Alchian–Allen effect

https://en.wikipedia.org/wiki/Alchian%E2%80%93Allen_effect

by henryaj • 8 years ago

122 53 8 years ago

en.wikipedia.org

dfee• 8 years ago

when the prices of two substitute goods, such as high and low grades of the same product, are both increased by a fixed per-unit amount such as a transportation cost or a lump-sum tax, consumption will shift toward the higher-grade product. This is true because the added per-unit amount decreases the relative price of the higher-grade product.
Another example is that Australians drink higher-quality Californian wine than Californians, and vice versa, because it is only worth the transportation costs for the most expensive wine.

ashrk• 8 years ago

My unscientific observation has been that the opposite holds for French wine. They seem to ship the US their worst, but it hits our shelves marked up 4x what it'd be in France.

Maybe it's a branding thing, and French wine's got enough power in that department that they don't need to send us good stuff to move product, at least at the lower end. Maybe with some French writing on the label low-price-shopping consumers are equally likely to buy it whether or not it tastes good, so there's no reason to send anything but the worst they have. Other European countries seem to have no trouble shipping us nice wine in the lower price tiers, but then they're not France.

zwaps3• 8 years ago

In reality there's a lot more in a product than just good and bad.

Especially with wine, where it is not actually so clear that a true quality ordering exists, all sorts of other effects are at play. For example, we know how the signaling value of price overrides some true quality variables.

In economics, its important to know these simple models and their outcomes, but it is equally important to know when other effects are at play, or when assumptions do not hold

andyjohnson03• 8 years ago

I don't see anything in the article about the extent to which this effect has been observed in the real world. Is it another example of economists assuming that people are idealised rational actors (who would prefer a 1.6x coffee bean premium to a 2x, to take the example in the article), or is there some hard evidence that it explains the real behaviour of real people?

dyeje• 8 years ago

They give 3 examples at the end:

"The effect has been studied as it applies to illegal drugs and it has been shown that the potency of marijuana increased in response to higher enforcement budgets,[2] and there was a similar effect for alcohol in the U.S. during Prohibition.[3] This effect is called Iron law of prohibition or Cardinal rule of prohibition.

Another example is that Australians drink higher-quality Californian wine than Californians, and vice versa, because it is only worth the transportation costs for the most expensive wine.[4]"

andyjohnson03• 8 years ago

Thanks. I did read the article and noticed the examples that you referred to.

To clarify: I just don't see where the evidence for any of these claims is accessible. Four of the six references don't appear to be available on-line. Of the two that are, one is a brief summary that leads to a paywalled FT article that I can't read, and the other is a ~200 pdf economics textbook.

I'm happy to accept that a price differential changes when a constant is added to both prices. But I do question whether people in the real world will switch to a different type of coffee bean because the differential goes from x2 to x1.6. And I wanted to know more about the real-world circumstances in which it had been observed.

I might be pursuaded that the market for illegal drugs/alcohol could exhibit this effect, but it would be interesting to know how how, given that illegal products are probably going to be subject to other effects not shared by legal products, biases and selectivities were compensated for.

Edit: Made first paragraph clearer.

srtjstjsj3• 8 years ago

Those don't seem to be the same effect as buying fancier coffee because it seems like a comparatively better deal per pound. Prohibition makes a product hard to get in desired volumes, beyond just price, so you want as much as you can get when a window opens; you can dilute it yourself at home later. It may be the case that demand an't be satisfied by buying weak lots in low quantities.

Matticus_Rex• 8 years ago

Most of the reaction against modern economists for assuming rationality is just meme repetition. Yes, this is an accurate criticism of a tiny portion of modern economists, and of many economists of prior eras. One of my favorite books, Kahneman's Thinking, Fast and Slow, is a prime example of how pervasive this meme is -- he seems completely unaware that the critique has largely been integrated into the field well before the time of publication. The assumption of rationality has been replaced by assumptions around bounded rationality and cognitive bias.

