nep-mst New Economics Papers
on Market Microstructure
Issue of 2017‒10‒15
seven papers chosen by
Thanos Verousis

  1. High Frequency Market Making with Machine Learning By Matthew F Dixon
  2. Trader Positions and Marketwide Liquidity Demand By Esen Onur; John S. Roberts; Tugkan Tuzun
  3. A buffer Hawkes process for limit order books By Ingemar Kaj; Mine Caglar
  4. The information content of short selling and put option trading: When are they substitutes? By Deng, Xiaohu; Gao, Lei; Kemme, David
  5. Fundamentals unknown: Momentum, mean-reversion and price-to-earnings trading in an artificial stock market By Schasfoort, Joeri; Stockermans, Christopher
  6. Liquidity risk in markets with trading frictions: What can swing pricing achieve? By Ulf Lewrick; Jochen Schanz
  7. Large-Scale Portfolio Allocation Under Transaction Costs and Model Uncertainty: Adaptive Mixing of High- and Low-Frequency Information By Hautsch, Nikolaus; Voigt, Stefan

  1. By: Matthew F Dixon
    Abstract: High frequency trading has been characterized as an arms race with 'Red Queen' characteristics [Farmer,2012]. It is improbable, even impossible, that many market participants can sustain a competitive advantage through the sole reliance on low latency trade execution systems. The growth in volume of market data, advances in computer hardware and commensurate prominence of machine learning in other disciplines, have spurred the exploration of machine learning for price discovery. Even though the application of machine learning to price prediction has been extensively researched, the merit of this approach for high frequency market making has received little attention. This paper introduces a trade execution model to evaluate the economic impact of classifiers through backtesting. Extending the concept of confusion matrix, we present a 'trade information matrix' to attribute the expected profit and loss of tick level predictive classifiers under execution constraints, such as fill probabilities and position dependent trade rules, to correct and incorrect predictions. We apply the execution model and trade information matrix to Level II E-mini S&P 500 futures history and demonstrate an estimation approach for measuring the sensitivity of the P&L to classification error. We describe the training of a recurrent neural network (RNN) and show (i) there is little gain from re-training the model on a frequent basis; (ii) that there are distinct intra-day classifier performance trends; and (iii) classifier accuracy quickly erodes with the length of prediction horizon. Our findings suggest that our computationally tractable approach can be used to directly evaluate the performance sensitivity of a market making strategy to classifier error and can augment traditional market simulation based testing.
    Date: 2017–10
  2. By: Esen Onur; John S. Roberts; Tugkan Tuzun
    Abstract: In electronic, liquid markets, traders frequently change their positions. The distribution of these trader position changes carries important information about liquidity demand in the market. From this distribution of trader position-changes, we construct a marketwide measure for intraday liquidity demand that does not necessarily depend on aggressive trading. Using a rich regulatory dataset on S&P 500 E-mini futures and 10-year Treasury futures markets, we show that this liquidity demand measure has a positive impact on prices. We then decompose our measure of liquidity demand into three components: aggressive, passive and mixed liquidity demand. Passive liquidity demand also has an impact on prices; a one standard deviation increase in passive liquidity demand is associated with 0.5 tick rise in prices for S&P 500 E-mini futures. In addition, we find that new information is incorporated into the prices when passive liquidity demanders take positions. By providing direct evidence, we contribute to the growing literature on the impact of passive limit orders.
    Keywords: Liquidity ; Passive Trading ; Price Impact
    JEL: G10 G13 G14
    Date: 2017–10–05
  3. By: Ingemar Kaj; Mine Caglar
    Abstract: We introduce a Markovian single point process model, with random intensity regulated through a buffer mechanism and a self-exciting effect controlling the arrival stream to the buffer. The model applies the principle of the Hawkes process in which point process jumps generate a shot-noise intensity field. Unlike the Hawkes case, the intensity field is fed into a separate buffer, the size of which is the driving intensity of new jumps. In this manner, the intensity loop portrays mutual-excitation of point process events and buffer size dynamics. This scenario is directly applicable to the market evolution of limit order books, with buffer size being the current number of limit orders and the jumps representing the execution of market orders. We give a branching process representation of the point process and prove that the scaling limit is Brownian motion with explicit volatility.
    Date: 2017–10
  4. By: Deng, Xiaohu (Tasmanian School of Business & Economics, University of Tasmania); Gao, Lei (Iowa State University, USA); Kemme, David (Iowa State University, USA)
    Abstract: Using January 2005 – June 2007 trading data for all NYSE stocks we identify the informational patterns and impact of exogenous shocks in short sales and option trades upon stock price changes. We find that short sales have more predictive power than put option trades. However, if short selling volume is low put options trading does have predictive power and thus may be a substitute used by informed investors.
    Keywords: Short selling; Put option trades; Informational patterns; Price discovery
    JEL: G12 G14
    Date: 2017
  5. By: Schasfoort, Joeri; Stockermans, Christopher
    Abstract: The use of fundamentalist traders in the stock market models is problematic since fundamental values in the real world are unknown. Yet, in the literature to date, fundamentalists are often required to replicate key stylized facts. The authors present an agent-based model of the stock market in which the fundamental value of the asset is unknown. They start with a zero intelligence stock market model with a limit-order-book. Then, the authors add technical traders which switch between a simple momentum and mean reversion strategy depending on its relative profitability. Technical traders use the price to earnings ratio as a proxy for fundamentals. If price to earnings are either too high or too low, they sell or buy, respectively.
    Keywords: Agent-based modelling,financial markets,technical and fundamental analysis,asset pricing
    JEL: C63 D53 D84 G12 G17
    Date: 2017
  6. By: Ulf Lewrick; Jochen Schanz
    Abstract: Open-end mutual funds expose themselves to liquidity risk by granting their investors the right to daily redemptions at the fund's net asset value. We assess how swing pricing can dampen such risks by allowing the fund to settle investor orders at a price below the fund's net asset value. This reduces investors' incentive to redeem shares and mitigates the risk of large destabilising outflows.Optimal swing pricing balances this risk with the benefit of providing liquidity to cash-constrained investors. We derive bounds, depending on trading costs and the share of liquidity-constrained investors, within which a fund chooses to swing the settlement price. We also show how the optimal settlement price responds to unanticipated shocks. Finally, we discuss whether swing pricing can help mitigate the risk of self-fulfilling runs on funds.
    Keywords: Financial stability, mutual funds, regulation, liquidity insurance, trading frictions
    JEL: G01 G23 G28 C72
    Date: 2017–10
  7. By: Hautsch, Nikolaus; Voigt, Stefan
    Abstract: We propose a Bayesian sequential learning framework for high-dimensional asset al-locations under model ambiguity and parameter uncertainty. The model is estimated via MCMC methods and allows for a wide range of data sources as inputs. Employing the proposed framework on a large set of NASDAQ-listed stocks, we observe that time-varying mixtures of high- and low-frequency based return predictions significantly improve the out-of-sample portfolio performance.
    JEL: C52 C11 C58 G11
    Date: 2017

This nep-mst issue is ©2017 by Thanos Verousis. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.