January 17, 2013

Teenagers less likely to have a driver's license than any time in the past 50 years

I've already got a few other graphs up like the one below (search for "license"), but now the 2011 data are in from Highway Statistics. The percent of eligible teenagers, 16 to 19, who actually have a driver's license continues to tumble:


At 51% in 2011, and at the rate it's falling, it's probably already below 50% in 2012. In 2011, 28% of 16 year-olds had a license, 45% of 17 year-olds (!), 60% of 18 year-olds, and 69% of 19 year-olds. So now, probably one-third of 19 year-olds do not have a license. How do they go through security at the airport to-and-from college, during spring break, or any kind of trip? If they're going through all the trouble of getting a passport or other government ID, why not just take that time and learn how to fucking drive?

For anyone who has even minimal contact with real teenagers today, this can come as no surprise. They feel so uncomfortable around other people that they'd rather just stay home and wrap themselves in their cyber-cocoon. Feeling so awkward in a social context, they prefer that their friends exist only in a virtual world where they don't have to actually meet up with them, restricting their "interactions" to online video games, Facebook, and texting.

But enough ragging on today's defenseless dorky young people. For this installment, I've managed to dig up some older data that go back to 1963. I didn't cover this before because the driver's license data that far back only say how many people have a license in some age group, not what percent of the age group does. But in the meantime I've gone through lots of data on the age structure of the population, so I can compare the number with a license to the number overall within some age group.

My age structure data groups all 15-19 year-olds together, so I can't look at just the 16+ teenagers. But the driver's license data has a category for anyone 19 and under. Not a huge change in what we're looking at, then.


This graph only goes up through 2010, but the decline over the past 20 years is still evident. The neat thing to notice here is that it wasn't at some constant high value before the decline. That's a common misconception about declines -- that everything had been cruising along nice and steady, when suddenly something disrupted the world and we've been in decline ever since.

As it turns out, teenagers in the early 1960s were about as likely to have a license as their counterparts in the declining period of the late '90s. Before the decline, young people became more and more likely to want a license to drive, before reaching an uneven plateau during the late '70s and the '80s. The protracted decline does not begin until 1990, another example of how cocooning behavior seems to slightly precede a fall in the crime rate. (Indeed, cocooning causes falling crime rates, as the predators have a harder time finding vulnerable prey out in the open.)

Economics does not explain the pattern, as cars have only gotten cheaper during the entire period, and parents only more willing to pay for their kid's insurance and even gas & maintenance. Or more typically, let their kid drive the family car, which again have only gotten cheaper and more ubiquitous over the past 50 years.

What about how youthful the population is? Perhaps when it's growing younger and younger, there's a feeling of excitement in the air, and especially the young people themselves don't want to feel like it's passing them by. Better jump on board while you can, and you need a driver's license to take part in the unsupervised socializing of your fellow teenagers.

Well, close but not quite. Below is a graph of the number of 15-29 year-olds compared to the number of 30-59 year-olds. This ratio reflects the difference between the strength of the force pushing for more excitement, and the opposite force pushing for more containment.



It rises through a peak in the late '70s and begins a steady fall after 1980, yet the sharp drop in driver's licenses doesn't begin for another decade. Also, there's been a steady if small rise in the youthfulness ratio since the early 2000s, yet young people have only withdrawn more and more from other people.

Again, the changes over time look more like the outgoing vs. cocooning pattern. It's more or less the crime rate graph shifted earlier by a few years. For example, youngsters seemed to become more outgoing and rambunctious a few years before the 1959 start of the most recent crime wave. And on the flip-side, they begin to withdraw into their own private spaces a few years before the 1993 start of the falling crime rate. These kind of cycles that are slightly out of synch with each other suggest that the "ecology" of criminals and normal people out in public is like that of predators and prey, perhaps not so shocking of a view but still worth keeping in mind.

31 comments:

  1. I don't believe computers/helicopter parents are to blame for kids not wanting to drive. Driving is supposed to be about freedom, but for young folks it is solid oppression. Young drivers get special selection for ticket revenue. Have special no texting laws(because at 19 you can suddenly safely text and drive.) And the big one is the world of hurt if they get any sort of alcohol/car/police interaction under 21. Not driving is perfectly rational behavior for young people. Why put up with all that when they can put a couple beers in a backpack and skateboard to the girlfriends house. Much better then getting popped under age with beer in a car. Or make Mom drive.

