James Traina writes:
I’m an economist at the SF Fed. I’m writing to ask for your first thoughts or suggested references on a particular problem that’s pervasive in my field: Aggregation of micro effects into macro effects.
This is an issue that has been studied since the 80s. For example, the individual-level estimates of wages on employment using quasi-experimental tax variation are much smaller than aggregate-level estimates using time series variation. More recently, there has been an active debate on how to port individual-level estimates of government transfers on consumption to macro policy.
Given your expertise, I was wondering if you had insight into how you or other folks in the stats / causal inference field would approach this problem structure more generally.
My reply: Here’s a paper from 2006, Multilevel (hierarchical) modeling: What it can and cannot do. The short answer is that you can estimate micro and macro effects in the same model, but you don’t necessarily have causal identification at both levels. It depends on the design.
You’ll also want a theoretical model. For example, in your model, if you want to talk about “the effects of wages,” it can help to consider potential interventions that could affect local wages. Such interventions could be a minimum-wage law, it could be inflation that reduces real (not nominal) wages, it could be national economic conditions that make the labor market more or less competitive, etc. You can also think about potential interventions at an individual level, such as a person getting education or training, marrying or having a child, the person’s employer changing its policies, whatever.
I don’t know enough about your application to give more detail. The point is that “wages” is not in itself a treatment. Wages is a measured variable, and different wage-effecting treatments can have different effects on employment. You can think of these as instruments, even if you’re not actually doing an instrumental variables analysis. Also, treatments that affect individual wages will be different than treatments that affect aggregate wages, so it’s no surprise that they would have different effects on employment. There’s no strong theoretical reason to think that effects would be the same.
Finally, I don’t understand how government transfers connect to wages in your problem. Government transfers do not directly affect wages, do they? So I feel like I’m missing some context here.
“Such interventions could be….etc.”
Ay yi yi, those three little letter mean so much! :) Your list offers the most vague hints at the massive spectrum of interventions that affect wages. Comparing your suggestions with reality is something akin to comparing PreCambrian vision to modern vision! I’m not criticizing the list either: it would take days or possibly weeks to put together a complete list for any given jurisdiction.
Daniel wondered recently why economics as a science has so few absolutes. The reason is because there are so many variables, and because there are so many policies that manipulate economic outcomes that it’s almost impossible to even conceive of the type of controlled tests that are used in normal sciences like physics (mass, angle, friction, push calculate!). The amazing thing about economics is not that it fails so frequently, but that it succeeds so frequently in producing useful information – even if it’s not absolute – given the myriad challenges it faces.
The “its so complicated argument” is often used in bio. In that case it is a cop-out. Researchers aren’t even trained in the tools required to develop a quantitative model (numerical simulation, calculus). They are also actively discouraged from checking each others results because it “lacks novelty”.
So you have people trying to study phenomena with 75% of the “facts” are false and no ability to check the predictions made by their theory. Why in the world would anyone expect this to work? And indeed it doesn’t, the momentum from earlier generations got outpaced by the drag from using this bizarro science ~2014, which is when life expectancy peaked in the US.
This question is about wages, which is measured in dollars per hour. We have pretty good, although not pefect, measures of time. But what about dollars?
Do we have any idea of the actual number of dollars in existence, let alone where they have accumulated and are most likely to flow to?
Eg, there is ~100 trillion in public/private debt in the US. But then there is all the offshore dollar-denominated debt that functions as dollars too (“eurodollars”). No one knows how much of this exists within 2 orders of magnitude. Ie, its probably $10 trillion to $1 quadrillion, most likely around $100 trillion but no one knows.
Then there is the Cantillon effect, one of the most relevant economic topics today, that doesn’t even have a wikipedia page.
It is like trying to do physics without the concept of a force or way to measure mass within a factor of 10. Not going to work.
And that all describes Econ too. Econ has a lot of math prerequisites but it seems like dynamics is rare to consider. Everyone is working with theories about “equilibrium” but lots of people acknowledge that at no point in time is anything in equilibrium.
I’m finally writing a book, and in the book I give an econ example in a chapter that discusses differential equations. The example is a chair manufacturer buys bolts, and has to test the boxes and throw away counterfeits. A wholesaler can’t tell the difference between counterfeit and real and prefers to buy from just slightly cheaper counterfeiter. A main bolt manufacturer produces a lot of bolts but a counterfeit competitor has some fraction of their production capacity but is willing to undercut their prices because they’re selling trash that costs less to make.
