Does the AI Productivity Dividend Diminish? Does the AI Productivity Dividend Diminish? A Validated Framework for Identifying the Shape of Firm-Level Returns to Artificial Intelligence
Abstract
Forecasts of the productivity dividend from artificial intelligence (AI) differ by an order of magnitude, partly because they extrapolate average gains observed among early, highly exposed adopters. Whether those gains scale linearly, flatten, or are competed away is an empirical question about the shape of the return to AI, not its average. This paper develops and validates a design-based framework for measuring that shape with linked firm-worker data. The framework combines a pre-determined, occupation-based measure of firm exposure to generative AI with a continuous-treatment difference-in-differences design around the release of ChatGPT in November 2022, three pre-specified tests of concavity, a decomposition of exposure effects into an adoption margin and a per-adopter return, a doubly robust analysis of use intensity, a staggered-adoption analysis of earlier AI adopters, and a production-function estimator that places AI in the law of motion of productivity. We validate every component on a calibrated synthetic panel of 20,000 enterprises that reproduces the structure of Statistics Canada's linked business microdata and embeds a known concave effect. The estimators recover the true dose-response curve and detect its concavity (slope difference -0.010, standard error 0.003, one-sided p = 0.002); Monte Carlo coverage of the main confidence interval is 0.97, and the concavity test has power of 0.93. The validation also exposes a practical pitfall: pre-trend tests clustered at roughly 40 industries over-reject a true null. The framework, released as open code, is ready for application to Canadian enterprise microdata, where it will inform estimates of potential output and the persistence of Canada's productivity gap.
Related
- Prioritizing next-day readiness decline in elite female soccer using readiness, workload, and wellness data: A temporally validated and calibrated monitoring framework
- Identifying Representational Biases in Datasets Using PCA: A Max-Disparity Partition Framework
- Causal Inference with Generative Artificial Intelligence: Application to Texts as Treatments
- A statistical framework for identifying subgroup vulnerability to predictive multiplicity in clinical AI
- The Akaike Information Criterion and Artificial Intelligence
BibTeX
@misc{agyekum2026does,
title = {{Does the AI Productivity Dividend Diminish? Does the AI Productivity Dividend Diminish? A Validated Framework for Identifying the Shape of Firm-Level Returns to Artificial Intelligence}},
author = {Louis Agyekum},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.05629},
eprint = {2610.05629},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.05629},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Does the AI Productivity Dividend Diminish? Does the AI Productivity Dividend Diminish? A Validated Framework for Identifying the Shape of Firm-Level Returns to Artificial Intelligence
AU - Louis Agyekum
PY - 2026
DA - 2026/10/04
PB - arXiv
DO - 10.48550/arxiv.2610.05629
UR - https://arxiv.org/abs/2610.05629
AB - Forecasts of the productivity dividend from artificial intelligence (AI) differ by an order of magnitude, partly because they extrapolate average gains observed among early, highly exposed adopters. Whether those gains scale linearly, flatten, or are competed away is an empirical question about the shape of the return to AI, not its average. This paper develops and validates a design-based framework for measuring that shape with linked firm-worker data. The framework combines a pre-determined, occupation-based measure of firm exposure to generative AI with a continuous-treatment difference-in-differences design around the release of ChatGPT in November 2022, three pre-specified tests of concavity, a decomposition of exposure effects into an adoption margin and a per-adopter return, a doubly robust analysis of use intensity, a staggered-adoption analysis of earlier AI adopters, and a production-function estimator that places AI in the law of motion of productivity. We validate every component on a calibrated synthetic panel of 20,000 enterprises that reproduces the structure of Statistics Canada's linked business microdata and embeds a known concave effect. The estimators recover the true dose-response curve and detect its concavity (slope difference -0.010, standard error 0.003, one-sided p = 0.002); Monte Carlo coverage of the main confidence interval is 0.97, and the concavity test has power of 0.93. The validation also exposes a practical pitfall: pre-trend tests clustered at roughly 40 industries over-reject a true null. The framework, released as open code, is ready for application to Canadian enterprise microdata, where it will inform estimates of potential output and the persistence of Canada's productivity gap.
ER -