.72 | 0.0447** |
| At most 6 | 23.70 | 0.0963 | 23.70 | 0.0963 |
| Full sample | ||||
| None | 198.1 | 0.0000** | 61.35 | 0.0000** |
| At most 1 | 134.0 | 0.0000** | 43.09 | 0.0003** |
| At most 2 | 92.20 | 0.0000** | 27.10 | 0.0404** |
| At most 3 | 69.79 | 0.0000** | 25.68 | 0.0587** |
| At most 4 | 50.55 | 0.0000** | 19.73 | 0.2325 |
| At most 5 | 40.29 | 0.0007** | 21.57 | 0.1576 |
| At most 6 | 32.94 | 0.0075** | 20.79 | 0.1868 |
| At most 7 | 42.09 | 0.0004** | 42.09 | 0.0004** |
| Unrestricted cointegration rank test (trace and maximum eigenvalue) | ||||
|---|---|---|---|---|
| Hypothesized no. of CE(s) | Fisher stat* (trace test) | Prob | Fisher stat* (max-eigen test) | Prob |
| Pre-GFC | ||||
| None | 177.1 | 0.0000** | 679.7 | 0.0000** |
| At most 1 | 380.0 | 0.0000** | 204.7 | 0.0000** |
| At most 2 | 362.6 | 0.0000** | 258.2 | 0.0000** |
| At most 3 | 224.9 | 0.0000** | 126.7 | 0.0000** |
| At most 4 | 127.0 | 0.0000** | 68.73 | 0.0000** |
| At most 5 | 80.31 | 0.0000** | 60.21 | 0.0000** |
| At most 6 | 52.55 | 0.0000** | 52.55 | 0.0000** |
| Post-GFC | ||||
| None | 151.7 | 0.0000** | 70.96 | 0.0000** |
| At most 1 | 91.02 | 0.0000** | 28.38 | 0.0285** |
| At most 2 | 69.73 | 0.0000** | 34.41 | 0.0048** |
| At most 3 | 46.86 | 0.0001** | 19.23 | 0.2569 |
| At most 4 | 39.09 | 0.0011** | 22.35 | 0.1322 |
| At most 5 | 31.82 | 0.0105** | 26.72 | 0.0447** |
| At most 6 | 23.70 | 0.0963 | 23.70 | 0.0963 |
| Full sample | ||||
| None | 198.1 | 0.0000** | 61.35 | 0.0000** |
| At most 1 | 134.0 | 0.0000** | 43.09 | 0.0003** |
| At most 2 | 92.20 | 0.0000** | 27.10 | 0.0404** |
| At most 3 | 69.79 | 0.0000** | 25.68 | 0.0587** |
| At most 4 | 50.55 | 0.0000** | 19.73 | 0.2325 |
| At most 5 | 40.29 | 0.0007** | 21.57 | 0.1576 |
| At most 6 | 32.94 | 0.0075** | 20.79 | 0.1868 |
| At most 7 | 42.09 | 0.0004** | 42.09 | 0.0004** |
Note(s): The table reports Johansen–Fisher panel cointegration results (Trace and Max–Eigen statistics with Fisher values and p-values) for pre-, post-GFC and full samples, showing multiple significant cointegrating relationships at the 5% level**
4.2.2 Kao panel cointegration test
Furthermore, the Kao residual panel cointegration test (Kao, 1999) was employed to confirm the existence of long-run relationships among the variables. This residual-based test evaluates the equilibrium adherence of variables over time by testing the null hypothesis of “no cointegration”. Analyses were conducted for the Pre-, During-, Post-GFC and Full sample datasets. The results are summarized in Table 5. The outcomes consistently disprove the null hypothesis of “no cointegration” affirming the presence of strong long-term relationships between Sukuk issuance and economic growth across all periods.
