Observations for the variables included in the short-run cost function were compiled on a
quarterly basis within the period 1980–1995, thus yielding 64 observations. Data were provided by
the balance sheets of the Italian railway company, FS.
Passenger output Y and freight output Q were measured by passenger-kilometers and ton-
kilometers, respectively. Since these data are collected monthly by FS, they were suitably aggre-
gated in order to be considered in the time series of the regression model.
The unit price of labour P L was derived from the available statistics as the outlay for that factor
divided by the quantity consumed of the same factor. An analogous computation was performed
to derive the unit price of fuel P F . Such a computation was possible because the consumption of
carbon, electricity and oil by FS are all expressed in terms of a common standard called TEP (see
Ferrovie dello Stato, 1996).
A number of operations were necessary to derive the unit price of equipment P E . In principle,
such an estimation would require to evaluate the capital stock of rail equipment and the cost
attributable to the use of that stock. Nevertheless, previous studies have found a very high cor-
relation between the unit price of locomotives and the unit prices of other equipments, such as
coaching stock, freight vehicles and freight containers (McGeehan, 1993; Braeutigam et al., 1982).
Thus, we felt it to be appropriate to carry out model estimation by representing the unit price of
equipment through the unit price of locomotives. In this framework, the purchase contracts of the
company were used to derive quarterly profiles of the number and types of locomotives in use, on
the basis of an estimate on the life of the assets. Since data on capital items are reported for the
purchase of an asset, it was necessary to convert the purchase price to an equivalent rental price
using information on the life of the asset and the interbank market interest rate. Thus, we defined
the unit price of equipment P E in the following manner: