Pricing software offers implicit collusion by providing competitors a way to agree on a price that is equivalent to monopoly prices, without them need to formally meet and agree on a price - which would be easily prosecutable.
This paper discusses how algorithmic pricing produces outcomes that are equivalent to collusion-based pricing in the German gasoline market.
Recent experiences with grocery pricing across North America and Europe look very similar to me. I know some computer scientists who have worked with major grocery retailers to implement fairly sophisticated automatic pricing tools. The goals they work towards are typically fairly simple metrics, like average spend. Even though the goal isn't to gouge consumers, it's easy to see how how this might be the unintended result.
I don't understand the mechanism here. The thing that makes competition work isn't that competitors don't know each others' prices. It's that each participant in the market makes more money by lowering their price below the monopoly price.
No one is selling exactly the same goods anyway, not at that level.
All purchases (wholesale, retail, whatever) are actually auctions, and they remain auctions even if you aren't aware that they are. So if someone was saying to themselves "Dammit, Kroger sells lettuce so high, I could that I could sell it for 4% over cost and still be 20¢ cheaper!" then they are deluding themselves. Ignore the overhead of opening a nationwide chain and all that other stuff.
When they go to purchase lettuce wholesale, as a new competitor, more lettuce doesn't pop up in the fields instantly. Or even quickly. So you end up causing wholesale prices to rise, at least until the agricultural sector catches up, if it does. They were always careful to grow just enough that none of it ends up unsold. Even if you could source it, likely you're getting the lower quality lots (the stuff Walmart goes after just so they can have theirs priced a penny cheaper, that turns slimy 24 hours after putting it in your fridge).
This discourages new entrants, competitors. And gives Kroger some slack if someone does decide to try it.
> They were always careful to grow just enough that none of it ends up unsold.
It is a market. However much they grow, exactly that much will end up sold. They'll drop the price if there is unsold inventory to sell it.
They can't be doing a physical estimate of how much lettuce people eat. If there was any variance at all in how much lettuce gets eaten (which there is) then there'd be constant lettuce shortages where the shelves were empty. That doesn't happen, so we can tell that they are using price signals to control how much gets sold.
The mechanism isn't additional information; it's enabling the collusion the individual participants would already prefer to participate in. With simple inventories (like gasoline), computers aren't essential for collusion. In more complicated markets it's helpful to let the machines do the "heavy" lifting.
When all else is equal, the preference for sellers is to charge the same price a monopolist would choose (assume temporary price excursions don't have long-term ramifications like brand loyalty, and assume that "all else equal" extends far enough that competitors would also drop prices to match to avoid greatly reduced sales).
Realish-time updates, apparently with very simplistic rules, seem to converge near that price point, and the inequalities of various sellers seem to not matter much (because there aren't many, because they average out, because they don't impact price preferences much, ....)
As one player raises their prices incrementally, this information is passed on to all the other players immediately. The software will then "recommend" that other players raise their rates to match. Or the software recommendations a slightly higher price to all players. With players of this size, small increases make a big difference.
This is only true when prices change slowly. If I use a computer to do my pricing, the here’s what can happen in real time.
Player A raises prices. Algorithm automatically raises everyone’s prices to the new average.
Player A lowers prices to be in line with new average.
everyone’s prices come down very slightly because A is now towing the line.
A was able to raise the prices of everyone before anyone can adjust their spending accordingly.
This is a real problem, and big players are going to continue to learn how to game it to the detriment of all.
BTW I’ve seen this strategy play out in online games with heavy botting. It’s funny to see there. Not funny to have potential to happen in real life.
There is no magic about the speed of price changes. If slow or fast price changes made a difference on amount bought, then companies would do that speed. But they don't.
Similarly, prices rising fast doesn't trick people in aggregate to ignore the increase - they still react to price increases.
This is all pretty basic econ. If this "new" system you are afraid of allowed rampant price increases, everything would cost more, yet things have not increased (except for a recent inflationary period completely predicted due to free COVID cash giveaway and Russian oil shocks).
Algorithmic price discovery lowered consumer cost in stock markets big time. It lowers risk to producers as they get up to dat information, and lower risk means lower spread required.
>A was able to raise the prices of everyone before anyone can adjust their spending accordingly.
Not "everyone" will buy at an increased level - if that were true, then the producer would raise the prices anyways. The algorithms only outprice people until those running them realize they screwed up and improve the algorithms.
>big players are going to continue to learn how to game it to the detriment of all
Again, if prices could be raised to get more profit, than that is what companies would have already done. They cannot because people take their dollars elsewhere.
>BTW I’ve seen this strategy play out in online games with heavy botting. It’s funny to see there.
And if it costs people real money, some price out and stop. Same as in reality.
This seems like a huge problem without an easy solution. Now that grocery stores are starting to deploy electronic price tags, I don't see why they wouldn't implement exactly that.
The only ideas I have are to limit how often prices can change or mandate prices, but both of those have some huge implications and downsides.
Did people buy less food after grocery stores raised prices? I doubt it. If you have something people require then you can increase prices a ton as long as you ensure everyone else also does it.
The reason this doesn't happen for groceries is that starting a grocery store is very easy, if everyone increases grocery prices a competitor with lower prices will appear extremely quickly. But for services that are harder to setup you can't do that, replacing a pharma or a chip company isn't something anyone can do even if they had lots of money.
And the reason we don't want a monopoly to set these prices is that it blocks progress. Imagine if you had to pay the maximum you'd be willing to pay for food instead of the cost it takes to produce? You'd be forced to spend most of your salary on food or ration it, that isn't a society you want to live in.
>Did people buy less food after grocery stores raised prices? I doubt it.
Yes they did. So your premise needs checked. Read any news about what happens to food sales as inflation outpaces income and it's abundantly clear stores cannot simply raise prices without losing sales. There's such easy literature to find you don't need to "doubt it" when you can simply check it. For example [1]
Some consumers purchase at the limit of what they can spend - there's no elasticity for them.
> The reason this doesn't happen for groceries is that starting a grocery store is very easy, if everyone increases grocery prices a competitor with lower prices will appear extremely quickly.
But a) it is not clear that this is, in fact, true at this point in time, and b) even if it is, "extremely quickly" is still going to be on the order of a year or two.
The reason I say (a) is because the existing grocery store chains are very large, and very willing to lower prices locally to prevent a competitor from getting an edge on them. They also have significant economies of scale that allow them to drop prices lower than a new local upstart could and still make at least some profit.
In short, the barriers to entry are high enough that a few months of raised prices aren't enough to cause a competitor to appear out of nowhere, and any would-be competitor would need to either have massive resiliency to outlast the incumbents undercutting them (again, locally, such that it wouldn't make a blip in the overall inflation numbers), or somehow start up enough locations all at once that such an undercutting attack would be less feasible and much more visible.
This is the problem with all players raising prices to the same rate at nearly the same time. There are simply no alternatives and people need to pay the higher price. That's kind of the whole point of the article.
The point of the article is to get views. If it were a decent study of facts, it would be better as a peer reviewed article. But it wouldn't stand up to that scrutiny, due to the fact that in such fake collusion any player can gain by breaking the pact.
Even OPEC cannot keep all their members in check to set oil prices - countries routinely undercut them to sell more oil. This has been demonstrated time and time again in industries where people claim long term widespread collusion - such things don't last long.
Not if the product is controlled by an explicit or an implicit cartel. The point being made is that cross company pricing services leads to an implicit and deniable cartelization.
High profit doesn't necessarily mean the players aren't at rock bottom. Some industries need high profits to justify the effort. If all you can eek out is a small profit, you can do just about anything else.
But, assuming there is room to go lower, then the market isn't yet competitive and has room for someone to swoop in and take the spoils. If there is some regulatory barrier that is preventing that, then there was no illusions of it being competitive in the first place.
