That's the magic number. There have to be four competitors of significant size before prices go down. There are both EU and US studies substantiating this.
It's been seen in cellular phone networks, ISPs, drugstores, and banks. Drop below the threshold of four, and the magic happens. Prices go up, margins go up, and consumers lose.
Four should be the basis of antitrust policy. Less than four, and there are two options - break up, or become a regulated public utility.
Two or three isn't enough. Collusion happens. Explicitly or implicitly, price competition doesn't happen with only two or three players.
Four seems to be enough that cartels usually break up on their own. Somebody won't play ball.
I was thinking about this when I saw that UAE only has du and e& which both don't provide the best customer service to formulate it nicely and prices which they could not ask for if they had to compete with their EU counterparts.
Interestingly, the US cell phone industry would seem competitive - major players (Verizon, AT&T and T-Mobile) and a bunch of smaller and/or regional players including the incumbent telcos like xfinity and spectrum. Then you dig a little deeper and see that all that “competition” runs on the towers of one of the three and unlike with the reform of British Telecom (BT), there are no requirements for this infra to be offered at any kind of lower price or at cost enabling competition for the last mile service.
5 years ago or so I had T-Mobile for $50/month unlimited everything. I switched off and just went back for $170 for a phone for my wife and I. To pretend to be comparable on a per line basis they gave me a phone number that I literally could not refuse. Before they bought Sprint we had the magic 4 but now it's 3 and a bunch of resellers.
Sprint was not going to survive. They were circling the drain and needed to be acquired to be rescued. I was a customer of theirs and they could literally only compete on price. T-mobile wasn’t in much better shape. Post merger the improvement in network quality and customer service has been night and day. Sprint owned tons of spectrum but was terrible at using it. T-mobile was spectrum starved and would not have the network they have today without purchasing Sprint.
Sprint literally bet on the wrong technology every single generation. They also had trouble getting many popular phones because they used such oddball network technology.
My understanding is that T-mobile solved their spectrum issue in the failed deal with AT&T. As part of the penalties of the deal not going through, AT&T had to give them a bunch of spectrum, and some cash to build out towers to deploy that new spectrum.
You just explained why prices went up - not because Sprint exited, but because the market bifurcated into prepaid and postpaid. You can get unlimited everything on T-Mobile's network from Mint, for about $30 per month. You just have to pay up front, and you have to bring your own phone. The "majors" have raised prices on 1- and 2- line plans because they are positioning themselves as a premium product especially for families and groups with lots of lines. No point in competing to a race to the bottom vs Mint, Cricket, Metro, etc.
4 is still an oligopoly[1]. It’s a little better than 3, but we didn’t have super cheap cell plans. Compared with most of the rest of the world, it was still quite expensive.
Speaking generally, not to you specifically, it seems like a ton of people see the problems created by capitalism and decide the solution is even more capitalism.
I don't think there's a magical number of competitors that fixes things. We see an interesting phenomenon in commercial landlords where we have a ton of players but they all use the same software [1], which creates an effective monopoly even though it's not a classical monopoly.
The problems we see in so many sectors were accurately described in the 1800s by just analyzing the workers relationships to the means of production. Internet access is a textbook cexample. National ISPs should not exist. They are rent-seeking parasites. The best Internet is in places with municipal broadband.
> Speaking generally, not to you specifically, it seems like a ton of people see the problems created by capitalism and decide the solution is even more capitalism.
"Capitalism" is a descriptive model, not an entity with causal agency. Everything boils down to the intentions and actions of the human beings involved, and those do not change by paying lip service to a different body of abstract dogma.
To the extent that "capitalism" accurately describes the structural incentives and constraints inherent in the reality that these problems are manifesting in, then there is nothing other than "even more capitalism" on the table as a solution.
Problems can only be solved from within the world that generates them -- there is no "outside", as devotees of various dogmas often mistakenly think.
> It's just not feasible or economical to have 4+ broadband networks being built in an area
Actually, it is. When telephone systems were invented, lots of telephone companies sprang up and ran wires everywhere. The barrier to this happening with broadband is regulation. Regulators work hard to ensure there is no broadband competition.
> it seems like a ton of people see the problems created by capitalism and decide the solution is even more capitalism.
I find that rather funny, as most people see the solution to failed government programs is expanding those programs. Anyone who wants to cut back such failures cannot get elected.
There is not enough space to dedicate to electrical and internet wiring with a competition model like this. The physical space is a natural monopoly. What should be done is to treat the wiring as a utility but allow for unlimited competition on the actual service, which is how the UK operates their internet service iirc
Which proves my point. It was not cost that prevents multiple sets of wires being laid. It's regulation, and frankly rather poor management by the government.
For example, a road nearby was widened for bike lanes. While they were at it, a trench was dug before the repaving to enlarge the water/sewer pipes. I emailed the people in charge that they should consider burying the power lines, too. The power lines ran on telephone poles along the road, and the poles were festooned with wires. The power lines regularly get dismantled by trees and wind storms. Laying wire in an existing trench is cheap as dirt. Digging a trench and laying wire and repaving costs about a million dollars a mile (I was told this a few years ago by a power employee).
