The lovely and talented Scott Alexander has a posting on Cost Disease: the costs of some things, notably education and medical care especially in the USA, have increased in the last few generations to a really unfathomable extent. He gives detailed statistics, but it's typically about a factor of 10 after accounting for general inflation. Why has this happened? He gives some hypotheses, and in a followup posting shares some ideas contributed by readers, but it's not at all clear what's going on. And it seems like knowing might be valuable, because the fact of this phenomenon's occurrence (whatever the cause) is causing a great deal of misery for a whole lot of people, bearing on many other important issues.
I don't know either, but it made me think of some things.
The Horror of the Mall
I don't like shopping malls. When I go to one, I can feel my mental protective filters kicking in. It's like I don't even really see a majority of the stores - because the mall is mostly clothing stores. The fraction of storefronts devoted to clothing alone feels grossly disproportionate. If I go to a mall's Web site and visit the alphabetical list of tenants, maybe there'll be a name on it I don't recognize. So I click on it, thinking it might be something interesting - but no, it's just another damn clothing store. What is with all these clothing stores? How many do we need?
Clothing is a basic necessity. Everybody needs to buy it on an ongoing basis. I don't keep exact records of this, but I figure I myself spend a few hundred dollars per year on clothing, out of my income which is a few tens of thousands of dollars per year. So, maybe I spend 1% to 3% of my income (probably nearer the low end of that range) on clothing. On that basis at first glance it would seem we need somewhere around one clothing store per mall complex. Maybe not every mall really needs to have a clothing store. So when I go to the mall I mentally do that calculation and then am horrified at how it differs from reality.
But I don't spend all my money at the mall. Very much of my income goes to housing - and okay there might be a real estate agent's storefront in the mall, but it's understandable that there wouldn't be many of them per mall in proportion to the money involved, because real estate is handled in a few large transactions instead of many smaller ones like clothing. We don't need so much infrastructure to process the small number of real estate transactions even if they involve a lot of money.
Similarly comparing number of visits instead of dollars, I visit clothing stores a little more often than dentists. I'm not sure which I enjoy more. So maybe that can help you understand my perspective if you don't already: suppose you went to a mall, 30 storefronts in it, and you noticed that even though you thought it was just a regular mall, 18 of the storefronts are dentists' offices, and nobody but you seems to think there's anything weird about that! How many dentists does our culture really need? That's how I feel every time I go to the actual mall and see all the clothing stores.
Other necessities, like food, often aren't bought at malls, and when people do buy food at the mall, they'll do it at an "anchor tenant" supermarket that actually does take up a large fraction of the total floor space after all. I'm applying an unfair standard when I count number of store fronts instead of total floor space. And for ready-to-eat food, the food court is also a significant percentage of total floor space in the mall. Similarly, the costs of transportation don't show up indoors at the mall even though they're significant in most people's budgets, and the mall does have a large parking lot, a dedicated bus stop, and so on.
It still feels like the amount of resources devoted to clothing retail is grossly disproportionate, though. There've been times in my life when I've seen that as a strong argument in favour of nudity, not that we needed any more reasons, but it's hard to get anyone to listen to that particular line of thinking right here and now - I'm in Toronto and it's February.
So let's look at a different angle. It would be impossible for all those clothing stores to remain in business if everybody bought clothes the way I do. So my way of life must not be the only one. It must be the case that not everybody buys clothes the way I do. There must be people whose entire relationship to the clothing they wear is fundamentally different from mine, who spend a qualitatively different fraction of their money just on clothing and on keeping all those storefronts in business. And these people must be quite a large fraction of the general population, maybe even a majority, to the point that if anyone's clothing spending is unusual it's mine, much more so than theirs.
At this point it's easy to snicker and say, "Well, yes, Matthew, these strange creatures of which you speak are called 'women'; you must not get out very much!" But I don't think it's that simple - sure there are a lot of stores in the mall that exclusively sell clothing aimed at women, but there are also a really bizarrely large number (from my point of view) of clothing stores with products aimed at all genders and even just aimed at men exclusively. The general population without regard to gender, and each gender considered separately, all necessarily contain large fractions of people whose clothing-buying habits are night-and-day different from mine.