I hypothesized control rules for the management of the warehouse and of the chair manufacturer (who also manufactures an alternative product and can decide how much resources to put into the two products). The dynamics are very interesting. No-where in the model will you find a demand curve, or a supply curve, or marginal elasticities or any of that crap. What you have is a very simple plausible model for the behavior of 2 managers.
Daniel
I think you dismiss econ too broadly. While equilibrium analysis has dominated, there is plenty of disequilibrium and dynamic modeling in economics. In particular, dynamic adverse selection models have been explored (e.g., see https://msbfile03.usc.edu/digitalmeasures/zhanghao/intellcont/PolicyGraphAlg-2.pdf). I’m not saying the work is particularly good, but it isn’t fair to say that economists ignore dynamic issues. It is true that Econ 101 ignores dynamics – primarily due to the mathematics required exceeding what students in that course typically can handle – and there is a significant loss in understanding the real world that results. But that is similar to the standard Stat 101 course that presents clean and simple data sets.
Dale, I’m sure there are people who specialize in dynamics, but the few people I have talked to with PhDs in Economics have never even had an ODE course, and I’ve looked at a couple different UC school economics undergraduate degree descriptions in course catalogs and none of them had any ODE or even discrete time dynamics materials. So it’s not just Econ 101 but actually graduating with a BS in Econ and zero background in dynamics that I could see. It’s possible that material is hidden inside the courses where the course descriptions wouldn’t make it obvious.
Of course they’ve heard of demand and supply curves.
Dynamics may be out there but I think it’s fair to say “rare to consider”. All you need to do to convince me otherwise is show me someone explaining the price of eggs over the last 2 years in terms of a dynamic model (or anything really, the price of copper over the last 20 years, the production of computer chips over the last 2 years… mean household monthly expenditures… whatever, as a discrete time or ODE model using some economic theory, and containing some parameters, fit to the first half of the time series, and then predicting for the second half).
There is a long history of disequilibrium models – see, for example https://cemi.ehess.fr/docannexe/file/2360/diamond.pdf. Most graduate programs will expose students to such models – you are correct that most undergrad programs do not. The “justification” is usually made that insights are gained from equilibrium models and that the adjustment process is important in exceptional cases – sort of like the reasons why behavioral economics is not the main paradigm taught in economics. Of course, the real reason may have more to do with the mathematics required – differential equations require more of a math background than algebra, geometry, and basic calculus. The simpler mathematics of behavioral models has permitted it to permeate undergraduate courses more readily than dynamic modeling has.
I probably agree with your general criticism that economists have overemphasized equilibrium modeling, but I’m not sure your examples are particularly convincing. I don’t see that the price of eggs over the past 2 years necessarily requires dynamic modeling, nor necessarily does the price of copper or computer chips. Supply and demand can accomplish a lot, although it often feels like an ex post story about why prices moved as they did (i.e., it must be that supply decreased and demand was relatively inelastic, or some such story). Estimation of demand and/or supply functions is usually taught using fake data designed to fit the theory. I always despised textbooks that use fake data rather than real data (though I agree with Andrew’s frequent urges to recommend simulated data as an important exploratory tool). As soon as you use real data, you run into the problem that the statistical background of most undergrad econ students is not sufficient to the task.
So I think the more accurate criticism of economics is that it oversimplifies reality, losing too much important detail in the process. If agricultural labor costs rise, we would predict the price of eggs to rise; similarly an outbreak of bird flu would cause prices to rise (due to decreased supply) – unless consumers are put off and demand also decreases – then it depends on the shapes of the supply and demand curves. Missing from those stories is any explanation of how we move from one equilibrium to another or how long it takes. The question, then, is how important those questions are compared to the comparative statics. I’m not sure whether we would disagree about that or not. But we can probably agree that the subject is a lot more interesting if it includes such questions.
Dale, I’m in strong agreement with your characterizations. Of course the equilibrium stories give some insight. But questions like the price of eggs do interest me, and I think millions of others, and I think it’s no good to use the equilibrium story to address most of them.
