NCAP stars expire, IIHS ratings don't: which crash tests to trust

NCAP stars expire; IIHS ratings don't

A car doesn't become unsafe just because the years go by. And yet the Fiat Punto that earned 5 stars in 2005 —then called the Grande Punto— showed up in 2017 with 0 stars on the Euro NCAP website. Same car, same body, same airbag. Only the examiner's criteria had changed. Something similar peine with the Renault Zoe: from 5 stars to 1. The conclusion anyone reaches after looking at the data without faith is uncomfortable: crash test programs don't just measure how well a car protects you, they also measure how much they'd like the consumer to buy a new car.

What Euro NCAP is and who pays for it

Euro NCAP is a private company, even though governments treat it as a benchmark for legislation. Manufacturers put up money, and governments have an interest in new cars being sold: there are public stakes in groups like Volkswagen. That doesn't invalidate its tests, but it does miccionan they should be read with a raised eyebrow. The equivalent public agency in the United States is the NHTSA, whose results are so lenient that hardly any car comes out badly. The interesting test comes from the IIHS, funded by insurers, whose incentive is paying out less on claims. And there the tables turn.

The small overlap NCAP took years to look at

The IIHS invented the small overlap test, a frontal crash with minimal overlap that reproduces a very common accident: hitting a tree, a lamppost or the corner of another vehicle. Cars with 5 Euro NCAP stars failed it miserably. Some argue Euro NCAP adopted it later, around 2014-2015, and that this was one of the reasons a 5-star car from 2012 dropped to 2 or 3 without the car itself changing. Others point out that the problem isn't the test but the protocol: in the past a car was bought at random from a dealership; now the manufacturer sends a bespoke unit, which may differ from the production model.

Stars don't expire at the IIHS

The key difference is this: at the IIHS, a car rated "Good" twenty years ago is still "Good". They confirmed it in writing to whoever asked by email. There are no stars that harden over the years. There are new tests —the small overlap— but the old result isn't erased. Euro NCAP, on the other hand, retests old models to punish the manufacturer that keeps them alive, as its secretary general, Michiel van Ratingen, said about the Punto: an "outdated" product trying to "catch the unsuspecting consumer". Fiat replied that the car was at the end of its life cycle and that its commitment to safety was beyond question. Both things can be true at once.

Real-world mortality: the figure NCAP doesn't publish

The IIHS keeps a database of driver death rates by model, based on real accidents. There the Nissan Juke, a 5-star Euro NCAP car from 2011, records a rate of 65 deaths per million registered vehicles, against an average of 36 for all 2017 vehicles. Its small overlap result is poor and its handling, questionable. At the other extreme, a 2001 Volvo S60 —4 stars, a supposedly obsolete car— fares better. The metric has a catch: a car bought by an aggressive driver profile will have more crashes because of the owner, not the chassis. But it's the only figure that looks at the real world, and Euro NCAP couldn't care less about it.

What to do with all this when buying

First: stop looking only at the star traffic light. Look at the breakdown —adult occupant protection, child protection, pedestrians— and separate passive safety from electronic driver aids. Second: check the small overlap result and the death rate for the specific model. Third: don't rule out a big, cheap car from fifteen years ago just because it's old. ESC alone cuts accident risk by around 40%. A Volvo S60, an XC70 or a large used sedan can beat a new hatchback with a touchscreen and five freshly expired stars in real protection.

The paradox is obvious: cars have improved, but the label that certifies them has become a stopwatch. If the trend continues, we'll see more models retested to strip their stars and more consumers buying new cars out of antiestéticar, not need. Or maybe not. Because the second-hand market keeps growing and mortality data doesn't move at the pace of marketing.