## tnmea2025008

## Source details

**Canonical URL:** [tnmea2025008](https://www.imf.org/-/media/files/publications/tnm/2025/english/tnmea2025008.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/tnm/2025/english/tnmea2025008.pdf.md)
- [Structured JSON version](/-/media/files/publications/tnm/2025/english/tnmea2025008.pdf.json)

---

### Motivation: rationale, context, and gendered tax/benefit design
- Addressing gender inequality is essential to support economic growth.
- Tax and benefit policies affect social inclusion, public debt, and growth prospects, with importance underscored by the COVID-19 pandemic.
- Policymakers must identify revenue-mobilizing measures that are inclusive and growth friendly, particularly for vulnerable groups including women (Coelho and others 2022).
- Women tend to be more responsive to changes in net incomes compared with men—particularly in labor supply decisions—so tax and benefit policies affect both gender disparities and broader economic outcomes (Alesina, Ichino, and Karabarbounis 2011).
- TaxFit scope and availability:
  - TaxFit has been developed for 21 countries, with public access available for 16 of them.
  - Modules for Argentina, Brazil, Chile, Mexico, and Peru are not yet published.

- Tax-design features with gendered implications:
  - Unit of Taxation:
    - Income taxes (IT) can be applied at the individual or household (HH) level.
    - Individual taxation: each person’s tax is determined based on their own income.
    - HH systems tax families on a joint schedule; HH-level taxation raises marginal tax rates for secondary earners.
    - Family-based provisions in some individual systems can generate similar disincentives as HH taxation systems.
  - Tax Base:
    - Taxable base = value of assets subject to taxation, adjusted by exemptions and deductions.
    - Deductions of work-related expenses can encourage labor force entry.
    - Allowances/deductions/benefits to households with young children are common; design matters to ensure well-targeted work incentives and minimize foregone revenues.

- Benefit-design features affecting labor supply and gender outcomes:
  - Adequacy:
    - Inadequate social protection benefits can create significant work disincentives, especially for women with higher caregiving burdens.
    - Comprehensive benefits (childcare, family allowances, health coverage) make formal workforce participation financially feasible for women.
    - Without sufficient support, costs of paid employment—particularly for secondary earners—may outweigh potential income gains, pushing women toward informal work or out of the workforce.
  - Eligibility and Accessibility:
    - Benefits conditioned on specific income levels or hours worked may exclude those in need, particularly women working part-time or in low-income positions.
    - Accessible, universal benefits and streamlined access encourage formal participation and reduce administrative barriers.
  - Duration:
    - Longer entitlements provide income stability during unemployment or childbirth; overly long coverage may disincentivize return to work; too-short benefits may force premature reentry.
  - Gender-Sensitive Design:
    - Paid parental leave, childcare subsidies, and shared parental leave directly affect women’s labor market participation and caregiving balance.
    - Benefits contingent on a woman’s husband’s work status tend not to incentivize women’s labor supply.
  - Interaction with Work Incentives:
    - In-work benefits, employment subsidies, or contributory family allowances increase net income from formal employment—especially for secondary earners.
    - Coordination of benefits with the tax system is essential: benefit phase-outs and high marginal benefit withdrawal rates can create substantial disincentives to increase hours or earnings.
    - Combined effects of taxes and benefit withdrawals can significantly diminish returns to formal employment for secondary earners.
    - Cost of formalization = change in net income when transitioning from informal to formal work; can discourage formal employment.
    - Net income defined as income remaining after accounting for added cash benefits and deducting tax liabilities and social security contributions (SSC).

- Role of microsimulation:
  - OECD TaxBen and EUROMOD have studied gender implications in advanced economies.
  - TaxFit facilitates similar analyses in emerging market and developing economies by providing country-level diagnostics.

---

### Components of Income: definitions, inclusions, exclusions, and assumptions
- TaxFit uses individual gross income to compute:
  - income taxes (IT),
  - social security contributions (SSC),
  - family, care, in-work and unemployment benefits, and
  - social assistance.
- Benefits currently out of scope (excluded):
  - disability benefits, housing benefits, childcare provisions, sickness and injury benefits, and old-age pensions.
- Taxable Income:
  - Computed within every country module to facilitate IT calculation.
  - Generally calculated at the individual level, except where IT are levied at the household level with no option for individual filing (example: Libya) or where joint taxation is optional and HH average tax rate is lower under joint than individual taxation.
  - Taxable income = gross employment earnings less available deductions or allowances.
  - Deductions commonly include mandatory SSC and health expenses or health insurance premiums.
  - Allowances commonly include a basic personal allowance and allowances for dependent spouses and children.
- Employment Income Tax:
  - IT computed from taxable income; progressive systems use rates and thresholds increasing with taxable income; minimum tax levels applied where relevant.
- Business Income Tax:
  - Applied to net business income (revenues less eligible expenses) or to turnover where turnover tax exists.
  - Eligible business expenses are not directly included; users may specify business income or turnover and an amount of eligible expenses defined as a share of turnover.
    - Default share = 30 percent of turnover.
  - Country rules override user input in case of caps on eligible expense shares.
  - Presumptive taxation for very low-revenue firms is included (examples: Democratic Republic of the Congo, Malawi, and Nepal).
- Social Security Contributions:
  - Computed for employees, the self-employed (on business income, where applicable), and employers.
  - Employee contributions generally based on gross earnings and often up to a maximum threshold; some countries apply minimum earnings thresholds (examples: Jordan and Türkiye).
  - Employer contributions included for analysis of total labor costs and labor tax wedges.
  - Non-tax compulsory payments (mandatory contributions paid to some entity other than the government) are computed separately and included in social security totals (example: mandatory contributions to private health care providers).
- Benefit categories defined:
  - Family Benefits: transfers to ease costs of raising children and maintaining HH well-being; include benefits for dependents and lone parents; maternity and parental benefits are classified under Care Benefits.
  - Care Benefits: maternity, paternity, parental leave, and transfers targeted at pregnant women or new mothers.
  - Unemployment Benefits: insurance, assistance, and individual savings accounts (forced savings).
  - Social Assistance: targeted to most vulnerable groups, eligibility via (proxy) means test and possible behavioral requirements.
  - In-Work Benefits: encourage labor force participation, often targeted to specific groups and may be time-limited or phased out.
- Assumptions (selected, preserved exactly where numeric):
  - Age: Adults are aged 18–64.
  - Children: Are unmarried, without a disability, and do not work if they are school-aged (25 or younger).
  - Business Income: Corresponds to revenues of a small business with no employees.
  - Work History: Workers are assumed to have worked continuously because of the age of 20 with no unemployment spells (overridable).
  - Minimum Wages:
    - When multiple minimum wages exist, model uses the lowest wage applicable to nonagricultural sectors or the wage applicable in the capital city.
    - Daily, weekly, and monthly minimum wages are annualized, assuming 5 working days per week and 52 working weeks per year.
  - Effective Date: July 1 of given year.
  - SSC: Contribution rates based on experience are set at their lowest levels.
  - Pregnancy: Pregnant women successfully give birth to a single healthy child in a hospital setting.
  - Maternity Leave: Mothers are assumed to take maternity leave after giving birth.
  - Paternity Leave: When fathers have the option but are not legally obliged, they are assumed to take the maximum paid leave.
  - Parental Leave: Mothers take all parental leave that can be shared between parents.
  - Proxy Means Testing: TaxFit assumes presence or absence of other factors that provide the most favorable eligibility outcome for the HH.
  - Single Parents: Single parents are divorced from the parent(s) of their children.
  - Alimony: Alimony or child support payments because of former partners are assumed to be withheld.

