Changelog · how the belief was versioned
Peltzman Effect
The tendency for safety regulations or protective devices to induce people to behave more recklessly, so that some or all of the intended safety gain is offset by riskier behavior.
# README · how it works
The proposed mechanism is a cost-benefit / utility tradeoff: a safety device or regulation lowers the perceived (or actual) cost of risky behavior, so a rational or partly-rational actor 'spends' some of the new safety on benefits they value (speed, convenience, thrill), partially offsetting the intended protection. Peltzman framed this in economic terms as demand for both safety and 'driving intensity'; psychologists frame it as adjustment toward a target risk level. The magnitude depends on whether the actor perceives the safety change, attributes it to themselves, is motivated to take more risk, and controls the behavior.
Release stability — contested. Peltzman's strong 'complete offset' claim is largely rejected: Robertson (1977) found data errors and unstable, non-predictive regressions, and later panel studies (Cohen & Einav 2003) show seat-belt laws clearly cut fatalities with no detectable offsetting driving. The weaker risk-compensation phenomenon replicates well in some domains (ABS taxi drivers; some PrEP observational data) but poorly in others — bicycle-helmet risk compensation has little empirical support and key 'evidence' has been criticized as a false conclusion from uncritical citation (Radun, Radun, Esmaeilikia & Lajunen 2018). Net: real but partial and highly context-dependent, not the universal, complete offset Peltzman claimed.
Issues filed in production
#001 open · 1994
Munich anti-lock brake (ABS) taxi fleet
Professional taxi drivers given ABS-equipped cars drove less cautiously (more abrupt braking, smaller safety margins) and showed no significant reduction in crash rates versus conventionally braked taxis — a real-world demonstration of offsetting behavior. The 1994 write-up is a book chapter: Aschenbrenner KM & Biehl B (1994), 'Improved Safety Through Improved Technical Measures?' in Trimpop RM & Wilde GJS (eds), Challenges to Accident Prevention: The Issue of Risk Compensation Behaviour, pp. 81–89, Styx Publications, Groningen. A related peer-reviewed Oslo study of behavioral adaptation to ABS among taxi drivers is Sagberg, Fosser & Saetermo (1997), Accident Analysis & Prevention 29(3):293–302.
#002 open · 2018
HIV pre-exposure prophylaxis (PrEP) and STI rates among MSM
Meta-analysis of observational studies found PrEP initiation associated with increased condomless sex and higher STI rates; rectal chlamydia reached statistical significance (OR 1.59, 95% CI 1.19–2.13), while the increase in any STI (OR 1.24, 95% CI .99–1.54) was not statistically significant (P=.059). Consistent with partial risk compensation in observational data — though RCT-based analyses generally found no such increase, illustrating how the effect depends on study design.
#003 open · 2003
U.S. mandatory seat-belt laws (counter-case)
A documented case where the predicted Peltzman offset did NOT materialize: panel analysis of all U.S. jurisdictions (FARS 1983–1997) found belt laws reduced occupant fatalities ~4–6% with no significant offsetting driving behavior, undercutting Peltzman's complete-offset claim.
Recommended migration
Treat the offset as an empirical magnitude, not a binary: measure net outcomes (fatalities, infections) directly rather than assuming engineering benefits are fully cancelled, and design interventions that don't make the risk feel free (e.g., keep the safety gain less salient, pair devices with feedback, or target settings where users lack control over the risky behavior).
Cohen & Einav (2003) found seat-belt laws cut fatalities ~4–6% with no significant offsetting behavior, and bicycle-helmet risk compensation has little empirical support — showing that net benefit often survives, contrary to complete-offset assumptions.
Catch it in the act
Watch for arguments that a safety measure 'won't help because people will just be more reckless,' especially when the speaker assumes the offset is total. Real Peltzman effects are partial and depend on whether people notice the safety change, feel in control, and have a reason to take more risk — so the honest question is 'how much is offset?' not 'is it all offset?'
Dependencies
- Risk homeostasis (target risk theory)siblingWilde's (1982) stronger claim that people hold a fixed target risk level so safety gains are fully offset in equilibrium; the Peltzman effect is the economic cousin and does not require full offset.
- Risk compensationparentThe broad behavioral-science category — adjusting behavior to perceived safety — of which the Peltzman effect is the economics-of-regulation instance.
- Moral hazardmechanistically-linkedBoth involve taking more risk when protected, but moral hazard centers on shifting costs to others (e.g., insurers), whereas the Peltzman effect centers on physical safety devices/regulation.
- Behavioral adaptationsiblingTraffic-safety term for any behavior change in response to system changes; risk compensation is the subset where the change erodes a safety benefit.
- Lulling effect / safety complacencyeasily-confusedOften conflated with Peltzman; refers to over-reliance/reduced vigilance from a safety feature rather than a deliberate risk-for-benefit tradeoff.