    ReplyDelete
  2. Umm, is this broken down by race?

    ReplyDelete
  3. It's a change in their underlying preferences, not in how rationally / efficiently they pursue those goals. Kids today just don't want to have fun or become independent.

    Cars give teenagers a whole lot more freedom than all alternatives, and again they're generally paid for by the parents because they're driving the family car. There's a pretty simple way to avoid getting busted with beer in the car...

    ReplyDelete
  4. "Umm, is this broken down by race?"

    No racial change would explain a drop from 65% to 50% of 16-19 year-olds having a license. Nor would it explain the rising fraction of licensed teenagers during the '60s and '70s -- America wasn't getting whiter.

    ReplyDelete
  5. "No racial change would explain a drop from 65% to 50% of 16-19 year-olds having a license. Nor would it explain the rising fraction of licensed teenagers during the '60s and '70s -- America wasn't getting whiter."

    The rise and the later fall need not have the same cause.

    The non-White share of the population has grown considerably since the '60s and '70s. As well, they disproportionally constitute the younger generations. They are also more likely to be poorer and live in urban areas (where driving is less necessary). This could easily explain a 15% drop in the share of teens with drivers licenses.

    It would seem to be prudent HBD wisdom to like to see statistics like these separated by race....

    ReplyDelete
  6. "The rise and the later fall need not have the same cause."

    That goes against Occam's Razor. We want to explain the greatest amount of variation that we can, using the fewest number of parameters. The alternative is a proliferation of ad hoc just-so stories.

    "The non-White share of the population has grown considerably since the '60s and '70s."

    Right, and despite that, during the '70s and '80s there were *more* teenagers with licenses.

    "They are also more likely to be poorer and live in urban areas (where driving is less necessary)."

    Actually over the period of declining licenses, blacks have been pushed out of the cities by more affluent whites moving back. Part of the falling-crime pattern -- less dangerous cities attract more whites, while rising-crime sends them running for the suburbs.

    You can see that in the GSS by looking at SRCBELT (urban metro, suburban, rural, etc.) by YEAR, control for RACE, and restrict AGE to 18-29. Since the 1990s, young whites have become more likely to live in the top 100 metro areas, while black urbanization has plummeted.

    ReplyDelete
  7. "they disproportionally constitute the younger generations."

    The percent of 15-19 year-olds who are white goes like this during the period of declining licenses:

    1990 - 80%
    2000 - 78%
    2009 - 76%

    The percent of that age group with a license goes like this:

    1990 - 52%
    2000 - 48%
    2011 - 42%

    A quantitative argument must say what the rates of having a license are among whites and non-whites, their fractions of the population in the relevant age group, and put those together to show that the overall decline follows from the changing demographics rather than changing license rates within whites (or non-whites).

    You said non-whites are less likely to have a license, so let's put the white license rate at 55% in 1990. To balance out to 52% overall, that makes the non-white license rate 40%. You also said the white rate isn't changing, so we'll keep the white license rate at 55% for 2000 and 2011.

    Now we have two choices for the non-white license rate. You didn't say whether you thought it was staying the same, just that their greater numbers were overwhelming whites, or whether their license rate was falling.

    Assuming it stayed the same at 40%, the overall license rate would fall from 52% to 51.6% to 51.4% over those 20 years -- not even a full percentage point, when we needed it to drop to the 40-45% range.

    Then assume that the non-white rate changed, while the white rate stayed the same, and getting them to balance out within each year. This requires the non-white license rate to go from 40% to 25% to 1%.

    ReplyDelete
  8. So neither quantitative argument about racial demographic change holds up. This is sadly a general result -- racial change has not been that great to explain the decay of our country. And obviously not in all other Western nations, many of which have hardly any non-whites, or even non-locals, like Scandinavia. Yet across the entire West, it's gotten less exciting, trusting, and independent since the '80s, particularly among young people. Jeez, even the Swedes made some bouncy dance music back then.

    In general, if you're hardcore into HBD, you probably hail from a diverse part of the country and find it plausible to blame the decay on that. But go to New England, or the Plains or Mountain states were there are few NAMs -- there too things have fallen a lot since the good old days. Heartland America isn't as free-wheeling and fun-loving as it was during the age of Footloose.

