No, this isn’t about the California wildfires. It’s about the Root IPO. Confused? Don’t worry, keep reading. I explain it all.
So I was going to write a detailed page by page guide to the Root S-1 like I did for Lemonade, but two things got in the way. First, the document just doesn’t have all that much detail.
Second, the Insider did a great summary already and addressed most of the topics I was going to write about. So, instead, this will be more focused.
Root has two big questions to answer:
- Will they ever be profitable?
- If so, can they prove it before they run out of money?
All the other stuff about the power of telematics and the flywheel and are they really different from Progressive or not…none of it matters. At least not until they prove they have a model that can produce an underwriting profit.
How Root Loses Money
Root is trapped in a box of their own making. The whole reason to be excited about Root is their pricing algorithm. Yet, they’re running over a 100 LR with little improvement as they scale (while it was better in Q2, that appears to be Covid driven).
Normally, you might say, well, that’s not uncommon for a startup. They’ll figure it out as they refine their pricing and get more experience. But they’re already on their third version of the pricing model – which they say is 10X more accurate – and the results have hardly changed!
Think about that for a second. They improved the power of the model 10X and it improved the LR 0X!!!
That suggests one of two things. Either, the original model was of so little value that improving it 10X still barely moves the needle (e.g. it was worth 0.1% on the LR and now it’s worth 1.0%) or, more likely, they returned all the benefit of the improved accuracy in lower prices to drive growth.
Would You Rather
So my question to a prospective Root investor is would you rather…
- hope the algo really isn’t that good yet, but maybe it will get better someday and loss ratios will drop by half or…
- believe the algo really is as good as they claim but they are purposely underpricing business to show investors top line growth, aka the Uber strategy aka the Insurance Hare.
The issue with the first option is if they’ve gotten it wrong for four years now, why should I believe it will get better in the future?
It also suggests that management overpromises and can’t deliver which begs the question of why should I believe any of the other aspirations they claim they can achieve…in which case, I should probably value it around book (which is actually a negative number at the moment).
If you pick door #2, then the only way to produce adequate returns is to assume some day they can raise prices 50-100%…without any sales or retention impact.
Uber is still trying to figure out how to do this (maybe Root can copy Uber’s move into the delivery game and offer Pizza Insurance???), so it’s not clear what playbook Root will emulate.
However, we know insurance has inelastic demand, so they can’t grow the pie. All they can do is steal share from others. Doubling your price isn’t a recipe to do that.
Furthermore, there is nothing about Root that can’t be copied. Others have telematics offerings (Root’s claims of preeminence notwithstanding) and others can create a mobile app. As I have argued before, they are a product, not a company. There is no obvious path to price elasticity in the future.
So if you asked me to choose, I’d pick door #1. Hope that there is still some amazing pricing discovery to come in the future because the Uber idea of giving away the product now to raise prices later isn’t going to fly in insurance.
Betting On the New Business Penalty
To be fair, Root does make the case for there being a third door…the famous new business penalty!
Now, is there such a thing as a new business penalty? Most certainly. That said, has every auto insurer that’s drastically underpriced their book always initially blamed it on the new business penalty? Yup.
So is there a case for dramatic improvement in Root’s LR as the book matures? Let’s investigate…
First, to have a meaningful improvement from new customers seasoning, you first need to actually keep those customers around. Depending on whose analysis you want to believe, Root’s annualized retention is somewhere between 33-50%.
If the new business penalty was that significant and they could offer better prices on renewal, you would think retention would be a lot higher. Also, let’s remember a key premise of Root’s model…the Test Drive!
They use the initial test drive period to weed out the worst drivers from ever becoming customers. This allows them to avoid the worst 10-15% of the driving pool. In other words, they should have NO new business penalty.
The reason the new business penalty exists is because you inadvertently write bad drivers…because you don’t have enough information about them! Root claims to have best in class information about their customers…so why is the new business penalty so large???
Nonetheless, let’s give some benefit of the doubt. Root says loss ratios decrease 27% over a two year period (given the age of the company this is likely a small pool but I’ll take it at face value for now).
Now, the whole book is running over 100% and some of that is already renewal, so I would guess to get to the right weighted average it’s probably something like new runs at 120% and renewal runs at 93%. Even if the whole book gets to 93%, we still have a problem.
Room For Improvement
We’ve established improving the loss ratio is going to be a big lift. But there’s another problem. The expense ratio is close to 100! But maybe they can improve that with growth? Let’s take a look…
The obvious first place to start is customer acquisition. The marketing ratio last year was about 40%. While that declined in 1H, that was heavily influence by cutting back spending due to Covid.
