0.00, 0.04 and 0.32 of the total variance of μ is explained by the total variance
in GFCF in, respectively, the (1) 7-10, (2) 5-7 and (3) 3-5 year frequency
ranges. Although statistical significance is established in the 7-10 and 5-7 year
ranges, the magnitudes are not economically meaningful.
More generally, the results in Table 3 suggest the following. The over-
time mark-up, μ is the only wage component that associates significantly with
changes in inventories, IF Without any doubt, most of its explained varia-
tion occurs within the shortest 3-5 year range. This variable is dominated by
changes in overtime hours and we speculated earlier that the inventory vari-
able offers a suitable choice of indicator. While other ranges are found to be
signihcant, they account for very little of the total variation of the premium
mark-up.
The proportion of overtime workers, ʌ, displays strong associations with
the middle bands of the GFCF, Y and N indicators. More than any other
wage component, ʌ reveals the usefulness of testing associations over a range of
cycle ranges. Both long- and, especially, short- cycle ranges also significantly
co-vary with this wage component.
The dominant cyclical influence on the consumption wage occurs within the
middle (5-7 year) time band. This is true for three of the cyclical indicators -
i.e. the GFCF, Y and N. These findings are consistent with the univariate
results. Note that, while explaining considerably less of the total variance, the
longest (7-10) range is also significant for the GFCF and Y indicators. Again,
this establishes that more than one length of cycle may significantly influence
wage earnings components.
Finally, the production wage relates exclusively to the 5-7 range and is
confined only to the GFCF indicator. Few earlier studies have been able to
link producer wages to the business cycle. These results point to the view - as
discussed in Section 3.2 - that such wages may be more appropriately explained
by different types of economic cycle. In particular, we highlighted a potential
connection with movements in Fxed capital formation.
19