Work

Does the Facial Trustworthiness of Hosting Staff Lead to Forecast Bias? Evidence from Analysts’ Site Visits

Year: 2026

Type: article

Abstract: We use unique data about corporate site visits in China to examine whether face‐to‐face interactions introduce bias in analyst forecasts. We find that analysts tend to issue more optimistic forecasts when accompanied by hosting staff with high facial trustworthiness, indicating a cognitive bias. This effect is amplified when firms convey more positive information during visits, but is attenuated for firms with high disclosure quality, for experienced and star analysts, for analysts with science, technology, engineering, and mathematics backgrounds, and for analysts visiting industry peers concurrently. Furthermore, the bias is stronger for sole‐visiting analysts but diminishes over time as additional information becomes available or with repeated interactions. Additional analyses show that forecast accuracy also declines, but the effect of facial trustworthiness does not extend to analysts’ stock recommendations or market reactions. Overall, our findings highlight that even sophisticated analysts are susceptible to cognitive bias in face‐to‐face interactions, while alternative information sources, professional experience, and repeated engagement help mitigate such bias.

Source: Abacus

DOI: 10.1111/abac.70061

Authors: Jiyuan Li, Zihui Li

Institutions: Chongqing University of Technology, University of International Business and Economics

Language: English

Topic: Auditing, Earnings Management, Governance

Subfield: Accounting

Field: Business, Management and Accounting

Domain: Social Sciences

Open Access status: bronze

Funder: National Natural Science Foundation of China

Award: G3544056166