Anmerkungen:
Nach Informationen von SSRN wurde die ursprüngliche Fassung des Dokuments June 30, 2022 erstellt
Beschreibung:
We analyse the price discovery process in bitcoin-dollar trading on Coinbase, the most established cryptocurrency spot exchange. Using a modified reinforcement learning framework, we find that market orders individually carry more information than limit orders, but due to their abundance, limit orders are the main driver of price discovery. Moreover, we observe a significant asymmetry in permanent price impact, with sell market orders carrying more information than buy market orders. We relate this asymmetry to unidirectional retail order flows. Owing to a negligible relative tick size, most of the price discovery contribution of limit orders originates from orders undercutting the best quotes. Consistent with theoretical predictions, the state of the order book has a significant impact on price discovery, with orders submitted along the order book imbalance showing an increased contribution and the importance of market orders being more pronounced in case of a binding bid-ask spread. Finally, the price discovery process on Coinbase is strongly influenced by information flows from other exchanges