One of my long term favourite firms.
It’s not dirt cheap yet, but it has been falling. Underperforming the market pretty steadily since the start of December.
Down about 26% from 52 week high so far.
For anyone following valuation via PE, which generally works for them as they are usually such a steady earner,
be sure to go through the quarters and adjust for the various extra loan loss provisions and unwindings.
I haven’t jumped in on this current dip, but watching. Maybe soon.
Jim
(already long, but not nearly as much as I’d like)
Kmx profit is mostly transactional in nature. The price of cars only matters if the prices are so high they suppress demand. The wholesale cost of used cars fluctuates weekly through an auction system,making kmx a middleman business.
Kmx profit is mostly transactional in nature. The price of cars only matters if the prices are so high they suppress demand. The wholesale cost of used cars fluctuates weekly through an auction system,making kmx a middleman business.
Indeed. Like Visa.
The stock sells off quite often, when (somewhat less informed) investors think the price of used cars will fall in an economic slowdown and hurt them.
But they make about $700 per transaction, through thick and thin.
The price of the cars doesn’t matter much, as long as there is turnover.
Economic slowdowns are often good for business. More churn in the auto market, more potential used car buyers relative to new car buyers.
Others worry about their securitization, which is a more rational concern.
On the surface it looks like their balance sheet is terrible, but a lot of the consolidated debt is non recourse.
KMX brick and mortar retailer, how are they going to compete with digital competition?
By not being a brick and mortar retailer?
30% of their revenues last quarter were on line, up from 20% a year earlier, counting wholesale.
Over half the vehicles they bought from customers last quarter were on line transactions.
Their online offering is growing fast, and unlike the obvious competition Carvana, they have deep knowledge, roots, data, market share and inventory.
That doesn’t guarantee they’ll come out on top, but it sure helps.
People who want a physical used car lot will go to one, and they lead in that space.
They’ll also do well with the clients who want a half-and-half deal.
For the pure online, they’re there and it’s theirs to lose, but they’ll have to earn leadership.
Their advantage is that people know that a car bought from Carmax won’t be a lemon.
They sell all the duds at wholesale, at least the ones they can detect, so they never hit their retail channels at all.
Superficially, Carvana customers seem not to be enjoying the experience as much as they hoped. https://bestcompany.com/car-finders/company/carvana
Delivery seems to be the big weak spot…because they’re essentially on-line only and don’t have a physical presence where needed.
I find it surprising that people would buy a used car without a test drive or a physical inspection, esp. if they haven’t owned that same make and model before. Unlike new cars every used car is unique. The online sellers allow buyers to return within 7 days, but I wonder how smoothly that works, if financing and trade in is involved in the purchase.
It is interesting to read the latest KMX earnings call. When questioned about profit margins, the CEO emphasized they are focused on profit per unit, not percentage gross margin, which has held up well, even with surging used car prices. Also he mentioned their multi channel (wholesale, physical, online) which gives them an advantage over online only. And their expertise in pricing and financing used cars.
I recall that a few years ago Amazon was supposed to kill B&M grocers with same day and 2-hour home delivery. They found out that the logistics of just-in-time perishable food delivery from wholesale warehouses, substitutions for out of stock items, micro-local preferences were not easily overcome. They ended up buying Whole Foods to improve their offerings. Supposed “road kills” like Kroger and Target leveraged their well-located stores to solve the last mile delivery problem with curb pickups.
Caravana type solutions are a risk,but just another level of competition. The true existential risk,if there is one,is transportation as a service. If this concept is fully deployed,total automotive demand will drop precipitously,and personal ownership will as well. I think this is not a threat today or soon,but could be in the longer run. Twenty years or so.
General Motors Company (GM +0.2%) is launching an online used vehicle market this spring called CarBravo. While the platform is expected to largely facilitate sales to dealers, it places GM in closer competition to online sellers like Carvana (CVNA +2.9%), CarMax (KMX -0.7%) and Vroom (VRM +5.7%).
The new site is expected to debut this spring with Chevrolet, Buick and GMC dealers offering GM-branded vehicles as well as used vehicles from other automakers.
