[Paper Review] Cross-Sectional Variation of Intraday Liquidity, Cross-Impact, and their Effect on Portfolio Execution

Seungki Min, Costis Maglaras+1 more|arXiv (Cornell University)|Nov 13, 2018

Financial Markets and Investment StrategiesEconomics, Econometrics and Finance18 references3 citations

TL;DR

This paper proposes a linear cross-asset market impact model that captures cross-sectional intraday variation in liquidity due to portfolio trading strategies, showing that coupled execution—simultaneously optimizing trades across multiple assets—reduces execution costs by up to 6% compared to standard separable VWAP-style execution. The model derives optimal execution schedules that exploit end-of-day portfolio liquidity, driven by ETFs and benchmark-rebalanced mutual funds.

ABSTRACT

The composition of natural liquidity has been changing over time. An analysis of intraday volumes for the S&P500 constituent stocks illustrates that (i) volume surprises, i.e., deviations from their respective forecasts, are correlated across stocks, and (ii) this correlation increases during the last few hours of the trading session. These observations could be attributed, in part, to the prevalence of portfolio trading activity that is implicit in the growth of ETF, passive and systematic investment strategies; and, to the increased trading intensity of such strategies towards the end of the trading session, e.g., due to execution of mutual fund inflows/outflows that are benchmarked to the closing price on each day. In this paper, we investigate the consequences of such portfolio liquidity on price impact and portfolio execution. We derive a linear cross-asset market impact from a stylized model that explicitly captures the fact that a certain fraction of natural liquidity providers only trade portfolios of stocks whenever they choose to execute. We find that due to cross-impact and its intraday variation, it is optimal for a risk-neutral, cost minimizing liquidator to execute a portfolio of orders in a coupled manner, as opposed to a separable VWAP-like execution that is often assumed. The optimal schedule couples the execution of the various orders so as to be able to take advantage of increased portfolio liquidity towards the end of the day. A worst case analysis shows that the potential cost reduction from this optimized execution schedule over the separable approach can be as high as 6% for plausible model parameters. Finally, we discuss how to estimate cross-sectional price impact if one had a dataset of realized portfolio transaction records that exploits the low-rank structure of its coefficient matrix suggested by our analysis.

Motivation & Objective

• To analyze how the rise of passive and systematic investment strategies alters intraday liquidity patterns across stocks.
• To model cross-impact arising from portfolio liquidity providers who trade in baskets of assets.
• To derive optimal execution strategies that jointly schedule trades across multiple securities to minimize market impact costs.
• To quantify the cost advantage of coupled execution over separable execution in the presence of time-varying portfolio liquidity.
• To develop a method for estimating cross-sectional price impact from realized transaction data using low-rank matrix structure.

Proposed method

• Formulates a stylized model where natural liquidity is provided by both single-stock and portfolio traders, leading to a linear market impact model with a coefficient matrix that combines diagonal (single-asset) and low-rank (portfolio) terms.
• Derives a quadratic cost function for portfolio liquidation under risk-neutral, cost-minimizing objectives, with cross-impact terms explicitly modeled via a Woodbury matrix identity for computational efficiency.
• Solves a multi-period optimization problem to determine the optimal coupled execution schedule, which jointly determines trade sizes and timing across assets to exploit end-of-day portfolio liquidity.
• Uses a reduced-form impact model with estimated parameters (γ_id, γ_f,k) from realized transaction data, assuming multivariate normal errors and applying maximum likelihood estimation.
• Applies the Woodbury identity to invert large N×N impact matrices efficiently by inverting smaller K×K matrices corresponding to portfolio weight vectors.
• Extends the model to include practical constraints such as side constraints (only trading in target securities and directions), and shows the optimal schedule remains coupled even under these conditions.

Experimental results

Research questions

• RQ1 How does the intraday variation in cross-sectional trading volume correlations reflect underlying portfolio trading activity?
• RQ2 To what extent does cross-impact from portfolio liquidity providers affect optimal execution strategies for multi-asset portfolios?
• RQ3 What is the cost advantage of coupled execution over separable VWAP-style execution when portfolio liquidity varies intraday?
• RQ4 How can cross-sectional price impact be estimated from realized transaction data using a low-rank structure?
• RQ5 How do practical constraints like side limits affect the structure and performance of optimal execution schedules?

Key findings

• The correlation between volume surprises across S&P500 stocks increases by a factor of two in the last 1–2 hours of the trading day, indicating heightened co-movement due to end-of-day portfolio trading.
• Cross-impact arises naturally from portfolio liquidity providers who trade in fixed baskets of assets, such as market or sector portfolios.
• Optimal execution is inherently coupled: joint scheduling of trades across assets is necessary to exploit end-of-day portfolio liquidity and minimize costs.
• The worst-case cost reduction from using the optimal coupled schedule over a separable VWAP approach can reach up to 6% under plausible model parameters.
• The coefficient matrix of cross-impact exhibits a low-rank structure, enabling efficient estimation from realized transaction data using maximum likelihood with the Woodbury identity.
• Even under side constraints (e.g., only trading in target securities), the optimal schedule continues to benefit from cross-impact effects, though it may deviate from the unconstrained solution.