Hacker News Reader

Comment by Kranar | Hacker News Reader

The fallacy is you stating something that's just plain wrong. The basic law of economics is that price is determined by BOTH supply and demand, whereas your statement is that price is determined only by demand.

The cost to manufacture a product is one of the key components that determines the supply of a product; the higher the cost to produce, the lower the supply and hence assuming demand doesn't change then the price will increase. Of course demand may not stay the same, could increase or decrease.

View on HN

> The basic law of economics

It's not a law. It's a model. A very simple model. Of a very complex system.

Computer Scientists never invoke Turing machines or the simply typed lambda calculus in debates about specific pieces of hardware or compiler implementations, except in cases where those models are actually applicable and in ways that comport with the pieces of those systems for which the model is usefully faithful.

The Econ 101 "supply & demand curve" is way simpler than TMs/lambda calculi, AND it's modeling a much more complex system than a computer/compiler. Yet, somehow, the model is constantly invoked.

There's a really good reason that every single financial analyst in the world goes directly to "pricing power" whenever they are asked to analyze the effect of "supply chains".

GGP said the price is only determined by how much people are willing to pay. This is obviously false. If the cost of producing a product rises above the current price, even if the price people are willing to pay doesn't change, the price has to increase or else the business will start losing money.

And, sometimes the business starts losing money.

We saw that when oil prices went below many producers' extraction cost. They continued selling at the lower price, because they had to service existing debt, and accumulated more debt until the price went back up.

This is also known as the "Groupon Business Model of Imminent Failure". Cash today is way more important than accounts receivable.

Also sometimes supply or production is inelastic and not possible to scale down and up. Thus to have chance to generate money in future they had to accept loss now.

My understanding with respect to the oil industry is that they have relatively high capital costs, and lower marginal costs. If prices drop so that they become unprofitable, it still makes sense to drill because the fixed costs are already sunk. It could also make sense if you expect the low prices to be a temporary situation, and shutting down and restarting would be expensive. Either way, producing at a loss is not sustainable in the long run, so without costs falling or prices increasing, they would eventually go out of business.

This isn't really a factor for Arizona, since as a brand, not a commodity, they can set their price, and since they're cheaper than many of their competitors, they have room to work with here, too. If their unit cost is higher than their unit price, they don't expect costs to decrease, and they don't intend to raise their price, they would simply make more money by not making the product.

They will never raise the price higher than people are willing to pay or else they won't sell the product at all. Sometimes they will sell at a loss temporarily and sometimes they will just sell off all remaining product at the highest amount people are willing to pay. Even if at a loss because getting something is better than the product sitting on shelves. Some products are sold at a loss permanently as a loss leader. Like hot dogs or roast chicken at Costco.

> They will never raise the price higher than people are willing to pay or else they won't sell the product at all

That isn't true, at least for some definition of "people". They will raise the prices until selling is profitable even if most people aren't willing to pay that, since there will be some who want the product more who pay the premium. If it was profitable to sell for the lower price they probably would make more money thanks to higher volume though, which is why they had the price originally.

Supply vs demand isn't always right, but it is closer to right than wrong in most cases. You can look at profit margins to see how well supply/demand works in a specific field, in software supply/demand doesn't exist no. But in most fields profit margins doesn't look like that, they are sub 10% meaning that for every 10 dollars you pay less than 1 dollar goes as profits and the rest goes to pay for the service you bought, that wouldn't be true if prices usually followed what people are willing to pay instead of being a mix of supply/demand.

Different people have different prices they're willing to pay for a product. That's the theory behind a demand curve. And in reality, if they raise the price to $1.29 or $1.49, there will still be people buying it. If the unit cost of Arizona tea ever went above the unit price, it simply would not make sense to continue manufacturing the product and selling it at that price, outside of contractual obligations.

