« Don't assume that mandates are cheaper | Main | Assorted links »
Antitrust Protectionism
Best evidence that the merger between Microsoft and Yahoo will increase competition?
Publicly, Google came out against the deal, contending in a statement that the pairing, proposed by Microsoft on Friday in the form of a hostile offer, would pose threats to competition that need to be examined by policy makers around the world.
Here's an Open Letter on Antitrust Protectionism (pdf) that I drafted nearly ten years ago during the Microsoft trial. It's still relevant today.
Posted by Alex Tabarrok on February 4, 2008 at 03:48 PM in Economics | Permalink
Comments
How much longer before most people realize that anit-trust laws are the most counter productive laws on the books? Anti-trust laws effectively criminalize doing business, trying to increase your market share, and thus bring the whole range of business activity under the government's yoke, so they can extract contributions and cater to rival businesses. For anyone to still believe the idea that these rules have anything to do with consumer protection takes a massive amount of trust in the benevolence of government.
Posted by: scott clark at Feb 4, 2008 6:19:26 PM
With Yahoo's stock price near what microsoft offered, investors believe the deal will go through.
Posted by: thehova at Feb 4, 2008 6:53:16 PM
Is this obvious? It seems to me that many IT markets have been winner-takes-all. If that's true, we may face a choice between a market with niches dominated by Google and Microsoft and a market where Microsoft dominated most of the major niches. In that case, the merger could be bad for Google and worse for competition.
Tom
Posted by: Tom G. at Feb 4, 2008 7:42:31 PM
tom, it is important to keep a distinction between competition, the degree of effort expended by a firm in an effort to increase marketshare by serving customers, and the competition, all the firms a company competes against. Even if Microsoft dominated all major niches, this would not convince me of a monopoly. Even potential competition can be as effective as actual competition in keeping firms innovating and producing. And even if Microsoft dominated all major niches, there will still be niches that it does not dominate, niches that haven't even been carved out yet, that a new firm could develop, exploit, and grow into a major firm and then rival Microsoft. Which is kinda what Google did. This degree of competition in the industry is only brought down by increased government intervention and regulation. Needing to have compliance departments and government relations departments then becomes another barrier to entry stifling potential competition.
It is important not to get hung up on the number of firms in the industry when determining monopoly status. It should be more about productive outputs, innovations, pricing, and firm behavior, it should be about outcomes.
Posted by: scott clark at Feb 4, 2008 8:24:37 PM
It's odd that you link to that Antitrust Protectionism PDF that mentions the MSFT/Netscape case given that, in retrospect, it's crystal clear that Microsoft's bundling of Internet Explorer was an anti-competitive move that basically killed innovation in the browser market for a good 6 to 7 years.
Just look at how badly IE stagnated as soon as it became clear that MSFT had "won" the browser war - active development on the product basically ceased. It's still far behind its competitors today (coasting on bundling and a massive installed user base), but at least it's being actively developed. If Firefox and Safari hadn't come into existence, it's likely that IE would still be stuck at version 6 (at least in terms of features, if not in name).
As for MSFT/YHOO - it's clearly a move of desperation on MSFT's part (their online services strategy is entering its 10th year of failure now, so merging is basically a "Hail Mary" play to try to make something happen). But GOOG's response tells us nothing one way or the other about anything, because it's always GOOG's dominant strategy to object. If GOOG thinks the merger will increase competition, then clearly it's optimal to object. But even if GOOG doesn't think the merger will increase competition, they should still object just to tie up YHOO and MSFT in additional red tape while the courts sort it out.
Posted by: mike at Feb 4, 2008 8:39:13 PM
Google is just returning the "favor": nothing more, nothing less. Yahoo is a has-been search engine and its mail features are second rate: Google knows this.
