Something you don't directly mention, but imply is training/documentation costs. People don't live forever, stay in the same job forever, and even if they do they forget things. So we often must make documents/procedures/aids for our future selves, new staff, reactivated staff, and of course our children.
In a very real sense this documentation/systemization is "capital" (personal, org, etc.) but unlike some physical capital, it requires a lot of attention to remain viable.
We see this in the costs of using automated devices like plastic injection molds or cnc milling machines - where it's quite common for setup costs to dominate until quite a large number of parts have been made. These setup costs are part specific - they aren't really "capital" - they're fixed startup overhead.
There's a long history in railroad economics with lengthy literature concerning price discrimination to deal with the problem of overhead. In the 19th century it was often called "value of service pricing." Moreover, there's also a classic article about this problem by one of Keynes's first students, Frank Ramsey: "A Contribution to the Theory of Taxation," Economic Journal (1927) 37, pp. 47-61.
And they use price discrimination so as to maximize revenue--except for contract rates which are more nearly negotiated economic equal to economic equal, Cargill to Union Pacific.
Milton Friedman would agree with your comments, but with an important caveat. Businesses and consumers don't literally make decisions at the margin. But participants in competitive market situations behave *as if* they make decisions at the margin. The tools of marginal analysis help to explain how markets arrive at equilibria nonetheless. The HR director at a large corporation doesn't need to make decisions at the margin. Accountants command a certain level of compensation because of all the millions of decisions that accountants and their employees make over time and employers are willing to pay that level of salary after they figure out the work that the last accountant hired does is necessary to the company's success. At the margin, the (public) company that hires too many accountants will be less profitable, and investors (at the margin!) will be less willing to invest additional capital in such companies, etc., etc. It's as von Hayek suggested, the state of the economy is the result of millions of decisions made by millions of people (even if decisions are made by simple rules). One can still describe the state of the economy using marginal analysis if no one agent uses marginal rules for decision-making.
Thank you for this lucid piece. Something to add - it seems to me anyway - is the Bureaucracy Factor. All large businesses become bureaucracies and those bureaucracies develop a life of their own and one that does not necessarily align with the best commercial interests of the business. An example is the commissioning of advertising. My father, who worked in advertising all his life, used to tell me that advertising often didn't have any commercial benefit; it worked to satisfy the advertising department - to make them feel useful, sophisticated, trendy , to provide contract handouts to their mates (in their social circle) etc etc. I think this principle (if widened out from my specific example) goes a long way to explaining the apparent paradox of huge corporations embracing 'Woke' - often with very negative consequences for their bottom line.
Agreed. One of the constant themes of working in industry I have seen is that the people who make the products/services sold are generally underpaid and hard to retain while those who do everything else (HR, various management bits) tend to proliferate. The root thought seems to be that makers are a cost to be minimized but people like us thought workers are a benefit to be pursued. Of course not all companies pull this, but as they get bigger it seems more constant. You see it in universities too these days, with 1:1 administrator to student ratios.
In Sweden we have minimized this. We see heirarchies as bad-in-themselves. We have the flattest corporate structure in the world, and we like this, a lot. We want the people on the top to be as close to the bottom, where the knowledge of "how the thing gets done, really" lives. So, you stuff the safety officer (you have to have one) as a very much part time job of an existing worker, who has a real job. Ditto with the 'environmental impact' officer.
All of these people get a half day a week off work to write yaps at the government, more if you think that more yapping is needed. These things tend to be spread out, if we scheduled them all for 'Friday after lunch' -- people would just go home early thinking that the government got too many yaps as it is to really need their input more than they needed 3 hours of unscheduled time.
That’s a damned good system. I have noticed many of the better run companies I have seen have very flat structures, while the worst have a conglomerate, mishmash of different businesses with an overhead machine... er management structure on top.
Sorry, I couldn't answer right away and then forgot your question...
I didn't really think through what I said and maybe it's not exactly chicken and egg. What I was thinking is instead of better run companies choosing a flat structure, the only companies that can choose it and succeed are the better run ones.
Like all things with deep historical roots, it is hard to point to any one thing. Or even a few things. In the 17th century we were blessed with a truly great Chancellor -- Axel Oxenstierna. He organised the Swedish military bureaucracy (though it is unclear if Max Weber would call it a bureucracy) to the point where Sweden could become and was a World Military Power. One of the reasons why Oxenstierna's institutions were so effective was that they weren't a matter of the nobility running the show and basically only interacting with each other. Competence was rewarded, and Oxenstierna was in a position to see that the more capable people succeeded despite the howls of protest from the more well connected.
