A quantitative, rules-based approach to discovering undervalued Indian mid-cap and small-cap companies
Most institutional investors focus on the top 200 companies by market capitalization. That leaves thousands of fundamentally sound businesses in the mid-cap and small-cap universe that trade at valuations far below their intrinsic worth — simply because nobody is paying attention.
JIP Horizon India uses a 5-layer signal system with 3 score modifiers to systematically identify these companies before institutional coverage begins. The 5 layers capture different dimensions of emerging quality — from promoter conviction and operating leverage inflection to corporate intelligence and policy tailwinds. Three modifiers then adjust the composite score: a valuation gate penalizes overpriced stocks, smart money tracking confirms institutional interest, and a degradation monitor flags deteriorating fundamentals. When multiple layers fire simultaneously, it creates a convergence signal that has historically been the strongest predictor of upcoming re-rating events.
Five independent layers feed into a weighted composite score
Layer 1: Promoter Insider Trading
ActiveLayer 2: Operating Leverage Inflection
ActiveLayer 3: Corporate Intelligence
ActiveLayer 4: Policy Tailwind
ActiveLayer 5: Quality Emergence
ActiveWeighted signals converge into a single conviction score per company
When promoters invest their own money into their company through open market purchases, warrant allotments, or creeping acquisitions, it signals genuine confidence in the business’s future. This layer tracks all insider transactions filed with stock exchanges and scores them based on the type of transaction, the amount relative to company market cap, and the pattern of purchases over time. Conversely, pledge increases and open market sells are treated as negative signals that reduce the score.
Open Market Buy
Promoter purchases shares from the open market at prevailing prices — the strongest signal of conviction
Warrant Allotment
Promoter converts warrants to equity, committing additional capital to the company
Creeping Acquisition
Gradual increase in promoter stake over time, signaling long-term accumulation intent
Preferential Allotment
Company issues shares to promoters at a fixed price, indicating capital infusion plans
Pledge Decrease
Promoter reduces pledged shares, strengthening the balance sheet and reducing risk
Pledge Increase (Negative)
Promoter pledges more shares as collateral — a warning signal that reduces the score
Open Market Sell (Negative)
Promoter sells shares on the open market — treated as a negative signal
ESOP Exercise
Key management personnel exercising stock options, a minor positive signal
Each insider transaction is scored based on its type (buy types score higher than sells), the transaction value relative to market cap (larger relative amounts score higher), the person category (promoter buys weigh more than KMP buys), and recency (transactions in the last 90 days are weighted more heavily). The layer produces a 0–100 signal score that is multiplied by the 30% layer weight in the final composite calculation.
This layer identifies companies at the inflection point where revenue growth begins to outpace cost growth, leading to nonlinear profit expansion. It analyzes financial statements to detect transitions from debt-laden balance sheets to net cash positions, structural margin expansion, improving capital efficiency, and accelerating revenue growth. These signals often precede significant re-ratings as the market recognizes improved business quality.
Debt to Net Cash Transition
Company moves from a net debt position to net cash — a powerful de-risking signal
EBITDA Margin Expansion
Sustained expansion in operating margins indicating improved pricing power or cost control
ROCE Inflection
Return on capital employed crosses above cost of capital, creating shareholder value
Revenue Hypergrowth
Revenue CAGR exceeding sector peers by a significant margin over 2–3 year windows
Receivables Compression
Declining debtor days indicating improved cash collection and business quality
Equity Value Creation
Growing net worth driven by retained earnings rather than equity dilution
Financial data from the last 3–5 years is analyzed across 6 sub-signals. Each sub-signal uses specific thresholds — for example, EBITDA margin must expand by at least 300 basis points over 3 years to qualify as a structural expansion. Sub-signals are combined into a 0–100 layer score. The debt-to-cash transition and margin expansion signals carry the highest internal weight since they most reliably predict future re-rating events.
NSE corporate filings — board outcomes, investor presentations, credit ratings, auditor changes, and management appointments — are automatically fetched, categorized, and scored using Python rules + Claude AI analysis for high-priority filings. This layer captures forward-looking signals from official disclosures that rarely make it into consensus estimates.
Management Sentiment Shift
Detects quarter-over-quarter changes in management tone from cautious to bullish
Capex Announcement
Identifies new capital expenditure plans that signal capacity expansion or modernization
Order Book Growth
Extracts order book or pipeline commentary indicating future revenue visibility
New Client/Geography Win
Detects mentions of new customer additions, export markets, or geography expansion
Guidance Upgrade
Identifies explicit upward revisions to revenue, margin, or growth guidance
Filings are classified into three buckets: earnings & strategy (board outcomes, investor presentations), capital actions (acquisitions, buybacks, credit ratings), and governance (auditor changes, management appointments). Each bucket generates a sub-score based on the signal content. The top filings are analyzed by Claude AI for hidden insights. The three sub-scores are combined into a 0–100 layer score with a 25% weight in the composite calculation.
Government policy decisions — PLI schemes, infrastructure spending, regulatory changes, budget allocations — create multi-year tailwinds for specific sectors. This layer maps policy announcements to sector-level beneficiaries and identifies companies positioned to benefit disproportionately. Small/mid-cap companies in policy-favored sectors often see sustained re-rating as order flows materialize.
