Observatory · macro-financial risk

Argus

Open-source risk intelligence

Framework for macro-financial risk analysis. Integrates banking supervision, systemic risk, macro indicators, and cross-border flows into a composite score, backed by 1.3 million rows of real regulatory data. Composite scoring backtests against the GFC, Euro Crisis, Taper Tantrum, COVID, and SVB collapse.

285
Modules
11
Data Collectors
7
Briefing Types
6
Risk Layers

What problem this solves

Fragmented risk data

Banking risk, systemic risk, macro indicators, and cross-border flows live in separate silos. Analysts spend more time assembling data than analyzing it. Argus integrates all four into one composite score.

No open tooling

Proprietary platforms dominate macro-financial risk analysis. Methodologies are opaque. Reproducibility is impossible. Every model in Argus is open, documented, and testable.

Supervisory gap

Existing open-source finance tools focus on trading and portfolio analytics. None offer CAMELS-calibrated bank scoring or AML/TBML detection. Argus covers what regulators actually need.

Six analytical layers

Each layer is independently useful. Together they produce a composite risk score that backtests against the major crises of the last two decades.

Layer 1

Banking Supervision

CAMELS-calibrated supervisory scoring (MSPS), distress prediction, CRE/HTM concentration, NIM compression, wholesale funding risk.

24 modules
Layer 2

Systemic Risk

CoVaR, SRISK, MES, network clearing vectors, G-SIB scoring, fire sale externalities, repo market stress, bank run dynamics.

33 modules
Layer 3

Macro-Financial

Financial conditions index, credit impulse, BVAR, yield curve dynamics, housing, inflation expectations, fiscal sustainability.

35 modules
Layer 4

Cross-Border

Trade-based money laundering, AML network analysis, FATF compliance, sanctions screening, capital flow reversals, dedollarization.

27 modules
Layer 5

Integration

Composite scoring engine, sensitivity analysis, risk attribution, backtesting framework, Granger causality across layers.

10 modules
Layer 6

Crosscutting

Climate transition risk, cyber risk scoring, geopolitical risk index, fintech disruption, pandemic preparedness.

5 modules

Real data, no mocks

Every metric is backed by a public, verifiable source. Eleven collectors pull from regulators, central banks, and statistical agencies into a single SQLite warehouse.

FDIC Call ReportsBank financials
Yahoo FinanceBHC equities
FREDMacro series
BISCross-border banking
UN Comtrade / BACITrade flows
OFACSanctions lists
FATFAML/CFT ratings
FinCENSuspicious activity
TICCapital flows
CPMIPayments data
SEC EDGAR10-Q / 10-K filings

Built with

Python FastAPI Jinja2 SQLite NumPy / SciPy pandas statsmodels scikit-learn Plotly 6 Claude AI Apache 2.0

Domain experience, not just code

Each of Argus's four core risk layers maps directly to professional experience. Banking supervision comes from five years at Bangladesh Bank, implementing Basel III compliance and building early-warning systems for banking sector oversight. Cross-border risk comes from three years in the Financial Intelligence Unit, investigating money laundering, trade-based financial crime, and presenting FATF compliance reviews. Macro-financial analysis is informed by a PhD in Economics and ongoing postdoctoral research on trade policy at the University of Tennessee. Systemic risk ties the layers together: the question regulators most need answered is how one institution's failure cascades into the system.

Argus exists because the analytical tools regulators and researchers need should be open, auditable, and reproducible. Every module has tests. Every data point comes from a public, verifiable source.