''' get_baseline = True calibrate_model = False #taxes = 'baseline' # policy# #none ## some flag for do we run without taxes, with baseline # taxes or do we use user inputted taxes? #globals().update(ogusa.parameters.get_parameters_from_file()) globals().update(ogusa.parameters.get_parameters(baseline=True, guid='abc123')) # Generate Wealth data moments output_dir = "./OUTPUT" wealth.get_wealth_data(lambdas, J, flag_graphs, output_dir) # Generate labor data moments labor.labor_data_moments(flag_graphs) # List of parameter names that will not be changing (unless we decide to # change them for a tax experiment) param_names = ['S', 'J', 'T', 'BW', 'lambdas', 'starting_age', 'ending_age', 'beta', 'sigma', 'alpha', 'nu', 'Z', 'delta', 'E', 'ltilde', 'g_y', 'maxiter', 'mindist_SS', 'mindist_TPI', 'b_ellipse', 'k_ellipse', 'upsilon', 'chi_b_guess', 'chi_n_guess','etr_params','mtrx_params', 'mtry_params','tau_payroll', 'tau_bq', 'calibrate_model', 'retire', 'mean_income_data', 'g_n_vector', 'h_wealth', 'p_wealth', 'm_wealth', 'get_baseline', 'omega', 'g_n_ss', 'omega_SS', 'surv_rate', 'e', 'rho'] '''
def runner(output_base, input_dir, baseline=False, analytical_mtrs=True, reform={}, user_params={}, guid='', run_micro=True): from ogusa import parameters, wealth, labor, demographics, income from ogusa import txfunc tick = time.time() #Create output directory structure saved_moments_dir = os.path.join(output_base, "Saved_moments") ssinit_dir = os.path.join(output_base, "SSinit") tpiinit_dir = os.path.join(output_base, "TPIinit") dirs = [saved_moments_dir, ssinit_dir, tpiinit_dir] for _dir in dirs: try: print "making dir: ", _dir os.makedirs(_dir) except OSError as oe: pass if run_micro: txfunc.get_tax_func_estimate(baseline=baseline, analytical_mtrs=analytical_mtrs, reform=reform, guid=guid) print("in runner, baseline is ", baseline) run_params = ogusa.parameters.get_parameters(baseline=baseline, guid=guid) run_params['analytical_mtrs'] = analytical_mtrs # Modify ogusa parameters based on user input if 'frisch' in user_params: print "updating fricsh and associated" b_ellipse, upsilon = ogusa.elliptical_u_est.estimation( user_params['frisch'], run_params['ltilde']) run_params['b_ellipse'] = b_ellipse run_params['upsilon'] = upsilon run_params.update(user_params) # Modify ogusa parameters based on user input if 'g_y_annual' in user_params: print "updating g_y_annual and associated" g_y = (1 + user_params['g_y_annual'])**( float(ending_age - starting_age) / S) - 1 run_params['g_y'] = g_y run_params.update(user_params) globals().update(run_params) from ogusa import SS, TPI # Generate Wealth data moments wealth.get_wealth_data(lambdas, J, flag_graphs, output_dir=input_dir) # Generate labor data moments labor.labor_data_moments(flag_graphs, output_dir=input_dir) get_baseline = True calibrate_model = False # List of parameter names that will not be changing (unless we decide to # change them for a tax experiment) param_names = [ 'S', 'J', 'T', 'BW', 'lambdas', 'starting_age', 'ending_age', 'beta', 'sigma', 'alpha', 'nu', 'Z', 'delta', 'E', 'ltilde', 'g_y', 'maxiter', 'mindist_SS', 'mindist_TPI', 'analytical_mtrs', 'b_ellipse', 'k_ellipse', 'upsilon', 'chi_b_guess', 'chi_n_guess', 'etr_params', 'mtrx_params', 'mtry_params', 'tau_payroll', 'tau_bq', 'calibrate_model', 'retire', 'mean_income_data', 'g_n_vector', 'h_wealth', 'p_wealth', 'm_wealth', 'get_baseline', 'omega', 'g_n_ss', 'omega_SS', 'surv_rate', 'e', 'rho' ] ''' ------------------------------------------------------------------------ Run SS with minimization to fit chi_b and chi_n ------------------------------------------------------------------------ ''' # This is the simulation before getting the replacement rate values sim_params = {} glbs = globals() lcls = locals() for key in param_names: if key in glbs: sim_params[key] = glbs[key] else: sim_params[key] = lcls[key] sim_params['output_dir'] = input_dir sim_params['run_params'] = run_params income_tax_params, wealth_tax_params, ellipse_params, ss_parameters, iterative_params = SS.create_steady_state_parameters( **sim_params) ss_outputs = SS.run_steady_state(income_tax_params, ss_parameters, iterative_params, get_baseline, calibrate_model, output_dir=input_dir) ''' ------------------------------------------------------------------------ Run the baseline TPI simulation ------------------------------------------------------------------------ ''' ss_outputs['get_baseline'] = get_baseline sim_params['input_dir'] = input_dir income_tax_params, wealth_tax_params, ellipse_params, parameters, N_tilde, omega_stationary, K0, b_sinit, \ b_splus1init, L0, Y0, w0, r0, BQ0, T_H_0, tax0, c0, initial_b, initial_n = TPI.create_tpi_params(**sim_params) ss_outputs['income_tax_params'] = income_tax_params ss_outputs['wealth_tax_params'] = wealth_tax_params ss_outputs['ellipse_params'] = ellipse_params ss_outputs['parameters'] = parameters ss_outputs['N_tilde'] = N_tilde ss_outputs['omega_stationary'] = omega_stationary ss_outputs['K0'] = K0 ss_outputs['b_sinit'] = b_sinit ss_outputs['b_splus1init'] = b_splus1init ss_outputs['L0'] = L0 ss_outputs['Y0'] = Y0 ss_outputs['r0'] = r0 ss_outputs['BQ0'] = BQ0 ss_outputs['T_H_0'] = T_H_0 ss_outputs['tax0'] = tax0 ss_outputs['c0'] = c0 ss_outputs['initial_b'] = initial_b ss_outputs['initial_n'] = initial_n ss_outputs['tau_bq'] = tau_bq ss_outputs['g_n_vector'] = g_n_vector ss_outputs['output_dir'] = input_dir with open("ss_outputs.pkl", 'wb') as fp: pickle.dump(ss_outputs, fp) w_path, r_path, T_H_path, BQ_path, Y_path = TPI.run_time_path_iteration( **ss_outputs) print "getting to here...." TPI.TP_solutions(w_path, r_path, T_H_path, BQ_path, **ss_outputs) print "took {0} seconds to get that part done.".format(time.time() - tick)
def test_sstpi(): import tempfile import pickle import numpy as np import numpy as np import cPickle as pickle import os import ogusa ogusa.parameters.DATASET = "REAL" from ogusa.utils import comp_array from ogusa.utils import comp_scalar from ogusa.utils import dict_compare from ogusa.utils import pickle_file_compare import ogusa.SS import ogusa.TPI from ogusa import parameters, wealth, labor, demographics, income, SS, TPI globals().update(ogusa.parameters.get_parameters()) # Generate Wealth data moments output_dir = TEST_OUTPUT input_dir = "./OUTPUT" wealth.get_wealth_data(lambdas, J, flag_graphs, output_dir) # Generate labor data moments labor.labor_data_moments(flag_graphs, output_dir=output_dir) get_baseline = True calibrate_model = False # List of parameter names that will not be changing (unless we decide to # change them for a tax experiment) param_names = [ "S", "J", "T", "lambdas", "starting_age", "ending_age", "beta", "sigma", "alpha", "nu", "Z", "delta", "E", "ltilde", "g_y", "maxiter", "mindist_SS", "mindist_TPI", "b_ellipse", "k_ellipse", "upsilon", "a_tax_income", "chi_b_guess", "chi_n_guess", "b_tax_income", "c_tax_income", "d_tax_income", "tau_payroll", "tau_bq", "calibrate_model", "retire", "mean_income_data", "g_n_vector", "h_wealth", "p_wealth", "m_wealth", "get_baseline", "omega", "g_n_ss", "omega_SS", "surv_rate", "e", "rho", ] """ ------------------------------------------------------------------------ Run SS with minimization to fit chi_b and chi_n ------------------------------------------------------------------------ """ # This is the simulation before getting the replacement rate values sim_params = {} for key in param_names: try: sim_params[key] = locals()[key] except KeyError: sim_params[key] = globals()[key] sim_params["output_dir"] = output_dir sim_params["input_dir"] = input_dir income_tax_params, wealth_tax_params, ellipse_params, ss_parameters, iterative_params = SS.create_steady_state_parameters( **sim_params ) ss_outputs = SS.run_steady_state( ss_parameters, iterative_params, get_baseline, calibrate_model, output_dir=output_dir ) """ ------------------------------------------------------------------------ Run the baseline TPI simulation ------------------------------------------------------------------------ """ ss_outputs["get_baseline"] = get_baseline