Another facet is that Sweden never had the sort of feudalism you saw in England or France, with lords and vassals. The farmers had their own house in the parliament and successive kings greatly played the farmers off against the nobles to preserve royal power and prerogative. The way that the army was organised was part of this plan. Each piece of land in Sweden was listed and expected to produce a certain number of soldiers. The farrmers, and not some local lord, were directly responsible for producing the soldier. With the rank came a plot of land. If the soldier was not needed for war, he'd work the land same as any other farmer. If he was at war, his neighbours would see that his wife and family and the host of hired hands were enough to get the crops in and keep the farm well maintained. Often the son of a deceased soldier-farmer would inherit the role (and thus the farm) but the farm went with the role. Should the farmers decide to promote somebody else to the role, the new person got the farm.
In more recent times, the military threat has been from the USSR (and now Russia). The defence policy has always assumed that the 'chain of command' would be destroyed early in any war, thus Sweden had to be able to run a guerilla war with autonomous local regions acting independently. Part of the once compulsory -- then wasn't for a generation -- now back again male military service was to train you for just that eventuality.
All of this means that there are almost no servile people in Sweden, -- or at least it it is expressed entirely differently than other places -- and that 'sucking up' is vastly less successful as a strategy for success than most other places. There is little sense that a 'bigger title' means that more people will defer to you -- hardly anybody is going to defer to you for any other reason than they think you are competent in the first place. Everybody from the CEO on down can expect pushback whenever he or she does something that the other people in the room think is unwise.
This has left out the non-military historical effects, which must be significant, but the problem with asking an amateur military historian a question about Swedish history is that you get a military history answer. You'll need a commercial historian to balance this approach but I am not sure I can round one up on short notice.
Sub-Dunbar hierarchy is inevitable for effective team efforts, but bigger companies need more hierarchy, and more bureaucracy. Most orgs then get too much of org oriented bureaucrats becoming dominant in the decision making hierarchy. I’m more interested in Sweden in general, but also IKEA in particular on hierarchy.
We get more bureaucracy in bigger companies, here too, but I am more familiar with how Ericsson and Volvo and the tech companies do things. I will ask around about how Ikea functions and report back if I learn anything interesting.
Volvo -- at least before it was bought -- had problems of the org oriented bureaucrats building their own little fiefdoms, and the drastic step was taken to entirely reorganise all the middle managers every, I forget now, but I think 7 years.
Everybody is managing a completely different group doing different things from what they were managing last week. This was hard on the managers who had expertise in certain areas, and who naturally wanted to work there, but overall the effect was very positive in terms of increased productivity. Goes to show how much nest-feathering and I-shall-passive-agressively-make-it-harder-for-groups-lead-by-rivals-to-succeed was going on.
Mostly Sweden has a problem of not enough medium sized companies. How to make the jump from a company with 300 people to one that has 3000 people is not well enough understood here, with the consequence that it far more common to fork your company into 2 companies when the employee count hits 500 or so.
We seem to keep finding ourselves commenting on the same Stacks. Yours and mine have different themes... but not so different. Would you be interested in doing a free Sub to each others?
The quart of milk may cost more per ounce because much of the cost is in the packaging and transport rather than the quantity of the product. So getting a quart of milk to the store shelf may cost nearly as much as the half gallon or full gallon, allowing the seller to make an acceptable profit while selling the milk in the larger package at a lower price per ounce.
Coincidentally, I stopped reading the blog entitled "Marginal Revolution", with links that are curated by Tyler Cohen and Alex Tabarrok, in favor of this substack. This entire post has a double-meaning
1 using very round numbers, I think Apple decided long ago that in order to double their market share, they'd have to cut their unit profit more than in half. I'm not sure where that is captured in what you say.
2 I think you are mostly right about the marginal unit cost being near zero in many cases (including railroads of over 150 years ago) but I'd say your examples leave a lot to be desired.
- milk: What you say makes zero sense. The store with the very lowest prices still charges way more per ounce for a quart than gallon. It has nothing to do with shoppers for the gallon being more price discriminate.
- airplane seat (and student). While airlines charge differing amounts to various passengers, the cost isn't per seat. It is per plane. And even on that plane there is a cost of adding a passenger if the plane is full enough that it could become oversold. And lately that has been a real risk as most flights are very full, even if in recent years airlines have gotten better at avoiding oversold situations.