Posted by: meter at Feb 4, 2008 8:44:52 PM
Upon further consideration, Alex, are you being sarcastic, or are you just being incredibly dense? Perhaps an example will explain the extreme fallacy in your logic:
Best evidence that granting the entire search engine market to me will increase competition:
Publicly, Google, Yahoo, and Microsoft came out against a deal that would legally grant 100% of the search engine market to a single unnamed private individual living in the San Francisco Bay Area, contending in a statement this legislation, proposed by that individual on Friday, would pose threats to competition that need to be examined by policy makers around the world.
Now, the YHOO/MSFT merger may or may not increase competition (I tend to guess not, since MSFT's online division is extremely incompetent, and any attempt to merge it with YHOO's totally different technology will be a multiyear headache, so they'll probably just both die even faster). But the argument you give tells us absolutely nothing - you might as well just have said "I think the merger will increase competition because the sky is blue!"
Posted by: mike again at Feb 4, 2008 8:46:54 PM
The post and earlier letter really don't give any suggestions that would help someone actually decide whether an anti-trust action would be justified or not in such cases, so some quantifications would be useful. Right now the level of argument is at about the level of the WSJ editorial page, but I know Alex can do better than this.
Posted by: jonm at Feb 4, 2008 9:24:56 PM
The sad thing is that you really have to have some pretty good computer science chops to understand the tragedy of the MS monopoly.
It takes some architecture skills to understand how their infrastructure limits the shape of innovations.
MS can say they innovate (and they do, within their framework), but most people don't see their framework.
(For the comp sci folks in the audience, this might be a blast from the past ... Rob Pike on why systems software research does is irrelevant. (PDF))
Posted by: odograph at Feb 4, 2008 10:18:38 PM
Out of curiosity, what would a consumer based antitrust movement look like? I doubt that there were street protests each time Standard Oil crushed and absorbed a competitor. Indeed customers were happy to be paying less for oil while the crushing was going on (only to have the price go up once the competitors were out of the way.)
If you are against _all_ antitrust proceedings in principle, I can respect that. Instead you seem to be singling out the modern batch on the basis that it is just sour grapes from competitors, though I would guess that is the historical norm (and intent of the framers.)
I'd also like to point out that Microsoft's position of strength is almost entirely based on the government granted monopoly of Copyright (with some patents and now DMCA nastiness.) so it is disingenuous to shout "laissez fair" too loudly.
There is no technical reason why the various instant messaging systems (AIM, ICQ, MSN Messenger, Yahoo! Messenger, Gmail chat, etc.) can't interoperate except that the lawyers will be unleashed on anyone who dares to interoperate uninvited. If MS pledged that they would allow interoperation between MSN+Yahoo Messenger and other systems, Google would feel a lot better about blessing the merger. They sure as heck aren't worried about competing on 'search'.
Posted by: Ed C at Feb 4, 2008 10:20:24 PM
Ed C,
All the data I've seen shows that during the existence of Standard Oil, output rose steadily and price continued to fall, firms and refiners that couldn't remain profitable in that environment were bought out and those assets were then back to being productively employed. At the same time, Standard Oil was spending money on research and development to make new products and find new uses for oil. If their competitive advantage came from being able to negotiate with rail roads, that still doesn't change the fact that this was the organization that was able to get the most value out of a barrel of oil. Same with Alcoa in the 50's, they continued to increase output year after year, and the price of Aluminum continued to drop. And its true that there were no other aluminum producers that could match alcoa, and Alcoa wasn't worried about other aluminum producers, but they were worried about plastics, so Alcoa continued to get more efficient and more productive. They got broken up too.
There is no doubt that MSFT knows how to work the patent and copyright system, and there is no doubt that real reforms are needed in the digital age, but just because one government law package is screwy doesn't mean that another screwy government law package can unscrewy it.
Posted by: scott clark at Feb 4, 2008 11:09:19 PM
Maybe Google thinks that if it doesn't oppose the deal, regulators would thereby conclude that the deal will reduce competition.
But what if the regulators think this is what Google thinks?
Posted by: Tyrone Slothrop at Feb 4, 2008 11:12:32 PM
I agree with Scott.
