Income smoothing of short line railroads

This study was undertaken to analyze the spending behavior of Class II railroads. It was hypothesized that expenses are functional to funds available beyond the influence of the cost determinants of traffic volume, traffic density, load density, length of haul, and wage rate. Variation in spending from the level predicted by these factors was studied with relation to operating revenues and current assets. Income smoothing was defined as spending manipulation to hold earnings near at normal levels. Increased accuracy in expenditure prediction by adding revenues or assets to the costing model was interpreted as evidence of the hypothesized smoothing behavior.

The study population was Class II railroads with independent ownership. Thus direct operating costs are reported as expenses. Data was taken from the carriers' reports to the Interstate Commerce Commission for the years 1968-1977. In all, 300 data points were transcribed representing 10 years operation for each of 30 short line railroads. These represent every member of the population for which reports were on file at the Commission's office in Washington, D.C.

A cost model was developed regressing operating statistics on various categories of railroad operating expense. The model estimates average elasticity with respect to each regressor and provides control figures for further analysis. Particularly noteworthy, the chosen model is robust with respect to carrier size and time period across the full range of the population. Separate time series and cross section hypothesis tests were conducted by comparing estimates of the control model to those generated from the same model after incorporating the influence of revenues and assets. Thus, smoothing between firms and within firms was measured.

The hypothesized behavior was substantiated for both cross section and time series dimensions. The categories of expense differed in how they exhibited smoothing. Cross section revenue effects between firms were greatest for maintenance of way and traffic expenditures, while time series smoothing was concentrated in traffic and equipment maintenance. Current assets significantly improved the revenue predictions only for maintenance of way in the cross section test. Thus, it appears firms spend what they can afford on track every year and smooth income by spending on or deferring equipment maintenance.

The significance of the study is essentially its method. The control model for cost estimation is superior to those in the literature and offers promise for application to railroads of larger size. The spending explanation technique for analyzing income smoothing can be applied to other industries for which cost control modeling is feasible. The findings are hardly unanticipated, but the coefficients developed allow for predictions about short line railroad spending responses to changes in traffic, rates, or even subsidies.