Architecture and metrics of the macro-fundamental scoring engine for G7 currencies. Analysis of structural value, cyclical dynamics, and risk regimes.
Technical Note
This model rests on the "Caveman Quant" philosophy: absolute vectorization, zero look-ahead bias, and an explicit macro rationale for each driver.
Executive Summary
This paper presents the design of a macro-fundamental scoring engine for G7 currencies. The model decomposes the FX signal into three layers: structural value (L1), cyclical dynamics (L2), and risk regime (L3), validated on more than 20 years of historical data.
1. Theoretical Foundations
The value of a currency is determined by the equilibrium of international investment flows, governed by four fundamental forces:
Interest rate differential: Based on Uncovered Interest Rate Parity (UIP). Over horizons of 1 to 3 months, carry reversion tends to dominate.
Trade balance: Real flows acting as a structural bullish driver for the domestic currency.
Monetary policy direction: A cyclical signal anticipating changes in real rates.
Growth momentum: A cyclical signal where economic overheating may signal a future deterioration of the current account.
2. Model Architecture
The model (Engine 2) uses currency-specific weights, reflecting the distinct mandates of central banks (e.g. the Fed’s dual mandate vs the ECB’s price stability).
The Three Signal Layers:
Layer 1 (Structural Value): Level of the policy rate and change in the trade balance.
Layer 2 (Cyclical Dynamics): Direction of monetary policy and GDP momentum.
Layer 3 (Risk Regime): Use of a Gaussian Mixture Model (GMM) on the VIX and the S&P 500 to classify Risk-On / Risk-Off environments.
3. Results and Performance
Out-of-sample (OOS), the Engine 2 model shows robust results:
Daily IC: 0.071 (significant at 5%).
Grinold IR: 0.60.
Portfolio Performance: +59.6% cumulative return over the history with a Sharpe of 0.48 and a Max Drawdown contained at -13.8%.
To access the detailed mathematical formulas, the Newey-West HAC implementation, and the performance charts by zone, download the full report at the top of the page.
Disclaimer: This document is provided for informational and educational purposes only. It does not constitute investment advice.
Premium Research — €10
Access the full research
This detailed research paper is available per unit or via subscription. Unlock our quantitative models and exclusive investment theses for €10.
Passionate about market analysis and statistical modeling, Léo oversees the strategic allocation of the model portfolio and the development of Horacle Capital's quantitative frameworks, as well as writing weekly articles.
