Optimal execution strategies aim to execute trade orders at a minimized cost and are key components of algorithmic trading systems. Quantifying the cost of trading process provides firms with the ability to use expected results as benchmarks to evaluate their traders performance and trading strategies in a more controllable and systematic way in a post-trade analytical context. The optimal execution is especially important for institutional investors in the buy-side context since their large size orders can affect the market prices of the securities temporarily or even permanently.
In this work, we calculate the optimal execution strategy for liquidation of a large position in a single asset. We focus on minimizing the order execution cost based on the future uncertainty in prices and market impact costs arising from the market impact of the trade. For order execution we employ the the Almgren-Chriss framework which consider the temporary and permanent market impact of the order together with the risk aversion of market traders for optimal execution. Our sample consists of securities listed in Borsa Istanbul XU030 index between 2019 and 2020. We estimate parameters of linear temporary impact function and permanent impact function by using regression technique on limit order book dataset (ITCH protocol). This dataset includes comprehensively dynamics of market represented by behaviour of market participants and dynamics of supply demand connection. We run backtests for the XU030 index securities. Our major contribution is to use the historical limit order book dataset in the backtests whereas prior studies mostly use simulated stock price processes in their model tests. Our simulation analysis suggests that the realized transaction costs and the cost intervals for order executions are parallel and closely track ex-ante theoretical cost and interval estimates provided by the Almgren-Chriss framework.
Anahtar Kelimeler: Optimal Execution, Market Impact, Limit Order Book, Algorithmic Trading, Almgren- Chriss Framework, Implementation Shortfall, High Frequency Trading