---

### Methodology: inputs, function arguments, reference wages, program/parameter files, and data handling
- Primary inputs (four required, one optional):
  - yReference wages: Average wage (AW) for a full-time formal employee or the statutory minimum wage.
  - yFunction arguments provided by the model’s users which specify hypothetical households.
  - yProgram files outlining tax and benefit policy structure for a given country and year.
  - yParameter files documenting rates and thresholds related to modeled policies.
  - Optional: observational data (HH survey or administrative microdata) to replace simulated households.
- Function arguments: purpose and examples
  - Generate synthetic households; types: text-based (strings) or number-based (numlists, integers, numeric inputs).
  - Capabilities: define HH characteristics, earnings levels, eligible expenses, countries and years, and practical parameters for input/output locations.
  - Employment status options:
    - emp — A formal work arrangement. Assumed to be full-time work.
    - inf — An informal work arrangement.
    - unemp — An unemployed person.
    - oow — An out-of-work individual.
  - Earnings handling:
    - For workers, earnings represent current earnings.
    - For unemployed, earnings represent previous earnings used to compute unemployment benefits.
  - Selected default values and argument examples (preserved exactly):
    - year: numlist — List of modeled years
    - country: string — List of publicly available modeled countries
    - surveyfile: string — Survey data input file location and name (dta, csv, xls, xlsx)
    - spouse: string — Default: No — Presence of legal cohabitating partner: “yes”/ “no”
    - earn: string — Default: AW — Principal: Annual earnings
    - earnsp: string — Default: 0 — Spouse: Annual earnings
    - empp: string — Default: EmpPrincipal: Employment status [emp, unemp, oow, inf]
    - emps: string — Default: OowSpouse: Employment status [emp, unemp, oow, inf]
    - age: integer — Default: 40 — Principal: Age (in years)
    - agesp: integer — Default: 40 — Spouse: Age (in years)
    - gender: string — Default: M — Principal: Gender [M, F]
    - gendesp: string — Default: F — Spouse: Gender [M, F]
    - binc: string — Default: 0 — Principal: Annual gross business income
    - bincsp: string — Default: 0 — Spouse: Annual gross business income
    - medexpenses: numeric — Default: 0 — Principal: Annual out-of-pocket medical expenses
    - medspouse: numeric — Default: 0 — Spouse: Annual out-of-pocket medical expenses
    - contributionhist: integer — Default: 240 — Principal: Social security contribution history (months)
    - contsphist: integer — Default: 240 — Spouse: Social security contribution history (months)
    - udur: integer — Default: 2 — Principal: Duration in unemployment (months)
    - udursp: integer — Default: 2 — Spouse: Duration in unemployment (months)
    - time: integer — Default: 2 — Household: Program duration (months)
    - nchildren: numlist — Default: 2 — Number of dependent children—max 10
    - ach1..ach10: integer — Defaults: ach1 = 2, ach2 = 4, ach3 = 6, ach4 = 8, ach5 = 10, ach6 = 12, ach7 = 14, ach8 = 16, ach9 = 18, ach10 = 20 — Age of children—max 25
    - bincr: numeric — Default: 0.7 — Ratio of gross business income to revenues (for inference)
    - outvars: string — CoreOutput variable format: trim, core, extended, all
    - progloc: string — Location of PROG and PARM files
    - yeararg: Use survey data: use the year function argument as opposed to the year listed in the survey data
  - Note: Function arguments with partially bolded names can be provided using only the bolder portions. AW = average wage.

- Reference Wages:
  - External database contains AWs, minimum wages, and average incomes by decile for each country-year selected.
  - AW definition: average annual wage of a full-time formal worker who has worked all year long; follows OECD methodology.
  - Adjustments when only compensation of employees reported: impute wages by dividing by one plus the employer’s contribution rate.
  - Data sources: ILO’s Labour Income Distribution Database (ILO 2022), labor share of nominal GDP estimates, total employment (ILO-modeled estimates), and nominal GDP from the IMF’s World Economic Outlook (IMF 2024).
  - Countries reporting only compensation of employees: Angola, the Democratic Republic of Congo, Egypt, Georgia, Indonesia, Iran, Jordan, Philippines, Thailand, and Türkiye.
  - Countries not publishing income-approach components of GDP: Libya, Madagascar, Malawi, Montenegro, Nepal, and Vietnam.
  - TaxFit countries that rely on imputed data: Angola, Madagascar, Malawi, Montenegro, Georgia, and Libya.
  - Annualization conventions:
    - Monthly values × 12.
    - Weekly amounts × 52.
    - Daily amounts × 5 × 52.
    - Hourly amounts × 40 × 52.

- Parameter and Program Files:
  - Naming convention required:
    - Parameter files: [CNT]_PARM_[YYYY].do
    - Program files: [CNT]_PROG_[YYYY].do
  - Parameter file prefixes:
    - p_* : Parameter
    - r_* : Rate
    - t_* : Thresholds
  - Program files: main program calling smaller programs for specific policy areas; individual-based policies run separately; HH-level programs run once with benefits attributed to HH or partners based on share of earnings.

- Observational Data:
  - Option to import survey or administrative microdata; each row must represent a HH with variables aligning with function arguments.
  - Example required variable names (preserved exactly): country, iso3, year, earn_p, earn_s, binc_p, binc_s, brev_p, brev_s, Spouse, emp_p, emp_s, age_p, age_s, gender_p, gender_s, Nchild, ach1..ach10, medical_exp_p, medical_exp_s, health_ins_p, health_ins_s, travel_exp_p, travel_exp_s, cont_hist_p, cont_hist_s, udur_p, udur_s.
  - Missing variables in uploaded data: model computes based on function arguments provided when loading data (e.g., binc_p and binc_s default to the binc and bincsp function arguments, zero by default).
  - Recommended workflow: construct two data sets (minimal required to run simulations; second with additional HH info) and merge outputs after running simulations.
  - Units: wages and expense amounts interpreted as local currency; all financial amounts reported in annual terms.
  - Users can specify ISO3 code in either country or iso3 variable.

---

### Uses of the Model: comparisons, reform analysis, indicators, and fiscal analysis
- Cross-Country Comparisons:
  - Rebase net income components from local currency to a common comparator such as AW or annualized minimum wage.
- Reform Analyses:
  - Explore counterfactual reform scenarios by adjusting structural components of the model (tax rates, benefit amounts, or introducing new measures).
  - Model can evaluate packages of multiple reforms on HH net income, labor tax wedges, marginal effective tax rates (METRs), fiscal costs, and revenue generation.
  - Recent uses include:
    - Childcare support for parents of young children in Türkiye.
    - Introduction of progressive income taxation in Montenegro.
- Key Indicators produced by TaxFit (definitions and formulas preserved):
  - Effective Tax Rates (ETR):
    - ETR =  (gr oss − netinc) / gr oss  =   (benefits − taxes) / gr oss
    - METR = 1 − (∆ benefits − ∆ taxes) / ∆ gr oss
    - Users compute METRs by taking the difference in ETR of two simulated households that differ in gross income.
  - Labor Tax Wedge:
    - Tax wedge =  (it + SSC + SSCR) / (gr oss + SSCR)
    - Components: it (employee income taxes), SSC (employee social security contributions), SSCR (employer social security contributions).
  - Marginal Participation Tax Rate of Secondary Earner (MPTS):
    - MPTS =  ( nt_sp −  nt_s0 ) / ( y_sp −  y_s0 )  =   ( nt_sp −  nt_s0 ) / y_p
    - Uses sole-earner HH as benchmark; note nt_sp ≠ nt_s + nt_p in general because of joint filing or HH provisions.
  - Cost of Jointness for Secondary Earners (CJS):
    - CJS =  ( nt_sp −  nt_s0 ) / ( y_sp −  y_s0 )  −  ( nt_p ) / y_p
  - Formal Work Disincentives (FWD):
    - For single individuals:
      - FWD_p =  ( nt_p + sscl_p − b_p ) / y_p
    - For secondary earner in married couple:
      - FWD_sp =  ( nt_sp + sscl_sp − b_sp  −  ( nt_s0 + sscl_s0  −  b_s0 ) ) / y_p
    - Interpretation: positive FWD value provides an incentive to work in the formal sector for a risk-neutral individual; negative value encourages informality.
  - Progressivity Indicators:
    - TaxFit can produce the Kakwani index where data exist: ratio between the Gini coefficient of net tax liabilities and the Gini coefficient of gross income.
- Fiscal Policy Analyses:
  - TaxFit supports fiscal analysis and policy reports on gender disparities and tax-benefit effects.
  - Example applications referenced in the source include IMF TaxFit work on Latin American countries and a comprehensive gender-related analysis for Greece.