For anyone following valuation via PE, which generally works for them as they are usually such a steady earner, be sure to go through the quarters and adjust for the various extra loan loss provisions and unwindings.
Since looking up the loss triangle is a nuisance, another quick and dirty yardstick is sales per share, and by extension price to sales ratio.
There is no huge reason to believe there is a big long term shift in the net margin either way.
So the shortcut of sales lets you skip past short term vagaries in net earnings due to the pandemic,
loan loss provision shifts, or extra large investments in the online initiatives.
For whatever it’s worth, with the price at $116, P/S is about 16% below its post-credit-crunch average.
Then there is ongoing growth…earnings per share have risen about inflation+11.2%/year on trend up till now.
Obviously there are much better ways to value the firm, but it’s good to know which shortcuts are likely to work better than others.
$110 and still falling, both in absolute terms and relative to market.
Starting to edge into the “interesting” zone, I think.
Maybe sales are having a bit of a spike lately, but FWIW price-to-sales ratio is now about 20% below its post-crunch average.
Earnings are usually better, but they have been a bit volatile lately with credit adjustments.
I really should look up the operating earnings and build a better metric, but I’ve been busy.
Carmax at $116ish… … $110 and still falling, both in absolute terms and relative to market. Starting to edge into the “interesting” zone, I think.
I’m accumulating at these $107-110 levels.
That doesn’t mean it’s a good idea, but hey, at least I’m calling my shots.
It’s not dirt cheap. Not quite as cheap as it looks as earnings are a pinch elevated.
But I think there is money to be made starting here, and I like their long run past and future prospects.
In December 2014, I bought a Honda Fit right off the lot. It had 23 miles, and I paid $20,814.80, including accessories and an extended warranty. This December, a buzzy startup called Carvana drove away with my car, cutting me a check for $20,905 — leaving me with a profit of $90.20.
Not only that, but Carvana’s offer was $5,000 higher than Vroom, $6,000 higher than TrueCar, and $7,500 higher than CarMax. Carvana’s offer changed day by day, too: the final one I accepted was $1,338 higher than its lowest quote.
I knew I had everything going for me — low mileage, no accidents, and desirable trim at a time when car prices are going through the roof on a model that Honda discontinued. And yet, it sounded ludicrous. Used cars almost never sell for more than their original price, and the company knew next to nothing about me. Yet, Carvana’s algorithm had agreed to pay $20K for my car sight-unseen, even bring a pre-printed check to my door, before any inspection took place. The online quote arrived so fast, I knew a human couldn’t have been involved.
The rap against CarMax has been online only used car dealers like Carvana is going to steal business from them. Maybe Carvana will follow along the trail blazed by Zillow.
In December 2014, I bought a Honda Fit right off the lot. It had 23 miles, and I paid $20,814.80, including accessories and an extended warranty. This December, a buzzy startup called Carvana drove away with my car, cutting me a check for $20,905 — leaving me with a profit of $90.20.
Great story. I guess it’s only an anecdote, but it takes some of the sheen off Carvana, and maybe CarMax doesn’t have to be too worried about that competition.
On the other hand, isn’t Carvana just setting up a trade? The car can’t really be ‘worth’ 20k, i.e. zero depreciation after 7+ years, but if someone wants to buy it at that price, who’s to say it’s not the right price? It’s not as though Carvana is buying it to put it in its inventory. So why do you think CarMax failed to take advantage of what will likely be a short-lived opportunity?
On the other hand, isn’t Carvana just setting up a trade? The car can’t really be ‘worth’ 20k, i.e. zero depreciation after 7+ years, but if someone wants to buy it at that price, who’s to say it’s not the right price? It’s not as though Carvana is buying it to put it in its inventory. So why do you think CarMax failed to take advantage of what will likely be a short-lived opportunity?
It’s a lot like the insurance business.
At first, you think: OK, we’ll have some winners and some losers, but it all averages out over time, as outlined by Carvana in that article.
This works fine, until you figure in the fact that people comparison shop. Also as in that article.
Clients are taking your bids that are on more generous terms than the competition, and passing on the less advantageous ones.
They are actively seeking out the errors and limitations of your system compared to the system of
your competitors, and closing only those transactions that are to your detriment.
