This study examines the relation between text-based firm complexity and labor performance.
Using 11,568 firm-year observations of US publicly listed firms from 2000 to 2024, we measure labor performance as the sum of net income and labor cost scaled by the number of employees. We measure text-based firm complexity using complexity-related language in firms' 10-K filings and estimate its association with labor performance.
Text-based firm complexity is positively and statistically significantly associated with labor performance. This association remains robust when using an alternative labor performance measure that excludes labor expense, lagged measures of firm complexity, a reduced set of controls excluding profitability-related variables and sales per employee as an alternative employee performance proxy. The positive relation also persists across subperiods and across both high- and low-labor-cost industries. Additional analyses using first-difference specifications, instrumental-variable estimation and propensity score matching provide further supportive evidence, although these results are interpreted cautiously because they do not fully eliminate endogeneity concerns.
Our study provides evidence that firms with greater text-based complexity exhibit higher labor performance. The findings suggest that firms operating in more complex disclosed environments may possess or develop internal information systems, control processes and coordination mechanisms that support labor productivity. By linking text-based firm complexity to labor performance, our study extends prior research on firm complexity beyond external information environments and capital-market outcomes to internal operating performance.