---

### User interface, data outputs, examples, and warnings/error codes
- Stata graphical interface:
  - After installing user-defined program, GUI available at Users > TaxFit.
  - Panel sections: Main options, Adult characteristics, Incomes, Children characteristics, Input data, Miscellaneous options.
  - Installation instructions included in the readme file of the TaxFit Stata function (reference provided in source).
- Main options and model year:
  - Currently, only the year 2022 is available in the model.
  - Country selection: single ISO3 code or multiple via Ctrl+Click or regional aggregations: AFRICA, ASIA, MENA, EUROPE.
- Income syntax and examples (preserved verbatim):
  - AW = 100% of AW = 100AW
  - 50AW = 50% of AW
  - 1/100AW = 1AW, 2AW, 3AW, ..., 100AW
  - 0(50)200AW = 0AW, 50AW, 100AW, 150AW, 200AW
  - MIN = 100% of minimum wage = 100MIN
  - 20(40)180MIN = 20MIN, 60MIN, 100MIN, 140MIN, 180MIN
  - 55/56MIN = 55% of MIN and 56% of MIN
- Examples of command-line use (high-level highlights):
  - Example 1 — Most Basic Run: year(2022) country(AGO); missing inputs reset to defaults.
  - Example 4 — Adding Countries and Wages: earn(0(20)100AW) equivalent to “0AW 20AW 40AW 60AW 80AW 100AW”; example yields 108 lines of output: 36 lines per country (six scenarios × six principal-earner wages).
  - Example 5 — Loading Survey Data: specifying surveyfile() causes country and year to be taken from survey data (example uses Türkiye 2021 survey); output rows equal survey row count.
- Output levels and identities:
  - Trim output: outvars(trim) — reports 14 variables per HH; identity preserved exactly:
    - netinc = gross + sa + fb + care + ub + iw − it − ssc
  - Core output: outvars(core) — reports 35 variables per HH; includes inctax_p, inctax_s, udur_p, udur_s, run_id, and other indicators.
  - Extended output: outvars(extended) — decompositions into high-level components (e.g., care_mat_ben).
  - Country-specific output: outvars(all) — complete list of variables used; caution about errors when running multiple countries due to too many output variables.
  - run_id variable format concatenates many HH characteristics (list preserved in source).
- Warnings and Error Codes (selected and preserved):
  - Warnings examples:
    - “SPouse() contained no valid values. SPouse() reset to default ‘no’ value.”
    - Warning: surveyfile() is specified, but datalocation() is not.
    - Warning: [ISO3 Code] is not currently a country included in the TaxFit model, it will be excluded from the results.
    - Warning: `gender_test_p’ value(s) of principal gender (gender_p) missing or incorrectly specified. They have been set as ‘M’ for Male.
  - Error Codes (selected):
    - Code 701: Issue: Survey data file type not specified. Potential Resolution: Provide a valid value to the surveyfile() function argument.
    - Code 702: Issue: Valid year not specified and not viable alternatives available from survey data. Potential Resolution: Ensure a valid column titled “year” or enter a valid year value to the year() function argument.
    - Code 705: Issue: Could not find PARM file. Potential Resolution: Verify that the Parameter file for the modeled country and year is included in the file location specified in the progloc() function argument.
    - Code 710: Issue: Could not save output file to output location. Potential Resolution: Check outname() and outloc() values for validity; ensure no file with same name is open.
    - Code 711: Issue: Net income validation error. Potential Resolution: Contact the TaxFit maintainer.

- Inputs and predefined household specification (selected):
  - Users can choose number of children, newborn indicator, pregnancy, marital status, woman’s and husband’s annual earnings.
  - Earnings expressed relative to national AW in four categories: no income, low income, middle income, and high income.
  - Income-category mapping:
    - “Low Earnings” corresponds to the average of HH-type earning 20, 40, 60, and 80 percent of the national average wage (AW).
  - Constraints and validation rules (examples):
    - Users cannot set marital status to “Married” when the husband’s earnings are greater than 0 or “No Income.”
    - Users shall not specify households with zero children while selecting “yes” for the newborn option.
  - Example input specification (Table 6.2) preserved:
    - Country: DRC, Indonesia, Jordan, Philippines, Vietnam
    - Year: 2022
    - Number of children: 2
    - Newborn: Ye s
    - Pregnancy: No
    - Marital status: Married
    - Woman’s annual earnings: Low earnings
    - Husband’s annual earnings: High earnings

- Participation Tax Rate (PTR) and cost-of-formalization diagnostics:
  - PTR = additional tax a woman owes when entering workforce as a share of additional income she brings to the HH.
  - Observations in example: PTR of a woman married to a high-income spouse higher than that of a single woman with same characteristics in many countries; PTRs of low-income women in Philippines and Vietnam are zero in example.
  - Cost of formalization = change in net income moving from informal to formal work, holding gross income and HH characteristics constant:
    - Negative value: financially beneficial to enter formal labor market.
    - Positive value: tax-benefit system creates a net cost to formal work.

---

*Source: IMF TaxFit Model Methodology and User Guidebook (tnmea2025008).*

### 1.      Motivation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 

### 1.      Motivation.

### Rationale and context
- Addressing gender inequality is essential to support economic growth.
- Tax and benefit policies play an important role in improving social inclusion, tackling high levels of public debt, and improving growth prospects—issues underscored by the COVID-19 pandemic.
- Policymakers must identify measures that can boost revenue mobilization in a way that is both inclusive and growth friendly, particularly among vulnerable groups, including women who are often over-represented at the bottom of the income distribution (Coelho and others 2022).
- Women tend to be more responsive to changes in net incomes compared with men—notably in terms of labor supply decisions—implying that tax and benefit policies affect both gender disparities and broader economic outcomes (Alesina, Ichino, and Karabarbounis 2011).

### Tax-design features with gendered implications
- Unit of Taxation
  - Income taxes (IT) can be applied at the individual or household (HH) level.
  - Individual taxation: each person’s tax is determined based on their own income.
  - HH systems tax families on a joint schedule. HH-level taxation ensures equal treatment across households but raises marginal tax rates for secondary earners.
  - Family-based provisions in some individual systems can generate similar disincentives as HH taxation systems.
- Tax Base
  - The tax base is the value of all assets subject to taxation, often adjusted by exemptions and deductions to favor certain activities.
  - Deductions of work-related expenses can encourage labor force entry.
  - Allowances, deductions, or benefits to households with young children are common; specific design is key to ensure well-targeted work incentives that minimize foregone revenues.

### Benefit-design features affecting labor supply and gender outcomes
- Adequacy of Benefits
  - Inadequate social protection benefits can create significant work disincentives, especially for women with higher caregiving burdens.
  - Comprehensive social benefits—childcare, family allowances, and health coverage—make formal workforce participation financially feasible for women.
  - Without sufficient support, costs of paid employment—particularly for secondary earners—may outweigh potential income gains, pushing women toward informal work or out of the workforce.
  - Adequate benefits also affect poverty reduction and economic stability, particularly in low-income households.
- Eligibility and Accessibility of Benefits
  - Eligibility criteria and ease of access shape labor supply incentives.
  - Benefits conditioned on specific income levels or hours worked may exclude those in need, particularly women working part-time or in low-income positions.
  - Accessible, universal benefits and streamlined access encourage women’s participation in the formal workforce and reduce administrative barriers to claiming support.
- Duration of Benefits
  - Longer benefit entitlements provide income stability during unemployment or childbirth.
  - Benefits that cover complete absence from work for too long may disincentivize return to work (for example, parental benefits encouraging extended leaves).
  - Time-limited or too-short benefits may force premature reentry into the workforce.
- Gender-Sensitive Benefits Design
  - Benefits targeting gender-specific needs—paid parental leave, childcare subsidies—directly affect women’s labor market participation.
  - Childcare subsidies reduce caregiving costs, making formal work more appealing.
  - Flexibility (for example, shared parental leave) encourages balanced caregiving roles.
  - Gender-insensitive designs may reinforce traditional roles and discourage labor force participation; benefits contingent on a woman’s husband’s work status tend not to incentivize women’s labor supply.
- Interaction with Work Incentives
  - In-work benefits, employment subsidies, or contributory family allowances increase net income from formal employment, affecting labor supply—especially for secondary earners who are frequently women.
  - Contributory allowances provide additional income under special circumstances (for example, breastfeeding, childrearing).
  - Coordination of benefits with the tax system is essential: benefit phase-outs and high marginal benefit withdrawal rates can create substantial disincentives to increase hours or earnings.
  - Combined effects of taxes and benefit withdrawals can significantly diminish returns to formal employment, particularly for secondary earners.
  - The cost of formalization—the change in net income when transitioning from informal to formal work—can further discourage formal employment.
  - Net income is defined as income remaining after accounting for added cash benefits and deducting tax liabilities and social security contributions (SSC).

