POVERTY & EQUITY NOTES DECEMBER 2020 · NUMBER 37 A Customizable Microsimulation Tool to Analyze Distributional Effects of Country Fiscal Policies Jia Gao, Gabriela Inchauste Microsimulation modelling has become a powerful tool to analyze the effects of fiscal policy changes. The World Bank’s Equity Policy Lab (EPL) has developed a customizable microsimulation tool to assess the distributional effects of tax, benefits, and other fiscal reforms. This Note explains why and how countries use the microsimulation tools, using examples from Ecuador and Armenia—2 of the more than 20 countries that have developed and used the tool over the past 2 years—to demonstrate its effectiveness in engaging government officials and informing policy making. Policy makers are eager to analyze distributional The World Bank Equity Policy Lab (EPL) has effects of potential fiscal policy reforms. They developed fiscal policy microsimulation tools to want to understand the potential effects on poverty help engage policy makers in designing tax and and inequality, populations most affected, the benefit policies. Built on the current fiscal policy progressivity of taxes, “winners and losers” from system in the country, these tools combine tax and reforms, and impacts on government fiscal budgets. benefit policies with household survey and model the Governments are increasingly using tax and details of each fiscal intervention, such as eligibility benefit microsimulation 1 tools to inform fiscal criteria for benefit programs, as well as allowances, policy. This is part of a broader movement towards exemptions and deductions for tax interventions. evidence-based public policy formulation and These tools show how tax and benefit policies reform. Microsimulation tools enable assessment of affect household incomes and display the the effects of changes in fiscal policies prior to distributional effects of alternative fiscal policies national implementation, where pilots are not cost- on income and welfare of the population, by effective or feasible. The information requirements simulating individual and household tax liabilities for developing these tools are relatively low, and and benefit entitlements. Using outcomes generated assumptions and modeling approaches are relatively with the tool, policy makers can understand what transparent. The tools are designed to be simple policies mean, in concrete terms, for poverty, enough for policy makers to use directly, alongside inequality, and government’s revenues and other evidence-based approaches, such as post- expenditures. The tools developed by EPL are more program evaluations and social/natural experiments. flexible compared to a harmonized multi-country Although microsimulation tools have been widely tool, such as the European Union’s EUROMOD, used in developed countries, few developing Modelers can include country-specific fiscal policies countries have developed this type of tool, either or characteristics, and even incorporate sub-national because policy makers are not fully aware of such tax policies for local policy makers. tools or due to lack of experts and funding to develop the tool. 1 There are many types of simulation tools in economics study. Tax and benefit microsimulation tools (or fiscal policy microsimulation tools) are the tools used in studying fiscal policies. administrative data, usually collected by the Ministry Methodology of Finance. A microsimulation tool produces several standard The microsimulation model is built on figures and tables, including: comprehensive fiscal incidence analysis 2 to study -Poverty and inequality estimates. existing and potential fiscal interventions. The -Net cash position by income decile, which assesses model is based on the Commitment to Equity (CEQ) the net receivers or net payers to the fiscal system. methodology 3, a rigorous and widely-used fiscal -Incidence by income decile, which displays taxes incidence diagnostic method. However, it goes borne (or benefits shared) by the population group. beyond fiscal incidence analysis, which simply -Progressivity of each intervention and its marginal identifies tax and benefit components observed in contribution to poverty and inequality. the survey. Microsimulation requires that each tax -Concentration curves for each intervention, which and benefit intervention can be modeled with shows proportion of taxes paid (or benefits received) enough detail to enable users to change policy by the bottom x% of population. parameters for simulation purposes. 