skills/economic-impact-report/references/workflow.md

# Economic Impact Workflow

Use this reference for full-form reports and high-stakes public-equity market-event work.

1. Define the event exactly and separate it from surrounding noise.
2. Identify what is genuinely new relative to baseline or consensus.
3. Name the public-equity decision affected: issuer diligence, sector view, earnings setup, valuation support, position sizing, hedge/watchlist, or portfolio review.
4. Explain why now: immediate repricing, medium-term earnings/policy effect, structural shift, or likely fade.
5. Build the first-pass transmission map: event -> channel -> issuer/sector driver -> financial line item -> stock/portfolio implication.
6. Write hypotheses before over-collecting facts. For each key channel, define hypothesis, first equity-market variable, confirmers, and falsifiers.
7. Gather evidence to test and rank the hypotheses. Prefer primary material and evidence that answers the highest-value unresolved questions.
8. Rank impacts by sign, magnitude, timing, confidence, directness, and priced-in status.
9. Cover materially relevant equity domains: listed issuers, peer groups, sectors, estimate paths, valuation/multiple support, factor/benchmark/ETF exposure, positioning, and portfolio action.
10. Use macro, rates, FX, commodities, options, futures, and credit signals only as public-equity inputs. If the best work product is a non-equity trade or credit-security view, route out.
11. Identify what is priced and what is not across spot price, consensus, volatility, financing, supply-chain/regulatory precedent, ownership, passive flows, short interest, and crowding.
12. Write the strongest counterargument and explain whether it is stronger, weaker, delayed, already priced, or unresolved.
13. Build base, upside, and downside cases with approximate probabilities where supportable.
14. End with monitoring signals, falsifiers, and what would change portfolio action.
15. Run the checker in `--mode draft` while iterating, then `--mode delivery` before final delivery. Do not deliver if source/freshness posture fails for missing, stale, or weak evidence.

## No Portfolio / Watchlist Fallback

When the user provides no holdings, thesis, watchlist, benchmark, or active weights, build a general exposure screen instead of pretending to know the book. The fallback must:

- infer the affected industries, countries/currencies, commodities, public companies, relevant private companies, supplier/customer groups, and second-order peer groups;
- label portfolio-specific action as unavailable without holdings, active weights, mandate, and risk limits;
- use `watchlist`, `wait for proof`, `pass`, `research queue`, or `re-underwrite candidate` rather than add/trim/exit recommendations;
- identify the next data needed to convert the general screen into portfolio action;
- still end every channel in issuer, sector, earnings/estimate, valuation/multiple, positioning/flow, and next-workflow implications.

## Common Failure Modes

Reject or revise outputs that:

- summarize an article instead of analyzing economic impact
- list impacted names without explaining transmission channels
- confuse headline intensity with economic importance
- ignore what was expected or priced
- treat rates, FX, options, futures, commodities, or credit as standalone outputs instead of public-equity transmission inputs
- over-focus on the named company and miss suppliers, customers, competitors, or financing channels
- collapse time horizons into one blurry conclusion
- present speculative higher-order effects with the same confidence as direct effects
- use qualitative language where a rough range is possible
- omit issuer/sector, earnings, valuation, positioning, or portfolio implications
- omit what would falsify the thesis