Essays on Textual Analysis and Analysts’ Forecasting

Mondal, Ujjal (2024) Essays on Textual Analysis and Analysts’ Forecasting. Doctoral thesis, University of East Anglia.

PDF Restricted to Repository staff only until 31 May 2028. Request a copy

Abstract

This thesis examines the association between the contents of sustainability reports, credit ratings reports, and equity analysts’ forecasting. The thesis consists of two empirical studies. In the first empirical chapter, I examine the association between sustainability reporting quality and equity analysts’ forecast accuracy and dispersion. I develop a dictionary of 337 keywords spanning economic, environmental, social, and governance topics, using manual content analysis of the Global Reporting Initiative (GRI) standards for assessing sustainability reporting (SR) quality. Applying this dictionary to 1,896 sustainability reports of 411 firms from 19 countries issued between 2015 and 2020, I find a positive association between SR quality and equity analysts' forecast accuracy but not forecast dispersion. The association with accuracy is stronger in countries with mandatory SR regulations. My dictionary-based measure of reporting quality subsumes the effect of simply following GRI guidelines used in prior research. By capturing stakeholder perspectives encoded in GRI standards, my measure reflects SR quality in a more transparent and potentially less biased manner compared to proprietary ratings. Overall, my externally validated sustainability dictionary can aid investors, analysts, researchers, and policymakers in assessing corporate SR. In the second empirical chapter, using neural network models, I examine whether net sentiment and readability of credit rating reports are associated with equity analysts’ forecast accuracy and dispersion. My sample includes 9,781 firm-report observations of 1,471 firms, in the US between 2002 and 2019. As predicted, I find that the net sentiment of credit reports is positively associated with analysts’ forecast accuracy and negatively associated with forecast dispersion. I also find limited evidence that the readability of reports is negatively associated with analysts’ forecast dispersion. My analysis of the impact of implementing the Dodd-Frank Act and Rule 17g-7 shows that regulation moderates the association between the net sentiment and readability of credit reports and analysts’ forecast accuracy and forecast dispersion.

Item Type:	Thesis (Doctoral)
Faculty \ School:	Faculty of Social Sciences > Norwich Business School
Depositing User:	Kitty Laine
Date Deposited:	30 May 2025 14:14
Last Modified:	30 May 2025 14:14
URI:	https://ueaeprints.uea.ac.uk/id/eprint/99353
DOI:

Actions (login required)

View Item