Stock Market Development and Economic Growth Nexus: Evidence from the Fragile Five Countries

Abstract

In emerging markets, stock markets play a crucial role in supporting long-term growth. This study explores the causal relationship between stock market development and economic growth in the Fragile Five countries—Brazil, India, Indonesia, South Africa, and Turkey—covering the period from 2001 to 2024. To ensure robust findings, it uses second-generation panel cointegration and causality tests that account for cross-sectional dependence and structural heterogeneity. The model includes three parameters representing financial depth, liquidity, and efficiency. Results indicate significant long-term cointegration, suggesting causality from stock market development to economic growth, supporting the supply-led growth hypothesis. This aligns with recent evidence highlighting the importance of institutional quality and sectoral interconnectedness in emerging markets. Furthermore, Panel DOLS and FMOLS analyses reveal that stock market capitalization has a notable positive effect on domestic productivity. Overall, these findings underscore that stock market parameters are vital for accurate economic forecasting and that strengthening capital markets is essential for sustainable growth in the Fragile Five.

1. Introduction

Emerging market economies have attracted the attention of researchers and economists as their per capita incomes have increased, partly due to their resilience during the 2008 global financial crisis, which affected countries differently. The term “Fragile Five” was coined by Morgan Stanley in May 2013 following the US Federal Reserve’s (FED) announcement to taper its bond purchases. It referred to Brazil, India, Indonesia, South Africa, and Turkey, whose currencies suffered the largest declines, and the term has since become widely used. The primary reason for the fragility of these countries is the expectation that foreign capital flows to highly dependent nations will decline after the FED ends its monetary easing policy. This expectation has, on one hand, driven up interest rates and, on the other hand, caused their currencies to depreciate quickly, raising financing costs. In November 2017, credit rating agency Standard & Poor’s (S&P) released its “New Fragile Five” report, highlighting high interest rates that increase borrowing costs and identifying Argentina, Qatar, Egypt, Pakistan, and Turkey as the new group.

The high interest rates have negatively impacted the real sector, which aims to borrow long-term—mainly from banks—to fund new investments or expand existing ones, particularly to meet external financing needs. Consequently, sustainable economic growth has been impeded. Coşkun et al. (2017) noted that, in developing countries like Turkey, capital markets contribute less to economic growth than the banking sector. As many nations transition from a focus on commercial banking to a production model based on technology and knowledge, it is crucial for the stock market—an alternative to interest rates and part of the capital market—to grow and become more efficient.

The 21st century has seen rapid changes in emerging market stock exchanges. Key factors such as financial liberalization, globalization, and technological advances have created networks that enhance market integration and facilitate trade growth. Around 20 years ago, many European companies preferred listing on U.S. exchanges over domestic ones due to less stringent accounting standards, weaker shareholder protections, and higher listing costs. Rajan and Zingales (2003) observed a shift from a banking-centric to a capital-oriented model within the European financial system, predicting that this trend would persist. Conversely, Kandil et al. (2015) reported that, aside from the UK, most European countries still primarily maintain banking-based financial systems.

The proliferation of financial liberalization policies has led to increased foreign capital flows into developing economies; for these economies, it has begun to play a significant role in stock market and economic growth indicators. Levine and Zervos (1996) and Demirgüç-Kunt and Levine (1996) have found a causal relationship between stock market indicators and growth. The development of the stock market facilitates investors’ access to capital by encouraging efficient resource allocation. Therefore, it enables rapid increases in domestic and foreign investment. A well-functioning stock market can provide substantial support for sustainable economic development by making the national economy attractive to foreign investors (Carp, 2012). Theoretically, a developed stock market is expected to support economic growth by promoting savings and reducing transaction costs, thereby enabling efficient resource allocation to investments. Meanwhile, risk sharing increases liquidity and reduces the cost of capital allocated to investments (Cooray, 2010). The impact of emerging stock markets on growth is transmitted to the real sector through specific channels, primarily liquidity, market value, risk sharing, and diversification. The study by Bencivenga et al. (1996) provides evidence that economic growth is positively affected by increased stock market liquidity, mainly due to improvements in business information and corporate governance.

The relationship between financial liberalization, the stock market, and economic growth has been discussed in the literature (Goldsmith, 1969; Shaw, 1973; Helhel, 2017). The theory of the “supply-led hypothesis” emphasizes the positive contribution of financial development to economic growth with financial openness at the forefront. The view that increased efficiency and production resulting from investment and innovation in the real sector, as proposed by Robinson (1952) and Patrick (1966), leads to the development of financial markets is known as the “demand-led hypothesis” in economics and finance literature. It is possible to find many studies that conclude that the interaction results from the combination of these two hypotheses, in other words, “financial development is the result of economic growth, and financial development is also a determinant of economic growth.” (A. E. Akinlo & Egbetunde, 2010; Hou & Cheng, 2010; Ndako, 2010; Marques et al., 2013).