### Role of microsimulation and TaxFit
- The gender implications of the taxation of labor income and benefits have been studied in advanced economies using microsimulation models such as the OECD TaxBen model and the European Commission’s EUROMOD.
- TaxFit facilitates similar analyses in emerging market and developing economies by providing country-level diagnostics of tax and benefit regimes for 21 countries.
  - TaxFit has been developed for 21 countries, with public access available for 16 of them. Modules for Argentina, Brazil, Chile, Mexico, and Peru are not yet published.

*IMF TaxFit Model Methodology and User Guidebook — Section 1. Motivation*

### 4. Components of Income

### 4. Components of Income

### Overview
- The TaxFit model uses individual gross income to compute:
  - income taxes (IT),
  - social security contributions (SSC),
  - family, care, in-work and unemployment benefits, and
  - social assistance.
- Several cash benefits are currently out of scope, including disability benefits, housing benefits, childcare provisions, sickness and injury benefits, and old-age pensions.
- The following sections outline simulated outputs common to all country modules.

### Taxable Income
- Taxable income is computed within every country module to facilitate the calculation of IT.
- Taxable income is generally calculated at the individual level, except:
  - where IT are levied at the household (HH) level with no option for individual tax filing (for example, Libya), or
  - where joint taxation is optional and the HH average tax rate is lower under joint than under individual taxation.
- Taxable income is generally computed as gross employment earnings less any available deductions or allowances.
  - Deductions commonly include mandatory SSC and health expenses or health insurance premiums.
  - Allowances commonly include a basic personal allowance and allowances for dependent spouses and children.

### Employment Income Tax
- Employment income tax (IT) is computed from taxable income.
- Progressive IT systems involve a series of tax rates and thresholds, increasing in taxable income.
- Minimum tax levels are implemented where applicable.

### Business Income Tax
- Business IT is applied to net business income—defined as revenues less eligible expenses—or directly applied to business revenues where a turnover tax is in place.
- Eligible business expenses are not directly included in the model; users may specify business income or turnover (revenues) and an amount of eligible expenses defined as a share of turnover.
  - By default, this share is set to 30 percent of turnover.
  - Note: some countries implement caps on the share of turnover that is eligible to be expensed; country rules override user input in case of conflict.
- Presumptive taxation for firms with very low revenues (for example, Democratic Republic of the Congo, Malawi, and Nepal) is included.

### Social Security Contributions
- SSC are dedicated taxes paid by workers and firms to finance the social safety net (commonly old-age pensions, unemployment benefits, and sometimes medical coverage).
- Contributions are computed for employees, the self-employed (on business income, where applicable), and employers.
- Employee contributions are generally calculated based on gross earnings and often up to a maximum threshold. In some cases (for example, Jordan and Türkiye), minimum earnings thresholds are also applied.
- Employer contributions, while not deducted from a worker’s gross income, are included for analysis of total labor costs and labor tax wedges.
- The model includes non-tax compulsory payments (mandatory contributions paid to some entity other than the government), which are computed separately and included in social security totals for reporting purposes (example: mandatory contributions to private health care providers).

### Family Benefits
- Government transfers to households to ease costs of raising children and maintaining HH well-being.
- Include benefits for dependents (spouse or young children) and for lone parents when applicable.
- Do not encompass maternity and parental benefits, which are classified as care benefits.

### Care Benefits
- Care benefits and paid leave allowances support care of newborn children and are provided by government or employers.
- Include maternity, paternity, and parental leave, and transfers targeted at pregnant women or new mothers (such as prenatal and nursing benefits).

### Unemployment Benefits
- Three types of benefits included: insurance, assistance, and individual savings accounts.
- Unemployment insurance:
  - Requires formally employed workers to make contributions.
  - Benefits usually calculated as a share of earnings prior to separation and program contribution history, often capped for high earners and available for fixed durations.
- Unemployment assistance:
  - Provides support for workers not eligible for insurance.
  - Not tied to prior contributions, often minimal amounts, and limited duration.
- Individual savings accounts:
  - Forced savings where contributions accrue in individual accounts workers can draw from under specific circumstances (including unemployment).
  - Often serve as primary support for jobseekers in some Latin American countries.

### Social Assistance
- Addressed to the most vulnerable groups, often emergency assistance.
- Can include supplementary amounts for households with children.
- Eligibility subject to a (proxy) means test and may include behavioral requirements (e.g., regular medical checkups, mandatory school attendance).

### In-Work Benefits
- Seek to encourage labor force participation.
- Often targeted to specific groups (for example, low-income women or youth).
- May be time-limited or phased out as income increases.

### Benefits Currently Out of the Scope of the Model
- Housing benefits paid to households to assist with housing costs.
- Disability benefits paid to people who have a disability that interferes with ability to earn an income.
- Sickness and injury benefits such as short-time leave.
- Childcare benefits, subsidies, or tax incentives provided directly to parents or indirectly to childcare providers to offset the cost of childcare.
- Old-age and survivor pensions paid to retired workers.
- Future developments may introduce some of these benefits.

### A. Assumptions
- Income Types: Income other than employment earnings, self-employment income, and cash benefits is excluded (includes capital income such as dividends and royalties, capital gains, and other forms).
- Age: Adults are aged 18–64.
- Children: Are unmarried, without a disability, and do not work if they are school-aged (25 or younger).
- Business Income: Corresponds to revenues of a small business with no employees. Paid voluntary or mandatory SSC are not deducted from business income because these deductions are explicitly calculated in the model.
- Eligibility: In addition to universal benefits, individuals and households who meet eligibility criteria regarding income, employment status, and HH composition are assumed to receive benefits. Selective or geographically targeted programs fall outside the scope.
- Work History: Workers are assumed to have worked continuously because of the age of 20 with no unemployment spells. It is possible to override this assumption for computing contribution history.
- Employer Nationality: Employers are domestic firms and are not affiliated with foreign enterprises.
- Tax Credits: Any accumulated tax credits or carryovers from previous tax years are ignored.
- Work Type: Employees are assumed to work full-time hours on permanent contracts.
- Unemployment: Unemployment is involuntary, and the unemployed actively search for a job.
- Joint Taxation: When some conditions are to be met for joint taxation, they are assumed to be met.
- Minimum Wages:
  - When multiple minimum wages exist, the model uses the lowest wage applicable to nonagricultural sectors or the wage applicable in the capital city.
  - Daily, weekly, and monthly minimum wages are annualized, assuming 5 working days per week and 52 working weeks per year.
- Effective Date: July 1 of given year.
- Timing: The model ignores timing of tax payment schedules and does not consider credits from taxes withheld at source during the tax year, or tax prepayments made on a previous year’s tax bill.
- SSC: Contribution rates based on experience are set at their lowest levels. Occupation-specific rates are generally excluded unless there is a rate for every occupation, in which case the rate covering the largest portion of workers is applied.
- Social Security Administration: When workers have an option between public and private systems, they opt for the public system (relevant for health care providers).
- Pregnancy: Pregnant women successfully give birth to a single healthy child in a hospital setting.
- Maternity Leave: Mothers are assumed to take maternity leave after giving birth. Adoption of young children is generally covered but paid leave may be shorter.
- Paternity Leave: When fathers have the option but are not legally obliged, they are assumed to take the maximum paid leave.
- Parental Leave: Mothers take all parental leave that can be shared between parents.
- Proxy Means Testing: For programs employing proxy means testing, TaxFit relies on income and HH characteristics included in each proxy means test and assumes presence or absence of other factors that provide the most favorable eligibility outcome for the HH.
- Single Parents: Single parents are divorced from the parent(s) of their children.
- Alimony: Alimony or child support payments because of former partners are assumed to be withheld.