"It is important not to get hung up on the number of firms in the industry when determining monopoly status. It should be more about productive outputs, innovations, pricing, and firm behavior, it should be about outcomes."
Is the consumer suffering? That's the question to ask before any government intervention takes place. I like the point that several niches exist where Microsoft fails to perform. With the exception of its XBox Live and Windows Live (which both pale in comparison), what online success can Microsoft offer in comparison to Google. Google no doubt knows this and could have brought about this accusation just to slow down the process. Microsoft in the past did have monopolistic control over several niches of the industry and in a lot instances still does, but not all are controlled by the beast.
My biggest worries however are mentioned in the article:
“The potential concern would be that Microsoft, if it acquires Yahoo, could do on the Internet what it did in the personal computer world — make technical standards more Microsoft-centric and steer consumers to its products,” said Stephen D. Houck, a lawyer representing the states involved in the consent decree against Microsoft.
Perhaps this isn't the case...
Let's not forget to give credit to the consumer. Firefox was created as a response to Internet Explorer's inadequicies and consumers used the product. Mac seems to be gaining more and more of a foothold. And how can we forget all the attacks against Microsoft that scared so many people into considering Linux as an operating system?
Posted by: Jarrod C at Feb 4, 2008 11:15:41 PM
scott clark,
Nobody disagrees that Stanard Oil reduced costs and increased efficiency, but they were broken up, in part, because they were perceived to have raised prices in areas where they had safely run the competition out of business. I have no position on whether either Standard Oil or the government were in the right. I was merely trying to learn whether Alex is against antitrust laws in general or only when a case is backed by competitors which is how the link colored things. My post was not pro- (or anti-) antitrust law...I just have no sympathy for MS in this case.
Posted by: Ed C at Feb 4, 2008 11:41:01 PM
As a computer science guy, I like a lot of what Microsoft does. They've done well with both XBoxes (although the 360's red light problem was an issue, it was no worse than the PS2's disk read problem). The Windows 2000 and XP operating systems were very stable and great for the average user when compared to the alternatives (which were pretty terrible). Vista didn't launch well at all, which cost Microsoft market share. I still don't use Vista; I have no reason to.
Lets face it, OS/2 was terrible for your average user. Linux is still terrible for your average user (I don't want my mom using a case-sensitive file system). Competing operating systems for the desktop market would, in my opinion, also be terrible. Sure they can try to conform to the same specs and have compatible software, but we all know how well many supposedly compatible products work together. Maybe today, with more people doing test-driven development and the overall higher quality of software, we could see some things work. But 10 years ago, I'm not so sure. Your average user doesn't want to mess with the issues associated with running different sorts of software on different distributions of an OS.
The fact that there may not be much innovation in OSes and file systems is great and all, but the market doesn't serve the innovator necessarily. The market serves the customer, and your average customer wants a stable operating system with a wide range of available software. He doesn't give a damn about microkernels or journaled file systems; he just wants things to work. Maybe certain innovations help, but I doubt they are as preferred as a mature, stable platform. Gamers, to some extent, like innovation because of their high-performance requirements, but much of that innovation is channeled into consoles which are not bound by the less flexible specifications of PCs.
I'm not trying to defend Microsoft's use of patents or copyrights; I don't know enough about their actions to do that. I don't think Microsoft does a very good job on a lot of things, and their reputation will probably come back to bite them when other OS alternatives become viable. I am trying to defend their near-monopoly in desktop operating systems because I believe that is what serves the consumer best. Don't listen to us computer science guys when we think we know what the market should look like; only the consumer knows that.
At some point I think you just have to trust that the market process has produced a desirable outcome. You shouldn't have to built a swimming pool to prove that a bowling ball sinks, and every business with a lot of market share shouldn't have to prove that things could theoretically be better if it were smaller.