Recent empirical evidence further explores the multidimensional impact of financial development on various economic sectors. Konstantakopoulou (2025) utilized a comprehensive panel of 126 countries to demonstrate that financial depth and efficiency are critical catalysts for sectoral growth, emphasizing the role of stock market liquidity in long-term development. In the context of emerging markets, Kwabi et al. (2024) highlighted how institutional quality moderates the negative effects of political uncertainty on stock market size and transaction costs. Furthermore, research on the Indian stock market by Chatziantoniou et al. (2021) identified dynamic connectedness across financial and industrial sectors, while Kök and Nazlıoğlu (2022) confirmed significant causality between stock market performance and energy security in the BRICS-T group. These recent studies collectively reinforce the need to use advanced panel methodologies to account for cross-sectional dependence and structural heterogeneity in the finance–growth nexus.

This study investigates whether there is a long-term interaction between the stock market, a submarket of capital markets, and economic growth, and if so, the direction and degree of this interaction for the Fragile Five group, as named by Morgan Stanley, using annual data from 2001 to 2024 and panel data analysis methods. While this study focuses exclusively on the relationship between stock market development indicators and economic growth, we acknowledge that other macroeconomic factors, such as inflation, trade openness, and foreign direct investment, also play vital roles in economic performance. Our choice of a more focused model was intended to prevent multicollinearity and to provide a clear, isolated analysis of capital market depth and efficiency. Future studies could expand this framework by incorporating a broader set of control variables to validate these findings across different emerging-market classifications further. The primary objective of this study is to determine whether stock market development acts as a leading indicator of economic growth in the Fragile Five countries. Specifically, we investigate the causal links between market depth, liquidity, and GDP per capita to test the validity of the supply-led growth hypothesis. By employing second-generation panel methodologies, this research aims to provide a robust empirical roadmap that accounts for cross-sectional dependence and structural heterogeneity. Consequently, the study seeks to determine whether strengthening capital markets can serve as a sustainable policy tool for these emerging economies.

Pedroni and Kao cointegration tests indicated a long-term relationship between variables. The direction of this relationship was identified using the VECM causality test, while DOLS and FMOLS developed by Pedroni (2001) were used to estimate the long-term coefficients. Unlike other studies in the literature on financial development and economic growth, this research highlights key differences: First, in developing countries, research mainly focuses on banking-related variables because their financial markets are primarily composed of banks and lending institutions. As a result, the relationship between capital markets, their components, and economic growth has been less explored. Levine and Zervos (1998) and Rousseau and Wachtel (2000) noted that even if the share of the real sector going public is low, a robust and organized capital market still significantly contributes to economic development. Second, most research emphasizes the stock market’s impact on growth, with few studies, especially in developing nations, analyzing the causality direction between the stock market and economic growth (Rousseau & Wachtel, 2000; Shahbaz et al., 2008; A. E. Akinlo & Egbetunde, 2010; Hou & Cheng, 2010; Hoque et al., 2017). Third, some existing studies have looked at the link between stock market indices and economic growth. This study, however, uses three measures of stock market financial development from the World Bank’s Global Financial Development Database, which categorizes indicators into four main areas: depth, access, stability, and efficiency (Demirgüç-Kunt & Maksimovic, 1996; Boubakari & Jin, 2010).

Organization of the Manuscript: The remainder of the paper is organized as follows. Section 2 reviews the relevant literature and develops the hypotheses. Section 3 describes the data and methodology. Section 4 presents the empirical findings. Section 5 discusses the results, and concludings the study.

2. Literature Review

Empirical studies examining the direction and effect of the relationship between the stock market and economic growth have yielded very different results across developed and developing countries, due to differences in the models, variables, and time periods used, and no consensus has emerged. One of the first studies to examine only the relationship between the stock market and economic growth, the analysis by Atje and Jovanovic (1993) used data from 40 countries and found that the stock market affects growth through four factors: liquidity, corporate governance, risk diversification, and information acquisition. Levine and Zervos (1998) demonstrated that increased liquidity in the stock market over the 1973–1993 period for 48 countries positively affected long-term growth by increasing capital accumulation and, consequently, productivity. Studies analyzing the relationship between the stock and banking markets and per capita national income, an indicator of economic growth, have produced more accurate growth forecasts (Rousseau & Wachtel, 2000; Beck & Levine, 2002). Harris (1997), Arestis et al. (2001) and Caporale et al. (2004) have reached conclusions supporting the supply-led hypothesis approach, demonstrating that a developed stock market triggers economic growth.