You’re exactly right about the ex-post nature of it all “oh prices went up so either demand increased or supply decreased” is precisely the wrong direction that a theory should operate in. Seeing the outcome and hypothesizing that therefore the un-observable explanatory variable must have done precisely what your theory predicted it to do, is precisely what’s wrong with economics at least in the public sphere today imho (ie. the kind of economics that is discussed in news media or policy positions or in Federal Reserve press releases etc)
The value of a dynamic model is it can predict into the future, and therefore it has some kind of testability. Also, dynamic models predict the duration of things. Saying that “Eventually the sea will be flat” to paraphrase Keynes is not enough. A dynamic model makes testable predictions like “how long will the price of eggs stay above its previous price?” and “how much will the total number of eggs consumed per month decline?” and soforth.
Also, dynamic models need to have multiple causes involved, and then you can ask questions about what does the model say about the relative importance of the causes.
There are reasonable arguments to be made that the bird flu had virtually nothing to do with the price of eggs skyrocketing. The fraction of chickens that were killed was some really small percentage. I don’t remember the precise numbers but the kinds of numbers I remember were something like 20M chickens culled out of 1 Trillion chickens in the US, or something ridiculous like that. Even if you assume the 1 Trillion is chickens produced per year, and most of them are fryers and whatever, it’s probably on the order of 1% reduction in egg laying capacity or less, orders of magnitude changes that were very typical in the years prior to this recent bird flu. But the argument is what DID have a big effect is the vast consolidation of chicken production and the fact that it’s an oligopoly which was able to take advantage of consumer expectations regarding increased prices due to both widespread food inflation and media knowledge of the bird flu event.
I don’t want to say that was definitely the case, I just think it’s a very legitimate question and one that economists SHOULD care a lot about, and because the whole event was “transient” you really can’t address it in any decent way without a dynamic model that can say how things changed in time and how much those changes were driven by different factors.
to me, these questions like how are oligopoly and monopoly powers influencing pricing and rent extraction are THE questions economics needs to be addressing. Unfortunately everything appears to be backwards ad-hoc looking when it comes to questions like this: “prices went up, so it must be supply was constrained” kind of thing. Not “supply was unconstrained in any sense by comparison to recent history and yet prices went up… why?”
Also, my personal view of supply and demand curves is that they simply don’t exist. Every outcome in economics comes about because of people making choices to change the rate at which they consume various goods based on their knowledge of the prices today and their expectations of future prices and the future availability of goods. It’s entirely plausible for prices to go up and for people to *temporarily* purchase a lot more of a thing because they’re assuming it’s going to go up even more, etc. You can’t describe that with any static curve.
The way economic outcomes come about are much better thought of in my opinion in terms of systems dynamics: there are stocks of goods in warehouses or stores, there are flows of goods from producers to sellers to consumers, there are physical constraints on the rate at which goods can be produced and moved, and how much people need of certain goods etc. People set prices based on how much stock they have of stuff and how much flow they’re observing through the system… they increase prices when there’s relatively little of it available, they decrease prices when they have more than they’d like in their warehouses or stores. People buy stuff at a rate associated with their consumption desires. Their desire to consume changes with price but also with other factors.
In the eggs example, to the extent that birds were culled, there’s a biological constraint on how fast you can replace them, this gives a timescale over which supply should change. But it doesn’t really seem that the tiny perturbation to production of a couple percent could cause tripling of egg prices by itself. And did the price change correspond to the timeline for supply constraint? And did pandemic food relief efforts cause changes in willingness to pay for food specifically? Was overall food price inflation a result of pumping money specifically for food into consumer pockets (for example my family received thousands of dollars to pay for our kids to eat, in lieu of school lunches). Did “chickenized” producers push a narrative into the news media that the bird flu was a big deal precisely because they knew with widespread inflation in general they could get away with quite large price increases in the short term?
I’m with Keynes when he said “The long run is a misleading guide to current affairs. In the long run we are all dead. Economists set themselves too easy, too useless a task if in tempestuous seasons they can only tell us that when the storm is past the ocean is flat again.
