CHANGELOG.md

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

0.22.0 - 2026-10-27

Added

• Adds two optional parameters for the initial distribution of household wealth on the transition path, initial_hhwealth_dist (an S x J matrix of shares of aggregate wealth, summing to one, that sets the shape of the initial distribution) and initial_wealth_ratio (the target ratio of initial aggregate household wealth to initial GDP, which sets its level). Both default to null, in which case TPI starts from the steady-state distribution of savings as before: get_initial_SS_values uses initial_hhwealth_dist for the shape only when it is not None, and run_TPI updates B[0] toward initial_wealth_ratio * Y[0] in each outer-loop iteration (rescaling the initial savings and refreshing initial_values for the inner loop) only when the ratio is not None. A reform run with either parameter set starts from the first-period wealth of the baseline transition path. Specifications.compute_default_params normalizes a null or cleared value of either parameter to None and checks the shape and sum of a supplied distribution. Documented in the calibration chapter (SecCalibOther_InitW) and the parameters list, with a test of run_TPI with both parameters set.
• Adds TPI._retry_root_if_failed, which re-solves a household problem with the Levenberg-Marquardt method whenever the root finder selected by FOC_root_method stops short of a root, and keeps the result with the smaller maximum residual. The hybr method can return "not making good progress" after a handful of evaluations when started far from the solution, in which case the returned point is just the initial guess. The fallback is applied to the oldest household's problem (firstdoughnutring), the partial-lifetime problems of households alive at t=0, and the full-lifetime problems (twist_doughnut).
• Adds TPI._euler_error_message so that a failed Euler-error check at the end of run_TPI reports the largest savings and labor FOC residuals and the (period, age, income group) at which each occurs, flagging households that were alive when the time path began.
• utils.get_initial_path now accepts scalars and 1 x M vectors as well as S x J arrays, returning paths of size T+S, (T+S) x M, or (T+S) x S x J. Optional S and J arguments declare the S x J layout for the ambiguous case S=1, J=M.

Updated

• Renames the initial-period wealth objects in TPI.py to make the age timing explicit. In get_initial_SS_values, the baseline steady-state savings rescaled to the period-0 population are now initial_b_sp1 (savings chosen at age s, held at s+1), and initial_b is the corresponding S x J array of wealth held at age s, with zeros in the first row. b_sinit and b_splus1init are copies of these two arrays, as before. This is a naming change only; the values of B0, b_sinit, and b_splus1init are unchanged.
• Removes the redundant initial_b element (identical to b_splus1init) from the initial_values tuple returned by TPI.get_initial_SS_values and unpacked in TPI.inner_loop and TPI.run_TPI. The tuple is now (B0, b_sinit, b_splus1init, factor, initial_n). All downstream uses in inner_loop (the arguments to firstdoughnutring and twist_doughnut) and in run_TPI (the initial guess for the savings path, BQ0, bmat_s in period 0, and b_mat_shift) now use b_splus1init.
• Changes the initial guesses in TPI.run_TPI. The household savings and labor supply paths now use the "quadratic" shape in utils.get_initial_path rather than "ratio". Aggregate household wealth B follows a quadratic path from its initial value to the steady state, the capital stocks K, K_d, and K_m scale with that path, and aggregate output Y is built up from industry output. Starting a reform instead from the period-0 capital stocks of the baseline time path (as in PR #1250) sent the outer loop onto a divergent track in examples/multi_industry_example.py, so baseline and reform runs use the same capital guesses; the loop recomputes these paths from household wealth and the fiscal rule after the first iteration, so they only affect the starting point. These changes affect only the starting point of the time path iteration, not the equilibrium it converges to (see the fix below for why the previous expected results differed).
• The Euler errors returned by TPI.inner_loop are now stored in a (T+S) x 2S period-by-age layout aligned with b_mat and n_mat: entry [t, s] is the savings FOC residual and entry [t, S+s] the labor FOC residual of the age-s household in period t. The euler_savings and euler_labor_leisure outputs of run_TPI are T x S x J in the same layout, so they now cover every household solved along the path, including those alive at t=0.
• Regenerates the expected results for test_run_TPI_full_run, test_run_TPI, test_run_TPI_extra, test_get_initial_SS_values, and test_inner_loop in test_TPI.py, and rebuilds the tests/test_io_data/OUTPUT2 baseline fixture (used by the two-iteration reform tests and the serial test) from a converged baseline run of the tests' small S=40, J=2 model with the current code. The old fixture predated industry capital and lacked the K_m entry the reform start-up now reads. The tests' variable-name mapping now falls back to the name itself for current output variables, and the comparison loops in test_run_TPI_full_run and test_run_TPI handle scalar outputs such as factor.
• Updates examples/multi_industry_example.py so that it runs with the current parameters: adds c_min for the three consumption goods, moves the start year to 2026, and sets the initial steady-state interest rate guess to 0.115. All of its output now goes under examples/MultiIndustryExample/, with the plots in example_plots/ and the summary table in example_output.csv, and the three industry plots are saved to separate files (industry_output_path.png, industry_labor_path.png, industry_capital_path.png) rather than overwriting one another.
• Renames the output of examples/run_ogcore_example.py from OG-Core_example_plots/ and OG-Core_example_output.csv to example_plots/ and example_output.csv, and updates test_run_example.py to read the renamed table.
• Removes the extraneous requirements.txt file. It was out of date (it listed os, which is part of the standard library, and taxcalc, which is not a dependency of OG-Core, and omitted the dependencies in pyproject.toml) and nothing in the build, tests, or workflows reads it.
• Bump certifi from 2026.7.22 to 2026.10.19 in PR #1259.