income_tax_params, wealth_tax_params, ellipse_params, parameters, N_tilde, omega_stationary, K0, b_sinit, b_splus1init, L0, Y0, w0, r0, BQ0, T_H_0, tax0, c0, initial_b, initial_n = TPI.create_tpi_params( **sim_params ) ss_outputs["output_dir"] = output_dir ss_outputs["income_tax_params"] = income_tax_params ss_outputs["wealth_tax_params"] = wealth_tax_params ss_outputs["ellipse_params"] = ellipse_params ss_outputs["parameters"] = parameters ss_outputs["N_tilde"] = N_tilde ss_outputs["omega_stationary"] = omega_stationary ss_outputs["K0"] = K0 ss_outputs["b_sinit"] = b_sinit ss_outputs["b_splus1init"] = b_splus1init ss_outputs["L0"] = L0 ss_outputs["Y0"] = Y0 ss_outputs["r0"] = r0 ss_outputs["BQ0"] = BQ0 ss_outputs["T_H_0"] = T_H_0 ss_outputs["tax0"] = tax0 ss_outputs["c0"] = c0 ss_outputs["initial_b"] = initial_b ss_outputs["initial_n"] = initial_n ss_outputs["tau_bq"] = tau_bq ss_outputs["g_n_vector"] = g_n_vector TPI.run_time_path_iteration(**ss_outputs) # Platform specific exceptions: if sys.platform == "darwin": exceptions = {"tax_path": 0.08, "c_path": 0.02, "b_mat": 0.0017, "solutions": 0.005} else: exceptions = {} # compare results to test data for old, new in zip(oldfiles, newfiles): print "trying a pair" print old, new assert pickle_file_compare(old, new, exceptions=exceptions, relative=True) print "next pair"
------------------------------------------------------------------------ Setup ------------------------------------------------------------------------ get_baseline = Flag to run baseline or tax experiments (bool) calibrate_model = Flag to run calibration of chi values or not (bool) ------------------------------------------------------------------------ ''' globals().update(ogusa.parameters.get_parameters()) # Generate Wealth data moments output_dir = "./OUTPUT" wealth.get_wealth_data(lambdas, J, flag_graphs, output_dir) # Generate labor data moments labor.labor_data_moments(flag_graphs) get_baseline = True calibrate_model = False # List of parameter names that will not be changing (unless we decide to # change them for a tax experiment) param_names = [ 'S', 'J', 'T', 'lambdas', 'starting_age', 'ending_age', 'beta', 'sigma', 'alpha', 'nu', 'Z', 'delta', 'E', 'ltilde', 'g_y', 'maxiter', 'mindist_SS', 'mindist_TPI', 'b_ellipse', 'k_ellipse', 'upsilon', 'a_tax_income', 'chi_b_guess', 'chi_n_guess', 'b_tax_income', 'c_tax_income', 'd_tax_income', 'tau_payroll', 'tau_bq', 'calibrate_model', 'retire', 'mean_income_data', 'g_n_vector', 'h_wealth', 'p_wealth', 'm_wealth', 'get_baseline', 'omega', 'g_n_ss', 'omega_SS', 'surv_rate', 'e', 'rho' ]
def runner(output_base, baseline_dir, baseline=False, analytical_mtrs=True, age_specific=False, reform={}, user_params={}, guid='', run_micro=True): #from ogusa import parameters, wealth, labor, demographics, income from ogusa import parameters, wealth, labor, demog, income, utils from ogusa import txfunc tick = time.time() #Create output directory structure saved_moments_dir = os.path.join(output_base, "Saved_moments") ssinit_dir = os.path.join(output_base, "SSinit") tpiinit_dir = os.path.join(output_base, "TPIinit") dirs = [saved_moments_dir, ssinit_dir, tpiinit_dir] for _dir in dirs: try: print "making dir: ", _dir os.makedirs(_dir) except OSError as oe: pass if run_micro: txfunc.get_tax_func_estimate(baseline=baseline, analytical_mtrs=analytical_mtrs, age_specific=age_specific, start_year=user_params['start_year'], reform=reform, guid=guid) print ("in runner, baseline is ", baseline) run_params = ogusa.parameters.get_parameters(baseline=baseline, guid=guid) run_params['analytical_mtrs'] = analytical_mtrs # Modify ogusa parameters based on user input if 'frisch' in user_params: print "updating fricsh and associated" b_ellipse, upsilon = ogusa.elliptical_u_est.estimation(user_params['frisch'], run_params['ltilde']) run_params['b_ellipse'] = b_ellipse run_params['upsilon'] = upsilon run_params.update(user_params) # Modify ogusa parameters based on user input if 'g_y_annual' in user_params: print "updating