John Maynard Keynes, A Tract on Monetary Reform (1923) Ch. 3
English economist (1883 – 1946)
“
Also my own experience and those of friends in the area was quite interesting when it came to the pandemic shutdowns and demand and supply of food. Of course everyone bought a lot of toilet paper etc, but the thing that was interesting was all of a sudden people bought a lot of producer type food goods… flour, rice, canned beans, frozen meat etc, the goal being to shift from eating “out” more to 100% eating in and without having to go shopping. In fact, between early March 2020 and early June 2020 my family didn’t buy ANY groceries. I had gone out and done 3 or so full sized shopping runs on about March 5th or so, buying like 50 pounds of flour and 25 pounds of black and pinto beans, 25 pounds of rice, and frozen meats and pasta and jam and peanut butter and frozen fruits and whatever. This was while doctor friends from my kids soccer team were still “poo pooing” the pandemic saying it wasn’t going to be a big deal and prices were totally undisturbed. (One of them went out of his way to apologize to me a few months later for having dismissed my concerns).
Anyway the point is, my family’s demand for food went to ZERO for like 8 or 9 weeks because of the stock we had on hand purchased at “normal” prices prior to pandemic craziness. When we did finally need to purchase food, again the kinds of food we were buying were very different from before, the prices were different, the types of goods demanded were different, etc. Without a dynamic theory that includes a lot of people “stocking up” and then running through their stocks, it’d be hard to describe food consumption in that time period. And let’s not forget the bald-faced rent-seeking of Tyson chicken or whoever it was that took out a full page ad in the NYT about “the food supply is collapsing” Or whatever. Dynamics are where all the interesting stuff is I think.
Daniel:
I find your criticism of equilibrium economics weak and misguided. Equilibrium is an appropriate assumption at some scale of observation or some relatively low level of resolution, and much of economics works fairly well with that assumption at the appropriate resolution of time and money. It’s hard to imagine you have a serious argument with that view
OTOH, of course, as the need or demand increases for higher resolution observations in ever smaller segments of the economy, equilibrium may not hold. Fine, but first you have to have data at that resolution then you have to work out the dynamics. However you’re not necessarily making progress because the smaller the segment of the economy you want to analyze the more variables will emerge and be relevant at that scale. So then you have to find a way to deal with those new variables some of which will be measurable and manageable, some of which will not. So now you’re going to wind up making new assumptions to rid yourself of many of those variables.
Eggs *are* interesting! Did I mention this already? Last week or so I paid $3.50 for one dozen at Albertsons. That surprised me because I had read prices were falling. Then just a few days later in WMt I found the double packs of 18 – so a total of *three* dozen eggs – for $4.50!!! Hilarious! Obviously there is a disequilibrium effect: Wal-Mart wants to have the lowest prices so it passes cost reductions immediately to consumers. Albertson’s is going to bilk $3.50 a dozen out of their customers until the customers stop paying it. At some point that difference will have to level off, presumably within a month or two – so thats the resolution of equilibrium pricing on eggs. Just the same, ABT’s egg prices were quite a bit higher than WMT before the pandemic, so not all of that is disequilibrium.
chipmunk, here’s the global price of copper since 1990. At what points was it in equilibrium? Was it in equilibrium at any point between 2002 and 2023?
https://fred.stlouisfed.org/series/PCOPPUSDM
If so, how do we know?
In the theory of dynamical systems, equilibrium is essentially a situation in which dy/dt is approximately zero. It doesn’t have to be zero exactly. But for example, you could hang a weight from a spring. If you lift it off the floor slowly it will come to an “equilibrium” distance from your hand. If you raise and lower your hand very slowly, the weight will rise and lower slowly and the distance between your hand and the weight will stay approximately constant. These are equilibrium/static approximations. If you move your hand quickly, the spring stretches, the weight lags, and the distance between your hand and the weight will oscillate and do all sorts of “dynamics”.
If you jerk your hand up and then leave it there, you’ll get an upward velocity of the weight, it will overshoot the equilibrium and then oscillate and come to a final position. The “equilibrium” position for the weight is determined like the theoretical econ equilibrium by the intersection of two curves, namely the spring force vs displacement curve, and the gravity force vs displacement curve (a constant).
The entire content of Equilibrium economics is essentially “although we don’t know absolutely anything from verifiable observations about the shape of the demand or supply curves and if they were to exist they are clearly time dependent, if we can get them to stay still long enough, the price will be at the intersection point of the two”.
the content of the physics equation is at least verifiable, you can pull the spring and measure the force.
