Gordon Stokes Transport  

Age and generational effects in travel behaviour

Transport Home Photography Home Transport Photos

Last updated - March 2013



This note shows various aspects of the way in which travel behaviour varies by age, gender, and the generation into which one has been born. It uses data from the National Travel Survey for England from 1995 to 2014.

It is exploratory analysis and aims to present a range of findings for others to interpret rather than to provide a complete or 'peer reviewed' paper. The aim is for others to pick up on ideas within it and expand if they see use in doing so. None of the analysis is complex or groundbreaking, but tends to look at aspects of travel behaviour that many researchers overlook since other more detailed aspects are of more interest to them. It is mainly concerned with one very simple measure - whether or not people used different travel modes in a travel diary week. Some of the graphics, however, can be confusing for some people, so I've tried to explain them in the text.

The page currently has the following sections:-

  • 1. How the use of modes varies with age
  • 2. How use of modes varies with age and gender
  • 3. Change in the use of various modes over time
  • 4. Suggestions for reasons for changes
  • 5. How have cohorts/ generations change over time?
  • 6. Can we predict what will happen to the generation who are driving less in their youth?
  • 7. Conclusions

The paper notes that different generations have travelled differently, mainly in terms of their car use, and that car travel habits formed generally appear to 'stick'. It is suggested that younger generations who are not using cars to the same extent as those born from the 1940s to 1960s may not get to use them to the extent that this older generation did. But it also concludes that it is not really possible to predict what they will do.

1. How use of modes varies by age

The chart below shows the percentage of people of each year of age who reported using each of the main travel modes during the NTS 'diary week' for all surveys between 1995 and 2014. It doesn't show how much they used each mode, just whether or not they did.


Figure 1 - Percentage of respondents recording use of a travel mode by age (1995-2014 inclusive)

A few points can be drawn from the chart:-

  • Car passenger (light blue) is the most likely mode to be used throughout childhood, but falls away from the age of 10, and even more from the age of about 17. But it carries on being used at least weekly by 50% of people throughout life.
  • Car driving (dark blue) becomes the most commonly used mode from about the age of 25, having risen very rapidly from age 17 to 19 and then more slowly. Its use tails off in in later life. But - do not assume that this is due just to ageing. Later graphs will show the importance of generational effects.
  • Walking (mid green) is recorded by around 50% of people (but see italic note in bullet below). There is a peak at age 15, with higher rates from about 10 to 17. Walking falls off in later life from about age 75, and this is probably due to health and fitness.
  • It should be pointed out that the NTS records walk trips of less than one mile on only one of the seven travel diary days. If it were treated the same as other modes the line would be higher, but it's difficult to say by how much. While data weighting can be used for measures such as numbers of trips or distance travelled it doesn't work for whether a person walked or not. All we know is that the line will be higher, and probably quite a lot.
  • Cycling (light green) is never used by more than about 12% (around age 14-15) and generally by under 10%.
  • Bus (red) is most used between ages 11 to 25, and then again after age 60. At other ages it is used by about 20%. For a couple of years at age 15 and 16 it's used by over 50%, and the peak in later life (from about 70 to 80) is around 35%.
  • London Underground (Bakerloo Line brown) is used most from about age 25 to 30. Although only 10% record its use, it is only available in London where about 20% of the population live (and some from outside London work)
  • Rail (mauve) is not used by a huge proportion, maximum 15% around age 20 to 25, and is used very little by those age 10 or less. But it's seen some interesting changes which I'll come onto later.
  • Finally, Taxi (New York yellow) has a peak at age 20, having risen rapidly from around 13, and then falls off, before rising a bit from age 80.

To make sense of the above, it seems apparent that the ages between about 10 and 20 and then from about 60 onwards are key. This covers the period of 'life changes' from primary to secondary school, and then through to working age - and then from retirement onwards.

During the teenage years, independent travel becomes the norm, and travel to school often involves cycle or bus. From 17 driving becomes a possibility and is taken up by many. But in the scheme of things car driving is not the 'universal' mode that many accord to it. At its peak, only 75% drive in any one week, around the early 40s.

