Spotting the stock and crypto markets’ rings of fire: measuring change proximities among spillover dependencies within inter and intra-market asset classes

Crypto assets have lately become the chief interest of investors around the world. The excitement around, along with the promise of the nascent technology led to enormous speculation by impulsive investors. Despite a shaky understanding of the backbone technology, the price mechanism, and the business model, investors’ risk appetites pushed crypto market values to record highs. In addition, pricings are largely based on the perception of the market, making crypto assets naturally embedded with extreme volatility. Perhaps unsurprisingly, the new asset class has become an integral part of the investor’s portfolio, which traditionally consists of stock, commodities, forex, or any type of derivative. Therefore, it is critical to unearth possible connections between crypto currencies and traditional asset classes, scrutinizing correlational upheavals. Numerous research studies have focused on connectedness issues among the stock market, commodities, or other traditional asset classes. Scant attention has been paid, however, to similar issues when cryptos join the mix. We fill this void by studying the connectedness of the two biggest crypto assets to the stock market, both in terms of returns and volatility, through the Diebold Francis spillover model. In addition, through a novel bidirectional algorithm that is gaining currency in statistical inference, we locate times around which the nature of such connectedness alters. Subsequently, using Hausdorff-type metrics on such estimated changes, we cluster spillover patterns to describe changes in the dependencies between which two assets are evidenced to correlate with those between which other two. Creating an induced network from the cluster, we highlight which specific dependencies function as crucial hubs, how the impacts of drastic changes such as COVID-19 ripple through the networks—the Rings of Fire—of spillover dependencies.