Large orders are usually split into small pieces and traded over long time horizons. The order splitting reduces the price impact but increases the risk of non-execution. I show that the cost of non-execution can be substantial. By ignoring this cost, the literature underestimates the total transaction cost of large orders. Additionally, this article shows that investors can lower the cost of non-execution by speeding up the trading pace, which, however, increases the cost of successfully filled orders.  Investors’ choice of trading pace is associated with a trade-off where lowering one cost increases the other.

Toxic arbitrage opportunities can arise when the prices of two related securities move sequentially rather than simultaneously. Relatively slow liquidity providers could then incur losses to the arbitrageurs, known as snipers. We investigate arbitrage activity between futures and exchange-traded funds in the Swedish blue-chip index. We find that trading firms that otherwise behave as market makers morph into snipers as toxic arbitrage opportunities emerge. In contrast to common belief, market makers are net beneficiaries of toxic arbitrage. The finding can be rationalized by economies of scope, since both arbitrage and market-making strategies rely on low-latency market monitoring technology.

I estimate the price impact implied by limit order book (LOB) depth, using methodology from Sandås (2001). The idea is that LOB depth reflects market makers’ beliefs about the price impact of market orders. I find the price impact estimates are slightly lower than those from a structural vector autoregressive model, but slightly higher than those from price impact regression. Thus, the LOB-implied price impact estimates match those from benchmark models, contradicting earlier research. I shed new light on limit order informativeness, contributing to contemporary literature on the increased information content of limit orders relative to market orders.

We investigate the economic rationale behind limit order cancellations from the perspective of liquidity suppliers. We predict that an order is cancelled whenever its expected revenue no longer exceeds the expected cost and we model how order profitability variation can be determined from changes in the state of the order book and the order queue position. Our empirical evidence supports the predictions in general and for orders submitted by high-frequency trading firms in particular. Consistent with our model approach, we find that order cancellation patterns are more consistent with market making than with liquidity demand strategies.