Masking Causality and Conditional Dependence

TL;DR

This paper investigates the limitations of using averaged statistical constraints to enforce causal masking in decision policies, revealing that such methods often fail to meet stratum-specific requirements and are difficult to detect.

Contribution

It formulates causal masking as a linear program, demonstrating the limitations of averaged constraints and highlighting the need for decision rule-level enforcement.

Findings

Averaged constraints often violate stratum-wise requirements.

Masked policies can recover most reward while being hard to detect.

Regulating direct dependence via averaged statistics is structurally limited.

Abstract

Many regulatory and analytic problems require that a prohibited variable influence a decision only through a designated allowable channel -- a conditional-independence requirement that arises in path-specific fairness, the handling of classified information, and the regulation of trading on non-public information, among other settings. Such requirements may be enforced either stratum-by-stratum or, more commonly (and more efficiently), through a single averaged constraint on the conditional effect. We study the resulting enforcement problem from two perspectives. From the regulator's side, we formulate causal masking as a linear program and show that averaged-constraint optimization almost surely produces policies that violate the stratum-wise requirement while satisfying the averaged one exactly. The gains from masking grow with confounding and outcome heterogeneity, and detection requires precisely the conditional-independence tests that average constraints aim to avoid. From the optimizer's side, the same construction shows that masked policies recover most of the reward of unconstrained exploitation while being far harder to detect, making them attractive in any setting where the basis of decisions is itself sensitive. Together, these results argue that regulating direct dependence through averaged statistics on observed decisions is structurally limited, and that meaningful enforcement must operate at the level of the decision rule itself.