In the econ story, the supply and demand curves play the role of angels dancing on the head of a pin. They are not observable, they certainly change in time, and we have no idea of the duration of time it would take for the actual price to reach the intersection in any given case. There is no information we could conceivably collect which would tell us their shape, or their dynamics, furthermore we know that in many cases the supplies and prices of various commodities are hardly constant for any time at all. Copper since 2002 for example.
If the supply/demand curve intersection tale is unverifiable (and it is) and can’t even begin to describe the trajectory of the price of copper for the last 20+ years, what is it other than a religious belief?
My own view (not a consensus by any means) is that supply and demand curves do exist and are useful constructs for many questions. What I don’t find helpful is the appeal to an underlying utility theory for consumer demand. Static analysis is limited, and at times misses the point, but dynamic analysis does not require rejecting the concepts of supply and demand. Given enough quality data, these curves can be identified though that is rarely the case. But they can still be useful conceptual tools. It is a mistake to use them ex post to rationalize what was observed, just as it is bad practice to choose your hypothesis after looking at the data.
The issue of market power is somewhat different. I agree that economists often overlook (or purposely ignore) issues of market power – this might reflect an ideological stance in the profession that business is good and government is bad. I happen to believe that market power has not been taken seriously enough by economists although the appeal to government regulation as a “solution” is quite naive. In the case of the recent price rises – for example, supermarket prices – I think the explanation of market power is a stretch. I will admit I have been puzzled by some of the price rises. Packaged goods, in particular, seem to have shot up quickly in price and I have a hard time finding a good reason. However, the idea that this has suddenly resulted from market power seems unconvincing. Did market power suddenly increase? Or did those with market power somehow not realize they could be charging much higher prices? I’ll admit that I don’t know – it is a mystery and one worth investigating and it isn’t clear whether the existing economic theories are adequate to the task.
Regarding copper prices, I actually won an international competition a few years ago for predicting copper prices. I based my analysis on an understanding of supply and demand, but my analysis was totally a time series analysis that had little to do with any economic theories. It was a statistical analysis and thoroughly dynamic. I’m not sure there is any other way to do it.
One other thought on equilibrium theories. I’ll use housing markets as an example. One approach is to assume that there is a long run equilibrium (when we’ll all be dead, as Keynes said), but that the convergence (if it even converges) is slow so that we are always out of equilibrium. The other extreme is to assume we are always in equilibrium and that supply and demand are constantly shifting to change that equilibrium. The latter view becomes less than useful when faced with real data – the data we have is not granular enough nor of good enough quality to model these constantly shifting curves. The former view has been implemented in dynamic models that allow for mean reversion, momentum, as well as shifting equilibrium prices. Those models fit the data much better and have a lot of intuitive appeal but are also somewhat limited by lack of good data (mostly for the last effect – the changes in underlying equilibrium prices). To some extent the choice between the extreme views is semantic, and to some extent it hinges on what data is available, But I think the only credible analysis of housing prices would require a dynamic model – the durability of housing assets and extended price adjustment times is an important reason. I’m not sure the price of eggs has features that require a dynamic model, however. The nature of the goods or services should be an important consideration when choosing whether an explicitly dynamic model is necessary.
Dale, again I agree with you. Of course the ideas of supply and demand have merit as heuristics. Prices go up, people buy less etc. Where I think the whole thing goes wrong is precisely the kind of thing you talk about with copper or housing prices. The supply and demand story if it doesn’t lead to near instant equilibrium is just a way of saying we pretend to know how things would go in other worlds (where the curves shift/change) we just don’t know what’s going on in this one.
When doing the dynamics, the existence of multiple participants ensures that as prices change for example go up, some people pull back on their purchasing rate before others do. This results in something sort of like a demand curve but it also is not a statically shaped curve. I may stop buying eggs because I’ve got 2 dozen at home and hope to ride through a brief uptick. I may substitute powdered egg for a while, I may have more potatoes and sausage for breakfast, I may bake fewer cakes… Etc but after a couple weeks of doing all that I may need to just go ahead and buy eggs again because my son’s birthday requires cake and they aren’t THAT expensive…
So it’s a bit like a chemical equilibrium, for every person suddenly needing a cake there’s a different person swearing off breakfast, you may get constant aggregate behavior for a while. But you might not! It’s not like atoms which always have their fixed properties. You may get more and more people shifting away from eggs through time, and more and more retailers doing shenanigans like Costco which I was told sent people to buy up eggs at local groceries and bring them back to Costco to resell. In essence all the action is right there at the margin, imagining there’s “a curve” at any given time is imagining an entire infinity of counterfactuals that you just happen to know and can predict accurately. All while being unable to successfully predict the actual price of eggs 1 week ahead.