So let’s look at the FY19 result. They spent $100M on marketing. While it’s reasonable to think that doesn’t scale 1:1 with growth, we do know that $100M is peanuts in the consumer insurance space.
To get to where they want to be as a company, they’re going to need to spend at least $1B on advertising annually. So can the ratio come down from 40%? Sure. Is it likely to get under 25% anytime soon? Doubt it.
How about G&A and technology spend? Certainly, those will scale with growth! Not so fast, my friend. Root suggests that, while in the long term they will decline as a % of revenue, they will actually increase the ER first in the near term.
The other category of expense is, well, “other” which is admittedly harder to forecast, but saw less growth in the 1H than the other categories, so perhaps assuming some leverage there is reasonable.
Put that all together and it appears the ER can improve to the 65-75% ratio over the next few years. Maybe an optimist will say 50%??? Regardless, it’s going to be a long time before it’s competitive with the industry.
Fiddling While Root Burns
While Root fiddles with its algorithm, cash is being lit on fire daily. With a CR of roughly 200%, every $ of premium written destroys a $ of capital. This is what we call unsustainable.
Even if they can get that LR down to 93% and the ER into the 60s, it’s still 150% CR so burning 50c for every $ of premium they write. This isn’t Buffett’s concept of float. This is inverse float!
A customer might sign up, not because they are risk averse and are willing to pay a premium for diversification, but because they know they rationally see the expected value of what they will get in claims payments is greater than they will pay in premiums! That’s one way to acquire customers!!!
But in all seriousness, Root is hemorrhaging cash with no end in sight. Back to our question #2, can Root make $ before they run out of cash, the answer is “unlikely”. Let’s look at the numbers.
For 2019, net income was -$282M. Cash used in operating + investing (more on why I use both below) was -$241M so about 85%. Earned premium was $275M so about the same as the net loss.
In 2018, off of a smaller base, EP was $40M, cash used was $47M, and net loss was $69M.
And for 1H 20, net loss was $145M, cash used was $162M, and EP was $233M. Note, my best guess is net loss and cash would have been at least $50M higher ex-Covid.
So, all-in, we have a pretty reliable relationship = $1 of EP = $1 of net loss = $1 of cash burned. Again, some improvement can be expected over time, but there will be a burn associated with growth for the forseeable future.
Note, this all ties out pretty well with their fundraising history. In early 18, they raised a $50M Series C which funded 2018 premiums. Their Series D later in 18 covered 1H 19 writings and their 2019 Series E covered growth through 1H of 20. Now, they are raising capital again.
| $ Raised | Subsequent EP | |
| Series C 3/18 | $51 | $40 (FY 18) |
| Series D 8/18 | $100 | $99 (1H 19) |
| Series E 8/19 | $350 | $410 (2H19/1H20) |
Aside: I used CFO + CFI to estimate burn earlier since a) this is what they actually need to finance and b) using just CFO only makes sense for a steady state company. CFO currently benefits meaningfully from growth (the IBNR is added back to net loss) and the investments being made with the float will be paid back out in a year or two, so CFI is essentially a proxy for next year’s paid claims.
The Capital Flywheel
Why does all this matter? I mean hopefully this is obvious, but to drive the point home, Root’s current business model is more like Tesla’s than an insurance company. It needs constant capital to continue growing.
Sure, all insurers need capital to grow but that is to fund capital requirements (which typically can be covered through earnings), not to cover losses.
The one piece they leave out of the flywheel analogy is that most insurance companies have a capital flywheel. Growth produces earnings which grows equity which funds growth and, oh, by the way, the premium creates positive float which produces investment income which produces more earnings.
A decent rule of thumb is you can grow top line the same as your ROE, so a 10% ROE can support 10% top line growth. If you ROE is -100%, well, you have to shrink to $0 unless you raise more money.
This means Root’s capital flywheel goes something like…ask investors for $ to chase growth, lose $ producing that growth, ask investors for more $ to chase more growth, lose those $ too producing even more growth, etc.
See the problem? One day investors say “no more” and your flywheel cracks. That’s not a business model. That’s a hope.
Now, of course, this ties back to the whole Tortoise vs. the Hare thesis and that as long as Root can be a Hare and have free capital, it has a chance to win the game. But let’s be clear, it’s a confidence game. If investors ever lose confidence, the game is over. Until then, burn, baby, burn…