New evidence on income and the velocity of money

Time series and cross-country empirical results suggest that cash holding as a percentage of income rises, or alternatively that velocity falls, as income increases. Numerous cross-sectional findings at many points in time, in several countries conclude oppositely. It is argued here that the former findings suffer from omitted variable bias by ignoring socio-demographic variables affecting the demand for cash balances. When one incorporates such demand shifters into the analysis the time series and cross-country are seen as consistent with the critically reexamined result that velocity increases with income.