The take-away lesson from my mall horror, for Cost Disease: it's not just across time. At the same time, in the same general population, in the same income bracket and the same gender, there can be two people who both appear to be ordinary people and yet have drastic, qualitative differences in their spending patterns.
Consider this painting by Pieter Bruegel the Elder.
These are ordinary average people very much like myself, except for being Dutch about 450 years ago. Those guys at the left are writing Perl scripts for an Internet pornography Web site. In the middle, the one on the ground under the tree is proving a theorem about extremal graphs. At the right, there's a group listening to someone lecture about data structure analysis. And in the back you can see a woman bent over with a soldering iron assembling a synthesizer module. All things I've done for a living at different times in my life - and even though it's 450 years and an ocean away from where I am, the clothing they're wearing for it is not all that much different from mine. It doesn't look notably more or less expensive. How much do people like them spend on clothing?
As I type this I'm wearing a shirt that I bought for about $35.00.
Eve Fisher wrote an article about that which made the rounds a few years ago, and a followup you should also read. She figures shirts like those in the painting in the mid-1500s, in proportion to the wealth of the people involved, cost the equivalent in 2013 of something like $3500 each. Note the decimal point. That is one hundred times the $35.00 price of my shirt. And these are not fancy shirts and not fancy people wearing them.
What would my life be like if my clothing cost that much? Well, I'd own a lot less of it, for a start - even though as my previous comments demonstrate, I already am less involved with clothing than others in my culture seem to be. I would also see my clothing, and clothing in general, in a very different light. My relationship with my shirt would be so different, if it cost $3500 instead of $35.00 and if it were normal for a shirt to cost that, that it's hard for me to even imagine how such a life would feel. For a start I would be really impressed by anyone wearing clothing noticeably finer than mine, in a way that I'm not, here and now.
But that $3500 number wouldn't actually have been paid in money, because those shirts in that painting wouldn't actually have been bought. It was paid in labour - people made clothing for themselves and their families. Gender comes back in here: it was just a basic fact through much of human history in the cultures that preceded mine and nobody thought anything of it that what women did was make and maintain clothing. All. The. Time.
And here and now, that is not the case. Clothing just isn't a big deal for us in North America today the way it was for our ancestors. The increase in the cost of education and health care in 50 years is almost unimaginable, causing those things to become Big Deals in a way they weren't before and completely warping our financial lives and our lives in general; the decrease in the cost of clothing in 450 years is also unimaginable and has had similarly extreme effects.
I think the history of clothing once having cost much more may be part of the key to the puzzle of the clothing stores in the mall, but I still don't really understand that. As for Cost Disease, I think the lesson is that it's possible for overwhelming cost changes over time to go in either direction. It's not a pure trade-off. We're not simply spending the money on education now that we used to on clothing. For one thing, with the shirts we're talking about a factor of 100 in about 450 years and with education and health care it's more like a factor of 10 in 50 years, a much faster increase than decrease over a much different time interval, which makes it less than natural to draw any kind of connection between them. But even without a connection between these two specific phenomena, we can say that in more general terms, overwhelming cost changes in both directions, not only increases, are known to occur; and maybe we need to think about both directions to really understand what's going on with overwhelming cost changes.
Doing it yourself
There's a scene in Anne of Green Gables where Anne is punished by not being allowed to attend a picnic, for stealing and subsequently losing Marilla's amethyst brooch. Then, shortly before the picnic is due to start, it's discovered that Anne is in fact innocent of this crime - so she can go after all. But she doesn't have a dress to wear. So Marilla, who feels badly about the false accusation, quickly "runs up" a dress from fabric she has. I don't think the book says how long that takes (it's been a while since I read it), but it doesn't seem to be very long. Maybe an hour.