*IMF TaxFit Model Methodology and User Guidebook — Chapter 4: Components of Income*

### 5.  Methodology

### 5.  Methodology

### Overview of model inputs
- The TaxFit model takes four primary inputs, with an optional fifth input:
  - yReference wages: Average wage (AW) for a full-time formal employee or the statutory minimum wage.
  - yFunction arguments provided by the model’s users, which specify the hypothetical households for which the model will provide results.
  - yProgram files outlining the structure of tax and benefit policies in a given country and year.
  - yParameter files that document the rates and thresholds related to modeled policies.
  - Optional: observational data—taken either from household (HH) survey or administrative microdata—which replace the simulated households created by the function arguments.
- The master TaxFit program incorporates these inputs to generate cross-country comparable indicators representing the burden and generosity of taxes and benefits.

### A. Function Arguments
- Purpose: generate a database of synthetic households to which tax and benefit rules are applied.
- Types: text-based (strings) or number-based (numlists, integers, numeric inputs).
- Capabilities:
  - Define HH characteristics, earnings levels, and eligible expenses for each potential worker.
  - Specify countries and years to be modeled.
  - Specify practical parameters for input/output file locations.
- Employment status options and interpretations:
  - emp — A formal work arrangement. Assumed to be full-time work. These workers pay taxes and social security contributions (SSC) on their earnings or self-employment income.
  - inf — An informal work arrangement. These workers generally do not pay taxes or SSC but may be eligible for some benefits.
  - unemp — An unemployed person. Not working but actively looking for a job and fulfills behavioral criteria for jobseekers to receive unemployment benefits (if relevant).
  - oow — An out-of-work individual. Not working and not looking for a formal job.
- Earnings handling:
  - For workers, earnings represent current earnings.
  - For unemployed, earnings represent previous earnings and are used to compute unemployment benefits.
- Selected default values and argument examples (preserved exactly as presented):
  - year: numlist — List of modeled years
  - country: string — List of publicly available modeled countries
  - surveyfile: string — Survey data input file location and name (dta, csv, xls, xlsx)
  - spouse: string — Default: No — Presence of legal cohabitating partner: “yes”/ “no”
  - earn: string — Default: AW — Principal: Annual earnings
  - earnsp: string — Default: 0 — Spouse: Annual earnings
  - empp: string — Default: EmpPrincipal: Employment status [emp, unemp, oow, inf]
  - emps: string — Default: OowSpouse: Employment status [emp, unemp, oow, inf]
  - age: integer — Default: 40 — Principal: Age (in years)
  - agesp: integer — Default: 40 — Spouse: Age (in years)
  - gender: string — Default: M — Principal: Gender [M, F]
  - gendesp: string — Default: F — Spouse: Gender [M, F]
  - binc: string — Default: 0 — Principal: Annual gross business income
  - bincsp: string — Default: 0 — Spouse: Annual gross business income
  - medexpenses: numeric — Default: 0 — Principal: Annual out-of-pocket medical expenses
  - medspouse: numeric — Default: 0 — Spouse: Annual out-of-pocket medical expenses
  - contributionhist: integer — Default: 240 — Principal: Social security contribution history (months)
  - contsphist: integer — Default: 240 — Spouse: Social security contribution history (months)
  - udur: integer — Default: 2 — Principal: Duration in unemployment (months)
  - udursp: integer — Default: 2 — Spouse: Duration in unemployment (months)
  - time: integer — Default: 2 — Household: Program duration (months)
  - nchildren: numlist — Default: 2 — Number of dependent children—max 10
  - ach1..ach10: integer — Defaults: ach1 = 2, ach2 = 4, ach3 = 6, ach4 = 8, ach5 = 10, ach6 = 12, ach7 = 14, ach8 = 16, ach9 = 18, ach10 = 20 — Age of children—max 25
  - bincr: numeric — Default: 0.7 — Ratio of gross business income to revenues (for inference)
  - outvars: string — CoreOutput variable format: trim, core, extended, all
  - progloc: string — Location of PROG and PARM files
  - yeararg: Use survey data: use the year function argument as opposed to the year listed in the survey data
- Note: Function arguments with partially bolded names can be provided using only the bolder portions. AW = average wage.

### B. Reference Wages
- Wages are a core input; an external database bundled with TaxFit contains:
  - AWs, minimum wages, and average incomes by decile for each country-year selected.
- AW definition and computation:
  - AW captures the average annual wage of a full-time formal worker who has worked all year long.
  - AWs follow OECD methodology: divide total wages and salaries listed in the National Accounts by the number of employees in the total economy, then multiply by the ratio of total hours worked to the number of hours worked in a scenario where all employees work full time.
- Adjustments for data gaps and employer SSC:
  - Some countries report only compensation of employees (which includes employer’s SSC); because TaxFit simulates employers’ contributions explicitly, employer contributions should not be included in wage figures.
  - In such cases, wages and salaries are imputed from compensation of employees by dividing by one plus the employer’s contribution rate.
- Data sources and imputation:
  - For countries that do not publish income-approach components of GDP, wage estimates rely on the ILO’s Labour Income Distribution Database (ILO 2022), estimates of the labor share of nominal GDP, information on total employment (ILO-modeled estimates), and nominal GDP in local currency from the IMF’s World Economic Outlook (IMF 2024).
  - An adjustment to account for employers’ SSC is made.
- Countries noting reporting issues or imputed series (as enumerated in the source):
  - TaxFit countries that report only compensation of employees include Angola, the Democratic Republic of Congo, Egypt, Georgia, Indonesia, Iran, Jordan, Philippines, Thailand, and Türkiye.
  - Some countries do not publish the income-approach components of GDP; these include Libya, Madagascar, Malawi, Montenegro, Nepal, and Vietnam.
  - The following TaxFit countries rely on imputed data: Angola, Madagascar, Malawi, Montenegro, Georgia, and Libya.
- Minimum wage and decile data:
  - Minimum wage data are taken from official sources and adjusted to reflect annual minimum wages of a full-time worker.
  - The earnings database provides income deciles derived from the ILO’s Labour Income Distribution Database (ILO 2022), including both employment and self-employment income.
- Annualization convention:
  - Monthly values are annualized by multiplying the monthly amounts by 12.
  - Weekly amounts are multiplied by 52.
  - Daily amounts are multiplied by 5 × 52 (representing 5 working days per week).
  - Hourly amounts are multiplied by 40 × 52 (representing 40 work hours per week).

### C. Parameter and Program Files
- Purpose: construct a simulation model of the tax-benefit system for a given country and year.
- Naming convention (required for running the model):
  - Parameter files: [CNT]_PARM_[YYYY].do
  - Program files: [CNT]_PROG_[YYYY].do
  - Both files are required for a given country–year combination.
- Parameter files:
  - Outline parameters, rates, and thresholds used by program files.
  - All parameters are stored as scalars.
  - Prefix conventions (preserved exactly):
    - p_* : Parameter (for example, number of months)
    - r_* : Rate (for example, tax rates)
    - t_* : Thresholds (for example, tax brackets)
  - Parameter values are annualized to facilitate cross-country comparison.
- Program files:
  - Outline the structure of tax and benefit policies and rely on parameter files for values.
  - Organized as a main program calling smaller programs for specific tax or benefit policy areas (for example, unemployment benefits, family benefits).
  - Individually based policies are run separately; HH-level programs are run once with benefits attributed to the HH or to partners based on share of earnings within the HH.

