The Problem With Conventional Macro
Most institutional macro research rests on frameworks that assume equilibrium, rational expectations, and mechanistic central-bank control. DSGE models treat economic cycles as responses to exogenous shocks around a stable baseline. Statistical and ML-based approaches identify correlations across historical data. Both share a critical weakness: they describe what has happened without modeling why it happened.
When the causal structure of the economy shifts — regime changes, fiscal pivots, balance-sheet stress, credit-cycle inflections — correlation-based approaches break down precisely when they matter most. The 2008 financial crisis, the post-COVID inflation surge, and the resilience of growth through the 2022–23 rate-hike cycle all defied consensus models built on these foundations.
DeepMinsky exists because a different approach is required.
What DeepMinsky Is
DeepMinsky is our proprietary macroeconomic simulation platform built on system dynamics — a methodology that models the economy as a network of interconnected stocks, flows, and feedback loops across all major sectors: households, firms, banks, the fiscal authority, the monetary authority, and the foreign sector.
Rather than fitting regressions to past data, DeepMinsky models the causal mechanisms that generate macro outcomes. Every variable in the system — output, employment, credit, profits, prices, interest rates, government balances — is connected through explicit stock-flow accounting identities and behavioral equations grounded in economic theory. The result is a unified simulation environment where shocks propagate through the same channels they do in the real economy: through balance sheets, income flows, credit creation, and policy feedback.
Multi-sector simulation
Households, high-net-worth investors, firms, banks, central bank, treasury, and the foreign sector — all linked through consistent balance-sheet accounting.
Endogenous credit cycles
Driven by profitability and borrower creditworthiness, not imposed externally.
Scenario analysis
Simulate fiscal consolidation, oil shocks, rate paths, credit tightening, or supply disruptions and trace how they propagate across the entire system.
Forward-looking regime identification
Detect balance-sheet fragility and credit-cycle turning points before they appear in lagging indicators.
DeepMinsky Sandbox
Explore a live demo of the DeepMinsky simulation environment. Analyze baseline forecasts, test alternative policy paths, and see how shocks propagate across growth, inflation, leverage, credit, and asset markets.
Simulation Settings
Primary Simulation
Secondary Simulation
Visualization
Why System Dynamics Is Superior
Feedback loops, not static assumptions
Real economies are defined by reinforcing and balancing feedback loops. Rising profits drive investment, which drives employment, which drives consumption, which feeds back into profits — until costs compress margins and the cycle turns. DeepMinsky captures this entire dynamic endogenously. DSGE models, by contrast, rely on exogenous shocks to generate cycles and assume economies return to equilibrium on their own — an assumption contradicted by decades of financial crises, secular stagnation, and policy-dependent recoveries.
Balance-sheet causality, not equilibrium
In our framework, the binding constraint on economic activity is not market-clearing but balance-sheet feasibility: whether the existing structure of nominal claims — debts, deposits, assets — can be carried forward without forced deleveraging. Asset prices trade not on intrinsic value alone but on what breaks if they fall. This lens explains phenomena that equilibrium models cannot: persistent multiple expansion, the Minsky moment, and why the same policy can be expansionary in one regime and contractionary in another.
Two money printers, not one
Our theoretical foundation recognizes two distinct sources of money creation. Government deficit spending creates the private sector's net financial assets — the Lévy-Kalecki profit channel that seeds corporate profitability and private-sector savings. Endogenous bank credit then amplifies this base through loan-deposit creation, financing investment and production. Neither the quantity of reserves nor the level of interest rates mechanically determines credit creation; profitability and creditworthiness do.
Regime-dependent analysis
A single macro indicator — the saving rate, the yield curve, inflation — has no invariant meaning. Its significance depends entirely on the underlying balance-sheet regime: whether private credit is expanding or contracting, and whether public balance sheets are growing or shrinking. DeepMinsky tracks four distinct regimes defined by these axes, each with different implications for growth, fragility, and asset prices.
The Theoretical Edge
DeepMinsky integrates insights from Post-Keynesian economics, Minsky's Financial Instability Hypothesis, endogenous money theory, and applied Modern Monetary Theory into a single operational platform.
