RDP 9612: External Influences on Output: An Industry Analysis Appendix B: Estimation Methodology for the Error-Correction Equations

The error-correction equations are estimated using the Seemingly Unrelated Regression (SURE) technique, a systems form of estimation which makes use of the cross-correlation of error terms in the equations to obtain more precise coefficient estimates. With the data set at hand, there is a strong expectation that ‘shocks’ to sectoral output are highly correlated. For example, a tightening of monetary policy would enter Equation (1) in the text as a shock but this would most likely be felt in all sectors. Similarly, a technological shock would most likely be experienced by a number of manufacturing sub-sectors, rather than be isolated to just one sub-sector. While equations estimated by OLS are consistent and unbiased, efficiency is improved by SURE estimation and making use of the contemporary correlation of the residuals of the equations. The downside is that precious degrees of freedom are lost in calculating the covariance matrix of residuals of the OLS equations. Our judgment is that using SURE substantially increases the efficiency of the estimates: as shown in Tables B.1 and B.2, correlations between the residuals of the OLS equations at the one and two-digit levels, respectively, are often high and significant. Accordingly, the standard errors of the SURE estimates tend to be substantially smaller than those of the OLS estimates, presented in Table B.3. Whatever the case, the interpretation does not change qualitatively between the two methods.

Table B.1: Correlation of OLS Residuals at Digit 1 Level
  Agriculture Mining Manufacturing Utilities Construction Wholesale Retail Transport Communications Finance Recreation
Agriculture 1 0.37 −0.13 0.31 −0.21 −0.27 −0.15 −0.02 −0.16 0.18 −0.60
Mining   1 −0.62 0.18 −0.72 −0.46 0.32 0.12 −0.38 0.13 −0.38
Manufacturing     1 −0.12 0.45 0.27 −0.35 0.14 0.80 −0.33 0.19
Utilities       1 0.00 0.12 −0.28 0.16 −0.41 −0.52 0.35
Construction         1 0.69 −0.03 −0.54 0.21 −0.37 0.41
Wholesale           1 −0.18 −0.44 −0.04 −0.21 0.40
Retail             1 −0.14 −0.22 0.40 0.22
Transport               1 0.01 0.09 −0.02
Communications                 1 −0.28 −0.10
Finance                   1 −0.37
Recreation                     1
Table B.2: Correlation of OLS Residuals at Manufacturing Digit 2 Level
  Food Textiles Clothing Wood etc Paper etc Chemical Minerals Basic metal Fabricated metal Transport equipment Other Miscellaneous
Food 1 −0.49 0.09 −0.36 −0.57 0.37 0.16 0.27 −0.58 −0.07 0.26 0.19
Textiles   1 −0.24 0.24 0.32 −0.34 −0.61 −0.45 0.56 −0.19 −0.15 0.00
Clothing     1 0.24 0.09 −0.40 0.57 −0.06 0.06 −0.53 −0.23 0.38
Wood etc       1 0.71 −0.08 0.29 −0.37 0.61 −0.60 −0.38 −0.15
Paper etc         1 −0.10 −0.13 −0.28 0.72 −0.40 −0.26 0.18
Chemical           1 −0.15 −0.2 0.02 0.11 0.41 0.32