Posted by: Grant at Feb 5, 2008 12:57:26 AM
Grant makes an extremely important point in answer to Mike, above. Mike is entirely correct that IE bundling killed Netscape, resulting many years of browser stagnation. But it doesn't follow that consumers were harmed. Indeed the prima facie explanation of those events is that consumers prefered getting a simple, static, standardized browser for free over a having a choice of feature-rich, evolving, and incompatible browsers that they had to pay for. What the geeks want isn't always (in fact, usually isn't) what the mass of computer users want!
Microsoft's compulsory OEM licensing and discriminatory pricing strategies did harm consumers and were a legitimate target of anti-trust action. But bundling their browser with their OS at no additional charge was an entirely legitimate, competitive tactic that certainly harmed Netscape but certainly benefited consumers.
Posted by: David Wright at Feb 5, 2008 7:10:52 AM
Isn't Grant's argument that the market would favor the first commercial operating system to gain critical mass?
That's exactly what happened.
Network Effects are very strong in computer science, and especially operating systems. And if you really understand the guts of OSes, you can imagine all of the OSes you never saw "built out" or polished for the customer because ... the niche was filled.
To name just one, Distributed Operating Systems promised a system where backup was automatic, and power grew as each new computer was added to your office (or home). Everyone got the cumulative power of the installation, without the rigid distinction of "pc" and "file server" etc.
A few examples were built, and they worked well, but they faded when they couldn't leap from the university.
History is full of "lock in" of course. You may have heard that the screw-in compact fluorescents we buy share the same thread design as an ancient kerosene can. The shape limits their innovation. How did they get that shape? They inherited it from the incandescent bulb. And when a guy in Edison's lab was making the first incandescent he found it easier to cut up a kerosene can he had there, and re-use the cap and spout, than make a set of threads from scratch.
As an engineer I'd like a little more of a "period of flux" for good ideas to percolate before lock-in and infrastructure is decided.
I think in OSes we decided WAY too soon.
Posted by: odograph at Feb 5, 2008 7:53:35 AM
Microsoft's bundling of Internet Explorer was an anti-competitive move that basically killed innovation in the browser market for a good 6 to 7 years.
No. Firefox is a direct descendant of Netscape and Mozilla -- Netscape never went away (though the name changed). And the Opera browser has been under continuous development since 1994. There was never a time when bundling Internet Explorer killed off the competition. And Internet Explorer is still bundled, as is Safari as is Opera on many non-computer devices. In fact, bundling is standard procedure -- all internet capable computers, phones, and other devices come with browsers.
Trying to sell such an internet-capable operating system or device without a browser would be as absurd as trying to sell an automobile without a steering wheel in order to protect the manufacturers of 3rd-party wheels.
Posted by: Slocum at Feb 5, 2008 8:02:14 AM
"Trying to sell such an internet-capable operating system or device without a browser would be as absurd as trying to sell an automobile without a steering wheel in order to protect the manufacturers of 3rd-party wheels."
We computer scientists sometimes make the pedantic point that an "operating system" does not in fact include any applications. It is the system that allows applications to operate.
Now I get your point totally that a computer system sold today is expected to do certain things.
And it is convenient for someone to bundle all that stuff on one CD. It might have been the computer maker (Dell or HP or Apple) or it might have been a software vendor.
Today Apple does it all, and HP/Dell negotiate a bit with MS about what they can put on their PCs. (A point of the anti-trust case was that some PC manufacturers wanted to bundle Netscape and MS used OS-monopoly to fight it.)
So sure ... expect your computer to do a bunch when you open the box ... but be aware that the consumer idea of an "operating system" is actually a large bundle of operating system, drivers, applications, etc., etc.
Posted by: odograph at Feb 5, 2008 8:12:13 AM
BTW, I hope plain folk are aware that when MS said Explorer was not separable from Windows they lied, totally.
And then they went back and violated every rule of good computer programming to make it true. It was like bolting the refrigerator to the floor of the kitchen and then saying that the architecture required it.
They were better off (legally) with a bad program that they could claim was inseparable.
Posted by: odograph at Feb 5, 2008 8:16:47 AM
We computer scientists sometimes make the pedantic point that an "operating system" does not in fact include any applications. It is the system that allows applications to operate.
