This study starts from a Social Accounting Matrix (SAM) based on 1982 U.S. data,, using a sector aggregation designed for examining agriculture. Multipliers are derived which measure the impact on demand and institutional incomes of .changes in government expenditure and exports. To explore the nature of intersectoral and inter-institutional structure, a multiplier decompositon is derived which separates the total multiplier into components measuring the contribution of input-output linkages and net-SAM linkages. The decomposition calculations indicate that leakages from agriculture to the rest of the economy are very large and that leakages back into agriculture from the rest of the economy are very small. Input-output effects typically account for only 15 percent of the overall multiplier on agricultural gross output. Policy experiments with increases in agricultural exports, income increases in agriculture resulting from transfers, increases in nonagricultural exports, and increases in economywide household incomes are presented. We find that increases in agricultural value added are most sensitive to transfers, next most to agricultural exports, and least to measures designed to improve economywide prosperity. Extensions of the SAM framework to a Computable General Equilibrium model are discussed. We conclude that such an extension is a desireable next step in the research agenda.