Why Hospitals Aren't in the Drug Business

(Bloomberg View) -- Fed up with high drug prices? Why not make your own?

Well, okay, for most of us that’s not very practical. But hospitals have decided that’s just what they’re going to do: get into the business of manufacturing generic drugs.

“This is a shot across the bow of the bad guys,” Marc Harrison told the New York Times. Harrison is the chief executive of Intermountain Healthcare, the nonprofit hospital group that’s leading this effort. “We are not going to lay down. We are going to go ahead and try and fix it.”

Business school professors are probably watching with a more jaundiced eye. Getting into the supply business is not exactly a revolutionary idea; it’s a well-known strategy called “vertical integration” that has been tried -- and often rued -- by many a company. While vertical integration can certainly work, it can also turn out to be a great way to lose a lot of money and screw up your core business. (Have you seen “Weeds”?)

Companies tend to pursue vertical integration for the same reasons that these hospitals are: because they are fed up with high prices and short supplies of some key input. They figure that if they got into the business, they’d produce as much as was needed, and eliminate all the grotesque profits their suppliers are enjoying -- or better yet, capture those profits for themselves. Rarely do they seem to stop to ask the reasons for the high prices and the scarcity, or to wonder whether those reasons will magically go away once they own their suppliers (or their buyers, a kind of vertical integration known as “forward integration”).

So let’s think through this for them. Why are supplies constrained, and prices high, in the generic drug business? Remember that we’re not talking about the much-maligned major drug manufacturers, who have valuable patents preventing others from manufacturing their drugs. The hospitals are talking about making generics, the basic instructions for which are in the public domain.

In an economics-textbook universe, we’d expect to see competitors looking at the high profits that companies are able to charge, and eagerly pile into the market, driving down prices until they are once again at competitive levels. Why doesn’t this happen?

Well, often it does. One thing that is underappreciated about the evil-genius strategy pursued by folks like Martin Shkreli is that it’s a short-term strategy. You may be able to corner the market for Epi-Pens or diabetes meds or anti-malarials for a little while and charge astronomical fees for your product. But if you’re selling something at a 5,000 percent markup, competitors are going to hustle in and eventually, your margins will fall back to normal.

But how do they even manage to get that kind of markup for a little while? Alex Tabarrok, an economist at GMU, explained it best at a panel in 2016: They’re exploiting the barrier to entry created by FDA approvals.

Essentially, these price increases tend to happen in generics market with some very specific characteristics: A product has only one manufacturer (often because the market is so small it can only really support one manufacturer at competitive prices); the drug is necessary to treat some condition; and the drug has no good substitutes.

These markets are vulnerable to evil geniuses who grasp their market power and decide to jack up the price to kingdom come, knowing that competitors will be afraid to enter. And why would anyone fear to enter a market that currently offers a 5,000 percent markup? Because they know that they’ll never get that markup. If they enter and offer, say, a 4,500 percent markup to steal the former monopolist’s customers, that provider will cut prices to a 4,000 percent markup, and the newcomer will have to go down to a 3,000 percent markup to get the customers back, and pretty soon the drug will be trading at a competitive price.

In the meantime, you’ll have expenses. You’ll have to figure out how to make the drug (harder than it sounds, even though the basic information is available from the patent filing). You’ll have to set up a factory to produce it. And you’ll have to get the FDA to certify that the drug you’re making is reasonably pure and safe for consumption by American patients. This doesn’t cost as much as getting a new drug approved, but it’s not free, and it takes quite a bit of time. And if the market is small, it’s not necessarily clear that you’ll ever recoup those expenses, because splitting a small market with another manufacturer may mean that you both lose money.

That’s not the only reason that drug prices spike; there can also be manufacturing problems that temporarily shut down one or more manufacturers in a multi-firm market, causing supply to crater and prices to soar. Or shortages of key inputs to the manufacturing process. Or the FDA can pull a competitive product off of the market, handing new market power to a longtime manufacturer (this is one of the things that happened with the Epi-Pen).

