workflow-er-idea-screening

name

workflow-er-idea-screening

description

WHAT: Quantitative and thematic investment idea generation — applying Piotroski F-Score, Beneish M-Score, FCF yield, ROIC vs WACC, and revenue growth filters to a broad universe, then overlaying thematic qualitative criteria to produce a ranked list of deep-dive candidates with one-paragraph thesis summaries. WHEN: Invoke when generating new investment ideas for a coverage universe or sector, performing quantitative stock screening, or responding to a mandate such as "find quality growth names in [sector]" or "screen for value with fundamental momentum".

Equity Research: Investment Idea Screening

What this skill covers

A two-stage funnel — quantitative filters followed by thematic overlay — that narrows a broad universe to a ranked shortlist of deep-dive candidates. Each surviving candidate receives a one-paragraph thesis summary and a composite score.

Inputs

• Universe definition: broad index, sector, or custom ticker list
• Thematic mandate (optional): sector focus, factor tilt (quality, value, growth, momentum), or ESG constraint
• Screening stringency: standard (Piotroski >= 7, Beneish < -1.78) or custom thresholds

Workflow

Step 1 — Define the universe

Start with the broadest feasible universe. Options:

• A sector via fmp_stock_screener (filter by sector, market cap floor)
• An index via fmp_index_constituents
• A custom ticker list provided by the user

Record the starting universe count.

Stage 1 — Quantitative screens

Apply hard filters sequentially. Record pass/fail count at each gate.

Gate 1: Fundamental quality — Piotroski F-Score

Call piotroski_fscore for each company. Pass criterion: score >= 7 (strong fundamentals on profitability, leverage, and operating efficiency dimensions). Fail: score < 7.

Gate 2: Earnings integrity — Beneish M-Score

Call beneish_mscore for each passing company. Pass criterion: M-Score < -1.78 (no manipulation signal). Fail: M-Score >= -1.78 (potential manipulation — exclude or flag for extra scrutiny).

Gate 3: Value creation — ROIC vs WACC

Call fmp_key_metrics for ROIC. Call wacc_calculator for WACC estimate. Pass criterion: ROIC > WACC (company destroys or creates value). Fail: ROIC <= WACC.

Gate 4: Cash generation — FCF yield and revenue growth

Call fmp_key_metrics for FCF yield. Call fmp_financial_ratios for revenue growth vs sector median. Pass criteria: FCF yield > 5% AND revenue growth > sector median. Fail: either criterion misses.

Record the universe count after Stage 1: [N] companies passed all four gates.

Stage 2 — Thematic overlay (qualitative filters)

For each Stage 1 survivor, apply the thematic mandate:

1. Sector trends: is the company operating in an industry with secular tailwinds (not just cyclical recovery)?
2. TAM growth: is the addressable market expanding >5% annually? (use industry sources or World Bank data)
3. Regulatory tailwinds: are pending policy changes a net positive for this company's model?
4. Management alignment: recent insider buying or low turnover at senior level (check fmp_insider_latest)

Each criterion scores 0 or 1. Minimum thematic score: 2 of 4 to advance to the shortlist.

Stage 3 — Composite ranking

Compute a composite score for each shortlisted company:

composite = (piotroski_fscore / 9 × 0.35) + (fcf_yield_normalised × 0.30) + (roic_spread_normalised × 0.20) + (thematic_score / 4 × 0.15)

Rank all shortlisted companies by composite score. Top 5-10 become deep-dive candidates.

Step 4 — One-paragraph thesis per idea

For each deep-dive candidate, write a one-paragraph thesis:

• Ticker, market cap (from fmp_market_cap), sector
• Key fundamental strengths (from Piotroski sub-scores)
• Valuation hook (FCF yield, ROIC spread)
• Thematic alignment
• Single-sentence investment thesis

Output format

1. Funnel table — universe start count and pass count at each gate
2. Shortlist table — ranked by composite score with sub-scores per dimension
3. Thesis summaries — one paragraph per deep-dive candidate
4. Next steps — which candidates are recommended for full workflow-er-initiating-coverage

Quality gates

• All Piotroski scores from piotroski_fscore — not manual calculation
• All Beneish scores from beneish_mscore
• WACC from wacc_calculator; ROIC from fmp_key_metrics
• Composite score formula applied consistently to all shortlisted candidates
• Shortlist limited to 5-10 candidates — not a padded list

Related skills

• workflow-er-initiating-coverage — full coverage initiation for shortlisted candidates
• workflow-er-sector-overview — sector context that informs thematic criteria in Stage 2
• workflow-er-thesis-tracker — thesis framework applied once a candidate advances to coverage