Various studies in the literature find no consistent relationship between the stock market and economic growth. Some even suggest that developing stock markets may hinder growth. Singh (1997) argued that stock market volatility could harm growth, particularly in developing countries, due to macroeconomic instabilities. Devereux and Smith (1994) pointed out that risk-sharing through stock market integration might negatively affect growth in these nations. Boubakari and Jin (2010) found that small stock markets with low trading volumes do not impact their countries’ economic growth. Other research indicates that the stock market capitalization ratio and liquidity can promote long-term growth but may slow short-term growth (Wu et al., 2010; Dell’Ariccia & Marquez, 2006). The GMM (Generalized Method of Moments) analysis by Naceur and Ghazouani (2007), based on data from 11 MENA countries—including the Middle East and North Africa—found no significant link between banking, capital markets, and economic growth.

On the other hand, studies have also found a two-way relationship between the stock market and growth. Hou and Cheng (2010) used quarterly data from Taiwan for the period 1971–2007 to demonstrate a long-term relationship between the banking sector, the stock market, and growth. According to the researchers, the stock market contributes more to growth, and the VECM Granger causality test indicates a two-way causal relationship between the stock market and growth. This result supports Patrick’s (1966) view that “financial development is a result of economic growth, and financial development is also a determinant of economic growth”. Marques et al. (2013) analyzed the relationship between the banking sector, the stock market, and growth in Portugal using quarterly data from 1993–2011 and a VAR Model. The analysis results show that there is a two-way causal relationship between the stock market and growth, and no evidence of a causal relationship from banking development to growth. In the study by A. E. Akinlo and Egbetunde (2010), the ARDL bounds test was applied to sub-Saharan African countries to examine the relationship between the stock market and economic growth, yielding different results across countries. The results of the VECM Granger causality test revealed a unidirectional relationship from the stock market to growth in South Africa and Egypt, a bidirectional relationship in Côte d’Ivoire, Kenya, Morocco, and Zimbabwe, and a weak relationship in Nigeria. According to the authors, it is crucial to implement macroeconomic policies that support the development of the stock market to stimulate growth in these countries.

3. Data, Model, and Hypothesis

In practice, the stock market development criteria for the group known as the Fragile Five (Brazil, Indonesia, South Africa, India, and Turkey) include the “market capitalization ratio,” which represents market depth; the “stock turnover rate,” which represents efficiency; and the “stock transaction value,” which represents liquidity. To determine the variables, the conceptual, yet relatively simple, framework developed by the World Bank for its Global Financial Development Database was used. These measures have also been used by Demirgüç-Kunt and Maksimovic (1996), Levine and Zervos (1996), and Cooray (2010), in their studies evaluating the relationship between the stock market and growth.

Table 1. Descriptive Statistics (2001–2024).

Variables	Mean	Std. Dev.	Minimum	Maximum
MC (%)	85.25	78.92	14.00	323.00
ST (%)	87.67	79.99	15.00	323.00
TR (%)	79.12	72.72	18.00	389.00
GDPpc ( )	5684.43	3456.37	779.00	14,933.00
lnPC	8.40	0.77	6.66	9.61

shows descriptive statistics for the variables. The analysis reveals significant differences. There is a high volatility in the stock market development criteria across the examined group.

Before proceeding with the empirical analysis, it is essential to define the baseline econometric model and the functional forms of the variables. The relationship between stock market development and economic growth is estimated using the following panel specification:

$$ln{PC}_{it}={\beta }_0+{\beta }_1{.MC}_{it}+{\beta }_2{.ST}_{it}+{\beta }_3{.TR}_{it}+{\beta \epsilon }_{it}$$

In this model, the dependent variable lnPC represents the natural logarithm of real GDP per capita, serving as a proxy for economic growth and long-term living standards. The explanatory variables are selected to capture different dimensions of stock market development: Market Capitalization (MC) reflects financial depth, Stock Turnover Ratio (ST) represents market liquidity, and Total Value Traded (TR) acts as a measure of market efficiency. To ensure cross-country comparability and minimize scale effects, all three stock market indicators are expressed as a percentage of GDP. This framework enables a robust examination of the supply-led growth hypothesis in the context of the Fragile Five economies.

The variables are defined as follows:

MC = Ratio of the total market value of selected domestic companies traded on domestic stock exchanges to national income.

ST = Ratio of the transaction value of domestic and foreign shares traded on domestic stock exchanges to national income.