2. How use of modes varies by age and gender

The following chart show the same as the one in the last section, for males, and females. If you pass the mouse over it, it should change to an animated image showing men and women. On an i-pad a clck should do the trick, and then click elsewhere to stop it. If it doesn't start changing after a second or so, something has gone wrong and your browser won't show 'animated gif' files.
Assuming you can see them change it's apparent that use of travel modes is different for male and females at most ages.

  • Up to age 10 there is little difference between boys and girls in terms of the use of modes
  • From 11 onwards boys use bikes a lot more, while girls use buses and travel as a car passenger more. They also walk somewhat more.
  • From the late teens and through most of adulthood a higher use of driving amongst men is 'offset' by lower use of most other modes.
  • Only rail and London Underground are similar for both genders.
  • Driving my men also remains high well into old age, compared with women. While men use driving more than any other mode up to their mid 80s, for women walking is more common from the mid 60s. BUT this is not due to differences in 'taste' but due to historical differences in how many men and women learnt to drive and get cars in decades past. It almost certainly won't always be like this, and I'll come back to the issue shortly.


Figure 2. Hold the mouse over the image to show animated version - or tap on a tablet

3. Changes in the use of various modes over time for different ages

So far we've looked at all the data for a twenty year period from 1995 to 2014, but things have changed during that time. The next charts (animated again) show how the use of various modes has changes for different age groups over time. In these charts a 5 year 'rolling' average is used. People labelled as, say, 20, actually mean that it's the average for people aged 18 to 22. This has been done because, as Figure 2 showed, the lines were getting a little bit jagged, showing that the statistical reliability was becoming questionable. What it does mean is that it might appear that, for example, people aged from 15 are driving a car, while this would only be the case for anyone driving illegally (and they'd be unlikely to admit to that in a government survey even if they were!).

Walking

Fewer people of most ages have been making a walk trip during the diary week as time goes by. The falls have generally been greater in older age for men, and in middle age for women.


Figure 3a&b. Hold the mouse over an image to show animated version (or tap on a tablet)

Bus

Bus use has not changed significantly for many age groups, but use in teenage years has fallen, especially for women. What is somewhat surprising is that the introduction of free bus travel has not had a major impact of increasing likelihood of using a bus in later years. In fact the differences in trends for bus use for London and elsewhere are large due to different systems and policies, but are not shown here.


Figure 4a&b. Hold the mouse over an image to show animated version (or tap on a tablet)

Rail use

Rail use has been rising since around 1995, and it seems that an increase in the likelihood of use has risen for most age groups. While men seemed to show an increase between 1995-99 and 2000-04 the increase for women only seemed to take off after the 2000-04 period.


Figure 5a&b. Hold the mouse over an image to show animated version (or tap on a tablet)

Driving a car

Car driving by age has shown quite remarkable changes for both sexes, and these changes are of a very different nature to other modes shown above. While for most modes any increase or reduction has been relatively uniform across ages, the pattern for driving shows some age groups increasing use while others reduce, and the changes seem to follow generational patterns. To explain this we have to look at how driving became more common from the 1960s onwards.

  • For men the graph looks like a wave moving forward, and that is a good description of what has happened. In the 1990s driving was done by about 80% of males up to the age of about 55 and then dropped off for older people. In 1997 (mid point of that period) those people would have been born in the 1940s, including the baby boom period and it was this generation who learnt to drive and acquire cars in large numbers. By 2010-14 the age of this level of 80% use had risen to 70 implying that the same generation were still as likely to drive as they were 15 years earlier.
  • At the younger end of the male chart the picture is different. The 'slope' upwards to a high level of likelihood of driving started rapidly in 1995-99 but by 2010-14 showed a similar start at age 17-19, but then slows much more. Whereas in 1995-99 around 80% were driving by age 30, this level wasn't reached till age 45 by 2010-14. It seems that becoming a regular car driver has become somewhat less likely for the generations born after about the mid 1960s.
  • For women there are elements of the same trends, but in the 1960s women were still much less likely to learn to drive and acquire cars. For them the rise in car use came later, and the overall wave has got bigger.