Minerals             1 −0.12 −0.18 −0.26 −0.32 −0.22
Basic metal               1 −0.53 0.48 0.21 0.01
Fabricated metal                 1 −0.44 0.10 0.19
Trans equipment                   1 0.41 −0.31
Other                     1 0.07
Miscellaneous                       1
Table B.3: Australian and US Sectoral Output Error-Correlations (1977–93), OLS
  Constant  
β0
Sector
adjustment
β1
US sector adjustment
β2
Aggregate adjustment
β3
US aggregate
adjustment
β4
US sector
impact
β5
Aggregate impact
β6
US aggregate
impact
β7
Lag sector
impact
β8
Inline Equation    
Total GDP(P) 1.72**
(0.51)
−0.74**
(0.19)
0.92**
(0.23)
0.54**
(0.19)
0.62
Agriculture 1.49
(0.97)
−0.85***
(0.23)
0.79***
(0.22)
1.40 **
(0.50)
0.44**
(0.20)
0.59
Mining −6.47***
(2.00)
−1.01***
(0.32)
0.48
(0.38)
1.66***
(0.47)
0.28
(0.23)
1.55**
(0.56)
0.49
Manufacturing 2.16*
(1.09)
−0.39**
(0.16)
0.31**
(0.11)
0.07
(0.16)
0.93***
(0.28)
0.68
Food 1.50
(1.30)
−0.43*
(0.22)
0.31*
(0.15)
0.13
Textiles 0.45
(1.72)
−0.61
(0.38)
0.45
(0.44)
1.31
(0.91)
0.16
Clothing 5.70***
(1.65)
−0.35**
(0.13)
−0.24***
(0.08)
0.46*
(0.24)
0.88*
(0.43)
0.61
Wood & furn. 2.21**
(0.95)
−0.42*
(0.19)
0.32
(0.24)
0.23
(0.21)
1.85**
(0.77)
0.51
Paper 0.88
(0.53)
−0.14
(0.11)
0.07
(0.20)
1.93***
(0.53)
0.49
Chemicals 2.32**
(1.07)
−0.41**
(0.17)
0.27***
(0.09)
0.07
(0.29)
0.35
Non–met min. 0.67
(1.44)
−0.58**
(0.26)
0.15
(0.16)
0.26*
(0.14)
1.85**
(0.71)
0.55
Basic metals −2.24
(1.71)
−0.45**
(0.20)
0.17*
(0.09)
0.44*
(0.21)
0.26**
(0.10)
0.36
Fabr' d met. 0.05
(0.88)
−0.28*
(0.14)
0.54
(0.41)
2.46***
(0.72)
0.85
(0.85)
0.77
Trans. equip. 15.60***
(3.60)
−1.66***
(0.37)
−0.27**
(0.12)
0.81**
(0.31)
0.56
Other mach. 4.96***
(1.50)
−0.86***
(0.25)
0.46**
(0.15)
0.53**
(0.22)
0.53**
(0.20)
0.61
Misc manuf. 2.64**
(1.05)
−0.41**
(0.17)
0.22
(0.14)
0.47***(0.11) 0.56
Utilities −0.77
(1.04)
−0.17
(0.13)
0.19
(0.17)
– 0.45**
(0.17)
0.35
Construction 2.35**
(1.07)
−0.50**
(0.21)
0.21*
(0.12)
2.06***
(0.41)
0.42**
(0.16)
0.83
Wholesale 1.54
(1.31)
−0.38
(0.25)
0.18
(0.12)
– 1.47***
(0.37)
0.65
Retail −0.83
(0.71)
−0.58**
(0.23)
0.55*
(0.25)
  0.26
(0.18)
0.74**
(0.25)
0.45
Tran & storage −4.58***
(1.40)
−1.32***
(0.26)
0.36***
(0.10)
1.22***
(0.29)
0.96***
(0.19)
0.77
Rail −6.55***
(2.03)
−0.88***
(0.26)
1.02***
(0.30)
1.66***
(0.48)
0.58
Water −1.79
(1.03)
−0.76**
(0.28)
0.59**
(0.21)
– 1.05**
(0.46)
0.45
Air 2.62**
(1.13)
−0.53**
(0.21)
0.47**
(0.18)
0.25
Road −7.69***
(1.39)
−1.44***
(0.23)
1.65***
(0.27)
1.03***
(0.25)
0.70**
(0.28)
0.81
Communic'ns −0.26*
(0.14)
0.13
(0.08)
−0.17
(0.13)
0.47**
(0.20)
0.27
Finance −0.45
(0.75)
−0.25*
(0.12)
0.37
(0.24)
0.60*
(0.30)
0.58*
(0.31)
0.59***
(0.14)
0.77
Recreation & pers. services −1.59***
(0.37)
−1.21***
(0.25)
1.05***
(0.22)
0.23**
(0.10)
0.54***
(0.11)
0.25
(0.20)
0.77

Note: *, ** and *** denote significance at the 10%, 5% and 1% levels, respectively, using the standard t-distribution.