The total amount of new vehicles sold in the U.S. in 2015 was 17.5 million [1]. Since the average is $33666, we can assume half are above, and half are below that price. In reality, there are far fewer than half above that price, since an outsized sale price pushes that average up.
Half of 17.5 million is 8.75 million. The U.S. Census estimates the 2015 population was 323,995,528 [2], with 116,211,092 households [3]. With every household replacing their car every five years or so (total guess), that leaves us 23.2 million households that should have been buying a car in 2015. That means that about 37.7% of the eligible U.S. households bought a car worth over $33666.
That is bizarre enough that it leads me to believe there is a serious fault in my methodology or it just so happened that car sales were abnormally high in 2015 due to an enormous rise in consumer confidence. Either way, I'm more confused now than when I started scribbling on my metaphorical cocktail napkin.
[1] https://www.google.com/search?q=site%3Awsj.com+U.S.+Car+Sale...
I didn't understand why you split the market into 'above-average' and 'below-average' prices though? I highly doubt the distribution is symmetric -- most of the probability mass is probably in cheaper cars, and there are smaller numbers of very expensive luxury vehicles pulling up the average price [1]. In that case, a much smaller number of households buy the 'above-mean-price' vehicles. Also, wealthy households tend to replace their cars more frequently; luxury car dealerships do much of their business in rolling people over every 3 years into a new vehicle lease.
In any case, 116mm / 17.5mm cars per annum households is 6.6 years per car per household, which seems totally reasonable to me since many households own multiple vehicles, and many vehicles are owned outside of households.
[1] http://fortune.com/2015/02/17/luxury-car-sales/ - 'Top-shelf brands, with cars generally selling for more than $40,000, represent a big and growing slice of the car business. They now account for 11% of total new car sales, when measured by number of units sold'
This context helps to explain what's going on, especially the cars per year metric you added. It would be nice to have more information, but it appears that 11% luxury segment can have an outsized effect on the mean sales price of the vehicle sales.
In a real interview, you normally wouldn't have access to a calculator or a computer. The point is to show you can arrive within a factor of 2-3x of the actual real-life answer just by smart guesstimating and doing arithmetic in your head.
One reason is because consultants often get asked these sorts of on-the-fly questions by clients, and they need to provide reasonable answers without looking them up.
Anyway, if the technique is good enough for Fermi [1] or Drake [2], what's the problem?
Even in 2008-2010 annual sales never dipped below 10M.
Detailed stats - https://www.bea.gov/national/xls/gap_hist.xlsx
116,211,092 * 2 / 15 ~= 15.5 million new cars per year which is within the ballpark of 17.5 million new cars.
Next, averages tend to be well above median. One BMW might bring the average up relative to 20 Honda Civics. Further, the low end of the market is more likely to buy used than new. So ~1/3 of the market is buying new cars, 15% of that is buying the expensive new cars and your looking at top 5% income earners at $167,000+k/year.
Granted, not everyone making 150k is buying a 50k car, and some people making less than 150k are buying 50k cars. But, if you think in terms of that kind of income then a 50k car is less ridiculous than you might think.
Maybe car sales are highly variable from year to year?
Having multiple, older vehicles that are 1. fully paid off, 2. dirt cheap to insure, and 3. old enough that you can actually work on them yourself is a good way to make sure you have something usable at any one time, even if the general maintenance needs are higher than average.
PS: Don't forget all those Verizon van's are part of that 2 average.
My new neighbors, though, are ridiculous in their car ownership (in my opinion). They are a married couple with only a new baby, but for some reason they have three cars. I've never seen them drive the third car, it's just in their garage and they park the two cars they do drive in the driveway.
Besides, there have been a couple of times where simply due to bad luck both of our newer cars were in the shop at once. Having the backup of the CVCC for a few days was genuinely useful.
Many suburban empty-nesters I know have three. I don't know why that's popular or makes any sense.
On the other hand most urban twenty-something's I know don't have any, so I'd agree the 2 car average seems possibly high.
