The Political Economy of Shitty Generic Drugs

FDA approval and the dangerous valley of mediocre regulation

I finished Katherine Eban’s book Bottle of Lies (Amazon) a thorough and narrative-driven, if perhaps lopsided exposé of the corporate greed and regulatory failings fueling and enabling a generic drug industry that endangers and shortchanges patients across the globe.

Amazon.com: Bottle of Lies: The Inside Story of the Generic Drug ...

The book covers a lot, but one central point is that the generic versions of prescription drugs are supposed to be virtually identical to the name-brand version, save perhaps for a few inert additives. In other words, they must contain almost the same amount of active ingredient, and release that ingredient into the body at almost the same rate as the original patented version.

Unfortunately for patients, however, this is not always the case. Eban documents some pretty shocking cases of patients with life-threatening diseases who came close to dying after switching to a generic version of their medication, but obviously smaller and less lethal differences are far more common. A few years back, patient complaints over a newly-released generic version of the antidepressant Wellbutrin (bupropion) led the FDA to take the rare step of running its own study, and found that the generic it initially approved wasn’t “bioequivalent” after all. Want to know about whatever it is you’re taking? Search around for “generic vs brand ____”, especially on Reddit (perhaps the epicenter of condition-specific forums).

But one thing the author doesn’t really tackle is why the FDA is the only thing standing between patients and an onslaught of inferior drugs. I’m pretty sure there are few parts of the economy more rigorously regulated than the prescription drug industry and - setting aside externalities, monopolies, and such - most industries are subject to the market force of consumer choice.

Just take the darknet markets for (mostly) illegal recreational drugs. I’m no free-market absolutist, but I’ll admit that they seem to be a case study in the self-regulating capacity of consumer freedom. According to this article:

Drug dealers care about their online brand and reputation and customer satisfaction as much as Airbnb hosts or eBay sellers. A typical vendor’s page will be littered with information, including: how many transactions they have completed; when the vendor registered; when the vendor last logged in; and their all-important pseudonym. It will also feature a short description about why a user should buy drugs from them, refund policy information, postage options and “stealth” methods (measures used to conceal drugs in the post). Even if you’re not in the market for what they’re selling, it’s hard not to be impressed; vendors put in real effort to demonstrate their trustworthiness.

For the sake of this post, I’m going to take the book’s word that U.S. brand name pharmaceutical manufacturing is basically safe and competent while U.S. generic manufacturing regularly fails to follow industry-standard manufacturing practices. Both are subject to the same laws and FDA regulations. If the FDA’s regulations and enforcement were sufficient, then it wouldn’t be possible for generic manufacturers to get away with sketchy manufacturing. But if neither consumer power nor FDA oversight are sufficient to ensure safety, then why are brand-name drugs generally exempt from these problems?

I’m well aware that the entire U.S. healthcare system is a byzantine labyrinth of distortions, rules, regulations, and market failures. There are probably dozens or hundreds of factors I’m not addressing or considering that are contributing factors.

That said, I think part of the answer is what I’ll call the dangerous valley of mediocre regulation. That is, a regulatory (or certification, without the force of law) body that is competent enough to create some amount of trust in the minds of consumers but not competent enough to sufficiently enforce their regulations can partially substitute for brand reputation from old-school consumer experience.


Imagine you’ve just been prescribed 40 mg/day of atorvistatin, America’s most-prescribed medication and one of the medications highlighted in Bottle of Lies as having had inferior generic manufacturers, but your doctor will order the brand-name drug Lipitor if you ask her to. You’re one of 8% of Americans without health insurance, so you’re paying out of pocket. According to GoodRX, you can take the generic for about $10/month or Lipitor for $470/month. I have little idea how much insurance plans tend to cover, but presumably most Americans without gold-tier coverage would be paying quite a bit more for the brand as well.

Seems like a no brainer, right? Even after reading specifically about the shady manufacturing squirreled away from easy FDA oversight and a rare but telling $500 million penalty levied on a generics company in 2012, I still trust the FDA enough that I’d take the generic if my out of pocket costs for Lipitor were anything like the retail price.

An Oversimplified Model

Here’s an oversimplified model of what I think is going on:

Any company making made-up drug “dioprexamine” can choose one of three levels of diligence and safety: “Completely Reckless,” “Cutting Corners,” and “Industry Standard.” To be allowed for sale, the company must get cleared by the FDA and is subject to plant inspections. “Completely Reckless” manufacturers can make a year’s supply of pills for $100, “Cutting Corners” ones for $200 and “Industry Standard” guys for $300, and a year’s worth of pills from each will make patients sick 5%, 1% and 0% of the time respectively.

Say patients value not getting sick at $10,000, and value a year’s supply of the drug at $350 (and for the sake of the model, all three levels make drugs of equal quality besides the illness risk). If patients could know which manufacturer they were buying from, here’s how they’d value a year of dioprexamine from each level of manufacturer:

  1. Completely Reckless: $350-$10,000*0.05 = -$150

  2. Cutting Corners: $350-$10,000*0.01 = $250

  3. Industry Standard: $350-$10,000*0 = $350

So we can easily imagine a two-tiered market where Cutting Corners firms sell for about $225 and Industry Standard firms for $325 (depending on elasticity and oligopoly dynamics and all that), and everyone gets around a $25 surplus.

