Used car depreciation cliff--is it a fable or real?

Question

I keep reading about this sudden drop in value on transferring a vehicle from new to used. I'd like to ask the validity of this, how do we know it to be true? Is this an actual phenomenon backed by publicly shared figures or something more like a myth, rule of thumb, or just simply case by case and anecdotal?

Just poring over this site we see statements like the following:

A new car loses about 30% of its value the second you drive it off the lot. source

Couple thousand when you drive off the lot and half that in 3-5 years source

This, along with the "Drive-off the lot depreciation"-factor is why my dad told me to always buy almost-new cars instead of brand-new. source

A brand-new car arguably loses a large percentage of its value the moment it leaves the dealer's lot. source

That's the same amount you would instantly lose on 2010 model after it leaves the dealership. source

BRAND new car loses thousands in price when it drives off the lot. source

A new car loses 15-20% of its value the second you drive it off the lot. source

Note that car (especially new one) is one of things that lose values very quickly - so you can't really go with "I'll buy it and see how it goes, and sell it if it is not working financially" - you'd easily lose 20-30% of the value the moment you buy it. source

Given that a car loses at the very least 10% (a very conservative estimate, usually much more) in the first few months of owning it source

But if you can show me some actual examples of people selling a year old car for virtually the same price as new, I'd be very impressed.. also very interested because that seems like a great deal for me. Might depend on the country, but it's certainly not the case in the US or Austria source

That sounds like a new-car price. If it is, you can kiss $4k-$5k of that price goodbye the moment you drive it off the lot. You'll pay the worst part of the depreciation on that vehicle. source

This answer gives the most context regarding justification for this, it basically boils down to risks from unknown history that apparently come with a steep hit accordingly:

Buying products (not just cars) new means you are getting a product straight from a trusted supply chain. Buying used, even nearly new, raises a whole bunch of questions about the item. Has it been abused in some way? does it have some hard to pin down problem? Has it been involved in some form of crime? are the manufacturers warranties/support packages fully transferable?

You can try to answer these questions of course, but there is both a cost to answering them and an uncertainty in the quality of the answers. So in normal circumstances, used products, even nearly new ones, will trade at a significant discount over new products. source

But I'm curious about how people know this hit translates to thousands of dollars, and even 30% of the buy price of a car or any other percentage. How have these figures become part of our common knowledge and how can we verify any of this remains true?

Answer 1

An Empirical Calculation of Actual Sales Data

This aggregate statistic could be calculated independently if one had access to VIN level sales data as told by a Certificate of Title form where the transfer data is registered. That form has the date, the mileage, and the sale price listed. Using California as an example we can see a sample title here https://www.dmv.ca.gov/portal/dmv-virtual-office/title-transfers/, the front has the Vehicle ID Number, YR Model, Make and mileage (the mileage now reads), the back has the fields PURCHASE DATE as well as PURCHASE PRICE OR IF GIFT, SO STATE. If each of those fields formed the columns of a table we could filter to those purchase dates separated by some threshold time interval for the same VIN. We’d have two price points at which the price delta could be calculated and expressed as a percent diff to the previous price. Collecting many matching percent diffs would allow us to estimate this depreciation cliff objectively for a good sample size, and could even group by make and model if the counts are sufficient.

So now I ask, do DMVs submit to FOIAs or anything like that. For California it would seem they do support this use case, to some extent, consider the following page

Requesting Records in Bulk

If you need to request more than 15 records at once, please contact the DMV Materials Management Section. https://www.dmv.ca.gov/portal/customer-service/request-vehicle-or-driver-records/

I need to follow through to ask what scale they support precisely.

Price Based Analysis

So it appears KBB tracks pricing data rather than pure sale data according to this source:

How does Kelley Blue Book get its data and information

KBB employees a team of about 200 statisticians and researchers. Every day, the KBB employees review the prices of different vehicles and see what they are selling for at dealerships, auctions, and private sales. https://www.motorbiscuit.com/where-does-kelly-blue-book-get-its-data-from/

This second source further backs this pricing based analysis

Kelley assesses a wide array of different values to both used and new vehicles. For used vehicles some of them include: Trade-in value, typical listing price, fair purchasing price (CPO), and private party.

Kelley determines Blue Book values by analyzing pricing information from real-world used car prices, as well as industry developments, economic conditions, and location. https://www.investopedia.com/articles/personal-finance/113015/are-kelley-blue-book-values-accurate-and-reliable.asp

That’s probably the next closest best thing obviously if states aren’t willing to export population wide figures to the public. The drawback I see with this approach is that one never gets exposed to the actual negotiated price just a bunch of asking prices.

Still there are price diffs to find through KBB even using those asking prices. I’ll run a brief sampling off KBB with a few models to share with others soon to demonstrate this.

Fair Price Value

The price criticism is addressed through KBB which has a great page on what they call the fair price value and they share some high level details for the inputs of their pricing values (italics mine)

Fair Purchase Price The Fair Purchase Price reflects a vehicle’s actual selling price and is based on tens of thousands of recent real sales transactions from auto dealers across the United States. The Fair Purchase Price is not calculated or based on a proprietary formula; instead it is derived from actual new vehicle sales and extensive knowledge of the marketplace.

So KBB does have some component based on actual sales figures, if even just a subset of all sales in a state which are those which passed through a dealer. That’s very impressive to build those relationships with so many sellers to accomplish this. This probably isn’t transparent as open data would be but it’s still worth knowing about.