Loss leaders are a different analysis, and in order for that to work, Arizona would have to convince all retailers to sell the product at a loss; Costco's business is not selling hot dogs or rotisserie chicken, and that's why they can afford a loss on those. Even these, they still put in a lot of work to keep costs low (e.g., making the hot dogs themselves instead of buying from Hebrew National), so the loss in recent years has been either small, or they break even. But obviously even this isn't something that could be kept up forever, given long run trends in inflation; the purchasing power of a dollar was ten times greater in 1950, and if in 2090 they keep selling hot dogs for $1.50 when a comparable meal costs $15, it might make sense to have a Costco membership just for hot dogs and rotisserie chickens (unless they set the membership cost so high that it's unrealistic to recoup the cost in a year without, e.g., eating there every single day), and at that point their loss leader just becomes a loss.

One final note I'll make is that average hourly earnings over the last couple decades have outpaced the rise in prices for most goods, even as the price of services outpace average hourly earnings; intuitively, the former should make sense, since with improved technology, the same good requires less labour to make. So food has gotten cheaper in real terms over the last few decades, as has shipping, while the opposite effect happens with services because of Baumol's cost disease. Because of this, even with inflation, more efficient products has somewhat blunted the impact on these manufacturers' costs, even though they too will one day have to increase their prices as well.

Laws in the sciences are all components of models. Newton's three laws are a component of Newtonian mechanics, so are the laws of thermodynamics. Arguing semantics about what a law is is the least interesting aspect of this discussion, with that said you are free to use whatever term you want but should at least be familiar that in general the term law of supply, law of demand, and that much of economics uses the term law to codify a set of common observations that are used as the basis for further derivations:

[1] https://en.wikipedia.org/wiki/Law_of_demand

[2] https://en.wikipedia.org/wiki/Law_of_supply

[3] https://en.wikipedia.org/wiki/Category:Economics_laws

The important difference I'm commenting on is that engineers and physicists rarely get confused about when Newtonian mechanics or thermodynamics are useful models. Even when they do, their "belief" in these laws is not taken as religious or political. Beliefs in models are quickly dispelled when it becomes obvious that the model is too wrong to be useful.

I rarely see economists invoke Econ 101 models such as [1],[2] except in extraordinarily constrained settings.

If you aren't invoking theorems about beta reduction for the simply typed lambda calculus to reason about C's memory model, then consider also refraining from using [1],[2] to analyze pretty much anything in a real economic system.

This isn't pedantic. This is a real difference in how engineers and software people treat laws in the natural sciences vs how laws get treated in economics by the same set.

You can call it a law, or a model, or whatever. The point -- and, not a pedantic one -- is that [1],[2] are almost never accurate enough to be useful and calling them laws doesn't change that fact.

I'm no expert myself, but my layman's understanding of some of the pushback behavioral economics has for the field of economics is relevant for your point.

Specifically, from what I've seen, behavioral economics precisely wishes to reject predicting behavior via these simple economic models in favor of more nuanced understandings of complex behaviors, derivable only through empirical observation of what consumers really do, not what they should do or a simplistic law predicts they should do.

(It's an interesting comparison to CS too.)

> (It's an interesting comparison to CS too.)

Particularly since the first behavioral economist to get a Nobel was also an early pioneer in many fields which are now housed in CS departments: programming languages, artificial intelligence, theorem proving.

> the first behavioral economist to get a Nobel

Possibly referencing https://en.wikipedia.org/wiki/Herbert_A._Simon

Thanks for your input in this thread!

You're not wrong, but GP's error is much more object level than that. The neoclassical paradigm has flaws; claiming that all markets are perfectly competitive is not one of them. Sweet tea is a textbook case of monopolistic competition, and short run inelasticity is to be expected.

Supply & Demand does not require perfectly competitive markets.

I've seen a lot of people theorize that the law of Supply+Demand price setting doesn't apply, or that it can be repealed by the legislature.

They're always wrong.

Feel free to prove me wrong with a case history.

> I've seen a lot of people theorize that the law of Supply+Demand price setting doesn't apply, or that it can be repealed by the legislature.

