Table 3.

The best model

Cointegration equation	Cointegration equation 1
Lnprivate (−1)	1
Lnpublic (−1)	−0.81
	(0.14)
	[−5.81]
Lnsale (−1)	0.12
	(0.20)
	[0.63]
Intercept	−2.55
Error correction	D(lnprivate)	D(lnpublic)	D(lnsale)
Cointegration equation 1	−0.10	0.02	−0.00
	(0.03)	(0.01)	(0.01)
	[−3.58]	[2.70]	[−0.49]
D(lnprivate(−1))	−0.20	−0.02	0.01
	(0.06)	(0.01)	(0.01)
	[−3.45]	[−1.72]	[0.87]
D(lnpublic(−1))	0.29	0.04	−0.01
	(0.27)	(0.06)	(0.06)
	[1.08]	[0.67]	[−0.16]
D(lnsale(−1))	0.28	0.07	0.15
	(0.26)	(0.06)	(0.06)
	[1.10]	[1.09]	[2.52]
Intercept	0.05	0.04	0.06
	(0.03)	(0.01)	(0.01)
	[1.71]	[6.22]	[7.72]
R2	0.12	0.04	0.03
Adj. R2	0.10	0.02	0.01
Sum sq. resids	27.52	1.57	1.56
SE equation	0.34	0.08	0.08
F-statistic	8.00	2.57	1.87
Log likelihood	−78.92	279.44	280.00
IC	0.67	−2.20	−2.20
Schwarz SC	0.74	−2.13	−2.13
Mean dependent	0.07	0.05	0.07
SD dependent	0.35	0.08	0.08
Determinant resid covariance (dof adj.)	4.48 × 10–06
Determinant resid covariance	4.22 × 10–06
Log likelihood	482.77
Akaike information criterion	−3.72
Schwarz criterion	−3.46

Standard errors in () & t-statistics in []

Sample adjusted for a period between 1984 and 2008

There are 250 observations included after adjustments

D(lnprivate): first difference of log private sector expenditure; D(lnpublic): first difference of log public sector expenditure; D(lnsale): first difference of log sales; Lnprivate (−1): log private sector expenditure with one year lag; Lnpublic (−1): log public sector expenditure with one year lag; Lnsale (−1): log sales with one year lag; D(lnprivate(−1)): first difference of log private expenditure with one year lag; D(lnpublic(−1)): first difference of log public expenditure with one year lag; D(lnsale(−1)): first difference of sales with one year lag