Segment 1 — The Cold Open

In the spring of 1966, a new federal agency began writing standards that would force automakers to install lap belts, padded dashboards, and other features in every car sold after 1968. The explicit goal was to cut the roughly 50,000 annual highway deaths recorded in the United States at the time. Instead, economist Sam Peltzman later found that drivers appeared to treat the added protection as a license to change their behavior behind the wheel.

Segment 2 — The Good Intention

The push came from a coalition that included physicians, state motor-vehicle officials, and consumer advocate Ralph Nader, whose 1965 book Unsafe at Any Speed documented how vehicle design contributed to injury severity. They were responding to a problem that had grown steadily worse since the 1920s: rising vehicle miles traveled and rising fatalities, with little regulatory attention to the car itself. The 1966 National Traffic and Motor Vehicle Safety Act created the National Highway Safety Bureau (later NHTSA) and gave it authority to issue Federal Motor Vehicle Safety Standards. At the time, the dominant engineering view held that most serious injuries resulted from the “second collision”—the occupant striking the interior—rather than the initial impact. Requiring belts, energy-absorbing steering columns, and stronger door latches therefore seemed a straightforward way to reduce harm without asking drivers to change habits. The people writing the rules were not trying to alter behavior; they were trying to make the existing behavior less lethal. They focused on measurable physical forces inside the passenger compartment because those forces could be modeled in laboratories and written into standards that applied uniformly across manufacturers.

Segment 3 — The Implementation

The first standards took effect for 1968 model-year cars. Lap belts became standard equipment, followed by shoulder belts in 1968 and dual front airbags mandated in the 1998 model year after years of regulatory battles. Early data showed clear gains: the National Highway Traffic Safety Administration estimated that lap-and-shoulder belts reduced fatality risk by 45–50 percent when used. Proponents pointed to these engineering margins in congressional hearings and press releases. A few skeptics, including some economists, warned that drivers might respond to perceived safety by driving differently, but those cautions received limited attention amid broad bipartisan support for the new rules. The rollout occurred through successive model years rather than a single overnight mandate, allowing manufacturers to phase in energy-absorbing steering columns, reinforced door beams, and improved windshield retention while regulators tracked compliance through production audits.

Segment 4 — The Unintended Consequences

In 1975 Sam Peltzman published “The Effects of Automobile Safety Regulation” in the Journal of Political Economy. Using state-level data before and after the 1968 standards, he found that the reduction in occupant deaths per accident was smaller than the engineering estimates had predicted, while deaths among pedestrians, cyclists, and motorcyclists rose. Peltzman interpreted the pattern as evidence that drivers offset some of the safety gain by taking greater risks—driving faster, following more closely, or paying less attention. A later field study in Munich examined taxis equipped with anti-lock braking systems; drivers using the ABS-equipped cars braked later in traffic and maintained shorter headway distances than drivers in conventionally braked taxis. The offsetting behavior did not erase the benefits of the devices—NHTSA continues to attribute tens of thousands of lives saved to belts and airbags—but it narrowed the net effect. Secondary consequences included pressure on insurers to offer discounts for cars with new safety features, which in turn encouraged further adoption, and a gradual shift in enforcement priorities toward speed and impairment rather than vehicle design alone. The causal chain begins with the driver’s updated perception of personal risk; once the probability of severe injury in a given crash declines, the marginal cost of speed or inattention falls, and some drivers reallocate that margin into faster travel or reduced vigilance. Because the adjustment occurs at the level of individual trip decisions rather than aggregate statistics, the offset appears only when data are examined across many drivers and many miles rather than in controlled crash tests. The rise in non-occupant fatalities suggests that the extra risk was not contained inside the vehicle; higher speeds and closer following distances increased the kinetic energy transferred to anyone struck outside the car.

Segment 5 — The Aftermath

Regulators did not abandon the standards; instead, they incorporated behavioral adaptation into later designs. Electronic stability control, introduced in the 1990s and later mandated, was engineered to intervene even when drivers took larger risks. Seat-belt reminder systems and ignition interlocks were refined rather than abandoned. Today most new vehicles carry five-star crash ratings and multiple airbags, yet traffic deaths per mile traveled remain far lower than in the 1960s. The partial offset identified by Peltzman is now treated as a design constraint rather than a surprise. Engineers routinely test new systems under the assumption that drivers will adjust their behavior once they sense added protection. When stability control was added, for example, the systems were calibrated to correct yaw even if drivers entered curves at speeds they would not have attempted without the technology, effectively closing one channel through which risk compensation could operate.

Segment 6 — The Lesson

Risk compensation does not mean safety measures are pointless; it means their net effect depends on how people actually respond once the technology is in place. When designing incentives or protections, it helps to measure downstream behavior rather than rely solely on laboratory or engineering projections. The same principle appears in helmet laws, cybersecurity defaults, and public-health interventions: the first-order technical gain can be altered by second-order human adjustment. How might current efforts to automate driving account for the ways drivers will test the limits of the new systems once they feel the car is watching the road?

💬 Reply to this email — Patrick reads every one.

Share: X · LinkedIn · WhatsApp

Forwarded this email? Subscribe here — it's free.