Liquidity shocks and intraday price reaction

• Published date: 01 June 2023
• Author: Tao Chen
• Date: 01 June 2023
• DOI: http://doi.org/10.1111/jfir.12315

Received: 16 January 2022

|

Accepted: 30 November 2022

DOI: 10.1111/jfir.12315

ORIGINAL ARTICLE

Liquidity shocks and intraday price reaction

Tao Chen

Faculty of Business Administration,

University of Macau, Taipa, Macau

Correspondence

Tao Chen, Faculty of Business Administration,

University of Macau, Taipa, Macau.

Email: torochen@um.edu.mo

Funding information

Universidade de Macau,

Grant/Award Numbers: MYRG2020‐00042‐

FBA, MYRG2022‐00008‐FBA

Abstract

Using a global sample of high‐frequency data, I investigate

how liquidity shocks affect intraday price movements. I find

a negative association between liquidity shocks and price

impact. This finding remains robust after considering the

exogeneity of liquidity shocks, using alternative windows

to measure liquidity shocks, and controlling for volume

shocks and volatility shocks. Additional tests show that the

documented relation stems from idiosyncratic shocks and

sell‐order shocks. Moreover, I find that liquidity shocks are

likely driven by uninformed traders. My evidence suggests

thatthemarketrequires30 mintoaccomplishprice

adjustments when meeting liquidity shocks.

JEL CLASSIFICATION

G12, G15

1|INTRODUCTION

Stock‐level liquidity is increasingly crucial to market participants who demand a return premium for undertaking

illiquidity risks (Amihud & Mendelson, 1986). In extreme circumstances, shrinking liquidity presumably triggers a

considerable price plunge across asset classes, which ultimately results in a marketwide crash, such as that during

the 2008–2009 global financial crisis. Given its importance to the stability of financial markets, numerous studies

have been devoted to disentangling the role of liquidity in asset pricing.

A majority of earlier studies have paid close attention to the cross‐sectional differences in liquidity

(Amihud et al., 2015; Bekaert et al., 2007). Consistent with this strand of research, illiquid assets tend to trade

at a discount for high transaction costs, parallel to the perception that average liquidity influences price

formation. In contrast, Bali et al. (2014) take a dissimilar perspective by probing the market reaction to

time‐variant liquidity shocks. In line with their argument, investors are susceptible to misprocessing the signal

embedded in liquidity shocks, eventually leading to market underreaction. Although liquidity shocks have an

impact on long‐term price dynamics (Bali et al., 2014), little is known about their effect on the intraday horizon.

J Financ Res. 2023;46:573–599. wileyonlinelibrary.com/journal/JFIR

|

573

© 2022 The Southern Finance Association and the Southwestern Finance Association.

To fill this void in the literature, I examine whether liquidity shocks affect intraday price impact. Such a

research question is appropriate in the contemporary financial market, wherein technological development

prompts exponential growth in high‐frequency trading (Hasbrouck & Saar, 2013). On this occasion, liquidity is

prone to change rapidly and substantially. Therefore, it is worthwhile to unravel how traders react to liquidity

shocks within such a shorter window.

In addition to the previously mentioned consideration, my article is inspired by a typical theoretical framework

in the market microstructure literature (Easley & O'Hara, 1987; Kyle, 1985). According to these asymmetric

information models, market makers can garner both public and private signals after monitoring order flows in a

persistent manner. When liquidity shocks emerge, their privileged position enables them to observe this

phenomenon quickly and incorporate inferred information into prices. In other words, such a learning mechanism

helps rationalize why intraday liquidity shocks deliver a price impact.

Notwithstanding the theoretical link, I may predict an equivocal association between liquidity shocks and

price impact depending on who is responsible for liquidity shocks. As Rindi (2008) argues, either uninformed or

informed traders are capable of accommodating liquidity needs in the market. Specifically, the literature

attributes liquidity provision to competitive uninformed traders (Glosten & Milgrom, 1985;Glosten,1994),

strategic uninformed traders (Goettler et al., 2005;Rosu,2009), and strategic informed traders (Goettler

et al., 2009;Kaniel&Liu,2006).

Given the positive liquidity shocks provided by uninformed traders, their informed counterparts prefer to

exploit this opportunity to act more aggressively. This is because the highly liquid market renders favorable timing

for informed agents to conceal their actions (Collin‐Dufresne & Fos, 2016). It follows that such informed trades

have a lower price impact. Furthermore, Admati and Pfleiderer (1988) craft a similar model in which a higher noise

trading activity motivates myopic investors with short‐lived information to exercise more trades because of the

lesser price impact. Likewise, the rationale behind this action suggests that positive liquidity shocks are connected

with a weaker price impact. In sum, it is plausible to anticipate a negative association between liquidity shocks and

price impact.

When the market confronts positive informed liquidity shocks, a sudden spike in informed order flow

discourages discretionary uninformed investors from trading (Foster & Viswanathan, 1990). As a result of reduced

liquidity traders, informed agents find it harder to look for the counterpart of a transaction. Under these

circumstances, informed trades are expected to have a greater impact on prices. Moreover, this line of reasoning

bears a resemblance to Dufour and Engle (2000), who ascribe a larger price impact to a lack of uninformed traders

taking the opposite side of a transaction. Combined, these studies imply a positive association between liquidity

shocks and price impact.

In view of the mixed conjectures following previous research, I am inclined to believe that how liquidity shocks

relate to price impact remains an empirical question. Unlike the burgeoning literature on liquidity shocks, to the best

of my knowledge, this article is the first to examine their instantaneous influence on price–liquidity dynamics.

I quantify liquidity shocks by the negative standardized difference between the 5‐min Amihud (2002) illiquidity

measure and its past 5‐trading‐day average. Using a sample of tick‐by‐tick data from 22 countries, I find that price

impact is negatively associated with liquidity shocks. This conclusion is valid when using alternative windows to

construct liquidity shocks and controlling for volume shocks and volatility shocks. Meanwhile, the documented

association depends on stock market size and gross domestic product (GDP). As additional investigations

demonstrate, idiosyncratic shocks and sell‐order shocks contribute to the price impact of liquidity shocks. In

addition, I uncover a stronger negative relation given higher investor sentiment, thereby substantiating that liquidity

shocks may emanate from uninformed traders. Collectively, my findings are compatible with the predictions of

Admati and Pfleiderer (1988) and Collin‐Dufresne and Fos (2016).

In light of the preceding reasoning, uninformed liquidity shocks prompt informed investors to trade more. Put

differently, liquidity shocks always accompany informed trades, which may convey value‐relevant information.

Therefore, I investigate how long it takes to incorporate this information into the market. Based on impulse

574

|

JOURNAL OF FINANCIAL RESEARCH