### D. Observational Data
- Option to import survey or administrative microdata and run the model on real microdata.
- Data structure requirements:
  - Each row must represent a HH, with information broadly aligning with function arguments.
  - Variable names must follow specific naming conventions (examples preserved from Table 5.3):
    - country: String — Country name
    - iso3: String — Three-digit country code according to ISO 3166 standard
    - year: Integer — Year of analysis
    - earn_p, earn_s: Numeric — Principal/Spouse: Annual earnings
    - binc_p, binc_s: Numeric — Principal/Spouse: Annual gross business income
    - brev_p, brev_s: Numeric — Principal/Spouse: Annual business revenues
    - Spouse: String — Presence of legal cohabitating partner: “yes”/ “no”
    - emp_p, emp_s: String — Employment status [emp, unemp, oow, inf]
    - age_p, age_s: Integer — Age (in years)
    - gender_p, gender_s: String — Gender [M, F]
    - Nchild: Integer — Number of dependent children—max 10
    - ach1..ach10: Integer — Ages of children—max 25
    - medical_exp_p, medical_exp_s: Numeric — Annual out-of-pocket medical expenses
    - health_ins_p, health_ins_s: Numeric — Annual health insurance premiums
    - travel_exp_p, travel_exp_s: Numeric — Commuting travel expenses
    - cont_hist_p, cont_hist_s: Integer — Social security contribution history (months)
    - udur_p, udur_s: Integer — Duration in unemployment (months)
- Handling missing variables in uploaded data:
  - If uploaded data do not include some required variables, the model will compute them based on function arguments provided by the user when loading the data (for example, binc_p and binc_s default to the binc and bincsp function arguments, zero by default).
- Recommended workflow when using survey data:
  - Construct two data sets:
    - One containing only variables required to run simulations and identify households (to load into TaxFit).
    - A second containing all additional HH information (unique identifiers, weights, other HH characteristics).
  - After running the model with the first data set, merge the resulting output with the second data set.
- Units and conventions:
  - Wages and expense amounts are interpreted as being in the local currency of the country being modeled.
  - All financial amounts are to be reported in annual terms.
  - Users can specify the ISO3 code in either the country variable or the iso3 variable.

*Source: IMF TaxFit Model Methodology and User Guidebook — Section 5. Methodology*

### BibliographySources and classification of benefits

### BibliographySources and classification of benefits

### Inputs and predefined household specification
- Users can choose:
  - number of children in the HH
  - if the child is a newborn
  - if the woman is pregnant
  - marital status
  - woman’s annual earnings and husband’s annual earnings (when marital status is “married”)
- Earnings are expressed relative to the national AW, comprising four categories: no income, low income, middle income, and high income.
- Income-category mapping from econometric analyses:
  - “Low Earnings” corresponds to the average of HH-type earning 20, 40, 60, and 80 percent of the national average wage (AW).
  - When an aggregate income category is chosen, outputs are averaged across the corresponding earnings levels (for example, outputs for 20, 40, 60, and 80 percent of AW are averaged when “Low Earnings” is selected).
- Constraints and validation rules:
  - Users cannot set marital status to “Married” when the husband’s earnings are greater than 0 or “No Income.”
  - Users shall not specify households with zero children while selecting “yes” for the newborn option.
  - The Inputs tab controls country and HH-type selection for the entire tool; slicers in other tabs are synced but cannot change those selections.
- Example input specification (Table 6.2):
  - Country: DRC, Indonesia, Jordan, Philippines, Vietnam
  - Year: 2022
  - Number of children: 2
  - Newborn: Ye s
  - Pregnancy: No
  - Marital status: Married
  - Woman’s annual earnings: Low earnings
  - Husband’s annual earnings: High earnings

### Taxes and benefits comparison
- First display compares taxes (and social security contributions [SSC]) owed and benefits the selected HH type is entitled to across the selected countries; countries are sorted by HH net income.
- Interactive features:
  - Users can adjust marital status and earnings of both earners.
  - Clicking on individual components (for example, the dark red circle denoted by “Income Tax”) highlights corresponding data in the figure to identify which countries apply that component.
- Example observation:
  - Only the Democratic Republic of the Congo (DRC) and Indonesia levy IT for the HH configuration in the illustrative selection.

### Participation Tax Rate (PTR) analysis
- PTR definition: the additional tax a woman owes when entering the workforce as a share of the additional income she brings to the HH.
- Comparative observations:
  - In the predesigned HH example, the PTR of a woman married to a high-income spouse is higher than that of a single woman with otherwise the same characteristics.
  - This disparity is evident across the income distribution, except in Philippines and Vietnam, where the PTR of low-income women is zero.
  - PTRs of individuals married to low-income spouses are not always lower than for single individuals.
- Analytical note:
  - In the selected countries, PTRs are very similar across marital status in most countries, implying that work disincentives for secondary earners for lower-earning households are less of a concern.
  - Differences between PTRs of single and married women are often explained by tax system structure, specifically the presence of joint tax filing and family-based tax provisions (Coelho and others 2022).

### Benefits decomposition
- Benefits included in the decomposition: family benefits, social assistance, in-work benefits, and care benefits.
- Care benefits are further disaggregated into: birth grants, maternity-, paternity-, parental leave, nursing benefits, and prenatal benefits.
- Cross-country comparison (example):
  - For low-earning married couples, care benefits appear to consist of the largest share of the benefits package, with a higher contribution from family benefits.
  - If households without children or pregnancy are specified, care and family benefits drop to zero.

### Cost of formalization
- Definition: change in net income when moving from informal to formal work, assuming gross income and HH characteristics remain unchanged.
- Interpretation:
  - A negative value implies it is financially beneficial to enter the formal labor market.
  - A positive value implies that the tax-benefit system creates a net cost to formal work.
- Interactive feature:
  - Users can adjust the spouse’s income to explore alternative scenarios.
  - Hover tooltips provide explanations for interpreting the cost of formalization.

### Background data and outputs
- The Background Data tab shows, based on selected inputs, the benefits and tax obligations (as a share of the AW) simulated for the user’s country and HH-type selections.
- The table displays:
  - AW in local currency units
  - net income
  - gross income
- The table is sorted by HH net income and contains the underlying data used in the previous charts.

*Source: Authors.*

### Bibliography

### Bibliography

### Bibliography tab and documentation access
- The Bibliography tab allows users to explore the underlying laws and tax codes used to construct the TaxFit model. Users can choose the country and policy category (tax, benefit, and so on) for which they want to examine the associated documentation.
- Documentation tab: provides detailed information about the methodology and country documentation, outlining the policies and parameters included for each country model.
- Stata program tab: allows users to download the underlying Stata-based simulation model.
- Project Overview tab: lists the project’s developers and contributors and directs users to contact the project team at TaxFit@imf.org for difficulties, more information, or access to underlying data.

### Stata graphical interface (TaxFit)
- Installation: After installing the user-defined program, Stata users may use a graphical interface available at Users > TaxFit in the main menu bar.
- Panel sections presented to users:
  - Main options — Countries and model years to be simulated.
  - Adult characteristics — Employment status, age, gender, working history, and eligible expenses for each adult.
  - Incomes — Employment and self-employment income for each working adult.
  - Children characteristics — The age and number of dependent children in the HH.
  - Input data — Location and name of relevant input data.
  - Miscellaneous options — Administrative model options and options not easily classified.
- Note: Installation instructions are included in the readme file of the TaxFit Stata function, available at: https://data-download.imf.org/iDataContent/Taxfit/TaxFit.zip

### Main options and model year
- Required inputs to run the model: countries to be examined and the model year.
- Currently, only the year 2022 is available in the model.
- Country selection: single country via ISO3 code or multiple countries via Ctrl+Click or regional aggregations: AFRICA, ASIA, MENA, EUROPE.