But that’s not the world we live in. Instead, the only information consumers have is:

  1. That a single particular seller makes the “brand name” version, which is made with Industry Standard practices.

  2. That every seller has cleared FDA approval.

In this world, the FDA demands Industry Standard practices, but those of lesser quality can sometimes sneak through the approval process. Suppose - from lots of experience and history with other medications - consumers know that FDA approval of a firm means there is a:

  1. 40% chance it is Industry Standard

  2. 50% chance it is Cutting Corners

  3. 10% chance it is Completely Reckless

A humble patient deciding whether to buy dioprexamine from an unknown but FDA-approved firm values the product at:

-$150*0.1 + $250*0.5 + $350*0.4 = $250

This means that until information disseminates about the particularities of the dioprexamine market, it makes no sense for any Industry Standard firm besides the brand-name manufacturer to enter the market. Yes, eventually consumers will learn that the 40% chance of their drug being of the highest quality is actually 0% and adjust accordingly, but this could take a very long time - especially because in real life, detection of adverse effects or substandard efficacy as causally attributable to cost-cutting manufacturing practices requires a cocktail of patient complaints and hunches, laboratory tests, regulatory infraction findings, scientific studies, and investigative journalism.

What about the other two types of firms? Will it make any sense for a manufacturer to level-up from Reckless to merely Cutting Corners? That depends on how expensive it is to get shut down by the FDA and how much better the FDA is at detecting the former type than the latter. But given what I’ve read, FDA detection of substandard manufacturing doesn’t mean that your $10 billion upfront investment in a plant goes to waste; it means that you have to sink a million dollars or something into upgrades, at least until the FDA forgets about you and moves on.

Even the absolute worst, most egregious offenders who falsified records, faked lab tests, and systematically deceived regulators across the world like Ranbaxy (linked above), get hit with a $500 million fine, and no one goes to jail! For context, their parent company Sun Pharmaceuticals makes that much in revenue in about 6 weeks. And I strongly doubt that many Americans started hassling their pharmacy to avoid using Sun or Ranbaxy generics either, so odds are the generic dioprexamaine market has at least a few Completely Reckless firms.


Mediocre Regulation and Goodhart’s Law

Back to the point: in a transparent, unregulated market, consumers would never have to worry that their medications were made by the most reckless type manufacturer. The FDA, however, creates a regulatory cover which any firm clever enough to trick their way into stands to profit. I think it is a sort of modified case of Goodhart’s Law, which states that

When a measure becomes a target, it ceases to be a good measure.

The key insight is that things that can be measured and optimized for are seldom the thing that whoever implemented the target cares about. Americans don’t care about FDA approval intrinsically (I don’t, anyway); they care about drug safety and efficacy, which FDA approval is supposed to track. But there is always some daylight between something valued (like safety) and its proxy measure (like regulatory approval).

Firms optimize for the latter instead of the former, sometimes to the point of breaking the law. The thing about Goodhart’s Law is that it isn’t literally true. If a measure is close enough to the thing it is supposed to track, then the measure might actually work pretty well.

All this is to say that the amount of daylight between actual drug safety and the Goodhart’s Law-prone measure/target of FDA approval depends on the competence of the FDA. The pharmaceutical industry might be powerful and amoral, but there does exist some finite degree of oversight and enforcement that will lead it to make safe drugs even with these newly-created market distortions.

Why This Matters

I am absolutely not implying we should abolish FDA oversight. There are many plausible reasons why the government should demand safety practices above and beyond what the market would bear. For instance, it might take many years and many patients harmed for a particular sketchy manufacturer to develop an appropriate reputation, and perhaps this isn’t worth cheaper drugs.

What I am implying is that lax, incompetent, or inconsistent regulation can be just as distortionary and harmful as excessive regulation. The problems I am outlining here exist only because the FDA is incapable of enforcing what it claims to but is still good enough to catch some proportion of offenders and garner some degree of consumer trust.

I have no idea whether the FDA is more or less competent than other U.S. or international regulatory agencies; no agency can do a perfect job, and policymakers have to strike a balance between making a regulatory regime expensive/high standards/(potentially) innovation-inhibiting/high false-positive/credible and cheaper/low standards/less distortionary/high false-negative/untrustworthy.

The lesson that should be learned is that the relationship between some valued outcome (cleaner air, fewer traffic accidents, safe drugs, whatever) and degree of regulatory stringency and competence may not be monotonic. It is completely possible for the worst outcome to arise from a mediocre regulatory regime.

Not Just Regulation

As a final point, if this explanation is correct then it applies to voluntary review or certification systems as well, like Consumer Reports or a movie critic. With no professional movie critic, people ask friends for recommendations and make informed guesses about what to see. With a really good critic, they can do better than that by accessing even more useful information, even if this critic prevents people from collecting recommendations from friends. With a mediocre critic (who is not yet revealed to be mediocre), it is quite plausible that his existence leads viewers not to seek information from friends and family such that their viewing choices are ultimately worse than both prior scenarios.

Importantly, the mechanism here is not that the reviewer is providing bad information; it very well may be that his reviews are better than no information at all, so that the ordering of information quality goes:

No Information < Mediocre Critic < No Critic < Good Critic.

Rather, he crowds out the existing information source (friend recommendations) thanks to his perception as a competent authority without sufficiently replacing that information with his reviews. Analogously, a mediocre FDA might be worse than ‘no FDA but patient internet forums’ but better than ‘No FDA and atomized, ignorant consumers.’


Actual American Politics

Not all regulatory bodies are subject to exactly the same dynamics described in this post, but I suspect that even those not involved with consumer products are also worse for the country when mediocre. A weak-but-existent EPA, for example, might generate a false sense of satisfaction in the American electorate such that meaningful action never takes place.

I haven’t done much research to confirm whether this is true, but my sense is that the Republican Party has, in recent years, worked to weaken Federal regulatory bodies and state capacity in general without having the capacity to eliminate them entirely. For instance, Consumer Financial Protection Bureau actions are way down in recent years. This is clearly suboptimal from both liberal and libertarian perspectives. Hopefully, recognition of the “dangerous valley of mediocre regulation” will discourage the trend going forward.