Fixed

• Fixes the application of replacement_rate_adjust on the household path along the time path. TPI.inner_loop pre-scaled theta by the adjustment before passing it to firstdoughnutring and twist_doughnut, and pensions.SS_amount then scaled it again, so for a uniform benefit cut the household first order conditions saw the factor squared while the aggregates and the saved pension_benefits applied it once. The cohort-path branch of SS_amount also indexed the pre-scaled vector by j and applied the factor for the cohort's entry period rather than the year of payment, so under a phased-in cut a cohort entering before the cut planned on full benefits that were never paid. inner_loop now passes the unadjusted theta[j], and SS_amount and pensions.replacement_rate_adjustment index the cohort path by the year each benefit is paid, matching the (T, S, J) pension array. Adds tests of the household-path multiplier under a phased-in cut and of consistency between the household path and the (T, S, J) pension array. (Reported in connection with comparing OG-USA Social Security reforms to Microcosm Dynamics.) See PR #1261.
• Corrects the note on replacement_rate_adjust in default_parameters.json and the parameter docs: the steady state uses the final value of the parameter rather than assuming it is one, so a permanent adjustment carries into the steady state that the time path converges to. See PR #1261.
• Fixes a silent failure of the household solution for cohorts alive when the time path begins. In TPI.inner_loop, the residuals of the partial-lifetime problems (the upper triangle solved with twist_doughnut and the oldest household solved with firstdoughnutring) were never written to euler_errors and so were never checked by the Euler-error test at the end of run_TPI, which covered only cohorts born during the time path. With the previous "ratio" initial guess, which drops savings by half the gap to the steady state in the first period, hybr failed immediately (status 5) on many of these problems -- 76 of the 546 in the default baseline, for cohorts initially aged 63 to 78 in every income group -- and the initial guess was carried through as if it were a solution, with FOC residuals of order 0.3 to 2. Because the outer loop's damped update kept feeding the failed guess back in, these unconverged household paths survived to convergence and were embedded in the previously expected test results, moving early-period aggregates by about 0.1 percent. All household residuals are now stored and checked, failed problems are re-solved with Levenberg-Marquardt, and the regenerated expected results satisfy every household first order condition to machine precision.
• Fixes docstring formatting in output_tables.py, aggregates.py, and utils.py that caused warnings in the documentation build, and replaces the equation in the aggregates.get_B docstring with the stationarized aggregate household wealth expression.
• Fixes the stationarized aggregate household wealth equation {eq}EqStnrz_Bt in stationarization.md, which was missing the hat on the first savings term.

0.21.2 - 2026-10-14

Fixed

The cache of auto-detected household Jacobian sparsity patterns in TPI (used when use_sparse_FOC_jac is True) was keyed only on the residual function name and problem size, so a pattern detected for one parameterization was reused for every later run_TPI call in the same process, including on long-lived dask workers. A pattern detected under US-Style Social Security (banded) was then applied to a Defined Benefits run, whose retirement-age conditions depend on working-age labor, and the household solves converged to slightly wrong values. The cache is now scoped to a single run_TPI call via p._jac_cache_scope and emptied when a new run begins.

• Adds tests/test_io_data/run_TPI_outputs_DB.pkl, which was not committed with the additional test in test_run_TPI_extra in test_TPI.py committed previously.

0.21.1 - 2026-10-12

Fixed

• pensions.DB_amount raised a TypeError on the transition path with the Defined Benefits pension system. TPI.firstdoughnutring solves the household in its final period of life at t = 0 and passes t = None with method 'TPI_scalar', and the short-horizon branch of DB_amount used t to index the replacement rate path. DB_amount now treats t = None as period 0 in that branch, consistent with the other 'TPI_scalar' branches.
• The conversion of avg_earn_num_years and yr_contrib from calendar years to model periods subtracted one period, so the Defined Benefits system averaged earnings over one fewer year and credited one fewer year of contributions than the parameters state, and avg_earn_num_years=1 gave a zero-length window: a 0/0 NaN replacement rate under US-Style Social Security and a ZeroDivisionError under Defined Benefits. A new pensions.pension_periods helper rounds S/80 * years with no shift and is used by DB_amount and deriv_DB, so the DB code matches the documented formula. replacement_rate_vals keeps the one-period shift in the Social Security averaging window so that pinned steady-state and time-path results are unchanged, but floors the window at one period so a one-year window no longer produces NaN; removing that shift (the highest 35 years with the default avg_earn_num_years=35 rather than 34) requires regenerating the pinned model outputs in tests/test_io_data and is left for a separate change.
• The partial-lifetime branch of pensions.DB_amount (cohorts alive when the time path begins) used the steady-state earnings profile e over the whole lifetime. It now uses the profile in effect in the period each age is lived in, with ages before the path anchored to the period-0 profile (as the wage is) and ages beyond the path using the final profile. The full-time-path (T x S x J) branch and the NDC_amount and PS_amount partial-lifetime branches still use the steady-state profile (see Issue #1014).
• pensions.pension_periods raises a ValueError when the averaging window is shorter than one model period or longer than the working life. The avg_earn_num_years validator allows values up to retirement_age, but a window longer than retirement_age - starting_age indexed before the first period of life and silently averaged in earnings from the oldest ages.
• Adds test_pension_periods_conversion_and_bounds to test_pensions.py and updates the expected values of the replacement_rate_vals, DB_amount, deriv_DB, pension_amount, and deriv_theta tests to the unshifted period counts. Fixes a typo in the Incomplete case of test_DB that set avg_earn_num_years on the wrong parameters object.
• Adds test_DB_amount_first_doughnut_ring_t_none to test_pensions.py and a Baseline, Defined Benefits pension parameterization of test_run_TPI_extra in test_TPI.py (expected output in run_TPI_outputs_DB.pkl) so that the transition path is solved with the Defined Benefits system.

0.21.0 - 2026-10-07

Added

• Allows the defined benefit pension replacement rate alpha_db to vary across lifetime income groups and over time. The parameter is now a T x J array (number_dims changed from 0 to 2 in default_parameters.json), extrapolated like the other T x J parameters, so a single value ([[0.02]]), one value per income group, or a time path can be entered. The rate applied to a benefit is the one in effect in the period in which the benefit is paid, so a change in the rate affects current as well as future retirees. Reform JSON files or scripts that previously set alpha_db as a scalar (e.g., 0.02) must now use the nested list form (e.g., [[0.02]]).
• Adds tests in test_pensions.py of DB benefits that vary by group in the steady state, of an incomplete lifetime straddling a change in the rate along the time path, of the derivative of benefits with a time-varying rate, and of consistency between the cohort-by-cohort and full-time-path computations of DB benefits with a rate that varies over time and by group.

Updated

• pensions.DB_amount now takes t and method arguments (DB_amount(w, e, n, t, j, method, p)) so that it can select the replacement rate for the period in which each benefit is paid. pensions.DB_1dim_loop, DB_2dim_loop, and DB_3dim_loop take the replacement rate as an array indexed by age (or by period for the full time path) rather than a scalar.
• pensions.deriv_DB and pensions.deriv_theta now take t and j arguments. With a time-varying rate, the derivative of benefits with respect to labor supply at each working age is evaluated at the rate in effect when the household retires. deriv_theta is not yet used in the household first order conditions, so this has no effect on model solutions.
• Fixes the title and description of alpha_db in default_parameters.json, which referred to a defined contribution system.
• Updates the defined benefit pension equations in the theory documentation (government.md) to use $\alpha^{DB}_{j,t}$.