### Adult characteristics
- Options include specifying presence of a second adult in the HH (“Secondary”), employment status, age, gender, and employment history of all adults.
- Users can specify eligible expenses for tax deductions in local current units (affects cross-country comparability).

### Incomes
- Employment earnings default to the AW in each country; can be set as fractions of AW or MIN (examples: “120AW” = 120 percent of the AW; “200MIN” = 200 percent of the minimum wage).
- Raw income values (e.g., 10,000) represent local currency units but reduce cross-country comparability.
- Self-employment income may be specified as turnover (gross revenues) or business income (turnover net of eligible expenses). If both provided, business income is used.
- Income ranges allowed, using one-unit increments (e.g., 0/100AW) or user-defined steps (e.g., 0(10)100AW). Incremental incomes compatible with reference wages and raw values.

### Children characteristics
- Users can specify up to a maximum of 10 children and their ages.
- Age constraints: greater than 0 and less than 25.
- Children aged less than 1 year are considered newborns born in the model year.
- Graphical interface currently allows only whole-number ages (cannot specify ages less than one via GUI).

### Input data
- Users specify input earnings data (reference wage) and the country directory holding parameter and program files; both fields are required and default to folders included with the TaxFit Stata package.
- To conduct reform analysis, users can point the model to altered earnings data or country files in alternate locations.
- Optional: specify microdata input sources (HH surveys or administrative data) to override function arguments. Survey rows must contain HH information (employment status, income, HH characteristics).
- Output preferences: options “Export data” and “Logfile”; default output file names if not specified: “TaxFit_output.dta” and “TaxFit_log.log”.

### Miscellaneous options
- Administrative specifications: choice of output variables and option to override survey year found in the data in favor of the year in function arguments.
- Other options: specify if an adult woman is pregnant; set assumed expense ratio between gross turnover and business income for self-employment; allow self-employed to opt into SSC; indicate whether to include mandated employer-paid benefits.
- Note: SSC = Social Security Contributions.

### Submitting commands via GUI
- “OK” button: submits the command and closes the window.
- “Submit” button: submits the command but does not close the window.
- Output is identical to running from the Stata command line and will be loaded into the data browser.

### Stata command structure and syntax
- TaxFit consists of a master ado file plus required inputs: an earnings database, and program and parameter files for each country.
- Users may call TaxFit from the Stata command line or via graphical interface.
- Condensed syntax supported for income variables (earn, earnsp, binc, bincsp, brev, brevsp).
- Income percentage and range syntax examples:
  - Use “/” for one-step increments (e.g., 1/100AW or 30/50MIN).
  - Use parentheses to specify increment size (e.g., 0(5)100AW or 100(25)200AW).
- Table examples of alternative syntax for income variables (preserved verbatim):
  - AW = 100% of AW = 100AW
  - 50AW = 50% of AW
  - 1/100AW = 1AW, 2AW, 3AW, ..., 100AW
  - 0(50)200AW = 0AW, 50AW, 100AW, 150AW, 200AW
  - MIN = 100% of minimum wage = 100MIN
  - 20(40)180MIN = 20MIN, 60MIN, 100MIN, 140MIN, 180MIN
  - 55/56MIN = 55% of MIN and 56% of MIN
- Users can type “help taxfit” in Stata for additional details.

### Command-line examples and behavior (selected highlights)
- Example 1 — Most Basic Run:
  - year(2022) country(AGO) specified; missing inputs are reset to defaults (spouse, employment statuses, number of children).
  - Console warnings example: “SPouse() contained no valid values. SPouse() reset to default ‘no’ value.” and others.
  - Default behavior when no output file location specified: data output displayed in Stata data browser only.
- Example 2 — Explicit Arguments:
  - Same as Example 1 but missing values provided explicitly (spouse(no) earn(AW) ...), console does not show defaults being supplied.
- Example 3 — Adding Household Members:
  - Multiple values in function arguments generate additional households; example yields six lines of output corresponding to scenarios with/without spouse and varying numbers of children.
- Example 4 — Adding Countries and Wages:
  - Multiple countries specified (AGO COD IDN) and earnings specified with condensed syntax earn(0(20)100AW) equivalent to “0AW 20AW 40AW 60AW 80AW 100AW.”
  - This example yields 108 lines of output: 36 lines per country (six scenarios × six principal-earner wages).
- Example 5 — Loading Survey Data:
  - Specifying surveyfile() causes country and year parameters to be taken from the survey data (example uses Türkiye 2021 survey); output rows equal to survey data row count.

### Stata output levels and key identities
- Supported output levels:
  - Trim output: reports 14 variables per HH; can be specified with outvars(trim).
    - Trim output includes eight key indicators (components of HH net income) and net income itself.
    - Identity preserved exactly:
      - netinc = gross + sa + fb + care + ub + iw − it − ssc
    - Trim output variables (listed verbatim examples): country, iso3, year, gross, sa, ub, fb, care, Iw, It, ssc, netinc.
    - run_id variable format (concatenates HH characteristics):
      - [iso3] − [year] − [spouse] − [nchild] − [ach1] − [ach2] − [ach3] − [ach4] − [ach5] − [ach6] − [ach7] − [ach8] − [ach9] − [ach10] − [gender_p] − [age_p] − [cont_hist_p] − [emp_ p] − [earn_ p_ ref] − [binc_p_ref] − [gender_s] − [age_ s] − [cont_hist_s] − [emp_ s] − [earn_s_ref] − [binc_s_ref] − [time] − [udur_p] − [udur_s] − [pregnant]
  - Core output: reports 35 variables per HH; can be specified with outvars(core).
    - Adds indicators such as spouse presence, nchild, earn_p, earn_s, binc_p, binc_s, emp_p, emp_s, sscr, aw, min, pearn_p, pearn_s, earn_p_ref, earn_s_ref, binc_p_ref, binc_s_ref, inctax_p, inctax_s, udur_p, udur_s, run_id.
    - Note: AW = average wage; IT = income taxes; HH = household.
  - Extended output: builds on core output and provides decompositions of inputs into individual high-level components (e.g., care_mat_ben for maternity benefits, care_pat_ben for paternity benefits); specify outvars(extended).
  - Country-specific output: outvars(all) returns complete list of variables used in computations (labels attached only for variables common to most country modules).
    - Caution: using outvars(all) when running multiple countries may cause Stata to throw an error because of too many output variables.

*Source: IMF TaxFit Model Methodology and User Guidebook (selected sections provided in the Bibliography chapter).*

### 7.     Uses of the Model

### 7.     Uses of the Model

### A. Cross-Country Comparisons
- Cross-country comparisons are possible when users rebase all net income components from local currency to a common unit of comparison.
- This common unit can be:
  - the national average wage (AW),
  - the annualized minimum wage, or
  - some other basis.

### B. Reform Analyses
- Advanced use: explore counterfactual reform scenarios by altering one or more aspects of the tax structure in a country by adjusting the structural components of the model.
- Reforms can:
  - examine past reforms, or
  - consider future changes to the tax and benefit system, either planned or hypothetical.
- Implementing reforms can be:
  - as simple as adjusting existing parameters—such as tax rates or benefit amounts for pre-existing programs, or
  - introducing structural policy changes which introduce new measures.
- The model’s flexibility allows policymakers to consider packages of multiple reforms to gauge overall effects on outcomes such as:
  - HH net income,
  - labor tax wedges,
  - marginal effective tax rates (METRs),
  - fiscal costs, and
  - revenue generation.
- Recent uses of TaxFit in reform or policy assessments have included:
  - Childcare support for parents of young children in Türkiye.
  - The introduction of progressive income taxation in Montenegro.

### C. Key Indicators
- The TaxFit model is equipped to generate a range of indicators, including common indicators and novel diagnostics capturing implicit biases in the law and policy-induced formal work disincentives (FWD).

