Revisiting stock market development and economic growth nexus: The moderating role of foreign capital inflows and exchange rates

Figures & data

Figure 1. Research model for stock market and economic growth.

Figure A1. ARDL model selection for Model A.

Figure A2. ARDL model selection for Model B.

Figure A3. CUSUM test for regression stability (Model A).

Table 1. Descriptive statistics

	ln(Econ. Growth)	ln(MKTCAP)	ln(FCI)	ln(ER)
Mean	1.7861	4.7985	4.1128	1.1172
Minimum	−0.6539	3.9468	3.450672	0.8329
Maximum	5.7714	2.3025	5.304215	1.4929
SD	0.5980	0.4785	0.546198	0.1966
Skewness	−0.7421	−0.1079	0.815802	0.2211
Kurtosis	2.2687	2.5961	2.494501	1.7151
Jarque–Bera	4.0036	0.2083	4.171596	2.9198
Probability	0.2002	0.9010	0.232257	0.2323

Table 2. Results of correlation among variables

	Correlation matrix
	ln(Econ. Growth)	ln(MKTCAP)	ln(FCI)	ln(ER)
ln(Econ. Growth)	1
ln(MKTCAP)	0.4970	1
ln(FCI)	−0.0552	−0.4728	1
ln(ER)	−0.3098	0.3171	−0.4255	1

Table 3. Unit root test results

	Log level		First differences			Z-Break test
	AIC lag	DFGLS stat	AIC lag	DFGLS stat	I (d)	AIC	Z stat
ln(Econ. Growth)	1	−1.896	1	−7.265**	I(1)	1	−4.307
ln(MKTCAP)	1	−1.501	1	−4.848**	I(1)	0	−4.583
ln(FCI)	2	−0.996	0	−4.465**	1(1)	0	−3.103
ln(ER)	2	−0.374	0	−4.712*	I(1)	1	−4.1655

Notes: All variables are in logs in the series. The DF-GLS statistic is compared to the critical values from the simulated MacKinnon table in Elliott (1996, Table 1, p. 825).

Results obtain from EViews 9.5.

^* Rejection of the null at 5% significance level.

^** Rejection of the null at 1% significance level.

Table 4. Granger causality between economic growth and stock market development

Null hypothesis	F-statistic	p-value
Stock market development does not Granger cause economic growth	13.20845	0.0000
Economic growth does not Granger cause stock market development	0.17231	0.8430

Notes: Period 1981–2016, lag length 2 (according to AIC).

The null hypothesis is rejected at the 5% significance level.

Table 5. Results from bounds tests on Equations (1) and (2)

Dep. var.	AIC lags	F-statistic	p-value	Outcome
Econ. Growth	1	13.0566	0.0020	Co-integration
MKTCAP	1	0.2762	0.7624	No-co-integration

Notes: Asymptotic critical value bounds are obtained from Table C(I) of Pesaran and Pesaran (2009). Case: Unrestricted intercept and no trend for k = 1 (Pesaran & Pesaran, 2009). The lower and upper bounds for the F-test statistic at the 10, 5, and 1% significance levels are [4.04, 4.78], [4.94, 5.73], and [6.84, 7.84], respectively. Model selection summary presented with Appendices A^{Appendix B}–C.

Table 6. Estimated long-run coefficients using the ARDL approach

Variables	Model A		Model B
	ln(Econ. Growth)	t-statistic	ln(MKTCAP)	t-statistic
Constant	0.3744	0.3586	4.46301	16.871***
ln(Econ. Growth)			0.191223	1.584
ln(MKTCAP)	0.2938	1.9889**
R^2	0.2218		0.18741
F-statistic	3.9912***		2.1945**

^* p-value at 10%.

^** p-value at 5%.

^*** p-value at 1%.

Table 7. Error–correction representation for the selected ARDL model

Variable	Model A		Model B
	D(ln(EconGrowth)	t-statistic	D(ln(MKTCAP)	t-statistic
C	0.0679	0.958	0.3247	0.749
D(ln(Econ. Growth (−1))	−0.1584	−1.691	−0.0459	0.964
D(ln(MKTCAP (−1))	0.2050	−2.016**	−1.0836	0.292
ECM (−1)	−0.6404	−3.60***	−0.043	−1.1345
R^2	0.8143
Akaike info criterion	0.4385
Schwarz criterion	0.7369
Durbin–Watson stat	2.176767		2.0810
Log likelihood	−22.86083		−6.4823
F-statistic	13.1635		1.7250
Prob. (F-statistic)	0.0005		0.1798
Wald test (F-statistic)	3.276**		0.9432
LM test (F-statistic)	1.313087		0.2561

^* p-value at 10%.

^** p-value at 5%.

^*** p-value at 1%.

Table 8. Regression-based moderating test results

	Model 1	Model 2	Model 3	Model 4
Constant	0.3750	0.2512	0.3011	−0.3051
ln(MKTCAP)	0.293**	0.331**	0.321**	0.416**
ln(FCI)		−0.0724*		−0.1471***
ln(MKTCAP)* ln(FCI)		0.0026**		0.0052**
ln(ER)			−2.698*	−3.4418
ln(MKTCAP)* ln(ER)			0.0014	−0.3021**
ln(FCI)*ER				0.5365**
ln(MKTCAP)* ln(FCI)* ln(ER)				0.0189**
R^2	0.2218	0.3788	0.4112	0.5321
R^2 increase due to interaction		0.0912***	0.0093	0.0755**
Critical value (F)	3.991**	7.016***	5.891***	7.9212***

Notes: Economic growth as the dependent variable for all models. Equations (3)–(6) employed for Models 1–4, respectively.

^* p-value at 10%.

^** p-value at 5%.

^*** p-value at 1%.

Elliott, G., Rothenberg, T. J., & Stock, J. H. (1996). Efficient tests for an autoregressive unit root. Econometrica, 64, 813–836.10.2307/2171846

 Web of Science ®Google Scholar

Pesaran, H. M., & Pesaran, B. (2009). Microfi t 5.0. Oxford: Oxford University Press.

 Google Scholar