Agent-Based Computational Economics
Cognitive Macroeconomics is developing a computational simulation platform that models financial markets, payment systems, and regulatory environments as living ecosystems of autonomous agents — enabling institutions to stress-test policy, understand systemic risk, and explore outcomes before they happen.
The Problem We Solve
Traditional economic models assume rational actors, equilibrium states, and predictable behaviour. Real markets don't. They are made up of thousands of heterogeneous agents — banks, regulators, traders, governments — each acting on incomplete information, interacting in non-linear ways.
Our digital twin approach simulates this complexity from the ground up: autonomous agents with bounded rationality, AI-driven decision-making, and emergent macroeconomic behaviour — so institutions can see what happens before they act.
01 — USE CASE
Simulate proposed regulations across thousands of market conditions before implementation — identifying unintended consequences, procyclicality, and systemic risks invisible to static models.
02 — USE CASE
Model participant behaviour in high-value payment systems under liquidity shocks, counterparty failures, or structural changes like de-tiering and system migration.
03 — USE CASE
Trace how shocks propagate through CDS networks, interbank lending, and clearing systems — mapping fragility before it becomes crisis.
04 — USE CASE
Model stablecoin dynamics, crypto arbitrage behaviour, and CBDC introduction scenarios — stress testing market design choices in silico.
Platform Architecture
Real-world data — payments, balance sheets, market prices, regulatory parameters — is used to calibrate agent behaviour and system topology.
Thousands of autonomous agents are instantiated — banks, CCPs, regulators, traders — each with distinct decision rules, risk appetites, and information sets.
The artificial economy runs. Agents interact, markets clear, shocks propagate, regulations bite. Emergent macroeconomic outcomes arise from the bottom up.
Outputs — systemic risk metrics, price dynamics, network topology, liquidity flows — are surfaced for analysis, comparison, and policy insight.
Scientific Foundation
The digital twin is not a concept without roots. It is built on over a decade of peer-reviewed research in computational economics, market microstructure, financial networks, and regulatory simulation — published in leading journals and central bank research programmes.
Numerical approximation of the risk-neutral daily valuation of LVTS Tranche 2 (2005–2016), identifying conditions for participant withdrawal and specifying "risk-based access" to clearing.
First theoretical framework for assessing optimisation decisions in payment systems from a market microstructure perspective using ACE and stochastic games.
GARCH, SVAR, and TVP-VAR analysis of USD-backed stablecoin volatility. Finds heterogeneous responses to macro-financial shocks with growing systemic integration post-2021.
Demonstrates significant ML-based predictive ability for real GDP and CPI using near-real-time payments system data, reducing the lag of official statistics.
Shows how Basel II CRT frameworks unintentionally intensified systemic risk through growing CDS market clustering — calibrated against FDIC call report data.
Persistent triangular arbitrage of $2–$25 identified across crypto and stablecoin pairs using genetic algorithms. Published in the Journal of Financial Market Infrastructures.
Why Cognitive Macroeconomics
Unlike commercial risk platforms built on parametric assumptions, our approach is rooted in computational economics research — with peer-reviewed foundations, real institutional data, and models that generate emergent behaviour rather than interpolating from historical correlations.
Founder
Dr. Oluwasegun Bewaji
Founder · Chief Economist
An applied computational economist whose research sits at the intersection of market microstructure, agent-based simulation, and institutional finance.
Segun has spent over a decade building the intellectual and empirical foundations for what Cognitive Macroeconomics is now productising. His research has modelled Canadian LVTS dynamics, stablecoin volatility, CDS network fragility, CDO market formation, payment volume migration, and cryptocurrency arbitrage — always using real institutional data and agent-based methods.
He has presented at the Bank of Canada and Payments Canada joint conference on agent-based modelling in financial market infrastructures, and has contributed to thinking at ACEFINMOD. His working models — functioning Java simulators for CDS networks, RMBS markets, and CDO dynamics — represent a decade of active prototype development now being rebuilt in Python for broader accessibility and integration.
Cognitive Macroeconomics is the company that brings that body of work into a coherent, institutional-grade simulation platform.
Stay Connected
We are in active development and not yet taking client engagements. If your institution is interested in the digital twin platform — as a future user, research partner, or collaborator — register your interest and we will keep you informed as we progress.
Where we are: The scientific models are proven and published. We are currently rebuilding the simulation engine in Python, expanding the agent framework, and designing the institutional API. Early partnership conversations are welcome.
No commitment. We'll be in touch as the platform develops.