    ReplyDelete
  9. It was assumed everyone would get a license when I was that age, but I lived in the suburbs. Now that I live in the city I sometimes meet people who never learned to drive. How do urbanization patterns match up.

    I found the reference to airports odd, is it assumed that's how kids get to college? I had to fly (one of surprisingly few times for me post 9/11) checking out West Point, but even that point in time I was pretty sure I'd attend a school in my own state. The modal student is attending junior/community college or a commuter school, not some nationally known school they have to fly to.

    Is there some method of getting away with beer that I never heard of? When I was a teenager there were two steps 1: Drive drunk 2: Hope you don't get caught.

    ReplyDelete
  10. "But go to New England, or the Plains or Mountain states were there are few NAMs -- there too things have fallen a lot since the good old days."

    You're talking to someone who lives two hours from the Canadian border in Maine...

    "'The rise and the later fall need not have the same cause.'

    That goes against Occam's Razor."


    It drives me crazy when people misuse Occam's Razor. Occam's Razor doesn't say "seek the simplest possible explanation, complexities be damned" it says that "entities should not be multiplied beyond necessity", which could be also put (to paraphrase Judith Harris) make things as simple as you can, but no simpler. It certainly doesn't mean, with all due respect, burying your head to sand and not looking for more data, when you can, and when it'd be useful and clearly relevant.

    "Actually over the period of declining licenses, blacks have been pushed out of the cities by more affluent whites moving back. Part of the falling-crime pattern -- less dangerous cities attract more whites, while rising-crime sends them running for the suburbs."

    Just because Blacks live in the suburbs doesn't mean that they drive.

    "Assuming it stayed the same at 40%, the overall license rate would fall from 52% to 51.6% to 51.4% over those 20 years -- not even a full percentage point, when we needed it to drop to the 40-45% range.

    Then assume that the non-white rate changed, while the white rate stayed the same, and getting them to balance out within each year. This requires the non-white license rate to go from 40% to 25% to 1%."


    Fair enough point.

    Note, for the record, I'm not saying that you're necessarily wrong. All I'm saying is that I'd like to see this broken down by race to see just how of this is due merely to changing demographics.

    Before we start getting too deep proclaiming the implications of these data, would it not be prudent to simply like to see what it looks like broken down by race?

    Additionally, even if there is a decline in White teen license rate, how much that is an effect of the aforementioned urbanization?

    After we've factored that out, we may have to consider the decline in purchasing power of the average White family, thanks to stagnant wages. That, more than anything, I suspect is the likely culprit.

    ReplyDelete
  11. " it says that "entities should not be multiplied beyond necessity""

    Right, I've got one variable explaining the up-and-down pattern -- the crime rate. Not getting a license is a form of cocooning behavior. Every form of cocooning behavior that I've looked at with data over the past 2-3 years shows the same timing, so they're all really reflections of a single underlying variable.

    Cocooning (or being out-and-about) interacts dynamically with the crime rate in a predator-prey type of way, producing oscillations (or "cycles" in an informal sense).

    So with only having to appeal to one variable, the crime rate, I can explain most of the up-and-down pattern in cocooning/sociability. You don't get more bang for your buck by introducing two, three, or more explanatory variables in the case of cocooning.

    The formalization of Occam's Razor is one of those Information Criterions, like the Akaike Information Criterion. It rewards a better fit to the data, but it penalizes the number of parameters in the model. When you include more parameters, it has to be worth the cost, or you have to get more bang for your buck.

    Capturing more complexity of a real-world pattern is not necessarily better -- otherwise we're stuck with no repeating patterns over time and space, because each case is too complex to look like any other.

    ReplyDelete
  12. "All I'm saying is that I'd like to see this broken down by race to see just how of this is due merely to changing demographics."

    If the data were out there, it might be interesting to look at, however they aren't broken down by race.

    The larger point, though, is that it's not necessary to go that route, and is generally not only non-productive but misleading. The reason is that the West as a whole shows so much similarity in changes over time, despite how they rank with respect to one another at any given point in time.

    Yet they don't have anything like the changing racial demographics of America. If Swedes became more exciting from the '60s through the '80s, then more boring since, that's not due to racial changes (it would not be here either, since the demographic change has been one-way, unlike the up-and-down pattern in sociability).

