This week, Oscar Health is going to IPO. It will attempt to raise about a billion dollars from the public markets after raising over $1.5 billion from non-public sources.
Oscar has had massive losses (from their S-1 p.23)
We have not been profitable since our inception in 2012 and had an accumulated deficit of $1,012.9 million and $1,427.1 million as of December 31, 2019 and 2020, respectively. We incurred net losses of $261.2 million and $406.8 million in the years ended December 31, 2019 and 2020, respectively. We expect to make significant investments to further market, develop, and expand our business, including by continuing to develop our full stack technology platform and member engagement engine, acquiring more members, maintaining existing members and investing in partnerships, collaborations and acquisitions. In addition, we expect to continue to increase our headcount in the coming years. As a public company, we will also incur significant legal, accounting, compliance, and other expenses that we did not incur as a private company. The commissions we offer to brokers could also increase significantly as we compete to attract new members. We will continue to make such investments to grow our business. Despite these investments, we may not succeed in increasing our revenue on the timeline that we expect or in an amount sufficient to lower our net loss and ultimately become profitable. Moreover, if our revenue declines, we may not be able to reduce costs in a timely manner because many of our costs are fixed, at least in the short-term. If we are unable to manage our costs effectively, this may limit our ability to optimize our business model, acquire new members, and grow our revenues. Accordingly, despite our best efforts to do so, we may not achieve or maintain profitability, and we may continue to incur significant losses in the future.
I will try to give Oscar’s argument about why they are a multi-billion dollar company and then I want to work through what I see as my sources of doubt.
Oscar has identified a real problem: Health insurance is not a consumer friendly business.
Oscar’s solution is to throw a ton of technology and front-end user interface development at consumer facing applications to make the consumer experience way nicer and smoother. Oscar then thinks that it can apply machine learning to its ever growing stack of data consisting of both claims data and user reported data to more effectively direct their covered lives to highly effective providers or to direct earlier upstream interventions to save long term money.
The back-end secret sauce relies on a high level of engagement (p.118 of the S-1 ~44% of members are “engaged” monthly.) Over time, the data gets better and the recommendations improve as members are engaged and active and they stick to OSCAR longer which allows for capture of the downstream gains from the back-end secret sauce even as the consumer experience is pleasant and not-painful which is a second source of value.
They are overwhelmingly doing this on the ACA individual health insurance markets.
I think this is a fair description of their pitch.
I have significant doubts about the probability that OSCAR can be a good enough insurance company as it is currently configured to make long term sustained profits. I have several concerns:
- ACA Individual Market is a tough market for the back-end value proposition
- Increased competition
- Ability to execute in 2 parts
- Policy, price shocks and attention
TLDR: OSCAR is making a lot of big bets that highly engaged membership because of an easy to use and user friendly experience can lead to better data that leads to better interventions and a source of private and social value over time. They have spent at least $1.5 billion dollars to learn how to be an insurance company and buying membership. They want to spend a lot more to keep on buying membership and building out their back-end. I have a hard time seeing how this strategy works well on highly competitive ACA individual marketplaces where the marginal buyer is buying almost entirely on cost. If they succeed, that is great, but I still don’t understand what their operations looks like when they are profitable.
More below the fold.
1 — ACA Individual Market is a tough market for the back-end value proposition
The ACA individual market from 2014-2021 is a bifurcated market. Individuals who earn between 100 to 400% Federal Poverty Level receive price linked subsidies that fill the gap between a household’s expected contribution and the local benchmark plan’s gross premium. Many insured individuals will over the course of their contracts never have a claim. Most costs are at the extreme end of the distribution. The marginal new buyer in the market is extremely price sensitive. Many buyers are on the market for a year or two at the most. Some buyers are long term individual market buyers, but the market as a whole is a churning maelstrom of price chasers.
