The Goods You Can't Check
Your car makes a noise. You take it to the shop. The mechanic says it needs a new part. You pay. The noise stops.
Did you need the part? You don't know. You will never know.
Your doctor says you need a test. Then a follow-up. Then a prescription. You feel better. Or you don't.
Did you need all of it? Did you get all of it? Was the price fair? You don't know that either.
Economists have a name for this. Most people have never heard it. They're called credence goods.
There are three kinds of things we buy.
Search goods you can judge before you buy. A shirt. A couch. You look, you touch, you decide.
Experience goods you can judge after. A restaurant meal. A haircut. You find out when you're done.
Credence goods you can't judge even after. The term goes back to two economists, Michael Darby and Edi Karni, in 1973. The classic examples are the mechanic and the doctor. The person who tells you what you need is the same person who sells it to you. You're left with one tool. Trust.
That's not an insult to doctors or mechanics. Most are honest. But the setup invites trouble. Economists list three ways it goes wrong. You get more than you needed. You get less than you needed. Or you get charged for more than you got.
Researchers have tested this. In one study, people took identical broken phones to repair shops in Turkey. When they said "I don't know what's wrong," they paid more than 40% more than when they named the broken part. Same phone. Same fix. The only difference was what the customer knew.
Now add insurance.
You'd think a third party would help. Someone big, checking the bill.
It doesn't work that way. In a study published in PNAS, researchers took computers with the same small, simple problem to 61 repair shops in Austria. The average bill was about 70 euros. When the customer mentioned that insurance would cover it, the bill went up more than 80%. Extra parts. Extra hours. When asked why, the shops' top answer was that insured customers don't watch the cost, because somebody else pays.
That was computers. But the economists who ran it put medical care right next to repair work as a prime example of a credence good.
Here's the problem. The insurance company can't verify the care either. It wasn't in the room. So it does what it can. It adds forms. Codes. Prior authorizations. Denials. Paperwork in place of trust.
The patient still can't check. The doctor spends the afternoon proving things to a stranger. And the price climbs to pay for all the checking.
I am not anti-insurance. Keep a policy for the wreck. But for everyday care, insurance didn't close the trust gap. It added a third party who can't see inside it either.
So what do you do when the customer can't check your work?
You make it so they don't have to.
You go overboard on the things they care about most. You take away the reasons to cheat before anyone has to wonder.
At AtlasMD, that looks like this. Unlimited visits, so there's no reason to book you more or squeeze you in less. No copays. All procedures free. Meds and labs at cost plus 10%, with the price posted where you can see it.
Unlimited visits take away the pull to overtreat or undertreat. A posted cost-plus price takes away the room to overcharge. You don't have to trust my diagnosis of your bill. You can read it.
None of that is easy as a business. Doing less for more is easier. It's also not fair.
A credence good asks the customer for faith. The least we can do is earn it on the parts they can see.
You shouldn't need faith to read a bill. So we post ours.
Sources:
Kerschbamer, Neururer & Sutter, PNAS (2016): www.pnas.org/doi/10.1073/pnas.1518015113
Phone repair field experiment, Management Science: pubsonline.informs.org/doi/10.1287/mnsc.2022.02666
Investopedia, Credence Good: www.investopedia.com/terms/c/credence-good.asp