4 -Effects on government revenues and expenditures. Generally, the tools simulate a set of taxes and In addition to these standard outputs, modelers benefits. The taxes and benefits typically include can customize the outcomes based on client direct taxes (such as personal income tax, payroll needs. The tool can be used for analysis at the taxes, agriculture taxes), social security contributions, national level or at state/provincial levels. It can be direct transfers (such as cash and food transfers, customized to show results for population subgroups unemployment benefits, child benefits, other social defined by gender, household type, etc. assistance and income-related benefits), indirect taxes (such as VAT and excise), indirect subsidies (such as energy and water subsidies), and education Data and health “in-kind” transfers. Following the CEQ approach, 4 income concepts (market income, disposable income, consumable income, and final Input requirements for the EPL tool are relatively income) are constructed by including or excluding low. The main input is data from a nationally certain fiscal interventions in order to assess the representative household survey, which contains overall impact of the fiscal system on poverty and sociodemographic characteristics, labor market inequality. outcomes, income from various sources, and detailed consumption information. For countries that collect Following best practices, baseline model results income and consumption information in different are validated and tested at both the micro and surveys, a survey-to-survey algorithm can match macro level. At the micro level, simulated taxes and different household data. For surveys with transfers are compared with respondent survey employment questionnaires rather than individual answers. At the macro level, the estimates for total income information, market income can be imputed tax collection, transfers, and number of with a “consumption-to-income” procedure using taxpayers/beneficiaries are compared with 2 Fiscal incidence analysis studies the economic impact of taxes and expenditures on households. 3 For more information on CEQ, please see Lustig, Nora, editor. 2018. Commitment to Equity Handbook: Estimating the Impact of Fiscal Policy on Inequality and Poverty. Washington, DC: Brookings Institution Press and CEQ Institute, Tulane University. 4 The caveats of the approach are, first, the microsimulation model does not consider general equilibrium effects; therefore, results should be interpreted as short-term effects. Second, as with most static models, the simulations usually do not consider behavioral responses. Finally, household surveys typically do not have families with extremely high incomes, so the tool cannot evaluate effects on these households unless the survey is merged with tax and administrative records on top earners. December 2020 · Number 37 2 Mincer regressions 5 and household consumption. Ecuador Main macro inputs are an input-output matrix, which shows the interdependencies between different The Government of Ecuador (GoE) invited the sectors of the economy and recent macroeconomic World Bank to assess the distributional impact of statistics, such as GDP growth, inflation, poverty, and various fiscal reform proposals, compare unemployment. In addition, modelers need to collect different tax-subsidy systems, and help select the information on tax and benefit policies, usually from most efficient and effective policies. In July 2019, the Ministry of Finance, including revenue collection upon request from the Ministry of Economics and and expenditures and the number of taxpayers and Finance (MoEF), EPL staff traveled to Ecuador to help beneficiaries. build a fiscal policy microsimulation tool (ECUTOOL). The GoE planned comprehensive tax-subsidy reforms—including changes in income, corporate, Programming Language and User Interface value-added and excise taxes—and on energy subsidies. These reform proposals were scheduled to be submitted for National Assembly approval in Currently, tool modelers can choose from 2 November 2019. programming languages and user-interfaces to develop the tool. One uses Stata software combined The EPL collaborating with World Bank Ecuador with an Excel interface, and the uses R software with country team developed the ECUTOOL over 2 an R Shinny interface. Developing the Stata-plus- weeks to address the MoEF’s urgent request. The Excel option is easier for a larger share of Bank staff, team communicated with policy makers frequently to while the R-plus-R Shinny interface is more user- understand their needs and revise the tool. Following friendly, allowing a broader set of clients to use the a 2-hour training session to demonstrate the tool and tool. provide instructions to the MoEF economics team, the tool was transferred to Ecuadorian authorities. Case Studies ECUTOOL allows users to compare distributional effects of different tax and subsidy policy In the past 2 years, EPL has assisted 22 countries scenarios and to estimate the magnitude of to conduct CEQ benchmark analysis and build shocks resulting from fiscal reforms. The tool can microsimulation tools (see Table 1). Case studies for also perform sensitivity analysis by changing model Ecuador and Armenia illustrate the use of these tools assumptions and/or policy parameters. The team and team engagement