I have looked at this on and off during the day. I am having trouble with the first equation.
If all of the savings are placed in the tax-sheltered option (meaning phi is 1), then utility in period 1 is
[MATH]u_1( y - s + \tau_1s) = u_1(y - (1 - \phi \tau_1)s).[/MATH]
And utility in the second period is
[MATH]u_2( \psi y + Rs - \tau_2Rs) = u_2( \psi y + (1 - \phi \tau_2) Rs).[/MATH]
Looks good if phi = 1.
But if none of the savings are placed in the tax-sheltered option (meaning phi is 0), then utility in the first period is still
[MATH]u_1(y - s) = u_1(y - (1 - \phi \tau_1)s[/MATH], but utility in the second period is
[MATH]u_2( \psi y + Rs - \tau_2rs ) > u_2( \psi y + Rs - phi \tau_2rs) > u_2(\psi y + (1 - \phi \tau_2)Rs)[/MATH].
Where does equation 1 come from?