128
W .289 + .687 W + .709 P .372 P .024 FP
t (.139)2 (.0591)ti (.0451)t"A (.0711>t"i (.030) C
- .0016 BSFF + .0167 131 .0020 01 .0008 TMCI
(.0007)L t (.0221) C (.0026) C (.0011) C
where the values in parentheses are standard errors and the subscripts 1
and 2 refer to significance at the .01 and .05 level of confidence. The
2
model explained approximately 99 percent (R .9937) of the monthly
variation in wholesale price. Five of the nine estimated coefficients
were statistically significant at the .05 level or greater.
The significance of the estimated parameter representing wholesale
price lagged one percent (Wt_j) indicates (as did the analogous estimate
in the retail price expression) that factors which cause a change in
wholesale price (Wfc) have an effect on wholesale price in the next
period. Specifically a one cent change in wholesale price causes whole
sale price in the next period to change in the same direction by -.69
cents. This is indicated by the estimate of the parameter being
significant at the .05 level. In addition, the lagged dependent vari
able was included due to the presence of first order serially correlated
errors. With the inclusion of this lagged dependent variable, the
residuals reduced to an approximate white noise process (Appendix Table
B).
Asymmetry was found not to characterize the price response rela
tionship between current ex-vessel (Pt) and current wholesale prices.
This is indicated by the estimate of current falling ex-vessel price
being insignificant (null hypothesis of zero value not rejected using a
one-tailed test). Thus, the estimate on current ex-vessel price reverts
to a test of symmetric price response between ex-vessel and wholesale
prices, which was found to be significant at the .05 level. This