Team Moral-Hazard Awareness 2.3B.3

Detecting where the incentive architecture rewards gaming over contribution, and installing the structural mechanisms that make honesty the path of least resistance

Published 19 March 2026; revised 28 August 2026.

The Problem Structure Cannot Avoid

The previous two units built the team's cooperation architecture. Engineered and bounded interdependence (2.3B.1) made the terms of cooperation visible and exit genuinely voluntary. Encoded objective law and selective transparency (2.3B.2) encoded enforcement into process and made the team's operations auditable without surveilling its members. On paper, the team's incentives are aligned and the structure protects against coercion.

But incentive structures do not remain aligned on their own. A metric that tracked value creation this quarter will incentivize gaming next quarter. A decision right that was clear in writing becomes ambiguous when stakes rise and speed is demanded. An interface standard that everyone agreed to slips when no one is auditing. And slowly — not through malice, but through the silent accumulation of misaligned incentives — the team becomes a mechanism for converting individual effort into collective extraction.

This is moral hazard — not individual vice, but structural failure. The people on the team are not uniquely prone to exploitation. The structure invites it. A developer can write unmaintainable code if velocity metrics reward throughput and penalize refactoring. A manager can avoid difficult conversations if performance reviews are opaque. A team member can hide failures if the penalty for reporting them exceeds the penalty for silence.

Systemic moral-hazard awareness — one of Resilient Objectivism's (RO) design patterns, constructed in Phase I (1.2.8) — is the design discipline of installing hazard detection as a standard institutional practice, identifying where incentive architecture separates consequences from agents, where metrics have become targets, where risk is being shifted to parties who cannot see it, and where oversight has been captured. At team scale, this means running a quarterly hazard scan, mapping the pathways by which metrics lose their meaning, and installing counter-gaming patterns that harden the structure against drift.

The Four Stages of Capture

Phase I defined the capture pathway (1.2.8) as the four-stage mechanism by which institutional architecture degrades from incentive misalignment through normalization and oversight accommodation to structural capture. At team scale, the same mechanism operates — smaller, faster, and closer to the people it affects.

Stage 1: Incentive misalignment. The metric does not perfectly track the value it was designed to serve. A sales team is rewarded for deal volume, not profitability. An engineering team is rewarded for lines shipped, not code quality. The misalignment is visible to anyone who looks, but because the metric and the underlying value have not yet decoupled, it is treated as a technical problem to solve later.

Stage 2: Normalization. People discover they can optimize for the metric faster than they can optimize for the underlying value. Gaming becomes normal — discussed in meetings, praised as cleverness, taught by senior members. The gap between the metric and its purpose widens, but the system does not yet see this as failure because the metric is improving.

Stage 3: Oversight accommodation. The people responsible for monitoring notice the drift. Metrics hit targets, but underlying value declines. They have three options: fix the incentive structure (hard), acknowledge the failure (costly), or accommodate the drift and lower the bar. Most choose the third. The oversight mechanism stops detecting failure and starts normalizing it.

Stage 4: Structural capture. The team's purpose has been fully redirected. The system produces the value it is incentivized to produce, not the value it was designed to produce. Everyone knows this. The original purpose is still stated in the mission, but the structure makes it impossible to pursue. The capture is invisible to outsiders because the metrics are being met.

The structural implication: capture can be interrupted at stages 1 or 2 by installing counter-gaming patterns. By stages 3 and 4, repair requires structural replacement. Early detection is the cheapest intervention.

The Goodhart Map

Phase I admitted Goodhart's Law — when a measure becomes a target, it ceases to be a good measure (1.2.8) — as one of the diagnostic patterns grounding systemic moral-hazard awareness. The Goodhart map (2.3B.3-1) is the team-level tool that operationalizes that diagnostic quarterly.

The map works in three steps. First, identify the team's high-stakes metrics — those tied to compensation, evaluation, promotion, or resource allocation. A metric is hazardous when people are evaluated based on it, when optimizing for it without serving the underlying value is possible, and when gaming is hard to detect. In a typical team, three to five metrics meet these criteria.

Second, map the gaming pathways for each high-stakes metric. For each, ask: How could someone satisfy this metric while failing its underlying purpose? Velocity gaming inflates estimates and marks work done before integration. Quality gaming narrows defect definitions. Delivery gaming cuts scope to hit dates. Each pathway shows where the structure is exposed.

Third, install counter-gaming patterns — six structural mechanisms drawn from Phase I's construction (1.2.8) that make honest behavior cheaper than gaming:

Randomized audits defeat gaming that depends on predictability. Audit velocity not only at sprint's end but mid-sprint, mid-project, months later.

Adversarial testing checks whether the metric still tracks the underlying value. An independent team audits for defects using a different methodology.

Capped leverage limits the payoff for gaming. No single metric determines more than forty percent of compensation, promotion, or evaluation.

