The aim of this paper is to appraise a few of the key innovative features of the early work in compiling SAMs for development policy analysis; to set out and review some recent methodological advances; and to identify those areas where compilation continues to be problematic. It briefly re-visits the features of the SAM as an integrating framework and sets out its relationship to the SNA 1993. The main compilation problems faced in practice arise from assembling the household accounts from household survey data where income data are especially unreliable and are difficult to link to the factor accounts and to income transfers. Experience is drawn from the construction of a Ghana SAM. In the literature relatively more attention has been devoted to balancing and data reconciliation methods, which are briefly reviewed, although these are second order adjustments and much still depends on the quality of the initial estimates