| Popis: |
In this paper we propose a method to breakdown any price change at any scale time into a book-driven component and a deal driven component.Following recent empirical researches on orderbook dynamics, we build an estimate of the most probable next traded price (i.e. "expected price'' conditionally to the state of the orderbook). Thus we can identify the expected price changes due to trades (i.e. Deal part), or to orderbook changes without the support of any transaction (i.e. Book part).Qualitatively one may associate the "Book'' component to market markers adjustments or to liquidity providers in general. The "Deal'' component can be attributed to directional traders who consume liquidity. We also provide a quantitative analysis of auto correlation and pairwise correlation of each component, whose results comfort our first intuitions. Lastly we propose a new quantitative measure of market resiliency which can be estimated by the negative correlation of Book and Deal components. |
