Association for Public Policy Analysis & Management

California introduced a state earned income tax credit (CalEITC) in 2015, and state policymakers have put forth multiple proposals to expand the credit to benefit more low-income households. While most state EITCs are set as a flat percentage of the federal EITC, California’s credit has a distinct structure, as it excludes self-employment earnings, provides no larger credits for married joint tax filers versus single or head of household filers with the same number of qualifying children, and targets only the lowest-income earners (with 2016 eligibility capped at earnings of $14,160 for a single parent with two children). For eligible tax filers, the CalEITC provides a credit generally equivalent to 85 percent of the federal EITC. The unique structure of the CalEITC increases the complexity of analyzing potential effects of changing or expanding the credit – analysis which is important to inform policymakers, administrators, and advocates as they actively explore alternatives to the credit’s current structure, such as increased income eligibility cutoffs and inclusion of self-employment earnings. Census American Community Survey (ACS) data represent a valuable resource for examining CalEITC policy alternatives, allowing for microsimulation of the effects of potential expansions of the CalEITC in a large dataset representative of the state population. Modeling the credit in ACS microdata also allows for estimation of dynamic effects of the credit in incentivizing increased employment, and for examining how different credit structures might interact with other state income and employment policies, such as California’s recent commitment to phase in a state minimum wage of $15 per hour. This paper applies a detailed income tax simulation model to ACS microdata for California to create estimates of CalEITC eligibility, credit amounts, filers’ net income taxes, and total policy costs under different scenarios for expanding the CalEITC. The income tax model was initially developed for use in the California Poverty Measure, a state-specific poverty measure modeled on the Supplemental Poverty Measure. The tax model involves construction of tax units in ACS data, which are then used to develop filer and income data used both as inputs into the NBER’s Taxsim tax calculator for calculation of income taxes under existing law, and as inputs to calculate CalEITC credits under alternative credit structures. The paper reviews the challenges of accurately modeling EITC claiming behavior in survey data, illustrating how tax and credit estimates are sensitive to different assumptions for the creation of tax units in ACS data, and showing that a more complex algorithm for assigning children among tax filers within extended households produces federal EITC claim estimates that more closely match totals from IRS administrative data. The analysis also accounts for ineligibility for the credit among tax filers who are undocumented immigrants, incorporates anticipated employment incentive effects, and models potential interactions with scheduled increases in the state minimum wage. Overall, the paper provides estimates of potential policy effects of the CalEITC that are directly relevant to policymakers and practitioners, while also advancing knowledge of methods for simulating tax policies that affect low-income populations using government survey data.