5 Myths About Zero-Accident Targets That Make Underreporting More Likely

Zero-accident targets are management goals that use an injury count, often zero, as a visible measure of safety performance. The goal may express a legitimate commitment to protect people, yet the number can distort reporting and decision-making when leaders reward a clean record more strongly than they reward honest evidence about exposure.

A production review ends with a green safety slide. No recordable injuries appear for the quarter, so the discussion moves to output, cost, and delivery. Then a supervisor mentions that a contractor stopped a lift after a load shifted, but nobody logged the event because the team did not want to “break the record.” The dashboard is clean. The operating reality is not.

That tension sits at the center of Andreza Araújo’s Far Beyond Zero, a critique of zero accidents as a management target. Across 25+ years leading EHS programs in multinational operations, her work has kept the distinction visible between a value and a score. The value is protecting people. The score is only one imperfect signal about how the system is performing.

Myth 1: A zero-accident target proves that the safety system is working

A zero in the injury column tells leaders that a defined outcome did not enter the reporting system during a defined period. It does not, by itself, show that high-risk work was controlled, that near misses were discussed, or that workers could challenge a weak barrier without career or social cost.

The distinction matters because serious harm is an outcome, while control effectiveness is a condition. A lifting plan can be approved and still be poorly understood. An isolation can be documented and still be unverified. A confined-space rescue plan can exist and still fail the first practical question, which is whether the team can execute it in the time the hazard allows.

James Reason’s work on organizational accidents explains why visible events are only the final expression of deeper conditions. When the dashboard treats the final event as the entire definition of performance, latent weaknesses can remain invisible until the organization gets lucky or someone gets hurt.

Executives should therefore ask what the zero is sitting on. The answer should include verified controls, credible escalation, useful near-miss information, and decisions that reduce exposure before an injury becomes the evidence.

Myth 2: People report honestly when the target is clearly explained

Communication does not cancel incentives. If promotion, bonuses, public recognition, or operational reputation depend on preserving a zero, employees and supervisors receive a practical lesson even when the policy says “report everything.” They learn which behavior is celebrated and which behavior creates friction.

Underreporting rarely begins with a formal instruction to hide information. It can begin with a small social signal, such as a manager asking whether an event is “really recordable,” a meeting that praises a clean month but ignores a weak investigation, or a corrective action that punishes the person who surfaced the problem.

A reporting system becomes credible when bad news receives a faster and more disciplined response than good news receives applause. That means the person who raises a concern can see what happened next, including who owns the decision, what changed, and when the control will be checked again.

The practical test is not whether the organization has a reporting channel. It is whether people use it when the information is inconvenient, incomplete, or likely to challenge a production assumption.

Myth 3: More reporting means worse safety performance

A rise in reports can mean that exposure is increasing, but it can also mean that visibility is improving. Treating every increase as failure teaches the organization to reduce the count instead of improving the work.

Leadership needs to separate at least three questions. Did the hazard become more common? Did the reporting threshold become more reliable? Did the response improve enough to remove the underlying condition? A single trend line cannot answer all three.

This is where leading indicators become useful. A near-miss rate is not automatically positive, and a low rate is not automatically negative. The quality of descriptions, the seriousness of the potential outcome, the speed of escalation, and the closure of corrective decisions reveal more than the count alone.

The same logic applies to safety conversations and control verifications. An organization can report many observations that merely praise compliance, while the most serious exposure remains absent from the record. Leaders should reward useful information, which is information that changes a decision, not activity that only increases the volume of forms.

For a practical distinction between activity and decision value, compare the four interpretations of leading indicators with the thresholds that can make critical-risk exposure look acceptable.

Myth 4: The injury rate is the fairest way to compare leaders

Comparisons feel objective because they produce a ranking. The problem is that injury counts are influenced by exposure, workforce composition, reporting practice, chance, and the classification rules applied after an event. Two operations can show the same result while carrying very different levels of uncontrolled risk.

A leader who inherits a stable operation may have a clean rate without having improved a critical control. Another leader may receive more reports after creating trust, improving field verification, or raising the standard for escalation. If the first leader is rewarded and the second is questioned, the measurement system teaches exactly the wrong lesson.

Fair accountability looks at decisions within the leader’s control. Did the leader fund the engineering change? Did the team have authority to stop work? Were overdue actions escalated? Did supervisors understand which controls were non-negotiable? Were repeated signals connected to a change in work design?

Andreza’s documented PepsiCo result, a 50% reduction in accident ratio in six months under a 180-day plan, should not be read as an argument for chasing a number in isolation. The useful lesson is that a measured outcome becomes meaningful when it sits inside a deliberate operating plan with clear ownership and decisions.

Myth 5: Replacing the target with a longer list of indicators solves the problem

Measurement overload can preserve the same weakness under a more sophisticated appearance. A dashboard with twenty indicators may still reward leaders for explaining the numbers rather than changing the conditions behind them.

A credible set of measures has a visible connection to action. If critical-control verification falls, someone knows what decision follows. If high-risk actions are overdue, escalation is defined. If workers stop raising concerns, leaders examine trust, response quality, workload, and retaliation risk instead of declaring that the operation has become safer.

The board should be able to trace each important indicator to an operating question. What exposure is changing? Which control should interrupt it? Who owns the decision? What evidence will show that the decision worked? When an indicator cannot answer those questions, it may be reporting activity rather than managing risk.

That is also why a safety dashboard should show uncertainty. A zero may mean no event occurred, no event was reported, no event met the definition, or the period was too short to reveal the exposure. Treating those possibilities as identical creates false confidence.

What leaders should do when the score is still zero

Keep the aspiration to protect people, but stop asking the injury count to carry the entire argument. Review the conditions that produce trustworthy information and the decisions that reduce serious exposure.

• Ask which high-consequence exposures were identified this period and how quickly they were escalated.
• Review whether critical controls were verified in the field, not only marked complete in a system.
• Compare reports with overtime, turnover, contractor changes, work interruptions, and unresolved concerns.
• Check whether a worker who raises inconvenient information can see a respectful response.
• Require every leading indicator to have an owner, a decision threshold, and a defined follow-up.

The executive conversation becomes stronger when the number is treated as evidence, not as a trophy. Leaders can then celebrate a clean record without making people feel that reporting a weak signal would damage the organization.

For a related decision framework, see how to audit a leading safety indicator before a board review. Safety performance improves when the measurement system makes reality easier to see and harder to dismiss.