The reply I received was they'd already made the plans and couldn't be bothered.
>Which proves my point. It was not cost that prevents multiple sets of wires being laid. It's regulation, and frankly rather poor management by the government.
Do you look at the picture I linked, and think this is a preferable end state over some extra regulatory cost? That you would rather have those sets of cables all over the place and think the government is preventing this better state?
If you do, that's an opinion you can have, but I also think we have fundamentally different views of the world.
> Do you look at the picture I linked, and think this is a preferable end state over some extra regulatory cost?
Alternatively, we can dispense with the false dichotomy (along with the presumption that the technical constraints of 19th-century telegraph lines are applicable to modern telecom) and identify ways to incentivize competitive markets without dealing with externaliies by imposing regulatory barriers that ultimately generate oligopolies.
OTOH, the modern version of your picture would probably consist of dozens of fiber lines all running through the same network of underground conduit, so the most direct answer to your question is "yes, absolutely".
> Do you look at the picture I linked, and think this is a preferable end state over some extra regulatory cost? That you would rather have those sets of cables all over the place and think the government is preventing this better state?
Yes, I saw the picture. It means it is not prohibitively expensive to run multiple wires.
Yes, it is unsightly. But there are many, many ways to run multiple signals today. Technology has advanced a great deal. For example, one trench or pipe can provide space for a large number of wires. The community can provide a pipe for such purposes, like they provide pipes for other purposes. Or the providers can agree to share a wire. Or use microwaves. Or the cell towers. Or Starlink. And on and on.
> For example, one trench or pipe can provide space for a large number of wires. The community can provide a pipe for such purposes, like they provide pipes for other purposes.
Congrats on describing municipal broadband while arguing against government involvement.
Nothing in the preceding comment remotely resembles any proposal for municipal broadband. Even if a municipal government were doing the work he assigns to "the community", that work consists of maintaining a system of conduit within which other parties can run their own fiber lines connected to their own networks, and wouldn't at all imply that the municipality would own or operate any of the physical network infrastructure, nor operate ISP services.
And there's no necessity that role of "the community" be handed off to a municipal government, either. People can and do solve complex coordination problems without relying on political authority all the time -- "the community" could easily be a non-profit organization, a mutual owned by local residents, or just a series of reciprocal agreements to maintain the baseline infrastructure needed for other parties to run cables, without any municipal government being involved.
The idea that any non-trivial coordination problem can only be solved by centralized political authority is one of the principal drivers of corruption and stagnation in modern society, and is itself one of the main causes of competitive markets degenerating into monopolies or oligopolies.
There are enormous economic forces that push for standardization. For example, your keyboard is likely connected via USB. USB isn't a government standard. Neither is Standard C. And endless other things that the market standardized. Even 2*4 lumber is standardized. Gasoline is standardized. And so on.
Ever since the iPhone, USB stopped being an actual standard to connect accessories to computers, until it became a (EU) government standard.
It is actually a great example to the opposite. The history of USB as a connector is of a direct consequence of the IBM PC, itself a standard. The only reason USB became a widespread standard is that there was a widespread standard computer and, as we see today with many other forms of computers, that is largely an accident : smartphones do not have a widespread standard, neither do gaming consoles, etc...
I never claimed as much. The IBM PC was a short-lived accidental scandal that IBM didn't fully intend, but couldn't stop when they failed to sue clean-room BIOSes out of existence. Only because of this was USB possible, and even then only for 10 years before it stopped being really universal, until the EU stepped in.
And yet somehow those incompatible telephone systems, rail systems, power systems, and everything else all came into existence.
USB is widely highlighted as an unusual success. But Apple has notably created their own incompatible connectors, and made money that way, which is why Europe is busy forcing a charging standard.
Gasoline is very obviously government regulated. 2 by 4s are only sort of standardized.
> Gasoline is very obviously government regulated.
The chemical makeup of it? The government banned tetra ethyl lead in it, and mandated some ethanol in it, but the rest is various mixes made by the gas companies.
> 2 by 4s are only sort of standardized.
I've bought 2 by 4s for most of my life. They're a standard size.
> The chemical makeup of it? The government banned tetra ethyl lead in it, and mandated some ethanol in it, but the rest is various mixes made by the gas companies.
There's a mil-spec gasoline that has to contain exact proportions of specific alkanes and ethanol to very tight tolerances, that could be taken as the golden standard gasoline.
Let's say it costs $1000 to bring fiber to each household in a given network build. The take up rate might be 35%. That is 35% of connected households become customers. Well that customer has to recoupe $3000 to pay for that build.
Now imagine there are 2 last mile networks. The take up rate hasn't changed. If they cost the same then each customer has to recoup $6000.
Network overbuilds make absolutely no sense.
Government programs don't fail per se. They're designed to fail by underfunding them or adding layers of administration (eg state block grants to replace direct Federal funding) and then that failure is used to justify more cuts. It's called starving the beast [1].