Pfft -- even in computer science there's no absolute, objective line between what's part of the operating system and what's an application. In your view is the windowing system part of the operating system? Is the file browser? How about the command-line shell? If not, should all of those be unbundled as well?
In any case, that's all irrelevant to the question of what an operating system is as a *product*, and that's what matters for this discussion.
BTW, I hope plain folk are aware that when MS said Explorer was not separable from Windows they lied, totally.
They didn't. The reliance on the Internet Explorer HTML rendering engine for the help system or as a component with public APIs that could be embedded in third party applications, for example, was perfectly reasonable. Now, it's possible to include that HTML rendering engine in the package but leave out browser shell that contains the HTML renderer. It's possible, but dumb. And even if you do that, you've included most of the browser -- you've just left out the relatively trivial container for it.
Posted by: Slocum at Feb 5, 2008 9:51:57 AM
Friedman says monopolies can't exist without government collusion. People eventually find ways around them.
Posted by: jorod at Feb 5, 2008 10:05:36 AM
A funny thing about search engines is: you could write a better one today, show it to the world tomorrow, and have the whole freaking world using it the day after that.
There's not a more unstable thing to monopolize on the planet.
Cheers,
Posted by: Michael Giesbrecht at Feb 5, 2008 10:27:28 AM
jorod...
Ah ha - so Friedman said it so its true? Interesting little religion you got there.
Posted by: reason at Feb 5, 2008 10:36:53 AM
So Michael Giesbrecht -- why haven't you done it?
Posted by: reason at Feb 5, 2008 10:38:20 AM
Slocum, you are moving the ball a bit when you choose a particular component (windowing system) that is viewed differently. I see that windowing is called out in this comp sci dictionary, whereas "the browser" and other applications are not.
On "they lied", I'd think the video tape they faked before federal court is pretty solid proof.
Posted by: odograph at Feb 5, 2008 10:58:28 AM
"But after a brief retort from Bristol's chief attorney Patrick Lynch, who refuted Tulchin's claims, the judge - looking thoughtful - said it was her decision not to dismiss the case and that she would reserve judgment until the court had heard Microsoft's side of the story. With that, she ushered in Redmond's first witness, Jim Allchin. Compared to his grueling couple of days on the stand during the Washington trial, where the government's attorney, David Boies, had shown that an Allchin-approved video testimony had actually been a fake, day one in Connecticut was a breeze. Bristol is promising a lively cross examination tomorrow."
Posted by: odograph at Feb 5, 2008 11:02:49 AM
Let's not romanticize pre-Mozilla Netscape too much, either. It took the Mozilla folks a long time to make Mozilla competitive with IE -- the 1.0 release is a fair estimate of when they pulled ahead. A lot of that monopoly period came from the competition not being particularly good.
Posted by: Zach at Feb 5, 2008 11:12:12 AM
On "they lied", I'd think the video tape they faked before federal court is pretty solid proof.
If Microsoft was trying to claim removing the browser was going to impair performance, that's absurd. But removing the HTML rendering would have screwed up other functions, not because Microsoft was doing shady things, but because an HTML renderer is an extremely useful component (for help, for rendering HTML-formatted emails, etc). And not removing the HTML rendering engine would have meant the browser was effectively there, but hidden.
I'm not defending Microsoft's *conduct* but rather pointing out that, technically speaking, the idea of removing the browser and the ability to display HTML from Windows would have been ridiculous (indeed, every other general purpose operating system and every internet-capable computer or other device bundles these things into the basic package).
Posted by: Slocum at Feb 5, 2008 11:15:29 AM
I think the right way to write an OS is with pluggable services, for things like HTML, and later XML. You have a bad architecture if you can't plug, and unplug, those (as fashions change and better inventions come along).
Now the case hinged on MS charging two prices for their "OS." You paid more if you unplugged IE and put in another browser, and you paid less if you kept it plugged. There is where it comes back to Alex's original subject and the value of anti-trust regulation. They ran afoul of anti-trust law by leveraging with their OS.