TR = Ratio of the trading volume of shares bought and sold on domestic stock exchanges to the total share value (market capitalization) registered on the stock exchange.

Per capita national income level was taken as the economic growth variable and included in the analysis after logarithmic transformation (lnPC). We used the natural logarithm of per capita GDP as our proxy for growth. This variable reflects long-term living standards and economic development. Logarithmic transformation helps to reduce heteroscedasticity and stabilize the variance. Many previous studies also use this proxy for consistent cross-country comparisons. It allows us to examine the long-term impact of financial depth. This measure is suitable for analyzing the sustainable development of Fragile Five economies. Therefore, our variable selection aligns with the study’s core research objectives.

With the increase in the dataset’s sample size (N = 115), the variables’ volatility (standard deviation) was more pronounced than in the previous study (Helhel, 2017; Helhel & Helhel, 2026). In particular, the differences in stock market depth (MC) and trading volume (TR) between South Africa and Turkey are striking. This expanded data set provides a stronger econometric basis for the next step: testing for unit roots and cointegration.

To empirically examine the causal nexus between stock market development and economic growth, the study specifies two functional panel models. Based on the theoretical frameworks of the “supply-led” and “demand-following” hypotheses, the relationship is modeled as follows:

$$lnPC={\alpha }_{0i}+{\alpha }_{1i}{SFD}_{it}+{\epsilon }_{it}$$

$${SFD}_{it}={\beta }_{0i}+{\beta }_{1i}{lnPC}_{it}+{\mu }_{it}$$

In these equations, i (1, 2, …, 5) represents the Fragile Five countries, while t(2001, 2002, …, 2024) denotes the time period. The term “${lnPC}_{it}$” signifies the natural logarithm of per capita GDP (in USD) as the proxy for economic growth. The term “${SFD}_{it}$” represents the stock market development measures (MC, ST, and TR). ${\alpha }_{0i}$ and ${\beta }_{0i}$ are country-specific fixed effects, while ${\epsilon }_{it}$ and ${\mu }_{it}$ represent the stochastic error terms.

4. Methodology and Findings

The study examines the relationship between stock markets and economic growth. First, cross-sectional dependence among the Fragile Five countries was tested. In panel data analysis, unit root tests are categorized into two generations. First-generation tests, such as the Levin, Lin, and Chu (LLC) and Im, Pesaran, and Shin (IPS) tests, assume cross-sectional independence. The LLC test is a pooled approach where the null hypothesis (H0) assumes that all series contain a unit root with a common autoregressive parameter. In contrast, the IPS test allows for heterogeneity by assuming that while some series may have unit roots, others may be stationary under the alternative hypothesis. Global shocks affect these five emerging economies similarly. Despite their foundational importance, these tests become inconsistent when cross-sectional dependence is present. Therefore, this study primarily relies on the second-generation CIPS test, which accounts for common factors by augmenting the Dickey–Fuller regressions with cross-sectional averages. We used the Pesaran (2004) CD test for this specific purpose. The results in

Table 2. Pesaran (2004) CD Test Results.

Variable	CD-Statistic	p-Value	Decision
MC	8.0504	0.0000	Dependence (Reject H0)
ST	7.4459	0.0000	Dependence (Reject H0)
TR	2.2987	0.0215	Dependence (Reject H0)
lnPC	12.7625	0.0000	Dependence (Reject H0)

show significant dependence for all variables. This finding makes first-generation unit root tests unreliable for our data.

Therefore, we applied the second-generation Pesaran (2007) CIPS unit root test. This test provides consistent results under cross-sectional dependence conditions. The CIPS test results are presented clearly in

Table 3. Pesaran (2007) CIPS Panel Unit Root Test Results.

Variable	Level (Constant)	First Difference (Δ)	Result
MC	−2142	−6.250 ***	I(1)
ST	−1985	−5.926 ***	I(1)
TR	−2014	−6.457 ***	I(1)
lnPC	−1624	−2.963 ***	I(1)

Note: *** indicates significance at the 1% level. Critical values for CIPS (N = 5, T = 24) are −2.85 (1%), −2.28 (5%), and −2.02 (10%) for models with an intercept-only.

below. All variables exhibit a unit root at their level. However, they become stationary after taking the first difference. Thus, all series are integrated of order I(1).

Following the unit root tests, we examined the long-term relationship. The Pedroni and Kao panel cointegration tests were performed accordingly. These tests determine if the variables move together over time. Our findings confirm a significant cointegration relationship among the variables. The stock market and economic growth are long-term linked.

