modeling-tax-attribute-preservation

name	modeling-tax-attribute-preservation
language	en
description	Analyzes NOL and tax attribute preservation strategies under Section 382 limitations in ownership change scenarios. Use when modeling tax attribute preservation, analyzing 382 limitations, or structuring ownership change thresholds.
tags	["modeling","distressed-and-restructuring","tax"]
metadata	{"author":"casemark","practice_areas":["Restructuring","Distressed Investing","Turnaround"],"document_types":["Financial Model"],"skill_modes":["Modeling","Forecasting"]}

Modeling Tax Attribute Preservation

When To Use

• Evaluating whether a proposed transaction (debt-for-equity swap, rights offering, plan of reorganization) triggers an "ownership change" under IRC §382
• Quantifying the annual §382 limitation on post-change NOL and credit utilization
• Comparing restructuring alternatives by their net tax-attribute value after §382 haircuts
• Advising on shareholder threshold management to avoid or defer an ownership change
• Modeling the incremental value of a §382(l)(5) bankruptcy exception versus §382(l)(6) election

Inputs To Gather

• NOL and credit schedule: Federal and state NOL carryforward balances by vintage year, expiration dates, and any built-in loss/gain positions (NUBIL/NUBIG) [VERIFY state-specific NOL rules and carryforward periods]
• Equity capitalization table: All 5%+ shareholders (direct and constructive under §382 attribution rules), option pools, convertible instruments, and warrant holders with strike prices
• Transaction terms: Proposed equity issuance amounts, conversion ratios, exchange mechanics, and closing timeline
• Company value inputs: Current equity market cap or agreed enterprise value, long-term tax-exempt rate published by IRS for the month of the ownership change [VERIFY current applicable federal rate]
• Historical ownership shifts: Rolling 3-year testing period log of 5%+ shareholder percentage-point increases
• Built-in gain/loss data: Fair market value versus tax basis of assets at the change date to determine NUBIG or NUBIL status under the §1374 approach or the §338 approach [VERIFY which recognition period method applies]

Workflow

1. Map the ownership change test

   • Identify all 5%+ shareholders under the constructive ownership and aggregation rules of Treas. Reg. §1.382-2T
   • Build a rolling 3-year testing-period schedule tracking each 5%+ shareholder's lowest and current percentage
   • Sum percentage-point increases; flag if aggregate exceeds or approaches the 50-point threshold

2. Calculate the §382 annual limitation

   • Multiply the loss corporation's equity value immediately before the ownership change by the applicable long-term tax-exempt rate
   • Adjust for any recognized built-in gains during the 5-year recognition period (increases the limitation) or built-in losses (subject to the limitation)
   • Apply the continuity-of-business-enterprise (COBE) requirement — if not met for 2 years post-change, limitation drops to zero [VERIFY COBE fact pattern]

3. Model the bankruptcy exceptions

   • §382(l)(5): If the loss corporation is in a Title 11 case and historic shareholders/qualified creditors own ≥50% post-change, no annual limitation applies — but NOLs are reduced by interest deductions on converted debt for the 3 prior tax years, and a second ownership change within 2 years zeroes remaining NOLs
   • §382(l)(6): Standard limitation applies but equity value is measured post-restructuring (typically higher), which may yield a larger annual cap
   • Compare net present value of usable NOLs under each path over the carryforward horizon

4. Build the NOL utilization forecast

   • Layer the §382 cap over projected taxable income by year
   • Apply vintage-year expiration dates — oldest NOLs used first (FIFO)
   • Incorporate state-level §382-equivalent limitations where applicable [VERIFY state conformity — e.g., California limits NOL usage independently]
   • Discount the tax savings at an appropriate after-tax rate to derive the net present value of preserved attributes

5. Run scenario and sensitivity analysis

   • Vary equity value ±10–20% to show limitation sensitivity
   • Model alternative transaction structures (partial conversion, staged closings) to test whether ownership change can be avoided
   • Test threshold management strategies: transfer restrictions, poison pills, §382 rights plans, or shareholder caps
   • Compare outcomes with and without NUBIG/NUBIL adjustments

6. Compile deliverable

   • Summary table: NOL balance, annual §382 cap, usable NOLs by year, NPV of tax savings
   • Scenario comparison matrix (l)(5) vs. (l)(6) vs. no ownership change
   • Key assumptions log with [VERIFY] flags for all rate, valuation, and jurisdictional inputs

Output

• §382 Limitation Model (Excel-ready table or structured output): annual cap, cumulative usage, expiration waste, and NPV of tax shield under base and alternative cases
• Ownership Change Testing Schedule: 3-year rolling shareholder log showing whether the 50-point threshold is breached
• Scenario Comparison Summary: Side-by-side of (l)(5), (l)(6), and non-change structures with NPV differential
• Assumption and Risk Register: All inputs, sources, and [VERIFY] items requiring confirmation

Quality Checks

• Confirm that the long-term tax-exempt rate used matches the IRS-published rate for the correct month [VERIFY]
• Validate that constructive ownership attribution (family, entity, option) rules are applied consistently across the shareholder schedule
• Ensure NOL vintage ordering follows FIFO and that expiration dates align with the applicable carryforward period (20-year pre-2018, indefinite post-TCJA with 80% taxable income cap) [VERIFY which NOL vintage rules apply to each tranche]
• Cross-check that NUBIG/NUBIL classification uses the correct measurement method and that recognition-period items are properly included or excluded
• Verify that the model handles a second ownership change correctly — §382 cap is the lesser of the new and old limitations
• Flag any state where §382 conformity is partial or nonexistent so state-level modeling is adjusted accordingly