Kao residual cointegration test
| Pre-GFC | ||
|---|---|---|
| t-statistic | Prob. | |
| ADF | −4.7848 | 0.0000** |
| Residual variance | 0.0165 | |
| HAC variance | 0.0228 | |
| Pre-GFC | ||
|---|---|---|
| t-statistic | Prob. | |
| ADF | −4.7848 | 0.0000** |
| Residual variance | 0.0165 | |
| HAC variance | 0.0228 | |
| Variable | Coefficient | Std. error | t-statistic | Prob. |
|---|---|---|---|---|
| RESID(−1) | −0.2436 | 0.0333 | −7.3159 | 0.0000** |
| D(RESID(−1)) | 0.4882 | 0.0534 | 9.1384 | 0.0000** |
| Variable | Coefficient | Std. error | t-statistic | Prob. |
|---|---|---|---|---|
| RESID(−1) | −0.2436 | 0.0333 | −7.3159 | 0.0000** |
| D(RESID(−1)) | 0.4882 | 0.0534 | 9.1384 | 0.0000** |
| During-GFC | ||
|---|---|---|
| t-statistic | Prob. | |
| ADF | −2.5416 | 0.0055** |
| Residual variance | 0.0359 | |
| HAC variance | 0.0468 | |
| During-GFC | ||
|---|---|---|
| t-statistic | Prob. | |
| ADF | −2.5416 | 0.0055** |
| Residual variance | 0.0359 | |
| HAC variance | 0.0468 | |
| Variable | Coefficient | Std. error | t-statistic | Prob. |
|---|---|---|---|---|
| RESID(−1) | −0.3995 | 0.0839 | −4.7584 | 0.0000** |
| D(RESID(−1)) | 0.4639 | 0.1090 | 4.2538 | 0.0001** |
| Variable | Coefficient | Std. error | t-statistic | Prob. |
|---|---|---|---|---|
| RESID(−1) | −0.3995 | 0.0839 | −4.7584 | 0.0000** |
| D(RESID(−1)) | 0.4639 | 0.1090 | 4.2538 | 0.0001** |
| Post-GFC | ||
|---|---|---|
| t-statistic | Prob. | |
| ADF | −5.4990 | 0.0000** |
| Residual variance | 0.0389 | |
| HAC variance | 0.0689 | |
| Post-GFC | ||
|---|---|---|
| t-statistic | Prob. | |
| ADF | −5.4990 | 0.0000** |
| Residual variance | 0.0389 | |
| HAC variance | 0.0689 | |
| Variable | Coefficient | Std. error | t-statistic | Prob. |
|---|---|---|---|---|
| RESID(−1) | −0.1825 | 0.0225 | −8.0870 | 0.0000** |
| D(RESID(−1)) | 0.5005 | 0.0446 | 11.212 | 0.0000** |
| Variable | Coefficient | Std. error | t-statistic | Prob. |
|---|---|---|---|---|
| RESID(−1) | −0.1825 | 0.0225 | −8.0870 | 0.0000** |
| D(RESID(−1)) | 0.5005 | 0.0446 | 11.212 | 0.0000** |
| Full sample | ||
|---|---|---|
| t-statistic | Prob. | |
| ADF | −8.0339 | 0.0000** |
| Residual variance | 0.0367 | |
| HAC variance | 0.0638 | |
| Full sample | ||
|---|---|---|
| t-statistic | Prob. | |
| ADF | −8.0339 | 0.0000** |
| Residual variance | 0.0367 | |
| HAC variance | 0.0638 | |
| Variable | Coefficient | Std. error | t-statistic | Prob. |
|---|---|---|---|---|
| RESID(−1) | −0.1786 | 0.0168 | −10.604 | 0.0000** |
| D(RESID(−1)) | 0.4672 | 0.0334 | 13.980 | 0.0000** |
| Variable | Coefficient | Std. error | t-statistic | Prob. |
|---|---|---|---|---|
| RESID(−1) | −0.1786 | 0.0168 | −10.604 | 0.0000** |
| D(RESID(−1)) | 0.4672 | 0.0334 | 13.980 | 0.0000** |
Note(s): The table reports Kao residual cointegration results for pre-, during- and post-GFC, as well as the full sample. Reported are ADF t-statistics, variances and regression outputs for RESID(−1) and D(RESID(−1)). Findings confirm cointegration across all periods at the 5% level of significance**
4.3 Pooled Mean Group of Autoregressive Distributed Lag
The study utilized the Pooled Mean Group (PMG) estimator developed by Pesaran et al. (1999) to assess both long- and short-run coefficients as well as the adjustment dynamics. Model selection was based on the Akaike Information Criterion (AIC), ensuring appropriate lag length. The significant and negative sign of the error correction terms confirm the presence of an adjustment mechanism that restores long-run equilibrium following temporary fluctuations. These outcomes affirm cointegration among all the four models.
Table 6 presents the PMG/ARDL results for the Pre-GFC period, identifying the optimal model as (2,2,2,2,2,2,2). The long-run equation shows that “LSUK” has a positive and statistically significant impact on “LGDPG” with a coefficient of 0.1650 and a p-value of 0.0000 (p ≤ 0.05). This indicates that a 1% increase in “LSUK” is associated with a 0.165% increase in “LGDPG” supporting the hypothesis H1a. This finding aligns with studies of Gürbüz et al. (2023), Novitasari and Arundina (2023) but contrasts with studies of Bella and Inas (2023), who reported no long-term impact of Sukuk on economic growth. Our outcomes for the pre-crisis period highlight that active corporate participation drove Sukuk growth, with global issuance rising from $1 billion in 2001 to $56.5 billion by 2006 (CAGR: 21.23%), led by Malaysia, Bahrain and Indonesia. Sukuk supported infrastructure financing and market diversification, enhancing economic stability, though effects on large-scale projects and foreign investment remained uneven, even in high-growth countries like Malaysia and the UAE.