But. I'll stand on my comp sci opinion that if you aren't pluggable you aren't doing good design.
Posted by: odograph at Feb 5, 2008 11:28:45 AM
History is full of "lock in" of course.
Except that no one has ever identified even one example of a superior good being unable to displace an inferior one due to 'lock-in'.
Which reminds me, there's a new paper by the bundled economists Liebowitz and Margolis. I downloaded it from the Stanford Law site.
Posted by: Patrick R. Sullivan at Feb 5, 2008 1:14:26 PM
"Except that no one has ever identified even one example of a superior good being unable to displace an inferior one due to 'lock-in'."
That is a very interesting assertion. I'll have to think about it a bit, but my first reaction is that we can't see the dog that didn't bark ... if you get what I mean.
(Aren't U-shaped CF bulbs better than spirals?)
Posted by: odograph at Feb 5, 2008 1:33:48 PM
I thought of my answer. "Lock-in" as I am using it here is really a dominant design, or a de-facto standard. It's just one that arises by happenstance (or market power) as opposed to by conscious agreement.
Surely there is a body of work on the trade-offs between standards and innovation? The good old NTSC standard allowed an American television industry to grow, but it restricted the ways it could grow.
Windows is similar to NTSC in this regard. Once in place, it allowed the PC industry to grow, but it restricted the ways it could grow.
For more on this see James Utterbeck and Clayton M Christensen on innovation, or David Bank's book "Breaking Windows":
on Page 71: " ... Inside Microsoft, such sacrifices were known as paying the "strategy tax." "I agree that making sure applications are primarily on Windows is something we have lost site [sic] of," Gates wrote ... "
Posted by: odograph at Feb 5, 2008 1:58:02 PM
There is indeed a body of work on standards and innovation. A large part of it written by Stan Liebowitz and Steve Margoliz. I linked to their latest paper above.
Most every paper they've published is available online. Including their famous paper on lock-in.
Posted by: Patrick R. Sullivan at Feb 5, 2008 2:10:35 PM
odograph,
That is a very interesting assertion. I'll have to think about it a bit, but my
first reaction is that we can't see the dog that didn't bark ... if you get what I mean.
I agree, but it doesn't necessarily follow from that that we should break up businesses that DO get "locked-in". The theory that maybe some new products would come out if you destroy the market share of the current one seems like a pretty shaky reason to break up a company with reasonable customer satisfaction. When those new shiny OSes you mentioned never saw the light of day, your average user was still trying to figure out how to send an email.
I don't think its clear that mitigating strong network effects is necessarily good for consumers or producers, although I admit to not having read any papers on the subject.
If you want innovation, I think you just have to look beyond the desktop sector (although I'm not familiar enough with OS technology to know if Linux and SunOS are really that innovative).
Posted by: Grant at Feb 5, 2008 2:19:38 PM
Let's not romanticize pre-Mozilla Netscape too much, either. It took the Mozilla folks a long time to make Mozilla competitive with IE -- the 1.0 release is a fair estimate of when they pulled ahead. A lot of that monopoly period came from the competition not being particularly good.
It's not entirely fair to blame Mozilla's doldrums on the Netscape-era code base. The Mozilla team made the controversial decision to do a near-total rewrite of the core browser technology. That, combined with some unfortunate political disputes, prevented a clearly superior Mozilla-based browser (arguably Firefox 1.0) from appearing for 4-5 years. Meanwhile, many users continued to prefer Netscape 6.0, even as the commercial browser market stagnated.
Posted by: Chris Conway at Feb 5, 2008 2:34:06 PM
Trying to sell such an internet-capable operating system or device without a browser would be as absurd as trying to sell an automobile without a steering wheel in order to protect the manufacturers of 3rd-party wheels.
So you agree that bundling is important and that MSFT leveraged its Windows monopoly to gain an unfair advantage in the browser market. Good.
