Trade openness and non-income poverty in Southern African Development Community (SADC) countries: A panel Autoregressive Distributive Lag (ARDL) analysis

Figures & data

Figure 1. Trade Openness and Non-Income Poverty in SADC Countries.

Table 1. Definition of variables

Indicator	Symbol	Measurement	Source
Non-income poverty	NPOV	Human development index	WIID; World Bank
Trade openness	RTO	imports plus exports in exchange rate US$ relative to GDP in purchasing-power-parity US$ (PPP GDP)	WITS, World Bank
Economic growth	RGDP	Real gross domestic product	WITS, World Bank
Income inequality	INEQ	The discrepancies in income distribution in a given economy. It is measured by the Gini coefficient, which ranges from 0 to 1	WIID; PovCalnet World Bank
Human capital Index(education level	HIND	The education level which is measured based on the years of schooling and returns to education.	World Penn Tables 10
Financial Development	FD	Domestic credit to private sector, percent of GDP	Global Economy Data
Institutional Quality	PS	Political Stability	Global Economy
Inflation	INFL	Consumer Price Index: Inflation as measured by the consumer price index reflects the annual percentage change in the cost to the average consumer of acquiring a basket of goods and services that may be fixed or changed at specified intervals, such as yearly.	IMF; World Bank
UNEM	Unemployment	The unemployment rate represents the share of the labor force that is without work but available for and seeking employment.	IMF, World Bank; Global Economic data

Source: Author’s Compilation.

Table 2. Descriptive statistics

Variable	Obs	Mean	Std. Dev.	Min	Max
npov	545	.4979193	.1257558	.195	.801
rto	635	43.63315	26.44985	2.064291	143.05
rgdp	640	3.453979	4.775	−24	26.8
ineq	467	.6546142	.0695477	.5279146	.8326869
hind	568	2.05886	1.38092	1.041064	12.3336
fd	584	22.57449	24.17033	.45	131.0482
ps	592	1.245439	6.506114	−10	10
infl	640	83.97343	981.3996	−72.7	23773.1
unem	354	12.03738	9.565821	.599	37.94
noda	610	8.361598	9.08675	−.2509203	62.1866

Note: hdi is the human development index, which indicates poverty reduction, rto-real trade openness, rgdp is the real gross domestic product(economic growth), ineq-income inequality, ps-institutional quality, fd-financial development, infl-inflation rate, hind- human capital index, unem-unemployment rate, noda-net official development assistance.

Source: Author’s compilation from STATA descriptive statistics results.

Table 3. Unit root test results

Method	IM Pesaran		ADF		CIPS		CADF
Order of integration	1(0)	1(1)	1(0)	1(1)	1(0)	1(1)	1(0)	1(0)
Variable	t-stat	t-stat	t-stat	t-stat	t-stat	t-stat	t-stat	t-stat
npov	8.0271	−6.9488***	5.0263	−3.3943***			−0.675	−2.346**
Lrto	0.1928	−10.6364**	0.5794	−8.4632**			−2.539***	−11.374***
Lrgdp	−108744**	−177964**	−9.5936**	−23.3245**	−4.472***	−6.163***	−6.826*	−17.720***
Lineq	3.3633	−2.8617**	−2.9787**	−3.2506**			−4.420**	−2.160***
Lhind	3.1536	−0.7440*	0.9857	−0.5363**			−0.369*	−0.948
Lfd	−1.6141***	−13.2695**	−2.2086***	−10.9661**			−2.476***	−8.976***
Lps	-	-	−0.7347	−11.6289**			−3.630***	−9.180***
Linfl	−5.5139**	−16.2761**	−4.9003**	19.0415**	−3.093*	−6.120***	−3.418***	−15.245***
lunem	0.4025	−5.8733***	−1.7106*	−6.6249***			−1.692*	−1.993***
lnoda	−1.6586***	−15.228***	−1.1214	−14.494***			−1.130	−11.371***

p<0.01**,p<0.05***, p<0.01*

Note: CIPS - Cross-sectionally augmented Im, Pesaran, and Shin; CADF- Cross-sectionally, Augmented Dickey-Fuller.

Source: Author’s compilation from STATA unit root test results.