The only possible reason for barriers to entry is regulation? There are many industries where startup costs are high for reasons intrinsic to the business and margins are relatively low so you are unlikely to get investment from outside sources. Grocery stores for example. You are trying to apply a toy model from econ 101 to explain the behavior of complex real world markets, and when the model doesn't fit you invent bogeymen to blame. This is more like religious fundamentalism than any kind of science.
Grocery stores are already operating at rock bottom – in most markets, at least.
It is not that difficult to try opening your own grocery store. In fact, many restaurants did exactly that during COVID-19 shutdowns. Realistically, succeeding is going to be nigh impossible, though, as there is not much you can compete on. You are not going to be able to sell the product for less.
It is not meaningfully bound by a regulatory barrier, but it is limited by there being no further room for competitiveness, as also spoken to in the previous comment.
I find it very funny that a forum of people who ostensibly work in the highest profit margin businesses in the world, enjoying the highest industry wide compensation to quality of life at work ratios, claim that the lowest profit margin businesses in the world with some of the lowest compensation to quality of ratio work has enough pricing power to allow them to reap undue windfalls of profit.
When it's entirely enabled by your vein of work, it is absolutely fair game to call out your fellow practitioners. Further, not everyone here is FAANG. Nor does everyone here necessarily get the majority of motivation on what they work on from the paycheck rather than the effects of the work done.
As the ostensible computer scientists in the room. Sitting back and not pointing out that there exist "monopoly pricing indirection mechanisms" implemented as businesses is really failing to do one's moral and ethical responsibility.
> As the ostensible computer scientists in the room. Sitting back and not pointing out that there exist "monopoly pricing indirection mechanisms" implemented as businesses is really failing to do one's moral and ethical responsibility.
And yet the empirical evidence is not there due to nonexistent profit margins. Show me the sustained increase in profit margins if you are going to claim malfeasance. And I am not claiming there is not malfeasance, I just don’t want to see innocent parties (those with low single digit profit margins) get accused of it for no reason.
Is it possible businesses are colluding resulting in increased profit margins? Obviously.
But we have publicly listed companies with public financials showing non material increases, or even decreases, in profit margin. Which means those businesses are just increasing prices to cover their own increasing cost of goods sold.
Thank you. Much too often the replies in threads on prices are just a conspiracyfest of people alleging things that make no economic sense without providing any kind of evidence.
Because it is of course completely obvious that collusion among very large, very complex competing entities is a very easy proposition, you just need a magic algorithm to do it! /s
Yeah I'm skeptical. When you say "maintain high profits", maybe that is true maybe it isn't. But they are sharing the market with all of their competitors.
Retail business profit margins are 1% to 5%. Mathematically, how can the prices be any lower without the business failing? Hence we see similar prices everywhere.
Grocery stores especially have 1% or 2% profit margins, so logically, the things they sell must be priced as low as they can. And also why a mom and pop grocery store cannot compete with Walmart/Kroger/Costco/Target/etc, you need those huge economies of scale otherwise your prices will be uncompetitive.
How can grocery store have a 1% to 2% profit margin, when the price of the same item across grocery store chains can vary by 10-30% or more? I can't imagine it coming down to a store having higher costs alone.
>How can grocery store have a 1% to 2% profit margin, when the price of the same item across grocery store chains can vary by 10-30% or more?
1. Individual price discrepancies of "10-30% or more" doesn't really matter. What does matter is overall markups.
2. The stores themselves might be upscale/higher tier, which also makes their stuff more expensive. I'm not talking about whole foods carrying organic products, I'm talking about stores that have better selection, full service butcher/deli, better cleaning, better interior design/decoration, or better location (richer neighborhood).
Ask yourself why the stores that charges 30% more are still around? Likely they operate in different areas, or they offer some services the cheaper place doesn't, or they have higher quality wares etc. Otherwise everyone would go shop at the cheaper store.
Now ask yourself, why doesn't the cheaper store offer those things? Maybe because they costs something?
In 2022, Kroger made $2.2B on $34.8B sales. That’s more than 6% _profit_, never mind margins. These companies are making way more margin than you are saying they are.
