¹⁵Because permanent income was not directly measurable, Friedman anticipated that researchers would challenge him to propose a measurable proxy for permanent income. One of his proposals for such a measure was average household income over a three year period. Friedman was explicit in rejecting the idea that permanent income should defined as “lifetime” income, as Modigliani and Brumberg (1954) had proposed; this seems to be because he had the (correct) intuition that liquidity constraints and uncertainty could prevent distant future income from influencing current choices (Carroll (2001)).