Abstract
We provide an alternative estimation method which can be used to estimate the coefficient of intertemporal substitution and risk aversion separately, without assuming log-normal distribution or using a social market portfolio. We find the two parameters are significantly positive. This finding is the first case in this field.