g_y_annual and associated" g_y = (1 + user_params['g_y_annual'])**(float(ending_age - starting_age) / S) - 1 run_params['g_y'] = g_y run_params.update(user_params) from ogusa import SS, TPI # Generate Wealth data moments wealth.get_wealth_data(run_params['lambdas'], run_params['J'], run_params['flag_graphs'], output_dir=output_base) # Generate labor data moments labor.labor_data_moments(run_params['flag_graphs'], output_dir=output_base) calibrate_model = False # List of parameter names that will not be changing (unless we decide to # change them for a tax experiment) param_names = ['S', 'J', 'T', 'BW', 'lambdas', 'starting_age', 'ending_age', 'beta', 'sigma', 'alpha', 'nu', 'Z', 'delta', 'E', 'ltilde', 'g_y', 'maxiter', 'mindist_SS', 'mindist_TPI', 'analytical_mtrs', 'b_ellipse', 'k_ellipse', 'upsilon', 'chi_b_guess', 'chi_n_guess','etr_params','mtrx_params', 'mtry_params','tau_payroll', 'tau_bq', 'retire', 'mean_income_data', 'g_n_vector', 'h_wealth', 'p_wealth', 'm_wealth', 'omega', 'g_n_ss', 'omega_SS', 'surv_rate', 'e', 'rho'] ''' ------------------------------------------------------------------------ Run SS ------------------------------------------------------------------------ ''' sim_params = {} for key in param_names: sim_params[key] = run_params[key] sim_params['output_dir'] = output_base sim_params['run_params'] = run_params income_tax_params, ss_parameters, iterative_params, chi_params = SS.create_steady_state_parameters(**sim_params) ss_outputs = SS.run_SS(income_tax_params, ss_parameters, iterative_params, chi_params, baseline, baseline_dir=baseline_dir) ''' ------------------------------------------------------------------------ Pickle SS results ------------------------------------------------------------------------ ''' if baseline: utils.mkdirs(os.path.join(baseline_dir, "SS")) ss_dir = os.path.join(baseline_dir, "SS/ss_vars.pkl") pickle.dump(ss_outputs, open(ss_dir, "wb")) else: utils.mkdirs(os.path.join(output_dir, "SS")) ss_dir = os.path.join(output_dir, "SS/ss_vars.pkl") pickle.dump(ss_outputs, open(ss_dir, "wb")) ''' ------------------------------------------------------------------------ Run the baseline TPI simulation ------------------------------------------------------------------------ ''' sim_params['input_dir'] = output_base sim_params['baseline_dir'] = baseline_dir income_tax_params, tpi_params, iterative_params, initial_values, SS_values = TPI.create_tpi_params(**sim_params) # ss_outputs['income_tax_params'] = income_tax_params # ss_outputs['wealth_tax_params'] = wealth_tax_params # ss_outputs['ellipse_params'] = ellipse_params # ss_outputs['parameters'] = parameters # ss_outputs['N_tilde'] = N_tilde # ss_outputs['omega_stationary'] = omega_stationary # ss_outputs['K0'] = K0 # ss_outputs['b_sinit'] = b_sinit # ss_outputs['b_splus1init'] = b_splus1init # ss_outputs['L0'] = L0 # ss_outputs['Y0'] = Y0 # ss_outputs['r0'] = r0 # ss_outputs['BQ0'] = BQ0 # ss_outputs['T_H_0'] = T_H_0 # ss_outputs['factor_ss'] = factor # ss_outputs['tax0'] = tax0 # ss_outputs['c0'] = c0 # ss_outputs['initial_b'] = initial_b # ss_outputs['initial_n'] = initial_n # ss_outputs['tau_bq'] = tau_bq # ss_outputs['g_n_vector'] = g_n_vector # ss_outputs['output_dir'] = output_base # with open("ss_outputs.pkl", 'wb') as fp: # pickle.dump(ss_outputs, fp) w_path, r_path, T_H_path, BQ_path, Y_path = TPI.run_TPI(income_tax_params, tpi_params, iterative_params, initial_values, SS_values, output_dir=output_base) print "getting to here...." TPI.TP_solutions(w_path, r_path, T_H_path, BQ_path, **ss_outputs) print "took {0} seconds to get that part done.".format(time.time() - tick)
