Agriculture accounts for over half of Ethiopian GDP, yet the case for agriculture as a focus of economic growth strategies must rely on identifying a set of intersectoral linkages through which agricultural growth contributes to the growth of nonagriculture in the Ethiopian economy. This article develops a four-sector numerical simulation model of economic growth in Ethiopia which permits the calculation of macroeconomic growth multipliers resulting from income shocks to agriculture, services, modern industry, and traditional industry. The resulting growth multipliers are 1.54 for agriculture, 1.80 for services, 1.34 for modern industry, and 1.22 for traditional industry. These results depict an economy in which intersectoral linkages operate on a highly uneven basis. These limits are reflected in the wide disparity between sectoral growth multipliers and by substantial differences in the patterns of their decomposition. The policy relevance of these findings relate, in part, to the distributional implications of growth in particular sectors. Poverty in Ethiopia is disproportionately rural. An income shock to agriculture is clearly the most progressive choice, indicating the need to highlight agricultural development in growth strategies for Ethiopia. Yet, the simulation results further indicate that doing so imposes relatively little trade off against total benefit. While a $1 service sector income shock generates $0.80 in indirect benefits, a $1 agricultural income shock still generates $0.54 in indirect gains - a somewhat smaller benefit, bnt one likely to make the greatest possible impact on poverty reduction. © 1999 Elsevier Science B.V. All rights reserved.