Skip to content

Commit

Permalink
update msm and min
Browse files Browse the repository at this point in the history
  • Loading branch information
@alanlujan91
alanlujan91 committed Mar 13, 2024
1 parent b9c64a3 commit 4411900
Show file tree
Hide file tree
Showing 28 changed files with 411 additions and 319 deletions.
4 changes: 2 additions & 2 deletions code/do_all.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -45,7 +45,7 @@
still run.
"""

from estimark.estimation import estimate
from estimark.min import estimate_min
from estimark.options import (
all_replications,
high_resource,
Expand Down Expand Up @@ -143,7 +143,7 @@ def run_replication():
replication_specs["subjective_labor"] = True
print("Adding subjective labor market beliefs...")

estimate(**replication_specs)
estimate_min(**replication_specs)


if __name__ == "__main__":
Expand Down
179 changes: 44 additions & 135 deletions code/estimark/estimation.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -34,7 +34,6 @@
# Parameters for the consumer type and the estimation
from estimark.parameters import (
age_mapping,
bootstrap_options,
init_calibration,
init_params_options,
init_subjective_labor,
Expand All @@ -47,13 +46,6 @@
from estimark.scf import scf_data
from estimark.snp import snp_data

# Pathnames to the other files:
# Relative directory for primitive parameter files
tables_dir = "content/tables/"
Path(tables_dir).mkdir(parents=True, exist_ok=True)
# Relative directory for primitive parameter files
figures_dir = "content/figures/"
Path(figures_dir).mkdir(parents=True, exist_ok=True)

# =====================================================
# Define objects and functions used for the estimation
Expand Down Expand Up @@ -149,6 +141,40 @@ def get_weighted_moments(
return emp_moments


def get_moments_cov(agent_name, emp_moments):
moments_cov = em.get_moments_cov(
scf_data,
get_weighted_moments,
moment_kwargs={
"variable": "wealth_income_ratio",
"weights": "weight",
"groups": "age_group",
"mapping": age_mapping,
},
)

if "Port" in agent_name:

for key1 in emp_moments:
# Check if key1 exists in moments_cov dictionary
if key1 not in moments_cov:
# If it doesn't exist, create a new dictionary for this key
moments_cov[key1] = {}

for key2 in emp_moments:
# Check if key2 exists in the nested dictionary under key1
if key2 not in moments_cov[key1]:
# If it doesn't exist, we need to add it
if key1 == key2:
# If key1 is equal to key2, set the value to 1.0
moments_cov[key1][key2] = 1.0
else:
# Otherwise, set the value to 0.0
moments_cov[key1][key2] = 0.0

return moments_cov


def get_empirical_moments(agent_name):
emp_moments = get_weighted_moments(
data=scf_data,
Expand Down Expand Up @@ -420,10 +446,7 @@ def calculate_se_bootstrap(


def do_estimate_model(
agent,
emp_moments,
initial_guess,
minimize_options=None,
agent, emp_moments, initial_guess, minimize_options=None, tables_dir=None
):
fmt_init_guess = [f"{key} = {value:.3f}" for key, value in initial_guess.items()]
multistart_text = " with multistart" if minimize_options.get("multistart") else ""
Expand Down Expand Up @@ -481,7 +504,12 @@ def do_estimate_model(


def save_results(
res, agent_name, time_to_estimate, dir, params_key=None, keys_to_save=None
res,
agent_name,
time_to_estimate,
dir,
params_key=None,
keys_to_save=None,
):
model_estimate = getattr(res, params_key)

Expand Down Expand Up @@ -521,6 +549,7 @@ def do_compute_se_boostrap(
time_to_estimate,
bootstrap_size=50,
seed=0,
tables_dir=None,
):
# Estimate the model:
print("~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~")
Expand Down Expand Up @@ -568,7 +597,7 @@ def do_compute_se_boostrap(
)


def do_compute_sensitivity(agent, model_estimate, emp_moments):
def do_compute_sensitivity(agent, model_estimate, emp_moments, figures_dir=None):
print("``````````````````````````````````````````````````````````````````````")
print("Computing sensitivity measure.")
print("``````````````````````````````````````````````````````````````````````")
Expand Down Expand Up @@ -614,7 +643,7 @@ def do_compute_sensitivity(agent, model_estimate, emp_moments):
plt.show()


def do_make_contour_plot(agent, model_estimate, emp_moments):
def do_make_contour_plot(agent, model_estimate, emp_moments, figures_dir=None):
print("``````````````````````````````````````````````````````````````````````")
print("Creating the contour plot.")
print("``````````````````````````````````````````````````````````````````````")
Expand Down Expand Up @@ -655,123 +684,3 @@ def do_make_contour_plot(agent, model_estimate, emp_moments):
plt.savefig(figures_dir + agent.name + "SMMcontour.png")
plt.savefig(figures_dir + agent.name + "SMMcontour.svg")
plt.show()


def estimate(
init_agent_name,
params_to_estimate,
estimate_model=True,
compute_se_bootstrap=False,
compute_sensitivity=False,
make_contour_plot=False,
subjective_stock=False,
subjective_labor=False,
):
"""Run the main estimation procedure for SolvingMicroDSOP.
Parameters
----------
estimate_model : bool
Whether to estimate the model using Nelder-Mead. When True, this is a low-time, low-memory operation.
compute_standard_errors : bool
Whether to compute standard errors on the estiamtion of the model.
make_contour_plot : bool
Whether to make the contour plot associate with the estiamte.
Returns
-------
None
"""
############################################################
# Make agent
############################################################

agent = make_agent(
init_agent_name=init_agent_name,
subjective_stock=subjective_stock,
subjective_labor=subjective_labor,
)

############################################################
# Get empirical moments
############################################################

emp_moments = get_empirical_moments(agent.name)

############################################################
# Get initial guess
############################################################

initial_guess = get_initial_guess(
agent.name,
params_to_estimate,
tables_dir,
)

############################################################
# Estimate model
############################################################

if estimate_model:
model_estimate, time_to_estimate = do_estimate_model(
agent,
emp_moments,
initial_guess,
minimize_options=minimize_options,
)

# Compute standard errors by bootstrap
if compute_se_bootstrap:
do_compute_se_boostrap(
agent,
model_estimate,
time_to_estimate,
**bootstrap_options,
)

# Compute sensitivity measure
if compute_sensitivity:
do_compute_sensitivity(
agent,
model_estimate,
initial_guess,
)

# Make a contour plot of the objective function
if make_contour_plot:
do_make_contour_plot(
agent,
model_estimate,
emp_moments,
)


if __name__ == "__main__":
# Set booleans to determine which tasks should be done
# Which agent type to estimate ("IndShock" or "Portfolio")
local_agent_name = "IndShock"
local_params_to_estimate = ["CRRA", "DiscFac"]
local_estimate_model = True # Whether to estimate the model
# Whether to get standard errors via bootstrap
local_compute_se_bootstrap = False
# Whether to compute a measure of estimates' sensitivity to moments
local_compute_sensitivity = False
# Whether to make a contour map of the objective function
local_make_contour_plot = False
# Whether to use subjective beliefs
local_subjective_stock = False
local_subjective_labor = False

estimate(
init_agent_name=local_agent_name,
params_to_estimate=local_params_to_estimate,
estimate_model=local_estimate_model,
compute_se_bootstrap=local_compute_se_bootstrap,
compute_sensitivity=local_compute_sensitivity,
make_contour_plot=local_make_contour_plot,
subjective_stock=local_subjective_stock,
subjective_labor=local_subjective_labor,
)
Loading

0 comments on commit 4411900

Please sign in to comment.