A Short-term Intervention for Long-term Fairness in the Labor Market

Lily Hu, Yiling Chen

Abstract

The persistence of racial inequality in the U.S. labor market against a general backdrop of formal equality of opportunity is a troubling phenomenon that has significant ramifications on the design of hiring policies. In this paper, we show that current group disparate outcomes may be immovable even when hiring decisions are bound by an input-output notion of "individual fairness." Instead, we construct a dynamic reputational model of the labor market that illustrates the reinforcing nature of asymmetric outcomes resulting from groups' divergent accesses to resources and as a result, investment choices. To address these disparities, we adopt a dual labor market composed of a Temporary Labor Market (TLM), in which firms' hiring strategies are constrained to ensure statistical parity of workers granted entry into the pipeline, and a Permanent Labor Market (PLM), in which firms hire top performers as desired. Individual worker reputations produce externalities for their group; the corresponding feedback loop raises the collective reputation of the initially disadvantaged group via a TLM fairness intervention that need not be permanent. We show that such a restriction on hiring practices induces an equilibrium that, under particular market conditions, Pareto-dominates those arising from strategies that statistically discriminate or employ a "group-blind" criterion. The enduring nature of equilibria that are both inequitable and Pareto suboptimal suggests that fairness interventions beyond procedural checks of hiring decisions will be of critical importance in a world where machines play a greater role in the employment process.

Introduction: The Trap of Statistical Discrimination

The persistence of racial inequality in the U.S. labor market remains a critical challenge, even when firms adopt "group-blind" or ostensibly meritocratic hiring practices. This paper argues that standard "individual fairness" metrics—which focus on a single snapshot of hiring decisions—are often insufficient because they ignore the dynamic feedback loops inherent in the labor market.

The core of the problem lies in statistical discrimination. When firms perceive a specific group as less likely to be qualified, they often set higher hiring thresholds for that group. This "higher bar" reduces the incentive for workers in that group to invest in costly skills or education, as the expected return is lower. This lack of investment then validates the firm's initial biased beliefs, creating a self-reinforcing cycle of disadvantage. To break this cycle, the authors propose a "jolt" to the system: a temporary intervention that shifts the market toward a self-sustaining, fair equilibrium.

Modeling a Dual Labor Market

To analyze these dynamics, the authors develop a reputational model centered on a dual-market structure. The career of a worker is modeled across several stages, from initial investment decisions to entry into different market tiers.

Timeline of the worker's lifecycle and hiring decisions in the dual-market model.

The model distinguishes between two phases:

  1. Temporary Labor Market (TLM): This serves as the entry point. Here, firms are constrained (either by law or algorithm) to ensure statistical parity, hiring groups in proportion to their population size.
  2. Permanent Labor Market (PLM): After gaining experience in the TLM, workers move to the PLM. In this stage, firms are free to hire based on their "best response," selecting candidates based on their individual reputations.

A critical component of this model is Reputational Feedback. An individual worker's success generates a "positive externality" for their group, improving the group's collective reputation. A higher collective reputation, in turn, lowers the cost of skill investment for future generations of that group by improving access to resources, mentorship, and social capital.

Key Insights: Fairness as a Pareto Improvement

The authors use recursive modeling and the Banach Fixed Point Theorem to demonstrate that the TLM constraint leads to a unique, stable, and symmetric steady-state where both groups are treated equally.

Convergence and Efficiency

  • Theorem 3.1 (Convergence): Enforcing statistical parity in the TLM eventually leads to a symmetric equilibrium. Crucially, as group reputations equalize, the "severity" of the intervention naturally fades; the market eventually maintains fairness without the need for permanent constraints.
  • Theorem 3.2 (Pareto Dominance): In markets with "unsaturated demand"—where firms desire more skilled workers than are currently available—the fair equilibrium is Pareto-dominant. It is strictly better than the discriminatory equilibrium because it unlocks the "untapped source" of talented individuals who were previously excluded by artificially high thresholds. This suggests that the perceived trade-off between efficiency and fairness is often a false choice.

Limitations and Future Directions

While the model provides a powerful theoretical framework, it relies on several stylized assumptions. It assumes binary performance outcomes and a simplified two-group society, whereas real-world markets are far more complex. Furthermore, the proof assumes "inherent equality" in the underlying ability distributions of the groups.

The authors also acknowledge that the transition period toward equilibrium may temporarily conflict with certain notions of individual fairness. Future research should explore more complex "pipelines" to identify where interventions are most effective and how "history-aware" algorithmic systems can be designed to avoid perpetuating historical cycles of disadvantage.

Summary updated

Abstract

The persistence of racial inequality in the U.S. labor market against a general backdrop of formal equality of opportunity is a troubling phenomenon that has significant ramifications on the design of hiring policies. In this paper, we show that current group disparate outcomes may be immovable even when hiring decisions are bound by an input-output notion of "individual fairness." Instead, we construct a dynamic reputational model of the labor market that illustrates the reinforcing nature of asymmetric outcomes resulting from groups' divergent accesses to resources and as a result, investment choices. To address these disparities, we adopt a dual labor market composed of a Temporary Labor Market (TLM), in which firms' hiring strategies are constrained to ensure statistical parity of workers granted entry into the pipeline, and a Permanent Labor Market (PLM), in which firms hire top performers as desired. Individual worker reputations produce externalities for their group; the corresponding feedback loop raises the collective reputation of the initially disadvantaged group via a TLM fairness intervention that need not be permanent. We show that such a restriction on hiring practices induces an equilibrium that, under particular market conditions, Pareto-dominates those arising from strategies that statistically discriminate or employ a "group-blind" criterion. The enduring nature of equilibria that are both inequitable and Pareto suboptimal suggests that fairness interventions beyond procedural checks of hiring decisions will be of critical importance in a world where machines play a greater role in the employment process.
BibTeX
@inproceedings{hu2018short,
  author = {Lily Hu and Yiling Chen},
  title = {{A Short-term Intervention for Long-term Fairness in the Labor Market}},
  booktitle = {The Web Conference},
  year = {2018},
  doi = {10.1145/3178876.3186044},
}
Figure