EVALUATION OF THE FORECASTING ABILITY OF RISK-NEUTRAL DENSITY IN BITCOIN OPTIONS

Abstract

This study assesses the out-of-sample forecasting capabilities of risk-neutral density models in Bitcoin options market, with a focus on the Normal Inverse Gaussian (NIG) density. Understanding forward-looking price dynamics becomes critical as cryptocurrencies continue to gain a reputation in financial markets. This research examines how the NIG model, with its capability to capture skewness and kurtosis, compares to the benchmark log-normal (LN) distribution. The analysis applies the likelihood ratio test to evaluate the predictive performance of the models. As a result, NIG model improves the accuracy of tail forecasts, outperforming LN in capturing extreme market movements, which holds implications for risk management and market timing in Bitcoin market.