Quality
and the Peltzman Effect: When Your Organization’s Safeguards Create the
Risks They Were Designed to Eliminate — and the Inspection You Added to
Prevent Defects Became the Reason Your Operators Stopped Paying
Attention
You added a second inspection station because defect rates were
climbing. The logic was airtight: if one inspector misses something, the
second one will catch it. The investment was modest — a used coordinate
measuring machine and a technician transferred from production. The
results were immediate and impressive. Defect rates dropped 40% in the
first month. Your quality manager presented the data at the monthly
review with justified pride. The slide showed a beautiful hockey-stick
graph plunging toward zero.
Six months later, the customer rejected an entire shipment.
Not for a subtle defect. Not for something that required electron
microscopy to detect. The parts were missing a hole. A hole that had
been on the drawing for eleven years. A hole that every operator on that
line had drilled ten thousand times. A hole that both inspectors — the
original and the new one — had signed off on.
When the quality team investigated, they discovered something
unsettling. The first inspector, knowing that a second inspector would
check his work, had gradually relaxed his standards. The second
inspector, aware that the first inspector had already approved the
parts, had never fully engaged with the process. Together, they had
created something neither could have created alone: an inspection system
that was less effective than a single inspector working without a
net.
This is the Peltzman Effect, and it is quietly undermining quality
systems across every industry on earth.
What Sam Peltzman
Discovered About Safety
In 1975, an economist named Sam Peltzman published a paper that
changed how we think about regulation. He had been studying automobile
safety mandates — specifically, the requirement that all new cars
include seatbelts, padded dashboards, and collapsible steering columns.
The conventional assumption was straightforward: safer cars would lead
to fewer deaths.
What Peltzman found was more complicated and more disturbing. Yes,
cars were safer. Yes, the severity of injuries in any given accident had
decreased. But the total number of accidents had increased. Drivers who
felt protected by seatbelts and padded dashboards drove faster, followed
more closely, braked later, and took more risks than they had before the
safety features were installed. The safety measures hadn’t eliminated
risk — they had relocated it. Some of the risk that drivers shed through
safer cars was absorbed by pedestrians, cyclists, and the drivers
themselves through increased accident frequency.
Peltzman called it “risk compensation.” The human brain maintains a
target level of risk. When you make an activity safer, people
unconsciously push the boundaries until the perceived risk returns to
their comfort zone. The safety equipment doesn’t eliminate the accident
— it changes the behavior that leads to the accident.
If you work in quality, this should make your stomach tighten.
The
Quality System That Protects Itself Into Failure
Every quality system is built on safeguards. Dual inspections,
automated checks, poka-yoke devices, statistical process control,
management reviews, layered process audits, golden samples, color-coded
templates, torque monitoring systems, vision systems, checklists,
sign-offs, and approvals. Each one was added for a good reason. Each one
addressed a genuine failure mode. Each one made sense in isolation.
But the Peltzman Effect operates in the spaces between these
safeguards. It exploits a fundamental feature of human cognition: when
we perceive that a system is protecting us, we allocate less of our own
attention to the task. We don’t do this consciously. We don’t decide to
be careless. The relaxation happens below the level of awareness, in the
same way that you drive closer to the car in front of you when you know
your brakes are excellent.
In a manufacturing environment, this plays out in patterns that most
quality professionals have seen but rarely diagnosed correctly.
The operator who used to measure every critical dimension begins
checking every third part when a vision system is installed. Not because
he’s lazy — because the system is watching. The inspector who used to
follow a twelve-step inspection protocol begins abbreviating it after a
poka-yoke device is added to the line. The supervisor who used to review
every first article begins spot-checking after the SPC charts show
twelve months of stability. Each adjustment is rational. Each adjustment
is small. Each adjustment is invisible.
Until the day the vision system’s camera gets coated with cutting
fluid and nobody notices because the operator stopped looking three
months ago. Until the poka-yoke sensor fails in the “pass” position and
the inspector’s abbreviated protocol doesn’t include the check that
would have caught it. Until the SPC chart’s control limits were
calculated from data that included a subtle shift that everyone
attributed to “normal variation.”
The safeguards didn’t fail. The safeguards changed the behavior of
the humans around them, and the changed behavior created the
failure.
The Hierarchy of Risk
Compensation
Not all safeguards trigger the Peltzman Effect equally. Through two
decades of consulting across automotive, aerospace, and pharmaceutical
manufacturing, I’ve observed a rough hierarchy of risk compensation
intensity.
At the top — the safeguards that trigger the most dramatic behavioral
relaxation — are those that are both highly visible and perceived as
comprehensive. A fully automated inspection system with reject bins and
alarm lights is the classic example. When operators see a machine
catching defects with mechanical precision, their mental model shifts
from “I am responsible for quality” to “the machine is responsible for
quality.” The psychological distance between the operator and the
outcome increases. Quality becomes something that happens to the parts
rather than something the operator produces.
