He had me at “hat.”
We can attempt to extrapolate from what we know, but that is very difficult. When we attempt to predict advances that have not yet been made, like the warp drive, that invites disaster. But it’s very tempting to try to do so. The inability of the mind to foresee its own advance is one of the reasons the future will always surprise us.
The other reason is that not all tastes, values, and desires of individual human beings are accessible even to them. That sounds very weird if it’s an unfamiliar concept to you. But I will give you an example that I find especially dramatic. Consider your own face. You probably know things about your face that your own spouse does not know. You probably have very considered opinions about what kind of eyewear looks better or worse, what kind of hat or cosmetics you favor, or what kind of shaving products you prefer. All these things are known to you, but only partially. Sometimes you walk into a store and see a hat that you’ve never seen before and say, “This is perfect! Hey, I did not know that, but this is the one!” And it is this inaccessibility of consumer tastes, values, and preferences that means economic planning is always impossible to do in advance when you want to try to plan for the entire society.
This is from Jason Kuznicki, “The Future History of Liberty,” Cato’s Letter, Fall 2018, Vol. 16, No. 4.
During my staycation, I’m catching up on things in my pile that had gone unread. The whole article is excellent. Kuznicki’s statement up front about the book that did a great deal to make him a libertarian is nicely surprising.
When he used the example of the hat, it immediately worked for me. Not that I’m a hat person that much, although when I play pickle ball I always wear a cap. Rather, I will see a shirt that maybe no one thought I would like–and I love it. My wife is very good at ascertaining my tastes. But even some shirts are ones that she wouldn’t have thought of.
Although truth be told, she hit a home run with her Father’s Day present, a T-shirt that says on the front:
Surely not EVERYONE was Kung-Fu Fighting.
Kuznicki’s piece, especially the quote above, is a nice application of Hayek’s “The Use of Knowledge in Society,” Michael Polanyi’s The Tacit Dimension, and my 7th Pillar of Economic Wisdom.
READER COMMENTS
Dylan
Jul 13 2020 at 8:16pm
I like this example and agree, but I’m kind of wondering about the flip side, which I find happens to me with more frequency, someone else, usually even people that aren’t close to me, will know those unknown tastes better than I will.
Back in the days when people got haircuts, I’d always have a big mop of hair when I’d go to the barber and tell them to do whatever they thought was best. Not because I didn’t care, but because I knew they were more likely to give me something I liked than I would be to think of it.
That’s true in a lot of areas, so, wherever possible I outsource my decision making to others on the theory that a) making choices is exhausting, and b) other people likely know me better than I do anyway.
Phil H
Jul 14 2020 at 2:06am
The argument about the impossibility of knowing detailed information about economic needs is… kinda old hat, isn’t it?
In the age of Google and Facebook and Amazon (not to mention WeChat where I live), and of course, PRISM, it’s time to acknowledge that a central organization, be it governmental or business, can in fact know enormous amounts about consumer demand and the needs of the economy more generally.
This was never the only libertarian argument. But it’s one that has passed its sell-by date in a very definite way, and can safely be retired now, I think.
john hare
Jul 14 2020 at 4:13am
I don’t care how much information you think you have on me, you don’t have the right to make my decisions without my consent. The exceptions are as Dylan notes above when I freely outsource certain choices.
Phil H
Jul 14 2020 at 7:32am
Exactly. That argument about control over your personal situation still works. But for a long time the libertarian argument was: “a central planner cannot possibly be effective because they cannot know enough about economy”. These days, many large companies have “economies” the size of a small country. And they do just fine.
Jon Murphy
Jul 14 2020 at 7:39am
There’s a difference between a firm planning for itself and an entire economy.
The fact that the ant colony runs well does not imply that the entire ecosystem could be run by the ants just as well.
Tyler Wells
Jul 14 2020 at 9:24am
Whether they “know” what is best for me or not our interests are not aligned, whether it be a government entity or private company. Certainly every company or government entity thinks that everyone “needs” what they are mandating/selling.
TMC
Jul 15 2020 at 10:39am
Google and Facebook and Amazon don’t do that great of a job. By their ads, they seem to know what I’ve already ordered, but don’t need anymore since I’ve just ordered it.
Jon Murphy
Jul 14 2020 at 7:30am
The Hayekian knowledge problem is not one of information and data collection. If it were, it’d be trivial. Rather, the Hayekian knowledge problem is of knowledge and transmission.