What constrains the supply of copies of a piece of software?

> what, exactly, constrains the supply of copies of a piece of software?

Nothing. Nearly all the software I use is free.

The answer is the legislature, via copyright and patent law. You know this, you know the point I'm making. Your choice to be coy and unserious makes me wonder if you even sincerely hold the position you're arguing.

I'm very serious. My Linux system is my main dev system. I didn't pay a dime for any of the software on it. I write and distribute, for free, software. It's all copyrighted and legal. You can distribute it, too, at zero cost. You don't even have to ask my permission. Feel free!

How about you being more specific?

You asked for an example of the legislature writing the laws of supply and demand, on the forum at least in large part about the software industry. An industry whose primary product has zero marginal cost but is none-the-yet scarce due to the acts of legislative bodies.

Supply, demand, and price can all be manipulated, including in ways that divorce one from the other two. Either by governments or by sufficiently large private actors. This... isn't controversial, even among the most staunch of neo-classical economists.

> How about you being more specific?

Believe it or not, there exists non-free software. So, to be more specific, I am referring to literally any machine that is not running a completely FOSS stack. Or even any machine that IS running an entirely FOSS stack but is being used in commerce and where some parts of that software stack are patent-protected!

I.e., basically every machine used for commerce and the vast majority of consumer machines. 99% of the global install base? More?

Again, this whole thread is an excellent demonstration of my original post. This is the Econ equivalent of getting into an argument with a political scientist over whether NVIDIA's latest chip is really just a <insert idealized model of computation>. These arguments don't even happen, except in econ with non-economists, because nobody has strong ideological attachments to Newtonian mechanics or the number of tapes on a turing machine or whatever.

> You asked for an example of the legislature writing the laws of supply and demand

Rent control is an obvious example. Minimum wage laws, another. Anti-gouging laws, still another. All of them fail.

You appear to be arguing that S&D only applies to the marginal cost of reproduction. This is not the case.

Let's take an extreme example. You need some custom software for your business. You come to me. I'm going to charge you what it takes to develop the software plus a profit. Not the cost of reproduction of the software. Whether you and I agree on the price or not is a function of S&D.

And there's every point in between that and Linux.

BTW, I bet you're using a free browser to access HackerNews (I am - Chrome), and HN itself makes itself available for free.

As for copyright law, it is not setting prices any more than a law making it illegal to steal apples from my orchard is interfering with S&D.

> nobody has strong ideological attachments to Newtonian mechanics

S&D is not an ideology. It's the Law. Just like NM is the Law. It applies whether you believe it or not. You can't escape it. The communists tried really, really hard to deny the existence of S&D. They had about as much success as the flat earthers. Today's Progressivists are currently doing their best to deny S&D, with as little result as the communists.

And yes, I know NM is an approximation of Quantum Mechanics.

There's a really important caveat which is explicit in any law: "conditioned on all else being equal".

In physics, we are able to isolate systems to the point where this caveat is true and the law can be experimentally validated.

In economics, we do not get such affordances, because it is not and cannot be a "laboratory science".

This is so true. The problem is that the price people are willing to pay can go to really, REALLY high for some items because there really is no alternative. (Cars, Wood/Housing, Raw materials like steel)

you're talking about price elasticity and the role of substitutes, for which economists also have simple models that always get over-applied :)

Theories in economics are not falsifiable, thus economics is not hard science. Some would argue it’s not science at all, since it provides no objective information.

It’s fun to look up economists opinions on markets, since you can almost always find opposing views. It’s like the weather people, where there’s no downside to being wrong.

Economics is inseparable from political economy. It's entirely ideology driven and in no way scientific.

>> price is determined by BOTH supply and demand

Most of economics teaching is about exploring just how wildly wrong econ101 is.

E.g. who passes on insulin since the last price increase?

E.g. why does a basic black and white rolex daytona cost around £12k at an authorized dealer? (For non-watch people there’s a 10 year wait list to buy at that price but you can buy 2nd hand immediately for a LOT more)

People certainly pass on specific kinds of insulin when their price increases.

