The Role of Causal Features in Strategic Classification for Robustness and Alignment

António Góis, Sophia Günlük, Nir Rosenfeld, Nidhi Hegde, Simon Lacoste-Julien, Dhanya Sridhar
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:2368-2376, 2026.

Abstract

In strategic classification, an institution (e.g., a bank) anticipates adaptation from users who change their features to increase utility in a classification task (e.g., loan repayment). Since a key challenge is the distribution shift induced by users, we turn to causal models, which have been shown to bound the worst-case out-of-distribution (OOD) risk, and establish several new results that link causality and strategic classification. First, we show that causal classification leads to optimal classification error after any sufficiently large adaptation, when the noise is bounded in a certain way. Second, when these assumptions do not hold, we show OOD cross-entropy risk of optimal classifiers decomposes into an OOD bias term and a term arising from not using all observable features, allowing us to understand when causal classifiers have an advantage. Finally, we show that the use of causal features can allow alignment of long-term incentives between institutions and users, contrasting with previous work that highlights social costs of such approaches. We validate our theory empirically on synthetic data, finding that our results predict behavior in practice.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-gois26a, title = { The Role of Causal Features in Strategic Classification for Robustness and Alignment }, author = {G{\'o}is, Ant{\'o}nio and G{\"u}nl{\"u}k, Sophia and Rosenfeld, Nir and Hegde, Nidhi and Lacoste-Julien, Simon and Sridhar, Dhanya}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {2368--2376}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/gois26a/gois26a.pdf}, url = {https://proceedings.mlr.press/v300/gois26a.html}, abstract = { In strategic classification, an institution (e.g., a bank) anticipates adaptation from users who change their features to increase utility in a classification task (e.g., loan repayment). Since a key challenge is the distribution shift induced by users, we turn to causal models, which have been shown to bound the worst-case out-of-distribution (OOD) risk, and establish several new results that link causality and strategic classification. First, we show that causal classification leads to optimal classification error after any sufficiently large adaptation, when the noise is bounded in a certain way. Second, when these assumptions do not hold, we show OOD cross-entropy risk of optimal classifiers decomposes into an OOD bias term and a term arising from not using all observable features, allowing us to understand when causal classifiers have an advantage. Finally, we show that the use of causal features can allow alignment of long-term incentives between institutions and users, contrasting with previous work that highlights social costs of such approaches. We validate our theory empirically on synthetic data, finding that our results predict behavior in practice. } }
Endnote
%0 Conference Paper %T The Role of Causal Features in Strategic Classification for Robustness and Alignment %A António Góis %A Sophia Günlük %A Nir Rosenfeld %A Nidhi Hegde %A Simon Lacoste-Julien %A Dhanya Sridhar %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-gois26a %I PMLR %P 2368--2376 %U https://proceedings.mlr.press/v300/gois26a.html %V 300 %X In strategic classification, an institution (e.g., a bank) anticipates adaptation from users who change their features to increase utility in a classification task (e.g., loan repayment). Since a key challenge is the distribution shift induced by users, we turn to causal models, which have been shown to bound the worst-case out-of-distribution (OOD) risk, and establish several new results that link causality and strategic classification. First, we show that causal classification leads to optimal classification error after any sufficiently large adaptation, when the noise is bounded in a certain way. Second, when these assumptions do not hold, we show OOD cross-entropy risk of optimal classifiers decomposes into an OOD bias term and a term arising from not using all observable features, allowing us to understand when causal classifiers have an advantage. Finally, we show that the use of causal features can allow alignment of long-term incentives between institutions and users, contrasting with previous work that highlights social costs of such approaches. We validate our theory empirically on synthetic data, finding that our results predict behavior in practice.
APA
Góis, A., Günlük, S., Rosenfeld, N., Hegde, N., Lacoste-Julien, S. & Sridhar, D.. (2026). The Role of Causal Features in Strategic Classification for Robustness and Alignment . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:2368-2376 Available from https://proceedings.mlr.press/v300/gois26a.html.

Related Material