Figure 6a&b. Hold the mouse over an image to show animated version (or tap on a tablet)

4. Possible reasons for changes

There could be a number of explanations for these changes, but the most plausible ones (to my mind) relate to the situations generations were brought up in and what we might call 'habit'. Some of these ideas put as hypotheses will be tested in later text. Ideas include:-

  • Men born in the 1940s, and 50s grew up at a time when cars were seen as 'the future', and gaining a driving licence and car was a sort of 'rite of passage' into adulthood. It's churlish to suggest that all younger men were 'petrol heads', but there is a likelihood that (to put it in economic parlance) men's propensity to drive was greater than the utility offered by cars. In day to day language, the desire to drive was greater than the benefit actually offered.
  • Men born from about 1970 onwards grew up being driven around by their parents with less independent travel, and having to be strapped in child seats and safety belts, at a time when congestion was increasing. 'Motoring' didn't seem quite the wonder that it was to the older generation. They were also very much more likely to go to university or college than their parents, and were likely to go to big cities where having a car was not a particular advantage, or campus universities where cars were restricted. If anything was replacing the car as a status symbol it was technology - computer games and, more recently, smart phones.
  • Added to this the cost of insurance for younger drivers increased rapidly, and the driving test became more complex - paid for lessons became more or less essential. Practice on quiet roads became more difficult to find. People were made aware, and accepted, that drinking and driving was, to say the least, unwise.
  • For women the situation was somewhat different. For generations up to those born in the 1940s driving was possibly seen as a 'male' activity, and few learnt to drive. But with increasing gender equality, and increasing working rates the practical advantages of driving became apparent. So driving amongst younger women increased. Since around 2010 it seems that the likelihood of men and women being the main driver of a car has reached 'equality' up to the age of about 40.
  • Older women were less likely to have ever got a driving licence or a car. A later section will show just how few people gain a licence after about the age of 30. So we are left with an older generation of women who are never likely to drive. In the 1990s this meant that women over the age of 50 were much less likely to drive but by the 2010s this age has reached nearer 70. There is still part of a generation of older men who were unlikely to have learnt to drive, but these were generally men born in the 1920s and earlier.
  • One interesting phenomenon seems to be that once people have access to a car they tend to not give it up. Cars are useful (in most situations for most people) and few people voluntarily decide that they might as well do without. It is usually ill health that stops people driving. Added to this, many more elderly people report having difficulty walking or using public transport than using a car - driving is easier on a failing body. Thus we have seen the large growth in older men and women driving, especially as facilities get larger and more distant from home, and if bus services are under threat.
  • This 'habit' is observable in NTS travel data for driving but not for other modes. The graphs in Section 3 seem to show a generational shift in car use, but not other modes. This raises the question of whether there is something in car use that encourages habitual behaviour (e.g. it's 'too' convenient to stop or reduce car use) or whether other modes have not yet gained a 'lifetime' habit status. Cycling is a mode that many see as a 'lifestyle', but few would say the same about buses, and rail doesn't offer travel for all the journeys someone might want to make for the vast majority.

5. How have cohorts/ generations change over time?

The charts above showed how travel behaviour varies with age. The next charts follow each age cohort in terms of how they use a mode throughout the survey years. The surveys covered are from 1985/6, and then the continuous survey from 1988. The early years have to be treated with caution, since uniform weighting of NTS data is only available from 1995 - statisticians who go by the text book would argue that they are suspect. However, as we are using a simple measure of whether or not people used a travel mode, it is felt that as long as the graphs are taken as pictorial, they provide more insight than if the earlier years were omitted. It's also important to note that these are not the same people in different surveys, but are random samples of people at different times - a series of snapshots, rather than a 'panel' reinterviewed.

5.1 How have cohorts changed their driving behaviour over time?

We start with how age cohorts have changed over time in terms of how likely they are to have driven a car in the diary week. Some people struggle with these diagrams - I hope you won't. Each is a series of lines charting how a cohort (or generation born at a set time) changed their travel behaviour at different survey times. Holding the mouse over the first set of charts (Figure 7)shows how the trajectories if each age cohort developed over time. And below the first two are different explanatory charts which may help (Figure 8). The left hand one has annotations for the male graph above (albeit in writing that is so small you can't read it - I'll try and fix it), while the right hand one shows, more generically, what different patterns mean in these charts.


