Hayek at the Hospital: The Use of Knowledge in Hospital Discharge Decisions

A live update from The New York Times on the COVID-19 situation in Peru:

Peru has more than 170,000 confirmed cases, despite taking the virus seriously early on. The president, Martín Vizcarra, ordered one of the first national lockdowns in South America. Though the official virus death toll stands at around 5,000, Peru had 14,000 more deaths than usual in May, suggesting that a growing number of people are dying at home as hospitals struggle to handle a flood of cases.

So should the Peruvians be taking to the streets to express outrage at the incompetence of all those heads of households where excess mortality has taken place? Or alternatively, should Peru’s hospitals just have warehoused up to 14,000 post-acute COVID patients in the excess space they had as they were struggling “to handle a flood of cases”?

We can only be thankful that the Peruvian government didn’t direct Peru’s nursing homes to accept expedited discharge of COVID positive patients from its hospitals. Because clearly, if only the Peruvians had lacked a credible answer as to where they would go, all those patients would have survived. It’s only when you have a credible answer as to the patient’s destination upon discharge that the patient suffers premature mortality.

“Hayekian” moral of the story: the key to keeping excess mortality low during a pandemic is not to have any determinate plan for discharge from hospitals for post-acute cases, as hospitals are managing an excess influx of acute cases. The minute some central planning authority insists on a discharge plan, people start to die. Best to rely on the dispersed, decentralized knowledge found in the market. As Hayek taught us, clinical outcomes are optimally brought about when there is no centralized coordination between ambulance squads, hospitals, and long-term care facilities. Surely the worst thing you could do is to bring all that knowledge together, and come up with a plan to deal with it in real time.

8 thoughts on “Hayek at the Hospital: The Use of Knowledge in Hospital Discharge Decisions

  1. Pingback: Nightcap | Notes On Liberty

  2. I admit I find your comments confusing, especially this line:

    … if only the Peruvians had lacked a credible answer as to where they would go, all those patients would have survived. It’s only when you have a credible answer as to the patient’s destination upon discharge that the patient suffers premature mortality.

    Maybe I just don’t understand what you mean by “credible answer”; the phrase does seem odd in this context. You also used the phrase, “determinate plan for discharge”; if that’s what you meant by a “credible answer” then it appears to me you’re misusing the term “credible”

    Certainly there must not be a one-size-fits-all rule for post-acute cases; but central planning does not imply such a thing; and we all can name individuals who all on their own have made stupendously bad decisions.

    Just like any individual; a centralized authority can act with awareness of their limitations. That is not an incredible expectation.

    sean s.

    Liked by 1 person

    • My post was meant to be a parody of recent uses of Hayek’s “The Use of Knowledge in Society.” It’s intended to be facetious from beginning to end. Either I played it so deadpan that the sarcasm didn’t come through, or I threw so much sarcasm on there that my point didn’t come through. The whole comment is deliberately meant to be absurd.

      Here’s a link to Hayek’s paper:


      It makes some good points, but I don’t regard it as a particularly well-written or well-argued paper. The thesis is that if “planning” is taken to mean centralized economic planning by a single authority, pictured as an omniscient and infallible being, then the price system will do a better job of allocating resources than “planning,” given the more modest demands a price system makes on people in terms of knowledge. True enough, but why understand “planning” in that extravagant way? The most charitable answer is that Hayek was arguing against a certain brand of socialist who understood it in that way, but if so, the paper has very limited utility.

      That’s not how libertarians usually read it. They read it as though it had argued that any plan offered by any government authority in any context for any reason is liable to fail simply because it is a “centralized plan.” They’ll say this even if the plan in question is operating in a system that allows for quite a lot of decentralization, e.g., a federal system like our own.

      There is, after all, a sense in which a state health commission is a “centralized authority.” It’s the single health commission for that state. For that matter, a town’s health department might be regarded as a “centralized authority.” It’s the single health department for that town. Etc. So if we accept this hard-core libertarian reading of Hayek, we reach the conclusion that health commissions always fail, whether at the state or municipal level. Why? Because they’re all centralized authorities, and those always fail.

      If “centralized authority” has this elastic meaning, it really just means “authority.” But if the claim being made about the futility of planning isn’t even tied in any essential way to planning-by-a-coercive-state-authority (and Hayek doesn’t do so), but just amounts to a criticism of the futility of planning per se, we’re led to the ridiculous conclusion that anytime any collective gets together to plan anything, it will be defeated by the impossibility of centralized planning.

