Optimal Execution and Macroscopic Market Making

Ivan Guo, Shijia Jin · 2025

Read Optimal Execution and Macroscopic Market Making by Ivan Guo, Shijia Jin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-06717.txt.

Introduction

Ivan Guo and Shijia Jin wrote this 2025 mathematics paper on the game between market makers and traders. In classic models, traders move prices from outside, and market makers take order flow as given. In real markets, each side reacts to the other, and the authors model that loop.

They build a stochastic game where price impact comes from the quotes makers choose, and order flow comes from the plans traders choose. They describe the balanced outcomes with a system of forward-backward stochastic equations. For a standard named model, they prove the solution exists across the whole horizon. Computer runs show quotes and strategic orders moving in opposite directions, unlike quotes and noise orders. The work joins execution and market making in one frame.

Catalog record

CREATORIvan Guo, Shijia Jin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE122 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-06717

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Optimal Execution and Macroscopic Market Making Thanks: Ivan Guo, Email: [email protected], Address: Centre for Quantitative Finance and Investment Strategies, School of Mathematics, Monash University, Wellington Rd, Clayton VIC 3800, Australia Thanks: Corresponding author: Shijia Jin, Email: [email protected], Address: School of Mathematics, Monash University, Wellington Rd, Clayton VIC 3800, Australia Ivan Guo, Shijia Jin Abstract . We propose a stochastic game modelling the strategic interaction between market makers and traders. From the trader’s perspective, the conventional exogenous permanent price impact is replaced by the endogenous quoting strategies of the market makers. Conversely, from the market maker’s perspective, order flows are no longer assumed to be exogenous, but are driven endogenously by the strategic traders. Characterizing the Nash equilibria via forward-backward stochastic differential equations (FBSDEs), we establish a local well-posedness result for the general game. For the specific ‘Almgren-Chriss-Avellaneda-Stoikov’ model, the decoupling approach guarantees the global well-posedness of the FBSDEs by reducing it to a backward stochastic Riccati equation with M + M_{+} -matrix coefficients. Finally, by introducing small diffusion terms into the inventory processes as an approximation to the general game, we establish its global well-posedness. Simulations reveal a negative correlation between quotes and strategic orders, in contrast to the positive correlation observed between quotes and noise orders. Keywords: Optimal execution; Market making; Stochastic differential game; Forward-backward stochastic differential equation; Decoupling field; Riccati equation 1. Introduction From the liquidity-taking side, this paper formulates and solves a novel trade execution problem: how to design a trading schedule to liquidate a large position efficiently, when the trade impact is determined by market makers. We focus on a macroscopic time horizon (minutes to hours), so trading schedules can be represented by trading rates. From the liquidity-provision side, the paper also addresses a new market making problem: how to dynamically quote limit orders in the presence of competing market makers and strategic order flows. In summary, we propose a unified game-theoretic framework that links optimal execution and market making through strategic interactions. Because of the time scale, we follow the macroscopic framework of [ 23 ] and [ 22 ] , which adapts the seminal market-making model of [ 4 ] to a macroscopic setting. Both works study market making problems with noise traders only, excluding strategic traders. Consequently, our model emphasizes quote-driven markets and order-driven markets where the bid-ask spread is large relative to the tick size. See [ 23 ] for further details on the macroscopic model. Introduced by [ 1 ] and [ 5 ] , the optimal execution problem arises from the trade-off between trading slowly to reduce market friction costs and trading quickly to avoid adverse price fluctuations. The framework rests on two pillars: (1) a permanent price impact that embeds trading activity into price dynamics, and (2) a temporary price impact that captures instantaneous frictional costs, such as the cost of ‘walking the book’ when submitting large orders over short intervals. Subsequent work has explored many variants of these impact functions. Nonlinear impact models are studied in [ 2 ] . Linear impact functions in game settings have attracted recent attention; see [ 13 ] and [ 18 ] for probabilistic approaches, and [ 7 ] and [ 27 ] for PDE methods. Empirical evidence also points to transient components of permanent impact, which are examined in [ 33 ] and [ 37 ] . For the optimal execution problem, our main contribution is the introduction of an endogenous permanent impact . Permanent impact captures how trading activity affects prices, and prices in turn depend on how market makers requote their limit orders partially. Motivated by this feedback loop, we model a game between liquidity takers and liquidity providers, in which permanent impact is not imposed exogenously but emerges from market makers’ quoting strategies. Specifically, if g is some permanent impact functional and \delta is a quoting strategy, we replace the classical term P_{t}+g(v_{[0,t]})\text{ \; by \; }P_{t}+\delta_{t} in market-price modelling. Here, the process P stands for a reference price and v represents a trading schedule. From the market-making perspective, this paper introduces a more comprehensive game framework among market makers by incorporating endogenous order flows . Building on the literature initiated by [ 25 ] and [ 4 ] , market making has been modelled as an optimization problem addressing two main issues: diminishing profit margins as transaction frequency increases, and rising inventory risk as positions deviate from zero. While [ 6 ] models more realistic order flows via a hidden Markov chain and [ 28 ] captures clustering and long-memory effects with general Hawkes processes, both treat order flow as exogenous. Recent game-theoretic studies (such as [ 30 ] , [ 15 ] , and [ 22 ] ) analyse how competition among market makers affects the quoting strategies, whereas the order flow continues to be modelled exogenously. Our contribution is to endogenize the order flow so that it interacts strategically with market makers within the game. Our proposed game framework, illustrated in Figure 1 , optimizes the actions of all market makers and of strategic traders. Noise traders with exogenous objectives are included in the model but are not optimized; they form part of the stochastic environment. Treating these flows as exogenous allows us to derive order-flow-driven execution strategies such as the volume-weighted average price. See [ 12 ] for related work on optimal execution with order flows. Figure 1. Three major components in the game Outline: In Section 2 , we begin with the optimal execution game, which sets up the framework for strategic traders. Moreover, while the existing literature predominantly employs individual temporary price impact, with the two-player analysis in [ 39 ] as a notable exception, we adopt a linear aggregated temporary impact to bridge this gap. Our work can be viewed as a generalization of [ 39 ] to an N -player environment (without tracking terms or constraints), and as an extension of execution games (e.g., [ 18 ] , [ 17 ] ) to a setting with joint temporary impact and a more general formulation. Numerical examples show that such joint impact discourages the crowding of orders. Continuing the optimal execution games, Section 3 incorporates the role of market makers. From the traders’ perspective, the permanent price impact component is replaced by market makers’ quoting strategies: market prices are determined by those quotes. Simultaneously, we extend the previous linear specification and introduce a class of aggregate temporary impact functions. On the other hand, following the competition modelling of [ 22 ] , each market maker competes against the best quotes of her rivals. Unlike previous market-making games, order flows here are endogenous, consisting of both noise traders and strategic traders. We apply the stochastic maximum principle to characterize Nash equilibria as a system of forward-backward stochastic differential equations (FBSDEs) and then establish a local well-posedness result. In Section 4 we analyse the ‘Almgren-Chriss-Avellaneda-Stoikov’ model under three specifications: (1) one strategic trader and two market makers, (2) a linear temporary impact for the trader as in [ 1 ] , and (3) exponential intensities for market-making competition as in [ 4 ] . Under the conditions above, the associated FBSDE system admits an explicit representation. Using the multidimensional decoupling method in [

Read the complete text