Myth: Small cars are unsafe and should be avoided

Fact: Some (but not all) small cars increase your chances of being injured if you’re in a crash. But they don’t make you any more likely to crash, and might even make you less likely. The best way to minimise your risk of being injured is to drive less.

Since the state of Melbourne’s public transport makes it impractical for most people to do without a car entirely, even many sustainable transport activists keep a car handy for occasional use. For this purpose small second-hand cars are attractive since they use less fuel, have lower ‘sunk costs’, and (as long as the supply of second-hand cars exceeds demand for them) don’t increase the energy expended in car manufacture. So while they’re still environmentally ‘bad’, small second-hand cars are less bad than the alternatives when a car is required.

If a new Prius were placed head-to-head with a used car, would the Prius win? Don’t bet on it. Making a Prius consumes 113 million BTUs [140 GJ], according to sustainability engineer Pablo Päster…. Buy a decade-old Toyota Tercel, which gets a respectable 35 mpg [6.7 l/100km], and the Prius will have to drive 100,000 miles to catch up. Better yet, buy a three-cylinder, 49-horsepower 1994 Geo Metro XFi, one of the most fuel-efficient cars ever built. It gets the same average mileage as a 2008 Prius, so a new hybrid would never close the carbon gap.

—Matt Power, Wired, 19 May 2008

So it’s of concern when ‘safety engineers’ and car industry representatives allege that people who drive small cars are putting themselves and their loved ones at undue risk. This, they say, is because when small cars crash, their occupants are injured more often and to a greater degree than the occupants of large cars. Here’s Stuart Newstead, leader of the Used Car Safety Ratings project at Monash University:

Stuart Newstead said small cars were not a safe bet when it came to crash worthiness. No light cars had crash worthiness better than average and of 37 light cars, more than half scored in the worst possible rating, Dr Newstead said. Buyers need to think carefully about a purchase in this class, especially young drivers who have a high crash risk and old drivers who are more susceptible to injury.

Rising fuel prices are also a factor, with some buyers opting for a smaller car believing them to be more economical. People need to be very conscious of the safety issue: you might save yourself a couple of extra dollars a week but what’s your life worth?

4WDs a risk to other road users: survey, The Age, 27 June 2006

A reader would take away the conclusion that all small cars are uncrashworthy: a conclusion that’s falsified by Dr Newstead’s own research. For a start, you have to read carefully to notice that Newstead is only referring to cars in the ‘light’ vehicle category and not those in the ‘small’ vehicle category. For example the popular Toyota Corolla, despite being considered ‘ultracompact’ in the USA, is classified as a ‘small’ car in the Monash study, and all Corolla models produced since 1994 rate better than the all-model average.

Even among ‘light’ cars four models (Daewoo 1.5i, Ford Ka, Kia Rio and Volkswagen Polo) had crashworthiness scores better than the all-model average in 2006, despite Dr Newstead’s comments to The Age. The researchers declined to give a ‘better than average’ rating to these models because the favourable scores were deemed not statistically significant, in part because the models were under-represented in the crash data.

Other models also have crashworthiness ratings at the low or high end of the scale, but their confidence limits overlap the all model average. Although such models may also have superior or inferior crashworthiness characteristics, the database did not contain sufficient numbers of these models for the data to represent scientific evidence that this is the case.

MUARC Vehicle Safety Ratings 2006 Report, p.48

So the real ‘fault’ of these particular light cars isn’t that they do badly in crashes, but that they don’t crash often enough to provide good statistics!

The good crash performance of at least some small cars is acknowledged elsewhere in the same report:

Points in the lower left quadrant [of Figure 1] represent vehicles with relatively low aggressivity as well as good (low) estimated crashworthiness. This area is populated by a number of small, luxury and medium vehicle models as well as some sports vehicles and compact 4WDs.

MUARC Vehicle Safety Ratings 2006 Report, p.58

Subsequent research bears out this conclusion. More recent updates to the Monash report include wording almost identical to the above. Just as in the 2006 report, however, the prima facie better-than-average performance of some small car models is not reflected in the final ‘consumer information’ due to the lack of plentiful crash data for these models. And a report focussing on young drivers takes the anti-light-car bias even further, with an explicit recommendation:

It was also discovered that the sports and light vehicle market groups did not have any vehicles that were significantly safer than average. It is important to actively encourage young drivers to avoid these vehicle market groups per se….

The study results have led to the following recommendations: …. Discourage young drivers from driving unsafe market groups – 4WDs, light cars and sports cars….

Vehicle Safety and Young Drivers, November 2009

Notice the way ‘insufficient evidence to confidently conclude this model is safer than average’ has suddenly become ‘unsafe’. But no matter how one manipulates the data, a better-than-average raw performance (as found for the VW Polo and several other light car models) can never serve as evidence for a worse-than-average rating.

It’s clear that there are many small and light car models with poor crash outcomes. But at worst, this indicates just that one should avoid those particular models; it doesn’t mean one has to avoid an entire class of vehicle. Indeed, in recent years an increasing number of small and light cars have been given the coveted ‘five-star’ rating for new models – including the Toyota Corolla and Yaris, the VW Golf, and the super-light Fiat 500. As a result, safety experts now suggest that in a crash, you’re better off in a new small car than in an old big one. (Though of course it’s not actually necessary to buy these cars brand new to benefit from features such as curtain airbags and electronic stability control.)

