LBO Stack

An applied leveraged-finance toolkit for transparent annual LBO scenario analysis.

LBO Stack turns a declared set of transaction and operating assumptions into a reconciled sources-and-uses schedule, debt and cash roll-forward, covenant view, exit-equity bridge, sponsor returns, sensitivities, and scenario analysis. It is designed to make the mechanics inspectable rather than hide them behind a spreadsheet.

This repository is the practical transaction-modelling project in the portfolio. The separate IFRS 16 LBO Engine focuses on model validation and threshold portability in a synthetic research benchmark.

Highlights

• Senior, mezzanine, bullet, revolver, and simplified IFRS 16 debt tranches.
• Separate cash and PIK interest, mandatory amortisation, and optional cash sweeps.
• Minimum-cash funding, revolver draws, and explicit cash/debt reconciliation.
• Net-debt leverage, cash-interest coverage, and cash-flow coverage.
• Operating and exit sensitivities plus seeded, assumption-driven Monte Carlo paths.
• A single-tier European-style fund waterfall, Streamlit interface, and PDF summary.

Workflow

deal assumptions
  -> sources and uses
  -> annual operating projection
  -> cash / debt / revolver waterfall
  -> covenants and sensitivities
  -> exit equity and sponsor returns
  -> fund-level distribution waterfall

Quick start

Python 3.10-3.12 is exercised in CI.

python -m venv .venv
# Activate .venv using your shell's activation command.
python -m pip install -r requirements.txt
python -m src.modules.orchestrator_advanced

The command reads the included illustrative Accor assumptions and writes output/lbo_analysis.pdf. Generated output is ignored so that a local run does not change the repository.

Launch the interactive model with:

streamlit run src/modules/streamlit_app.py

Model scope

The annual engine includes:

• transaction sources and uses;
• operating projections, capex, working capital, and simplified cash taxes;
• an annual NOL roll-forward;
• minimum-cash funding and revolver capacity;
• priority-ordered amortisation and cash sweeps;
• covenant monitoring and explicit failure states;
• exit-equity and sponsor-return reconciliation;
• operating/exit sensitivities and seeded scenario analysis.

The dashboard calls the same modelling functions used by the command-line workflow; it is a presentation layer rather than a separate engine.

Core reconciliation formulas

Cash

Closing cash
= Opening cash
+ Operating cash generation
+ Operating revolver draw
- Cash-funded mandatory amortisation
- Optional cash sweep

Revolver-funded mandatory amortisation does not pass through ending cash: the draw increases revolver debt and immediately repays the target debt tranche.

Debt

Closing debt
= Opening debt
+ Revolver draws
+ PIK interest
- Actual mandatory amortisation
- Optional cash sweep

Exit equity

Exit equity
= Exit enterprise value
- Sale costs
- Closing debt
+ Closing cash

The opening sponsor cash flow equals the equity cheque from the canonical sources-and-uses schedule. Transaction fees, financing fees, OID, and retained cash therefore enter sponsor returns directly.

Waterfall convention

The current waterfall supports one European-style whole-fund tier:

1. pro-rata return of LP and GP contributed capital;
2. compounded LP preferred return;
3. 100% GP catch-up;
4. residual LP/GP split;
5. optional cashless carry deferral;
6. an end-of-life clawback check.

Management fees are separate investor cash outflows and are not deducted from portfolio distributions. Cash distributions and LP/GP cash flows include the final clawback transfer. In cashless mode, carry is held in a simplified notional reserve until the final model period. Multi-tier waterfalls and hurdle resets deliberately raise NotImplementedError.

Monte Carlo reporting

The Monte Carlo inputs are uncalibrated scenario assumptions. Percentiles include every completed path with a mathematically defined IRR, including underperforming paths. Model failures, insolvencies, and undefined IRRs are reported separately; they are not replaced with an arbitrary return. Unconditional statistics are shown only when every simulated path has a defined IRR.

Validation

python -m pytest -q
python -m pytest -q --cov=src/modules --cov-report=term-missing --cov-fail-under=75
python -m ruff check src tests

Deterministic tests cover sources-and-uses equality, cash and debt roll-forwards, minimum cash, revolver capacity, covenant failures, exit-equity reconciliation, IRR cash flows, and fund-waterfall allocation.

Data and reproducibility

data/accor_assumptions.csv and data/accor_historical_recreated.csv are illustrative reconstructed inputs, not audited company or transaction data. The base workflow and Monte Carlo generator use fixed seeds where randomness is involved. Results still depend on the declared assumptions and installed numerical libraries.

Limitations

• The model is annual, not monthly or quarterly.
• Tax, interest deductibility, and NOL treatment are simplified and not jurisdiction-specific.
• IFRS 16 treatment is assumption-driven, not a complete lease-accounting engine.
• Lease principal follows an assumed amortisation period.
• Monte Carlo distributions are scenarios, not empirically calibrated forecasts.
• The fund waterfall is simplified and does not replace an LPA-specific legal model.
• Multi-tier waterfalls, detailed refinancing, and intra-period liquidity are not supported.

Repository structure

src/modules/lbo_model.py             annual cash and debt engine
src/modules/orchestrator_advanced.py transaction workflow, sensitivities, reports
src/modules/fund_waterfall.py        fund distribution mechanics
src/modules/streamlit_app.py         interactive presentation layer
tests/                               deterministic financial invariants
data/                                illustrative reconstructed inputs
output/                              ignored local reports

License and disclaimer

MIT licensed. This project is for education and scenario analysis; it is not investment advice, a valuation opinion, or a substitute for transaction, accounting, tax, or legal diligence.