Our data set indicates that the average—over the 3
years—cost for maintaining tracks is 3.8 (3.6) and 6.4
(5.4) million (m) SEK for the secondary and main lines,
respectively, with the standard deviation in parenthesis.
Our sample of 520 observations consists of 167
secondary and 353 main track units. From estimating
the full model (Eq. (3)) it is possible, using a
conventional F-test, to test for whether a restricted
model—where time constant parameters as well as the
parameters for secondary and main tracks are the same
over the 3-year period (39 restrictions)—provides a better
description of the data than the full model. We get F ¼
1:154 (P-value ¼ 0.25) and hence, these restrictions
cannot be rejected;
(i) there is no statistically significant difference in
coefficients between the years;
(ii) although average costs differ, we cannot detect any
difference in the structure of how the two types of
railway lines are maintained once we control for traffic,
track length, etc.
The final model used for analyzing the Swedish data is
therefore