Health care costs are a bit like the weather: Everyone talks about them, but no one ever does anything about it. They differ, however, in this regard: People want to do something about health care costs. And yet, those costs have long outpaced inflation and are projected to reach one-fifth of our gross domestic product by 2025.

Companies, which provide much of the health insurance in this country, are understandably nervous about this state of affairs. They would like to do something to stop it. And three very large companies, with a lot of market power, have just announced that darn it, they’re gonna.

A partnership of Amazon.com Inc., Berkshire Hathaway Inc. and JPMorgan Chase & Co. is forming an independent company, “free from profit-making incentives and constraints” to “provide U.S. employees and their families with simplified, high-quality and transparent healthcare at a reasonable cost.” They say they’ll be “tackling the enormous challenges of healthcare and harnessing its full benefits.”

While the initial focus will be on technology, and the efforts will initially be aimed only at employees of the three firms, one suspects that the ambitions are slightly bigger: building a business that can somehow tamp down the pressures that drive health care costs ever upward. In remaking the market for health care services, they might even divert some small fraction of that gross national spending into their own pockets.

And if they can, more power to them; paying Amazon et al. an outlandish profit would be well worth it if they can actually dam up the river of money that flows into the health care system every year. The question is, can they? Or will they, like many others before them, start building what they think is a mighty levee, only to see it collapse under the pressure as the flood waters roll on?

It’s not entirely crazy to think that they might.

Start with the little-known fact that most large companies self-insure. They generally pay outside firms to administer their health insurance for them, but they are financially responsible for the claims. A large employer is a little statistical universe, with unusually healthy employees balancing out the bills for the unusually sick ones. So while those companies are not experts in managing health insurance in the way that, say, Aetna is, they do have some experience with it.

Add in the fact that Amazon is really, really good at technology — and most health care companies aren’t. Much medical technology is wondrous, to be sure — but the systems that tie all that technology together are, by the standards of any other industry, a hot mess. There are a number of reasons for this, from privacy laws to provider fragmentation, but it’s hard to escape the conclusion that part of the reason health care IT is so bad is that it simply doesn’t have to be very good. Large parts of health care are sheltered from normal competition, because people don’t shop around for doctors and hospitals, and the companies paying the bills don’t have a great deal of control over the system.

The health care companies that do IT well don’t necessarily get rewarded for it, because the payment systems aren’t set up to deliver those rewards. So the normal incentives that drive companies to use information technology to make themselves more efficient simply aren’t as strong in health care as they are in other industries — at least on the provider side.

But as customers with a combined employment base of over a million people, Amazon and JPMorgan and Berkshire Hathaway may have the incentives, and the expertise, to do it right.

That said, there are other companies in the industry, with an incentive to get technology right, and so far, few of them have managed to overcome all of the obstacles that the system puts in their way. The dysfunctional incentives of third-party payer, where the people making the decisions seldom have any reason to reward efficiency … the incredible fragmentation of the market, which makes it hard to come up with big, unified solutions … the fierce resistance of providers to adopting new ways of doing things … and if you somehow manage to surmount all of those obstacles, and actually start rationalizing things, the tendency of legislators and regulators to come steaming in with some new law or regulation that renders your idea illegal.

Most important, you are dealing with human beings at their most stubborn and vulnerable. Your regime of evidence-based medicine will founder on the fact that human bodies are not very well standardized, and go wrong in all sorts of perplexing ways that will resist any attempt to neatly categorize them. Your behavioral modifications will run up against the fact that human behavior is awfully hard to change. And your attempts to beat down costs will run aground when you discover that many market participants enjoy being the only game in town — like rural hospitals and pharmaceutical manufacturers — and that you cannot avoid dealing with them unless you want some combination of legal trouble or employee revolt.

So while there are some reasons to think this company might succeed, there are also plenty of reasons to think that it will fail. The one thing we can say, however, is that if it succeeds, its success may help usher in an era of even tighter employer control over employees’ lives.

Right now, even when our employer is functionally purchasing health care services for us, that transaction is arm’s-length: We decide on the services, and an intermediary actually pays the bill.

There are probably considerable savings to be had if employers use their power to guide employees toward better decisions about everything from ER use to smoking.

But one big reason that our health care system is such an expensive mess is that Americans hate being told what to do. They demand maximal, expensive, freedom of choice about their health care. They rebel if they can’t get it. Worse still, if they are denied it, they call their legislators, who do things like telling insurers to stop denying so many claims for experimental treatments of dubious worth.

Maybe Amazon — with big data and smart algorithms like the one that recently enticed me to buy Russian cake piping tips, a product I had previously had no interest in or even awareness of — can get us to start acting like more responsible health care consumers. But then again, maybe consumers and providers will continue to be swayed by the perverse incentives that gave us the system we have today.

Megan McArdle is a Bloomberg View columnist (mmcardle3@bloomberg.net). She is the author of “The Up Side of Down: Why Failing Well Is the Key to Success.”

First Published: January 31, 2018, 5:00 a.m.