Does Talking More but Doing Less in Digital Transformation Affect Total Factor Productivity? Evidence from Large Language Models

Applied Economics. 2026.

Liang L., Wang J.

Amid the rapid expansion of digital transformation, understanding the economic consequences of corporate digital catering has become increasingly important. Combining large language model-based textual analysis with financial data, this paper constructs a measure of corporate digital catering and examines its association with firm productivity. Using Chinese A-share listed firms, we find a negative association between digital catering and total factor productivity, a pattern that is generally supported by alternative specifications and measurement checks. Further analysis provides supportive evidence consistent with two potential channels: higher operating burdens and tighter financing constraints. The negative association is more pronounced among firms with higher information asymmetry, weaker external governance, and greater digital transformation difficulty, and in regions with weaker market competition, poorer government–market relations, and less developed digital infrastructure. These findings provide new evidence on the authenticity and economic consequences of digital transformation.

Keywords: совокупная факторная производительность, total factor productivity, цифровая трансформация, digital transformation, большие языковые модели, large language models, digital catering, symbolic disclosure, символическое раскрытие информации, Цифровое приспособленчество

Publication based on the results of: Institutions, human capital and regional Development: historical analysis and contemporary challenges (2025)