Shashamane, Dire Dawa, Mekele and Jimma) is estimated as a system using SURE model and the
results are given in Table 8. For all the maize markets considered except Mekele the signs of the
coefficients on the liberalization dummy variable is found to be positive, implying that market
liberalization has increased the maize price levels in real terms. However, the increases are
significant only in the major surplus-producing areas of Shashamane and Bako markets. In these
markets, equilibrium maize prices have been 11 and 4 birr per quintal higher (in constant 1996
birr) in the post-liberalization period. For the deficit market of Mekele, market liberalization was
associated with a significant decline of 28 birr per quintal (constant 1996 birr) in equilibrium
maize price levels. These results may also be due to the ending of the civil war, which occurred
a year after liberalization occurred, and was especially disruptive in the northern part of the
country (Dercon 1994).
The impact of food aid on maize price levels was insignificant in all six markets for which data
were available. Given the magnitude of imported food aid wheat over the estimation period, these
results are perhaps surprising, but may be plausible if wheat and maize are not close substitutes
in consumption. The effect of rainfall on maize price levels is found to be, as expected, negative
in all cases. The impact of rainfall was statistically significant at a probability of less than 1% only
for Shashamane, at a probability of less than 5% for Addis Ababa and Mekele markets and at a
probability of less than 10% for Dire Dawa. Surprisingly, the effect of rainfall on maize price
level in the heavy rainfall areas like Bako and Jimma is not significant.
In the case of white teff, the effect of liberalization is negative in all three deficit markets and
positive in all three surplus markets. However, the effects of market liberalization is statistically
significant only for the surplus areas of Hosaenna and Bahir Dar at a probability of less than 1%.
In these markets, market liberalization was associated with a 25 b/q and 17 b/q rise in equilibrium
teff prices. Market liberalization was associated with a decline in teff prices (from 1 to 48 birr per
quintal) in the deficit markets of Addis Ababa, Dire Dawa and Mekele. The impact of
liberalization was statistically significant at the 1% level only for Mekele market.
A statistically significant seasonal effect is also observed for all the markets considered, January
through April being months of lower price for most of the markets (Table 9). The effect of
rainfall is negative for all markets considered. However, the effect is statistically significant only
for Addis Ababa, Bako and Mekele markets at a probability of less than 1%.
The volume of food aid released is negatively associated with white teff prices in four of six cases,
being statistically significant at the 5% level for Addis Ababa and Mekele markets. For these two
areas, wholesale prices in a given region and a given month declined by 2 to 5 birr per quintal for
every additional 30,000 quintals of food aid released within that region over the prior three-month
period. The volume of imported food aid wheat has ranged from 0.3 to over 1.1 million tons
since 1985, accounting for an estimated 20 to 50 percent of national annual marketed cereal
output. In some cases such as Tigray, the volume of food aid was of such a magnitude to depress
wholesale teff prices by 15% to 25%, other factors constant. The welfare effects of lower grain
prices (due to food aid) on food production incentives, input use, and rural livelihoods are
complex and clearly differ among different types of rural households. A large percentage of rural
households are net buyers of cereal on an annual basis (in 1995/96 this percentage was almost
50% nationwide); these households directly benefit from lower staple food prices. However,
lower prices due to food aid may impede input use and cereal production by rural households who
-13-