Healthcare Isn't One Market. It's Three.
I think we've been arguing about healthcare incorrectly for decades.
The debate almost always ends up in the same place.
"The free market will solve it."
"No, healthcare should be run by the government."
For a long time, I assumed those were the two sides of the discussion.
I don't anymore.
I think the entire debate begins with a faulty assumption:
Healthcare isn't one market.
It's at least three completely different markets masquerading as one.
If that's true, then expecting a single financing model to solve all three may be the biggest architectural mistake we've made.
Engineers Start With the Problem
As a software engineer, I rarely begin by choosing a technology.
I begin by asking:
What problem am I actually trying to solve?
Only after understanding the problem do I decide what architecture makes sense.
Healthcare deserves the same treatment.
We shouldn't begin by asking whether markets or government are better.
We should begin by asking whether we're even talking about the same problem.
Because I don't think we are.
Problem #1: Routine Care
Consider services such as:
- Annual physicals
- Basic blood work
- Treatment for a sore throat
- Dental cleanings and cavities
- Routine prescription refills
These services are generally predictable. Most are not emergencies, and patients can often compare providers, prices, and convenience.
That means routine care has many of the characteristics required for a functioning market.
But we rarely allow it to operate like one.
Imagine if your auto insurance paid for gasoline, oil changes, tire rotations, windshield-wiper fluid, and every car wash.
You would likely stop paying attention to what any of those services cost. A mechanic might charge $50 for an oil change or $500, and as long as the insurer covered it, the price would have little influence on your decision.
The mechanic, in turn, would have less incentive to reduce costs. Why invest in better equipment, streamline workflows, reduce labor time, or develop faster ways to perform an oil change if the person receiving the service neither sees nor directly pays the bill?
Over time, the system would almost certainly become more expensive and less efficient.
Additional administrative layers would emerge to determine which services are covered, which billing codes apply, and how costs are allocated. Customers would choose providers based on insurance networks rather than price or quality. Providers would optimize their behavior around reimbursement rules instead of the underlying service itself.
This kind of waste would not require bad intentions. It would arise naturally once the feedback loop between price, choice, and value had been weakened.
In ordinary markets, businesses are disciplined by a simple mechanism:
A customer sees the price, compares alternatives, and decides whether the service is worth it.
That decision rewards providers who deliver better value and penalizes those who do not.
When customers do not see the price, do not pay it directly, and cannot easily compare options, that discipline largely disappears.
This is, in many ways, what has happened with routine healthcare.
Insurance was originally designed to protect people from financial catastrophe. When it is used to pay for predictable, everyday services, it becomes an extremely complex payment intermediary—and it weakens the very forces that would otherwise drive lower prices, greater efficiency, and innovation.
Problem #2: Catastrophic Events
Now consider a completely different situation.
You receive a cancer diagnosis.
You're involved in a serious car accident.
You suffer a heart attack.
Nobody plans for these events.
Nobody shops around while unconscious in an ambulance.
And the financial consequences can be devastating.
This is exactly the kind of problem insurance was invented to solve.
Insurance is incredibly effective when protecting people from low-probability, high-cost events that would otherwise bankrupt a family.
Pooling those risks across millions of people creates enormous value.
Trying to pay for these events out of pocket makes little sense.
Unlike routine care, catastrophic events are not problems that markets alone handle particularly well.
Problem #3: Long-Term Catastrophic Disease
Diseases such as Multiple Sclerosis, ALS, Parkinson's disease, and certain lifelong genetic disorders aren't routine, but they also aren't one-time catastrophes. They're decades-long catastrophes.
Insurance works best when uncertainty exists. Once someone develops one of these diseases, the uncertainty largely disappears. The costs become known, ongoing, and potentially enormous.
This is a third engineering problem that deserves its own architecture.
Different Problems Deserve Different Solutions
Instead of asking:
Should healthcare be public or private?
I started asking:
Which of the three problems are we trying to solve?
Routine care benefits from transparency, competition, and price feedback.
Catastrophic events naturally lend themselves to insurance.
Long-term catastrophic disease may require an entirely different financing model because its incentives are fundamentally different.
Start With the Architecture
Before we debate markets versus government...
Before we argue about insurance versus single payer...
Before we redesign the entire healthcare system...
We should probably answer one simpler question first.
What problem are we actually trying to solve?
Because until we agree on that, we'll keep building elegant solutions to the wrong problem.