This paper is the first empirical analysis of U.S. short line railroad prof
itability using primary cost and revenue data. Models of short line profita
bility are developed using Earnings Before Interest and Taxes (EBIT) as the
profitability measure. The sample includes 34 short lines operating in 17
states in the midwest region of the U.S. for the fiscal years 1986-1995. Th
e models are estimated by OLS regression and explain up to 75% of the varia
tion in shea. line profitability. Nearly all the explanatory variables have
the theoretically expected sign and are statistically significant. The var
iable DENS (number of carloads per mile of main-line track) is the most imp
ortant influence on EBIT. However, several other variables are identified a
s important including management ability to control expenses, type of short
line ownership, size and ownership of the short line's network, compositio
n of traffic, and length of haul. The empirical results of the study indica
te that a short line operating at the mean values of the explanatory variab
les is likely to only break even. About 25% of the sample short lines have
a high probability of requiring government assistance to continue operating
. (C) 1998 Elsevier Science Ltd. All rights reserved.