0.20.4 - 2026-10-05

Added

• Allows the linear payroll tax rate tau_payroll to vary across lifetime income groups as well as over time. The parameter is now a T x J array (number_dims changed from 1 to 2 in default_parameters.json), extrapolated like labor_income_tax_noncompliance_rate and income_tax_filer, so a single value ([[0.1]]), one value per income group ([[0.1, ..., 0.2]]), or a time path ([[0.1], [0.12]]) can be entered. Reform JSON files or scripts that previously set tau_payroll as a flat list (e.g., [0.124]) must now use the nested form (e.g., [[0.124]]).
• Adds a vary tau_payroll by J case to test_FOC_labor in test_household.py, a tau_payroll extrapolation test to test_parameters.py, and a hand-computed check of get_payroll_tax_revenue with group-specific rates in test_aggregates.py.
• Adds a by_J keyword argument (default False) to the demographic plotting functions in parameter_plots.py (plot_imm_rates, plot_mort_rates, plot_population, plot_fert_rates, plot_mort_rates_data, plot_omega_fixed, plot_imm_fixed, plot_population_path) now that the demographic objects are T x S x J. With by_J=False, population distributions are summed across income groups and rates are averaged across income groups using the population weights in omega, so one line is plotted for the overall population. With by_J=True, one line is plotted per income group and labeled with the j value. The array-based functions take an omega (or omega_list) argument to supply the weights. Each function checks the dimensions of its inputs and continues to accept objects without an income-group dimension (e.g., T x S or S arrays).

Updated

• tax.income_tax_liab and household.FOC_labor now select tau_payroll by j (or across all J when j is None) in the same way as the tax noncompliance rates and filer status.
• aggregates.get_payroll_tax_revenue now takes household labor supply n, effective labor e, and population weights in place of aggregate labor L, and computes payroll tax revenue as the population-weighted sum of tau_payroll[t, j] * w * e * n across households, which is required when the rate differs across income groups. When tau_payroll is zero everywhere, the frac_tax_payroll fallback is unchanged.
• Specifications.compute_default_params now collapses a previously extrapolated T x J parameter (labor_income_tax_noncompliance_rate, capital_income_tax_noncompliance_rate, replacement_rate_adjust, income_tax_filer, wealth_tax_filer, tau_payroll) back to a single column when J changes and the values are the same across groups, so that these parameters need not be re-specified whenever J is changed via update_specifications (the T x S x J demographic and ability parameters still must be).
• Fixes the population weighting in plot_mort_rates, which multiplied mortality rates by the unnormalized population shares rather than taking a weighted average across income groups.
• Fixes the year indexing in plot_imm_rates, plot_mort_rates_data, and plot_population_path, which indexed years as start_year - year rather than year - start_year.

0.20.3 - 2026-10-01

Updated

• Removes the ceiling on the numpy version in pyproject.toml to "numpy>=2.5". All tests passed locally, except for the first test test_run_ogcore_example() (the time test) in test_run_example.py and the example run script ran successfully under the new repo (see output below). This resolves Issue #1233.
• The pytest output for the failed test showed that the baseline and reform of the example script ran and finished in less than 300 seconds, which resulted in an error in the test (our model is running too fast for the test). As such, I updated the test_run_ogcore_example() script in test_run_example.py to pass either if the model was running after 300 seconds or if it had a successful exit code.
• Updates the six GitHub Action files (build_and_test.yml, check_catalog.yml, check_ruff.yml, deploy_docs.yml, docs_check.yml, publish_to_pypi.yml) to only run on pushes and pulls to the master branch of the PSLmodels/OG-Core repository.
• Fixes some warnings and errors in the documentation. This entailed changes in some of the /ogcore/*.py file docstrings, *.bib files, *.rst files, and *.md files.
• Updates the uv.lock

0.20.2 - 2026-09-30 13:00:00

Added

• SAU UN codes to demographics.py
• Report aggregate hours worked in SS and TPI output dictionaries
• Parameterize the initial guesses for b and n in the SS solution algorithm
• Allow for a local currency conversion in output_tables.npv_table

0.20.1 - 2026-09-17 15:00:00

Added

• Fixes Issue #1202: get_r_gov clipped the interest rate on government debt at zero, so a sovereign that genuinely pays a negative real rate could not be modelled. The bound is now a parameter, r_gov_floor, with a default of 0.0 that reproduces the previous behaviour exactly.
• npv_table in output_tables.py (Issue #1131): builds a table of the net present value of the reform-minus-baseline change in flow variables (e.g. Y) over a horizon, evaluated at a list of discount rates. Values are un-stationarized by default so the NPV is taken over the actual (trend-inclusive) level path.
• New resource constraint error message, noting the maximum absolute resource-constraint error, the period it occurs in, and the tolerance, plus a one-line note on how to read it
• Scatter parameters object once per SS solve instead of one per residual evaluation. Pure overhead removal on any Dask-backed SS solve; the larger the Specifications object and the more workers, the more it saves. No numerical change — same solves, same results.

Bug Fixes

• Fixes Issue #1200: replacement_rate_adjust was read only inside SS_amount, so it applied to the US-Style Social Security system and was silently ignored under Defined Benefits, Notional Defined Contribution, and Points System. The adjustment is now applied to those three systems in pension_amount, via a replacement_rate_adjustment helper that mirrors the indexing SS_amount already uses, including the per-cohort t + tt offset along the time path. SS_amount is unchanged, so US-Style results cannot move.

0.20.0 - 2026-08-13 12:00:00

Added

• Addresses Issue #1205: the UN Data Portal API token can now come from a un_token argument to demographics.get_un_data, from a UN_API_TOKEN environment variable, or from a single per-user file ($XDG_CONFIG_HOME/og/un_api_token.txt, or %APPDATA%\og\un_api_token.txt on Windows). Sources are tried in that order, then un_api_token.txt in the working directory. This gives one token per user instead of one per directory.
• An og-token command, OG-Core's first console script, to manage that token without hunting for the file: og-token set saves one, og-token show reports where it lives and which source wins, and og-token rm deletes it. The token is read without echoing and is never printed back.