      That sounds silly, but in decades of experience with libertarian rhetoric I’ve heard people make that claim over and over and over ad nauseam. And I’ve heard it on the nursing home issue, as well.

      So, in New York and New Jersey, the health commissioners (and by implication the state governments) issued this nursing home directive saying that nursing homes must accept COVID positive patients for all the reasons I’ve been belaboring in these past few posts. The most extreme right-wing (including libertarian) criticism of the directive is that it amounted to “murder of the elderly,” knowingly sending elders to their deaths in a manner tantamount to putting Jews on trains and sending them to Auschwitz. (Yes, the claim is made in that exact form, invoking Auschwitz, so that Murphy and Cuomo, along with Whitmer in Michigan, all turn out to be Nazis.) The more moderate claim is the non-moral Hayekian version: the rule was bound to fail, because after all, it originated in a state health commission. Why was it bound to fail? Well, because “centralized planning” is always bound to fail, and centralized planning is what state health agencies do.

      I’ve been asking now for over a month where the “failure” was, and haven’t gotten a single cogent answer. If it was so obviously a failure, you would expect the people alleging failure to explain what this failure was. The enormous irony is that the people alleging failure are doing so at a remove from the action, both in space and time. Meanwhile, they are second-guessing state health authorities who were right here, directly proximate to the events in question. Yet somehow, the directives of people right here, with local knowledge, were “failures of central planning.” Whereas evidently, criticisms made from afar–by critics who can’t even identify the error being alleged weeks after the fact–make perfect sense, and are totally compatible with Hayek’s strictures.

      Hopefully now the post makes more sense, read as a parody of this attitude. If we can accuse the state for killing all those people in the nursing homes, can we accuse Peruvian householders for killing all the people who died at home in Peru? The straight answer to both questions is “no.” But people committed to thinking that they can attack Cuomo and Murphy should now be attacking the Peruvians. Naturally, they aren’t.

      Sometimes you’ll hear them say, cavalierly, that post-acute COVID patients should just have been held in the hospitals. But now that Peru is experiencing in real time what we experienced a few weeks ago, we should be able to see the absurdity of that suggestion. How do you warehouse 14,000 post-acute COVID patients while you’re processing an incoming deluge of acute COVID patients? I find it incredible that these critics have no answer to this question. Where did they propose to put all those post-acute patients? In the hospital’s basement? In the parking lot? In filing cabinets? The irony is that Hayek’s essay is about social scientists’ ignorance of matters of logistical detail. But that’s exactly what these critics are guilty of. They’re so focused on beating up the regulatory state that they haven’t asked themselves even the simplest question about where all those patients were physically supposed to go. Which only goes to show that the patients are an afterthought to the whole polemical exercise.

      In other words, these critics literally lack any credible answers to basic questions. Meanwhile, they feel free to criticize those who had and enacted answers to the very questions they themselves have refused to ask. It seems surreal, but it’s as though they thought that lacking answers was a better idea than having them. Why? Well, having them would be central planning, right? We can’t have that. Better to have no plan than to have one. If you engage in planning, you’re doing something Soviet-like. Whereas if you don’t…

      One reason you may have been confused by my post is that I’m parodying a line of thought that’s so absurd it’s hard to process. On what planet would anyone think that things go better if you have no systematic plan for dealing with the various contingencies you may face? Answer: on Planet Libertarian.

      The last paragraph of the original post is a parody as well. Every claim in it is meant to be absurd. If you really thought that planning was a futile, crazy activity, wouldn’t you just dispense with hospital discharge….plans altogether? If planning is such a terrible thing, why not discharge patients without planning anything at all? Fire all the social workers, and all the hospital-LTC facility liaison officers, and just have newly recovered COVID patients figure out their own discharge orders on their own, with the assistance of the one thing that Hayek valorizes as the solution to all problems: the price system. How would it work? I have no idea. Neither do they.

      The plan that New York and New Jersey came up with was not the insane piece of Soviet-style central planning that libertarians are making it out to be. Putting aside whether a given state agency should have a given function or not; if a given agency has good information, and good personnel, and good resources, nothing prevents it from enacting a reasonable plan for dealing with some social problem. It depends on the specifics of the case at hand. It just seems absurd to rail, as Hayekians so often do, against institutionalized “planning” as such, and insist that every problem on Earth must be solved by a price system, and only by a price system, on pain of imitating the Soviet Union. Prices have their place, but they’re not magic, and can’t be applied to everything. “Centralized” or institutionalized planning can go wrong, but it can also go right, and it’s sometimes indispensable.