Finally, we estimated the long-term coefficients using FMOLS and DOLS. These methods correct for endogeneity and serial correlation in panels. The results indicate that stock market capitalization positively impacts growth. Also, the turnover ratio contributes significantly to economic development. These findings support the supply-led growth hypothesis for these countries.

The stationarity of all variables was first tested using the CIPS method. We found that lnPC, MC, ST, and TR are non-stationary at levels. However, all variables became stationary at their first difference levels. This means that the variables are integrated of order one, or I(1). These results provide a strong basis for the following cointegration analysis.

In the next stage, we conducted the Pedroni panel cointegration test. This test identifies long-term relationships between variables across different panel sections. We defined the null hypothesis as no cointegration for all cross-sections. A rejected null hypothesis indicates that a long-term relationship exists between the variables.

We also applied the Kao Cointegration Test to ensure robustness. The results of CIPS tests are shown in Table 3. Most statistics indicate cointegration among all examined variables. Specifically, stock market depth and liquidity move together with economic growth.

Our findings are consistent with the studies of Puryan (2017) and Kandil et al. (2015). These results confirm that stock market parameters are linked to sustainable growth. Therefore, we can proceed to estimate the long-term coefficients for the models.

Pedroni (1999) developed seven tests for cointegration within a panel data framework. Four of these tests are referred to as within-dimension tests, assuming that the variables in the tests are cointegrated. The residuals from the four tests are combined across the panel’s time dimension to estimate the autoregressive coefficients. The other three between-dimension tests are based on the mean coefficient estimate, estimated individually for each cross-section unit.

Another cointegration test used in the study was developed by Kao (1999); Kao and Chiang (2000) and is based on the DF (Dickey–Fuller) and ADF (Augmented Dickey–Fuller) tests. Kao used the sequential limit theory to analyze the asymptotic distributions of four different DF tests.

Table 4. Pedroni and Kao Cointegration Test Results (2001–2024).

Test Statistics	lnPC−MC	lnPC−ST	lnPC−TR
Pedroni Cointegration
Panel v-Statistic	1.985 ** (0.023)	3.214 * (0.001)	0.412 (0.340)
Panel rho-Statistic	−2.561 * (0.005)	−4.110 * (0.000)	1.854 ** (0.032)
Panel PP-Statistic	−3.422 * (0.000)	−3.982 * (0.000)	−3.102 * (0.001)
Panel ADF-Statistic	−3.810 * (0.000)	−4.025 * (0.000)	−2.945 * (0.002)
Group rho-Statistic	−1.892 ** (0.029)	−3.220 * (0.001)	−0.612 (0.270)
Group PP-Statistic	−3.905 * (0.000)	−6.115 * (0.000)	−4.218 * (0.000)
Group ADF-Statistic	−3.712 * (0.000)	−6.230 * (0.000)	−4.110 * (0.000)
Kao Cointegration
ADF Statistic	−2.854 * (0.002)	−2.712 * (0.003)	1.522 (0.064)

Note 1: * and ** indicate significance at the 1% and 5% levels, respectively. Note 2: Values in parentheses ( ) represent p-probability values. Note 3: Lag length selection was based on the Schwarz Information Criterion (SIC).

shows the results of both tests. Overall, the Pedroni cointegration test results indicate that the majority of the seven statistics indicate cointegration among all three variables. Kao’s cointegration results, on the other hand, support the existence of a cointegration relationship, except for the stock trading volume (TR). In this context, it is accepted that the stock development variables are cointegrated with the growth variable. This result is consistent with the findings of Puryan (2017), T. Akinlo (2016/2016), Wu et al. (2010) and Kandil et al. (2015), but is inconsistent with the cointegration results of Kandil et al. (2015) and Gözbaşı (2015).

Cointegration Presence: The results of the Pedroni and Kao cointegration tests strongly support the existence of a long-term relationship between the variables examined (see Table 4). In our analysis, although the cross-sectional dimension is limited to five countries (N = 5), the relatively long time series (T = 24) provides sufficient degrees of freedom. To mitigate potential small-sample bias associated with asymptotic tests, we used both the Pedroni and Kao tests to assess the consistency of the long-term relationship before proceeding with DOLS and FMOLS estimations.

Statistical Significance: The vast majority of the seven Pedroni statistics are significant at the 1% and 5% levels, particularly in the relationship between Market Value ($lnPC-MC$) and Transaction Value ($lnPC-ST$). This indicates that stock market depth and liquidity move in tandem with economic growth.

Exceptions: While some statistics are weak in the Stock Turnover Ratio ($lnPC-TR)$ variable, the within-group and between-group ADF/PP tests generally indicate the presence of cointegration. Consistency with Literature: These results are consistent with studies by Kandil et al. (2015).