Note that the "old" kind of insulin is available for 25 $/vial. It sucks in a lot of ways (requires more frequent dosage, and has a smaller margin of error), but it is dramatically cheaper than the newer and better kinds.

So financially strapped diabetics do have a trade off available to them.

> why does a basic black and white rolex daytona cost around £12k at an authorized dealer?

Because they're buying a status symbol, not a watch.

> For non-watch people there’s a 10 year wait list to buy at that price but you can buy 2nd hand immediately for a LOT more

An excellent demonstration of Supply&Demand at work. Your example doesn't refute S&D at all :-)

>>> the sticker price has nothing to do with the cost of materials; it’s driven only by how much people are willing to pay

>> why does a basic black and white rolex daytona cost around £12k at an authorized dealer?

> Because they're buying a status symbol, not a watch

QED?

This is why the other poster got frustrated with you. The current watch market is a terrible example of supply and demand as a traditional model. Do you think S&D effectively models NFT marketplace? Because the watch market is a lot more like that, too complex to be models effectively by traditional S&D.

Source: am a reseller.

It's a perfect example. Supply is limited, demand is high => high prices.

> Do you think S&D effectively models NFT marketplace?

Of course it does. The supply is one item, and so the highest bidder gets it. I am baffled why you think it is not S&D.

If you have more questions, fire away!

>> and so the highest bidder gets it

The highest bidder is around £30k but the person who got it paid £12k.

You can walk into an AD and bid all you want, they’ll let you try it on or find your favourite colour version. You aren’t buying it though.

I think the point he was making is that change in supply doesn't necessarily lead to a change in price. I don't know if that's what they teach in basic economic s, but it's certainly true.

The post I replied to doesn't mention supply either directly or indirectly, rather it states the following:

"[prices are] driven only by how much people are willing to pay." Emphasis theirs.

That statement is false, prices are a function of both supply and demand and supply is a function of manufacturing costs. The cheaper something is to produce, the more of it one can produce and hence the greater the supply.

I think that quote is fine if you read 'how much people are willing to pay' as meaning the whole order book (as it were) - i.e. how many people are willing to pay how much.

The price is usually (an attempt at) optimising profit, where that may mean selling for more dollars to fewer people.

Ceteris paribus a reduction in the cost of supply doesn't change that calculation.

(Though, as the economist in my family says, 'ceteris never bloody is paribus'!)

> I think that quote is fine if you read 'how much people are willing to pay' as meaning the whole order book (as it were) - i.e. how many people are willing to pay how much.

It's really not though.

If I give you the full detailed order book of a secret product and ask you to set the price we should use as a company, you won't be able to do it. You need at least a rough idea of the production cost first.

If people will buy 10x as much at $100 compared to $200, then we should probably price at $100 if our cost is $40, and we should definitely not price at $100 if our cost is $120.

If you want to set a price off of one data lump, you need to replace the revenue numbers in that order book with margin per unit.

Yes, true, I suppose I'm assuming a pre-existing/otherwise somehow known wide margin.

The optimisation I describe is on profit, not revenue, and (as you say) you need to know the fixed and variable (marginal unit) costs to do that.

If people don't need it, they may simply choose not to buy.

A more accurate statement might be: prices are upper bounded by what people are willing to pay.

The cheaper something is to produce, the more of it one can produce and hence the greater the supply.

This only affects the price if there are competitors. If I want a widget, the total supply of widgets is irrelevant. For example Insulin, the fact that you have tons more of it won't make the price go down, you have it and I need it. You can sell it for whatever I am willing to pay irrespective of supply or production costs.

> The post I replied to doesn't mention supply either directly or indirectly

The post you replied to doesn't mention demand either. It only talks about price.

There is a difference between demand and price, and there is no generic function from price to demand. Even such basic assumptions as "demand falls as price goes up" fail in some situations.

