---
title: "Causal machine learning for deciding who gets the policy"
description: "Double machine learning, causal forests, conditional average treatment effects and policy learning appear on 50 recent papers, split between the AI and economics journals, with 14 in the last quarter against about 11 per quarter before. The methods promise to find who benefits most from a programme. Do the targeting rules they produce hold up in new samples, and are they fair?"
url: https://scholar4ai.com/ideas/causal-machine-learning-for-targeting
markdown_url: https://scholar4ai.com/ideas/causal-machine-learning-for-targeting.md
license: CC BY-NC 4.0 (https://creativecommons.org/licenses/by-nc/4.0/)
share_text: "Double machine learning, causal forests, conditional average treatment effects and policy learning appear on 50 recent papers, split between the AI and economics journals, with 14 in the last quarter against about 11 per quarter before."
date: "2026-10-23"
updated: "2026-10-23"
status: "seed"
tags: ["ai", "economics", "methods"]
keywords: ["double machine learning", "causal forest", "conditional average treatment effect"]
---

# Causal machine learning for deciding who gets the policy

Double machine learning, causal forests, conditional average treatment effects and policy learning appear on 50 recent papers, split between the AI and economics journals, with 14 in the last quarter against about 11 per quarter before. The methods promise to find who benefits most from a programme. Do the targeting rules they produce hold up in new samples, and are they fair?

## What prompted this

Since October 2025, 50 papers carry a double-machine-learning, causal-forest, heterogeneous-treatment, uplift, conditional-average-treatment-effect or policy-learning keyword (19 in AI journals, 19 in economics, 8 in finance): "double machine learning" on 13, "causal forest" on 11, "conditional average treatment effect" on 8, "policy learning" on 5. The quarter to September 2026 had 14 such papers against about 11 per quarter earlier.

## Questions

- When targeting rules estimated from one experiment are applied to a new experiment of the same programme, how much of the predicted gain over uniform treatment is realized?
- Do the rules concentrate treatment on groups that are already better off, and how much is lost by adding fairness constraints?
- Are the heterogeneity estimates stable across estimators, or do causal forests and double machine learning disagree about who benefits?

## Data that could help

Public experimental datasets with multiple waves or sites (job training, microfinance and cash transfer experiments), and replication code of the estimators.

## Next step

Estimate targeting rules on one site of a multi-site experiment with two estimators and evaluate them on the other sites.

## Related research activity

*The keywords this note follows, in the tracked finance, accounting, economics, management, operations research and AI journals as of 2026-10-26*

| Keyword | Papers, last 30 days | Rank | Papers per month, Oct 23 – Oct 26* | Papers, last 36 months |
|---|---:|---:|---|---:|
| double machine learning | 2 | #1303 | 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 2, 0, 0, 0, 2, 1, 0, 0, 0, 0, 0, 3, 1, 0, 0, 3, 1, 0, 0, 2, 2, 0, 1, 1, 1, 1, 2 | 24 |
| causal forest | 0 | – | 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 1, 0, 0, 2, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 2, 2, 1, 1, 0, 1, 2, 0 | 18 |
| conditional average treatment effect | 0 | – | 0, 0, 0, 0, 0, 0, 1, 2, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 2, 0, 1, 1, 0, 1, 0 | 13 |

## Cite this page

Scholar4AI. (2026). Causal machine learning for deciding who gets the policy. Journal Research Rankings. Retrieved October 23, 2026, from https://scholar4ai.com/ideas/causal-machine-learning-for-targeting

```bibtex
@misc{scholar4ai-ideas-causal-machine-learning-for-targeting,
  author = {{Scholar4AI}},
  title = {Causal machine learning for deciding who gets the policy},
  year = {2026},
  howpublished = {\url{https://scholar4ai.com/ideas/causal-machine-learning-for-targeting}},
  urldate = {2026-10-23},
  note = {Journal Research Rankings; data from OpenAlex (CC0); licensed CC BY-NC 4.0}
}
```

License: CC BY-NC 4.0 (https://creativecommons.org/licenses/by-nc/4.0/).