Minsky-Keen credit dynamics
Periods of stability breed rising leverage and deteriorating borrower quality until the credit cycle turns endogenously — not because of an external shock, but because balance-sheet feasibility is exhausted.
The Lévy-Kalecki profit identity
Government deficits are the primary source of private-sector profits, which in turn drive investment, credit expansion, and asset prices.
Endogenous money creation
Banks create money by extending loans; they do not lend out reserves or prior savings. Credit expansion is governed by profitability expectations, not by the central bank's balance sheet.
Balance-sheet price floors
Asset prices are constrained from below by the structure of outstanding claims — prices must remain high enough to prevent forced deleveraging, which is why fiscal flows lead asset-price movements mechanically, not narratively.
Four endogenous breakage channels
Credit systems fail through (1) cost-inflation eroding cash-flow validation, (2) marginal borrower degradation, (3) fiscal asset drains via taxation, or (4) adverse income redistribution away from borrowers. Our System Dynamic Risk Index (SDRI) tracks all four.
From Macro to Portfolio: AlphaTilt
DeepMinsky's simulation output feeds directly into AlphaTilt, our portfolio construction layer that converts macro regime forecasts into actionable multi-asset positioning across duration, equities, credit, commodities, inflation hedges, and defensive allocations.
| Regime State | Dominant Driver | Portfolio Implication |
|---|---|---|
| Growth ↑ / Inflation ↑ | Credit expansion + capacity pressure | Commodities, short duration, inflation hedges |
| Growth ↓ / Inflation ↓ | Fiscal drag + credit contraction | Cash, long duration, gold |
| Leveraging | Credit creation outpacing fiscal drag | HY credit, cyclical equity |
| SDRI Rising | Balance-sheet feasibility breakdown | Defensives, safe assets, reduce exposure |
This is not correlation-based factor rotation. It is causal-layer allocation: positioning based on which balance-sheet forces are expanding or contracting, and what that means for the feasibility of current asset prices. Because fiscal flows and credit impulses move asset prices before earnings and macro data confirm the shift, this framework provides a structural timing advantage over approaches that wait for backward-looking confirmation.
AlphaTilt
Explore a live demo of AlphaTilt. Translate expected changes in growth, inflation, fiscal impulse, credit conditions, and regime risk into tilts across asset classes and strategy sleeves.
Growth surprise is strong positive and inflation surprise is strong positive over a 6-month horizon. At 86% confidence, the portfolio applies significant portfolio repositioning.
Demand impulse is neutral — no strong directional push from fiscal or credit flows.
Governance risks are reducing maximum tilt intensity by ~17%.
No strong feasibility signals.
| Asset | Cap % | Tilt |
|---|---|---|
SGOVAll | 9.6 | Neutral -0.7% |
SCHPQ2/Q4 | 10.5 | Overweight +7.3% |
TLTQ3 | 3.6 | Underweight -10.0% |
IEFQ3 | 3.5 | Underweight -7.2% |
VTIQ1 | 17.3 | Neutral +0.5% |
IEFAQ1 | 8.1 | Neutral -0.1% |
VWOQ1/Q2 | 5.3 | Overweight +2.3% |
AVUVQ1 | 5.1 | Underweight -1.3% |
VNQQ1/Q2 | 5.2 | Overweight +2.2% |
HYGQ1 | 4.1 | Neutral -0.0% |
GLDQ2/Q4 | 15.1 | Overweight +8.1% |
DBCQ2/Q4 | 12.5 | Overweight +11.3% |
Gross: 100.0%(target ~100%)
Who This Is For
DeepMinsky and AlphaTilt are built for institutional investors who want:
- A causal macro framework that explains why regimes change, not just that they changed
- Forward-looking scenario capability — stress-test a macro thesis against fiscal, credit, and policy counterfactuals
- A direct, systematic bridge from macro research to portfolio construction
- A differentiated process that outperforms precisely when consensus models fail — at turning points, policy shifts, and regime breaks