Ye gods. :|
Mom and dad's daily drivers + dad's baby and/or project.
If you get outside of urban areas, car ownership is much higher, since you can't get anywhere without them.
Take a look at the size of high school and commuter college student parking lots on Google Earth, and then consider that every single one of those cars probably belongs to a household with 3+ cars (mom, dad, teenager).
Very common for parents to upgrade and pass their old car down to their 16-year-old instead of trading in.
Pretty much all the families I've known have at least two cars.
Now, we could get rid of the Scion and use the VW as the new DogMobile. But the Scion is long paid off, reliable as a hammer, and we don't put miles on a temperamental old VW that wasn't that reliable to begin with (in comparison to a Toyota-made Scion). The resale value on Leafs is horrible, and we like the car. Long since paid off, too, and since it's our "nice car": no dogs allowed.
Oh, and my Dad just texted me this weekend asking if I wanted his 1963 Austin-Healey 3000. As much as it pains me, I'm probably going to turn him down and let him sell it. Too many vehicles the way it is, and nowhere to put the Healey except a storage unit. But until I give a definitive "no", there's the possibility that we could be a four car family.
Ridiculous, yes. But for some the vehicles just kind of pile up.
A household next to me has within the last five years had three residents: mother, father, son then about twenty. They had three cars.
A household down the street: mother, father, employed daughter, employed son (the last there part time). All have their own vehicles:sports ute, truck, sedan, truck.
It is possible that a household at the end of the block, mother, father, daughter in high school, has three cars; but I'm not sure whether father and daughter use the same car or a different one.
We are in a city, close to public transportation. I suspect that as one gets farther out into the suburbs, one finds more households with three or more cars.
We're well below median household income; but we must have transportation at all times.
I have found the sweet spot for used cars to be about $4000. Won't need too much work for a couple years, and I refuse to take the depreciation on more expensive cars. Cash-and-carry, I will not have a car loan, either - not paying interest.
My ego is not in my ride.
A quick search shows there were 260,350,938 registered highway vehicles in 2014. The truth is US auto fleet age is older than it has ever been so car sales are not abnormally high and could actually accelerate (assuming robots don't drive them all!).
http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pub...
I think the issue is here. Yes, most households don't hold onto a car for longer than 5 years, but not all households upgrade to a brand new car every time. (Heck, I've never owned a brand new vehicle ever.) So one new car sale could result in 2-3 households upgrading their vehicle, as cars are passed down the line.
Maybe you're not stating a symmetry assumption about the distribution of prices about this mean, but in general you can't actually assume that.
I'm not sure where you're from, but this is Hacker News, so I'll just ask - Who are all these people who can afford to live in California? ;-)
It's all relative. Remember, the annual new car market is the number of cars on the road divided by the life of a car. This accounts for the sale of used cars as well.
People with families will probably spend a little more for a minivan or SUV, which will skew it higher. Students (or people in their 20s in the early phases of their careers) will spend less, balancing out the minivan purchasers. Totally guessing though.
As long as most middle class American don't total their cars at the end of 4-5 years, I'd say they probably get 1/2 of that value back as trade-in.
I'd wager it's more like 8-15k every 4 to 5 years in installments of $300/mo for those staying in a perpetual mode of making car payments.
And that is assuming you aren't leasing, which a lot of people do. No trade in value there.
For leasing, you aren't paying for the full cost of the car either, in fact, depends on the company you could actually come out ahead when compared to purchasing.
I've had friends lease recent model cars on 3year/12K terms and get about 60% residual value on their term sheet. I think 1/4 to 1/3 is an extreme case, unless you're choosing really crappy manufacturers that are known to fall in value quickly
Honestly, if you consider $1,000 on new tires to be a frustrating expense, then you really can't afford to purchase new cars in the first place. A new vehicle is a multiple of the cost of new tires and even more so when the balance of your last vehicle is rolled into the new financing and spread out over payments with interest. I worked at a car dealership while in college. There's a reason the "finance guy" drove a 10+ year old Honda that he paid cash for.
http://www.mrmoneymustache.com/2011/11/28/new-cars-and-auto-...
http://www.financialsamurai.com/the-110th-rule-for-car-buyin...