There was never a time when bundling Internet Explorer killed off the competition.
If MSFT considered the rotting corpse of Netscape (which, incidentally, was not called Firefox back then - it was called Phoenix and Firebird) to be serious competition, why did they effectively cease active development of IE?
More generally, you can't claim that bundling isn't an important competitive advantage unless you want to claim that Bill Gates is a complete moron. Because if bundling isn't an important competitive advantage, then why would he drag his company through years of lawsuits and billions of dollars of settlements when he could have defused the primary government argument by simply unbundling the damn browser?
It's one thing to argue that there are costs and benefits to anti-trust regulation and that you think the costs outweigh the benefits - that I can respect. But it's absurd to take the view that it's ALL costs, there are NO benefits, and it could NEVER be the case that a company is leveraging its monopoly to eliminate competition in other sectors (even when it obviously is). That's just crazy fanaticism.
Posted by: mike at Feb 5, 2008 2:56:21 PM
Except that no one has ever identified even one example of a superior good being unable to displace an inferior one due to 'lock-in'.
Because of course it's tautologically true that if the good was not replaced, then obviously the good trying to replace it was inferior. Wow, I'm so impressed.
Posted by: mike at Feb 5, 2008 3:01:05 PM
The Windows 2000 and XP operating systems were very stable and great for the average user when compared to the alternatives (which were pretty terrible). Vista didn't launch well at all, which cost Microsoft market share. I still don't use Vista; I have no reason to.
Lets face it, OS/2 was terrible for your average user. Linux is still terrible for your average user (I don't want my mom using a case-sensitive file system). Competing operating systems for the desktop market would, in my opinion, also be terrible.
OS/2 was vastly superior to Microsoft's primary operating systems at the time (DOS/win31, win95 and win98). It was Microsoft's monopoly power, Microsoft's profits and market network affects that prevented a much superior product from prevailiong. Even today eComStation, OS/2's successor, it much more stable than Microsoft's OS. www.ecomstation.com
Microsoft continues to skew the market to lockin Microsoft products. Open Source software and Linux would be much more widely used if not for numerous obstacles Microsoft has thrown out. Microsoft's purchase of Yahoo is a twofer. If helps it compete with Google and since Yahoo owns the Open Source, Zimbra email client server that competes with Microsoft Outlook/Exchange, Microsoft can kill another Open Source competitor.
Posted by: james at Feb 5, 2008 3:56:24 PM
Grant wrote:
"I agree, but it doesn't necessarily follow from that that we should break up businesses that DO get "locked-in".
Yes, I am ambivalent about what I'd even do if I held immediate power to do this to MS right now. I think the proper number of pieces is 4 (os, application, developer tools, extras inc xbox), but I'm it matters now.
Really the interesting point in time is long past, and even back then the most appropriate response might not have been a break-up.
Around the same time as MS gained a desktop monopoly buyers in the corporate data centers shifted to "open systems." This is something that predates and does not exactly match "open source." ("Open source" takes "open systems" even further.)
With open systems you had a standards based environment which let customers mix and match between a number of suppliers. Innovation was possible, both within the latitude of those open standards and by the ability to create new standards.
The UNIX systems, like IBM's AIX, Sun's Solaris, HP's HP-UX, and many more, plugged an played. (Linux "plugs into" this framework.)
... so I wonder what would have happened if the government had just been more shrewd in it's own purchasing, and it's on support of open standards. Would that have been enough to allow new entrants, even without explicit controls on MS?
Posted by: odograph at Feb 5, 2008 3:59:48 PM
For clarity:
"Yes, I am ambivalent about what I'd [] do if I held immediate power to [break up] MS right now. I think the proper number of pieces is 4 (os, application, developer tools, extras inc xbox), but I'm [not sure] it matters [at this late date]."
Better, standards based, computing is possible. And interestingly this is the battleground for MS (and IBM).