It’s understandable if hospitals and patients don’t really care why the prices are going up. But it does matter, because the reasons for the price increase tell us a lot about how successful this hospital-led effort is apt to be.

Take Epi-Pens. Given that every other kid these days seems to have some sort of terrible allergy, that’s a pretty large, robust market; I wouldn’t expect Mylan’s pricing power to last for all that long. But by that token, it’s unlikely that a hospital-owned manufacturer is going to do much good. There’s no reason to think that they’ll be any faster at creating their own Epi-Pen competitor than some company that has long experience in pharmaceutical manufacture and medical device development -- and indeed, there are a lot of reasons to think that the hospital-owned maker would be slower. Manufacturing is hard, and doing it well takes a specific set of skills, and a managerial focus, that hospitals don’t have.

Or consider IV fluids, which have also been in short supply because of Hurricane Maria, which devastated Puerto Rico’s large pharmaceutical and medical device industry. This new manufacturer is unlikely to be much help with these kinds of problems; it will not itself be somehow magically immune from acts of God that destroy its facilities, and it will not be able to tool up production to respond to supply shocks any faster than a normal manufacturer.

On the other hand, look at Daraprim, the drug that made Shkreli famous. In 2015, Turing Pharmaceuticals hiked the price from $13.50 a pill to $750 a pill; outside of the U.S., it often costs as little as one or two dollars. And last time I looked, it was still absurdly expensive. That’s the sort of abuse that a hospital-owned company might actually be able to fix. Not quickly, but eventually. And the existence of a company that can step in and take away all of your ungodly profits may deter firms from adopting this strategy in the first place.

But even so, they’ll also need to think hard about the costs of vertical integration. One obvious cost is … the costs. Developing a manufacturing competency will be expensive, in terms of both management expertise and capital expenditures. They will have to devote some of their own precious management time to the project, and to the inevitable arguments over exactly which drugs their new firm should make. There’s a good chance that they will turn out to be worse at manufacturing drugs than the existing manufacturers (just look at the sad fate of the insurance co-ops created by the Affordable Care Act), and will end up as a drain on hospital budgets rather than a gain.

But another cost is what vertical integration does to corporate decision-making. Once you have an expensive facility set up to do something, you start making bad downstream decisions. An airline buys an oil refinery to hedge against rising oil costs -- then decides to use the oil to fly planes, even though the company would make more money selling that oil on the open market.  A clothing manufacturer buys a textile mill to gain a pricing advantage in the production of fashionable double-knit clothing -- then keeps making double-knits long after they’ve gone out of style, because what else are you going to do with that mill? A maker of small consumer electronics sets up a chip manufacturer to control their biggest input cost -- then discovers that the minimum profitable scale of chip manufacturer is much larger than that of their products, and they either have to find some way to offload the extra chips, or sell a bunch of alarm clocks at a loss.

In the health care context, that’s more likely to look like “Well, there’s this new drug that does a better job, but we’ve got a huge investment in making this generic, and it would cost a bunch more money to do the approvals and the retooling to make something else, plus, we get this one cheap. So let’s stick with the drug we make in-house.” Or too-frequent retoolings to satisfy hospital needs, keeping the drug manufacturer running at a loss and draining hospital budgets.

That’s not to say that vertical integration never works. But it should always be approached with great caution. Beware of its side effects.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Megan McArdle is a Bloomberg View columnist. She wrote for the Daily Beast, Newsweek, the Atlantic and the Economist and founded the blog Asymmetrical Information. She is the author of “The Up Side of Down: Why Failing Well Is the Key to Success.”

  1. Yes, I understand that the hospitals are not literally going to be trying to make this stuff in the cafeteria; they will hire experts to do it. But there’s a reason that so many of the conglomerates of the 1950s and 1960s ended up breaking up; it turns out that a small, nimble company with a laser-like focus on a few core-competencies can often outperform a company that owns a bunch of very different businesses, because management never gets very good at any one thing.

For more columns from Bloomberg View, visit http://www.bloomberg.com/view.

©2018 Bloomberg L.P.

Stay Updated With Business News News On BloombergQuint