From this perspective, membership chases premium more than it chases other features. Vastly superior networks are not highly valued (~$25 per month willingess to pay) and adversely selected. Customer experience is likely a positively valued attribute but given that most people are barely touching the medical system over the course of the year, it is not going to make up a large pricing gap. Most of the value proposition from the buyer’s point of view is pricing, and more specifically whether or not the plan is either the cheapest silver plan if the buyer earns under 200% FPL or the cheapest plan in each metal tier for everyone else.
OSCAR’s pricing is widely variant. In 2021, they are selling in 128 counties served by Healthcare.gov. In 24 counties they offer the least expensive Silver plan. In another 10 counties, they offer the benchmark silver plan. Slightly less than 20% of their potential market by population on Healthcare.gov will a consumer who is subsidy eligible see OSCAR as either the best or 2nd best option for the most commonly bought plan type (silver). Assuming the House reconciliation bill passes, this will be even more important as both the least expensive and the benchmark silver plans will be effectively zero premium plans. Looking at Bronze plans (the lowest premium, highest cost-sharing plans) OSCAR only has the least expensive plan offering in 16 counties. They have low risk membership attractive pricing in Arizona including Maricopa County and then bits and pieces here and there. In Texas, they are operating at a significant pricing disadvantage to Medicaid managed care like entities such as Molina and Centene.
2 — Increased competition
From 2018 to 2020, the structural conditions of the ACA market were set up to allow insurers to make a whole lot of money. Policy and political uncertainty in 2017 created a massive number of monopolies in 2018. Silverloading led to premiums to be massively overpriced in 2018 and overpriced in 2019. Industry wide Medical Loss Ratios (MLR) were really low. By 2020, Silverloading probably had mostly worked its way through the system after two years of nearly flat industry wide premiums. And then COVID hit which meant a massive decrease in claims. These years are structurally set up for insurers in the ACA individual market to be making serious money through low MLR. OSCAR did not do this. They got their 2020 MLR down to 84.7% from their 2019 MLR of 87.6%. Their MLR is well above industry leaders.
There are two ways to get MLR down; increase premiums or lower claims costs.
And the possibility that they can increase premiums quickly without losing a lot of membership will be constrained by increased competition. Insurers, by 2021, had mostly returned to the marketplaces after running like hell in 2018. More insurers are projected to come back in 2022. Holding premiums constant (average $505 PMPM) and bringing MLR down to 80% only reduces the $1000 Per Member Per Year loss by about $300 PMPY. OSCAR needs to at least increase premium while still compressing claims and then get a handle on their admin costs. Given new entrants to the marketplace, the premium portion is going to be a significant challenge. And this challenge of increasing competition, assuming the Supreme Court does not remove more than 1 or 2 pages of the law in Texas v California, will be persistent.
OSCAR’s pricing is not amazing in most regions to compete solely on price. And these are prices that are generating significant operational losses, so OSCAR is trying to buy membership and scale right now so that they have more leverage against hospitals to get lower per unit costs (some areas they get good for commercial plan pricing, and other areas, they are paying for the privilege of having almost no utilizing membership in a region where the local hospitals can laugh at the concept of pricing concessions). If OSCAR has to price at a level where they break even, they have an even harder time getting to scale.
3 — Ability to execute in 2 parts
Strategy and Covered Population
I have repeatedly said that I have a hard time seeing what OSCAR does besides get good press, deploy a nice app and lose a lot of money that my former co-workers do not do. This specifically applies to their ability to run the basics of pricing risk and running the back-end of an insurance company in a regulated market.
OSCAR currently insures a population that on net is much healthier than the general ACA individual market. OSCAR is currently insuring a population is coded to be very healthy relative to the rest of the ACA universe. We know this because OSCAR is a significant net payer into risk adjustment. In 2020, each member, on average had about a $1700 transfer into risk adjustment (some markets OSCAR was a risk adjustment receiver) which is notably larger than the 2019 net outflow of just under $1,200 per member per year. About a quarter of premium goes out the door for risk adjustment.
THERE IS NOTHING WRONG WITH A STRATEGY THAT INVOLVES PAYING INTO RISK ADJUSTMENT!