Hayek discusses that that relevant knowledge needed in economic transactions and itshait’s interpretation has several characteristics:
1) it is tacit and elements are inarticulable
2) it is local to the precise time and place in which the choice made occurs
3) it is subjective, not objective, and cannot be recorded or understood by other people.
Furthermore, as pointed out by other thinkers in the same tradition, what objective information we do get doesn’t provide us with any useful knowledge. To say Joe spent $160 on a dog tells us nothing about how much he values the dog or what his other alternatives are.
There’s a lot more here than I can get into right now. I highly recommend you check out Hayek’s essay linked in the post, as well as Don Lavoie’s National Economic Planning and James Buchanan’s Cost and Choice
Phil H
Jul 14 2020 at 7:40am
“it would be trivial” – no, I disagree with that. Information collection is not trivial, it’s a massively important technology (just look at how profitable it is!), and one that has recently taken a great leap forward. Back in Hayek’s day, it was literally true that central planners could not get their hands on enough objective information to run the economy effectively. And at that time, I think the information argument stood up. Today, it doesn’t. Technology has changed the circumstances in which we’re operating.
I’m not sure your three points make sense. The information needed for economic planning may be tacit and unarticulated, but that doesn’t mean you can’t use it. We know this because supermarkets do it every day. They manipulate vast, global supply chains, in precise anticipation of what their consumers will buy tomorrow, next week, and next year. There’s nothing mystical about economic activity that means it cannot be predicted. The only question is what mechanisms are used to predict it.
Jon Murphy
Jul 14 2020 at 8:14am
The “it” there referred to Hayek’s insight, not information collection. Hayek’s insight would be trivial if he was only talking about information collection.
Indeed, Use of Knowledge in Society was written, in part, as a response to those who said his (and, earlier, Mises’) insights were merely about information collection.
robc
Jul 14 2020 at 9:14am
Grocery stores, even large ones, are examples AGAINST central planning, not for.
They tweak and adjust and HAVE SKIN IN THE GAME, meaning if they get it right, they profit and if they get it wrong, they fail. That is exactly what central planning doesnt have.
And grocery stores clearly can’t plan properly, as recent empty shelves have shown. Its like they didnt even plan for and predict a global pandemic.
Jon Murphy
Jul 14 2020 at 9:20am
The shortages aren’t really the fault of the grocery stores. Price controls are the bigger issue here
robc
Jul 14 2020 at 9:41am
Of course, but not all the items with shortages have price controls on them.
Rice? Yeast? I really don’t know, do either have price controls?
Jon Murphy
Jul 14 2020 at 10:06am
Yes. Ag is heavily price-controlled. It usually doesn’t matter.
Christophe Biocca
Jul 14 2020 at 8:30am
You go from “these things exist and are better than their predecessors” to “therefore they are sufficient” but that’s a huge leap.
Let’s tackle governmental first because the argument is simpler. PRISM’s effectiveness at detecting even the rather extreme consumer preference of “wants to commit terrorism” is near zero.
On the private side, let’s look at Google and targeted advertising, because it’s the subset I’m most familiar with:
Even today most of the ads presented to you are driven by relatively simple factors:
– Your current search and or web page.
– Your recent search and web visit activity.
– A vague notion of your profile as a consumer, built over a longer period.
The first one drives most of what ads you see at the moment. The CPM (cost per thousand impressions) can vary wildly between websites. Travel is famous for having really high revenue per impression (because by the time someone in New York is googling for things to do in Rome, the odds they’ll buy a plane ticket are extremely high). This means google isn’t very good at predicting you want to travel to Rome, until you start giving it massive hints.
The second batch is largely what we call “retargeting”, advertising by a website you recently visited but didn’t become a customer of, on the hopes that it will lead to an eventual sale. This can be a pretty smart strategy, but it is entirely context-unaware. When a friend of mine came back from an internship at Indochino and many of his friends looked the website up out of curiosity, that company ended up spending a lot of dough turning every ad we saw from that point on into ads for their suits, but didn’t get a single sale out of those university students.
The last one is fun because you can test how good Google is in practice at predicting you: https://adssettings.google.com/
In my case it gets my age and gender right, at least, but it believes I want to buy a Hyundai (I don’t have nor want to have a car at all), and products from Wizards of the Coast (nope) as my top two interests.
