There are two types of disruption. The virtuous kind where a new innovation improves the industry and provides a better product to the consumer. Then, there is the disastrous disruption, where a new entrant starts a race to the bottom that benefits nobody. One recent example of the latter is how Robinhood took trading commissions to zero. Today, I will address the possibility that we could see disastrous disruption in auto insurance.
The prime suspect to initiate this war would be Root, though it is possible Metromile could lead the way as well. Given Root has shown the faster growth and has the highest valuation of any insuretech, I am going to focus the analysis on them.
The Root Business Model
First, a quick primer on Root. Root is a direct auto insurer that relies on usage based insurance (aka UBI, aka telematics) to price drivers. In other words, they monitor how you drive and give you a price based on your behavior rather than allocating you to a pool of similar drivers and spreading risk. This arguably violates the principle of risk pooling that insurance is based upon, but that is a topic for another day.
The main thing to know is the use of UBI is not unique to Root, nor are they the leader. Most of the established auto insurers offer a UBI product. Progressive has aggressively promoted it over the last decade and began experimenting with UBI 20 years ago. Progressive has 27B of miles driven under UBI and as much as $10B of premium. This compares to $400M of annualized premium at Root. Progressive grew personal auto premium $4B in 2018 alone with perhaps half of this coming from UBI.
There is one unique thing about Root. They only need to observe you for two weeks before assessing your driving (compared to six months at Progressive). They also rate you primarily on your driving score, de-emphasizing the role of other proven metrics (e.g. credit) that Progressive and others still rely on, in addition to driving behavior. It appears there is no ongoing monitoring of driving after those first two weeks.
How Accurate Is Root’s Pricing?
Good question. I don’t know. I can tell you this much: there are many complaints on customer review sites of Root jacking up pricing on renewal. I don’t consider 10% “jacking”. Often, these people complain of doubling or more (with no claims activity).
This would suggest one of three things to me. First, there is the nefarious explanation that they intentionally underprice to get you in the door then hammer you on renewal. I can’t dismiss this, but it seems unlikely. After all, they surely understand they will lose a lot of customers this way and one of the keys to success in the direct auto channel is retention. It doesn’t make sense to spend a bunch of $ on acquiring customers to lose them after six months.
The second possibility is that as they grow and gain new customers, the incremental data leads to the algorithm improving and, for some drivers, this means large rate increases even if nothing apparently changed in their profile (whereas others get rate declines and they obviously don’t go to message boards to complain).
This seems like a reasonable hypothesis given the lack of data Root has compared to Progressive. Every six months the algorithm probably does improve given the small starting base. However, it’s not good business practice to have pricing be so volatile and they probably need a better strategy to address this.
Finally, the answer may be that, because they don’t have sufficient data, they have underpriced the product. As loss ratios came in worse than expected, Root could have recalibrated and raised prices across the board. This answer makes a lot of sense, but I am suspicious of it given their continued rapid growth, unless they are adding new customers so fast that the high churn isn’t observable yet. As I discuss below, they may be underpriced, but not unwittingly.
It also raises another question of how do they know who needs pricing and who doesn’t? As they have stopped monitoring you and there have been no claims in the first six months, it would seem they would have to take a blanket approach and raise prices across the board rather than just raise them on the worst drivers. Of course, this will lead to adverse selection as the true good customers leave and the true bad customers swallow the rate hike.
Does Root Care About Profitability?
Again, I don’t know. I suspect since they are a “tech” company, they subscribe to the “grow as fast as you can, we’ll worry about profits much later“ philosophy that is so common in traditional tech.
The problem with that is insurance companies have this pesky thing called a financial strength rating. If results are too poor, the rating agencies will downgrade them. Furthermore, growth requires capital. A lot of it. The best way to fund growth is with earnings. However, if the tech mentality has taken hold, then it may be the investors are willing to fund that growth solely with capital injections.
As an aside, Root has found one trick to reduce capital requirements. They set up an agency to run all the expenses through and charges what appears to be a fixed commission back to the insurance company. This reduced the expense ratio from over 100 to the 30s. This allows the losses to be taken outside the insurance balance sheet where they won’t depress capital tests.
Loss ratios are currently running around 100 (ex-reserve development) which means every $ of expense running through the insurance company is depleting capital. Obviously, LRs of 100 (they need to get to ~70 over time) suggest either a) Root has good data but is intentionally underpricing policies 30% to gain share or b) their algorithm isn’t performing well and they are missing their LR target. There has been some adverse development but it is more likely that a) is the greater contributor.
What Can We Learn From Uber & Lyft?
Well, nothing too comforting. Uber and Lyft created a “sustainable” business model that underpriced taxis supported by cheap venture capital that had a high tolerance for losses. Now, these companies are struggling to figure out how they will ever be profitable given the prices they charge and the realization they will never have pricing power because enough people will shift to taxis if Uber and Lyft decide to act as an oligopoly and raise price (in other words, they can never be a true oligopoly because there is such a close substitute).
The question of whether Root is stuck in this trap of never being able to charge a necessary price is one for another day. The question at the moment is how long are they willing to tolerate 100 LRs? The lesson from ridesharing (and many other tech startups) is a really long time.
How Much Money is Root Willing to Lose to “Succeed”?
In this case, success is defined as market share. They certainly can’t scale to several billion of premium at a 100 LR. I mean, maybe they can I guess, if they’re willing to lose $1B/yr on expenses???
There is a financial incentive to improve the combined. Because of the heavy level of losses, Root is running at <1X premium/surplus where a mature company can run at 3X. If they can get the CR to 100 or even maybe 105, they can start taking premium leverage up and reduce the need for more capital injections.
Let’s speculate that Root would be happy if five years from now they are at a 105 CR = 85 LR + 20 ER. If that allows them to grow to $2.5B of premium, that is $2B of market share being taken away from incumbents. With some investment income added in, they would be about break even on net income. There would still be a need to fund growth with capital injections, but I would guess that a breakeven net income would be enough for their investors to do that.
Why Would Auto Insurers Act Differently From Stock Brokers?
Would the major auto carriers really stand by and watch Root underprice them, especially if Root appears on its way to $5B of premium (or even more) a few years after that? Or would they have to start cutting prices to match and move their CR towards 100?
I think the answer to that is pretty obvious. They will do exactly what Schwab and Fidelity did. They will lower their return thresholds to protect their franchise.
So the real unknown here is not whether the incumbents would engage in a price war with Root, but will they have to? That will depend on a) whether Root knows what they are doing on pricing or not enough to continue their strong growth (unclear)and b) are their investors patient enough to endure a long period of losses on the hopes that there will be a long term payoff (likely).
If I were running one of the large incumbents, I would be spending a lot of time trying to answer these two questions. If I were an investor in a public personal lines insurer, I’d want to ask management how they are assessing the threat and how they are preparing for it if they think it’s potentially serious.