To say this all differently, if a seller produces thousands or millions of units and they might add a handful of units here or there at virtually no cost, that really doesn't tell us their marginal cost is near zero in what is a meaningful way.
Telecommunications regulators in Canada have been practicing Ramsey pricing (incremental cost plus a markup to cover fixed costs that reflects inverse price elasticities) for some thirty-five years now. Granted, they seldom have precise price elasticities, but the principle is there, e.g. higher markups to business customers than to residential customers). I note in passing that, for traffic-based services, the marginal cost is zero: you can always carry an extra gigabit of traffic without any capital outlay or operating expense. The proper measure is incremental cost, i.e. the extra cost of the volume of traffic that you are considering in your decision. The cost will vary according to the decision you are contemplating. (This originates in James Buchanan's Costs and Choice, an often overlooked work.)
The marginal cost is zero for network traffic, until you get congestion and then suddenly it isn't and you would need to lay a new cable to carry it. I'm not sure that model is a very good one.
Exactly. This is the mistake know-nothings make about solar. They look at the kwh price for solar panels and think that means we should convert 100% to solar. But generation is less than 10% of the cost of electricity: most of the cost is overhead (operation and maintenance of transmission and distribution, billing, safety).
If you tried to imagine an optimal economy, it would be mainly overhead. I'm thinking of the machine - SPOILER WARNING - in Forbidden Planet, an old science fiction movie based on The Tempest. It was an immense machine, self maintaining, self improving, that could provide anything anyone wanted by merely thinking about it. That expensive machine meant zero marginal cost. - MORE SPOILER - How expensive was it? It's makers the Krell were gone. A spaceship was reported lost there. As the movie begins, a rescue patrol ship has arrived to investigate.
It's a surprisingly good movie for its era, both good science fiction and a fun take on Shakespeare. It also has an economics lesson for which I'll thank this blog.
> In all such cases, if the seller were to follow the marginalist principle, its price would not be high enough to cover overhead. It would lose money and go out of business.
That also happens a lot. Hence why so many of these industries are either monopolies or near monopolies.
"In all such cases, if the seller were to follow the marginalist principle, its price would not be high enough to cover overhead. It would lose money and go out of business."
Economic theory suggests that a firm would never choose to set price equal to marginal cost. Either they have no choice over the price (price takers), or they choose to set price above marginal cost--above the point where MR=MC.
Economic models already fully account for "overhead", not sure what's wrong with these models.
As for milk, it's less than 4 times as expensive to retail a gallon of milk as a quart of milk, so the per ounce price would be lower even without price discrimination.
Yes, that wording was careless. A better sentence would have said that price has to be far above marginal cost, or else the firm will lose money or go out of business.
The reason that the simple "MR=MC" rule does not help in the real world where MC is close to zero is that there are so many strategic considerations: price discrimination, as I mentioned; entry deterrence; network and lock-in effects. etc.
Great insight on economics, marginal analysis, and overhead.
A huge amount of overhead is to help with management decision making. What is the real status, how does it compare to our expectations, what can we do to improve our (team’s) numbers? Improvement involves lots of uncertainty about alternative decisions, with unknown trade offs between costs & benefits. I hope AI helps top management make better decisions, and think lots of overhead labor is at risk of becoming redundant, much lower value.
Lousy example of milk, tho, because efficiency of scale explains it better; Walmart lower prices are based on this. Why does 7/11 milk cost more than Walmart? Because of convenience, the total price of everything bought includes time, locations to travel, and storage costs at home.
Price discrimination does explain real business decisions better than marginal prices, but comparison to competitors is also important. For small companies, most important. Thus very few 7/11s open up next to Walmart, tho Walmart might open up near an older 7/11.
Alternatives, opportunity costs, dominate most decisions.
Arnold
I’m a small business owner. Four employees. Service not manufacturing.
And your right, my cost of next customer is minimal. Cost of first customer is huge.
After forty years, I’ve noticed that most cost is now paying for - to ideas .
Like, insurance, car, liability, employee, social security.
Accounting. Advertising.
And taxes.
Minimal costs for supplies, truck, tools.