This is all still a massive advance over a paper ad targeted at the average (or maybe modal) reader of a newspaper, but it’s far from something I’d let make purchasing decisions on my behalf (which is a prerequisite for letting it run the economy).
Phil H
Jul 18 2020 at 12:51pm
Thanks, Christophe. This is a very well-targeted comment. I agree with all the weaknesses you mention, but… however bad this information is (compared to some good/perfect standard that you may have in mind), it seems to be good enough. By which I mean, it supports billions, maybe trillions of dollars of economic activity. That’s qualitatively different to how things used to be, and it damages the old libertarian argument.
robc
Jul 14 2020 at 9:18am
If google/amazon/etc were really good at this stuff, they wouldn’t be showing me ads for things I recently purchases. They would be showing me ads for things I havent yet purchases and don’t even know I want yet.
Do they really think I am going to buy ANOTHER pellet grill?
Dylan
Jul 14 2020 at 9:30am
That’s just what they do to give you a false sense of security. Meanwhile they’ve already shipped you the ribeyes for that grill that you haven’t even realized you were going to order from them yet.
KevinDC
Jul 14 2020 at 11:41am
Oh boy.
I’m genuinely flabbergasted by this comment and your various follow ups. It’s as if you think the knowledge problem is about the ability of organizations were able to gather enormous amounts of information. But…that’s just not remotely true. The argument has never been about something as trivial as how much information firms or governments can accumulate. If you think the knowledge problem is about mere “information collection,” you have not made a serious attempt to actually understand the argument. I can’t imagine how one could, for example, get through Thomas Sowell’s 400 page book on the topic, Knowledge and Decisions, and still walk away with the impression that the knowledge problem depends on how much information can be collected.
Even in the opening few lines of Hayek’s Nobel speech, he made this clear. He starts by pointing out central planners argue their system would be viable if, among other conditions, “we possess all the relevant information.” Not merely “far more information than we currently have.” All of it. He follows this up by declaring “This, however, is emphatically not the economic problem which society faces.” That is, even possessing complete and total information still does not make central planning viable. The rest of his speech outlines why, and as others have indicated there have been scores of scholarly articles and books expanding the argument. So to try to respond to that argument by pointing out that we can access more information these days is a totally impotent reply. That’s why it’s called the “knowledge problem,” not the “information problem.” Knowledge and information are not the same thing, and the difference between them for the purpose of this topic really, really matters. One can have enormous amounts of flawlessly accurate information and yet hold very little knowledge, and one can have a great deal of detailed knowledge without possessing much by way of information. Access to all the information in the world will do exactly zero good as a response to what the knowledge problem actually entails. So if you try to argue against the knowledge problem by appealing to the availability of information, all you’ll do in effect is hold up a giant sign saying “you don’t need to take my arguments seriously.”
Similarly, if you think the success of large firms with lots of information access like Amazon etc vindicates central planning – I’m sorry, but that’s just cringeworthy. Central planning called for the abolition of private industry, the banning of economic competition, an end to the price mechanism, and the cessation of the profit and loss system. So to act as though central planning is vindicated by pointing to a variety of private industries, in economic competition with each other, making decisions guided by market prices, and operating under a profit and loss system? I’m embarrassed just reading that. You’re usually better than this.
KevinDC
Jul 14 2020 at 1:26pm
Correction – the Hayek reference was to his article “The Use of Knowledge in Society,” published in the AER. I mixed it up with his Nobel prize speech, which was “The Pretense of Knowledge.” Ugh. Bad Kevin. Drink more coffee.
Jon Murphy
Jul 14 2020 at 1:41pm
I do that all the time
Phil H
Jul 18 2020 at 1:09pm
“I’m genuinely flabbergasted”
You’re always surprised by me, Kevin. So are sheep. It doesn’t make me more likely to think you’re right.
“The argument has never been about something as trivial as how much information firms or governments can accumulate.”
I’m sure you’re correct that no true Scotsman has ever made that argument, but plenty of Americans have. It took me 0 seconds of searching (even using decrepit old bing) to find this example: https://www.forbes.com/sites/norbertmichel/2014/04/26/dreams-of-a-central-planner/
“Most [economists] realize that large-scale central planning leads to economic disaster because planners can never have access to all the required information.”
Whether you and Hayek like it or not, this is a thing that many people have said. Here’s another example: Kuznicki, quoted above, “And it is this inaccessibility of consumer tastes, values, and preferences that means economic planning is always impossible to do…”
But consumer tastes are no longer inaccessible to Walmart. That’s why they can bet billions of dollars on next year’s products.