Edit: looks like that was just the quarter — still 1% profit which means margins must be far higher. Margins in retail, by the way, are mark-up over wholesale cost.
Those numbers are wrong. Yahoo shows their financials: revenue was 122b, 132b, 137b, 148b for years 2020-2023. Normalized the income was respectively 1.4b, 1.7b, 2.3b, and 2.8b.
Kroger buys other retailers. Their income (and their taxes) will be reduced by (amortization of) the intangible cost of the retailers bought. This is money paid to the shareholders of the retailers who sold out, which should also be counted as income of the retailing sector.
Furthermore, we now have an economy in which many so-called industries have a single-winner or have a race to become the single-winner now in progress. So just about every firm that advertises is paying uncompetitive rates for eyeballs in the media markets, every firm that accepts credit cards paying uncompetitive rates for payment processing, and seemingly every firm that wants to have more control over its pricing is paying exorbitant executive compensation for those who are supposed to bring that about. If the firm is at all profitable, the customers pay for all of that, too.
> Kroger buys other retailers. Their income (and their taxes) will be reduced by (amortization of) the intangible cost of the retailers bought. This is money paid to the shareholders of the retailers who sold out, which should also be counted as income of the retailing sector.
This does not make any accounting sense. Profit (net income) is not a function equity, and what if the prior owners lost money on the investment?
Also, what intangibles are you referring to in a grocery business? The buildings, real estate, supplies etc are all tangibles.
If a corporation buys some stores from another corporation it pays a price. It then values all the tangible assets to find out how much it puts on its books as tangible assets to be depreciated over their respective useful lives. It will quite typically find that it paid more in total for the acquisition than the tangible assets are worth, but it is not usually required to report a loss for paying more than the sum of the values of the tangible assets. The difference is the value of the acquired businesses as going concerns, sometimes called 'good will.' That becomes an asset on the acquirers balance sheet to be depreciated over some period of time. The acquirer may also be able to assign some of the purchase price to capitalize such assets as deferred tax credits, contracts signed by the sold business, trademarks and licenses, etc. Those assets may be written down over years to offset income of the combined entity that would otherwise be taxable. Then there is the possibility of buying a corporation, holding it for a few years and then spinning it off tax-free, which can equivalent to a huge dividend to the shareholders of the purchasing company that never was taxable as corporate income.
>which should also be counted as income of the retailing sector
This makes no sense.
If I own an asset worth X and sell it for X, I made zero profit. Paying shareholders to buy them out is not all profit, it's trading one asset for another.
Next, these deals are rarely simply cash giveaways, but include all sorts of other asset trades (stock in new company, payment over time... etc), also not being simply magic pure profit. Next, they're not simply taken out of one year's profits, but are generally financed by taking on more debt, so this is not some way to hide or reduce profits.
>we now have an economy in which many so-called industries have a single-winner or have a race to become the single-winner now in progress.
This is simply untrue. Pretty much every industry has lots of players. And there's constant churn. There's nearly zero product categories I cannot shop between many sellers.
The larges US banks have under 20% share, and there's literally thousands of banks.
The largest US grocery by dollar share include: Walmart 18%, Kroger 8.8%, Costco 6.4%, Albertson 6.4%, Delhaize 4.3%, Publix 3.7, Sams Club 3.6, Target 2.4, and literally hundreds more.
The same pattern of the largest company follows in pretty much every industry.
So pick some industries where you think there's a single winner that make up a decent amount of consumer sales and list them. I don't think you'll find any.
There are examples in many markets, here is a well-researched example that comes to mind: https://economics.yale.edu/sites/default/files/clark_acex_ja...
This paper discusses how algorithmic pricing produces outcomes that are equivalent to collusion-based pricing in the German gasoline market.
Recent experiences with grocery pricing across North America and Europe look very similar to me. I know some computer scientists who have worked with major grocery retailers to implement fairly sophisticated automatic pricing tools. The goals they work towards are typically fairly simple metrics, like average spend. Even though the goal isn't to gouge consumers, it's easy to see how how this might be the unintended result.