This paper formulates a permanent income model of consumer expenditures. Through use of this model, measures of permanent income are examined that are alternatively based on (a) measured income patterns, (b) stocks of consumer liquid assets, (c) stocks of consumer financial assets, and (d) stocks of consumer real and financial assets. Using assets in permanent income measures is found to improve upon using measured income patterns in explaining consumer expenditure. Consumer assets are further found not to be treated as one homogeneous stock by consumers. Money and risky financial assets contribute significantly to explaining consumer spending, while savings and real assets are of no value. Finally, asset effects on consumption are found to be inelastic but statistically significant.