Case Study 3
Extending Beyond Asset Prices
In our first two use cases, we demonstrated how the Skanalytix financial modeling platform can simulate realistic asset returns — both for individual equities and portfolios — by capturing features like volatility clustering and cross-asset dependencies. In this third use case, we extend the framework to simulate macro-financial environments: coherent time series in which interest rates and economic indicators evolve together.
Our goal is to generate synthetic yield curves that respond dynamically to macroeconomic conditions, such as inflation expectations, unemployment, and monetary policy decisions. Unlike traditional yield curve models that isolate rates from their economic context, our approach models the joint behavior of financial and macro variables — enabling more realistic scenario generation, stress testing, and policy exploration.
Generator Temporarily Unavailable
An improved version will be available soon. Thank you for your patience!
Key Variables in Macro-Financial Modeling
To capture the complex interactions between the economy and interest rates, our simulations focus on six core variables:
- unemployment rate
- inflation expectations
- federal funds rate,
- 2-year Treasury yield
- 10-year Treasury yield
- 30-year year Treasury yield.
Unemployment and inflation expectations serve as primary measures of economic slack and price pressures, which heavily influence the Federal Reserve’s policy decisions. The fed funds rate, as the central bank’s policy instrument, directly impacts short-term interest rates — especially the 2-year yield — while longer-term yields reflect market expectations of future inflation, growth, and monetary policy. This set of variables allows us to model the drivers of interest rate policy and the resulting yield curve dynamics across a range of economic conditions. We focus on inflation expectations rather than actual inflation, as these better reflect market beliefs and policy drivers in forward-looking simulations.
Note on Variable Selection:
Although GDP growth is an important macroeconomic indicator, we chose not to include it explicitly in our simulations. This decision was made to keep the model parsimonious while focusing on variables most directly linked to monetary policy and yield curve dynamics. In practice, the effects of GDP growth are often reflected through unemployment and inflation expectations — both of which are core inputs in our framework. Since the Federal Reserve typically responds to economic slack and price stability, rather than headline GDP figures, this streamlined selection still allows us to generate macro-consistent and policy-relevant scenarios without redundancy.
The figure below shows the values of these variables over the period from January1, 2003 to April 1, 2025. The top three panels display unemployment, inflation expectations, and the federal funds rate; the fourth panel shows the 2-, 10-, and 30-year Treasury yields.
Historical Patterns in Macroeconomic and Yield Curve Data
Before we explore synthetic simulations, it’s useful to understand how these variables behaved during key historical periods. Our dataset spans multiple economic cycles, including notable episodes that highlight the interplay between unemployment, inflation expectations, monetary policy, and yield curve dynamics.
- 2008 Global Financial Crisis: Prior to the crisis, the Fed raised rates gradually, narrowing the gap between short- and long-term yields (a flattening of the yield curve). As unemployment surged and inflation expectations fell, the Fed cut rates to near zero. Short-term yields dropped sharply, with long-term yields falling less, steepening the curve — a signal of recession and prolonged low rates.
- 2020 COVID-19 Shock: The pandemic caused a rapid spike in unemployment and a collapse in inflation expectations. The Fed again cut rates near zero. Short-term yields fell more dramatically than long-term yields, steepening the curve swiftly. By late 2020, long-term yields rose as markets anticipated recovery and inflation pressures.
- 2022–2023 Rate Hike Cycle: Inflation expectations surged starting in late 2020, with actual inflation peaking near 9% in mid-2022. The Fed responded aggressively, raising the funds rate and pushing short-term yields up quickly. Longer-term yields rose more gradually, leading to a yield curve inversion — a common recession indicator.
From Historical Data to Synthetic Simulations
While these historical episodes provide valuable insights into how economic variables and interest rates interact, they represent specific events shaped by unique circumstances and policy decisions. Our goal with synthetic data is not to replicate exact past crises or identify their specific causes. Instead, we focus on generating macro-consistent scenarios — simulated time series where economic and financial variables evolve together in ways that are realistic and coherent.
For example, a synthetic increase in unemployment might arise without an explicit backstory, but it should still lead to plausible responses in inflation expectations, policy rates, and the yield curve. This approach allows us to explore a wide range of possible economic environments, including rare or hypothetical conditions, helping investors, policymakers, and researchers better understand potential risks and opportunities beyond historical experience.
Although the variables differ from earlier use cases, the underlying modeling approach remains the same. We apply the Skanalytix modeling platform to macro-financial data in the same way we applied it to equities and portfolio returns: by capturing realistic dependencies, time dynamics, and systemic behavior without relying on rigid parametric assumptions. What’s new here is the domain — not the engine.
Synthetic Simulation Results: Macro-Consistent Scenarios
We now present a series of synthetic macro-financial simulations generated using the Skanalytix platform. These scenarios are produced without manual intervention or hard-coded rules — instead, they reflect the statistical relationships and time dynamics learned from historical data. Like the historical examples shown earlier, each simulation is visualized as a four-panel figure: the top three panels display unemployment, inflation expectations, and the federal funds rate; the fourth panel shows the 2-, 10-, and 30-year Treasury yields.
Although entirely synthetic, these scenarios exhibit realistic and internally consistent behavior. For example, a rise in unemployment often triggers a drop in the policy rate and a steepening of the yield curve, while persistent inflation expectations may lead to yield curve flattening or inversion. These responses are not hardcoded but instead arise naturally from the learned statistical structure of the model.
In addition to capturing directional effects, the simulations reflect key empirical properties of financial and macroeconomic data — such as volatility clustering and heavy tails in yield changes. These characteristics emerge from the generative process itself, without requiring separate volatility modeling.
This simulation mode provides a valuable tool for exploring macro-consistent alternatives to historical data — useful for stress testing, scenario design, or understanding how yield curves might behave under unfamiliar or hypothetical conditions.