def test_sstpi(): import tempfile import pickle import numpy as np import numpy as np import pickle as pickle import os import ogusa ogusa.parameters.DATASET = 'REAL' from ogusa.utils import comp_array from ogusa.utils import comp_scalar from ogusa.utils import dict_compare from ogusa.utils import pickle_file_compare import ogusa.SS import ogusa.TPI from ogusa import parameters, wealth, labor, demographics, income, SS, TPI globals().update(ogusa.parameters.get_parameters()) # Generate Wealth data moments output_dir = TEST_OUTPUT input_dir = "./OUTPUT" wealth.get_wealth_data(lambdas, J, flag_graphs, output_dir) # Generate labor data moments labor.labor_data_moments(flag_graphs, output_dir=output_dir) get_baseline = True calibrate_model = False # List of parameter names that will not be changing (unless we decide to # change them for a tax experiment) param_names = [ 'S', 'J', 'T', 'lambdas', 'starting_age', 'ending_age', 'beta', 'sigma', 'alpha', 'nu', 'Z', 'delta', 'E', 'ltilde', 'g_y', 'maxiter', 'mindist_SS', 'mindist_TPI', 'b_ellipse', 'k_ellipse', 'upsilon', 'a_tax_income', 'chi_b_guess', 'chi_n_guess', 'b_tax_income', 'c_tax_income', 'd_tax_income', 'tau_payroll', 'tau_bq', 'calibrate_model', 'retire', 'mean_income_data', 'g_n_vector', 'h_wealth', 'p_wealth', 'm_wealth', 'get_baseline', 'omega', 'g_n_ss', 'omega_SS', 'surv_rate', 'e', 'rho' ] ''' ------------------------------------------------------------------------ Run SS with minimization to fit chi_b and chi_n ------------------------------------------------------------------------ ''' # This is the simulation before getting the replacement rate values sim_params = {} for key in param_names: try: sim_params[key] = locals()[key] except KeyError: sim_params[key] = globals()[key] sim_params['output_dir'] = output_dir sim_params['input_dir'] = input_dir income_tax_params, wealth_tax_params, ellipse_params, ss_parameters, \ iterative_params = SS.create_steady_state_parameters(**sim_params) ss_outputs = SS.run_steady_state(ss_parameters, iterative_params, get_baseline, calibrate_model, output_dir=output_dir) ''' ------------------------------------------------------------------------ Run the baseline TPI simulation ------------------------------------------------------------------------ ''' ss_outputs['get_baseline'] = get_baseline income_tax_params, wealth_tax_params, ellipse_params, parameters, N_tilde, omega_stationary, K0, b_sinit, \ b_splus1init, L0, Y0, w0, r0, BQ0, T_H_0, tax0, c0, initial_b, initial_n = TPI.create_tpi_params( **sim_params) ss_outputs['output_dir'] = output_dir ss_outputs['income_tax_params'] = income_tax_params ss_outputs['wealth_tax_params'] = wealth_tax_params ss_outputs['ellipse_params'] = ellipse_params ss_outputs['parameters'] = parameters ss_outputs['N_tilde'] = N_tilde ss_outputs['omega_stationary'] = omega_stationary ss_outputs['K0'] = K0 ss_outputs['b_sinit'] = b_sinit ss_outputs['b_splus1init'] = b_splus1init ss_outputs['L0'] = L0 ss_outputs['Y0'] = Y0 ss_outputs['r0'] = r0 ss_outputs['BQ0'] = BQ0 ss_outputs['T_H_0'] = T_H_0 ss_outputs['tax0'] = tax0 ss_outputs['c0'] = c0 ss_outputs['initial_b'] = initial_b ss_outputs['initial_n'] = initial_n ss_outputs['tau_bq'] = tau_bq ss_outputs['g_n_vector'] = g_n_vector TPI.run_time_path_iteration(**ss_outputs) # Platform specific exceptions: if sys.platform == "darwin": exceptions = { 'tax_path': 0.08, 'c_path': 0.02, 'b_mat': 0.0017, 'solutions': 0.005 } else: exceptions = {} # compare results to test data for old, new in zip(oldfiles, newfiles): print("trying a pair") print(old, new) assert pickle_file_compare(old, new, exceptions=exceptions, relative=True) print("next pair")
def runner(output_base, input_dir, baseline=False, analytical_mtrs=True, age_specific=False, reform={}, user_params={}, guid='', run_micro=True): from ogusa import parameters, wealth, labor, demographics, income from ogusa import txfunc tick = time.time() #Create output directory structure saved_moments_dir = os.path.join(output_base, "Saved_moments") ssinit_dir = os.path.join(output_base, "SSinit") tpiinit_dir = os.path.join(output_base, "TPIinit") dirs = [saved_moments_dir, ssinit_dir, tpiinit_dir] for _dir in dirs: try: print "making dir: ", _dir os.makedirs(_dir) except OSError as oe: pass if run_micro: txfunc.get_tax_func_estimate(baseline=baseline, analytical_mtrs=analytical_mtrs, age_specific=age_specific, reform=reform, guid=guid) print ("in runner, baseline