Modern economic theory is not some abstraction that doesn't resemble the real world, even in cases where the empirical evidence is as-yet weak (which it's not in this case, as is apparent to anyone who reads the whole entry).

greggyb2• 8 years ago

Maybe I had pretty biased professors, but in my study of economics, there was no standing assumption that individuals are rational.

Rationality is an observation of aggregate behavior, rather than individual. This is pretty robustly observed. It is interesting because this is in spite of the glaringly obvious observation that most people are not rational at all times (though many people are sometimes rational). Thus it is a useful abstraction to reason about rational individuals, though I have never been exposed to a serious claim in academic economics that individual human beings are rational.

The interesting results of behavioral economics are many, but they primarily give us proof of the already-known fact that individual humans are often irrational. The interesting part is not that humans are individually irrational, but how.

Behavioral economics makes the observation of aggregate rationality even more interesting, because many irrational actors can cause aggregate rationality.

If you would like my musing about how it is that what I have said above is so often misunderstood, then I have some thoughts below.

In my academic experience, all of my intro level professors took pains to ensure we understood that we were learning useful models, and that rationality is an observed phenomenon in aggregate. My intermediate level professors often didn't take as many pains to make this clear, but did make similar explanations. For anything 300-level or above, we never really discussed this topic, as the professors assumed we had already learned such.

Based on my observations, I would guess that most people who do not take a full course load of intro and intermediate economics do not have drilled into them that rationality is not something we expect of individuals, despite it having a lot of explanatory power in aggregate observations.

Additionally, as with all fields, those who practice often forget what is taken for granted in the field. Among economists, unless one has significant experience teaching economics to beginners, I expect it is not a common thing to preface every statement with the type of 101-level introductory remarks that would help to remove such confusion.

Additionally, economics seems to me to be plagued with jargon that are homonyms with common words in the vernacular. E.g. "cost" to a layperson often means "value denominated in currency, which I might pay to someone to acquire a specific good or service" or "the amount my supplier paid for the thing I am purchasing, such that their cost is less than the price that supplier charges me", whereas to an economist, "cost" is pretty much always going to mean "opportunity cost"[0], which is the value (to the entity making the choice) of the next best alternative. This sort of homonymity is rampant in economics jargon.

Finally, there is a pervasive assumption in economic thinking of "ceteris paribus", or "all other factors being held constant" (technically "all else equal", but give me some leeway in the translation to capture intent a bit better). Most theories, models, and statements in economics are going to be statements about a single factor/observation with a magical "nothing else can or will change" condition applied to the entire universe. This is, obviously, a huge over-simplification, but one that is useful for discussion.

In summary, all models are wrong, but some are useful. Economists often discuss in terms of models and the way they discuss uses words you think you understand but you don't, unless you study economics (not a dig, just a truthful observation of laypeople using jargon - see a recent SMBC for non-economics examples[1]).

[0] https://en.wikipedia.org/wiki/Opportunity_cost - Key excerpt:

Thus, opportunity costs are not restricted to monetary or financial costs: the real cost of output forgone, lost time, pleasure or any other benefit that provides utility should also be considered an opportunity cost.

[1] http://www.smbc-comics.com/comic/physical-analysis

pwaivers• 8 years ago

The effect has been studied as it applies to illegal drugs and it has been shown that the potency of marijuana increased in response to higher enforcement budgets...

If the price of all marijuana goes up, then I bet overall usage of it goes down even if the individual potency goes up.

cbsmith6• 8 years ago

Depends... if quality goes up, sometimes that grows the demand for a product. Kraft Parmesan convinced me I didn't like Parmesan for the longest time...

trhway• 8 years ago

one of the reasons in play for the fentanyl and carfentanyl trafficking observed these days.