Interesting
Thanks
Clay
Great post. I was up at 4 am this morning, so got carried away with my response, looking at this idea in the context of the semiconductor industry, in case anyone’s interested in such things: https://nothingbadsofar.substack.com/p/overhead-revolution-semiconductor?sd=pf
Something you don't directly mention, but imply is training/documentation costs. People don't live forever, stay in the same job forever, and even if they do they forget things. So we often must make documents/procedures/aids for our future selves, new staff, reactivated staff, and of course our children.
In a very real sense this documentation/systemization is "capital" (personal, org, etc.) but unlike some physical capital, it requires a lot of attention to remain viable.
We see this in the costs of using automated devices like plastic injection molds or cnc milling machines - where it's quite common for setup costs to dominate until quite a large number of parts have been made. These setup costs are part specific - they aren't really "capital" - they're fixed startup overhead.
There's a long history in railroad economics with lengthy literature concerning price discrimination to deal with the problem of overhead. In the 19th century it was often called "value of service pricing." Moreover, there's also a classic article about this problem by one of Keynes's first students, Frank Ramsey: "A Contribution to the Theory of Taxation," Economic Journal (1927) 37, pp. 47-61.
And they use price discrimination so as to maximize revenue--except for contract rates which are more nearly negotiated economic equal to economic equal, Cargill to Union Pacific.
Milton Friedman would agree with your comments, but with an important caveat. Businesses and consumers don't literally make decisions at the margin. But participants in competitive market situations behave *as if* they make decisions at the margin. The tools of marginal analysis help to explain how markets arrive at equilibria nonetheless. The HR director at a large corporation doesn't need to make decisions at the margin. Accountants command a certain level of compensation because of all the millions of decisions that accountants and their employees make over time and employers are willing to pay that level of salary after they figure out the work that the last accountant hired does is necessary to the company's success. At the margin, the (public) company that hires too many accountants will be less profitable, and investors (at the margin!) will be less willing to invest additional capital in such companies, etc., etc. It's as von Hayek suggested, the state of the economy is the result of millions of decisions made by millions of people (even if decisions are made by simple rules). One can still describe the state of the economy using marginal analysis if no one agent uses marginal rules for decision-making.
Thank you for this lucid piece. Something to add - it seems to me anyway - is the Bureaucracy Factor. All large businesses become bureaucracies and those bureaucracies develop a life of their own and one that does not necessarily align with the best commercial interests of the business. An example is the commissioning of advertising. My father, who worked in advertising all his life, used to tell me that advertising often didn't have any commercial benefit; it worked to satisfy the advertising department - to make them feel useful, sophisticated, trendy , to provide contract handouts to their mates (in their social circle) etc etc. I think this principle (if widened out from my specific example) goes a long way to explaining the apparent paradox of huge corporations embracing 'Woke' - often with very negative consequences for their bottom line.
Agreed. One of the constant themes of working in industry I have seen is that the people who make the products/services sold are generally underpaid and hard to retain while those who do everything else (HR, various management bits) tend to proliferate. The root thought seems to be that makers are a cost to be minimized but people like us thought workers are a benefit to be pursued. Of course not all companies pull this, but as they get bigger it seems more constant. You see it in universities too these days, with 1:1 administrator to student ratios.
In Sweden we have minimized this. We see heirarchies as bad-in-themselves. We have the flattest corporate structure in the world, and we like this, a lot. We want the people on the top to be as close to the bottom, where the knowledge of "how the thing gets done, really" lives. So, you stuff the safety officer (you have to have one) as a very much part time job of an existing worker, who has a real job. Ditto with the 'environmental impact' officer.
All of these people get a half day a week off work to write yaps at the government, more if you think that more yapping is needed. These things tend to be spread out, if we scheduled them all for 'Friday after lunch' -- people would just go home early thinking that the government got too many yaps as it is to really need their input more than they needed 3 hours of unscheduled time.
That’s a damned good system. I have noticed many of the better run companies I have seen have very flat structures, while the worst have a conglomerate, mishmash of different businesses with an overhead machine... er management structure on top.
I think leaves a chicken and egg question.
How do you mean?
Sorry, I couldn't answer right away and then forgot your question...
I didn't really think through what I said and maybe it's not exactly chicken and egg. What I was thinking is instead of better run companies choosing a flat structure, the only companies that can choose it and succeed are the better run ones.
How did Sweden acquire the flattest corporate structure in the world?