Changed

• Answering the UN API token prompt now saves the token to the per-user file rather than to the current working directory, which used to leave a copy of the token in every directory a model was run from. An existing un_api_token.txt in the working directory is still read, with a notice that the location is deprecated.
• The token prompt now names both ways forward: where to generate a free token, and that pressing return uses the archived copy of the same data instead.
• The token is read with getpass rather than input, so it is no longer echoed into the terminal and its scrollback while being typed or pasted.
• An expired token is now named as the reason for falling back to the archived data, instead of failing the same way as a missing one. The portal issues JSON Web Tokens, so the expiry date is read locally with no extra request, and og-token show reports whether the stored token is still current. A token that is not a readable JSON Web Token is used as before, with no expiry reported.
• A run that falls back to the archived data now says once, not once per series, how to register a token: where to get one, the full path of the og-token command belonging to the running interpreter, and the token file to write. The full path matters because og-token is installed beside the interpreter and is normally not on the shell's PATH, so someone who obtains a token days later has something they can paste rather than a command that reports "not found".
• The token prompt is skipped when standard input is not interactive, so scheduled and scripted runs fall back to the Population-Data archive instead of waiting on input.

0.19.2 - 2026-08-18 17:00:00

Adds

• Updates .gitignore to include a line .claude/.
• Updates the uv.lock file to include upgraded pillow package version 12.3.0. This satisfies the dependabot PR update in PR #1176. I updated all the packages in the environment. This resulted in an update of the pytest package, which caused errors that led to the changes below.
• Updated the version in pyproject.toml, ogcore/__init__.py and CHANGELOG.md.
• Updated the version of pytest in pyproject.toml to pytest>=9.1. Some changes to logging happened in versions 9.1+, that are not backward compatible. All our testing started throwing errors once I updated that package. I changed the pytest miniversion in line 94 of pyproject.toml to 9.1, and I updated the filterwarnings in lines 94-98.
• I updated all the Makefile commands to fit with standard GNU formatting and to work with our updated usage of the uv package. I also got rid of some compute studio commands, the objects of which had no targets.
• I updated the format of the logger.info() warnings in household.py which was a source of errors from the updated pytest package.
• I fixed some errors that were happening in the building of our Jupyter Book documentation. There was a fight going on between our make_params.py and make_vars.py scripts and the Sphinx _toc.yml table of contents script. And we were missing config.rst, SS.rst, and TPI.rst files. I explicitly put the API chapters in the _toc.yml file. And I built an updated parameters.md file.
• We had some Jupyter Book warnings for headings that go from H2 to H4 in parameters.md and variables.md. However, this is what we want. So I added the _ext/supress_header_warnings.py file and referenced it in _config.py in order to surpress only those warnings associated with only the parameters.md and variables.md files.

Bug Fixes

• Fixes Issue #1199: aggregates.revenue counted payroll tax revenue twice. Each household's income_payroll_tax_liab already includes T_P = tau_payroll * labor_income, so payroll revenue is inside iit_payroll_tax_revenue; PR #1184 then added get_payroll_tax_revenue (the same take, re-derived from the aggregate wage bill) into that total again when tau_payroll was nonzero. Revenue was overstated, the government over-spent, and any model with a nonzero tau_payroll failed the steady-state resource constraint (OG-USA was unaffected only because its tau_payroll is 0). The erroneous addition is removed; get_payroll_tax_revenue still provides the income vs payroll reporting split, which is unchanged. (See PR #1204 and PR #1199.)

0.19.1 - 2026-08-10 12:00:00

Added

• Stall detection for the TPI outer loop (Issue #1177): when the best distance has not improved over the last TPI_stall_window iterations (default 50; 0 disables), run_TPI logs a diagnosis distinguishing a cycling outer loop (suggesting a lower nu or TPI_outer_method="anderson") from a diverging economy (usually an inconsistent fiscal block, which solver settings cannot fix). The default TPI_stall_action="warn" only logs, leaving model solutions unchanged; "stop" also ends the loop early, so a hopeless run fails through the existing non-convergence checks instead of spending the rest of maxiter. The window check lives in ogcore.solvers.diagnose_stall and works for both the picard and anderson update rules. The diagnosis is re-logged if it changes (e.g. escalates from cycling to diverging), and a run that ends unconverged while stalled carries the diagnosis in the RuntimeError message, so it reaches users who only see the traceback. PR #1178.
• Changes the default of stationarized to False in output_plots.plot_aggregates and output_plots.plot_industry_aggregates, so unstationarized values are plotted by default, resolving Issue #1133. PR #1183.

Bug Fixes

• Fixes Issue #1186: with demographics that vary across lifetime income groups (introduced in PR #1165), the use_zeta = False branch of household.get_bq divided each group's bequest pool by its birth share (lambdas[j]) rather than its actual population share, so bequests received did not sum to bequests left and the steady-state resource constraint failed. Receipts are now divided by the group's actual population (from omega_SS / omega), matching the use_zeta = True branch. Results are unchanged when demographics are common across groups. PR #1187.
• Fixes Issue #1015: fixes escape with special characters in parameter names that caused failures when reading from JSON files. Adds a test to prevent this in the future. PR #1193.
• Resolve Issue #1152: expose same keys in the steady-state and time path dictionaries. PR #1185.
• Correct payroll tax calculations, resolving Issue #1023. PR #1184.

0.19.0 - 2026-07-29 12:00:00

Added

• Fixes Issue #1169: the Notional Defined Contribution pension system could not run against a real Specifications object because its growth-rate settings were not parameters. ndc_growth_rate, dir_growth_rate, and points_growth_rate are now parameters in default_parameters.json (choices "r", "Curr GDP", "LR GDP"; default "LR GDP", matching the previous fallback behavior). g_ndc and g_dir now handle the scalar r and g_y of the steady state, and delta_ret falls back to the steady-state mortality rates implied by rho (averaged over lifetime income groups) when no mort_rates_SS attribute is set. The NDC system is added to the real-Specifications pension test.

0.18.1 - 2026-07-22 12:00:00

Bug Fixes

• Fixes Issue #1180: get_pop_objs raised an AssertionError for the homogeneous case (income_percentiles=None with no income-specific inputs), because expand_pop_obj_J asserted before reaching its homogeneous branch. income_percentiles now defaults to a single income group (J=1) when no income-specific inputs are supplied, restoring the pre-0.18 call signature; it remains required whenever gradients or immigrant income shares are provided. Also removes a stray debug print from that branch.