According to Granger (1988), if a cointegrated vector is detected between variables, there must be at least unidirectional causality between the variables in question. If the variables are first-order stationary and cointegrated, it is recommended to use the VECM (vector error-correction model) rather than the panel VAR (vector autoregression), which does not account for the error term that corrects short-run imbalances (Anwar & Nguyen, 2018). This method is advantageous because it allows distinguishing the direction of causality in the short and long term. The significance of the F-statistic of the lagged change in the independent variables measures short-term causality. In contrast, long-term causality is measured by the t-statistic significance of the ECTs. ECT represents the long-term dynamics and refers to the error correction term derived from cointegrated equations. The number of lags must be determined in the estimation process. The Schwartz–Bayesian Criterion (SBC), which is the most frequently used in the literature, was preferred to determine the optimal lag length in the study. The potential for endogeneity—often stemming from bidirectional causality or omitted factors—is a common challenge in growth models. To provide unbiased and consistent coefficients, this study prioritized FMOLS and DOLS estimators. These methods are superior to standard OLS because they adjust for endogeneity in the regressors and nuisance parameters, ensuring that the identified impact of market capitalization on growth is statistically reliable. Selecting the optimal lag length is a critical step before performing the VECM causality test. In this study, we primarily used the Schwarz–Bayesian Criterion (SBC) to determine the lags. To ensure robustness, we also cross-validated the selection using the Akaike Information Criterion (AIC), the Final Prediction Error (FPE), and the Hannan–Quinn Information Criterion (HQIC). The results in

Table 5. VECM Lag Length Selection Criteria Results.

Variable	AIC	BIC/SBC	FPE	HQIC
lnPC − MC	1 *	1 *	1 *	1 *
lnPC − ST	1 *	1 *	1 *	1 *
lnPC − TR	1 *	1 *	1 *	1 *

* Optimal Lag.

show that all criteria consistently suggest an optimal lag of 1 for the examined variables. This high degree of consensus confirms that our model specification is stable and reliable across different information criteria.

Table 6. VECM Granger Causality Test Results (2001–2024).

Direction of Causality	F-Statistics (Short-Run)	p-Value	ECT(t-1) (Long-Run)
$∆MC\to ∆lnPC$	5.312 **	0.015 *	−0.412
$∆MC\leftarrow ∆lnPC$	1.0125	0.342	0.215
$∆ST\to ∆lnPC$	6.120 *	0.008	−0.456
$∆ST\leftarrow ∆lnPC$	0.854	0.412	0.32
$∆TR\to ∆lnPC$	4.982 *	0.021	−0.395
$∆TR\leftarrow ∆lnPC$	1.110	0.315	−0.002

Note 1: * and ** indicate significance at the 1% and 5% levels, respectively. Note 2: ECT represents the coefficient of the error correction term derived from the cointegrated equation. Note 3: right arrow indicates the direction of causality, while (–) indicates no causality.

presents the findings of the causality relationship between the series. The findings show that the relationship between stock-based variables and economic growth is unidirectional, moving from stock market development to economic growth. Therefore, the “supply leading” hypothesis is valid in the relationship between the stock market and growth. The results of our study are consistent with those of Peia and Roszbach (2015) in their research on 22 developed economies. In that study, it was determined that in economies with capital-based financial markets, financial development drives economic growth. In contrast, in economies with banking-based financial markets, economic growth drives financial development. Dogga and Samantaraya (2014) investigated the relationship between growth and a financial development index composed of capital market parameters in the post-financial reform period in India, 1991–2012. They found a unidirectional causality from growth to financial development. Vo and Nguyen (2016), in their study specific to Vietnam, used quarterly data to conclude that there was no relationship between stock market variables and growth during the period 2001–2013. Puryan (2017), in his analysis using data from 1988–2012 for selected North African and Middle Eastern countries, concluded that there is a two-way relationship between stock market development and economic growth.

The estimation results show some differences between the DOLS and FMOLS methods. The TR variable is positive and significant only in the DOLS model. However, it becomes statistically insignificant in the FMOLS estimation. This discrepancy arises from the different technical approaches of these estimators. DOLS uses leads and lags to handle endogeneity more dynamically. FMOLS uses a semi-parametric approach to correct for long-run correlation.

The Fragile Five economies often exhibit high market volatility and heterogeneity. FMOLS results suggest that market efficiency does not impact growth uniformly. In contrast, DOLS captures the short-term dynamics of trading volume better. These results indicate that the growth impact of efficiency is country-specific. Therefore, capital market depth remains more reliable than efficiency for these nations.