Figures 7a&b Holding the mouse over the charts will show a dynamic version showing how each cohort has developed its use of cars over time


Figure 8 - Explanations of cohort trajectory graphs

I've written a page on which you can choose to show these cohort charts for different modes together. It's not yet refined and you'll have to use a bit of trial and error to get it to work. I'd also recommend reading a little bit further before trying it out.

Link to cohort chart mixing web page

For men:-

  • In middle age, likelihood of driving remains fairly steady (yellow line in Figure 8 - those born about 1950) while in older age the likelihood does reduce (e.g. red line).
  • For all younger age groups the likelihood grows, but if you track up vertically for those aged, say, 30, it's apparent that those born in 1980 are less likely to drive than those born in 1970 (red line compared with light blue line).
  • By age 30, 72% of males born in 1970 were driving, while only 60% of those born in 1980 were.
  • At age 70, 57% of males born in 1920 were driving, but 72% of those born in 1930 were.

For women the chart is different, but interesting in other ways.

  • Older women seem just as likely to have carried on driving over the survey periods. Only for those born in 1930 and earlier was there a significant fall, and this was much less than for men. This could be because women tend to live longer than men and stay in better health till a later age, but it may also relate to older women being much less likely to have learnt to drive, and probably, those who did learn were from better off backgrounds, and have maintained their health better. For those who did learn to drive they may have husbands who no longer drive or who are no longer with them so have taken over the driving role.
  • The distance between the parallel lines from about age 40 onwards (in 1985) points to the increasing likelihood of driving for those born up to about 1970. For men these lines only diverge from about age 60 (in 1985)
  • The trajectories in younger life are relatively similar for men and women, though in the earlier surveys women were less likely to drive. But in each case the likelihood of driving increases up to about age 40. Put crudely, if you're not driving by age 40 you're not very likely to (based on evidence up to the present time).

The graph below (Figure 9) looks at the related measure of the percentage who are main drivers of household cars, and how over time this has varied for people born in the same years. These show very clearly how, for those born from 1960 and earlier, the level of access to a car has remained almost constant between the surveys from 1995-99 till 2010-14. The percentage has fallen for men born before about 1935, but not for women. For those born after around 1960 the lines are of suspect value because of the ages of people in the different surveys.

Something that surprises many people is that the demographic most likely to be the main driver of a car are men aged between 60 and 70, most of whom no longer work. It may be that this 'peak' age will get older as this generation ages.


Figure 9 - Percentage who are main drivers of a car according to year of birth in different survey periods

5.2 How have cohorts changed travelling as a car passenger over time

The charts for travelling as a car passenger (below) can be seen, very roughly, as mirroring the charts for driving (in terms of being opposite). Both start at a very similar high rate of around 80% recording car passenger for one or more journeys in the travel diary for young children. But the male chart then dives much lower to around 30% in mid life years, while the female reduces much less.


Figures 10a&b

While the trends for different male cohorts are a bit random in places, for women from the age of 20 to about 50 to 60 show a trend towards much less likelihood of being a passenger over time. While, around 1985, about 74% of women aged 30 travelled as a passenger, by 2010-14 the figure was around 58%. This mirrors the rise in likelihood of driving. For males there is also less likelihood of being a car passenger (down from about 48 to 38% at age 30) but the changes are generally a lot less.

5.3 How have cohorts changed likelihood of walking over time

Because walks of less than one mile are only recorded on one day of the diary week we really don't know what percentage make a short walk in the dairy week. (I tried using probablility to scale up to a week, but it either doesn't work, or I'm not clever enough to do it right.) For this reason I've split walk trips into people who made a walk of less than one mile on the one day recorded, or one of over a mile on any day during the week. These graphs are from 1995 onwards (unlike the other modes which started at 1985).

Figures 11a and 11b show that falls in the percentage recording short walk journeys in the diary day have fallen for just about all cohorts of different ages. The only exceptions are for young children, who are at a stage of life when we would expect increasing likelihood of walk trips. For both men and women the reduction slowed or even stopped between 2005-09 and 2010-14. While falls from age 20 to age 35 might be regarded by some as an inevitable consequence of a move from 'youth' to a busier lifestyle which offers less time for walking, the consistent falls for other ages are almost shocking. At most ages the falls have been greater for women, except for older men, who used to show a high level. These changes for older people may be partly a consequence of increasing access to a car. Let's just hope that a lot of the younger ones are driving to gyms!