In the middle are layered systems — multiple checks performed by
different people at different stages. These trigger a subtler but
equally dangerous form of risk compensation: the diffusion of
responsibility. When three people are responsible for catching a defect,
each person unconsciously reduces their effort because they expect the
others to compensate. Social psychologists call this the bystander
effect. In quality systems, it means that three inspectors are
collectively less effective than one inspector working alone — exactly
what happened with the missing hole.
At the bottom are safeguards that augment human capability without
replacing human judgment. Go/no-go gauges, for example, don’t typically
trigger significant risk compensation because they don’t change the
operator’s perception of their role. The operator still decides whether
to use the gauge, still interprets the result, still owns the outcome.
The tool extends the operator’s capability without altering their
responsibility.
This hierarchy has profound implications for how you design quality
systems. But before we get there, we need to understand why the Peltzman
Effect is so difficult to detect.
Why You Can’t See It Coming
The Peltzman Effect is invisible to conventional quality metrics for
a specific reason: it doesn’t create new failure modes. It activates
existing ones.
When an operator stops measuring every part and starts measuring
every third part, the quality metrics don’t immediately change. The
vision system is still catching defects. The SPC charts still show
control. The customer still receives good parts. Everything looks fine —
often for months or years.
What’s actually happening is that the system’s margin of safety is
being consumed. The defense-in-depth that you carefully designed is
being quietly dismantled from the inside, not by sabotage or negligence,
but by the natural human tendency to calibrate effort to perceived risk.
Each individual adjustment is invisible. The cumulative effect is
catastrophic.
Think of it like a ship’s hull with multiple watertight compartments.
Each compartment can hold. But the Peltzman Effect is like tiny, slow
leaks through each bulkhead — leaks so small that no individual
compartment ever triggers an alarm. The ship doesn’t sink from one
compartment flooding. It sinks when all compartments have been slowly
filling for months and a single wave pushes the whole system past its
limit.
This is why the failures that result from risk compensation are
always shocking. They always involve “obvious” defects that “should have
been caught.” They always reveal that multiple safeguards failed
simultaneously — not because the safeguards were broken, but because the
humans around them had calibrated their behavior to assume the
safeguards would compensate for any relaxation in their own
attention.
The post-mortem always concludes with some version of “human error.”
And it always misses the point.
The
Automotive Case That Changed My Understanding
A tier-one automotive supplier I consulted with had invested two
million euros in a state-of-the-art automated inspection system for
their cylinder head line. The system used multiple cameras, laser
measurement, and AI-based defect recognition. It was impressive
technology, and for the first year, it worked beautifully.
Then the customer reported a field failure rate that was ten times
higher than the previous year — the year before the automated system was
installed.
The investigation revealed a cascade of risk compensation so elegant
it could have been designed as a case study.
Before the automated system, the line had three manual inspection
points. The operators at each point were experienced, attentive, and
took personal pride in their work. They knew that if a defective part
left their station, it would reach the customer. That knowledge shaped
their behavior every minute of every shift.
After the automated system was installed, management reduced the
manual inspections from three to one — a cost optimization that made
financial sense given the system’s proven accuracy. The remaining
inspector, knowing that the automated system was checking every part,
gradually shifted from a thorough inspection to a confirmatory glance.
The operators on the production line, aware that the automated system
would catch any defects they produced, began running their machines at
higher speeds, pushing tooling beyond recommended change intervals, and
skipping the in-process checks they had previously performed.
The automated system itself was operating at 97% detection efficiency
— an impressive number in isolation. But the manual inspections that had
previously contributed their own detection capability had been
effectively neutralized by risk compensation. The combined system —
automated plus the residual human oversight — was actually less
effective than the purely manual system it had replaced.
Two million euros spent to make quality worse.
The fix wasn’t to remove the automated system. The fix was to
redesign the relationship between the humans and the technology so that
the automated system augmented human capability rather than replacing
human responsibility. We did this by making the operators the owners of
the automated system’s performance — they reviewed its rejection
decisions, they were responsible for its calibration, and their
performance metrics included the accuracy of their override decisions.
The system became their tool rather than their replacement.
Within three months, field failure rates dropped below the
pre-automation baseline.
Designing Against Risk
Compensation
You cannot eliminate the Peltzman Effect. It is a feature of human
cognition, not a bug. What you can do is design quality systems that
account for it.
The first principle is to preserve psychological ownership. Every
person in the quality chain should feel personally responsible for the
quality of the output. This means designing safeguards that make human
judgment more important, not less. Instead of an automated system that
replaces inspection, implement one that flags borderline cases for human
review. Instead of a checklist that an inspector mechanically ticks, use
one that requires a decision at each step. The goal is to keep the human
mind engaged, not to replace it.
The second principle is to measure the measurement system, not just
the product. Most quality metrics track what comes out of the process —
defect rates, scrap rates, customer complaints. These are lagging
indicators. By the time they detect a problem, the behavioral shift that
caused it has been in place for months. You need leading indicators that
track the health of the safeguards themselves. How often does the
automated system reject parts? Is the rejection rate changing? How many
borderline decisions is each inspector making per shift? Are the manual
override rates consistent? These metrics tell you whether the humans
around the safeguards are still engaged or have mentally checked
out.
