This issue is to suggest and discuss the implementation of transitory policies for non-IIT policy changes. In particular, we should consider allowing all policy parameters to vary flexibly over the time path. These include:
tau_bdelta_tauretirement_agetau_bqtau_payroll- wealth tax parameters:
h_wealth, m_wealth, p_wealth
alpha_T and alpha_G (which would mean we can remove the T_shifts and G_shifts parameters)
We might also consider allowing some economic parameters to vary over the time path, such as:
Finally, given how social security retirement income replacement rates are computed from the SS, we should probably have a parameter, call it theta_adjust, that would adjust the replacement rate by some fraction for all periods in the time path (e.g., theta_adjust = [1.0, 1.0, 0.67, 0.67, 0.67] would scale down replacement rates by 1/3 beginning in the third period in the time path).
My proposition is that these parameters are specified as lists. The first value in the list corresponds to the value of that parameter in the first year of the model, the second value the second year, and so on. Values will not need to be specified for all T periods, but rather just until the parameter value is constant. e.g., tau_b = [0.35, 0.35, 0.21] would mean set the business tax rate equal to 35% in the first two periods of the model and 21% for all periods after. SS values will thus be inferred as the last value in this list.
@rickecon thoughts?
This issue is to suggest and discuss the implementation of transitory policies for non-IIT policy changes. In particular, we should consider allowing all policy parameters to vary flexibly over the time path. These include:
tau_bdelta_tauretirement_agetau_bqtau_payrollh_wealth,m_wealth,p_wealthalpha_Tandalpha_G(which would mean we can remove theT_shiftsandG_shiftsparameters)We might also consider allowing some economic parameters to vary over the time path, such as:
Zg_yworld_int_rateFinally, given how social security retirement income replacement rates are computed from the SS, we should probably have a parameter, call it
theta_adjust, that would adjust the replacement rate by some fraction for all periods in the time path (e.g.,theta_adjust = [1.0, 1.0, 0.67, 0.67, 0.67]would scale down replacement rates by 1/3 beginning in the third period in the time path).My proposition is that these parameters are specified as lists. The first value in the list corresponds to the value of that parameter in the first year of the model, the second value the second year, and so on. Values will not need to be specified for all
Tperiods, but rather just until the parameter value is constant. e.g.,tau_b = [0.35, 0.35, 0.21]would mean set the business tax rate equal to 35% in the first two periods of the model and 21% for all periods after. SS values will thus be inferred as the last value in this list.@rickecon thoughts?