0.18.0 - 2026-07-20 12:00:00

Added

• Demographic parameters (rho, omega, imm_rates) are now allowed to vary across lifetime income groups, J. See PR #1165

0.17.0 - 2026-07-16 12:00:00

Bug Fixes

• Fixes the Defined Benefits and Points System pension paths so they run against a real Specifications object (Issues #1014 and #1075): p.retire (an array since PR #433) was passed as the scalar S_ret into the numba loops, the scalar p.g_y was indexed as an array inside the loops, the scalar steady-state wage was indexed as a path, and the time-varying 3-D e matrix (PR #895) was sliced as 2-D. The systems use the steady-state retirement age and earnings profile for now; full time variation remains open in Issue #1014. The same coercions are applied to the NDC path, but the NDC system additionally requires growth-rate settings (ndc_growth_rate, dir_growth_rate) that are not yet parameters in default_parameters.json, so it still cannot run against a real Specifications object (see Issue #1169).
• Wires the pre-time-path inputs the DB/NDC/PS benefit formulas need into the TPI: labor supplied before the time path begins comes from the model's initial condition (the same baseline object that initializes wealth), and pre-time-path wages are anchored to the period-0 wage of the current path. Adds the full time-path (T x S x J) evaluation of Defined Benefits amounts (DB_3dim_loop) used when the TPI computes aggregate revenues, built from each cohort's own wage and labor history so aggregates are consistent with household behavior. With these changes a country model using the Defined Benefits system solves both the steady state and the transition path.
• Adds regression tests that call pension_amount with a real Specifications object per pension system (the existing tests pre-scalarized the inputs and so never exercised the real interface) and a local-marked steady-state solve test with the Defined Benefits system. See PR #1167.
• Fixes the tax-liability revenue calculation using the first year's tax noncompliance and filer rates for every period, even when those rates are set to vary over time (Issue #1168): households responded to the changing rates while government revenue, the budget, and debt stayed on the first-year values. income_tax_liab now slices the rates over the transition path. See PR #1174.
• Fixes the steady-state etr_ss and mtry_ss diagnostics using the labor-income tax noncompliance rate in place of the capital-income rate (capital_noncompliance_rate_2D in SS.py, Issue #1170). This affects only the post-solve SS diagnostics -- the solution itself already used the correct rates -- and is a no-op when the two rates are equal (the default). See PR #1171.
• Fixes alpha_FA (direct foreign aid, added in 0.16.0) not being extended over the model time path: it was never registered in tp_param_list, so a multi-year path stayed short and broke the transition solve. It now extends to T + S with the last value carried forward, like alpha_G/alpha_T/alpha_I. See PR #1166.

0.16.4 - 2026-07-02 12:00:00

Added

• Adds an opt-in accelerated update rule for the TPI outer loop, selected by the new TPI_outer_method parameter (default "picard", which leaves the model's historical damped functional iteration and all model solutions unchanged). Setting TPI_outer_method="anderson" uses limited-memory Anderson acceleration on the outer-loop residual history, guarded by a trust region anchored to the always-feasible damped point (controlled by TPI_anderson_m, TPI_anderson_beta, and TPI_trust_radius). On a stiff multi-industry reform that limit-cycles under constant dampening, this converged to the same equilibrium in 53 outer iterations vs 126 for constant nu=0.1 (about −37% wall-clock vs the best damped-nu schedule). The new solver lives in ogcore/solvers.py. See PR #1164.

0.16.3 - 2026-06-25 15:00:00

Added

• Better functionality and more country repositories with the OG installer. See PR #1162

Bug Fix

• Fixes an inconsistency with the pre-time path population distribution and growth rates. Note that the demographics.get_pop function has been changed and now only returns one object: the population distribution (not also the distribution prior to the start year). In addition, the get_pop and get_pop_objs functions no longer have the pre_pop_dist kwarg. See PR #1073.

0.16.2 - 2026-06-15 12:00:00

Added

• Adds output_tables.model_fit_table function that generates a table comparing model output to data for a set of target moments, and adds a test of this function in test_output_tables.py. The function takes as input a list of target moment descriptions, the TPI output dictionary, and the model parameters object, and returns a pandas DataFrame with the target moment descriptions, the model values for those moments, and the data values for those moments (where applicable). The function currently supports a set of macroeconomic moments (interest rate, capital share of output, labor share of output), inequality moments (Gini coefficient for before-tax income and after-tax income), and demographic moments (fraction of population 65+ and population growth rate). See PR #1138.
• Validates beta_annual and chi_b against J. (PR #1149)

Bug Fixes

• Fixes issue with reading UN data on Pandas >= 3.0 and UN token string. (PR #1151)
• Fixes math notation for plot labels. (PR #1148)
• Fixes reshaping issues with J=1 parameterization. (PR #1145)

0.16.1 - 2026-06-04 12:00:00

Added

• Adds a use_sparse_FOC_jac Specifications parameter (default True) that accelerates the time path iteration (TPI) household solve. With it on, scipy.optimize.root is given a sparse (banded) finite-difference Jacobian for the stacked household Euler and labor first order conditions: the sparsity pattern is auto-detected once per problem size and the solver then needs far fewer function evaluations per Jacobian build (about 20x fewer on the default S=80 cohort solve), with an automatic fallback to dense finite differences if the Jacobian is not sparse enough to benefit or if a solve fails. The result matches the legacy dense-finite-difference solution to within the model's resource-constraint accuracy floor on every calibration tested (OG-Core standard example, OG-ETH, OG-ZAF, OG-PHL, OG-IDN), giving roughly a 1.9-2.4x TPI speedup at no accuracy cost. Set use_sparse_FOC_jac=False to recover bit-identical agreement with v0.16.0 and earlier.

0.16.0 - 2026-06-02 12:00:00

Added

• Adds a new parameter, alpha_FA, that allows the user to specify the level of direct foreign aid as a percentage of GDP. See PR #1126.

0.15.13 - 2026-05-15 06:00:00

Added

• Increases the maximum value of the mindist_TPI parameter to 0.01 in default_parameters.json.
• Updates the uv.lock with package updates.

0.15.12 - 2026-05-14 12:00:00

Added

• Caches the e_long array in household.FOC_savings and household.FOC_labor rather than rebuilding it on every call. e_long is a pure function of p.e, p.S, and p.J, none of which change during a solve, so a _get_e_long helper now builds it once per worker and reuses it. Profiling identified this rebuild as the single most expensive operation in a TPI run. The change is a pure cache — model output is bit-for-bit identical to master — and gives roughly a 3x speedup on a single-reform TPI run. See PR #1128.
• Builds the per-period tax-parameter slices in TPI.inner_loop as numpy arrays via a new _params_to_array helper, and switches txfunc.get_tax_rates to np.asarray, so the repeated per-call list-to-array conversion is skipped on the hot TPI path. mono and mono2D tax functions store callables rather than numbers, so their nested-list form is passed through unchanged. Profiling identified this conversion as the next hot spot after the e_long rebuild. Model output is bit-for-bit identical to master, and the change gives roughly a further 10% speedup on a single-reform TPI run (about 3.3x cumulative versus master). See PR #1128.
• Changes the minimum of the allowable range for tau_c in default_parameters.py to allow for government consumption subsidies.
• Fixes a bug in the parameter_plots.plot_fert_rates function. See PR #1127.

