Majores Ventures

← All projects

fairness-ledger

Banks' disclosed valuation assumptions, extracted from merger proxies.

Python library, CLI and dataset · Python, NumPy, SciPy, pandas · MIT · v0.1.0 · github.com/Thesimpleex/fairness-ledger

Abstract

Extracts the discount rates, terminal growth rates, exit multiples, implied value ranges and fees that investment banks disclose in U.S. merger proxies, with every number linked to its sentence in the filing. For bankers and valuation professionals benchmarking DCF assumptions, and for researchers.

Difference between two advisors' discount-rate midpoints for the same target, 526 pairs in 356 deals
Figure 1. Difference between two advisors' discount-rate midpoints for the same target, 526 pairs in 356 deals.

1Motivation

Fairness-opinion summaries are the only large public record of how banks actually value companies, but they sit in unstructured 150-page filings. Existing datasets are private hand-collections.

2Method

  1. Find merger proxies with DCF analyses via SEC EDGAR full-text search.
  2. Segment the bank-opinion sections and attribute each analysis to its advisor.
  3. Extract ranges with a tested grammar and keep each value's exact text span.
  4. Check consistency, for example implied perpetuity growth from exit multiples.
  5. Evaluate precision on a frozen annotated set.

3Results

Discount-rate midpoints by year against the 10-year Treasury yield
Figure 2. Discount-rate midpoints by year against the 10-year Treasury yield.
Share of exit-multiple DCFs whose disclosed implied perpetuity growth exceeds nominal growth bounds
Figure 3. Share of exit-multiple DCFs whose disclosed implied perpetuity growth exceeds nominal growth bounds.
Precision and recall of the extractor by field on the frozen holdout set
Figure 4. Precision and recall of the extractor by field on the frozen holdout set.

4Usage

4.1 Fetch filings

Discovers and caches DEFM14A merger proxies from SEC EDGAR.

fairness-ledger fetch --years 2012-2025

4.2 Extract facts

Extracts discount rates, growth rates, multiples, value ranges and fees, each with its source sentence.

fairness-ledger extract

4.3 Inspect a deal

Shows one deal's analyses side by side with their sources.

fairness-ledger show <ticker-or-accession>

4.4 Report

Writes an HTML report over the dataset: 2,967 filings, 2,925 deals, 2012–2025, 29,242 extracted facts.

fairness-ledger report --out report.html

5Installation

pip install git+https://github.com/Thesimpleex/fairness-ledger

Python 3.11 or later. Not on PyPI; installs from GitHub.

6Specification

Runs
Python 3.11+, CLI
Built with
Python, NumPy, SciPy, pandas
License
MIT
Version
0.1.0

Source code, issues and releases: github.com/Thesimpleex/fairness-ledger. License: MIT.