Skip to main content

Table 4 Tests for panel data estimation techniques

From: Integration into regional or global value chains and economic upgrading prospects: an analysis of the East African Community (EAC) bloc

Estimation

Test and hypotheses

Results

Test for multi-

collinearity

Variance inflation factor (VIF)

Variable

VIF

1/VIF

LGDPCA

8.15

0.0985

LFVA

5.16

0.1937

LQS

3.80

0.2635

QI

2.67

0.3748

LHCA

2.28

0.4395

LFDI

2.13

0.4694

LDOC

1.87

0.5339

Mean VIF

4.01

Breusch-Pagan Lagrange multiplier (LM)

xttest0 command in STATA 12.

H0=Null Hypothesis: No

panel effect.

Chibar2 (01) = 674.24

Prob > F= 0.000

Hausman Test: selection between fixed or random effects

xtreg +

Hausman commands in STATA 12.

H0: Random effect model is appropriate

H1: Fixed effect model is appropriate

P.value>\(\mathop x\nolimits^{2}\)=0.000

P.value>\(\mathop x\nolimits^{2}\)=>

0.05, therefore fixed effect is

appropriate.

Tests for serial correlation

Wooldridge Test

H0: no first

order autocorrelation

F (1, 4) = 0.860

Prob > F = 0.4061

Test for

heteroskedasticity

Modified Wald Test

H0: Sigma(i)2=sigma2 for all i

\(\mathop x\nolimits^{2}\)(5) = 102.12

Prob= 0.0000