All else being equal, a change in supply will change the price, except in perfectly elastic demand, in which case less would just be sold at the same price.

Sure not everything is always equal, but the relationship is clear

The point of Arizona iced tea is volume sales, and they maintain or increase that volume by pricing under Coke/Pepsi. Bigger can, smaller dollars.

That business is all about scale. They probably make more per unit than coke as they spend zero dollars on marketing.

I did this with an email service. We had runway and sold at 30% under cost. After about two years in our little niche, we had pushed enough volume that we were making a healthy profit. Then we sold the business to a bigger fish.

That's what basic economics theory says, yes, but not what happens in practice to make Arizona Iced Tea $0.99. Nor why a baggie of 5 screws at Lowes costs $2.99. Supplier costs are important in commodities, especially where easily compared alternatives are readily available, but that's not true of all products.

There are a lot of factors that go into pricing, but for many consumer goods the vendor is able to set prices pretty much independently of the costs.

> Nor why a baggie of 5 screws at Lowes costs $2.99

Indeed—and one might well see that price remain stable even if the cost of producing the screws doubled, because the margin's already insanely high.

That law of economics applies to a market of products and assumes that supply decreases as additional manufacturers of the product find it less profitable at a given price and drop out of the market.

When applied to a single consumer packaged good product, the supply curve is 0 at a point below the manufacturing cost and then rises to infinite at some point shortly above that price.

eg: If I an providing dollar bills and it costs 10 cents for me ship one to you, I will sell you an infinite quantity of them if you're willing to buy them at $1.11.

Coca Cola's willingness to supply Coke at $2 is infinite, if demand doubled overnight, CC would build whatever factories it needs to meet that demand and would do it for each successive doubling until it started hitting material planetary constraints like consuming all of the world's sugar.

"The basic law of economics is that price is determined by BOTH supply and demand"

That is one basic law of economics, but in some markets price isn't only determined by supply and demand.

I think a similar/analogous example for the price on the can is minimum wage. Yes, in a lot of cases wages are set by the intersection of supply and demand, but if that intersection is below a minimum wage then the "price" of labor (the wage) is the minimum wage. Now there's economic implications of that too, but at the end of the day in the US you can't pay a McDonalds employee $2 an hour and a corner store can't sell an Arizona tea for $2.

Yes but in a modern industrialized economy if there is a profit to be had the supply will create itself, without barriers to entry. However in beverages there is a big barrier to entry in consumer brand awareness and in retail shelf space especially for refrigerated products.

The supply curve you are thinking about exists when you are talking about a market with multiple suppliers. If we’re discussing a single product from one supplier, the supply curve becomes a vertical line.

No it doesn't... You don't think Arizona Iced tea would produce more cans if everyone started wanting them for $2? Marginal costs increase as the company expands due to bureaucracy.

Arizona will sell you as many cans as you want, and as many as they can supply, at their asking price.

Their overhead costs per unit decrease as a company grows.

The increase as you travel right on a supply curve represents the entry of more participants offering supply. You can think of a supply demand curve like a bunch of dots, representing different suppliers asking prices.

Imagine you’re graphing a bid/offer spread chart for a stock. Arizona’s product offering is a single “offer”.

> Overhead costs per unit decrease as a company grows.

This is correct to an extent, at Arizona's size it is wrong. It would cost them a lot to increase production.

Sure, technically speaking, it’s not an infinitely vertical line, it has an lower and upper terminus, and might begin again at another y value.

But we’re already misusing the model to begin with. The line is intended to represent multiple suppliers participating in a market for a good with a fungible source. It is not and never has been a model for pricing decisions at a single company.

Arizona Iced Tea's unit per-economics are closer to software than to food.

They both have ridiculously high incremental profit margins, so talking about finite supplies doesn't make sense.

If you're talking about actual drinkable content, then yes. The majority of the costs are in packaging and logistics - and those are getting expensive.

Reply on news.ycombinator.com