The Tax-Calculator is capable of dealing with large amounts of input
data, so long as each filing unit has variable names that are in the
USABLE_READ_VARS set in the records.py
file.
More details about the data input requirement can be found here in the
DATAPREP.md
file.
Thanks to such capability, we are able to test the current tax logic with as many "extreme" units as we want, and thus are able to find a couple tax units that are considered counter-intuitive against common sense. For instance, if allowed by the tax law, some unit should get better-off in terms of individual income tax liability when that person choose to have no deduction, or zero deduction amount.
This document stores units that would yield counter-intuitive results we have found, where all units listed are artificially made-up. During make-up process, we try to use as few variables as possible so that our units will end up being rather simple. Adding more (valid) variables to those units, however, usually will not affect the counter-intuitive case we are trying to explain.
Generally speaking, we take it for granted that filing units will end up being better-off when they choose the higher deduction amount among either standard deduction or itemized deduction. The unit we described in this section, however, should yield a lower tax liability when he/she choose to have a lower deduction amount.
The reason that accounts for this phenomenon is that: when choosing to have no deduction, although this unit will pay higher regular tax, or '_taxbc', comparing to regular scenario that has deduction amount, his/her alternative minimum tax (AMT) liability will end up being much lower. And thus this yields a comparatively lower individual income tax liability. To be more specific, choosing no deduction amount will result in higher taxable income, this higher income will come into play in the AMT calculation (which is essentially introduced from Schedule D Tax Worksheet for form 1040) as an offsetting force, that eventually diminishes the AMT liability. That said, in these two cases, the difference for AMT part is larger than the difference resulting from regular tax, in terms of absolute value, and thus this unit is able to obtain lower individual income tax liability when no deduction is being applied.
In order to replicate this phenomenon, the following unit needs to be fed through the Tax-Calculator with 2013 tax logic, without any extrapolation/blowup for the data input. Also, to allow units having zero deduction, you shall suppress all the itemized deduction items. To do all these, you might find PR #749 helpful.
| Variable | Value |
|---|---|
| FLPDYR | 2013 |
| MARS | 2 |
| e00300 | 14300 |
| e00900 | 175000 |
| e02000 | -1795000 |
| e03270 | 20000 |
| e18400 | 40000 |
| e18500 | 2700 |
| e19200 | 15700 |
| e19700 | 3200 |
| p23250 | 4580000 |
| e24515 | 2090000 |
The Tax-Calculator offers the approximation of units' marginal tax rates (MTRs) using finite difference approach, where a small increase (a penny) will be added to some specified income type. For more details, you might find the code lines here helpful.
During the course of testing the Tax-Calculator, we found that, under some circumstances, filing units could end up having rather high (over 50%) MTRs. This means that, surprisingly, more than 50% of those units' extra income will contribute to their tax liabilities. Based on our current findings, one possible reason that would result in such high MTR situation is because of the foreign tax credit.
When some unit has foreign tax credit (FTC), such credit can possibly wipe out the income tax liability in the AMT calculation, and thus offset the AMT liability. Filing units with enough FTC that totally wipes out the income tax liability in the AMT calculation will need to take care of more tax burden when given some extra income. More explicitly, we use the following table to generalize such situation:
| Base | Base, Added a penny | |
|---|---|---|
| Adjusted Gross Income | AGI | AGI + 0.01 |
| Standard Deduction | STD | STD |
| Taxable Income | AGI - STD | AGI - STD +0.01 |
| Income Tax | intax | intax + delta |
| Foreign Tax Credit | FTC | FTC |
| AMTI | AGI | AGI + 0.01 |
| Tentative Minimum Tax | TMT | TMT + less_delta |
| AMT | TMT | TMT + less_delta |
| Tax Before Credits | TMT + intax | TMT + less_delta + intax + delta |
Notice that, in this example, we have intax + delta < FTC so that
FTC would wipe out the income tax, which should have neutralized the
tentative minimum tax amount if the inequality does not hold. Also,
delta > less_delta > 0 holds since, in brief, the less_delta is
generated by more iterations (deteriorations) in the tax form than
delta, and thus is less significant.
In order to replicate this situation, the following unit needs to be fed through the Tax-Calculator with 2017 tax logic along with the reform that doubles the standard deduction amount while, still, without any extrapolation/blowup for the data input. By the default MTR calculations in the Tax-Calculator, you shall get a MTR of 67.60%. You might find the discussion in issue #555 helpful.
| Variable | Value |
|---|---|
| FLPDYR | 2017 |
| MARS | 2 |
| e02000 | 648675 |
| e00600 | 24759 |
| e00650 | 24759 |
| e07300 | 199643 |
| f6251 | 1 |
| XTOT | 2 |