BHARATEQUITY
Loading…
The methodManthaan · editorial doctrine

A research letter, not a dashboard.Synthesis over summary.

very name in the ~5,000-stock universe is read the way a portfolio committee would read it — through four PhD-level analytical lenses and a forensic engine that reads the annual report so you don’t have to. What we publish is a view, argued from the filings — not a score, asserted. The numbers are the same everywhere; the work is in the reading.

01The lenses

Read like a portfolio committee

Before a single flag is raised, every name is examined through four frameworks drawn from the academic canon — each asking a different, hard question of the same financials.

Penman

Quality of earnings/Accounting analysis

Strips accounting noise to ask whether reported profit is backed by operating reality — separating returns earned from the business from returns manufactured by leverage.

Merton

Distance to default/Structural credit

Treats equity as an option on the firm's assets to measure how many standard deviations of value sit between the company and insolvency.

Mauboussin

Market-implied expectations/Expectations investing

Reverse-engineers the growth and margin the current price already assumes, then asks whether the business can plausibly clear that bar.

Ohlson

Bankruptcy O-score/Distress probability

A probabilistic read on financial distress built from size, leverage, liquidity and profitability — a sober cross-check on the optimists.

The market hands everyone the same numbers. The edge is not in having them — it is in reading the footnotes nobody else finishes.

Why the forensic engine exists
02The forensic engine

One filing in.
Six findings out.

A forensic engine that reads the annual report so you don’t have to — six passes per filing, every flag cited to its source.

Live demo · No sign-in
Run all 6 on DIXON
Input
Annual report
CARO annexure
Quarterly results
Shareholding pattern
Board filings
10y depth · 5,000+ listings
Output
38COMPOSITE / 100
Auditor Worry Score
DIXON · Watch band · peer pctl 61
6
Forensic engines
10y
Filing depth
60+
Forensic indicators tracked
Daily
Refresh cadence
03The pipeline
01
Ingest

10y of filings, every NSE + BSE listing

5,000+ cos
02
Normalise

Ind AS aligned, restatements reconciled

30+ metrics
03
Run engines

Six forensic passes per filing

6 engines
04
Score

Weighted composite, peer-percentiled

0–100 scale
05
Surface

Every flag cited to its source page

Daily refresh
04Engine deep dive

CARO Decoder

Auditor annexure flags
Reads
CARO 2020 annexureAuditor's reportIFC opinion
Signals
  • Loan defaults & evergreening clauses
  • Statutory dues in arrears
  • Fund-diversion qualifications
  • Fraud reporting under §143(12)
Flag incidence · market-wideper 100 cos
FY20
FY21
FY22
FY23
FY24
FY25
Composite weight22%
Live flag · DIXON
Clause (ix)(d) — short-term funds in long-term use
Engine score 31 / 100
05Scoring

The Worry Score

Five engines, weighted, peer-percentiled. One number per company.

38COMPOSITE / 100
Clean
0–25
Watch
25–50
Elevated
50–75
Severe
75–100

Engine weights

Re-calibrated annually against realised governance events.

CARO Decoder22%
Related-Party20%
Contingent Liability15%
Governance18%
Earnings Quality15%
+ Trend & peer-percentile overlay10%
Editorial north star

We are a research library, not an advisor. Everything here is descriptive analysis of public filings — built for your own thinking, never a recommendation to buy, sell, or hold.

Every flag cites its source filing and page.