Centene and other Medicaid-esque insurers routinely pay a significant portion of premium in net risk adjustment payables. Other insurers are profitable with significant risk adjustment receivables. Both of those courses are fine as long as pricing and strategy are aligned with the population being insured.
I don’t think there is alignment.
OSCAR’S membership is currently heavy on 18 to 34 year olds and light on 55 to 64 year olds (P. 117). Premiums are heavily concentrated in 55-64 year olds.
Profitable insurers with large risk adjustment payables are mainly competing on price because their membership is barely using services.
More importantly, OSCAR’S adjustment profile is saying that their covered population should be healthy and use few services but their MLR is telling me that they either have horrendous per unit costs or a lot of utilization. I am curious if their highly engaged members and OSCAR’s no cost-sharing virtual platforms generate a lot of low dollar, low acuity claims that are not generating risk-adjustable diagnoses for things that if there was either cost sharing or an expectation of in-person visits with the corresponding transaction and hassle costs would never have generated a visit. I know that if my son is feeling mostly fine but a little warm and had green boogers coming out of his nose and he wants to go to bed by 7:00, the diagnosis is likely a viral infection with a treatment course of sleep, hydration and a day of cartoons. If there are either cash or transaction costs to go to a doctor, we’re not going to the doctor unless it does not clear up in a week. If there is a no cost to the visit and we can do it in between conference calls, sure, I’ll get him checked out.
OSCAR has demonstrated that they can sell an experience at a loss. The population that is mostly buying this experience is young and comparatively healthy. They have not demonstrated that they can price their products competitively at a break-even point. There is little space left at current pricing to drive down MLR. Their 2020 membership gains were from a population that was even healthier relative to the rest of the market than their 2019 population which was still an extremely healthy population relative to the rest of the market.
OSCAR contends that its analytical engine’s special sauce driven by integrating both claims (which every insurer can do, and many can do so from far deeper and broader data sets) and member reported data which is somewhat unique to OSCAR will lead to better choices for lower cost bundles of care and better health. This sounds plausible if there are large populations with robust medical needs that may be amenable to management and intervention. But given OSCAR is insuring a population that is heavy on young people and light on old people and has a risk score that is well below market averages, where is the squeeze? OSCAR’s special sauce to better manage diabetes could be fantastic but OSCAR is unlikely to be covering a lot of people with Type 2 diabetes. Where is the squeeze with the current covered population?
OSCAR is bragging that they are rapidly growing. They are buying membership with large losses. Some of their markets have decent scale within the context of the ACA (California, Texas, Florida). Some of their markets are tiny with fewer covered lives than my high school had students. They have half a million total covered lives in 18 states. Their ability to leverage their membership to get good pricing locally is limited. Their administrative costs have been high and are not shrinking on either an absolute or percentage of premium basis. When I worked at UPMC Health Plan, the Medicaid line of business for a single state (Pennsylvania) had 400,000 covered lives and was profitable. We had scale in parts of the state and were an after-thought in other portions of the state. We had scale challenges in a single state with the same ballpark of covered lives as OSCAR has in a good chunk of the national footprint.
4 — POLICY
As I mentioned earlier, OSCAR has been operating in an environment from 2018-2020 that is conducive to pricing high and low MLRs. That is changing.
More importantly, operating with the assumption that the Biden Administration is able to pass the COVID-related American Rescue Plan with significantly enhanced subsidies for more people, large chunks of the current covered population will be receiving significant positive price shocks. Inertia has been the friend of any insurer that can gain membership by pricing low in their first year and then raise premiums modestly over time. Price shocks, even positive price shocks, will make people who are currently not paying attention and automatically renewing their policies in plans that are no longer the best priced options, pay attention. OSCAR’s pricing in 2021 for Healthcare.gov is the best for the under 150% FPL group that will see $0 premium Silver CSR plans. This will be helpful to OSCAR in 2022. However, its pricing in other markets will not be as good as other on-exchange carriers which people will pay attention to and plausibly switch to in 2022.