Research Library
Research
Working papers on market risk, tail risk and the validation of machine-learning models in finance, with the code and data needed to reproduce them.
Model validation
When Backtests Lie: A Governance-Grade Validation Framework for Machine-Learning Value-at-Risk Models under Regime Shifts
A governance-oriented validation framework for machine-learning Value-at-Risk models. It extends conventional backtesting with regime-conditional coverage analysis, the Probability of Backtest Overfitting and conformal recalibration, to test whether aggregate evaluation can hide weaknesses in stressed or changing markets.
Backtesting Value-at-Risk Models Under SR 11-7: A Comparative Analysis of Kupiec, Christoffersen, and Basel Traffic-Light Tests Applied to S&P 500 Returns
A walk-forward empirical study of Value-at-Risk models on S&P 500 daily returns, comparing unconditional coverage, conditional coverage and Basel traffic-light backtests, and their relevance to model validation and risk management.
Tail and systemic risk
Conditional Tail Episode Risk: A Path-Dependent Framework for Extreme-Loss Episodes Beyond Value-at-Risk and Expected Shortfall
A path-dependent approach to tail risk that treats extreme losses as episodes, measuring their persistence and cumulative severity rather than single-day breaches alone.
Systemic Tail Episode Risk: Persistence and Cumulative Severity of Synchronized Extreme-Loss Episodes Across Markets
Extends tail-episode analysis across markets to measure how often extreme losses arrive together, how long synchronized stress persists, and how severe it becomes in cumulative terms, a dimension that standard portfolio VaR can understate.
Machine-learning risk models
ML-LiqVaR: A Liquidity- and Regime-Aware, Extreme-Value-Calibrated Gradient-Boosting Value-at-Risk Model for US Equities
A framework for one-day-ahead Value-at-Risk forecasting that combines volatility modeling, Extreme Value Theory, gradient-boosted quantile regression, liquidity and volatility features, and regime-aware modeling, while keeping rigorous statistical validation.
Research in progress
When the Tail Lies Too: governance-grade backtesting of Expected Shortfall under regime shifts
A sequel to When Backtests Lie that applies the same governance lens to Expected Shortfall, the capital measure in the Basel Fundamental Review of the Trading Book. Results will be published here once the working paper is posted.
Analysis on Wall Street Feed
Research-driven market analysis draws on these methods. Browse the Quant & AI section, or read the author profile for the wider research agenda. All research code is on GitHub.
About these papers. The papers listed here are working papers by Niraj Neupane (Calderyn Institute). Unless a paper is identified as peer-reviewed or formally published, it has not completed journal peer review, and its findings may change as methods, data and validation evolve. The research reflects the author’s independent views and is not investment advice. See our Editorial Policy and Corrections Policy.