Pooled mean group pre-GFC
| PMG/ARDL model (2,2,2,2,2,2,2) | ||||
|---|---|---|---|---|
| Variable | Coefficient | Std. error | t-statistic | Prob** |
| Long run equation | ||||
| LSUK | 0.1650 | 0.0065 | 25.265 | 0.0000** |
| LGFCF | −2.8325 | 0.2482 | −11.409 | 0.0000** |
| LGGFE | 0.9335 | 0.1665 | 5.6057 | 0.0000** |
| LINF | 0.2927 | 0.0242 | 12.070 | 0.0000** |
| LTOPN | 0.9389 | 0.0960 | 9.7727 | 0.0000** |
| LFDI | 3.2056 | 0.3946 | 8.1220 | 0.0000** |
| Short run equation | ||||
| COINTEQ01 | −0.1650 | 0.0515 | −3.2055 | 0.0021** |
| D(LSUK) | 0.0251 | 0.0262 | 0.9550 | 0.3430 |
| D(LGFCF) | −5.2304 | 7.0610 | −0.7407 | 0.4615 |
| D(LGGFE) | 21.823 | 7.8041 | 2.7963 | 0.0068** |
| D(LINF) | 0.0221 | 0.3771 | 0.0586 | 0.9534 |
| D(LTOPN) | 1.0354 | 5.6673 | 0.1826 | 0.8556 |
| D(LFDI) | −6.1431 | 27.503 | −0.2233 | 0.8239 |
| PMG/ARDL model (2,2,2,2,2,2,2) | ||||
|---|---|---|---|---|
| Variable | Coefficient | Std. error | t-statistic | Prob** |
| Long run equation | ||||
| LSUK | 0.1650 | 0.0065 | 25.265 | 0.0000** |
| LGFCF | −2.8325 | 0.2482 | −11.409 | 0.0000** |
| LGGFE | 0.9335 | 0.1665 | 5.6057 | 0.0000** |
| LINF | 0.2927 | 0.0242 | 12.070 | 0.0000** |
| LTOPN | 0.9389 | 0.0960 | 9.7727 | 0.0000** |
| LFDI | 3.2056 | 0.3946 | 8.1220 | 0.0000** |
| Short run equation | ||||
| COINTEQ01 | −0.1650 | 0.0515 | −3.2055 | 0.0021** |
| D(LSUK) | 0.0251 | 0.0262 | 0.9550 | 0.3430 |
| D(LGFCF) | −5.2304 | 7.0610 | −0.7407 | 0.4615 |
| D(LGGFE) | 21.823 | 7.8041 | 2.7963 | 0.0068** |
| D(LINF) | 0.0221 | 0.3771 | 0.0586 | 0.9534 |
| D(LTOPN) | 1.0354 | 5.6673 | 0.1826 | 0.8556 |
| D(LFDI) | −6.1431 | 27.503 | −0.2233 | 0.8239 |
Note(s): This table presents PMG-ARDL (2,2,2,2,2,2,2) estimates for the pre-GFC period, reporting long-run coefficients, short-run dynamics through the error correction term and significance at the 5% level**
The short-run equation shows “LSUK” has a positive but insignificant coefficient of 0.0251 (p = 0.3430), indicating no significant short-run relationship between Sukuk issuance and GDP growth during the Pre-GFC period. Thus, the hypothesis H1b is not supported. This line up with Fuadi et al. (2022), contradicting Naz and Gulzar (2023). Our results indicate that limited liquidity, a small investor base, complex structures, underdeveloped regulations and long project gestation hindered Sukuk’s short-term economic impact.
Table 7 exhibits the PMG/ARDL results for the During-GFC period, identifying the optimal model as (1,1,1,1,1,1,1). The variable “LSUK” shows a negative and statistically significant impact on “LGDPG” with a coefficient of −0.8599 (p = 0.0101, p ≤ 0.05), indicating a negative long-run relationship. A 1% increase in “LSUK” is associated with a 0.85% decrease in “LGDPG” for the sample countries during this period. Thus, the hypothesis H2a is partially supported, as the relationship is negative. This finding aligns with study of Fatimah and Rahmayanti (2023) and others, but contradicts Bella and Inas (2023) who found no significant long-term impact. Our findings show Sukuk issuance peaked at $49 billion in 2007 but fell 35% in 2008 due to reduced demand, a credit freeze and waning investor confidence. Real estate declines, lower foreign investment and falling oil revenues further constrained Sukuk-financed projects, while fiscal stimulus measures offered little support, limiting long-term growth contributions.
