The addition of illumination where none was present is generally believed to have a positive effect on motor vehicle safety; reducing the frequency, as well as the severity of crashes. The operational cost of illumination, however, can make it a candidate for conservation during periods of high energy costs. In response to a forecasted energy shortage, the Oregon Department of Transportation selectively reduced illumination on interstate highways as part of an energy-saving effort. The reductions occurred at 44 interchanges and along 5.5 miles of interstate highway. This paper presents the results of a crash-based analysis of the changes in safety performance using an empirical-Bayes observational methodology. The study found an increase in reported crashes where the lineal lighting was reduced both in total crashes (28.95%, P=0.05) and injury night crashes (39.21%, P=0.07). Where full interchange lighting was reduced to partial lighting, a 2.46% increase (P=0.007) in total night crashes was observed. Injury night crashes, however, decreased by 12.16% (P<0.001) though day injury crashes also decreased at these locations. Unexpectedly, for interchanges where illumination was reduced from partial plus to partial, a 35.24% decrease (P<0.001) in total crashes and 39.98 (P<0.001) decrease in injury night crashes was found, though again, day crashes also decreased.