0.15.11 - 2026-05-08 12:00:00

Added

• Updates the uv.lock to include dependabot upgrades PRs #1118, #1119, #1120, and #1121, as well as some other package upgrades that were available.

0.15.10 - 2026-04-23 12:00:00

Added

• Updates uv.lock to include upgrade to package `nbconvert from 7.17.0 to 7.17.1. This was the subject of PR 1110. But we could not get its tests to pass.
• Fixes typo in CHANGELOG.md

0.15.9 - 2026-04-23 03:00:00

Added

• Create a ogcore logging environment so that logging specificed in ogcore.config is not overridden by other logging configurations in user code. See PR #1109.
• Includes PR 1111 and PR 1112 that change build_and_test.yml to not upload CodeCov statistics for PRs generated by dependabots and updates the codecov-action to v5.

0.15.8 - 2026-04-16 00:30:00

Added

• Updates the uv.lock file to the current version of ogcore package (0.15.8)
• Adds '**.lock' to the paths in build_and_test.yml so that any changes to the uv.lock file trigger the GH Action CI tests in build_and_test.yml
• Incorporates the dependabot updates to uv.lock versions from PR 1105 and PR 1106.

0.15.7 - 2026-04-15 20:00:00

Added

• Fixed some typos in CHANGELOG.md and intro.md
• Added dependabot package version updates to uv.lock from PR 1102 and PR 1103

0.15.6 - 2026-04-15 16:00:00

Added

• Changes to package files to use uv for packaging and running commands. See PR #1093

0.15.5 - 2026-01-27 12:00:00

Added

• Additional parameter metadata to default_parameters.json. See PR #1097

0.15.4 - 2026-01-27 12:00:00

Added

• Ability to simulate the model with a single type of household (J=1). See PR #1062

0.15.3 - 2026-01-24 12:00:00

Added

• Two new parameters, income_tax_filer and wealth_tax_filer, that determine whether certain types j pay income or wealth taxes, respectively. See PR #1084

0.15.2 - 2026-01-22 12:00:00

Added

• Two new parameters, r_gov_DY and r_gov_DY2, that allow the government interest rate to be a function of the debt-to-GDP ratio. See PR #1037

0.15.1 - 2026-01-19 12:00:00

Added

• A new parameter c_min to the Parameters class that allows the user to specify minium consumption amounts by consumption good. See PR #1085

0.15.0 - 2025-12-03 12:00:00

Added

• A new parameter baseline_theta to the Parameters class that allows the user to specify whether to use the steady-state replacement rate parameters from the baseline solution in a reform run. See PR #1077

0.14.14 - 2025-11-24 12:00:00

Added

• Create SS.SS_initial_guesses function to allow more flexible initial guesses for steady state solution ((PR #1061)[#1061])
• Robust steady-state solution used for reform solution ((PR #1061)[#1061])
• Test of SS.solve_for_j function ((PR #1061)[#1061])

Bug Fix

• Fixes to deprecated API calls avoid many warnings during testing ((PR #1061)[#1061])

0.14.13 - 2025-11-21 12:00:00

Bug Fix

• Fix calculation of consumption tax revenue with differentiated goods ((PR #1074)[#1074])

0.14.12 - 2025-11-07 12:00:00

Bug Fix

• Use data for pre-time path population distribution (rather than inferring it) (PR #1071)

0.14.11 - 2025-11-07 12:00:00

Added

• Adds Ethiopia demographic data mapping ((PR #1063)[#1063])

0.14.10 - 2025-09-11 12:00:00

Added

• Fixes nonconformable matrices in TPI.py introduced in version 0.14.9 ((PR #1054)[#1054])

0.14.9 - 2025-09-10 20:00:00

Added

• Fixes replacement_rate_adjustment parameter in the steady state ((PR #1053)[#1053])
• Adds some saved output to tpi_vars.pkl object ((PR #1054)[#1054])

0.14.8 - 2025-08-26 12:00:00

Added

• Adds a complete benchmark suite for measuring and optimizing Dask performance in OG-Core, with particular focus on Windows performance issues.
• New and updated files:
  • tests/test_dask_benchmarks.py: Mock benchmark tests with synthetic workloads
  • tests/test_real_txfunc_benchmarks.py: Real-world tax function benchmarks
  • tests/run_benchmarks.py: Automated benchmark runner with reporting
  • tests/BENCHMARK_README.md: Comprehensive documentation and usage guide
  • pytest.ini: Updated with benchmark test markers
• Key features:
  • Platform-specific optimization tests (Windows, macOS, Linux)
  • Memory usage and compute time benchmarking
  • Baseline establishment and performance regression detection
  • Comparison of different Dask schedulers and client configurations
  • Real tax function estimation performance measurement
  • Automated identification of optimal Dask settings per platform
• Benefits:
  • Establishes performance baselines before optimization work
  • Identifies Windows-specific Dask performance bottlenecks
  • Provides automated regression detection for future changes
  • Enables data-driven optimization decisions
  • Supports continuous performance monitoring
• Usage:
  • python tests/run_benchmarks.py # Run all benchmarks
  • python tests/run_benchmarks.py --quick # Quick benchmarks only
  • python tests/run_benchmarks.py --save-baseline # Save performance baseline
  • python tests/run_benchmarks.py --compare-baseline # Compare against baseline
• 🤖 Generated with help from Claude Code

0.14.7 - 2025-08-21 17:00:00

Added

• Refactor calls to dask in SS.py and TPI.py. See PR #1048

0.14.6 - 2025-08-15 14:00:00

Added

• Removes initial_guess_w_SS in default_parameters.json
• Updates environment and testing to cover Python 3.13

0.14.5 - 2025-07-08 22:00:00

Added

• Increases the maximum value of initial_guess_TR_SS in default_parameters.json

0.14.4 - 2025-06-23 18:00:00

Added

• Fixes the sign error on the remittances RM term in aggregates.py, resource_constraint() function.
• Added a test with positive remittances to test_aggregates.py, test_resource_constraint() function.