Pooled mean group during-GFC
| PMG/ARDL model (1,1,1,1,1,1,1) | ||||
|---|---|---|---|---|
| Variable | Coefficient | Std. error | t-statistic | Prob** |
| Long run equation | ||||
| LSUK | −0.8599 | 0.3155 | −2.7253 | 0.0101** |
| LGFCF | 9.6175 | 3.0160 | 3.1887 | 0.0031** |
| LGGFE | 17.202 | 4.9183 | 3.4977 | 0.0013** |
| LINF | 6.1681 | 1.3306 | 4.6354 | 0.0001** |
| LTOPN | −26.348 | 7.6977 | −3.4229 | 0.0016** |
| LFDI | 115.74 | 34.289 | 3.3756 | 0.0019** |
| Short run equation | ||||
| COINTEQ01 | −0.0678 | 0.0250 | −2.7078 | 0.0105** |
| D(LSUK) | −1.0697 | 0.7999 | −1.3371 | 0.1901 |
| D(LGFCF) | 5.9847 | 14.999 | 0.3989 | 0.6924 |
| D(LGGFE) | −72.216 | 51.268 | −1.4086 | 0.1680 |
| D(LINF) | −5.9133 | 6.5606 | −0.9013 | 0.3738 |
| D(LTOPN) | 18.774 | 17.199 | 1.0915 | 0.2827 |
| D(LFDI) | −58.189 | 39.778 | −1.4628 | 0.1527 |
| PMG/ARDL model (1,1,1,1,1,1,1) | ||||
|---|---|---|---|---|
| Variable | Coefficient | Std. error | t-statistic | Prob** |
| Long run equation | ||||
| LSUK | −0.8599 | 0.3155 | −2.7253 | 0.0101** |
| LGFCF | 9.6175 | 3.0160 | 3.1887 | 0.0031** |
| LGGFE | 17.202 | 4.9183 | 3.4977 | 0.0013** |
| LINF | 6.1681 | 1.3306 | 4.6354 | 0.0001** |
| LTOPN | −26.348 | 7.6977 | −3.4229 | 0.0016** |
| LFDI | 115.74 | 34.289 | 3.3756 | 0.0019** |
| Short run equation | ||||
| COINTEQ01 | −0.0678 | 0.0250 | −2.7078 | 0.0105** |
| D(LSUK) | −1.0697 | 0.7999 | −1.3371 | 0.1901 |
| D(LGFCF) | 5.9847 | 14.999 | 0.3989 | 0.6924 |
| D(LGGFE) | −72.216 | 51.268 | −1.4086 | 0.1680 |
| D(LINF) | −5.9133 | 6.5606 | −0.9013 | 0.3738 |
| D(LTOPN) | 18.774 | 17.199 | 1.0915 | 0.2827 |
| D(LFDI) | −58.189 | 39.778 | −1.4628 | 0.1527 |
Note(s): This table presents PMG-ARDL (1,1,1,1,1,1,1) outcomes for the during-GFC period. It reports long-run coefficients along with short-run dynamics captured by the error correction term and first-differenced explanatory variables. Statistical significance at the 5% level is indicated by**
The short-run equation reveals that during the crisis period, “LSUK” has a negative coefficient of −1.0697 but an insignificant p-value of 0.1901, showing no significant short-run relationship with GDP growth. Thus, the hypothesis H2b is not supported. This aligns with Fuadi et al. (2022), Yıldırım et al. (2020), Kartini and Milawati (2020) who also found no significant short-term impact of Sukuk on economic growth. These findings suggest that delayed Sukuk returns, market disruptions and liquidity shortages fostered risk-averse investor behavior, limiting Sukuk’s short-term economic impact.
Table 8 depicts the PMG/ARDL estimation for the Post-GFC period. The optimal model, identified as (2,4,4,4,4,4,4), reveals that “LSUK” has a positive and statistically significant long-run impact on “LGDPG” (coefficient = 0.2071, p = 0.0000, p ≤ 0.05). This indicates that a 1% increase in “LSUK” is associated with a 0.2071% increase in “LGDPG” for the selected Islamic countries post-crisis. These findings support the hypothesis H3a. The results align with studies emphasizing Sukuk’s positive contribution to economic growth, such as Fatimah and Rahmayanti (2023), Novitasari and Arundina (2023) while contrasting with research by Bella and Inas (2023), and similar studies arguing no significant long-term impact. The GFC triggered a surge in Sukuk issuance, peaking at $137 billion in 2012, as it funded large-scale projects in the GCC and Malaysia. Sukuk supported infrastructure and Vision-2030 initiatives, enhancing economic resilience, job creation, financial inclusion and Sharia-compliant investment opportunities.