Unidirectional Causality: The findings show a unidirectional causal relationship from equity market variables (MC, ST, TR) to economic growth (lnPC); see Table 4.

Supply-led Hypothesis: These results support the “supply-led hypothesis,” which holds that financial development triggers economic growth, for the Fragile Five countries.

Long-Term Equilibrium: The negative and statistically significant Error Correction Term (${ECT}_{t-1}$) confirms that short-term imbalances return to equilibrium in the long term.

Literature Comparison: The findings are consistent with Peia and Roszbach (2015) but differ from the work of Dogga and Samantaraya (2014), which found causality from growth to finance in countries such as India.

After applying cointegration tests and the VECM causality test, the DOLS (Dynamic Ordinary Least Squares) and FMOLS (Full Modified Ordinary Least Squares) methods were used to estimate the final unbiased coefficients for the independent variables based on the identified relationship between stock market variables and growth. Proposed by Stock and Watson (1993) and later developed by Kao and Chiang (2000), DOLS accounts for the endogeneity between the independent variables and the error term and eliminates the bias problem encountered in ordinary least squares cointegration. DOLS facilitates tracking the tendency toward simultaneity among regressors by creating a model that accounts for long-term cointegration and incorporates lagged, leading, and subsequent period values of the variables (Law et al., 2014). FMOLS is a model that produces consistent estimates in small samples and does not allow endogenous or heterogeneous dynamics to distort results. Developed by Pedroni (1999) for panels with cointegration, this model allows for significant heterogeneity across cross-sections, thereby satisfying group-internal and group-external asymptotic x^2-unbiased properties. FMOLS is a semi-parametric approach, and its most important advantage is that it allows for the estimation of different cointegration vectors for each horizontal cross-section in the presence of heterogeneity (Kandil et al., 2015).

Long-Term Coefficients: The DOLS and FMOLS results confirm that stock market capitalization (MC) has a positive, statistically significant effect on economic growth (see

Table 7. Panel DOLS and FMOLS Test Results (2001–2024).

Estimator	Independent Variable	Coefficient	t-Statistic	Std. Error	Prob.
DOLS
	∆MC	0.0142 *	3.652	0.0004	0.000
	∆ST	0.0118	1.410	0.0084	0.159
	∆TR	0.0015 **	2.514	0.0006	0.012
FMOLS
	∆MC	0.0028 *	3.780	0.0007	0.000
	∆ST	−0.0029	−1.815	0.0016	0.071
	∆TR	−0.0011	−1.050	0.0010	0.294

* and ** indicate significance at the 1% and 5% levels, respectively.

).

Market Depth (MC): The positive effect of increased market capitalization (MC) on growth indicates that deeper capital markets in the Fragile Five countries facilitate the reallocation of resources to the real economy.

Transaction Volume and Efficiency: According to DOLS results, stock transaction volume (TR) has a positive and significant effect on growth, whereas FMOLS results indicate a weaker effect. This suggests that the impact of market efficiency on growth may vary across countries.

Coefficient Magnitude: The relatively low coefficients (such as 0.0142 and 0.0028) indicate that the financial systems in these countries remain bank-centric and that the contribution of capital markets to growth has not yet reached its potential.

The study also considers the structural differences among the Fragile Five countries. Each country has a unique level of economic and financial market maturity. For example, South Africa has the most developed stock market depth. In contrast, Turkey relies more on the traditional banking sector. Brazil and India show stronger responses to changes in stock market capitalization.

Heterogeneity analysis reveals that market efficiency affects growth differently in Indonesia. Institutional quality and legal protection also vary significantly across these nations. These individual differences explain why TR results vary in some models. Aggregate results provide a general trend for the entire group. However, country-specific factors remain crucial for long-term sustainable economic development.

According to the panel DOLS test results examining the effects of stock market development indicators on economic growth, stock market capitalization (MC) and trading volume (TR) positively and statistically significantly affect growth. However, the effects are weak given the coefficients. The effect of transaction value (ST) on stock growth is positive but statistically insignificant. When the panel FMOLS test results are evaluated, only the market value of stocks, in other words, the capitalization ratio (MC), has a positive and statistically significant effect on growth. However, this effect appears to be quite weak. No effect of other stock market development variables, namely trading value (ST) and trading volume (TR), on economic growth was found. Kandil et al. (2015) applied the FMOLS test and concluded that the market capitalization ratio, stock trading value, and trading volume positively affect growth. DOLS and FMOLS tests are uısed to reveal a two-way relationship between the market capitalization ratio (MC) and growth. Per capita national income does not have a statistically significant effect on the dependent variable, the market capitalization ratio, using DOLS and FMOL.