I’m very willing to believe that Hayek’s argument is more sophisticated. But I wasn’t responding to Hayek. I was responding to Henderson/Kuznicki. And they (like many before them) absolutely do make the information collection argument.
KevinDC
Jul 19 2020 at 5:03pm
Hey Phil –
While I think basically everything you said was completely off base, in a way, it was also very illuminating. I was suddenly reminded of Paul Krugman’s classic article Ricardo’s Difficult Idea, where he describes the difficulty of getting intelligent people to understand what comparative advantage actually means. Forgive the extended quote, but it seems particular relevant in this context:
I say this with all the love in the world, but this is often how it feels when trying to talk to you about economics. Just like how discussions of comparative advantage simply take for granted the audience already understands all the underlying interconnected concepts (and therefore you don’t need to bother to spell them out), discussions about the economics of information, knowledge, and central planning assume that the audience has already learned the background arguments as well so it’s equally unnecessary to spell out such basic points. Especially if you’re reading the Cato newsletter. Reiterating that the argument uses a Hayekian framework in that publication is as pointless as reiterating the Darwinian framework on the Richard Dawkins message board. (I’m assuming he has a message board somewhere? Is that a thing?)
But enough of the general commentary. Some more specific points:
Your zero seconds of effort link doesn’t serve you well. I said in my initial reply that Hayek’s argument was that central planners were wrong to argue the economy could be centrally planned, even if they had all the relevant information. Your reply is that you’re not trying to respond to Hayek’s argument, and offered that link as an example of the kind of argument you are responding to. You quote all of one line from it to support your point. But in literally the very next sentence, the author goes on to point out that even if they did have all the information, central planning still wouldn’t work for a variety reasons – just like Hayek: “Even if they did have such access, the fact that people are motivated by their own self-interests makes them susceptible to special interests. All of these issues, as well as the absence of market prices, ensure that central planners will waste resources on (most likely) a massive scale.” Granted, he didn’t spell out the argument for why those factors would bring about that result in any detail, but there’s only so much you can explain in one article. Regardless, I’m not surprised to see that you ended the quote where you did.
Additionally, your contention that “consumer tastes are no longer inaccessible to Walmart” continues to be irrelevant. I don’t think anyone has ever denied that successful companies can have detailed knowledge of the preferences of their consumer base, even on a gigantic scale. That’s not what this discussion is about – hence why the quote you were responding to said “planning is always impossible to do in advance when you want to try to plan for the entire society.”
Part of the background that Kuznicki was assuming his audience would be familiar with was this section from Hayek’s article:
So yes, obviously large companies use information available to them and use that information to plan their own economic activity – that has been understood by proponents of the knowledge problem since the beginning. That observation brings exactly nothing to the table. But part of the argument is that these large companies making their plans and getting their information is itself dependent on participating in an ongoing and constantly shifting landscape of competition with other companies, following price signals set by supply and demand, and acting in accordance with incentives due to profits and losses. The fact that individual companies can gain lots of information about their consumer bases as a result of these processes, and that they use that information to plan their specific economic activity, does not entail that therefore we can eliminate those processes altogether and just “plan the entire society” through a central agency.
Mark Z
Jul 15 2020 at 1:18am
“In the age of Google and Facebook and Amazon (not to mention WeChat where I live), and of course, PRISM, it’s time to acknowledge that a central organization, be it governmental or business, can in fact know enormous amounts about consumer demand and the needs of the economy more generally.”
To the extent that this is true (I think it’s false; come the day when these companies’ machine learning algorithms never recommend things you don’t want and never fail to recommend things you know you want, I’ll accept it’s true), isn’t it moot? If technology has enabled us to effectively predict demand for goods and services everywhere, then it will be very profitable for private companies to do so inasmuch as the cost of collecting and using the information isn’t too high. The state has no real advantage in this. Unless the state patented the technology that enabled perfect market planning and prevented anyone else from using it, it’d have no reason to replace private businesses in doing the planning. That’s why I think this is more than just a practical objection to central planning.
Phil H
Jul 18 2020 at 1:13pm
Hi, Mark.
”False…recommendations”
The fact that the information isn’t perfect (or even very good on the individual level) doesn’t mean that it wouldn’t be good enough on the aggregate level.
