Income and expenditure measures are commonly, used to establish poverty lines representing, respectively, the availability of cash resources and the standard of living approaches to measuring the extent and composition of poverty. Using UK data we compare these two measures and show how the might be combined. Although overall poverty rates are similar whichever measure is used, the relativities they imply for different apes of household differ considerably. There is little overlap between income and expenditure poverty and very few households are both income- and expenditure-poor. The concept of poverty as constraint on choice or constrained expenditure is then defined as the absence of spending on durable goods and luxury items. Using logistic regression, income thresholds associated with the observed levels of constrained expenditure are derived for different types of household. Assuming all income is spent, these thresholds define a poverty line below which expenditure is severely constrained, Re extent to which social assistance rates limit or prevent household expenditure is also estimated, The method and the estimates illustrate the value of exploring the links between income and expenditure in the measurement of poverty, drawing attention to the limitations of the data, and identifying future research needs.