Scenarios
What happens to the financial stress index after a climate shock? This page relies on published estimation results as they are, without re-estimation: every figure is an estimated coefficient multiplied by the shock size you choose.
Estimation period: 1999-2025 (27 annual observations).
Updated : 6 November 2026
Shock simulator
Choose the rise in the climate risk index, or a historical scenario, then the state of the financial system (quantile τ). All results below update.
Effect the following year = coefficient of the lagged climate risk index in the financial stress equation × shock size. Dynamic response = VAR response to a one-standard-deviation shock (21.9 points), multiplied by the shock size divided by that deviation. In historical scenarios, the observed rise in the index is treated as a shock: this is an approximation.
The same shock depending on the state of the financial system
Each row gives the estimated effect of the chosen shock with its 95% confidence interval. When the interval crosses the grey line (zero), no effect cannot be ruled out.
| State | Coefficient | 95% CI | p | Significance |
|---|---|---|---|---|
| τ = 0.20 (low stress) | 0.667 | [-0.252 ; 1.288] | 0.180 | not significant |
| τ = 0.50 (median) | 0.687 | [-0.516 ; 1.369] | 0.358 | not significant |
| τ = 0.80 (high stress) | -0.027 | [-0.772 ; 1.719] | 0.922 | not significant |
| τ = 0.90 (maximum stress) | -0.233 | [-0.878 ; 1.922] | 0.860 | not significant |
| Average (VAR) | 0.624 | [0.141 ; 1.108] | 0.007 | significant |
At the quantiles, bootstrap confidence intervals include zero, and the test of equal coefficients across quantiles does not reject equality (p = 0.496). Only the average VAR coefficient is significant. The apparent difference between states is therefore not established, and these results describe predictive precedence, not a causal effect.
Response path over time
Horizontal axis: years after the shock (0 = shock year). Cholesky decomposition with the climate risk index ordered first. Only the one-year response is statistically significant at 5%.
Share of climate risk in financial stress fluctuations
After one year, climate risk shocks explain 8% of the financial stress forecast error variance, and about 46% at 10 years. This decomposition depends on the ordering of variables in the VAR.
Stress test card
Morocco Climate and Economy Observatory—
Go further
What each model forecasts for the financial stress index next year, with its past errors.
Advanced analyticsWhat a machine learning model relies on to reproduce the index, year by year.
Financial stabilityHow both indices are built and how they have evolved since 1998.
Granger test: F = 6.75, p = 0.016. Toda-Yamamoto test (climate risk to financial stress): p = 0.012; reverse direction: p = 0.742.