²⁸This paper assumes that each period corresponds to a quarter, while from Carroll (1992) is the value on an annual basis. Therefore, following Carroll, Slacalek, and Tokuoka (2008), 0.010 needs to be multiplied by since the variance of log transitory income shocks of quarterly data should be four times as large as that of annual data. (Note further that Carroll (1992)’s calibration of was considerably lower than his raw empirical estimate of , on the grounds that a substantial portion of the changes in measured income is likely to come from measurement error).