The supply and demand curves are ok as a heuristic, but should be explained as such. As a real world modeling technique it’s much better to hypothesize a distribution of individual behaviors, and then talk about how that distribution is shifting in time because that distribution is potentially probe-able, you could survey consumers and ask them to keep their paper receipts for 3 weeks, and then argue you’ve captured verifiable information about that distribution. You could survey egg suppliers and ask them about their efforts to replace lost laying hens or the number of eggs they shipped each day on 14 consecutive days etc.
Those kinds of efforts are like taking the spring and stretching it to verify it’s properties, except the spring is an unstable plastic material, so you had better reverify every few minutes.
Without the ability to connect the microscale facts to a macroscale observed effective behavior, supply and demand curves are just-so-stories. It’s precisely the ability to aggregate from a survey of individual behavior to a prediction for the aggregate dynamics which would make economics a science proper.
In order to get there economics needs to take dynamics more seriously, to the extent it relies on equilibrium it needs to be because the dynamic theory underlying predicts specifically that the market should equilibriate quickly, just like the Newtonian physics could tell you when the spring is stiff like a bar of steel you can ignore it’s extension for ordinary motions.
When it comes to housing you have long term contracts, local laws on eviction and rent control, private equity firms buying up distressed properties, etc. The distribution of individual situations is a scientifically verifiable fact. Work from that towards a theory that describes the dynamics. If it works out to effectively a supply and demand curve with a relatively static shape and overall small shifts ok, but you and I both know it won’t. It could for something like pencils though.
Dale, another thought. The analogy with chemical equilibrium is really useful, but even chemical equilibrium isn’t so simple, I had a really great HS chem teacher, and he ran a demo experiment where you added some chemical to a beaker and got a deep blue color…. And everyone oohed and aahed appropriately… But then like 30 seconds later it went orange all of a sudden! More ooh and aah… But then 30 seconds later it went blue again! It did this for minutes maybe something like 10-20 even I don’t know.
Chemical reactions take femtoseconds to occur yet here is behavior that takes 30 minutes to equilibriate. That’s a scale ratio of 10^17 or something!
By similar analogy, an individual buying eggs at a store takes a minute to check out, but why couldn’t the equilibriation of price take months? And if it does take months for eggs, why couldn’t housing prices take 20 years? And policy changes all the time, and varies from location to location. Out of equilibrium would be the rule!
https://www.youtube.com/watch?v=SCoLMfplVWs
YouTube video about the reaction. And yes, it’s unusual to get an oscillating reaction in tabletop chemistry, but these kinds of interacting reactions are pretty common in biology, with expression of gene A causing expression of gene B and then expression of gene B causes expression of gene C and that causes A to turn off, which causes B to turn off etc…
> Chemical reactions take femtoseconds to occur yet here is behavior that takes 30 minutes to equilibriate. That’s a scale ratio of 10^17 or something!
If you wait 17 orders of magnitude longer other interesting things will surely happen. All models are wrong, etc.
Back when I used to teach intro econ, I always began with running experiments. The simple double sided auction was a good way to introduce supply, demand, and equilibrium. It almost always worked well- equilibrium was established relatively quickly, all gains from trade were realized, despite most participants not really understanding what they were doing. When I introduced changes that shifted supply and/or demand, the new equilibrium would also be discovered – but it was interesting to see that this happened quickly sometimes and slowly at other times. All this with the same number of participants. I had no explanation for the speed of convergence, but it was always of interest to discuss that (and the fact that the theory doesn’t say anything about how quickly equilibrium might e found).
One of my first jobs ever was to work on data feeding a stock market “market impact” model. It would estimate the size of the change in price caused by placing a buy or sell order of a certain size (up to institutional sized orders). It was reasonably good, one of the most important aspects of the model was the role of market makers, people who will buy your shares a little lower than what is clearing right now and then hold them long enough to unload them on people who want to buy hoping to get a little higher sale price. The role of market makers in general is I think underappreciated, and in markets like housing one reason for it’s illiquidity is there aren’t market makers really.