The third principle is to periodically remove safeguards. This sounds
counterintuitive, but it is one of the most powerful tools available.
Every six months, temporarily disable one layer of your quality system —
with appropriate risk mitigation — and observe what happens. If defect
rates remain stable, your remaining layers are robust. If defect rates
spike, you’ve discovered that one of your other layers has been
compromised by risk compensation, and you’ve found it before it caused a
customer-facing failure. Think of it as a stress test for your quality
system.
The fourth principle is to vary the safeguard structure. Predictable
safeguards create predictable behavioral adaptation. If operators know
exactly when and how each check will be performed, they unconsciously
optimize their effort around those checks. Introducing variation —
random inspection timing, unannounced audit schedules, rotating
inspection protocols — keeps the system fresh and prevents the kind of
passive calibration that the Peltzman Effect thrives on.
The fifth principle is the hardest: accept that more safeguards do
not always mean more safety. Every additional layer you add to a quality
system has a cost, and that cost is not just financial. Each layer
changes the behavior of the humans it touches. Before adding a new
safeguard, ask yourself: how will this change the way people work? What
will they stop doing because this new system is doing it for them? What
margin of safety am I consuming by adding this protection?
The Audit Paradox
The Peltzman Effect is nowhere more visible than in the relationship
between audits and quality performance.
Organizations that prepare intensively for external audits — ISO,
IATF, customer audits — often show a pattern that should trouble every
quality leader. In the weeks before the audit, compliance is meticulous.
Nonconformances are escalated immediately. Procedures are followed to
the letter. The audit goes well. The certificate is renewed. The auditor
leaves.
And within two weeks, the organization relaxes back to its pre-audit
state. Not dramatically. Not in ways that show up on a dashboard. But in
the small, invisible ways that the Peltzman Effect specializes in:
slightly longer intervals between calibration checks, slightly less
thorough first article inspections, slightly more relaxed
interpretations of borderline dimensions.
The audit didn’t improve the quality system. It compressed it into a
temporary state of heightened compliance that was always going to
decompress. The organization’s baseline quality performance — the thing
the audit was supposed to elevate — remained unchanged.
This doesn’t mean audits are useless. It means that the way most
organizations use audits triggers the Peltzman Effect instead of
counteracting it. An audit that is perceived as an external test —
something done to the organization — creates the same dynamic as a
seatbelt: it makes people feel protected without changing their
fundamental behavior. An audit that is perceived as an internal
diagnostic — something the organization does to itself — preserves the
ownership that prevents risk compensation.
The difference is not in the audit itself. It’s in whether the
organization treats the audit as a safety net or as a mirror.
The Leadership Challenge
Here is the uncomfortable truth about the Peltzman Effect in quality
systems: it is primarily created by the people who design those
systems.
When leadership responds to a quality failure by adding a new
safeguard — a new inspection, a new approval, a new check — they are
acting on the same impulse that makes drivers speed up when they get
better brakes. The new safeguard feels like progress. It feels like
control. It feels like the problem has been addressed.
But if the root cause of the failure was a behavioral disengagement
caused by existing safeguards, adding another safeguard makes the
problem worse. You are treating the symptom of the treatment.
The leadership discipline required to counter the Peltzman Effect is
the discipline of subtraction. When a quality failure occurs, the first
question should not be “What safeguard do we need to add?” The first
question should be “What existing safeguard changed the behavior that
led to this failure?” Only when you’ve answered that question are you
ready to design an effective response.
Sometimes the answer is to add a new layer. Sometimes the answer is
to remove one. The wisdom is in knowing the difference.
The Paradox Worth Living
With
You cannot escape the Peltzman Effect. Every quality system you build
will change the behavior of the people who work within it. Every
safeguard you install will be met with an unconscious behavioral
adjustment. Every layer of protection will create its own form of
vulnerability.
But this is not a reason to abandon safeguards. It is a reason to
design them with the same rigor you apply to the products they protect.
A quality system that ignores human behavior is like a bridge that
ignores wind — technically sound in a vacuum, dangerously fragile in the
real world.
The organizations that build the most robust quality systems are not
the ones with the most safeguards. They are the ones that understand the
subtle, invisible, relentless way that human beings adapt to the systems
around them — and that design their safeguards to work with human nature
rather than against it.
The hole that was missing from those parts was never a manufacturing
problem. It was a systems problem. The operators could drill the hole.
The inspectors could see the hole. The technology could measure the
hole. What failed was the invisible web of assumptions, expectations,
and behavioral adaptations that the quality system had woven around
every person on that line.
Fix the system, and the hole will never go missing again. Fix only
the part, and you’ll find a different hole missing next month.
Peter Stasko is a Quality Architect with 25+ years
of experience transforming organizations across automotive, aerospace,
and pharmaceutical industries. He specializes in designing quality
systems that work with human psychology rather than against it — because
the most sophisticated safeguard in the world is useless if it changes
the behavior it was meant to protect.