Like all things with deep historical roots, it is hard to point to any one thing. Or even a few things. In the 17th century we were blessed with a truly great Chancellor -- Axel Oxenstierna. He organised the Swedish military bureaucracy (though it is unclear if Max Weber would call it a bureucracy) to the point where Sweden could become and was a World Military Power. One of the reasons why Oxenstierna's institutions were so effective was that they weren't a matter of the nobility running the show and basically only interacting with each other. Competence was rewarded, and Oxenstierna was in a position to see that the more capable people succeeded despite the howls of protest from the more well connected.
Another facet is that Sweden never had the sort of feudalism you saw in England or France, with lords and vassals. The farmers had their own house in the parliament and successive kings greatly played the farmers off against the nobles to preserve royal power and prerogative. The way that the army was organised was part of this plan. Each piece of land in Sweden was listed and expected to produce a certain number of soldiers. The farrmers, and not some local lord, were directly responsible for producing the soldier. With the rank came a plot of land. If the soldier was not needed for war, he'd work the land same as any other farmer. If he was at war, his neighbours would see that his wife and family and the host of hired hands were enough to get the crops in and keep the farm well maintained. Often the son of a deceased soldier-farmer would inherit the role (and thus the farm) but the farm went with the role. Should the farmers decide to promote somebody else to the role, the new person got the farm.
In more recent times, the military threat has been from the USSR (and now Russia). The defence policy has always assumed that the 'chain of command' would be destroyed early in any war, thus Sweden had to be able to run a guerilla war with autonomous local regions acting independently. Part of the once compulsory -- then wasn't for a generation -- now back again male military service was to train you for just that eventuality.
All of this means that there are almost no servile people in Sweden, -- or at least it it is expressed entirely differently than other places -- and that 'sucking up' is vastly less successful as a strategy for success than most other places. There is little sense that a 'bigger title' means that more people will defer to you -- hardly anybody is going to defer to you for any other reason than they think you are competent in the first place. Everybody from the CEO on down can expect pushback whenever he or she does something that the other people in the room think is unwise.
This has left out the non-military historical effects, which must be significant, but the problem with asking an amateur military historian a question about Swedish history is that you get a military history answer. You'll need a commercial historian to balance this approach but I am not sure I can round one up on short notice.
Sub-Dunbar hierarchy is inevitable for effective team efforts, but bigger companies need more hierarchy, and more bureaucracy. Most orgs then get too much of org oriented bureaucrats becoming dominant in the decision making hierarchy. I’m more interested in Sweden in general, but also IKEA in particular on hierarchy.
We get more bureaucracy in bigger companies, here too, but I am more familiar with how Ericsson and Volvo and the tech companies do things. I will ask around about how Ikea functions and report back if I learn anything interesting.
Volvo -- at least before it was bought -- had problems of the org oriented bureaucrats building their own little fiefdoms, and the drastic step was taken to entirely reorganise all the middle managers every, I forget now, but I think 7 years.
Everybody is managing a completely different group doing different things from what they were managing last week. This was hard on the managers who had expertise in certain areas, and who naturally wanted to work there, but overall the effect was very positive in terms of increased productivity. Goes to show how much nest-feathering and I-shall-passive-agressively-make-it-harder-for-groups-lead-by-rivals-to-succeed was going on.
Mostly Sweden has a problem of not enough medium sized companies. How to make the jump from a company with 300 people to one that has 3000 people is not well enough understood here, with the consequence that it far more common to fork your company into 2 companies when the employee count hits 500 or so.
We seem to keep finding ourselves commenting on the same Stacks. Yours and mine have different themes... but not so different. Would you be interested in doing a free Sub to each others?
Sure! I just subbed your way, and there is no paywall on mine.
The quart of milk may cost more per ounce because much of the cost is in the packaging and transport rather than the quantity of the product. So getting a quart of milk to the store shelf may cost nearly as much as the half gallon or full gallon, allowing the seller to make an acceptable profit while selling the milk in the larger package at a lower price per ounce.
Coincidentally, I stopped reading the blog entitled "Marginal Revolution", with links that are curated by Tyler Cohen and Alex Tabarrok, in favor of this substack. This entire post has a double-meaning
I always read Arnold, seldom MR.
1 using very round numbers, I think Apple decided long ago that in order to double their market share, they'd have to cut their unit profit more than in half. I'm not sure where that is captured in what you say.
2 I think you are mostly right about the marginal unit cost being near zero in many cases (including railroads of over 150 years ago) but I'd say your examples leave a lot to be desired.