Table 4. Short-run PMG estimation results: real trade openness and non-income poverty

	(Model_1)	(Model_2)	(Model_3)	(Model_4)	(Model_5)	(Model_6)
VARIABLES
ECT	−0.0670**	−0.0286**	−0.0538***	−0.0560**	−0.0535*	−0.0430*
	(0.0277)	(0.0126)	(0.0158)	(0.0243)	(0.0278)	(0.0225)
D.lrto	0.00304	−0.0112	−1.195	−0.00204	0.0352	−0.0626
	(0.00507)	(0.00943)	(1.286)	(0.121)	(0.0411)	(0.0635)
D.lrgdp	0.000941**	−0.00879	−0.000559	0.000962**	0.001**	0.001**
	(0.000387)	(0.0167)	(0.000386)	(0.000456)	(0.0003)	(0.0004)
D.lineq	1.778	0.215	−10.16	−1.013	−1.746	−0.642
	(1.715)	(0.586)	(11.21)	(1.186)	(2.302)	(0.807)
D.lhind	0.129	−0.184	0.296	0.275	0.279	0.346
	(0.560)	(0.483)	(0.278)	(0.471)	(0.260)	(0.222)
D.lfd	0.00848*	0.00260	0.00310**	0.00215	0.0265	0.00314
	(0.00468)	(0.00432)	(0.00155)	(0.00144)	(0.0455)	(0.00196)
D.lps	0.00215	0.00331	0.00269	0.000908	0.00280	−0.0356
	(0.00185)	(0.00275)	(0.00219)	(0.00136)	(0.00251)	(0.107)
D.linfl	−0.00105**	0.000193	−3.24e-05	−0.000577	1.85e-05	−0.000209
	(0.000516)	(0.000615)	(0.000503)	(0.000744)	(0.000771)	(0.000690)
D.lunem	−0.00599	0.000291	0.00560	−0.00633	−0.00970	0.00210
	(0.0143)	(0.0129)	(0.0106)	(0.00849)	(0.0108)	(0.0125)
D.lnoda	0.000695
	(0.00183)
D.lrtolrgdp		0.00213
		(0.00396)
D.lrtolineq			2.120
			(2.245)
D.lrtolhind				0.0100
				(0.0823)
D.lrtolfd					−0.00613
					(0.00967)
D.lrtolps						0.00934
						(0.0244)

Note: NPOV is the non-income poverty measured by the human development index, which indicates poverty reduction, rto-real trade openness, rgdp is the real gross domestic product (economic growth), ineq-income inequality, ps-institutional quality, fd-financial development, emp-employment, infl-inflation rate, hind- human capital index, unem-unemployment rate, noda-net official development assistance.

Standard errors in parentheses and p-values in squared brackets. The selected model is ARDL (1, 1, 1, 1, 1, 1, 1, 1, 0, and 0) with constant (according to the AIC criterion)*** p<0.01, ** p<0.05, * p<0.1.Source: Author’s Compilation from STATA PMG Regression Results.

Table 5. Long-run PMG estimation results: real trade openness and non-income poverty