Just think of that! An ordinary person - certainly not a professional dressmaker, and not a rich person by any means - just happens to have on hand, in the normal course of her life, all the materials, tools, and above all the skills such that she can produce a usable custom-made dress starting from raw fabric, in maybe an hour, maybe a little longer, anyway a short enough time that the kid wasn't very late to the picnic. And it probably wasn't a high-end dress, but it was good enough that a teenage girl was willing to be seen in it in public. And in the book's setting (late 19th Century Atlantic Canada) that's not a particularly remarkable thing. Of course Marilla can run up a dress. That's what people like her do when they need to. How many people do you know who could do it today?
Well, my mother probably could. But her level of dressmaking skill is higher than what's usually expected for her generation, and it would be even more unusual in a later generation. Another story about my mother: when they were married she told my father she wouldn't darn socks. One or two generations earlier that wouldn't have been open to negotiation, wives had to darn socks. In the subsequent generation (mine), most women let alone men can't darn socks whether they want to or not. And I think people significantly younger than me don't know what darning socks even means.
Do you understand the line in the Beatles song about Father MacKenzie darning his own socks? He's a celibate Catholic priest, without a wife to do that. Darning is a sewing technique for repairing holes, commonly used on socks, where you basically weave a new section of cloth that fits into the hole. It takes a fair bit of work, and the results aren't all that great unless you're good at it.
My socks come in 12-packs from Bangladesh and I throw them out when they get holes in them. If you are poor you buy cheaper socks and if you are rich you buy more expensive socks, but either way you throw them out. Yes, some people do still knit socks by hand and the results are valuable enough to be worth a darn (hey do you think that might be related to where that phrase comes from?) but then there are also people who camp out in the most miserable weather they can find to eat hardtack and pretend it's the Revolutionary War. These people marry each other, pretty often.
So there's a thing that seems to be related to the unimaginably extreme decrease in clothing prices: the decrease in prices has gone hand-in-hand with the shifting of labour, mostly to machines but also to human beings in other countries, away from the end users. Instead of your clothing being made by you or the women of your family, now it's made by machines operated by strangers very far away, and that is why it's so much cheaper, and also why you (or, again, the women of your family...) have lost the skills and even the mental models associated with it.
I don't know what exactly that has to do with education and health care. But I note that very much of the decrease in clothing costs seems to come from automation. In what way is automation relevant to education and health care? That's not clear to me. Scott Alexander talks about it a bit but I don't see a smoking gun. We've seen increases in class sizes and per-doctor patient loads, each at least partly attributable to automation, but we haven't seen tenfold such increases; and if it's exactly the same effect seen with clothing, then its consequences should be going in the opposite direction from what has actually occurred. (Less labour needed, therefore lower prices.) We could argue in the opposite direction that because education and healthcare cannot be automated (which is not obviously true, but let's go with it) their prices should go up as automatable things remain steady, but the numbers don't seem to work well for that either.
Geographic distance seems to make a difference for clothes. I wonder - I really don't know what the answer to this is - what the statistics are on how often people go to college far away from where they grew up (let's say, from the place where they graduated high school, as a useful definition of that). If there's been a significant change (either direction) in people going far away for college, over the same time period as the increase in tuition, we could tell a just-so story about how things that are supplied over increasing geographic distances will naturally have their prices undergo drastic changes in one direction or another.
Maybe it's about the limits of globalization. You can get your socks made in Bangladesh really cheap. There's a pretty good university there, too, but I don't think it has many students from the USA. Everything that has decreased dramatically in price in the USA is something routinely sold across national borders from very far away. (I will leave the prognosis of near-future US trade policy for another time.) Everything that has increased dramatically is somehow tied to geography - often because it's a service the recipient has to receive in person. So maybe that's relevant.
For education in particular I'm aware of a claim with some scientific support that it really just comes down to student loan subsidies: prices simply increase as necessary to negate the effect of the subsidies, no matter how large the subsidies are made. It may be possible to construct a similar theory for health care. I'm not sure it can be extended to some of the other examples Alexander considers, though. Some things have gotten extraordinarily expensive without being subsidized.
Chicken, steak, lobster, and aluminum foil
David Manheim comments on substitution of higher-quality goods, using the example of someone who eats chicken regularly, then gets more income, and instead of eating a larger quantity of the same thing (i.e. more chicken), switches to the same quantity of a more expensive thing: beef steak. This doesn't directly bear on his point - he could have used any two commodities that serve similar needs with a price difference - but I think it's an interesting choice of example because it highlights another significant price change over time.