Pooled mean group post-GFC
| PMG/ARDL model (2,4,4,4,4,4,4) | ||||
|---|---|---|---|---|
| Variable | Coefficient | Std. error | t-statistic | Prob** |
| Long run equation | ||||
| LSUK | 0.2071 | 0.0456 | 4.5432 | 0.0000** |
| LGFCF | −0.0556 | 0.2822 | −0.1973 | 0.8438 |
| LGGFE | −0.0615 | 0.4544 | −0.1354 | 0.8924 |
| LINF | 1.0175 | 0.0327 | 31.115 | 0.0000** |
| LTOPN | 2.6461 | 0.2686 | 9.8514 | 0.0000** |
| LFDI | −11.387 | 2.0141 | −5.6536 | 0.0000** |
| Short run equation | ||||
| COINTEQ01 | −0.4375 | 0.1312 | −3.3336 | 0.0010** |
| D(LSUK) | −0.0786 | 0.1954 | −0.4024 | 0.6878 |
| D(LGFCF) | 4.7011 | 3.0813 | 1.5256 | 0.1287 |
| D(LGGFE(-2)) | 2.3903 | 1.0521 | 2.2718 | 0.0242** |
| D(LINF(-2)) | −0.1608 | 0.0546 | −2.9443 | 0.0036** |
| D(LTOPN) | 6.0985 | 1.9323 | 3.1560 | 0.0019** |
| D(LFDI(-1)) | 12.748 | 6.3956 | 1.9932 | 0.0476** |
| PMG/ARDL model (2,4,4,4,4,4,4) | ||||
|---|---|---|---|---|
| Variable | Coefficient | Std. error | t-statistic | Prob** |
| Long run equation | ||||
| LSUK | 0.2071 | 0.0456 | 4.5432 | 0.0000** |
| LGFCF | −0.0556 | 0.2822 | −0.1973 | 0.8438 |
| LGGFE | −0.0615 | 0.4544 | −0.1354 | 0.8924 |
| LINF | 1.0175 | 0.0327 | 31.115 | 0.0000** |
| LTOPN | 2.6461 | 0.2686 | 9.8514 | 0.0000** |
| LFDI | −11.387 | 2.0141 | −5.6536 | 0.0000** |
| Short run equation | ||||
| COINTEQ01 | −0.4375 | 0.1312 | −3.3336 | 0.0010** |
| D(LSUK) | −0.0786 | 0.1954 | −0.4024 | 0.6878 |
| D(LGFCF) | 4.7011 | 3.0813 | 1.5256 | 0.1287 |
| D(LGGFE(-2)) | 2.3903 | 1.0521 | 2.2718 | 0.0242** |
| D(LINF(-2)) | −0.1608 | 0.0546 | −2.9443 | 0.0036** |
| D(LTOPN) | 6.0985 | 1.9323 | 3.1560 | 0.0019** |
| D(LFDI(-1)) | 12.748 | 6.3956 | 1.9932 | 0.0476** |
Note(s): This table presents PMG-ARDL (2,4,4,4,4,4,4) estimates for the post-GFC period, showing long-run coefficients and short-run dynamics via the error correction term and first-differenced explanatory variables. ** indicates 5% significance
The short run results underscore that “LSUK” has a negative and insignificant association with “LGDPG” (coefficient = −0.0786, p = 0.6878), suggesting no significant short-term relationship between Sukuk issuance and GDP growth during the Post-crisis period. Consequently, the hypothesis H3b is not supported. This aligns with findings by Fuadi et al. (2022), Yıldırım et al. (2020) which also report no significant short-term impact of Sukuk on economic growth. These short-run findings suggest that economic conditions, investor hesitancy, liquidity constraints and regulatory delays likely limited the rapid mobilization of Sukuk funds, restraining immediate economic impact.
Table 9 displays the PMG/ARDL estimation for the full sample period, with the optimal model identified as (4,8,8,8,8,8,8,8). The long-run equation reveals that “LSUK” has a positive and statistically significant impact on “LGDPG” (coefficient = 0.2270, p = 0.0000, p ≤ 0.05), indicating a 1% increase in “LSUK” is associated with a 0.227% rise in “LGDPG” in the long run. These results strongly support the hypothesis H4a. This aligns with studies by Novitasari and Arundina (2023), Gürbüz et al. (2023), Fatimah and Rahmayanti (2023) confirming a robust long-term relationship between Sukuk issuance and economic growth, while contrasting with findings by Bella and Inas (2023). These results show that Sukuk’s asset-backed structure enhances stability, investor confidence and risk mitigation while financing infrastructure, creating jobs and attracting ethical and Islamic investment. Instruments like Green Sukuk further promote sustainable growth.
