摘要：Food companies and analysts often need transportation rate data to explore market opportunities. In some cases, it may not be practical or necessary to obtain actual rates for all routes under consideration. This study provides analysis of truck rate patterns and alternative rate-estimating equations. The original objective was to provide shipping cost estimates for a national beef-marketing model involving 30 regions of the United States. The data set is 254 rates for refrigerated shipments of boxed fresh meat throughout the United States in 1988. Distance of the routes range from 50 to 2,923 miles. The average route is 1,181 miles, at a cost of $1,324 per load of $3.31 per cwt. The average cost per 100 miles was $0.28 per cwt. This simple average underestimates the short hauls and overestimates the long hauls. To improve on this simple procedure, both arithmetic and log equations were fitted to the data. Dummy variables were added to improve the estimates when shipping to different areas in the United States. Highest cost delivery for constant miles was for the areas including the Carolinas north through New England. The lowest cost delivery was the mid-section of the country. The estimating equations explained up to 79 percent of the actual cost. Most of the equations considered performed well in the 500 to 1,500-mile distance, but the quadratic and the double log functions, with dummy variables, performed the best over 2,500 miles. Obviously, these estimating methods are not sufficiently accurate (or necessary) for commercial transactions where only a few routes are under final consideration. Their main value should be to identify rate patterns and for market analysis and planning.