0.14.3 - 2025-04-25 10:00:00

Added

• Puts a ceiling on the version of the marshmallow<4.0.0 package in environment.yml
• Update txfunc.py for the estimation of the HSV and GS tax functions.
• Update test_txfunc.py, tax_func_estimate_outputs.pkl, and tax_func_loop_outputs.pkl files for testing
• Update dask and distributed client calls in SS.py and TPI.py to allow for updated versions.

0.14.2 - 2025-04-04 12:00:00

Added

• utils.pct_change_unstationarized replaced with utils.unstationarize_vars, which allows for more general use of a utility to find unstationarized values of time series output
• output_plots.py and output_tables.py are updated to all for plots and tables of unstationarized output for variables of any type, not just percentage changes.
• ouput_tables.tp_output_dump_table has been renamed output_tables.time_series_table
• utils.param_dump_json has been renamed utils.params_to_json
• API docs have been updated to include functions left out previously and for new function names

0.14.1 - 2025-03-16 12:00:00

Bug Fix

• Packages model_variables.json with ogcore

0.14.0 - 2025-03-16 07:00:00

Added

• Updates the output dictionaries for TPI.py and SS.py to use consistent variables names
• Adds a model_variables.json file that has metadata about the model variables and is used to build a new chapter in the documentation describing the model variables
• Replaces print commands in TPI.py and SS.py with logger.info commands for easier suppression of output

0.13.2 - 2024-12-08 12:00:00

Added

• Adds KOR, THA, BRA to the list of countries in the demographics.py module.

0.13.1 - 2024-10-02 12:00:00

Added

• Three new parameters to adjust government spending amounts in the case of baseline_spending=True:
  • alpha_bs_G: the proportional adjustment to the level of baseline spending on government consumption (time varying, default value is 1.0 for each model period)
  • alpha_bs_T: the proportional adjustment to the level of baseline spending on non-pension transfers (time varying, default value is 1.0 for each model period)
  • alpha_bs_I: the proportional adjustment to the level of baseline spending on infrastructure investment (time varying, default value is 1.0 for each model period)

0.13.0 - 2024-09-26 12:00:00

Added

• Updates all of the documentation.
  • Adds remittances to all instances of the household budget constraint
  • Rewrites bequests and transfers components of household budget constraint in terms of individual variables in all instances
  • Adds a household transfers section to households.md with subsections on bequests, remittances, government transfers, and universal basic income
  • Changes all instances of $p_t Y_t$ to $Y_t
  • Updates the steady-state equilibrium algorithm description in equilibrium.md
  • Added updates to the government pensions descriptions in government.md and added pensions to all instances of the household budget constraint.
  • Updates the docstrings in tax.py for the wealth tax ETR and MTR functions. The code is right. I just thought there was a clearer specification of the equations in LaTeX.
• Adds remittances to the OG-Core code
  • Adds aggregate remittances function get_RM() to aggregates.py
  • Adds household remittances function get_rm() to household.py
  • Adds four new remittance parameters: alpha_RM_1, g_RM, alpha_RM_T, eta_RM
  • We model aggregate remittances as a percent of GDP in the first period, then growing at a specified rate that can deviate from the country growth rate until the cutoff rule period, after which the remittance growth rate trends back to the long-run model growth rate (growth rate in population and productivity). We also model remittances in reforms as being a percentage of baseline GDP. In this way, if remittance parameters are not changed in the reform, remittances remain at their baseline levels. The only way they change is if their parameter values are changed.
  • Adds 3 tests using the test_get_RM() function in test_aggregates.py
  • Adds 4 tests using the test_get_rm() function in test_household.py
  • Changes the initial_guess_r_SS in two tests in test_SS.py because they were not solving with their current values
• Increases RC_SS steady-state resource constraint tolerance from 1e-9 to 1e-8 because two test_run_SS() tests were failing in test_SS.py with resource constraints errors just bigger than 1e-9 (-2.29575914e-09 for [Baseline, small open] and -2.29575914e-09 for [Reform, small open]).
• Increases RC_TPI transition path resource constraint tolerance from 1e-5 to 1e-4 in because one test_run_TPI_full_run() test was failing in test_TPI.py with a resource constraint error just bigger than 1e-5 (1.4459913381864586e-05 for [Baseline, M=3 non-zero Kg]).
• Updated two directory path references that were out of date in test_run_example.py.
• Updated expected value tuples and dictionaries in test_txfunc.py.

0.12.0 - 2024-08-20 12:00:00

Added

• Support for Python 3.12

0.11.17 - 2024-08-18 12:00:00

Added

• Description of alpha_I in docs
• Updates valid range of the nominal UBI parameters

Bug Fix

• Extrapolate alpha_I in parameters.py
• Ensure alpha_I shape conforms in TPI.py
• Fix formatting of labels in constants.py

0.11.16 - 2024-08-10 12:00:00

Added

• Added parameter script make_params.py that generates a markdown file for the documentation parameters.md. Updates Makefile, and GitHub Actions (PR #963)
• Updated debt-to-GDP plot labels (PR #962)

0.11.15 - 2024-07-30 12:00:00

Bug Fix

• Make OGcorePlots.mplstyle importable from the package by adding it to setup.py

0.11.14 - 2024-07-30 12:00:00

Added

• Aesthetic updates to plotting functions in parameter_plots.py and demographics.py

0.11.13 - 2024-07-28 12:00:00

Added

• Added three new pension types to the model: (i) defined benefits system, (ii) notional defined contribution system, and (iii) points system.

[0.11.12] - 2024-07-26 01:00:00

Bug Fix

• Fixes extrapolation of nested lists of tax function parameters.

0.11.11 - 2024-06-24 01:00:00

Added

• Add new parameters for resource constraint tolerances for steady state and time path solution.

0.11.10 - 2024-06-17 01:00:00

Added

• Add HSV to list of valid tax functions in default_parameters.json

0.11.9 - 2024-06-12 01:00:00

Added

• Update demographics.py in the case input prompt not work.
• Add new utility to dump the parameters to a JSON file

0.11.8 - 2024-06-09 01:00:00

Added

• Updates to demographics.py module to accept token for UN World Population Prospects database access or to download data from the Population-Data repository.