- milk: What you say makes zero sense. The store with the very lowest prices still charges way more per ounce for a quart than gallon. It has nothing to do with shoppers for the gallon being more price discriminate.
- airplane seat (and student). While airlines charge differing amounts to various passengers, the cost isn't per seat. It is per plane. And even on that plane there is a cost of adding a passenger if the plane is full enough that it could become oversold. And lately that has been a real risk as most flights are very full, even if in recent years airlines have gotten better at avoiding oversold situations.
To say this all differently, if a seller produces thousands or millions of units and they might add a handful of units here or there at virtually no cost, that really doesn't tell us their marginal cost is near zero in what is a meaningful way.
If Apple charges less, they will be seen as less of a luxury brand as well.
Telecommunications regulators in Canada have been practicing Ramsey pricing (incremental cost plus a markup to cover fixed costs that reflects inverse price elasticities) for some thirty-five years now. Granted, they seldom have precise price elasticities, but the principle is there, e.g. higher markups to business customers than to residential customers). I note in passing that, for traffic-based services, the marginal cost is zero: you can always carry an extra gigabit of traffic without any capital outlay or operating expense. The proper measure is incremental cost, i.e. the extra cost of the volume of traffic that you are considering in your decision. The cost will vary according to the decision you are contemplating. (This originates in James Buchanan's Costs and Choice, an often overlooked work.)
The marginal cost is zero for network traffic, until you get congestion and then suddenly it isn't and you would need to lay a new cable to carry it. I'm not sure that model is a very good one.
It would be interesting to see your list of revisionist economic ideas.
Exactly. This is the mistake know-nothings make about solar. They look at the kwh price for solar panels and think that means we should convert 100% to solar. But generation is less than 10% of the cost of electricity: most of the cost is overhead (operation and maintenance of transmission and distribution, billing, safety).
> But generation is less than 10% of the cost of electricity
This is the key fact that keeps getting left out of discussions about energy policy.
If you tried to imagine an optimal economy, it would be mainly overhead. I'm thinking of the machine - SPOILER WARNING - in Forbidden Planet, an old science fiction movie based on The Tempest. It was an immense machine, self maintaining, self improving, that could provide anything anyone wanted by merely thinking about it. That expensive machine meant zero marginal cost. - MORE SPOILER - How expensive was it? It's makers the Krell were gone. A spaceship was reported lost there. As the movie begins, a rescue patrol ship has arrived to investigate.
It's a surprisingly good movie for its era, both good science fiction and a fun take on Shakespeare. It also has an economics lesson for which I'll thank this blog.
> In all such cases, if the seller were to follow the marginalist principle, its price would not be high enough to cover overhead. It would lose money and go out of business.
That also happens a lot. Hence why so many of these industries are either monopolies or near monopolies.
I'm not sure I know what this means:
"In all such cases, if the seller were to follow the marginalist principle, its price would not be high enough to cover overhead. It would lose money and go out of business."
Economic theory suggests that a firm would never choose to set price equal to marginal cost. Either they have no choice over the price (price takers), or they choose to set price above marginal cost--above the point where MR=MC.
Economic models already fully account for "overhead", not sure what's wrong with these models.
As for milk, it's less than 4 times as expensive to retail a gallon of milk as a quart of milk, so the per ounce price would be lower even without price discrimination.
Yes, that wording was careless. A better sentence would have said that price has to be far above marginal cost, or else the firm will lose money or go out of business.
The reason that the simple "MR=MC" rule does not help in the real world where MC is close to zero is that there are so many strategic considerations: price discrimination, as I mentioned; entry deterrence; network and lock-in effects. etc.
Great insight on economics, marginal analysis, and overhead.
A huge amount of overhead is to help with management decision making. What is the real status, how does it compare to our expectations, what can we do to improve our (team’s) numbers? Improvement involves lots of uncertainty about alternative decisions, with unknown trade offs between costs & benefits. I hope AI helps top management make better decisions, and think lots of overhead labor is at risk of becoming redundant, much lower value.
Lousy example of milk, tho, because efficiency of scale explains it better; Walmart lower prices are based on this. Why does 7/11 milk cost more than Walmart? Because of convenience, the total price of everything bought includes time, locations to travel, and storage costs at home.
Price discrimination does explain real business decisions better than marginal prices, but comparison to competitors is also important. For small companies, most important. Thus very few 7/11s open up next to Walmart, tho Walmart might open up near an older 7/11.
Alternatives, opportunity costs, dominate most decisions.