Skip to content
Heterodata An Arcanum Research project Volcker
Volcker

Volcker

Comparative empirical testing of theories of banking

Arcanum Research — an agent-based research architecture

161years of US banking data
11competing theories
96atomic predictions
12formal models
11studies

What this project is

A comparative empirical test of ~11 theories of banking against 161 years of US banking data plus an international panel, run through Anwar Shaikh's real-competition framework. The method is a structured theory-prediction matrix: 11 schools, 35 thinkers, 96 atomic predictions, 12 formal models, and pre-registered horse races estimated on one harmonised panel so the schools compete on the same data and estimators.

The motivating fact

SVB 2023 met Basel III book-capital rules on the day it failed; the loss was driven by uninsured wholesale funding and held-to-maturity duration mismatch — the Minsky 'position-making' mechanism the dominant prudential framework leaves unnamed. Headline expectation: Minsky position-making (T02.P04) dominates Admati-Hellwig book capital (T06.P01) in failure prediction.

Source: Volcker project README

Headline findings

Explore

Browse all 11 studies with their evidence rollups, or open the data explorer to chart real derived outputs from the bank-failure taxonomy and the Correia replication's series dictionary.