Pooled mean group full sample
| PMG/ARDL model (4,8,8,8,8,8,8,8) | ||||
|---|---|---|---|---|
| Variable | Coefficient | Std. error | t-statistic | Prob** |
| Long run equation | ||||
| LSUK | 0.2270 | 0.0497 | 4.5624 | 0.0000** |
| LGFCF | 2.7912 | 0.4037 | 6.9127 | 0.0000** |
| LGGFE | 1.6730 | 0.5154 | 3.2458 | 0.0014** |
| LINF | 0.5645 | 0.0406 | 13.886 | 0.0000** |
| LTOPN | −1.2377 | 0.1843 | −6.7138 | 0.0000** |
| LFDI | 14.642 | 1.8053 | 8.1107 | 0.0000** |
| DUMMY_GFC | −0.5323 | 0.0629 | −8.4516 | 0.0000** |
| Short run equation | ||||
| COINTEQ01 | −0.3995 | 0.1577 | −2.5321 | 0.0121** |
| D(LSUK) | 0.0671 | 0.0993 | 0.6751 | 0.5004 |
| D(LGFCF) | 4.2471 | 2.5424 | 1.6704 | 0.0964 |
| D(LGGFE(-2)) | 2.8326 | 2.6373 | 1.0740 | 0.2841 |
| D(LINF(-6)) | 0.2954 | 0.1380 | 2.1398 | 0.0336** |
| D(LTOPN) | 3.5552 | 4.6708 | 0.7611 | 0.4475 |
| D(LFDI) | 1.8872 | 12.316 | 0.1532 | 0.8784 |
| D(DUMMY_GFC) | 0.0777 | 0.3711 | 0.2095 | 0.8342 |
| PMG/ARDL model (4,8,8,8,8,8,8,8) | ||||
|---|---|---|---|---|
| Variable | Coefficient | Std. error | t-statistic | Prob** |
| Long run equation | ||||
| LSUK | 0.2270 | 0.0497 | 4.5624 | 0.0000** |
| LGFCF | 2.7912 | 0.4037 | 6.9127 | 0.0000** |
| LGGFE | 1.6730 | 0.5154 | 3.2458 | 0.0014** |
| LINF | 0.5645 | 0.0406 | 13.886 | 0.0000** |
| LTOPN | −1.2377 | 0.1843 | −6.7138 | 0.0000** |
| LFDI | 14.642 | 1.8053 | 8.1107 | 0.0000** |
| DUMMY_GFC | −0.5323 | 0.0629 | −8.4516 | 0.0000** |
| Short run equation | ||||
| COINTEQ01 | −0.3995 | 0.1577 | −2.5321 | 0.0121** |
| D(LSUK) | 0.0671 | 0.0993 | 0.6751 | 0.5004 |
| D(LGFCF) | 4.2471 | 2.5424 | 1.6704 | 0.0964 |
| D(LGGFE(-2)) | 2.8326 | 2.6373 | 1.0740 | 0.2841 |
| D(LINF(-6)) | 0.2954 | 0.1380 | 2.1398 | 0.0336** |
| D(LTOPN) | 3.5552 | 4.6708 | 0.7611 | 0.4475 |
| D(LFDI) | 1.8872 | 12.316 | 0.1532 | 0.8784 |
| D(DUMMY_GFC) | 0.0777 | 0.3711 | 0.2095 | 0.8342 |
Note(s): This table shows PMG-ARDL (4,8,8,8,8,8,8,8) estimates for the full-sample period, reporting long-run coefficients and short-run adjustments through the error correction term and first- or lagged-differenced variables. Statistical significance at 5% is indicated by**
To assess the impact of the Global Financial Crisis (2007–2009) within the full sample period, a Dummy-GFC variable is introduced which exhibits a significant negative long-run relationship with GDP growth, with a coefficient of −0.5323 (p = 0.0000). The short-run equation demonstrates that during the full sample period, Sukuk issuance shows an insignificant short-term relationship with GDP growth, as evidenced by a positive coefficient of 0.0671 and a p-value of 0.5004. Thus, the findings do not support the hypothesis H4b. These results are supported by Fuadi et al. (2022) suggesting Sukuk has no significant short-term impact on economic growth. Short-run findings show Sukuk issuance has limited impact on economic growth due to global conditions, regulatory challenges, market misconceptions and macroeconomic factors, with project delays further restricting short-term benefits.
4.4 Robustness tests
To ensure robustness of the crisis window (During-GFC: 2007Q1–2009Q4), additional estimations were performed using FMOLS and DOLS. Given the study’s focus on crisis dynamics, these checks reinforce the validity of the results for this period. As the pre-crisis, post-crisis and full-sample models already exhibited stable outcomes under diagnostic tests Johansen Fisher Panel Cointegration along with Kao Residual Cointegration Test and PMG-ARDL, further robustness testing for those periods was deemed unnecessary.
In Table 10, the FMOLS results reveal that Sukuk issuance “LSUK” exerts a negative but marginally significant long-run effect on GDP growth “LGDPG” during the GFC sub-period. Specifically, a 1% increase in Sukuk issuance is associated with approximately a 2% decline in GDP growth (coefficient = −2.01, p < 0.10). This suggests that Sukuk issuance during the crisis years functioned more as a stabilizing or refinancing instrument rather than as a direct driver of real economic activity. These findings are consistent with the PMG-ARDL thus reinforcing the persistence of long-run cointegration as also evidenced by the Kao panel test.