    And you see the same thing as you go back through the centuries in Western Europe. Periods of rising homicide rates have very similar cultures, and so do periods of falling homicide rates. Again, with no change at all in the non-white population since there were none.

    Why does 1580-1630 feel so similar to 1780-1830, and the Age of Reason feel so similar to the Victorian era? It's not changing racial demographics.

    So, looking more closely at racial factors even within America won't buy you much, unless you've found a one-way trend. But especially when you try to capture as much variation as possible, i.e. by working Europe into the picture as well, then it buys nothing.

    I think this is why the Scandinavians and Nordic regions have outclassed us in things like behavior genetics. In America, if you're liberal you don't want to hear about genetics at all, and if you're conservative you're likely to fall into the primeval way of blaming The Other.

    They don't have The Other in Denmark, so they are more used to thinking about what distinguishes one Dane from another within the same time, and what makes the Danes of yesterday different from the Danes of today.

    ReplyDelete
  13. "I found the reference to airports odd"

    Not just for college, but every reason a young person would have for flying.

    "Is there some method of getting away with beer that I never heard of?"

    Well, like showing up to a party or get-together where there will be alcohol but not bringing any yourselves. Either bring something else, or throw in a little cash for the communal beer fund before/after.

    ReplyDelete
  14. "we may have to consider the decline in purchasing power of the average White family, thanks to stagnant wages."

    Inequality, wage stagnation for the median worker, etc., began increasing in the early-mid 1970s. If that were the culprit, we'd see a change very fast -- if you don't have the dough to make car payments, insurance, gas, etc., you respond almost right away to declining wages.

    But sociability doesn't start turning down until the early '90s, about 20 years later.

    Same on the flip-side. Declining inequality and rising wages for the average Joe should've made the later '40s and the '50s more free-wheeling and rambunctious. Once you've got all that extra cash, it should burn a hole in your pocket for things like going out and having fun.

    But it took until the late '50s for it to even get started. The later '40s through roughly the first half of the '50s was about the nadir of being out-and-about in America.

    Going back even further, rising inequality should've made the Roaring Twenties rather dull and anti-social. Yet that was the most outgoing period within 50 years on either side. It was the '80s before the '80s.

    ReplyDelete
  15. And just to round that out, only the homicide rate shows a similar pattern from 1900 through today -- up through the early '30s, down through the late '50s, up through the early '90s, and down through today.

    ReplyDelete
  16. "Inequality, wage stagnation for the median worker, etc., began increasing in the early-mid 1970s. If that were the culprit, we'd see a change very fast -- if you don't have the dough to make car payments, insurance, gas, etc., you respond almost right away to declining wages.

    But sociability doesn't start turning down until the early '90s, about 20 years later."


    Actually, think about it: Your model (looking at licensing and ignoring "sociability", unless you have other data for that point) follows the wage pattern delayed by 20 years (one generation).

    The people who gave us the Baby Boom were the children of the Great Depression, and they put their high wages into having large families. This was perhaps aided by the fact that the sense relative deprivation was likely at a low point then, since these people grew up in a world where everyone had little (hence, felt less of a need to acquire material goods).

    Fast forward 20 years. The Gen X'ers come online, and, coming from families where resources were abundant (albeit, starting to decline), felt a stronger need to keep up with the Jones. Hence, everyone gets their license. Indeed, for a teen, there is a deep practical reason for this: this is the age when they are beginning to compete for mates, and being able to have the basic markers of status are key, especially for males.

    In another 20 years time, the Gen Y's come about, coming from families where resources were not so abundant, having a driver's license (and a car) was less of a social necessity because fewer of their peers have such.

    "Going back even further, rising inequality should've made the Roaring Twenties rather dull and anti-social. Yet that was the most outgoing period within 50 years on either side. It was the '80s before the '80s."

    How exactly was this measured?

    ReplyDelete
  17. " and being able to have the basic markers of status are key, especially for males."

    No, cars don't make you more sexually attractive to teen girls, though more people might want to be your friend

    The real reason Gen Xers bought cars was because they needed them to drive to each other's houses and have sex.