Probably the most comprehensive survey of the evidence is this one by Austroads, with its firm conclusion that small cars are not inherently less ‘safe’ than large cars. Substantive ‘vehicle’ factors in crash outcomes include age and the presence of safety features, but not mass per se.

Crash Ratings vs. Safety Ratings

Ultimately the more important point is that, despite the titles of these study reports, none of the Monash figures or the ‘star’ ratings regularly presented in the motoring press actually represent safety ratings: rather, they are crashworthiness ratings. They indicate the expected proportion of crashes in which the driver will suffer a life-threatening injury. But your actual risk of being injured depends on two factors: it is the likelihood of being injured when the vehicle crashes, multiplied by the likelihood of being in a crash to begin with.

The Monash results are obtained by going through police and other official records of crashes and matching injuries against kinds of vehicles involved. So they say nothing whatsoever about how likely you are to be involved in a crash when driving a particular vehicle, because the researchers do not look at the entire population of vehicles, only those that have crashed. (Nor is it the intent of these studies to investigate how often particular vehicles crash; only what the expected outcome of a crash is.)

Suppose there are two kinds of vehicles, Model A and Model B, with a million of each on the road and driven equally often. Over a ten-year period ten Model A vehicles crash, and in each case the driver is seriously injured. Over the same period 1000 Model B vehicles crash, and the driver is seriously injured in 100 of these cases. On a crashworthiness rating Model B is ten times as good as Model A, because the driver was injured in just 10% of Model B crashes, compared with 100% of Model A crashes. Despite this, the ‘average’ driver of a Model B vehicle has a 10 times higher risk of being seriously injured in a crash than the driver of a Model A vehicle, simply because Model B cars are (for whatever reason) involved in so many more crashes.

In the real world of motoring, crash rates depend on many factors including driver behaviour, the road environment, and the behaviour of other drivers. Not all of these factors are within a driver’s control. But there are at least two ways in which the type of vehicle being driven can affect the likelihood of being in a crash:

  1. Design features of a vehicle affect how well it handles in difficult situations, which can make the difference between crashing and not crashing. It’s now well known that four-wheel drives are more vulnerable to rollover crashes than conventional cars; more subtle is the fact that smaller cars are more manoeuvrable and can have more near-misses in situations where larger cars would crash. A small car presents a smaller target in a crash situation, and its lower mass makes it somewhat easier to move out of the way of trouble.
  2. Vehicle factors can feed back into driver behaviour. If a driver feels safer in a particular vehicle (whether or not the driver really is safer), they may compensate by driving in a more risky manner – a phenomenon known as ‘risk compensation’. Some larger vehicles are also deliberately marketed with a ‘risk-taker’ image; very few small cars are marketed this way!

These factors are described most eloquently in an article by Malcolm Gladwell, that originally appeared in the New Yorker in 2004. As Gladwell explains:

Most of us think that SUVs are much safer than sports cars. If you asked the young parents of America whether they would rather strap their infant child in the back seat of the TrailBlazer [SUV] or the passenger seat of the [Porsche] Boxster, they would choose the TrailBlazer. We feel that way because in the TrailBlazer our chances of surviving a collision with a hypothetical tractor-trailer in the other lane are greater than they are in the Porsche. What we forget, though, is that in the TrailBlazer you’re also much more likely to hit the tractor-trailer because you can’t get out of the way in time. In the parlance of the automobile world, the TrailBlazer is better at ‘passive safety’. The Boxster is better when it comes to ‘active safety’, which is every bit as important….

[Automotive engineer David] Champion says that one of the occasions when he came closest to death was a snowy day, many years ago, just after he had bought a new Range Rover. Everyone around me was slipping, and I was thinking, Yeahhh. And I came to a stop sign on a major road, and I was driving probably twice as fast as I should have been, because I could. I had traction. But I also weighed probably twice as much as most cars. And I still had only four brakes and four tires on the road. I slid right across a four-lane road.

—Malcolm Gladwell, Big and Bad: How the SUV ran over automotive safety, January 2004.

As Gladwell’s article argues, it is not small cars per se but the proliferation of large vehicles on urban roads that poses a real threat to safety. The 2006 Monash study confirmed previous findings that large four-wheel drives are unusually ‘aggressive’, resulting in 6.1 serious injuries to other road users per 100 crashes, compared with an average of 3.9 across all car models. This is consistent with international evidence: in Britain, where there are more small cars and fewer large cars than in Australia, the fatality rate per kilometre driven is one-third lower than Australia’s. The aforementioned Austroads report likewise concludes that fatality risk is higher in collisions between 4WDs and regular cars (of any size) than in car-car collisions.

The other crucial factor in car crash risk is, of course, the amount of driving one does. For each kilometre travelled, the risk of being injured on public transport is around one-tenth that from driving a car, which means every public transport trip that replaces a car trip improves the safety of those travelling. Or to look at it another way, the most ‘dangerous’ car in the Monash study, if used only on weekends, carries no greater risk than the most crashworthy car driven to work every day.

In summary, there is a lot more to being safe in a car than is conveyed in half-truths about crashworthiness ratings. For years, Australia presided over a car industry that was geared almost exclusively to the production of large sedans, and that entirely missed the shift in consumer preferences toward small cars on the one hand and 4WDs on the other. Whatever might have been done to save this industry, scare-mongering about small cars being unsafe was no way to go about it. Nowadays, it continues to be simply misleading as advice to consumers.

Last modified: 23 April 2018