VAR(Ci-Cj) = VAR(Ci) + VAR(Cj) - 2 COV(Ci, Cj)
where VAR and COV stands for variance and covariance respectively and Ci and Cj are the
coefficients on a given variable for the ith and jth market. The standard error can be found by taking
the square root of the variance.
4. MARKETS AND SOURCES OF DATA
Three grain crops (maize, white tef, and white wheat) and eight markets are considered in this
analysis. Markets are chosen based on the availability of continuous time series price data
covering at least three years before and after market liberalization. The descriptive statistics cover
the period January 1986 to July 1996. However, the estimation period for the econometric
analysis covers the period from January 1986 to December 1993 due to the unavailability of
regionally-disaggregated food aid data after this time.
For maize six markets are considered, Addis Ababa, Dire Dawa and Mekele which are mainly
deficit markets for maize and Bako, Shashamane and Jimma which are surplus markets. For white
teff also six markets are analyzed: Addis Ababa, Dire Dawa and Mekele which are deficit
markets for white tef and Bako, Bahir Dar and Hosaenna which are surplus markets. In the case
of white wheat four markets are studied: Addis Ababa, Dire Dawa and Mekele which are deficit
markets and Hosaenna which is a surplus market.
The price data for the markets used in this analysis is obtained from the Ethiopian Grain marketing
Enterprise (EGTE) collected for selected markets in Ethiopia. For Bako market the price data
collected by Bako Research Center is used. Prices were deflated using the non-food consumer
price (1996=100) for Addis. The rainfall data for all markets except Bako market is obtained
from the National Meteorological Station of Ethiopia; for Bako area the rainfall data is obtained
from Bako Research Center. The food aid data is from the World Food Program (WFP) data files
and DPPC.
-8-