Chicken wasn't the cheap "budget" meat until the mid-20th Century. It was formerly an expensive luxury; beef was never exactly cheap but once would have been closer to being an everyday food. At one point there were significant geographic differences - my parents told me of a friend of theirs who pinched pennies eating chicken in the USA in the 1960s to save money for a trip to visit his family in Argentina, home of beef, only to find once he got there that they wanted to make a fuss over him by treating him to locally-expensive chicken. Just like the shirts that went from $3500 to $35 because of globalization and industrialization (much cheaper labour and much less of it), chicken's place in US lives was transformed by the industrial revolution, specifically factory farming techniques.
Herbert Hoover, President of the USA in 1929-1933, was famous for allegedly promising to put "a chicken in every pot and two cars in every garage," and what that meant was "economic prosperity": everybody would become rich enough to be able to afford chicken. You can maybe infer from the dates how well it worked. We can only hope that analogous, more recent, Presidential promises will turn out better.
Henry IV of France is thought to have said that he wanted "a chicken in every peasant's pot on Sundays" circa the late 1500s, so there was already some significant price deflation between Henry and Hoover even though both were before the advent of factory farming. This was relevant in pre-factory-farming Green Gables, too - there's a bit near the end of the book where Anne is flattered and impressed at being served chicken, in a way people today wouldn't be.
And that book is set in Prince Edward Island, now known for its lobsters. If you want to go all out on a luxurious meal, maybe you have "surf'n'turf": that is, steak and lobster, the latter quite possibly shipped from PEI. Lobsters have to be cooked alive because they start decomposing almost instantly upon death. So to eat fresh lobster outside the places where they're caught is possible only because they can be shipped, expensively, on jet planes in the relatively short time they remain alive after they're hauled up from the traps.
Before air freight was available, in the mid-20th Century (within the memories of people I've talked to who experienced it first-hand) lobster was famine food. If you lived in Atlantic Canada and you couldn't afford regular food, you'd go down to the shore and catch lobsters in the tide pools, and you'd do it on the sly because it was shameful. You wouldn't want people to know that you were so poor you were forced to eat lobster.
That's another commodity on which price has shifted in a game-changing way, though in the opposite direction from chicken. Air freight didn't make lobsters cheaper for those of us not on the coast; it made them available commercially at all. Then, even though catching lobsters had basically always been possible in the places where they live, there came to be a market for lobsters in places where there formerly wasn't such a market, which caused the prices to go up from basically zero to significantly more than zero. But if you go to PEI and have a "lobster supper" at a waterfront restaurant, it's actually not so expensive. Certainly less than steak. The social status of lobster has changed even in its home, but the price is only very high when you're somewhere else and you're paying for the jet plane.
Back on chicken, maybe after you eat part of your cheap meal you want to save the rest for later, so you wrap it in aluminum foil, right? Aluminum ore is common, but the metal itself was more precious than gold until the late 1800s when they finally figured out a way to smelt it that really worked. The technique involves a lot of electricity, which wasn't a commercial commodity until that time.
As of this writing, aluminum costs about $0.0058 per ounce troy (just over half a cent), while gold costs about $1234.60 per ounce troy: a factor of 21267 difference. (US dollars; figures from Kitco for gold and aluminum; note the prices on those charts are quoted per ounce troy (31.1035 grams) and pound avoirdupois (453.592 grams) respectively.) I don't know just how much more valuable than gold aluminum really was before the development of the electrolytic smelting process. If we assume they were at par weight-for-weight (the most conservative assumption available), this is equivalent to about 4.3 factor-of-ten cuts in the aluminum/gold price ratio in about 230 years - a speed of price decrease at least as fast as the recent increases in education and health care.