Figures 11a&b

For walks of over one mile (Figures 12a and 12b) the trajectories are more stable for each cohort, with most cohorts following those before them. And most lines are relatively parrallel so there hasn't been a reduction like there has been for shorter walks. It is seen that the peaks in walking for teenagers is related to walks over one mile, rather than short walks. It's also notable that from the ages of about 25 to 50 there have actually been increases for each age cohort (though some falls between 2005-09 and 2010-14) which go against the trajectories for short walks. Many show a rise early in the period and then a fall.


Figures 12a&b

5.4 How have cohorts changed likelihood of travelling by bus

The charts for bus (Figures 13a and 13b) generally show less generation cohort change than for the car ones. The lines for men follow roughly the same trajectories. For women, however, there has been a 'level' use of bus by those born from about 1935 to 1945, but with each cohort starting from a lower rate. This probably relates to increasing driving rates by women for these generations meaning fewer did use and continue to use buses.


Figures 13a&b

5.5 How have cohorts changed likelihood of travelling by train

For rail (Figures 14a and 14b) the rates are low, since rail is not a commonly used mode of travel, but the growth in patronage that has occurred since 1995 can be seen to have been strongest amongst those of working age. Unlike any of the other modes most cohorts are increasing their likelihood of travelling by train.


Figures 14a&b

6. Can we predict what will happen to the generation who are driving less in their youth?

Prediction is a difficult business. In the 1950s it was forecast that the UK's population would be 75 million by 2000. Less than 10 years later it was estimated to grow to about 53m and then level off. Neither have been right.

Many (including me) have talked about the notion of peak car. What this actually implies (to most transport researchers) is not a peak from which car use will now reduce, but that the amount of driver mileage per person has levelled off and future growth in car use will roughly relate to population growth rather than 'keep growing'. But the truth is that we don't know for sure whether a fundamental change in behaviour has occurred, or whether recent circumstances have lowered use amongst younger people and it will rise again. Or, whether it will fall.

This isn't the place for an in depth analysis of how and why forecasting is done, but, as in many social and economic areas, many people demand forecasts (or rather just one forecast) and get annoyed when they end up being wrong. Here we look at some analysis that points to how current generations may change their behaviour (or not) in the coming years, with some possible pointers as to how current children (and those not yet born) might travel when they reach adulthood. The future behaviour of current adults is difficult to predict, but a lot easier than for those who have not yet formed their independent travel behaviour, or are not yet born.

To my mind, accurate forecasting is never possible, or even sensible or desirable. The best forecast is more likely to be wrong than right, and lead to wrong decisions, and, at worst, accusations such as "experts are useless - we don't need them". Far more important is to formulate a plan that is robust to the range of plausible and likely futures that we can identify. If you take one thing away from this note, I think that should be it.

The slowdown in younger people acquiring licences and cars begs the question of whether they will do so at some point in the future. Some argue that when people have families and likely move from cities to small towns and rural areas, and gain wealth they will find cars useful, and acquire them. Others argue that once younger people are in the habit of not using cars they will tend to steer things so they don't need a car. Of course, some of these people will get cars, and some won't - the question is whether there is any evidence that helps suggest 'how many'.

6.1 How has licence holding developed over time for men and women?

It used to be that "men drove and women didn't", which was always a gross oversimplification, but Figure 15 shows that in the combined surveys from 1995 to 2014 for those born before the 1940s it was very much more likely that men had driving licences. For those born in more recent decades there has been growing 'equality'. It should be noted that people aged under 40 have been excluded from the graph since, although most people who are going to get licences tend to get them while youngish, we can take 40 as the age where just about everybody who is going to get a licence has done.


Figure 15

6.2 Does age of acquiring a licence have an effect on how much people drive?

Current evidence (based on past behaviour and shown in the next two charts) does seem to point to a reduced likelihood of gaining a licence in later life, and that those who acquire a licence later in left tend to drive less.

