![[2410.17587] Predicting Company Growth using Scaling Theory informed Machine Learning](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2410.17587] Predicting Company Growth using Scaling Theory informed Machine Learning
Summary
The paper presents a novel Scaling-Theory-Informed Machine Learning (STIML) framework for predicting company growth by integrating structural growth models with data-driven forecasting, analyzing financial data from over 31,000 North American companies.
Why It Matters
Understanding company growth dynamics is crucial for investors and policymakers. This research offers a new approach that combines theoretical and empirical methods, potentially improving predictive accuracy and informing strategic decisions in finance and economics.
Key Takeaways
- STIML framework integrates scaling theory with machine learning for growth prediction.
- The model captures both trend-driven and fluctuation-driven growth dynamics.
- Company size and volatility significantly influence predictability.
- Macroeconomic variables contribute less to predictive performance than expected.
- The study suggests refining mechanistic models to account for asymmetric deviations.
Computer Science > Computational Engineering, Finance, and Science arXiv:2410.17587 (cs) [Submitted on 23 Oct 2024 (v1), last revised 14 Feb 2026 (this version, v2)] Title:Predicting Company Growth using Scaling Theory informed Machine Learning Authors:Ruyi Tao, Veronica R. Cappelli, Kaiwei Liu, Marcus J. Hamilton, Christopher P. Kempes, Geoffrey B. Wes, Jiang Zhang View a PDF of the paper titled Predicting Company Growth using Scaling Theory informed Machine Learning, by Ruyi Tao and 6 other authors View PDF HTML (experimental) Abstract:Predicting company growth is a critical yet challenging task because observed dynamics blend an underlying structural growth trend with volatile fluctuations. Here, we propose a Scaling-Theory-Informed Machine Learning (STIML) framework that integrates a scaling-based growth model to capture the mechanism-driven average trend, together with a data-driven forecasting model to learn the residual fluctuations. Using Compustat annual financial statement data (1950--2019) for 31,553 North American companies, we extend the growth model beyond assets to multiple financial indicators, and evaluate STIML against growth model-only and purely data-driven baselines. Across 16 target variables, we show that company growth exhibits a clear separation between trend-driven predictability and fluctuation-driven predictability, with their relative importance depending strongly on company size and volatility. Interpretability analyses further show that STIML...