If the price of all marijuana goes up, then I bet overall usage of it goes down

in high margin (and supply limited) goods i think it works in the opposite direction - the price goes up in response to the demand going up.

cbsmith6• 8 years ago

I didn't realize they had a name for it... it seems so obvious.

azeirah• 8 years ago

Why shouldn't obvious things have names?

pc863• 8 years ago

They weren't saying obvious things shouldn't have names.

They were saying they didn't realize it had a name.

cbsmith6• 8 years ago

Usually if it is an obvious thing, you don't have a name tied to the people who discovered it... you know, because it predates recorded history.

jldugger• 8 years ago

it seems so obvious.

I'm not sure the effect is logically sound -- before the price hikes you weren't willing to pay a dollar more for name brand (or premium drugs etc), but after the price hike you are? I'm not sure if the authors have a mathematical model supporting the decision, or are just observing illegal drug markets and generalizing.

TezlaOil• 8 years ago

before the price hikes you weren't willing to pay a dollar more for name brand (or premium drugs etc), but after the price hike you are?

Imagine that 10 normal people and Bill Gates want to buy bottles of wine. Bill goes for a bottle of $6000 Richebourg Grand Cru, and the 10 normal people always go for Shitscreek Red, sold for $4. So about 9% of the buyers are willing to pay extra for the name brand.

After the Great Wine Tax Bill of 2021, the prices are both increased by a fixed per-bottle amount of $6000. So now the Richebourg costs $12000, and the Shitscreek costs $6004. The 10 normal people migrate to beer, and Bill Gates still buys Richebourg. So about 100% of the buyers are willing to pay extra for the name brand.

aqme28• 8 years ago

That's not what is being claimed. The _ratio_ of high quality to low quality products shifts in favor of higher quality as the prices of the two equalize due to fixed costs.

cbsmith6• 8 years ago

So let's say you'd have to pay a dollar more for a higher quality product that costs two dollars. That's twice the price, and hard for you to justify.

Now, imagine that someone puts a tariff in place so that the low quality product now costs ten dollars, and the higher quality product costs eleven dollars. Sure, you might still buy the lower cost product, but if you are going to buy the product at all, the incremental cost of spending eleven dollars instead of ten dollars seems much more worth it. As long as you have to spend ten bucks anyway, might as well spend eleven and get a higher quality good.

amelius2• 8 years ago

It doesn't sound obvious or universally true to me. For example: what if the expensive product is of only 1% better quality than the cheap product? Will people who use the expensive product not think that they are paying too much for too little gain, and switch to the cheaper product?

silvester23• 8 years ago

That's completely independent from an increase in price though, isn't it?

cbsmith6• 8 years ago

The point is that even when there is a fixed price gap between higher quality and lower quality goods, the rationale to buy the higher quality gets stronger as that gap becomes smaller proportional to the price of the products.

It's not that people always buy the more expensive product.

tlb• 8 years ago

The reverse effect works too. When an overhead cost drops, people consume crappier versions. Free online news is a striking example. If you walked down to the news stand and paid $0.75 for a physical newspaper and the content was all Clickhole listicles you'd never buy it again. But people consume plenty for free.

srtjstjsj3• 8 years ago

That's not the reverse effect, that's supply and demand, and people forgetting opportunity cost of time. Newsstands sell plenty of junk like US Weekly.

pc863• 8 years ago

It is literally the opposite of the effect described in the article.

mikeash• 8 years ago

There was and still is plenty of Clickhole-level crap on newsstands next to the newspapers. People buy it.

stonecraftwolf• 8 years ago

In fact IIRC people buy printbait considerably more than they buy actual newspapers.

gnulinux2• 8 years ago

the Alchian–Allen theorem

Uhm, I'm not an economist but how is this a "theorem"? What is the theory, what are the axioms, where is the proof? This seems like an empirically observed trend.

shitgoose2• 8 years ago

there is some basic arithmetic in the article. imagine the cost was raised by $100, then prices of beans would be $101.5 and $103, i.e. pretty much everyone would go for higher quality. gradually decrease the premium, and you will gradually decrease the number of people who prefer the higher grade product in favor of lower price. this is pretty much it. quite simple and neat observation, actually.