Fully modified ordinary least square during-GFC
| Variable | Coefficient | Std. error | t-statistic | Prob |
|---|---|---|---|---|
| LSUK | −2.0122 | 1.0598 | −1.8987 | 0.0615* |
| LGFCF | 30.5039 | 3.1246 | 9.7624 | 0.0000** |
| LGGFE | −38.9457 | 5.9543 | −6.5408 | 0.0000** |
| LINF | 1.9893 | 3.5908 | 0.5540 | 0.5812 |
| LTOPN | −27.5799 | 4.4501 | −6.1977 | 0.0000** |
| LFDI | −24.5337 | 34.9879 | −0.7012 | 0.4854 |
| Variable | Coefficient | Std. error | t-statistic | Prob |
|---|---|---|---|---|
| LSUK | −2.0122 | 1.0598 | −1.8987 | 0.0615* |
| LGFCF | 30.5039 | 3.1246 | 9.7624 | 0.0000** |
| LGGFE | −38.9457 | 5.9543 | −6.5408 | 0.0000** |
| LINF | 1.9893 | 3.5908 | 0.5540 | 0.5812 |
| LTOPN | −27.5799 | 4.4501 | −6.1977 | 0.0000** |
| LFDI | −24.5337 | 34.9879 | −0.7012 | 0.4854 |
Note(s): The table reports FMOLS estimates for the during-GFC period, coefficients, standard errors, t-statistics and p-values for the explanatory variables are presented, with significance at 10 and 5% denoted by * and **, respectively
Table 11 reveal DOLS results that Sukuk issuance “LSUK” also employs a negative and marginally significant long-run impact on GDP growth “LGDPG” during the GFC sub-period. Specifically, a 1% rise in Sukuk issuance is associated with approximately a 1.8% decline in GDP growth (coefficient = −1.84, p < 0.10). This finding implies that, similar to FMOLS, Sukuk served primarily as a refinancing or liquidity-support instrument in crisis years rather than directly fostering real sector expansion. The consistency of these results across DOLS, FMOLS and PMG-ARDL estimations strengthens the robustness of the evidence for long-run cointegration, as further confirmed by the Kao panel test.
Dynamic ordinary least square during-GFC
| Variable | Coefficient | Std. ERROR | t-statistic | Prob. |
|---|---|---|---|---|
| LSUK | −0.293186 | 0.092992 | −3.152797 | 0.0022** |
| LGFCF | 0.362718 | 0.250471 | 1.448141 | 0.1511 |
| LGGFE | −0.186990 | 0.249131 | −0.750568 | 0.4549 |
| LINF | 0.390215 | 0.134160 | 2.908591 | 0.0046** |
| LTOPN | −0.498259 | 0.435671 | −1.143660 | 0.2558 |
| LFDI | 7.562386 | 2.784607 | 2.715783 | 0.0079** |
| C | 3.609100 | 1.506876 | 2.395087 | 0.0187** |
| Variable | Coefficient | Std. ERROR | t-statistic | Prob. |
|---|---|---|---|---|
| LSUK | −0.293186 | 0.092992 | −3.152797 | 0.0022** |
| LGFCF | 0.362718 | 0.250471 | 1.448141 | 0.1511 |
| LGGFE | −0.186990 | 0.249131 | −0.750568 | 0.4549 |
| LINF | 0.390215 | 0.134160 | 2.908591 | 0.0046** |
| LTOPN | −0.498259 | 0.435671 | −1.143660 | 0.2558 |
| LFDI | 7.562386 | 2.784607 | 2.715783 | 0.0079** |
| C | 3.609100 | 1.506876 | 2.395087 | 0.0187** |
Note(s): This table reports DOLS estimates for the during-GFC period, with LGDPG as the dependent variable. It presents coefficients, standard errors, t-statistics and p-values for the explanatory variables, with significance at 5% denoted by**
4.5 Panel causality test
The study applies the Dumitrescu–Hurlin (2012) pairwise panel causality test to examine causal directions among the variables, whether unidirectional, bidirectional or absent, aligning with recent applications in panel causality research (Naz and Gulzar, 2022a). As an extension of the traditional Granger causality test, this approach incorporates cross-sectional dependence and heterogeneity in both the regression framework and causal dynamics. Table 12 presents the outcomes for the Pre-, During- and Post-GFC periods, as well as the full sample, with corresponding test statistics and probabilities under the null hypothesis.