So one take-away lesson is that after an industrial revolution has caused a dramatic price change, our repertoire of ideas will be changed so much that even the throwaway examples we use in discussing the way our viewpoint has changed, will be different. Switching from chicken to steak because you have more money sounds perfectly reasonable to us today, but that's only because we've been the beneficiaries of an overwhelming game-changing technological price shift related to chicken. That happened only a few decades back and we already naturally assume that chicken is a cheap meat just because it is. We don't naturally think, though it is true, that the cheapness of chicken is a remarkable and recent development; and we implicitly assume that things always will be this way. Another lesson is that, like the availability of air freight causing lobster to change from a famine food to an expensive luxury, it's possible for new technology to cause dramatic price increases, not only decreases, through complicated mechanisms that may not be easy to predict.
Stories in Heaven
As a sometime science fiction writer I've given a fair bit of thought to how to construct plots in settings that differ very much from our world, and one that interests me is the one I call Heaven, or someone else might call Utopia: a perfect world where everybody has everything they need. Or, anyway, they seem to... because if you're going to have a story, you have to have conflict of some kind, or else it'll be boring and not really a story at all. What is the final conflict - the last one to exist after improvements in the general state of humanity have destroyed all other conflicts?
We might assume that something like the most optimistic predictions of the nanotechnology speculators and transhumanists has come true. Instant matter-to-matter conversion with no resource limits. Any material object you want, you've got it. Artificial general intelligence comparable to humans with none of the singularity hazards that otherwise-intelligent philosophers have fooled themselves into worrying about. If there's a difficult or unpleasant task that needs to be done, you can get a machine to do it, and the machines are perfect p-zombies with no true consciousness; there is no moral issue about forcing them to work for you. Humans only have to think about hard things, or do any sort of work at all, if they want to for art or recreation. Everybody's immortal in the strongest sense, meaning that they not only don't age or get physically or mentally sick any more than they want to, but also cannot be killed or harmed at all - none of this "eagle eats your liver every day, oh the eternal agony" type of thing is possible. There's no money in this world because there's no need for it, so nobody's poor; money, and the concept of an economy at all, exists for the purpose of dividing up limited resources. What about a setting with no resource limitations - a Strong Heaven?
What almost everybody does to create story plots in that kind of setting, is compromise it into some sort of Weak Heaven. Things are somehow not really as good as they seem, or they cannot remain that way. It's always "Yes but." Yes but they forgot how to maintain the Machine and eventually it Stopped. (E.M. Forster) Yes but it all depends for unspecified mystical reasons on the misery of one child. (Ursula K. LeGuin) Yes but the robots are not really p-zombies after all. (Karel Čapek, Isaac Asimov, and basically everybody who has ever written about humanoid robots) Yes but the demons will eventually come back, perhaps when the stars are right. And so on.
I'm interested in finding plots that could work in the Heaven setting without compromising away its definition in a "yes but," or at least, without making the same compromises that have already been tried. My story Race is actually on the periphery of that project, even though it's not exactly the same, nor science fiction: in that story there are no villains, all the characters are friends and they all sincerely intend to be nice to each other, some characters maybe could have made better choices but nobody seems to be exactly at fault, there's no external disaster either, and yet someone ends up getting hurt pretty badly.
So how do you get conflict in Heaven? You can't pit characters against the environment because the environment by definition can never threaten or limit them. Maybe you can conceive of some sort of truly insoluble philosophical problem for gods to worry about, but it's hard to make that stick as a story motivator. You can try to pit characters against each other, and that seems ultimately to be what you have to do, but it's not easy to make that stick either. If we're both godlike beings in an unlimited environment, you can't kill me or harm me, it's not clear why you would want to anyway, all you can do is talk to me and if I find that annoying I can just go do my god thing somewhere else.
The author can start playing one-up with irresistible forces and immoveable objects. It's bad enough trying to define exactly what "omnipotent" is supposed to mean when you only think one entity might be it; worse if there are two or more. If you and I are both omnipotent, how able am I to prevent you from continuing to pester me? For anyone to be able to impose on another's omnipotence quickly becomes another compromise of the setting definition. It does seem to be the key to creating conflict, though. In an environment where we are all gods, the one thing I can't have in unlimited amounts is the real unsimulated and uncoerced recognition and attention of other beings equal to myself. The demand for that increases in principle exactly as the supply increases, supply never outstripping demand.