- Effective Tax Rates (ETR)
  - ETR calculates the effective tax that an individual or HH pays.
  - These ETRs can change as individuals enter work or increase their work intensity (increase either their wage or their hours).
  - The marginal change in the ETR resulting from this change is called the METR.
  - Note: Effective tax rates calculated at the household level are referred to as effective average household tax rates, which measures the tax burden to the household from the employment choices of all working-age members.
  - Formulas provided in the model:
    - ETR =  (gr  oss − netinc) / gr  oss  =   (benefits − taxes) / gr  oss
    - METR = 1 − (∆ benefits − ∆ taxes) / ∆ gr  oss
  - Users compute METRs by taking the difference in ETR of two simulated households that differ in gross income.

- Labor Tax Wedge
  - Definition: the ratio between an employee’s after-tax income and an employer’s labor costs.
  - Components: employee income taxes (it), employee social security contributions (SSC), employer social security contributions (SSCR).
  - TaxFit supports calculation at the HH level within core output and at the individual level using detailed country-specific output.
  - Formula:
    - Tax wedge =  (it + SSC + SSCR) / (gr  oss + SSCR)

- Marginal Participation Tax Rate of Secondary Earner (MPTS)
  - Captures the additional tax liability from a secondary earner entering the labor force, as a share of the additional income this decision brings to the HH.
  - Uses the sole-earner HH as a benchmark.
  - Notation: let nt_sp denote the net tax liability of a HH where one person earns p and the spouse earns s.
  - Formula:
    - MPTS =  ( nt_sp −  nt_s0 ) / ( y_sp −  y_s0 )  =   ( nt_sp −  nt_s0 ) / y_p
  - Note: nt_sp ≠ nt_s + nt_p in general because joint tax filing units or HH-level provisions create non-additive net liabilities.

- Cost of Jointness for Secondary Earners (CJS)
  - Measures implicit bias in the tax law, comparing the marginal PTR of a secondary earner entering the labor force with that of a single individual of the same characteristics.
  - Formula:
    - CJS =  ( nt_sp −  nt_s0 ) / ( y_sp −  y_s0 )  −  ( nt_p ) / y_p
  - Although TaxFit output produces MPTS and CJS for taxes, users can compute similar indicators for employee SSC and benefits in principle.
  - Social security combines tax and expenditure aspects, with complex interactions; estimates of social security benefits are under development.

- Formal Work Disincentives (FWD)
  - Measures the policy-induced cost of working formally (taxes and SSC) against the cost of working informally (foregone benefits).
  - Interpretation: For a risk-neutral individual, a positive FWD value would provide an incentive to work in the formal sector; a negative value would encourage informality.
  - Computable individually or jointly for spouses.
  - For single individuals:
    - FWD_p =  ( nt_p + sscl_p − b_p ) / y_p
  - For the secondary earner in a married couple:
    - FWD_sp =  ( nt_sp + sscl_sp − b_sp  −  ( nt_s0 + sscl_s0  −  b_s0 ) ) / y_p

- Progressivity Indicators
  - Where household or revenue administration data are available, TaxFit can produce indicators of progressivity, such as the Kakwani index.
  - Kakwani index defined as the ratio between the Gini coefficient of net tax liabilities and the Gini coefficient of gross income.
  - Higher values represent higher degrees of progressivity.

### D. Fiscal Policy Analyses
- TaxFit supports fiscal policy analysis, including analytical notes and policy reports on tax and benefit policy effects on gender disparities in the labor market.
- Example applications:
  - Underpinned work in the “The role of taxes and benefits to support female labor force participation in selected Latin American Countries IMF TaxFit box” (IMF 2024), highlighting roles of taxation, social assistance, and benefits such as unemployment and parental benefits in addressing gender gaps in labor force participation and employment.
  - Comprehensive gender-related tax-benefit policy analysis for Greece, exploring interplay between gender and cultural norms, parental and care policies, taxation and social spending, including the country’s Guaranteed Minimum Income.

### Annex A. Warnings and Error Codes
- The TaxFit Stata command produces warnings and errors to help users understand model functioning.

- Warnings (examples and interpretations)
  - Warning: surveyfile() is specified, but datalocation() is not.
    - Interpretation: TaxFit cannot load in external data. Survey data cannot be loaded, TaxFit will instead generate synthetic data from function arguments.
  - Warning: datalocation() is specified, but surveyfile() is not.
    - Interpretation: Survey data cannot be loaded, TaxFit will instead generate synthetic data from function arguments.
  - Warning: surveyfile() is specified, but file type is not csv, xls, or xlsx. TaxFit cannot load in external data.
    - Interpretation: Survey data cannot be loaded, TaxFit will instead generate synthetic data from function arguments.
  - Warning: [ISO3 Code] is not currently a country included in the TaxFit model, it will be excluded from the results.
    - Interpretation: A country listed is not included in the modeled countries and will be excluded (e.g., incorrect ISO3 code).
  - Warning: [YEAR] is not currently a year included in the TaxFit model, it will be excluded from the results.
    - Interpretation: A year listed is not included in the modeled years and will be excluded.
  - Warning: SPouse() provided a value other than ‘yes’ or ‘no’. Incompatible value ‘[VALUE]’ has been removed.
    - Interpretation: spouse() contained values other than “yes” or “no”; unrecognized values are dropped and may revert to default “no.”
  - Warning: EMPp() contained values other than ‘emp’, ‘unemp’ ‘oow’ or ‘inf’. Incompatible value ‘[VALUE]’ has been removed.
    - Interpretation: empp() accepts only “emp,” “unemp,” “oow,” or “inf”; other values are ignored and empp() reverts to default “emp.”
  - Warning: NCHildren() contained no valid values. NCHildren() reset to default value of ‘2’.
  - Warning: `gender_test_p’ value(s) of principal gender (gender_p) missing or incorrectly specified. They have been set as ‘M’ for Male.
  - Warning: `gender_test_s’ value(s) of spouse gender (gender_s) missing or incorrectly specified. They have been set as ‘F’ for Female.
  - Source note: DRC = Democratic Republic of the Congo; HH = household.

- Error Codes (selected codes, issues, and potential resolutions)
  - Code 701
    - Issue: Survey data file type not specified
    - Potential Resolution: Provide a valid value to the surveyfile() function argument.
  - Code 702
    - Issue: Valid year not specified and not viable alternatives available from survey data
    - Potential Resolution: If using survey data, ensure a valid column titled “year” in the data set. Otherwise, enter a valid year value to the year() function argument (see Table 5.2 for valid years for each country).
  - Code 703
    - Issue: Location of input wage file not specified
    - Potential Resolution: Verify the value provided to the earnloc() function argument.
  - Code 704
    - Issue: Location of PARM and PROG files not specified
    - Potential Resolution: Verify the value provided to the progloc() function argument.
  - Code 705
    - Issue: Could not find PARM file
    - Potential Resolution: Verify that the Parameter file for the modeled country and year is included in the file location specified in the progloc() function argument.
  - Code 706
    - Issue: Could not find PROG file
    - Potential Resolution: Verify that the Program file for the modeled country and year is included in the file location specified in the progloc() function argument.
  - Code 707
    - Issue: Error when running PARM file
    - Potential Resolution: There may be an error in the PARM file. Contact the maintainer or rerun with “set trace on” in Stata to locate the error.
  - Code 708
    - Issue: Error when loading PROG file
    - Potential Resolution: There may be an error in the PROG file. Contact the maintainer or rerun with “set trace on” in Stata to locate the error.
  - Code 709
    - Issue: Error with running the PROG file
    - Potential Resolution: There may be an error in the PROG file. Contact the maintainer or rerun with “set trace on” in Stata to locate the error.
  - Code 710
    - Issue: Could not save output file to output location
    - Potential Resolution: Check outname() and outloc() values for validity; ensure no file with same name is open.
  - Code 711
    - Issue: Net income validation error
    - Potential Resolution: Contact the TaxFit maintainer.

*Source: tnmea2025008 - 7.     Uses of the Model (IMF TaxFit Model Methodology and User Guidebook, TNM/2025/08).*

---


_Source: https://www.imf.org/-/media/files/publications/tnm/2025/english/tnmea2025008.pdf_