EITC: Observing weird things

Hello there,

I’m currently working with IPUMS USA data and there are some observations that I don’t understand regarding EITC (I am using spmeitc for the EITC amount, inctot for the total income of household, famsize for the family size, and nchild for the number of children):

  1. on the amount of EITC: I find individuals with an abnormally high EITC (e.g $12,862, while the maximum amount is around 7k unless I’m mistaken)
  2. on the wealth of those who get EITC: I find some individuals who, ex ante, shouldn’t be eligible, but still get EITC (e.g a household with 3 people including a kid, with total income $830,000, and getting $480 EITC).

Would anybody be able to explain those two weird things? Am I interpreting one of the variables in an incorrect way?

Many thanks for your help!

I replied to your email on this topic and am posting here in case others have similar questions.

The tax-related variables in the American Community Survey are not reported directly by respondents; the ACS does not collect any information on a household’s tax obligations or tax filing. These variables, including EITC and SPMEITC, are imputed using NBER’s TAXSIM tool. The tool estimates federal and state income tax liabilities from survey data, using variables including state of residence, year of filing, marital status, ages of the primary taxpayer and spouse, number of dependents, and wages and income to calculate the estimates. This Census Bureau working paper details how the supplemental poverty measure variables (such as SPMEITC) are calculated for the ACS. Imputation of tax information, including estimates of the value and receipt of tax credits, inherently includes some amount of error. There are individuals with imputed EITC credit amounts that are not possible or plausible given their income and/or the EITC limits. However, these cases are few, and users can impose restrictions on which EITC values they deem valid.