The Performance of the Agricultural Economics in BRICS Countries - The Agricultural Economics of the 21st Century - page 85

Table 6.1 Results obtained with time series econometric techniques, based on the function of
production model (linear model obtained with logarithms), for the agricultural output in the period
1961-2012

South Africa Brazil China India Russia

ARCH
family
regression

Prais-
Winsten
(Robus)

Prais-
Winsten

Prais-
Winsten

Prais-
Winsten Model

Constant 0.542*

(3.130)
[0.011]

10.309*
(3.080)
[0.002]

3.203*
( 4.990)
[0.000]

5.552*
(4.790)
[0.000]

Agriculture value added per
worker (constant 2005 US$)

0.797*
( 3.490)
[0.000]

0.0376*
( 2.020)
[0.052]

Employment in agriculture
(% of total employment)

0.272*
(3.350)
[0.007]

1.550*
(9.690)
[0.000]

0.829*
(4.660)
[0.000]

Augmented Dickey-Fuller
test for unit root

7.092*
[0.000]

6.115*
[0.000]

5.456*
[0.000]

5.765*
[0.000]

3.077*
[0.028]

EG-ADF test for
cointegration

2.033
[0.272]

2.489
[0.118]

2.063
[0.259]

1.973
[0.298]

2.320
[0.165]

Portmanteau test for white
noise for autocorrelation

296.018*
[0.000]

308.999*
[0.000]

322.740*
[0.000]

320.681*
[0.000]

30.878*
[0.000]

Durbin's alternative test for
autocorrelation

0.843
[0.358]

20.264*
[0.000]

9.810*
[0.001]

6.336*
[0.011]

0.600
[0.438]

Breusch-Godfrey LM test
for autocorrelation

1.144
[0.284]

12.646*
[0.000]

8.302*
[0.004]

5.754*
[0.016]

0.739
[0.389]

Breusch-Pagan/Cook-
Weisberg test for
heteroskedasticity

1.730
[0.188]

0.000
[0.984]

6.350*
[0.011]

0.010
[0.929]

0.040
[0.843]

Ramsey RESET test using
powers of the fitted values

18.570*
[0.001]

5.100*
[0.008]

10.490*
[0.000]

3.940*
[0.018]

2.470
[0.111]

LM test for autoregressive
conditional heteroske-
dasticity (ARCH)

0.551
[0.458]

4.386*
[0.036]

1.194
[0.274]

0.092
[0.761]

0.527
[0.467]

Note : *Statistically significant at 5 %

Winsten” time series estimation method was used. The “Breusch-Pagan/Cook-
Weisberg test for heteroskedasticity” test revealed problems with the heteroske-
dasticity in China and so the “Prais-Winsten robust” was considered to solve this
statistical infraction. The “LM test for autoregressive conditional heteroskedasticity
(ARCH)” test presents difficulties in Brazil and taking this information into account
the “ARCH family regression” estimation method was considered. Relative to the
values of the coefficients, it is possible to conclude that the proportion of employ-
ment in agriculture has had a positive effect in the agricultural output in
South Africa, China, and in the Russian Federation. The more important influence
comes from the China with a contribution from employment in the output of 1.550,