Measuring the return on household enterprise: What matters most for whom?

Return on assets (ROA) from household enterprise is crucial for understanding the well-being and productivity of households in developing economies. Yet the definition and measurement of household enterprise ROA remain inconsistent or unclear. We illustrate potential measurement problems with examples from various actual surveys. We then take advantage of a detailed integrated household survey to perform a robustness analysis, acting as if we had gathered less data than was actually the case, to see what matters and for whom. The three issues that matter most for accurate measurement of household enterprise ROA are the choice of accrual versus cash basis of income, the treatment of household's own labor in enterprise income, and the treatment of non-factor income. Also, this sensitivity matters most for a relatively poor region dominated by crop cultivation relative to a richer region with non-farm enterprises. Though the choice between accrued income and cash income matters less when the frequency of the data declines, there remains high sensitivity in longer-term and annualized data. We conclude the paper by providing recommendations on how to improve the survey questionnaires for more accurate measurement in field research.