traders. However, the magnitude of these checkpoint charges and their effects on prices received
by producers and prices paid by consumers (i.e., who ultimately pays the tax) is very unclear. To
overcome this information gap, enumerators were hired to ride on trucks carrying grain across
five major cereal trade routes: Addis to Harar; Addis to Mekele; Nekempt to Addis; Shashemene
to Addis, and Jimma to Addis. Enumerators rode on five different trucks on each trade route
during August 1996, and recorded information on the number of checkpoints encountered, time
spent at each checkpoint, and the tariff charges incurred, both registered and unregistered.
The number of checkpoints observed on each grain trading route varied from 8 to 18 and at each
check point the transporters were detained on average for about ten minutes (Table 7). The total
time taken for all checkpoint inspections from departure to destination varied from one hour to
three hours. The transporters paid 7.2 birr per quintal on average as check point taxes. There
was substantial variation in the magnitude of checkpoint taxes along different routes, being as low
as 3.8 birr per quintal from Shashemene to Addis, and as high as 15.0 birr per quintal from Addis
to Harar (including both official and unofficial charges). The checkpoint charges accounted for
20% to 33% of average price spread observed on these major grain trading routes.
The checkpoint charges are also considerable in terms of the prices received by the farmers,
accounting for roughly 10% of the price received by farmers for maize in Shashamene, and about
5% and 6% of the price received by farmers for teff and wheat in Dejene and Hosaenna
respectively for the months of July to August 1996.9
While the taxes support other government objectives, they increase grain marketing costs and
work against government efforts to stimulate incentives to use productivity-enhancing farm
technology. Other research has shown the value-cost ratio of fertilizer use on maize could be
increased by 8% in key producing regions if the elimination of checkpoint tariffs were half passed
on to producer prices (Mulat, Ali and Jayne 1996). Moreover, since the poor spend a
comparatively large proportion of their income on food, the taxation of grain is likely to be
regressive.
Results of Econometric Analysis
Changes in price levels due to liberalization and food aid releases
The analysis so far considered only the effect of liberalization on price levels and marketing
spreads without controlling for changes in other factors likely to affect price levels. In the
following sections the results of SURE estimations are presented to estimate the effect of
liberalization and food aid releases holding constant the influence of other factors.10
The effects of various factors affecting the maize price levels in six markets (Addis Ababa, Bako,
9 The producer prices are based on the new MIS data base run by GMRP.
10 The results of SURE model estimation under the different market integration assumptions are very
similar; here only the results under the assumption of weak market integration are discussed while the
results under the assumption of full market integration are given in Appendices 8 to 13.
-12-