Informed trading, investor beliefs consensus and volatility: Evidence from the Limit Order Book dynamics during COVID-19 and short-selling ban
Year: 2025
Type: article
Abstract: This study investigates the relationship between short-horizon volatility and two distinct sources of microstructural information: executed order flow, measured by VPIN, and the latent order book structure, proxied by its SLOPE. While VPIN captures the realized trade imbalances, SLOPE acts as a proxy for aggregated "belief consensus." The objective is to systematically compare the relative importance of these mechanisms—realized flow versus latent consensus—as drivers and predictors of market volatility. Using tick-by-tick data and the full limit order book for 32 IBEX-35 constituents during the 2019–2020 period, we employ a multifaceted econometric approach in event-time (volume clock), combining stock-level regressions with random-effects meta-analysis, robust fixed-effects panels (Driscoll–Kraay), conditional-probability tables (CPTs), and stock-level VARs with Granger tests and meta-IRFs. Three main results emerge. First, we find that informed trading has a dual role: it helps build belief consensus in the book (H1a) while simultaneously consuming internal liquidity (depth) (H1b). Second, and most critically, belief consensus is a markedly superior predictor of subsequent volatility than VPIN; Conditional Probability Tables confirm that a high degree of consensus sharply increases the probability of the lowest-volatility state (H2). Third, VAR analysis reveals a unanimous, bidirectional, yet asymmetric loop: belief consensus robustly reduces volatility, while volatility, in turn, erodes consensus (H3). The causal links for VPIN, in contrast, are sporadic and size-dependent. Our results establish a new informational channel, demonstrating that the market's latent belief structure is a more potent and reliable determinant of short-term risk than the realized toxicity of order flow.
Authors: Sandra Ferreruela, Daniel Martín
Institution: Universidad de Zaragoza
Language: English
Subfield: Finance
Domain: Social Sciences
Open Access status: hybrid
Estimated APC paid: $3,080
OpenAlex ID: W4417279797