Separation of duties prevents collusion. The team that ships features does not set the velocity target. The team that writes code does not write quality standards.

Disclosure rules raise the cost of hidden gaming. Publish how metrics are calculated, what assumptions are made, when definitions change.

Compensation clawbacks extend the payoff period beyond the gaming window. Quarterly bonuses based on velocity are not paid fully until code is stable twelve months later.

The Goodhart map runs quarterly. At each cycle: are these still the right metrics? Have gaming pathways changed? Are counter-gaming patterns holding? And critically — should any metrics rotate out of focus to prevent them from becoming sacred?

Capture Resistance

Once counter-gaming patterns are installed, the structural test is the capture resistance test: If the most powerful person on this team left tomorrow, would the accountability mechanisms still function?

The test asks four questions. Can anyone suppress audit results? If yes, the structure is captured — randomized audits are useless if results can be hidden. Can metrics be redefined in private? If yes, accountability is theater — publish when metrics change and require written reasoning. Would the team still report gaming if it made the team look bad? If no, the incentive to hide gaming has not been structurally removed. Could a faction capture the oversight function itself? If yes, separate the people who oversee accountability from those who benefit from misaligned metrics.

Each "yes" identifies a coercive power path hiding inside the team's structure. Each "no" confirms a structural defense. The test connects to Phase I's strain condition on politics under concentrated institutional failure (1.3.9) — capture resistance at team scale is the smallest version of the load case Phase I identified at institutional and civilizational levels.

Culture as Team Signal

A team's cultural signals — the stories it tells about itself, what it celebrates, what language it uses to describe success and failure, and what behaviors are informally rewarded — carry structural information about the state of its incentive architecture, ahead of what quantitative metrics register.

A team whose formal metrics reward quality but whose informal culture celebrates speed is showing aesthetic drift — the cultural signal diverges from the structural commitment. A team that celebrates "creative workarounds" is normalizing gaming. A team that treats compliance as bureaucratic overhead is signaling that encoded objective law has not been internalized as structural integrity.

The cultural signal audit asks three diagnostic questions: What does this team celebrate? (The answer reveals what is actually valued, regardless of what the metrics measure.) What language does the team use to describe success? (If "hacking the process" or "finding shortcuts" carries positive connotation, gaming is culturally endorsed.) What behaviors are informally rewarded — through recognition, attention, or social capital — that the formal structure does not measure? (Informal rewards that contradict formal metrics are the early-warning system for stage 2 normalization.)

Cultural signal auditing does not replace the Goodhart map. It provides the qualitative complement — the leading indicator that detects normalization before the quantitative metrics register it.

The Team Evidence Pack

The team evidence pack (2.3B.3-3) closes the sub-arc. It assembles the artifacts produced across all three team-level units into a single, structurally organized collection governed by the principle that runs through all of Arc 3: processes public by default; persons private by default.

The pack has two layers. Private artifacts are the team's own records — Goodhart map assessments, counter-gaming pattern audits, metric rotation logs, hazard scan worksheets, capture resistance test results, plus the charter amendments, decision audit logs, transparency matrix reviews, encoded objective law records, and entry/exit records from prior units (2.3B.1, 2.3B.2). Their purpose is accountability to self: is the team actually practicing what it claims, or has it drifted into performing the vocabulary?

Public attestations are outcomes that external parties can verify. Turnover and retention data: do departing members report fair treatment? Quality stability: does shipped quality match claimed metrics? Metric honesty: can external auditors identify gaming pathways, or are they absent? Decision auditability: do decisions follow published criteria with traceable reasoning?

The evidence pack is assembled incrementally — every discipline that produces an artifact feeds the pack. No additional documentation burden. Maintenance cadence is quarterly, synchronized with the Goodhart map cycle. The failure mode is pack theater: maintaining artifacts without maintaining the practices they represent. Stale artifacts are the signal that practice has drifted.

What This Unit Completes

This is the last unit in Sub-Arc 3B. Across three units, the team-level practice has been fully specified. Engineered and bounded interdependence (2.3B.1) built the structure — roles, decision rights, interfaces, consent validity, clean exit. Encoded objective law and selective transparency (2.3B.2) encoded enforcement — checklists, signatures, transparency matrices, dispute path. Moral-hazard awareness (this unit) installed the maintenance discipline — hazard scanning, counter-gaming, capture resistance, cultural signal auditing, and the evidence pack that proves it all.

The three units form a load path: structure → enforcement → maintenance. A team that has built all three carries a cooperation practice layered on top of Traditional Objectivism's (TO) character (Arc 1) and Open Objectivism's (OO) method (Arc 2) — one that can charter cooperation, enforce agreements, detect incentive failures, prove its practice, and do all of it without exposing the persons who make the team work. What remains is making these disciplines hold across organizations and enterprises — Sub-Arc 3C's work.