Also, there is a certain value in having a reliable car with predictable monthly costs. Sure, you can get an extended warranty to cover unexpected expenses, but there is still the hassle factor of having your car constantly in the shop after a certain point.
There's nothing wrong with buying a car and keeping it for 10 years, but there is also nothing wrong with having a new car every X years.
You're right of course about features, including safety features. Personally, I'm not going to get rid of a reliable running car for most new technology. But then I've always been a drive car until (or a little bit past :-)) the point where it's not reliable transportation any longer.
Really depends on the car. It isn't uncommon for some sets to breach $1000, even at Costco.
They eyeroll at my 2002 Honda, and I eyeroll at their $1000/mo payments.
To top it off, when you're financing, used car loans are rarely as attractive as manufacturer/dealer incentives - BankRate says they average about 3% annually nowadays, so an 18% mark-up to the value of the used vehicle over the 6-year ownership period.
Out of 330M people even if 5% of population is buying a new car in the given year that is whooping 15M cars per year. USA sells around 17M cars year +-2M. One simple proxy to arrive at the 5% figure is the number of people in 16-19 age group where they might get parent's old car and the parent might buy a new one. That population is around 20M in USA.
Of course I have not accounted for taxi fleets, car rental fleets, corporate purchases etc. In fact go to think of it 17M cars per year does look like a very low number.
Also, I'd like to think your average developer is more financially literate/savvy than your average American. Otherwise, the housing crash wouldn't have happened. You know, people with no money buying expensive houses. People carrying credit card balance and rolling them over.
Those people exist, and they are many.
You may like to think so, but I have seen absolutely no evidence of that...
How much do you drive?
What kind of vehicle do you want?
How long do you plan on keeping the car?
Etc...
Here's one case: I just bought a new Tundra a couple months ago (cash because I don't like debt). I could have saved $5K (15%) and gotten one 3 years old with 40K miles on it.
That's the problem with some models. The resale value is SO good that the value proposition for buying used is almost nil.
Since my trade-in Tundra was 10 years old with 178K miles, let's assume that's the life of the new vehicle too. I don't think saving 15% to buy a vehicle that's 20% through its life span is a good deal.
With unlimited resources you can keep anything on the road. But 8k is about 25% the cost of a new truck, so you'd have to be bonkers to put that much into something that old.
No, my assumption is fine. Oddly, yours is flawed. My assumption doesn't use national averages that include southern states and non-coastal states where cars last forever. In my unique living conditions, my assumption is fine.
Again, there are more factors at work here. The trucks practical life was over, not because it couldn't be repaired, but because it didn't make sense. When a repair bill becomes greater than the value of the car, you stop repairing it.
http://press.ihs.com/press-release/automotive/average-age-li...
There are some tax advantages with leases that some people can take advantage of, but for most people, leasing is not going to save them money but it does provide a convenient way to always have a new car.
For Maserati, some of their models actually carry a negative money factor, so you are being PAID to lease. You can lease an $80k Ghibli for $600/month for 3 years, that's way less depreciation than if you bought and sold the car after 3 years.
BMW only usually only does about $3k downpayment anyway.
Assuming you want to keep driving new cars, that is.
Between low-balling used-car dealers and Craigslist randos asking if you'd take personal checks, for many people selling a car is more stressful than buying a car.
There's an assumption there about getting higher residual value on a used vehicle than what the dealer has in the lease paperwork.
The marketplace for used cars (before Beepi and Shift) is hardly efficient, and you might extract optimal price if you're willing to drive to multiple used-car dealerships or meet random people off Craigslist at random hours to accommodate test drives.
Some people would rather not deal with that, hence leasing that provides a guarantee a used vehicle will be accepted at pre-negotiated residual value.