> Regulators work hard to ensure there is no broadband competition.
Regulators can be the tool to stifle competition. The driver though is inherent in business needs. Competition is the biggest cost to any business, so there will always be a drive to reduce the competition to as close to zero as possible. It's inherent in how businesses operate.
Capitalism is a system that needs to be within operating parameters to function. Prevention of monopolies and cartels is an important part of those operating parameters.
>That's the magic number. There have to be four competitors of significant size before prices go down. There are both EU and US studies substantiating this.
"Entry and Competition in Concentrated Markets" [1]
This is a paper from 1991, from some economists at Stanford. They say the key number is between 3 and 5. This was a study of smaller businesses in small towns - doctors, dentists, druggists, plumbers, and tire dealers.
"Competition Policy Brief (2021)" An EU study.[2] This is much more of a macro study.
"How European Markets Became Free: A Study of Institutional Drift"[3] This one compares US and EU antitrust policy changes over time, especially for telecom and airlines.
There's overall agreement that trouble begins below 5 competitors.
Cant wait for somehow the government to require four companies in order to ensure good competition by some well intentioned poorly reasoned and executed policies.
It’s a statistic, not an absolute, obviously 4 is probably the minimum. It’s a rule of thumb and there to say “2 or 3 probably isn’t enough, it takes at least 4 AND government oversight” otherwise, uh, capitalism finds a way [to cheat]
This. More, smaller companies mean more duplication and overhead which can manifest in higher prices. One of the biggest factors behind market consolidation and centralization is because it reduces duplication and overhead which can increase margins and might reduce prices.
For my area in Los Angeles, it wasn't viable because everyone (or someone you know) has a car. a friend ride is much cheaper than an Uber.
I only download the Uber app when I need to get to the airport and am leaving for weeks at a time. Any other time, I take myself, ask a friend, or even consider what remnants we have of public transportation (e.g. instead of taking a horribly expensive uber in/out of LAX, take a shuttle to some smaller area and then hitch a ride. can easily save 4x the cost for a $5 bus ride and maybe 30-45 extra minutes)
A lot of us won’t ask friends for rides because it’s kind of selfish, I’ll do an uber/lyft every time unless my buddy is right there saying he’s heading that way.
Well I have my car and we tend to alternate carpools, so there's not usually any pressure. I tend to give them $10-20 to cover gas and I still come out way ahead.
I just never get a good time to go to/leave an airport, so that's the only time I resort to Uber.
Try $150. Trying to Uber out of LAX is highway robbery and they know that. Going from my house to LAX is anywhere from $60-100. Going the other way is anywhere from $100-300, and only gets worse on holidays.
Shuttle to anywhere else is $10 and prices become normal again. Best advice I ever got. That and trying to use other airports where possible.
...when the government legislated pay is over $26 per hour. Plenty of business would see higher prices with business-specific pay legislation mandating pay substantially higher than minimum wage.
> Four should be the basis of antitrust policy. Less than four, and there are two options - break up, or become a regulated public utility.
3 options; usually monopolies are matched by regulatory constraints making competition expensive. Removing the barriers to other firms entering the market is always an option.
I'm not sure of many, if any, monopolies where regulatory constraints are the cause. Especially compared to causes like network effects. Do you have a modern example?
I can think of a couple of examples where regulations exist to manage other constraints on competition, like the FCC regulating the spectrum. But not where the regulations cause the number of major competitors to decrease dramatically.
One example I like is Internet Explorer. IE had a "monopoly" back in the early 2000s and made up something like 90% of the browser market. Then Firefox came along and broke the market open, then Google and Apple rolled in like bulldozers [0] and we got to a healthy 4-player market.
Everywhere except for South Korea. You can still read articles on it to this day [1]. Regulations had an unfortunate interaction that made IE into the only option and set up an actual monopoly situation.
The caveat is I personally think monopoly is called far too quickly; as far as I can tell every successful product is a monopoly to someone on HN. A lot of the things people want to call monopolies aren't, they are markets like the browser market that haven't finished settling. Unfortunately they can become ossified due to bad regulatory practice.
[0] What Steve Jobs did to Flash should come with 18+ violence content warnings. They were the days. Nothing could be done about that scourge until suddenly something had been done.
IE didn't lose in the market when Firefox came along. IE lost when the government fined Microsoft billions of dollars and forced it to relinquish it's monopoly position.
Meanwhile the current market is Chrome, reskinned Chrome and Safari. Hardly a healthy marketplace.
That's the magic number. There have to be four competitors of significant size before prices go down. There are both EU and US studies substantiating this. It's been seen in cellular phone networks, ISPs, drugstores, and banks. Drop below the threshold of four, and the magic happens. Prices go up, margins go up, and consumers lose.
Four should be the basis of antitrust policy. Less than four, and there are two options - break up, or become a regulated public utility.
Two or three isn't enough. Collusion happens. Explicitly or implicitly, price competition doesn't happen with only two or three players. Four seems to be enough that cartels usually break up on their own. Somebody won't play ball.
Four.