Corporate Finance
Decision Lab
Annual-report data, cash-generation analysis and a model of the capital required to fund growth. Microsoft, Target and Infosys, FY 2023–FY 2025.
Cash generation and investment
Inspect the reported financials before choosing a future growth assumption. The source spread preserves fiscal dates, consolidation and accounting definitions.
Growth funding and operating-policy value
New investment is paid before the next year’s incremental operating profit. The first payment is immediate. Change the assumptions and inspect the same ten-year schedule.
WACC and terminal-growth sensitivity
Reported EBIT’s expense and lease basis is retained. Growth, marginal ROIC, WACC and tax are illustrative assumptions. Values precede financing, nonoperating and equity adjustments. Historical CFO-minus-investment is a separate accounting proxy. These outputs are not market-calibrated enterprise values or target share prices.
Trace the numbers and abstentions
Both methods answered 33 of 36 final financial targets correctly, with three justified abstentions. The methods tie on the final fiscal-year cases. Development improvements are reported separately.
| FY 2025 report | Literal correct / targets | Strict correct / targets | Direct unavailable |
|---|---|---|---|
| Microsoft | 11 /12 | 11 /12 | Pure D&A |
| Target | 11 /12 | 11 /12 | Direct total liabilities |
| Infosys | 11 /12 | 11 /12 | Pure cash PPE |
Original-file identity
The final check contains three retrospective report clusters. Target FY 2025 cover highlights were incidentally displayed during source review; no blind or unseen-issuer claim is made. References are workflow transcriptions, not external human review. The three unsupported-control types are current share price, customer count and undisclosed future revenue. Their 9/9 final rejections per method are structural API-contract checks, not a language-model hallucination benchmark.
Reproduce and inspect the complete ledger
Six development reports: strict 66/72 financial targets correct versus literal 64/72, with six direct-unavailable targets retained. The two baseline errors selected prior-year Infosys D&A cells. Correct consolidated annual MD&A duplicates count as valid baseline provenance.
Code, protocol and reproduction instructions · Complete error ledger
Explore the seven-study portfolio
Credit and insurance risk, treasury, corporate finance, causal decisions, demand and Indian inflation research.