is ", baseline) run_params = ogusa.parameters.get_parameters(baseline=baseline, guid=guid) run_params['analytical_mtrs'] = analytical_mtrs # Modify ogusa parameters based on user input if 'frisch' in user_params: print "updating fricsh and associated" b_ellipse, upsilon = ogusa.elliptical_u_est.estimation(user_params['frisch'], run_params['ltilde']) run_params['b_ellipse'] = b_ellipse run_params['upsilon'] = upsilon run_params.update(user_params) # Modify ogusa parameters based on user input if 'g_y_annual' in user_params: print "updating g_y_annual and associated" g_y = (1 + user_params['g_y_annual'])**(float(ending_age - starting_age) / S) - 1 run_params['g_y'] = g_y run_params.update(user_params) globals().update(run_params) from ogusa import SS, TPI # Generate Wealth data moments wealth.get_wealth_data(lambdas, J, flag_graphs, output_dir=input_dir) # Generate labor data moments labor.labor_data_moments(flag_graphs, output_dir=input_dir) get_baseline = True calibrate_model = False # List of parameter names that will not be changing (unless we decide to # change them for a tax experiment) param_names = ['S', 'J', 'T', 'BW', 'lambdas', 'starting_age', 'ending_age', 'beta', 'sigma', 'alpha', 'nu', 'Z', 'delta', 'E', 'ltilde', 'g_y', 'maxiter', 'mindist_SS', 'mindist_TPI', 'analytical_mtrs', 'b_ellipse', 'k_ellipse', 'upsilon', 'chi_b_guess', 'chi_n_guess','etr_params','mtrx_params', 'mtry_params','tau_payroll', 'tau_bq', 'calibrate_model', 'retire', 'mean_income_data', 'g_n_vector', 'h_wealth', 'p_wealth', 'm_wealth', 'get_baseline', 'omega', 'g_n_ss', 'omega_SS', 'surv_rate', 'e', 'rho'] ''' ------------------------------------------------------------------------ Run SS with minimization to fit chi_b and chi_n ------------------------------------------------------------------------ ''' # This is the simulation before getting the replacement rate values sim_params = {} glbs = globals() lcls = locals() for key in param_names: if key in glbs: sim_params[key] = glbs[key] else: sim_params[key] = lcls[key] sim_params['output_dir'] = input_dir sim_params['run_params'] = run_params income_tax_params, wealth_tax_params, ellipse_params, ss_parameters, iterative_params = SS.create_steady_state_parameters(**sim_params) ss_outputs = SS.run_steady_state(income_tax_params, ss_parameters, iterative_params, get_baseline, calibrate_model, output_dir=input_dir) ''' ------------------------------------------------------------------------ Run the baseline TPI simulation ------------------------------------------------------------------------ ''' ss_outputs['get_baseline'] = get_baseline sim_params['input_dir'] = input_dir income_tax_params, wealth_tax_params, ellipse_params, parameters, N_tilde, omega_stationary, K0, b_sinit, \ b_splus1init, L0, Y0, w0, r0, BQ0, T_H_0, tax0, c0, initial_b, initial_n = TPI.create_tpi_params(**sim_params) ss_outputs['income_tax_params'] = income_tax_params ss_outputs['wealth_tax_params'] = wealth_tax_params ss_outputs['ellipse_params'] = ellipse_params ss_outputs['parameters'] = parameters ss_outputs['N_tilde'] = N_tilde ss_outputs['omega_stationary'] = omega_stationary ss_outputs['K0'] = K0 ss_outputs['b_sinit'] = b_sinit ss_outputs['b_splus1init'] = b_splus1init ss_outputs['L0'] = L0 ss_outputs['Y0'] = Y0 ss_outputs['r0'] = r0 ss_outputs['BQ0'] = BQ0 ss_outputs['T_H_0'] = T_H_0 ss_outputs['tax0'] = tax0 ss_outputs['c0'] = c0 ss_outputs['initial_b'] = initial_b ss_outputs['initial_n'] = initial_n ss_outputs['tau_bq'] = tau_bq ss_outputs['g_n_vector'] = g_n_vector ss_outputs['output_dir'] = input_dir with open("ss_outputs.pkl", 'wb') as fp: pickle.dump(ss_outputs, fp) w_path, r_path, T_H_path, BQ_path, Y_path = TPI.run_time_path_iteration(**ss_outputs) print "getting to here...." TPI.TP_solutions(w_path, r_path, T_H_path, BQ_path, **ss_outputs) print "took {0} seconds to get that part done.".format(time.time() - tick)