      Contemporary libertarianism has dispensed with such merely commonsensical observations, replacing them with grand pronouncements that often enough manage to run afoul of reality. When they do, reality starts to get the blame, not the grand pronouncements. That’s what the nursing home controversy has been like. A state agency acted, and people died. That one fact has been enough to get libertarians revved up and ready to stick it to The State. It doesn’t seem to matter to them that there’s no coherent accusation to make. It doesn’t matter that they had no better ideas about what should have been done. What matters is the eager willingness to capitalize on all those dead bodies. No parody could capture the absurdity involved.


  3. I’ll take the fault for not knowing Hayek well enough to catch the parody/sarcasm. Mea Culpa.

    On a Totally Unrelated Matter: what do you think the libertarian response will be to the “defund the police” movement. Seems like another opportunity for them to regret what they wish for.

    sean s.


  4. When I ask advocates of “defund” what the heck they mean, it becomes complicated. This demonstrates it’s at best a badly crafted slogan and needs to be ditched. “Demilitarize” works better. It remains complicated but that is the gist of it i think.

    I am about half way through that Hayek paper; too many other higher priorities just now to focus on it. I’ve read worse.

    sean s.


    • I agree; the Hayek paper is often genuinely insightful. The point he makes about the sheer the distance between their social scientific models/normative theories and the real world is profoundly right. It just doesn’t have the sweeping implications he takes it to have, much less the cosmic implications libertarians have drawn from it.

      There are some not-bad accounts of what “defund the police” means.

      My objection is not so much with the messaging, but with the message itself. I object to indiscriminate, on-the-fly “starve the beast” budget-cutting of any kind. It doesn’t matter whether what we’re talking about is funding for the public library, or Medicaid, or a university chemistry department, or enrollment management, or even something as bloated as the Pentagon or a university administration or NCAA athletics scholarships. A budget, once passed, is a promise to spend money a certain way. To undo budgets on the fly is to break promises in an ad hoc way, under ideological pressure. Call me old-fashioned, but I object to that way of doing things.

      Even if you wait until the next budget deliberation, so that you’re not literally breaking a promise, budget-cutting is a delicate process that should be done slowly, in a methodical, discriminate way. Unless we’re talking about an obviously immoral, quasi-criminal enterprise that should be stopped immediately, I think it’s very wrongheaded to make budget decisions under pressure. It’s one thing to resolve that the money should be re-directed. It’s another thing to just start eliminating whole budget lines on the spur of the moment. That’s a recipe for the subversion of trust and for everlasting resentment. You can’t do politics that way. You can only create the conditions for a civil war.

      My problem with the current protest movement is that it fixates on one truth at the expense of another. The one truth it recognizes is that we face a systemic problem in our police forces in this country. But the truth that gets lost in the shuffle is that we desperately need good policing. Policing is in some respects analogous to our foreign policy, but one major disanalogy is that if we simply did away with our foreign interventions, we’d be fine. By contrast, if we simply did away with policing, we would not be fine. You don’t lightly dismantle something that necessary to social existence unless you know how you’re going to put it back together.

      The example being invoked is Camden, New Jersey, whose police force was in effect reconstructed from the ground up. It’s a stirring, wonderful story.


      If it can be repeated, I’d be the first to say, “Let’s go.” But it is cheap rhetoric to just wave the “Camden example” in the face of skeptics and say, “Oh, if it worked in Camden, it will work everywhere.” That’s what was said about “Broken Windows Policing.” It supposedly “worked” in New York City. Actually, it kind of didn’t work anywhere, but its architects have not given up on defending it. You can’t just flourish an experiment with a sample size of 1, and say, “There’s the solution!”

      The thing to note about Camden is that it didn’t happen as the result of mass protests of an adversarial sort. It took place in a very different atmosphere. That’s why it got buy-in on both sides. It is not obvious to me that you can achieve what BLM wants by insisting on more adversariality. I’m the last person to object to adversarial proceedings, but they only go so far, unless your object is outright warfare. “Defund the police” is not a slogan calculated to inspire buy-in by the other side. It’s a slogan calculated to put them on the defensive. But you can’t really negotiate with people once you put them in that state. No matter how many paper victories we get out of these protests, those “victories” will not stick if the police don’t buy in. Like racism itself, their resentments will just go underground. That’s not my idea of success. I wish it wasn’t anyone else’s, either.

      There’s also the practical issue of how you engineer reform with less money than you had before you started, but let it go. I’m starting to sound like my accountant.


  5. I think we agree “defund” is a bad slogan as much as a bad idea. I’m seeing many thoughtful objections to it already for reasons that often parallel your own.

    sean s.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s