0.11.7 - 2024-06-07 01:00:00

Added

• Heathcote, Storesletten, and Violante (2017) tax functions to txfunc.py

0.11.6 - 2024-04-19 01:00:00

Added

• Scatters parameters once in TPI.py
• Removes Python 3.9 tests from build_and_test.yml

0.11.5 - 2024-04-11 12:00:00

Added

• Adds a list of file change event triggers to build_and_test.yml so that those tests only run when one of those files is changed.
• Updates the codecov GH Action to version 4 and adds a secret token.
• Adds a list of file change event triggers to deploy_docs.yml and docs_check.yml, and limits docs_check.yml to only run on pull requests.

0.11.4 - 2024-04-03 22:00:00

Added

• Add a function to utils.py to shift lifetime profiles of parameters
• Add a function to utils.py to compute percentage changes in non-stationary variables
• Add more functionality to parameters_plots.py, allowing the user to plot parameters from multiple parameters objects together

0.11.3 - 2024-03-08 12:00:00

Added

• Allow for demographics.py to save downloaded data directly.
• Retrieve population data from the UN World Population Prospects database through CSV rather than JSON to avoid rate limit errors.

0.11.2 - 2024-02-17 12:00:00

Added

• Updates demographics.py with more functionality across the time path.
• Allow the user to have the population distribution from the initial period forward inferred from given fertility, mortality, and immigration rates (functionality to infer immigration from a given evolution of the population is retained, the user specifies what they want to do via arguments to the relevant function calls).
• Extends all series returned from the get_pop_objs() function over the full transition path of T+S periods (except those that apply only to a single period).
• Addresses Issues #900 and #899

0.11.1 - 2024-02-12 15:00:00

Added

• Updated setup.py Python version requirement to be python_requires=">=3.7.7, <3.12"

0.11.0 - 2024-02-06 15:00:00

Added

• Allow chi_n parameter to vary over the time path (PR #897)
• Create a demographics module in OG-Core (PR #896)
• Create a time varying ability matrix (PR # 895)
• Simplify the extrapolation of arrays over the time path (PR #891)
• Update the copyright year of documentation to 2024

0.10.10 - 2023-10-25 17:00:00

Added

• Remove surve_rate parameter (PR #886)
• Updates to plot_2D_taxfunc (PR #881)

0.10.9 - 2023-09-08 12:00:00

Added

• PR #880 standardize the time path output length
• PR #878 fix tax function indexing, dimensions, and plotting. This PR also enables Python 3.11.
• PR #875 remove unused dependency

0.10.8 - 2023-04-22 12:00:00

Added

• Adds a 2D monotonic smoothing spline tax function estimation to txfunc.py
• Changes the tax function parameters objects from NumPy arrays to lists in order to accomodate the nonparametric functions that get passed with the mono and mono2D options

0.10.7 - 2023-03-31 12:00:00

Added

• Uses lists to pass and access effective tax rate objects etr_params and marginal tax rate objects mtrx_params and mtry_params

0.10.6 - 2023-02-15 12:00:00

Added

• Uses 300 dpi when saving plots to disk
• Better labels of the plot_industry_aggregates plots

0.10.5 - 2023-02-14 12:00:00

Added

• Fix to SS.py to use baseline solution on reform run if dimensions match
• Fix to test_basic.py dimensions for r_gov_scale

0.10.4 - 2023-02-06 12:00:00

Added

• New calibration section to documentation (PR #850)
• Allow government risk premia to vary across time path for parameters r_gov_shift and r_gov_scale (PR #852)

0.10.3 - 2023-01-21 12:00:00

Added

• Bug fixes for new tax function parameter estimation

0.10.2 - 2023-01-12 12:00:00

Added

• Adds a new minimum value to the r_gov_shift parameter of -0.3

0.10.1 - 2023-01-05 12:00:00

Added

• Removes hard coded year label in parameter_plots.plot_population_path() (PR #825)
• Fixes documentation (PR # 827)
• Adds "mono" specification to default_parameter.json and test_parameters.py (PR #830)
• Restricts Python version to be < 3.11 and removes the mkl dependency in environment.yml and setup.py (PR #833 and #840)
• Updates CI testing to include Mac, Windows, and Linux operating systems and Python 3.9 and 3.10, and solves some CI test issues (PR #836)
• Increases the maximum values for r_gov_shift and r_gov_scale in default_parameters.json (PR #838)
• Removes the mkl dependency from environment.yml and setup.py (PR #840)

0.10.0 - 2022-09-27 12:00:00

Added

• Adds matrix of tax noncompliance parameters to households (PR #816)
• Incorporate input/output matrix mapping production goods to consumption goods (PR #818)
• Adds a new monotonic tax function estimation method to txfunc.py (PR #819)

0.9.2 - 2022-08-21 12:00:00

Added

• Updates the form of the investments tax credit to be on a proxy for investment (depreciated capital) in order to satisfy theoretical requirements of static firms in each industry.
• Update the documentation
• Update the requirement for the m_wealth parameter in the wealth tax function to be strictly greater than zero.

0.9.1 - 2022-07-22 12:00:00

Added

• Adds an investment tax credit parameter to the model
• Adds a boolean that allows the option to compute a reform using a stored baseline solution rather than recomputing the baseline

0.9.0 - 2022-06-30 12:00:00

Added

• Adds multiple production industries to the model

0.8.2 - 2022-06-01 12:00:00

Added

• Formatting of source code with black
• Last tag before extension of model to include multiple industries

0.8.1 - 2022-04-01 12:00:00

Added

• Updates PyPI.org packaging setup and includes auto-publishing GH Action (PRs #790, #795, and #797)
• Cleans up documentation issues and old erroneous references to ogusa package (PR #797)

0.8.0 - 2022-02-18 12:00:00

Added

• Adds a public capital good (i.e., infrastructure) used to produce private goods and services
• Adds a financial intermediary that links domestic and foreign savings to investment
• Improves multiprocessing with Dask
• Updates documentation
• Moves testing files outside of the ogcore package source files directory
• Tests functionality with Python 3.10.

[0.7.0] - 2021-08-30 12:00:00

Added

• This is the first release of the OG-Core model (formerly the OG-USA model)

Previous versions

Summary

• Version [0.7.0] on August 30, 2021 was the first time that the OG-USA repository was detached from all of the core model logic, which was named OG-Core. Before this version, OG-USA was part of what is now the OG-Core repository. In the next version of OG-USA, we adjusted the version numbering to begin with 0.1.0. This initial version of 0.7.0, was sequential from what OG-USA used to be when the OG-Core project was called OG-USA.
• Any earlier versions of OG-USA can be found in the OG-Core repository release history from v.0.6.4 (Jul. 20, 2021) or earlier.