”Moot…comparative advantage”
Yep, I completely agree with that.
robc
Jul 14 2020 at 7:58am
For me the prime example of the calculation problem is the craft beer industry.
There is no way any central authority (or even google, if Phil H is reading this) could have predicted in the late 70s the demand for craft beer. And even into the 90s, no authority would have said “20% of craft beer should be IPA and variants”.
And who could have predicted a trend in gose?
Jon Murphy
Jul 14 2020 at 8:20am
Or the various forms of craft distilling. In DC, we have a company called One-Eight Distilling. They make craft spirits. Very unique flavors. They’ve been very successful because they’ve defied conventional wisdom and went with what the data didn’t show. I interviewed the owner. What he said was he was looking at the big sales of standard whiskeys like Jack Daniels, Knob Creek, Jim Beam, etc., and all that they were all mixing whiskeys. What people really wanted, in his experience, was flavor. They wanted tasting whiskeys. And boy has he gotten wealthy ignoring the data and sympathizing with customers rather than treating them as mere data points.
Dylan
Jul 14 2020 at 9:25am
Interesting example. I came of age in a region (PNW) where craft brewing has been mainstream since the 80s, I always thought that this would eventually spread around the country and world, but I certainly didn’t predict the level of saturation that would occur. A recent visit to Seattle I was simply astounded at the number of craft breweries in a small neighborhood and wonder how that can possibly be sustainable?
For IPAs, it feels like an industry that has kind of jumped the shark. Around here, IPAs appear to be much more than 20% of the craft market, based on shelf space I’d guess much closer to 80%. What’s weird is that in an industry with such competition, they seem to be leaving big chunks of the market out. I like IPAs, but I know a lot of beer lovers that don’t, and they’ve essentially exited the market now because it is hard to find a beer to their liking. 10-15 years ago the market seemed a lot more mixed, there were plenty of IPAs, but you would also see Amber, Rye, Porters, Wheat Beer, Belgiums, Barley Wine…all of these have essentially disappeared from the retail market and in place you have 100 different versions of IPAs and a bunch of ciders, and a handful of Gose labels.
robc
Jul 14 2020 at 9:44am
I agree. I love IPAs, but I struggle to find the variety I want at times. When I was brewing professionally, I called my IPA “Obligatory IPA” for that reason. But it was not my focus.
Mikk Salu
Jul 15 2020 at 6:22am
Off-topic, I do live in Europe and I see the same trend. They started with IPAs, then market became more mixed, but in recent years IPAs are again dominating. I have a couple of friends in the craft beer business and if I remember them correctly, one of the explanations was that IPAs are (technically) easiest to make, and to my surprise, lagers are the most challenging.
robc
Jul 15 2020 at 7:43am
Lagers are more expensive and time-consuming to make compared to ales. But they aren’t necessarily more challenging. Light lagers, on the other hand, are incredibly challenging, there is nowhere to hide any flaws.
I think Budweiser’s flagship products are crap, but they are amazingly engineered crap. The skill level to produce their beers with the exact same flavor profile over and over again in multiple locations is extremely high. Some of the best craft brewers honed their skills at the big guys.
IPAs aren’t the easiest style to make, but throwing a bunch of hops in can cover up a lot of flaws. But the best of the best IPAs are well made and the hops profile enhances instead of covering up.
Phil H
Jul 18 2020 at 1:16pm
I swear I have an excellent riposte somewhere, but talking about good beer has… uh distracted me a little. If you spot Brewdog anywhere, I highly recommend them!
Thomas Hutcheson
Jul 15 2020 at 8:27am
Did anyone in 2018 need to be reminded that
If not, it is rather not just an argument against any changes in economic policy?
robc
Jul 15 2020 at 11:11am
No, its an argument against having an economic policy.
David Seltzer
Jul 15 2020 at 5:47pm
Robust commentaries. Hayek’s “spontaneous order” is the result of human action. Not design. As for information collection, that activity tells us where we’ve been but does little to predict future prices. Gene Fama’s EMH, in all three forms, confirms this. If I could use historical price behavior of the S&P 500 to predict prices tomorrow, theoretically, I could outperform the market. As trader who made markets in several capital instruments, that seldom occurred. All the information known at time t is impounded in the price of the index. Price transmits the decisions of billions of market agents. When a new piece of information arrives expectations are spontaneously discounted to reflect a new equilibrium. Covid 19 changed those expectations. When news of an effective vaccine comes, markets will react.
Comments are closed.