def runner_SS(output_base, baseline_dir, baseline=False, analytical_mtrs=True, age_specific=False, reform={}, user_params={}, guid='', run_micro=True): from ogusa import parameters, wealth, labor, demographics, income from ogusa import txfunc tick = time.time() #Create output directory structure saved_moments_dir = os.path.join(output_base, "Saved_moments") ssinit_dir = os.path.join(output_base, "SSinit") tpiinit_dir = os.path.join(output_base, "TPIinit") dirs = [saved_moments_dir, ssinit_dir, tpiinit_dir] for _dir in dirs: try: print "making dir: ", _dir os.makedirs(_dir) except OSError as oe: pass if run_micro: txfunc.get_tax_func_estimate(baseline=baseline, analytical_mtrs=analytical_mtrs, age_specific=age_specific, start_year=user_params['start_year'], reform=reform, guid=guid) print ("in runner, baseline is ", baseline) run_params = ogusa.parameters.get_parameters(baseline=baseline, guid=guid) run_params['analytical_mtrs'] = analytical_mtrs # Modify ogusa parameters based on user input if 'frisch' in user_params: print "updating fricsh and associated" b_ellipse, upsilon = ogusa.elliptical_u_est.estimation(user_params['frisch'], run_params['ltilde']) run_params['b_ellipse'] = b_ellipse run_params['upsilon'] = upsilon run_params.update(user_params) # Modify ogusa parameters based on user input if 'g_y_annual' in user_params: print "updating g_y_annual and associated" g_y = (1 + user_params['g_y_annual'])**(float(ending_age - starting_age) / S) - 1 run_params['g_y'] = g_y run_params.update(user_params) globals().update(run_params) from ogusa import SS, TPI # Generate Wealth data moments wealth.get_wealth_data(lambdas, J, flag_graphs, output_dir=output_base) # Generate labor data moments labor.labor_data_moments(flag_graphs, output_dir=output_base) get_baseline = True calibrate_model = True # List of parameter names that will not be changing (unless we decide to # change them for a tax experiment) param_names = ['S', 'J', 'T', 'BW', 'lambdas', 'starting_age', 'ending_age', 'beta', 'sigma', 'alpha', 'nu', 'Z', 'delta', 'E', 'ltilde', 'g_y', 'maxiter', 'mindist_SS', 'mindist_TPI', 'analytical_mtrs', 'b_ellipse', 'k_ellipse', 'upsilon', 'chi_b_guess', 'chi_n_guess','etr_params','mtrx_params', 'mtry_params','tau_payroll', 'tau_bq', 'calibrate_model', 'retire', 'mean_income_data', 'g_n_vector', 'h_wealth', 'p_wealth', 'm_wealth', 'get_baseline', 'omega', 'g_n_ss', 'omega_SS', 'surv_rate', 'e', 'rho'] ''' ------------------------------------------------------------------------ Run SS ------------------------------------------------------------------------ ''' sim_params = {} glbs = globals() lcls = locals() for key in param_names: if key in glbs: sim_params[key] = glbs[key] else: sim_params[key] = lcls[key] sim_params['output_dir'] = output_base sim_params['run_params'] = run_params income_tax_params, wealth_tax_params, ellipse_params, ss_parameters, iterative_params = SS.create_steady_state_parameters(**sim_params) ss_outputs = SS.run_steady_state(income_tax_params, ss_parameters, iterative_params, baseline, calibrate_model, output_dir=output_base, baseline_dir=baseline_dir)
def runner(output_base, baseline_dir, baseline=False, analytical_mtrs=True, age_specific=False, reform={}, user_params={}, guid='', run_micro=True): #from ogusa import parameters, wealth, labor, demographics, income from ogusa import parameters, wealth, labor, demog, income, utils from ogusa import txfunc tick = time.time() #Create output directory structure saved_moments_dir = os.path.join(output_base, "Saved_moments") ssinit_dir = os.path.join(output_base, "SSinit") tpiinit_dir = os.path.join(output_base, "TPIinit") dirs = [saved_moments_dir, ssinit_dir, tpiinit_dir] for _dir in dirs: try: print "making dir: ", _dir os.makedirs(_dir) except OSError as oe: pass if run_micro: txfunc.get_tax_func_estimate(baseline=baseline, analytical_mtrs=analytical_mtrs, age_specific=age_specific, start_year=user_params['start_year'], reform=reform, guid=guid) print("in runner, baseline is ", baseline) run_params = ogusa.parameters.get_parameters(baseline=baseline, guid=guid) run_params['analytical_mtrs'] = analytical_mtrs # Modify ogusa