	(Model_1)	(Model_2)	(Model_3)	(Model_4)	(Model_5)	(Model_6)
VARIABLES
L.lrto	0.0243**	0.0907***	0.675***	−0.0350	0.178***	0.0617***
	(0.00943)	(0.0175)	(0.155)	(0.0237)	(0.0292)	(0.0112)
L.lrgdp	0.0140***	−0.372***	0.0360***	0.0251***	0.0140***	0.0244***
	(0.00202)	(0.0238)	(0.00638)	(0.00455)	(0.00269)	(0.00401)
L.lineq	−0.0681	−2.170***	3.025***	−0.443***	−0.285**	−0.227**
	(0.0941)	(0.119)	(0.905)	(0.128)	(0.111)	(0.113)
L.lhind	−0.0999***	−0.0744***	−0.0503***	−0.384***	0.0739***	0.0407***
	(0.0118)	(0.0148)	(0.0190)	(0.0740)	(0.00869)	(0.0145)
L.lfd	0.103***	−0.0610***	0.0376***	0.0594***	0.183***	0.0530***
	(0.00896)	(0.0130)	(0.0106)	(0.0107)	(0.0430)	(0.00838)
L.lps	0.00212	0.00222	−0.00166	0.0183***	0.0133***	0.0178
	(0.00159)	(0.00172)	(0.00358)	(0.00406)	(0.00299)	(0.0161)
L.linfl	−0.0190***	0.00967***	0.00653	0.0189***	0.0159***	0.0212***
	(0.00400)	(0.000826)	(0.00434)	(0.00426)	(0.00253)	(0.00353)
lunem	−0.00200	−0.0349***	−0.183***	−0.0225**	0.00114	0.00798
	(0.00409)	(0.00252)	(0.0653)	(0.00933)	(0.00606)	(0.00850)
lnoda	−0.0108**
	(0.00438)
lrtolrgdp		0.100***
		(0.00636)
lrtolineq			−0.946***
			(0.241)
L.lrtolhind				0.0783***
				(0.0152)
L.lrtolfd					−0.0388***
					(0.0104)
L.lrtolps						6.35e-05
						(0.00418)
Constant	0.0296***	0.0560**	−0.0982***	0.0419**	−0.0147*	0.00312
	(0.0108)	(0.0225)	(0.0290)	(0.0180)	(0.00870)	(0.00241)
Hausman p-val	[0.4851]	[0.8966]	[0.1353]	[0.2734]	[0.2006]	[0.4857]
Observations	277	277	396	394	394	394

Standard errors in parentheses and p-values in squared brackets. The selected model is ARDL (1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0) with constant (according to the AIC criterion).

*** p<0.01, ** p<0.05, * p<0.1.

Source: Author’s Compilation from STATA PMG Regression Results.

Figure 2. Plot of CUSUM of squares: Trade openness and non-income poverty analysis.

Source: Extract from Results from STATA ARDL diagnostic tests results.

Table 6. Diagnostic test results: trade openness and non-income poverty

Group/Country	Durbin-Watson	B-Godfrey	White	CUSUM
1. Angola	1.7766	1.356	18.00	stable
2. Botswana	1.4221	0.463	23.00	stable
3. Comoros	2.6528	11.535*	17.00	stable
4. DRC	2.1359	0.388	18.00	stable
5. Eswatini	1.5416	1.168	20.00	stable
6. Lesotho	2.1589	1.371	21.00	stable
7. Madagascar	2.2434	2.104	17.00	stable
8. Malawi	2.5215	2.793	18.00	stable
9. Mauritius	2.9282	15.65	19.00	stable
10. Mozambique	2.3874	1.244	19.00	stable
11. Namibia	2.0213	0.931	21.00	stable
12. Seychelles	2.9047	8.019	16.00	stable
13. South Africa	2.7545	5.073	20.00	stable
14. Tanzania	2.0111	0.128	19.00	stable
15. Zambia	2.2407	1.310	23.00	stable
16. Zimbabwe	2.7740	9.722	21.00	stable

Standard errors in parentheses.

*** p<0.01, ** p<0.05, * p<0.1.

Source: Author’s compilation from STATA ARDL diagnostic tests results.

Table

	lhdi	lrto	lkof	lexp	limp	lrgdp	lineq	lhind	lfd	lps	linfl	lunem	lnoda
lhdi	1.000
lrto	0.444	1.000
lkof	0.587	0.298	1.000
lexp	0.518	0.647	0.612	1.000
limp	0.380	0.650	0.552	0.699	1.000
lrgdp	−0.013	0.069	0.128	0.030	0.047	1.000
lineq	−0.021	0.143	0.076	0.130	0.147	−0.041	1.000
lhind	0.561	0.223	0.106	0.090	0.234	−0.114	−0.011	1.000
lfd	0.709	0.290	0.568	0.343	0.323	−0.122	0.058	0.234	1.000
lps	0.310	−0.100	0.329	−0.107	0.045	0.162	0.213	0.259	0.326	1.000
linfl	−0.205	−0.161	0.026	0.019	−0.087	0.074	0.096	−0.298	−0.317	−0.021	1.000
lunem	0.227	0.222	0.378	0.390	0.258	0.007	0.562	−0.089	0.287	0.123	0.135	1.000
lnoda	−0.733	−0.502	−0.436	−0.474	−0.203	0.109	−0.050	−0.278	−0.596	−0.047	0.134	-0.4298	1.000

Source: Author’s compilation from STATA panel Correlation Coefficients results.