Another question is about consumables. Things like food and lumber etc. The question is not a one time exchange with gains from trade but rather rates of consumption. I don’t think it would be hard to run a game/experiment where a market was locked up because of people willing to hold goods rather than sell them and people willing to wait on projects rather than buy at the moment. Then people warehousing are hoping for ever higher prices to offset their warehousing costs and people waiting on projects get nervous that prices will go up even more… Bubble like feedbacks etc all of that will alter price behavior. Lumber prices looked like that I think.
The usual story in economics is equilibrium occurs **quickly** like anywhere from milliseconds to a day or two. I don’t think anyone but research level economists discuss the idea that some processes could take a month or two years , or 20 years to fully play out. In my view changes to the economy made in 1970s are still ongoing today. For example women participation in the labor force and education.
One thing non-economists generally don’t appreciate is how readily markets can find equilibrium and how that equilibrium promotes efficiency (exhausts the potential gains from trade) – when the conditions are right.
One thing economists generally don’t appreciate is how often the conditions are not right.
Dale, as you know, the reason I criticize economics is because I think it’s really important and should do better. Market economies are really wonderful things, when they work well. They work well when governments actively promote competition and an equal playing field, and don’t actively enable monopoly and regulatory capture and barriers to entry, when they actively perform anti-trust activities, and don’t pass laws preferencing first movers or setting prices and soforth. In those kinds of conditions equilibrium can be achieved easier, I still think not nearly as fast as Economists tend to think, but at least easier.
One thing that I think is under-appreciated is the role of distribution of information through the economy. An example might be the replacement of plywood with Oriented Strand Board. The OSB is stronger, and cheaper to manufacture. There are various forms, and some of them really do poorly when they get wet, but if it doesn’t get soaking wet it’s by far a better product in terms of strength and price. Still, it’s looked down on by many builders as “cheap” and low quality, largely because in the past they didn’t protect it from rain, and used the wrong nails, and installed it incorrectly and soforth. As information trickles through the economy demand changes. You might say that this is all compatible with equilibrium theory, that at any given time the price and supply of the product is in equilibrium and the changes through time are like the person moving their hand keeping the weight on the spring about equal-distant from their hand. But then, I think that is just begging the question, because the thing to be answered is “how does a product enter a market where it competes with a product people have a long time history of trusting the old product and there’s a high risk to installing the new product which is supposed to last for 50 years or more?” Like, if that’s not the question you’re asking, then you’re asking the wrong question. OSB has been in the market since it was invented in 1969. It certainly could have taken time to come to be well understood, but that should have happened by say 1989 in terms of manufacture and engineering knowledge, yet it’s still somewhat distrusted in 2023, commanding only 66% of the market where it probably should be closer to 80% (plywood still has applications for aesthetics etc). If OSB sales are trending continuously for 50 years while plywood declines then in some sense the whole time it’s not in equilibrium. A similar effect goes on for women in the workplace. Women born before say 1960 will have had a dramatically different experience throughout their lives in terms of how the workplace worked for them, compared to women born after say 1990. And the change between 1960 and 2020 is huge. And men are removing themselves from higher education now maybe in part because they’re finding that competing with women in that environment is not productive for them perhaps.
If you say that a golf ball is in equilibrium while it’s flying through the air because you’re a chemist and the structure of the polymer inside it is stable and unchanging on the scale of 0.1 second, you miss the entire point, which is that it’s flying through the air with forces on it from spin and then bouncing off the ground and rolling towards the pin, or ricocheting off a tree. The interesting part isn’t that the polymer is not changing, it’s that the game is being played!
@Andrew: I understand that your correspondent is not talking about government subsidies affecting wages, but about separate {wages -> (un)employment} and {government transfers -> consumption} effects. The latter could be related to the debates on the Inflation Reduction Act.
To your question whether transfers affect wages: There is probably an indirect effect. Government transfer -> less time worked by individuals -> reduced labour supply -> increased wages, ceteris paribus. The effect would be heterogeneous across income brackets. It would also be filtered through wage rigidities. Also, labour supply is quite inelastic, which leads me to believe that the effect would be rather small and probably empirically unobservable at the macro level.
On a completely different note.
Since you have often commented on NPR stories and Dan Ariely, here is a story where they collide:
https://www.npr.org/2023/07/27/1190568472/dan-ariely-francesca-gino-harvard-dishonesty-fabricated-data