Use with Survey Data - Example 7
Of course the SOI is not the only source of income data and taxcalc can calculate tax liabilities from any source, provided variables are named according SOI conventions. It isn't necessary to have all the variables - taxcalc will assume missing data items are zero and continue calculating. In this respect it duplicates the procedures of the IRS in checking actual tax returns.
In this example we supply only 4 variables, from the SCF of 2001. With only wages, marital status and year used in this example the result is of low accuracy, but it makes the example a bit clearer. The more variables from the full list the more accurate the calculated result.
Note that "importance weights" are not documented, but seem to be appropriate for calculating sums in a probability sample such as the SCF. I don't know if they would be correct for higher moments.
Note that the 2001 survey provides data for 2000, and that the SCF does include most of the variables needed for a fairly accurate tax calculation although we only use 4 in this example. In the SCF x8023 specifies marital status, x4112 wages, and x42001 the record weight. With only 4 variables the result is of low accuracy, but it makes the example a small and clear. The more variables from the full list of variables actually used in the calculation the more accurate the final result will be. In addition to the wages and filing status variables used in the example, you might want to assign values to:
- xocah
- Number of children at home.
- agep
- 1 for primary taxpayer over 65, 0 otherwise
- ages
- 1 for secondary taxpayer over 65, 0 otherwise
- e00200
- wages for primary and secondary taxpayers together
- e32890
- wages for secondary taxpayer only
- e00650
- qualified dividends
- e00300
- interest, non-qualified dividends, schedule c,e and f, etc.
- e01700
- taxable pensions
- e02400
- gross social security income
- e18500
- Real Estate and other state and local taxes
- e32800
- child care expenses
- e02300
- unemployment insurance
- n24
- number of children under 17
- e19300
- mortgage interest and other itemized deductions that are not a preference for the AMT
- e22250
- net short term capital gains and losses
- e23250
- net long term capital gains and losses
among others. TAXCALC will assign zero to any variables you leave missing.
Note that (for example) in the SOI database e00300 is only interest, but here we suggest lumping most other categories of income in that field, because they will all receive the same tax treatment. xocah is "children at home" for SOI, but you could put the total count of dependents here.