    -Curtis

    ReplyDelete
  18. Rising gas prices, increased urbanization, the rise and fall of external media (going to the movies becomes downloading a film), stricter enforcement of drunk driving legislation, higher insurance rates for young people, more families where two parents work (needing their cars more often), increased environmental awareness, and lower total family income (selling the old car instead of giving it to the kid) -- these are the reasons I personally would not insist on my child getting a driver's license.

    Occam's razor is a useful principle, but we are talking about a hugely complex society that has changed radically over the last fifty years. There are innumerable trends that, taken singly, will match up with the rate at which teenagers get driver's licenses.

    There are certainly many pressures favoring an increase in the number of licenses and many pressures against. The balance has shifted, but it is extremely unlikely that only one factor was responsible.

    ReplyDelete
  19. Cars give teenagers a whole lot more freedom than all alternatives, and again they're generally paid for by the parents because they're driving the family car.

    Maybe Gen-X and Boomer parents are too selfish to pay the increased premiums and costs?

    Seems consistent with them being more narcissistic and helicopter parenting than their Greatest / Silent Generation parents.

    ReplyDelete
  20. Re: the Razor of Ockham,

    Driving licence rates are different between states and countries in a way that is not simply explained by crime rates.

    Thus a multi-variable explanation is obligated as a necessity (to explain driving licence rates) and if it explains the local diachronic change in the US, then that's one thing it should do.

    ReplyDelete
  21. Also, the growth in the number of driving licences increased steadily through 1930 to 1960, with the rate of increase being fundamentally alike to the 1960 to 1990 period, and unlike the flatline thereafter.

    (Although interestingly, vehicles per capita increased slightly more between 1960 to 1990. I expect cars and driving were getting relatively slightly cheaper during this period.)

    So how does this match up with a "violence and crime" explanation?

    1930 to 1960 should have a low velocity of licence increase if that were the case. But they don't.

    A crime rate explanation that doesn't correlate outside a narrow range even within a single country...

    ReplyDelete
  22. "Maybe Gen-X and Boomer parents are too selfish to pay the increased premiums and costs?"

    Nah, they pay for their kids' tutoring, college, living expenses once away from home, boarding when the kid returns after graduating, expensive cell phone units plus the monthly costs, etc.

    ReplyDelete
  23. "Driving licence rates are different between states and countries in a way that is not simply explained by crime rates.

    Thus a multi-variable explanation is obligated as a necessity (to explain driving licence rates) and if it explains the local diachronic change in the US, then that's one thing it should do."

    It's not "obligated as a necessity" (lol) because the rates are not what is being explained -- it's the derivative, or change in the license rates. I keep emphasizing rising vs. falling, whether for driver's license rates, crime rates, or whatever.

    I used to think that this focus on rising vs. falling was unnatural for the human mind, that that's why so many people misunderstand what the point is.

    But now I think it's totally natural, since human beings didn't evolve in a world where they could quantify things very accurately. They wouldn't know whether 20% or 30% of their fellows had some trait, but they sure would notice if it felt like they were more and more likely (or less and less likely) to have the trait compared to earlier on.

    It's some kind of nerdy insistence on measurement and quantification as an end in itself that makes modern people focus more on point estimates than on derivatives. Why is the temperature 17 degrees today instead of 15 or 20? Who knows? But why is it rising or falling? Probably because of seasonality.

    ReplyDelete
  24. "1930 to 1960 should have a low velocity of licence increase if that were the case. But they don't."

    Don't think like such a nerd. Obviously there's a longer-term upward secular trend due to the automobile's invention, commercial introduction, and lowering prices through firm competition.

    No one should have to spell out that we're talking about the de-trended pattern, i.e. the interesting pattern not due to obvious things like the invention of the car.

    ReplyDelete
  25. "(looking at licensing and ignoring "sociability", unless you have other data for that point"

    Like I said, you'll have to go through what I've written over the past several years. It's an ongoing project. Park visits, playing on an amateur softball team, young people going out to the roller rink, video game arcade, or other gathering spot, the rise and fall of malls, etc.

    "It was the '80s before the '80s."

    How exactly was this measured?"

    The dance craze of the '20s, dance marathons even, petting parties, speakeasies, the automat and cafeteria and diner vs. the drive-in restaurant, the drive for everybody (not just jocks) to play a sport, and similar things.

