VaR/ES Forecasts
Monthly stock-level tail risk forecasts with ticker, company name, VaR, and ES.
ReSGA Tail Risk Forecasts
This website is based on the paper “ReSGA: A Large Tail Risk Model for Learning Value-at-Risk and Expected Shortfall.” It provides monthly stock-level Value-at-Risk and Expected Shortfall forecasts for U.S. equities, together with realized returns, historical analogue states, latent risk networks, and reliability monitoring tools for interpreting current tail-risk conditions.
Research Outputs
Risk Drivers
Feature-group importance showing which characteristic families move forecast loss the most.
Explore DriversHistorical Analogues
Similar historical states with their subsequent tail-risk outcomes.
Explore AnaloguesLatent Risk Network
Explore latent risk neighborhoods, similarity links, and stock-level connected risk profiles.
Explore NetworksForecast Browser
Latest date
-
Stocks covered
-
Average VaR (%)
-
Average ES (%)
-
| Loading predictions... | |||||
Risk Drivers
Importance is computed by masking one characteristic group at a time and measuring the resulting change in forecast loss across the ReSGA forecast archive.
Overall
What matters most?
Time variation
-
Monthly driver map
Loading monthly importance...
Historical Analogues
Find Similar States
Compare the current forecast state with high-similarity historical states.
Open Stock DetailLoading historical analogue preview...
Risk Network
Loading latent neighborhoods...
Loading neighborhoods...
Backtesting
Forecast Status
Latest evaluated
Loading
Stocks evaluated
-
VaR CC test
Pass / Fail / Insufficient data
ES ASER test
Pass / Fail / Insufficient data
E-backtesting Alert
Green
Yellow
Red
GreenLoading
Yellow-
Red-
Green e-process < 2
Yellow 2-20
Red >= 20
Sequential ES evidence monitor loading...
Documentation