Pairwise Dumitrescu Hurlin panel causality test
| Null hypothesis | W-stat | Z-stat. | Prob. | Direction |
|---|---|---|---|---|
| Pre-GFC | ||||
| LSUK does not homogeneously cause LGDPG | 1.7915 | −0.5184 | 0.6042 | LGDPG → LSUK |
| LGDPG does not homogeneously cause LSUK | 8.4841 | 6.7845 | 1.E−11** | |
| During-GFC | ||||
| LSUK does not homogeneously cause LGDPG | 3.3335 | 2.2680 | 0.0233** | LSUK ↔ LGDPG |
| LGDPG does not homogeneously cause LSUK | 6.8553 | 6.2614 | 4.E−10** | |
| Post -GFC | ||||
| LSUK does not homogeneously cause LGDPG | 7.3932 | −1.9228 | 0.0545** | LSUK ↔ LGDPG |
| LGDPG does not homogeneously cause LSUK | 5.3812 | −2.6994 | 0.0069** | |
| Full sample | ||||
| LSUK does not homogeneously cause LGDPG | 6.4779 | −2.14126 | 0.0323** | LSUK ↔ LGDPG |
| LGDPG does not homogeneously cause LSUK | 6.5580 | −2.09709 | 0.0360** | |
| Null hypothesis | W-stat | Z-stat. | Prob. | Direction |
|---|---|---|---|---|
| Pre-GFC | ||||
| LSUK does not homogeneously cause LGDPG | 1.7915 | −0.5184 | 0.6042 | LGDPG → LSUK |
| LGDPG does not homogeneously cause LSUK | 8.4841 | 6.7845 | 1.E−11** | |
| During-GFC | ||||
| LSUK does not homogeneously cause LGDPG | 3.3335 | 2.2680 | 0.0233** | LSUK ↔ LGDPG |
| LGDPG does not homogeneously cause LSUK | 6.8553 | 6.2614 | 4.E−10** | |
| Post -GFC | ||||
| LSUK does not homogeneously cause LGDPG | 7.3932 | −1.9228 | 0.0545** | LSUK ↔ LGDPG |
| LGDPG does not homogeneously cause LSUK | 5.3812 | −2.6994 | 0.0069** | |
| Full sample | ||||
| LSUK does not homogeneously cause LGDPG | 6.4779 | −2.14126 | 0.0323** | LSUK ↔ LGDPG |
| LGDPG does not homogeneously cause LSUK | 6.5580 | −2.09709 | 0.0360** | |
Note(s): This table presents causality results across pre-, during-, post-GFC and full-sample periods, showing W- and Z-statistics, p-values and causality direction. Significance at the 5% level is indicated by**
The Panel causality tests across the Pre-GFC, During-GFC, Post-GFC and Full sample periods reveal varying causal relationships between Sukuk issuance and GDP growth in selected Islamic countries. In the Pre-GFC period, a unidirectional causal relationship was identified from “LGDPG” to “LSUK” supporting Robinson’s “Demand-following Hypothesis” which posits that economic growth drives Sukuk development, consistent with study of Gürbüz et al. (2023). In contrast, the During-GFC, Post-GFC and Full sample periods indicate a bi-directional causal relationship among “LSUK” and “LGDPG” supporting Patrick’s “Feedback Hypothesis” suggests that Sukuk issuance and economic growth mutually reinforce each other in congruence with the result of Novitasari and Arundina (2023), Fatimah and Rahmayanti (2023).
5. Conclusion
The Global Financial Crisis (2007–2009) highlighted the resilience of Sukuk, revealing its potential to support long-term economic growth despite short-term challenges. This research examined the link between Sukuk issuance and economic growth across the Pre-, During- and Post-GFC phases, as well as the full sample, in major Sukuk-issuing economies. Findings indicate that Sukuk significantly drives long-term growth, especially before and after the crisis, by funding infrastructure and development projects. During the GFC, Sukuk’s impact was constrained due to global financial instability. Short-term impacts were less significant, as the benefits of Sukuk-funded projects take time to materialize. The causality analysis indicated a one-way relationship from GDP growth to Sukuk issuance prior to the GFC, consistent with the “Demand Following Hypothesis”. In contrast, during, after the crisis and in full sample, bidirectional causality was observed, supporting the “Feedback Hypothesis”, indicating a mutually reinforcing relationship among Sukuk issuance and economic growth.
Sukuk’s sustained contribution to growth underscores its value as a long-term development tool. Strengthening its role could involve adjusting central bank capital-adequacy rules, allowing wider use as collateral, and improving market depth through uniform Shariah standards, active market-making and secondary-market guarantees, as seen in Malaysia and Indonesia. Enhancing investor understanding via financial literacy programs, and broadening appeal through innovative risk-sharing structures and Green Sukuk for sustainable projects, can further support growth. Internationally, regional platforms like ASEAN or OIC Sukuk clearinghouses could standardize issuance, ease cross-border settlement and expand the investor base. Together, these measures provide a practical roadmap to make Sukuk markets more resilient and responsive in both stable and turbulent conditions.
This study has several limitations. The during-GFC phase covers only 12 quarterly observations (36 months), which may limit capturing long-run dynamics; however, this falls within the 32–64 months horizon considered long run in prior studies (e.g. Hayat et al., 2021), and cointegration tests (Kao) and PMG-ARDL, FMOLS and DOLS estimates confirm a long-run relationship. Additional limitations include limited historical Sukuk data and cross-country heterogeneity in issuance patterns, regulatory frameworks, market maturity and reporting standards, which may affect comparability.
Future research should explore sector-specific impacts, compare Sukuk’s effects across different countries and investigate innovative Sukuk structures. Comparative studies with conventional financial instruments will offer valuable insights into Sukuk’s contribution to economic growth.