"the harder the punishment, the harder the drug" is another very important non-trivial observation, missed by most people.

gnulinux2• 8 years ago

quite simple and neat observation, actually.

Did you read my comment? That's exactly what I wrote. This is not a "theorem". Theorems live inside theories, they're logically implied by axioms of the theories. That basic arithmetic proves nothing of the sort.

daxat_staglatz2• 8 years ago

Google got me this [0], which seems to fit what you want

[0] https://mpra.ub.uni-muenchen.de/901/1/MPRA_paper_901.pdf

tr3523• 8 years ago

Similar to an even exchange in chess. If you’re behind it’s a good deal, if you’re ahead it’s bad. Of course evaluating value taking into account positions (and thus determining whether an exchange is really “even”) is not straightforward.

daxat_staglatz2• 8 years ago

This seems wrong: an even exchange does not change absolute balance but magnifies relative balance, which you do not want if you're behind.

Let's say your "strength" is simply the sum of the value of your pieces. The losing side has a strength of l and the winning side has a strength of L.

The losing side loses by l-L in absolute terms, or (l-L)/(l+L) in relative terms.

An even exchange of value k makes it go to (l-k)-(L-k) = l-k in absolute terms (no change); and to (l-L)/(l+L-2k) in relative terms. (l-L)/(l+L-2k) < (l-L)/(l+L).

To take an example, if the situation is white : two pawns and black: one pawn, going to white: one pawn and black: nothing is a bad deal for black (the losing side)!

tr3523• 8 years ago

You're right, I meant the exact opposite: an even exchange is good if you're ahead, bad when you're behind.

murftown• 8 years ago

I don't consider myself a good chess player at all, but I always thought the opposite way: When I'm ahead, I'm willing to whittle down both sides with even trades until they are left with no options and can be checkmated. When I'm behind, an even trade feels scary - I'm looking for a way to catch up.

Am I missing some reason why it would be the opposite?

tr3523• 8 years ago

Yes you're right, I meant the exact opposite.

robrenaud• 8 years ago

Another example is that Australians drink higher-quality Californian wine than Californians, and vice versa, because it is only worth the transportation costs for the most expensive wine.

As a person who loves Belgian beer, I came to this realization when I was on a bit of a beercation there. For the most part, the good Belgian beer makes it here. When I'd try beer from breweries I hadn't heard of while I was there, it was mostly worse than the Belgian beer I had tried at home.

anonytrary• 8 years ago

I thought this was just common sense. Suppose you're buying necklaces.

Today: Gold $10; Silver $0.10

Tomorrow: Gold $1010; Silver $1000.10

By adding a constant to both prices, the price ratio no longer remains aligned with the underlying value ratio. Assuming you have to get one necklace each day, there is less of a decision tomorrow than there is today, because silver is overpriced relative to gold. The correct prices for tomorrow should have been "Gold: $1010, Silver: $10.10", so you're effectively getting more scammed by going with silver.

jetrink• 8 years ago

I'd agree that once presented with as stark an example as yours, it is common sense, but it takes some insight to create such a thought experiment in the first place.

lurquer2• 8 years ago

So that's why I go broke on vacations... I want to eat somewhere cheap and the wife says, "No, let's go to the expensive joint... we've already spent $x getting here, so an extra $50 won't make a difference!" Now I have a name for the phenomena.

peterkshultz• 8 years ago

I would call what you're experiencing the sunk cost fallacy [1].

[1] https://en.wikipedia.org/wiki/Sunk_cost#Loss_aversion_and_th...

lurquer2• 8 years ago

No. The sunk cost comes into play in explaining why we've gone to Disney World three time too many simply because we bought some multi-day package years ago. I hate Disney. She hates Dianey. But, we gotta use up the stinking tickets.

Alchian–Allen effect

Related Stories