parameters based on user input if 'frisch' in user_params: print "updating fricsh and associated" b_ellipse, upsilon = ogusa.elliptical_u_est.estimation( user_params['frisch'], run_params['ltilde']) run_params['b_ellipse'] = b_ellipse run_params['upsilon'] = upsilon run_params.update(user_params) # Modify ogusa parameters based on user input if 'g_y_annual' in user_params: print "updating g_y_annual and associated" g_y = (1 + user_params['g_y_annual'])**( float(ending_age - starting_age) / S) - 1 run_params['g_y'] = g_y run_params.update(user_params) from ogusa import SS, TPI # Generate Wealth data moments wealth.get_wealth_data(run_params['lambdas'], run_params['J'], run_params['flag_graphs'], output_dir=output_base) # Generate labor data moments labor.labor_data_moments(run_params['flag_graphs'], output_dir=output_base) calibrate_model = False # List of parameter names that will not be changing (unless we decide to # change them for a tax experiment) param_names = [ 'S', 'J', 'T', 'BW', 'lambdas', 'starting_age', 'ending_age', 'beta', 'sigma', 'alpha', 'nu', 'Z', 'delta', 'E', 'ltilde', 'g_y', 'maxiter', 'mindist_SS', 'mindist_TPI', 'analytical_mtrs', 'b_ellipse', 'k_ellipse', 'upsilon', 'chi_b_guess', 'chi_n_guess', 'etr_params', 'mtrx_params', 'mtry_params', 'tau_payroll', 'tau_bq', 'retire', 'mean_income_data', 'g_n_vector', 'h_wealth', 'p_wealth', 'm_wealth', 'omega', 'g_n_ss', 'omega_SS', 'surv_rate', 'e', 'rho' ] ''' ------------------------------------------------------------------------ Run SS ------------------------------------------------------------------------ ''' sim_params = {} for key in param_names: sim_params[key] = run_params[key] sim_params['output_dir'] = output_base sim_params['run_params'] = run_params income_tax_params, ss_parameters, iterative_params, chi_params = SS.create_steady_state_parameters( **sim_params) ss_outputs = SS.run_SS(income_tax_params, ss_parameters, iterative_params, chi_params, baseline, baseline_dir=baseline_dir) ''' ------------------------------------------------------------------------ Pickle SS results ------------------------------------------------------------------------ ''' if baseline: utils.mkdirs(os.path.join(baseline_dir, "SS")) ss_dir = os.path.join(baseline_dir, "SS/ss_vars.pkl") pickle.dump(ss_outputs, open(ss_dir, "wb")) else: utils.mkdirs(os.path.join(output_dir, "SS")) ss_dir = os.path.join(output_dir, "SS/ss_vars.pkl") pickle.dump(ss_outputs, open(ss_dir, "wb")) ''' ------------------------------------------------------------------------ Run the baseline TPI simulation ------------------------------------------------------------------------ ''' sim_params['input_dir'] = output_base sim_params['baseline_dir'] = baseline_dir income_tax_params, tpi_params, iterative_params, initial_values, SS_values = TPI.create_tpi_params( **sim_params) # ss_outputs['income_tax_params'] = income_tax_params # ss_outputs['wealth_tax_params'] = wealth_tax_params # ss_outputs['ellipse_params'] = ellipse_params # ss_outputs['parameters'] = parameters # ss_outputs['N_tilde'] = N_tilde # ss_outputs['omega_stationary'] = omega_stationary # ss_outputs['K0'] = K0 # ss_outputs['b_sinit'] = b_sinit # ss_outputs['b_splus1init'] = b_splus1init # ss_outputs['L0'] = L0 # ss_outputs['Y0'] = Y0 # ss_outputs['r0'] = r0 # ss_outputs['BQ0'] = BQ0 # ss_outputs['T_H_0'] = T_H_0 # ss_outputs['factor_ss'] = factor # ss_outputs['tax0'] = tax0 # ss_outputs['c0'] = c0 # ss_outputs['initial_b'] = initial_b # ss_outputs['initial_n'] = initial_n # ss_outputs['tau_bq'] = tau_bq # ss_outputs['g_n_vector'] = g_n_vector # ss_outputs['output_dir'] = output_base # with open("ss_outputs.pkl", 'wb') as fp: # pickle.dump(ss_outputs, fp) w_path, r_path, T_H_path, BQ_path, Y_path = TPI.run_TPI( income_tax_params, tpi_params, iterative_params, initial_values, SS_values, output_dir=output_base) print "getting to here...." TPI.TP_solutions(w_path, r_path, T_H_path, BQ_path, **ss_outputs) print "took {0} seconds to get that part done.".format(time.time() - tick)