    ReplyDelete
  26. "Your model ... follows the wage pattern delayed by 20 years (one generation)."

    Driving a car is a here-and-now thing, though. Both the Baby Boomers and Gen X were part of the rise in teen license rates, yet the Boomers had parents of the Depression era, while most X-ers had later Silent parents who grew up during the prosperous mid-century.

    On the flip-side, Gen Y has Boomer parents (who grew up during the prosperous mid-century), while Millennials have Gen X parents (who grew up during the stagnation phase since the '70s). Yet both Gen Y and Millennials are part of the falling license trend of the '90s through today.

    So, differences in parental generation don't seem to help here. It's more what the zeitgeist is. From the parents' point of view, it's irrelevant what the world was like when you were growing up -- you want your children to adapt to the world of here and now.

    So whether you grew up during the Depression or mid-century prosperity, if you had teenagers in the '60s, '70s, and '80s, you wanted them to gain more independence. That's the way the wind was blowing, and you knew that you as an isolate parent couldn't keep them from having to adapt to that trend.

    Likewise, regardless if you grew up during mid-century prosperity or '70s-and-after stagnation, if you had teenagers during the '90s through today, you're want to keep them from gaining their independence (although you rationalize it in a more flattering way).

    The free-wheeling, entrepreneurial, trusting days are gone, and you realize that you as an isolated parent can't make your kid more independent when the entire cocooning, micro-managing, Peter Pan zeitgeist is pushing against you. He's probably dead set against getting a license (or a job), and you'd need a magic wand to change him, so just let him be.

    ReplyDelete
  27. Don't think like such a nerd. Obviously there's a longer-term upward secular trend due to the automobile's invention, commercial introduction, and lowering prices through firm competition.

    So basically you're happy to handwave away explanations you aren't interested in. "One variable, one explanation" when it suits "Lots of explanations, lots of variables" when it doesn't.

    ReplyDelete
  28. I assumed the readers here were not retarded and already understood that there's a secular increase that does not need to be accounted for.

    "Lots of explanations" = 2, 1 for the secular rise, 1 for cocooning.

    ReplyDelete
  29. This is purely anecdotal, but I'm very surprised at the number of people in their 20s who cannot drive. And these are people who live in the suburbs, away from a really convenient mass-transit system. Some of these people (most of them being women) are pushing into their late-20s now.

    ReplyDelete
  30. Matt Strictland2/4/13, 5:11 PM

    This doesn't surprise me much but I think its more related to lifestyle changes than anything else.

    What I think is causing the decline in even bothering to get a license is there really aren't public spaces people need to go to.

    There aren't hangouts and activities that were done in public spaces such as viewing movies are these days often done at home. Even things like getting together with friends and a radio are discouraged.

    Take the 80's stereotype of the mall rat, thats gone now many malls do not allow any minors to enter then without adult supervision

    Forget going there as a work place.

    A mall --name withheld-- where I worked for a time has staff in their 20's and 30's and of all the fellow employees at various I knew (quite a lot) only one was under 18.

    Shopping too is often done at home.

    While people still go to say video game stores, a lot of games are purchased as downloads, books bought from Amazon music from I Tunes and so on,

    So baring needing a car for work or college, why bother?

    Its a big social change and given the cost of energy, maybe a good one. Well unless you sell cars I guess.

    In the long term if things keep up like this we may end up a society without much of a sense of public spaces. This will have interesting consequences but thats for another post.

    ReplyDelete
  31. Hi, I'm currently conducting research in San Mateo County on this topic. Working in the transit industry, this phenomenon has gone largely unnoticed with no attempt to understand what is causing transit's number one competitor, the auto industry, to lose a significant share of their market. This may be a lost opportunity for transit systems to capture teenagers by planning and marketing transit service to meet their needs. Yes, teenagers are traveling much less with social media connections to friends, freedom to find themselves through the information highway, and more pressure to focus on academics. But, if you can provide a meaningful transit service for teenagers now, even when they eventually decide to get their license the'll be more apt to ride more often, vote on transit measures, or support transit as an elected official.
    https://www.facebook.com/SanMateoCountyTeenageMobilitySurvey

    ReplyDelete

You MUST enter a nickname with the "Name/URL" option if you're not signed in. We can't follow who is saying what if everyone is "Anonymous."