By Chris Atkins

I have lived, studied and worked overseas for just over 20 years and have recently moved back to the UK, where I was born and brought up. Not much has changed in that time – certainly not the famed British weather – but having settled into living back “home” I have noticed one thing: there seem to be a lot more fancy cars around than there were before I left. In particular, the Audi marque seems to be much more common on UK roads than in the mid-1990s.

Having read Daniel Kahneman’s excellent book, Thinking, Fast and Slow, I’m conscious of how this might just be a fast-thinking bias on my part and, not one to tolerate such doubt, decided to think slow and see if I was correct or not. This is consistent with my “two cultures” approach to life, taken from C.P. Snow, and described previously in Montgolfiere Weekly. I’m reminded of Jerry Maguire’s pleas to “show me the money,” although in my case it’s almost always show me the data.

Fortunately for me, developed-world, government-funded statistics are available to help. The UK’s Department for Transport publishes the number of licensed vehicles by quarter and provides data going back to 1994, the year I first left the UK. Furthermore, it breaks the data down by vehicle make and model. I happily started working on a simple analysis, taking the total number of a given marque as my primary measure.

What do the data show? First, the growth in licensed vehicles has been quite high. The total number of licensed cars in the first quarter of 2016 (the latest for which data are available) is close to 50% higher than in the fourth quarter of 1994.

But what about the more luxury end of the market? I selected a few makes, based on my observations of the local traffic and looked to the growth in those. The results are shown in the table below.

Marque Growth









Land/Range Rover












All cars




All excluding luxury




Rolls Royce




What does this tell us? First, there’s little doubt that there are more high-end cars on the road than there were over 20 years ago. Next, the growth in some of the brands is really very high. However, this is mainly explained by the introduction to the market of new marques (such as Infiniti) that had very few licensed models in 1994 (in fact, for Infiniti there were none, hence an infinite growth). While the overall growth is around 50% over the period, this includes the growth of the high-end vehicles. Excluding those from the analysis, as shown in the table, indicates that there has actually been a fall in the more run-of-the-mill vehicles during the analysis period. Finally, although there have been winners, there have also been losers – high-end vehicles whose growth has been below the overall trend, such as Triumph, Rolls Royce and Volvo. The first of these is probably explained by it not being available any longer, the opposite of the Infiniti effect described above (the last Triumph, the Acclaim, was introduced in 1981).

Having justified my intuition, and been pleased to find that the marque I noticed most, the Audi, has shown some of the highest growth, I turned to wonder why this is the case. Price seems an obvious factor so I looked for inflation data for motor vehicles, and found that motor vehicle inflation has been negative, at least for the period for which I could find data, as shown here.

Unfortunately, this does not tell the whole story. First, the data are for vehicles overall and not by marque, so it isn’t possible to tell if high-end vehicles have become relatively less expensive, although I suspect that more extensive research (and time to do so) would have been able to address this question. Also, as the article points out, during the same period fuel and insurance have become more expensive, so overall cost may have gone up. Finally, a “wealth effect,” where people spend more as the value of their assets rises, may have come into play – probably caused by the increase in UK house prices over the same period – driving the observed phenomenon.

What should we take away from this? First, intuition can be useful, although it should always be challenged and tested with data. Second, data are generally available, particularly from national statistics agencies and government departments. Finally, although it is often harder to find the supporting data, plausible factors can be hypothesized and tested, although it’s well known that correlation and causation are not the same thing and should not be relied on as an absolute explanation.

Having had my interest piqued, I find that there are other driving-related topics that come to mind and will form the basis for another article. This will be well into the other culture and be data free.

Photo credit: FotoSleuth Audi RS6