How we transformed Market Monetarist theory into an interactive tool with Claude AI
Recently, I discussed the risks of political interference in monetary policy and the potential for a 1970s-style inflation scenario in my Market Monetarist articles. Today, I want to share how these theoretical concepts have been turned into an interactive simulation that allows readers to explore these relationships themselves.
From Theory to Interactive Simulation
The simulation model emerged from two recent articles:
“Trump’s Tariffs and the Fed: A 1970s Velocity Nightmare in the Making” – examining how the politicisation of the Federal Reserve under G. William Miller led to an immediate surge in money velocity, well before external shocks like the Iranian Revolution.
“The Trump Superspike: If Trust in U.S. Governance Breaks, Interest Rates Will Explode” – exploring how market confidence in institutions underpins bond market stability, and how quickly that confidence can unravel.
These rather complex macroeconomic concepts benefit greatly from visual and interactive exploration, which is why I decided to create a simulator.
How the Simulator Was Created
Rather than coding this from scratch myself, I experimented with Claude 3.7 Sonnet to transform these concepts into a React-based simulation. The process was surprisingly effective:
Claude analysed the articles and identified the key economic relationships to model – particularly the connections between Fed credibility, money velocity, inflation expectations, and interest rates.
Using the historical examples from my articles (like the 7% velocity increase following Miller’s appointment and interest rates jumping from 9% to 15%), Claude developed mathematical relationships that could replicate these dynamics.
The final product is a React component with adjustable parameters, scenario presets, and dynamic visualisations that readers can use to explore alternative scenarios.
The Market Monetarist Model Behind the Simulator
The simulator captures several market monetarist principles:
1. Credibility and Inflation Expectations
When central bank credibility erodes, both short-term and long-term inflation expectations become unanchored. This relationship is at the heart of the simulator – reduce the “Fed Credibility” parameter and watch inflation expectations drift upward over time.
This reflects a core market monetarist insight: market expectations about future monetary policy are central to current economic outcomes. When people begin to doubt the central bank’s commitment to price stability, their behaviour changes immediately.
2. The Velocity Effect
Perhaps the most important insight from my “Velocity Nightmare” article is how quickly money velocity can surge when monetary policy credibility is questioned. The simulator shows this relationship explicitly.
When Fed credibility falls below certain thresholds, particularly with high political interference, money velocity accelerates – mimicking the historical pattern observed after Miller’s appointment. This captures the essential market monetarist insight that money demand is sensitive to expected inflation.
3. Interest Rate Dynamics and Threshold Effects
The simulator models both gradual and sudden changes in interest rates. Under normal conditions, interest rates rise gradually with inflation expectations. But when credibility falls below critical thresholds, we see non-linear “jumps” in interest rates – similar to what happened in the Liz Truss scenario.
This feature captures how financial markets don’t just price in gradual changes but can exhibit discontinuous repricing when confidence thresholds are breached.
4. Tariffs and Supply Shocks
Following a market monetarist approach, the model distinguishes between demand-side and supply-side inflation. Tariffs represent a supply-side shock that monetary policy cannot easily offset without causing further distortions.
Using the Simulator
The simulator includes several preset scenarios for readers to explore:
- Baseline Scenario: Current economic conditions with moderate risks
- 1970s Miller Scenario: Recreates conditions with declining Fed credibility
- Market Confidence Crisis: Similar to the Liz Truss event
- Trade War Scenario: High tariffs with political pressure on monetary policy
- Stable Policy Scenario: Strong Fed independence and low trade barriers
Beyond these presets, readers can adjust individual parameters:
- Fed Credibility (0-100%)
- Political Interference (0-100%)
- Tariff Levels (0-50%)
- Starting Inflation Rate
- Time Horizon (1-10 years)
The results display in real-time through three main charts:
- Interest Rates & Inflation Expectations
- Money Velocity
- Price Level (indexed to 100 at start)
Educational Purpose and Limitations
The simulator simplifies complex economic relationships for educational purposes. It’s meant to help readers understand the concepts from my articles, not to serve as a precise forecasting model.
Despite simplifications, it effectively demonstrates a core insight: monetary policy credibility can erode quickly, with severe consequences for inflation expectations, money velocity, and ultimately price stability.
AI-Assisted Economic Education
What’s particularly interesting about this project is how AI can help bridge the gap between economic theory and public understanding. By transforming market monetarist concepts into interactive simulations, we can make these ideas more accessible.
I expect we’ll see more AI-assisted economic education tools in the future. The conversational development process allows economists to focus on conceptual aspects while AI handles technical implementation.
I invite all Market Monetarist readers to explore the simulator, test different scenarios, and develop their own intuition about how monetary policy credibility, political interference, and trade policies interact to shape economic outcomes.
In an era of increasing challenges to central bank independence and rising protectionism, understanding these relationships is more important than ever.
You can try the Simulator here.
This entire article was written by Claude 3.7 Sonnet with input from Lars Christensen (LC@paice.io).