Table 8. Summary Comparison Table.

Feature	Findings of This Study	Other Studies (e.g., Konstantakopoulou, 2025)	Relationship
Dominant Proxy	Market Capitalization (MC)	Multidimensional (Depth/Efficiency)	Consistent
Hypothesis	Supply-led Growth	Mixed/Varies by Dimension	Contrast/Refinement
Methodology	Second-gen Panel (CIPS/CD)	VAR or GMM	Methodological Advance

highlights the key differences between the proposed study and the literature.

5. Conclusions

This study empirically investigates the theoretical link between financial development and economic growth for the “Fragile Five” group (Brazil, India, Indonesia, South Africa, and Turkey) as identified by Morgan Stanley. Utilizing an expanded dataset for the 2001–2024 period, the analysis aims to fill a significant gap in the literature where studies focusing exclusively on stock market parameters as indicators of financial development remain relatively scarce. Using panel data methods, three frequently cited stock market parameters—market capitalization, transaction value, and turnover rate—were analyzed in relation to per capita national income.

The main findings of the study are as follows:

The primary contribution of this study lies in its robust identification of the supply-led growth hypothesis for the Fragile Five economies through second-generation panel techniques. By accounting for cross-sectional dependence and structural heterogeneity, our research proves that stock market development serves as a critical leading indicator—rather than a mere consequence—of economic progress in these specific emerging markets.

Our empirical results, which distinguish between the impacts of market depth (MC) and liquidity (ST/TR), are consistent with the multidimensional evidence provided by Konstantakopoulou (2025). This confirms that for the Fragile Five, specific dimensions of stock market development play distinct yet complementary roles in fostering economic resilience. Specifically, the estimated coefficients from the Long-run models provide significant economic insights. For instance, the FMOLS results indicate that a 1% increase in stock market capitalization (MC) is associated with an approximately 0.28% increase in real GDP per capita. While this effect size may appear modest, it represents a substantial impact in the context of emerging economies, where marginal improvements in financial depth can trigger broad-based productivity gains through enhanced capital allocation efficiency.

Regarding the mixed evidence for trading activity (TR and ST) observed between DOLS and FMOLS estimates, our findings suggest that the structural size of the market is a more stable and reliable driver of growth than short-term trading frequency in these economies. Consequently, the proposed policy implications directly stem from the long-term cointegration relationships we identified. Policymakers should prioritize expanding local capital markets by incentivizing public offerings to deepen market depth. Furthermore, our causality analysis shows that market efficiency supports sustainable growth, and strengthening investor protection frameworks is essential to reduce information asymmetry and attract high-quality foreign direct investment. These targeted measures will create a more robust financial environment, ensuring that capital markets provide a reliable foundation for long-term development despite the inherent volatility of emerging market dynamics.

Long-term Cointegration: Following stationarity testing, Pedroni and Kao cointegration tests confirmed a stable, long-term relationship between stock market development parameters and economic growth.

Validity of the Supply-Led Hypothesis: The VECM Granger causality test results indicate a unidirectional causality running from stock market development variables to economic growth. This confirms that the “supply-led” hypothesis is valid for the Fragile Five countries during the examined period.

Impact of Capitalization: Estimations from DOLS and FMOLS tests indicate that stock market capitalization (MC) has a positive, statistically significant effect on growth. However, the magnitude of this effect remains relatively weak in terms of the coefficient, suggesting that the stock market’s full potential has yet to be realized in these economies.

Efficiency Divergence: A divergence was noted in the stock trading volume (TR) variable. At the same time, the DOLS test showed a positive and significant impact on growth, but the FMOLS test found the effect to be statistically insignificant.

Policy Recommendations

The findings suggest that stock market development parameters are vital indicators for economic growth forecasts. Although the financial structures of the Fragile Five are currently banking-centered, capital markets offer a crucial alternative for financing sustainable growth. Based on these results, the following strategic steps are recommended:

Deepening Capital Markets: Policies should aim to expand the volume and depth of capital markets to reduce reliance on the banking and credit sectors.

Incentivizing Public Offerings: Governments should implement policies to encourage companies to go public, such as reducing listing costs, streamlining pre- and post-issuance regulations, and providing technical support for stock price determination.

Enhancing Investor Confidence: To attract hesitant investors, transparency must be prioritized through strict adherence to international financial reporting standards and greater institutionalization of publicly traded firms.

Legal Protections: Establishing robust legal sanctions and controls is essential to protect household investors from speculative attacks.

Educational Outreach: Efforts to reduce public perceptions that stock market investing is overly complex will further support the market’s contribution to long-term economic stability.