Table

	Npov	Rto	rgdp	Ineq	hind	fd	ps	infl	unem	noda
r1	1	0	1	1	0	0	0	0	0	0
r1	1	1	1	1	0	1	0	1	1	0
r1	1	0	1	1	0	1	1	0	0	1
r1	1	1	1	0	1	0	1	1	1	0
r1	1	0	0	1	1	1	0	0	0	1
r1	1	1	1	0	1	1	1	1	1	0
r1	1	1	1	1	1	1	1	0	0	1
r1	1	1	1	1	1	1	1	1	1	0
r1	1	1	1	1	1	0	1	0	1	1
r1	1	1	0	1	1	1	0	1	0	1
r1	1	1	1	1	1	1	0	1	1	0
r1	1	1	0	1	0	1	1	1	0	0
r1	1	0	1	1	0	0	0	0	1	1
r1	1	1	1	1	1	0	1	1	1	0
r1	1	0	0	0	0	0	1	0	0	1
r1	1	0	0	1	0	0	0	1	0	0
C_lag	1	1	1	1	1	1	1	1	0	0

ARDL (1,1,1,1,0, 1,1,1,1, 1, 1,0,0) according to AIC criterion.

Source: Author’s compilation from STATA panel ARDL Optimal Lag Length Selection results.

Table

	(PMG)	(PMG)	(MG)	(MG)	(DFE)	(DFE)
VARIABLES	ECT	SR	ECT	SR	ECT	SR
ECT		−0.0670**		−0.874***		−0.0239*
		(0.0277)		(0.171)		(0.0125)
D.lrto		0.00304		0.0104		0.00127
		(0.00507)		(0.0226)		(0.00239)
D.lrgdp		0.000941**		−0.000249		0.00169***
		(0.000387)		(0.00126)		(0.000293)
D.lineq		1.778		23.33		0.0124
		(1.715)		(24.04)		(0.0461)
D.lhind		0.129		−1.848		−0.0126
		(0.560)		(1.302)		(0.00831)
D.lfd		0.00848*		0.0287		0.00440***
		(0.00468)		(0.0256)		(0.00131)
D.linfl		−0.00105**		0.00345		0.000433
		(0.000516)		(0.00283)		(0.000278)
D.lunem		−0.00599		0.0159		−0.000568
		(0.0143)		(0.0109)		(0.00203)
D.lnoda		0.000695		−0.00152		0.000472
		(0.00183)		(0.00229)		(0.000910)
L.lrto	0.0243**		0.0382		−0.0158
	(0.00943)		(0.0348)		(0.0629)
L.lrgdp	0.0140***		−0.0194		0.108*
	(0.00202)		(0.0140)		(0.0562)
L.lineq	−0.0681		−0.180		0.272
	(0.0941)		(3.641)		(0.684)
L.lhind	−0.0999***		1.145**		0.00959
	(0.0118)		(0.523)		(0.0820)
L.lfd	0.103***		−0.0241		0.0990**
	(0.00896)		(0.0540)		(0.0482)
L.linfl	−0.0190***		−0.00808		0.0658*
	(0.00400)		(0.0147)		(0.0393)
lunem	−0.00200		−0.00223		−0.0756
	(0.00409)		(0.0418)		(0.0680)
lnoda	−0.0108**		0.0205		0.0220
	(0.00438)		(0.0360)		(0.0375)
Constant		0.0296***		−3.269		0.00217
		(0.0108)		(4.000)		(0.0136)
Hausman p-value			[0.4851]		[1.000]
Observations	277	277	277	277	.	.

Standard errors in parentheses and p-values in squared brackets. The selected model is ARDL (1, 1, 1, 1, 1, 1, 0, 0) with constant (according to the AIC criterion).

*** p < 0.01, ** p < 0.05, * p < 0.1.

Source: Author’s compilation from STATA Panel ARDL Regression results.

Table

	Test Statistic	Absolute Value	P-value
Pesaran’s Test	16.130	0.501	0.000
Friedman’s Test	87.353	0.512	0.000

Source: Author’s compilation from STATA Cross-Sectional Dependence tests results.