99
Table 2. Haugh-Pierce (H-P) Causality Tests on Monthly Ex-vessel,
Wholesale, and Retail Prices for the 21-25 Size Class Using
ARIMA Filtered Data.
Series 1 Prewhitened Ex-vessel Prices
Series 2 = Prewhitened Wholesale Prices
Series 3 = Prewhitened Retail Prices
TEST I
(N =* 162)a
TEST II
(N =* 166)a
Lag (k) in
Causal
Series j
Cross Correlations (r^Ck)}
(a) r2}1(k) (b) r1>2(k)
Cross Correlations {r^
(a) r3>2(k) (b) r2>
0
.836*
.836*
.100
.100
1
.167*
-.037
.227*
.130
2
.005
.041
.389*
.059
3
-.010
-.063
.129
.061
4
.033
.020
.017
.014
5
-.026
.012
.022
-.080
6
-.027
-.044
.106
-.027
7
-.065
.087
.090
-.073
8
.017
-.064
-.027
-.078
9
-.038
-.012
.004
-.143
10
.100
.101
.040
-.081
11
-.069
-.141
.114
-.074
12
-.182*
-.128
.081
-.064
TEST I:
(a) Null Hypothesis: ex-vessel does not cause wholesale price,
H-P xf2 13.61***
(b) Null Hypothesis: wholesale does not cause ex-vessel price,
H-P Xi2 &.40***
TEST II:
(a) Null Hypothesis: wholesale does not cause retail price,
H-P X|2 44.02**
(b) Null Hypothesis: retail does not cause wholesale price,
H-P X12 13.10***
aN is the effective number of observations for cross correlation.
Greater than two standard errors (s), where s = N-1^.
Reject null hypothesis at the .05 level.
***Fail to reject null hypothesis at the .10 level.