Figure 16 shows that once car driving is established within a cohort, most get a licence when quite young. For men born from the 1950s, by the age of 40 and over, around 85% have a full licence, and over 50% of these gained their licence by the age of 18, and about two thirds by the age of 20. The proportion gaining a licence after the age of 30 is very small (around 5%). For women the situation is somewhat different, with many more born in the 1940s and 1950s getting a licence later, but still less than 15% getting one after the age of 30. By the time of those born in the 1970s the profiles look very similar to that for men.

It should be mentioned that a question in past NTS surveys implied that many without licences intended to get one within the next one or five years. However, I think there is a difference between intention and action in many cases. And I often worry that an interviewer asking you a question like that is likely to encourage a 'positive' answer, especially when you haven't a clue whether you will or not.


Figure 16

Figure 17 is a bit of a diversion but shows how the number of licences acquired is closely related to age of population booms. It shows the numbers of people recorded in the NTS surveys (from 1995 to 2014) by their year of birth, and also the number of licences gained in each year. It shows (blue and brown lines) that the population booms show up with higher numbers in the sample (post WW1, post WW2 and the 1960s - a prolonged rumble rather than a boom). It also shows that there were distinct booms in licence acquisition 17 or 18 years after the population booms. The post WW1 boom doesn't show for either gender; the post WW2 boom shows strongly for men, and forms the start of a 'plateau' for women; the 1960s resulted in peaks for both genders, with the highest in the late 1980s for women. It is likely that this peak for women was also the result of higher later licence acquisition related to higher rates of working for women. This period also saw the most rapid growth in car use that the UK has experienced.


Figure 17

Finally, Figure 18 strongly suggests that the later one learns to drive, the less distance one is likely to drive. Each line shows the mileage driven per year by people who gained their driving licence at similar ages. Those who gained licences at age 17 or 18 consistently drive further than others. For men and women the lines for those aged 19-20 and 21-24 are similar, but those who gain their licence later show markedly lower mileage.

There are a number of possible explanations for this, including:-

  • those who are smitten by the idea of driving will learn early
  • those who live in areas where a car is 'a necessity' such as rural areas will learn early
  • those who learn later have learnt other 'habits' for travel which mean that they are unlikely to use a car for all their journeys but use them more selectively, and with more discrimination
  • driving may be an activity for which learning is 'best done' (?) when the nerves of adulthood have not yet become ingrained. For those who learn later 'excitement' may be replaced by 'anxiety' - the two can be differing reactions to the same stimuli.


Figure 18

7. Conclusions

This note has covered a wide area of analysis, and the following would seem to be conclusions that are worthy of further research. Besides the obvious conclusions that need no expert analysis, none can really be proved by this work alone, but are areas that I hope to continue researching.

  • That age affects how we travel is beyond any reasonable doubt in that we use different modes at different ages because of our changing travel needs. The travel needs of a young child are very different to teenagers, students, those with children, those without children, and those who have retired.
  • That when one was born also has a separate effect. The most obvious is that driving a car is related to birth decade. While we are now at a time when nearly all men still alive grew up at a time when car driving was 'the norm' there is still a generation of older women who grew up at a time when it was not expected that they would drive. There are now younger generations who have been brought up when driving is not seen as such a 'natural thing to do' as those born earlier.
  • There’s strong evidence that having a car is ‘difficult to give up’ in that once people (shown here as similar age cohorts) have a car and drive, they tend to carry on. With this level of evidence we can't say whether it's 'habit', 'addiction', or just convenience that leads to this, but it is observable. There is much less evidence of ‘life habit’ formation for other modes

Younger generations are not using cars to the extent ‘baby boomers’ did. The question is whether they will they get ‘stuck’ on public transport and/ or cycling habits that they have learnt while in 'formative' years, whether they will start using cars as much as the ‘boomers’ once they have families and get rid of debt they may have accrued, or whether they will do whatever is most convenient, dependent on their circumstances?

The only sensible answer is that "we don't really know". But it does seem unlikely that they will embrace car use to the extent that many of the 'baby boom' generation did. It is also patently true that each generation is not a homogeneous group. For a long time there have been different 'sub-cultures' with different aspirations and cultures amongst the young. Although many of these may not become life long sub'cultures they probably tend to influence later life attitudes. We may quite easily see different groups within the same age groups behaving very differently.

Back to Transport Index

Gordon Stokes, 2016