Tech Shocks and Budget Optimization: The Impact of Generative AI on Corporate R&D and G&A Allocation

Abstract

This thesis investigates the impact of the 2022 release of ChatGPT and other LLMs on corporate budget allocations, specifically analyzing the shift from general and administrative (G&A) expenses towards innovation-driven research and development (R&D) expenses. To go about this, this study uses a panel dataset of 68 US firms evenly split amongst the S&P 500 IT and S&P 500 Consumer Staples indices for the years 2019-2025 while conducting a Difference-in-Differences framework with Two-Way Fixed Effects regressions on AI engagement metrics. The results reveal that strategic commitment to AI within firm 10ks is positively and significantly associated with higher R&D to G&A ratios, and more specifically that each additional mention of AI or related keywords in firm 10ks corresponds to a 0.2% increase in the R&D to G&A ratio following the 2022 shock. While the sector model (IT vs. Consumer Staples) showed insignificant results, the regressions suggest that firm size is a strong proxy for AI engagement, with larger firms leading the way in this movement towards innovation spending. Ultimately, the findings support the classification of AI as a modern General Purpose Technology that will only continue to create innovation in the years to come.