Closed codykallen closed 4 years ago
PR #118 revised the modeling of the PYMTC. Next up should be revising the modeling of the AMT revenue slightly to make it more easily adaptable to splitting by industry.
I'm not planning to make further changes to the AMT model, other than splitting by industry.
The current approach to the corporate AMT relies on a number of different estimates, as well as a steady-state assumption for the baseline. These estimates include values about transition rates between AMT liability status from the 1990s and accumulated PYMTC carryforwards. These will not be applicable if we split C corporations by industry. Consequently, the model should be modified to estimate these parameters based on observed data/trends, rather than to pull external estimates. The model also relies on assumptions about taxable income, which may not be entirely accurate for large changes to definitions of taxable income. However, given that the corporate AMT has been repealed, it is unlikely to matter much.
The TCJA made a handful of changes relevant to corporate AMT model. In addition to repealing the AMT, it made the PYMTC refundable for 2018-2021, and repealed it thereafter.
If we're improving the model, we should use data from https://www.irs.gov/statistics/soi-tax-stats-table-23-returns-of-active-corporations, which provides data on the corporate AMT and corporations subject to it.