An Analysis of Grain Production Decline During the Early Transition in Ukraine: Bayesian Inference

Introduction The first years of reforms in the former Soviet Union resulted in a sharp decline in agricultural production. Between 1989 and 1992, Ukrainian gross agricultural product dropped some 20%, and the output of grain, which is produced mostly in the collective sector, went down some 25% (World Bank). Several reasons for the fall have been advanced. In addition to weather variation across years, there was a drop in state deliveries of production inputs, especially fertilizer and pesticides (World Bank). The emergence of private opportunities resulted in labor and management migration from the large-scale collective system (Csaki and Lerman), thus reducing the average labor qua lity in the sector. The ability of farms produce efficiently was also affected by the transition-related break in old production ties and networks (Blanchard and Kremer). While all three explanations of production differences between years (weather variability, input quantities decline, and loss of production efficiency) are plausible, little is known about the relative contribution of these factors to the overall decline in production. Identifying the reasons for the decline in agricultural output in count ries like Ukraine is important to predict the likelihood of success of new policies and initiatives, such as accession to the European Union (Macours and Swinnen). We attempt to fill in the gap and improve the understanding of the decline in grain production in early transition Ukraine. Much of the literature devoted to the analyses of the changes in agricultural production in the former Soviet Union focuses on the later transition years and uses data aggregated both across outputs and farms (e. Arnade). Macours and Swinnen used country-level agricultural production data from several Central and Eastern European countries and found that severe drought and reductions in production factor use accounted for around 10% and 80%, respectively, of 3 total output decline between 1989 and 1995. Kurkalova and Jensen, who used farm-level survey data of Ukrainian farms, estimated that the decline in input quantities and technical efficiency decline contributed about half each to the overall drop of grain production between 1989 and 1992. Neither of the last two studies provided standard errors on the estimates of the shares of the components of output decline. Yet knowing the standard errors is important as underestimating the estimation error may lead to unrealistically precise inferences. This paper extends the analysis of Kurkalova and Jensen by (1) focusing specifically on the decomposition of output decline into …