A risk-based explanation of cryptocurrency returns

Abstract

We investigate the dynamics of returns in cryptocurrency markets through the lens of a small-scale latent factor model with time-varying factor loadings instrumented by individual cryptocurrency characteristics. We have three main empirical findings. First, our factor model excels in providing a risk-based explanation of daily realised and expected returns of cryptocurrencies, improving over both static latent factor models and pre-specified portfolios sorted on observable characteristics. Second, we show that expected returns are primarily driven by liquidity, volatility, and past performance. Third, our model provides evidence of a limited, although significant, spill-over of fundamental risk factors between equity and cryptocurrency markets.

Keywords: Instrumented PCA, Cryptocurrency markets, Asset pricing, Factor models, Risk premiums.

JEL Classification: G11, G12, G17, C23