There can only be one Top Dog or Top God; as the population of the world increases, the number of lists someone can be at the top of doesn't increase as fast if at all, so the possibility of being at the top of a list you care about, must decline.
If I want other beings equal to myself to listen to me, or talk to me, or have sex with me or whatever such post-human beings do instead of that, okay I can have the p-zombies perform those services for me, while assuming the physical or post-physical forms of whoever I want, but it won't be real. Note that wanting it to be real is key to the Turtle's issues in the Race story, and although all my characters are autobiographical, it's probably clear that in that story he's the one most like myself. And if I know the difference between real attention from others on my own level, and fake attention from the p-zombies, then scarcity of real attention can be a source of conflict to drive a story. It seems like the so-called "attention economy" could exist even in a Strong Heaven.
Sidelight: consider that I wrote "real unsimulated and uncoerced." Even if we have a setting where there's only one omnipotent entity who has no peers, could that entity be so starved for attention that it would choose to create others, lesser than itself but really conscious with their own free will and not p-zombies, just so that it would be possible even though by definition not guaranteed that the new beings might choose to give it attention? Does this sound like anybody we know? Yes, that's Haruhi Suzumiya. But it's also someone in a much older franchise you might have heard of.
How does all that connect to the Cost Disease? It doesn't have to connect perfectly. It's just something I think about a lot in general, and thought about again when I read the Alexander article. But some of my thinking is that even though we're far away from living in a Strong Heaven, we do by some measures have a lack of scarcity. Consider my $35 shirt from the point of view of those old Dutch people for whom shirts were a Big Deal. If you told them about my clothing, with appropriate translation of the exchange rate, it might seem that I am already living in an amazing transhumanist post-scarcity world without any resource limitations whatsoever. Anything I might still want and can't have would be something that the characters in that painting wouldn't care about or comprehend... except attention.
The pattern of Cost Disease seems to be related to things that inextricably require the unsubstitutable labour and attention not just of human beings but of human beings somehow comparable to the buyer. (Americans, for the US focus of most of this discussion.) Education not only requires teachers who are part of the same cultural milieu as their students, but it requires the attention of the students themselves, and attention is inherently expensive. As the only thing that can be expensive in the final Strong Heaven, attention predictably gets more expensive in a culture that moves more and more toward general post-scarcity. Health care similarly requires local human involvement.
Alexander also looks at infrastructure (tied to geography, not subject to much automation, but on the other hand not so obviously tied to attention either) and housing. Housing is tied to geography and seems to also be very much tied to what I'm calling "attention" if we take that in a broad sense. It's amusingly easy to imagine angels or deities bickering over who gets which apartment in a Strong Heaven where there are absolutely enough apartments for everyone and they're all objectively good. The value of housing comes in large part from the fact others want it and my having it means they can't, in a way that is independent of its intrinsic qualities, and in a way that can't be so easily inflated away just by building more.
My view of the Net is shaped by my choices of where I go, but consider what we fight about on the Net. The big fights today are ad-blocking; spam; "harassment" which on closer examination usually turns out to be stuff like "randos in my mentions" instead of what that word actually means; who does or doesn't get the magic blue checkmark; algorithmic curation... everything is about attention and who is allowed to impose communication on whom. Despite infernal proposals for "micropayments," attention is the only currency that means anything on the Net. Attention is the only thing that can always be scarce. The Net may not be a Strong Heaven, but it supports the same kinds of plots for fiction, and this even though the Net is just a small part of a larger world that for the most part is in nothing like post-scarcity.
I'll leave it there. I think attention is an important part of the puzzle. Geography may be even more so - it seems a common thread among the examples - but I haven't said much about that because I don't have a clear idea what its role actually is. There may be other pieces missing, too. But just the fact of these huge price changes, even if I'm all wet about where they may have come from, seems like an important thing to understand. I hadn't really thought through before now just to what extent drastic price shifts have happened in the last few decades and how big an impact it has had on the culture.