Range-Based Volatility Estimators for Monitoring Market Stress: Evidence from Local Food Price Data

Battle Count: 58

The paper primarily targets development economics and humanitarian monitoring rather than traditional quantitative trading. However, the methodology is directly transferable: range-based volatility estimators (especially Yang-Zhang) are standard tools in financial risk management, derivatives pricing, and trading strategy development. The insight that volatility captures symmetric shocks and offsetting supply-demand disturbances that momentum indicators (RSI) miss is relevant for commodity trading, pairs trading, and market-making strategies. The detection framework could inform volatility-based trading signals in thinly traded or illiquid commodity markets. The fiGARCH-GED framework for modeling long-memory volatility with heavy tails is applicable to commodity futures and spot markets.

The volatility estimators themselves are computationally simple (closed-form formulas over rolling windows) and require no model re-estimation. The detection rule is a straightforward two-condition threshold. However, the underlying data pipeline (fiGARCH-GED estimation, GED quantile computation, OHLC construction from conditional distributions, ML-based imputation) is moderately complex. The paper emphasizes that the operational workflow is lightweight once OHLC data is available. Implementation in R is straightforward for the estimators; the RTP data generation is the more complex component.

Code and data archived on Zenodo (https://doi.org/10.5281/zenodo.18846560). All estimators implemented in R. RTP data available through World Bank data portals. Methodology is transparent with explicit formulas for all estimators. However, the underlying fiGARCH-GED model parameters and imputation process are proprietary to the RTP system, and event timelines rely on ex post interpretation. Rolling window (n=10), annualization (N=12), and threshold parameters are clearly specified.