It’s a very intelligent distributed learning system: a million comparison shoppers systematically probing your algorithm for flaws or biases.
If you (and presumably your competitors) have differing systems, even if they all work well on average,
they will all do badly because transactions with random errors to the client’s benefit will be
executed and transactions with random errors to the benefit of the company will be skipped.
So, what has the insurance industry taught us about this dynamic?
First, there is an infinite demand for underpriced product. Revenues are not a problem, and short term profits aren’t that difficult either…for a while.
Second, it always bites you in the end, so the long term winners are the ones that price individual transaction risk/reward most accurately.
Third, those long term winners go through long stretches turning down bad business. Their short term revenue growth can be very bad for a while when the competitors are dumb.
This waxes and wanes with the hard and soft insurance market: the smart companies pass on bad business and wait till this cycle’s fools fade away.
This might conceivably answer your question on why Carmax hasn’t taken advantage of this short term opportunity: they know better what cars are worth, and what it takes to know that.
Since revenue growth is all the market seems to care about lately, I imagine Carvana will be very popular for the rest of the cycle.
But I suspect that at some point they will have serious profitability problems.
If customers are picking off your overbids and ignoring your underbids, that’s a much bigger issue than the cost of paying someone reasonably knowledgeable to really look at the car.
Moving to algorithm based car pricing sounds like a great idea that won’t work, because it’s based on statistical pricing.
Unlike telematics for risk assessment in insurance, which is a real game changer, because it’s based on improving individual transaction (customer policy) pricing.
I’m not sure everything above is right, but I think it’s a useful way to think about it.
Moving to algorithm based car pricing sounds like a great idea that won’t work, because it’s based on statistical pricing. Unlike telematics for risk assessment in insurance, which is a real game changer, because it’s based on improving individual transaction (customer policy) pricing.
I’m not sure everything above is right, but I think it’s a useful way to think about it.
That is useful, thanks.
But I don’t think I made my question very clear:
If Carvana is offering too much and having to sell the car at a loss, of course I agree the anecdote of paying 20k for a car that was bought 8 years ago for 20k would suggest that this is a business that will eventually fail.
But if Carvana happens to know that there is going to be a buyer for that car at 20k and is just paying that much because that is what it takes to get a product their buyer wants, even at 20k (1k more, or whatever Carvana’s markup is), then the price is still crazy, but the craziness is coming from the end-buyer, not from the middleman. Carvana is only being stupid if they end up being stuck with these overpriced cars. So we can laugh at the price, but maybe Carvana gets the last laugh, if they are just supplying what the market wants, and CarMax (thinking like you and me) doesn’t want to pay that much for what is, at least for the moment, a car in heavy demand.
But if Carvana happens to know that there is going to be a buyer for that car at 20k and is just paying that much because that is what it takes to get a product their buyer wants, even at 20k (1k more, or whatever Carvana’s markup is), then the price is still crazy, but the craziness is coming from the end-buyer, not from the middleman.
Absolutely true.
I was making the simplifying assumption that they do not have a buyer lined up when they make the offer for the car.
i.e., that they have no more than a model that thinks there’s a buyer willing to buy that kind of asset at a given price.
Or that, statistically, there is likely to be a buyer at that price fairly soon.
If there is no buyer lined up, the model might be right or not in any specific case. There will be error bars.
Let’s say it’s exactly right on average: the average offer is exactly the normal wholesale fair price for that vehicle.
My worry for them is that they are coming up with an offer price and committing to it for long enough that the client can get offers from others and shop it around.
Even if all the companies’ models are on average perfect, if each offer has a little random variation it’s be enough to crush profits for all of them.
Imaging three random number generators giving numbers from 0 to 1, and all have an average value of 0.5.
In this analogy, 0.5 represents the true fair value for the car at wholesale. All three are perfect models, statistically.
At each cycle, you receive a number from all three–an offer for your car–and pick the highest.
What’s the average number you pick? Somewhere around 0.75.
With just three such competitors, the average computer-model car buying firm will overpay by about
25% of the maximum variation in bids, on average, on every transaction where the client comparison shops.