when one takes into account the shifts in trucker and shipper
behavior: the assumed power has negligible effect on the MR&R
total revenue, but there are important distributional effects.
Furthermore, the MR&R total cost does not depend on the
assumed power. Pricing leads to large drops in annual traffic
loading, despite the increases in distances traveled. As a result,
pricing can lead to significant reductions in highway construction
costs. These results can be compared with the results obtained by
Johnsson (2004) using a computable general equilibrium model
for Sweden, keeping in mind that he calculates prices by
allocating the entire MR&R total cost to vehicles based on the
number of DEFs, so they are not MR&R marginal cost prices.
Johnsson finds that the MR&R total cost (and therefore the MR&R
total revenue) changes very little with the assumed power
(Johnsson, 2004) and concludes that ‘‘the cost of being wrong
appears to be modest’’. Although the findings in terms of MR&R
total revenue and MR&R total cost are very similar to those of
Anani (2008), the latter also looks at potential implications to
highway construction costs (by making the realistic assumption
that a highway agency constructs thinner pavement layers when
it expects lower traffic loading), and it studies distributional
effects. Other pricing studies attempt to look at the macroeco-
nomic impacts of pricing. For example, Doll and Schaffer (2007)
use an input–output model and find that the positive and negative
impacts on the German economy are almost equal, but they
recognize the significant microeconomic impacts.
A shortcoming of our paper is that it uses a simplistic method
to convert axle groups to equivalent single axles: a tandem
(tridem) axle group that weighs W is replaced with two (three)
single axles each weighing W/2 (W/3). Future research is needed
in order to make this conversion more realistic.