LAUSR

In this paper we introduce a model for trading of large market orders in limit-order books. We use a linear combination of self-exciting Hawkes processes to model asset-price dynamics, as done in [Bacry et al., 2013a]. We make the addition of a general price-impact function to this model, which allows for specification of an impact function that is concave, as concavity is known to be a characteristic of real-life order books. Such functions have a dynamic property that is driven by changes in the order-flow imbalance (OFI), that is, the immediate price-impact function changes according to the difference in the intensity of the Hawkes process. Dependence on OFI adds complexity but also makes for a more realistic model, as it is realistic to consider how order-flow imbalance will change the price impact of an incoming order. In this paper we introduce a criterion for a general price-impact function, and within this framework we examine the immediate and permanent impacts of large orders, we look at the potential for price manipulation, and we show the effectiveness of strategies such as time-weighted average price (TWAP). Overall this paper contributes toward an ongoing discussion on price impact for models with Hawkes processes.