Naive Macroeconomic Model Failure Dashboard

A simple linear fit looks stable in a quiet window, then breaks across zero-lower-bound policy, shifting inflation dynamics, and post-2020 regime change.

Quiet-period RMSE
0.32 pp
The model appears well-behaved during the calm 2005-2007 fit window.
Baseline fit window
GFC / ZLB RMSE
7.85 pp
Error surges once policy is pushed against the lower bound in the post-2008 regime.
24.7× quiet-period RMSE
COVID / inflation RMSE
9.34 pp
The same model misses badly again during the 2020-2022 inflation shock.
29.4× quiet-period RMSE
Rolling CPI-UN correlation
+0.85 → -0.86
The inflation-unemployment relationship swings across sign, so coefficients learned in one regime do not travel well.
Latest: -0.74
Why the naive model fails
1

The fit looks excellent in the calm 2005-2007 window (RMSE 0.32 pp), which creates the illusion that a simple linear mapping from inflation and unemployment to the policy rate is stable.

2

Once the economy hits the 2008-2015 zero-lower-bound regime, the model keeps wanting much lower rates than the policy framework can deliver, pushing GFC-era RMSE to 7.85 pp.

3

After 2020, the same coefficients are exposed again: unemployment, inflation, and the funds rate move through combinations that were not present in the quiet sample, driving the largest miss to 30.532 pp in Apr 2020.

How to read the dashboard
 GFC / ZLB band  COVID / inflation band  Actual Fed Funds  Naive prediction

The first three charts are time-series views: policy-rate fit, rolling CPI-unemployment correlation, and residuals. All three repeat the same regime shading so the breakdown is easy to compare across views.

The regime error bars quantify the overfitting problem directly, while the scatter plot shows how the inflation-unemployment feature space itself shifts across periods. Marker size scales with absolute prediction error, so the biggest points are the months where the naive fit breaks most visibly.

Model Failure Across Regimes
Actual vs. naive-predicted Fed Funds rate. The dashed zero line marks the lower bound, while shaded bands flag the 2008-2015 and 2020-2022 regimes where the simple fit fails structurally.
The Breaking Phillips Curve
A 36-month rolling correlation between CPI inflation and unemployment shows that the predictor relationship itself is unstable, flipping across sign and magnitude over time.
Model Residuals
Positive residuals mean the naive model predicted rates that were too low; negative residuals mean it predicted rates that were too high. Crisis bands make the error bursts explicit.
Overfitting to a Quiet Period
Grouped bars compare mean absolute error and RMSE across distinct macro regimes, making the quiet-period overfit visible in one view.
Feature Relationship Drift
Inflation vs. unemployment is not a single stable cloud. Regime coloring and error-scaled marker sizes show why one static coefficient vector struggles across the full sample.
Largest Monthly Misses
Date Regime Actual Fed Funds Naive Prediction Prediction Error