with governments. later learned that the MoEF used the tool during its decision-making by simulating approximately 1000 Table 1. List of Countries with Existing or Ongoing alternative policy proposals, with results presented to Microsimulation Tool Applications Ecuador’s Vice President. Argentina Colombia Jordan Paraguay Armenia Dominican Republic Mauritius Russia Armenia Bhutan Ecuador Mexico Senegal Brazil Egypt Moldova Ukraine EPL assisted the World Bank Armenia team to Cambodia Ethiopia Mozambique develop a web-based interactive tool (ARM Web Costa Rica Indonesia Myanmar Tool). This was in response to the Armenian Note: As of end of June 2020, the countries in bold had an Ministry of Finance’s (MoF) request to help established tool, while the rest are in progress. analyze distributional effects of various taxes and transfers. The ARM Web Tool was based on an 5 The Mincer earning function is a model that explains wage as a function of schooling and experience, named after Jacob Mincer. December 2020 · Number 37 3 updated fiscal incidence analysis. The tool allows Standardizing the tool design process, such as using estimating of the distributional effects of a proposed the same coding structure and user interface, would personal income tax reform and an upcoming simplify and speed tool construction, enabling users increase of the minimum wage, as well as a number to recognize similar tax-benefit components across of additional changes in taxes or transfers. For countries. instance, the tool helped the Government of Armenia• (GoA) evaluate effects of a potential universal health• Finally, the microsimulation tool can be combined insurance reform financed partly through direct and with other macro-micro models and tools—such as indirect taxes. ADePT 6 and Computable General Equilibrium models 7— to account for external shocks, such as One of the features of the ARM Web Tool is that COVID-19, in the distributional analysis. it has a “dynamic simulation” option. This enables users to simulate tax and transfer policies for a span The EPL team plans to improve and enhance the of 6 years to observe changes in distributional effects tool. The team has begun modelling the effects of over time. This feature relies on assumptions on GDP behavioral responses and fiscal and distributional growth and inflation. In November 2019, the team issues. This is critical for evaluating dramatic changes, delivered a 2-day training to MoF staff to introduce such as removal of energy subsidies or hyperinflation the fiscal incidence methodology and demonstrate scenarios. The team is also considering developing the use of the tool, and about 11 MoF staff dynamic models to capture household whole completed the training. The tool strengthened the lifecycle income or consumption instead of just MoF’s capacity to analyze welfare impact of policy focusing on one period, which is important for reforms before implementation. policies whose benefits are difficult to observe in the short-term, such as cigarette taxes or pension reform. In addition, the team plans to develop short-term Lessons and Moving Forward complementary models using high frequency data, such as monthly income or labor force participation data, and employing time-series regressions, to help A few key lessons have emerged after policy makers monitor poverty and distribution more constructing several microsimulation tools: frequently. • Once a customized country tool is developed, it can be used for multiple years with regular updates. • Building a user-friendly interface could allow a wider ABOUT THE AUTHORS audience to access the tool, especially users who do Jia Gao is an E T Consultant in the World Bank’s not have modelling or technical skills. Poverty & Equity Global Practice. • Collaboration across World Bank practices, such as between the social protection and the macro-fiscal or Gabriela Inchauste is a Lead Economist in the energy team, can improve tool design and ensure World Bank’s Poverty & Equity Global Practice. consistent messaging to policy makers. This note series is intended to summarize good practices and key policy findings on Poverty-related topics. The views expressed in the notes are those of the authors and do not necessarily reflect those of the World Bank, its board or its member countries. Copies of this note series are available on www.worldbank.org/poverty 6 Automated Development Economics Poverty Tables (ADePT) is a tool to simulate distributional impacts of macro-dynamics, which is developed by World Bank economists. For details, please see Olivieri et al. 2014. Simulating Distributional Impacts of Macro-dynamics: Theory and Practical Applications. Washington, DC: World Bank. 7 Computable General Equilibrium (CGE) models are large numerical models which combine economic theory with real economic data in order to derive computationally the impacts of policies or shocks in the economy. December 2020 · Number 37 4