PLI Scheme Beneficiary
Company belongs to a sector covered by Production Linked Incentive schemes
Infrastructure Spending
Direct or indirect beneficiary of government infrastructure investment programs
Regulatory Tailwind
Positive regulatory changes that expand addressable market or reduce compliance costs
A curated registry maps active government policies (PLI schemes, infrastructure programs, defence indigenisation, green energy mandates) to beneficiary sectors and sub-sectors. Each company is matched against relevant policies based on their sector and industry classification. Companies with direct exposure to multiple active policies score higher. The layer produces a 0–100 score with a 10% weight in the composite calculation.
Some of the best multi-baggers start as poorly-governed companies that undergo a quality transformation — hiring professional management, improving disclosures, reducing related-party transactions, or initiating dividend payments for the first time. This layer tracks early signs of improving corporate governance and financial discipline that institutional investors look for before initiating positions.
Board Independence Improvement
Addition of independent directors or formation of proper board committees
Disclosure Quality
Improved quarterly reporting, investor presentations, or annual report transparency
Dividend Initiation
First-time dividend payment signaling confidence in sustained profitability
Related-Party Cleanup
Reduction in related-party transactions indicating improved governance standards
Financial data from the last 3–5 years is analyzed across 6 quality signals: ROE breakout (crossing 15%), ROCE consistency (>15% for 2+ years), margin expansion streak (3+ years), deleveraging (D/E falling below 0.5), working capital tightening (declining debtor+inventory days), and earnings quality (margin expansion with >10% revenue CAGR). When 4+ signals fire, a 1.15x quality multiplier is applied. The 5% weight is intentionally small but acts as a quality filter.
After the base composite is calculated from the 5 layers, three modifiers adjust the final score. These are not weighted layers — they act as gates, bonuses, and penalties.
A stock with great signals but a 60x PE should score lower than the same signals at 10x PE. The valuation gate scores each company across 5 dimensions: PE vs sector median (25 pts), absolute PE (25 pts), price-to-book (15 pts), EV/EBITDA (20 pts), and 52-week price position (15 pts). The total score maps to a valuation zone that multiplies the base composite.
Deep Value
1.15x
Cheap
1.08x
Fair
1.00x
Expensive
0.90x
Overvalued
0.75x
Tracks institutional and superstar investor activity to confirm or deny the thesis. Monitors 25 known Indian superstar investors (Ashish Kacholia, Vijay Kedia, Dolly Khanna, etc.), mutual fund accumulation patterns, FII flows, and bulk/block deal data from NSE.
Superstar new entry: +10 | Superstar increased: +6 | Superstar exited: -8
MF accumulation >1% QoQ: +6 | MF exit >1%: -4 | FII accumulation: +4
Institutional bulk buy: +5 | Institutional bulk sell: -5
The system only identifies buys — but exit discipline matters. The degradation monitor scans all existing tables for 10 red flags. The penalty (up to -30 points) is subtracted directly from the composite score.
Net insider selling > buying (90d): -8 | Pledge % increasing: -6
EBITDA margin declining 2+ quarters: -6 | Revenue declining YoY: -5
Auditor change (mid-term): -8 | CFO/CEO resignation: -6
Credit rating downgrade: -5 | Price >15% below 200-DMA: -3
base = weighted_average(promoter:30%, OL:30%, corp_intel:25%, policy:10%, quality:5%)
base = apply_convergence_bonus(base)
final = clamp(base × valuation_mult + smart_money + degradation, 0, 100)
Convergence bonus: +6% for 2 layers firing ≥40, +10% for 3 layers, +15% for 4+ layers
Every company is classified into one of four tiers based on its composite score
Multiple signals converging — highest probability of re-rating
Strong quantitative signals — AI thesis generated
Moderate signals — worth monitoring for escalation
Early signals detected — added to watchlist
AI Investment Thesis
An AI-generated investment thesis (powered by Claude) is produced only for High and Highest conviction tier companies. This keeps AI costs controlled at approximately $0.005 per thesis while focusing analysis effort where it matters most.
Convergence Bonus
When two or more signal layers fire simultaneously, the composite score receives a 6% bonus. Three converging layers earn a 10% bonus, and four or more layers earn 15%. This rewards companies showing multi-dimensional improvement \u2014 the strongest indicator of an upcoming re-rating event.
All data is sourced from public filings and free/open data providers
NSE / BSE Corporate Filings
Official stock exchange filings including insider trading disclosures (SAST regulations), corporate announcements, and quarterly results
Insiderscreener.com
Aggregated insider transaction data parsed and normalized for pattern detection across promoters, directors, and KMPs
Screener.in
Financial statements data including 5-year P&L, balance sheet, and cash flow statements for operating leverage signal computation
Yahoo Finance
Historical price data and basic market metrics via yfinance with .NS (NSE) and .BO (BSE) suffixes
NSE Bulk & Block Deals
Daily bulk and block deal data from NSE, matched against 25 superstar investors and institutional keywords for smart money tracking
BSE Shareholding Patterns
Quarterly shareholding breakdown from BSE — promoter, FII, DII, mutual fund, and public holdings with quarter-over-quarter delta analysis
Disclaimer
This tool is for educational and research purposes only. It does not constitute investment advice. All data is sourced from public filings. Users should conduct their own due diligence and consult a SEBI-registered investment advisor before making investment decisions. Past performance of any screening methodology does not guarantee future results. JIP Horizon India is not a SEBI-registered research analyst or investment advisor.