Posted by: odograph at Feb 5, 2008 4:07:09 PM
Because of course it's tautologically true that if the good was not replaced, then obviously the good trying to replace it was inferior. Wow, I'm so impressed.
People have been offering examples for over 20 years, and not one has withstood scrutiny; the QWERTY keyboard, Betamax video, MS/DOS, the internal combustion engine....
There is an extensive professional literature available. This isn't just idle speculation on a blog's comments section.
Posted by: Patrick R. Sullivan at Feb 5, 2008 4:17:18 PM
QNX was an excellent MS/DOS alternative. It did real-time multitasking years before MS, and with a smaller footprint. I'd be very surprised if the "scrutinizers" considered it though.
Basically Patrick, I think it is reasonable to expect that some standards (de-facto and intensional) were as good as they could be, but that it begs incredulity to take an extreme - that either "all" or "none" of them were substandard.
The world does not contain that kind of absolute.
Posted by: odograph at Feb 5, 2008 4:44:43 PM
People have been offering examples for over 20 years, and not one has withstood scrutiny; the QWERTY keyboard, Betamax video, MS/DOS, the internal combustion engine....
"Inferiority" is an inherently subjective measure - we define Good B as inferior to Good A for consumer i at some moment in time if i prefers A over B at the same price at that moment. So of course the superior good always wins. Duh.
In at least three of the examples you cite, once can easily make the case that the alternative is better on purely objective grounds. But that doesn't mean anything to an economist because we don't know how objective measures map into preferences, short of simply observing whether a consumer chooses one good or the other.
There is an extensive professional literature available. This isn't just idle speculation on a blog's comments section.
There is also an extensive professional literature documenting the superiority of the Dvorak keyboard, Betamax, etc. Whoop dee do.
Posted by: mike at Feb 5, 2008 4:45:59 PM
Because if bundling isn't an important competitive advantage, then why would he drag his company through years of lawsuits and billions of dollars of settlements when he could have defused the primary government argument by simply unbundling the damn browser?
Why did MS fight so hard? Seems pretty obvious. First because a web browser is a fundamental element of a desktop operating system. It wouldn't be too much of an exaggeration to say that the rest of the crap exists to enable the user to connect to the internet and browse the web. If you had a computing device that could do *only* that it would still be pretty useful. But a desktop computing device without that ability would be severely crippled.
Second he fought so hard because it was clear that the browser was becoming, if not *the* application platform, then at least A critically important application platform. Microsoft, you may have noticed, is in the platform business.
Lastly, MS fought so hard because of the precedent it would set. If MS was not able to add new categories of functionality to its operating system as the computing world evolved, it's operating system would increasingly become a boring mass of software plumbing (which, BTW, is just what MS's competitors hoped for -- and continue to hope for -- in pushing for all the anti-trust activity).
Posted by: Slocum at Feb 5, 2008 4:46:41 PM
james,
OS/2 was vastly superior to Microsoft's primary operating systems at the time (DOS/win31, win95 and win98). It was Microsoft's monopoly power, Microsoft's profits and market network affects that prevented a much superior product from prevailiong. Even today eComStation, OS/2's successor, it much more stable than Microsoft's OS. www.ecomstation.com
What was so much better about OS/2 for the average user? The people I know who tried it were gadget geeks, and hated it for anything other than a novelty (dual-booting with win95; OS/2 was far slower). I agree Win95 wasn't very good, but 98 was a lot better, and 2000/XP are extremely stable. I admit to not having a lot of experience running Windows vs. Linux servers, but in my experience XP SP2 is every bit as stable as as stripped Gentoo distros. Stability hasn't been an issue for me at all, but I didn't upgrade to Vista when Microsoft wished me to either.
Where I went to school, we had computer labs full of Windows XP SP2 machines and Sparcs running SunOS. We'd sit down to code Java in Eclipse and use Firefox (all open source, runs the same on both platforms). Most people sat in front of the Windows PCs.
Microsoft continues to skew the market to lockin Microsoft products. Open Source software and Linux would be much more widely used if not for numerous obstacles Microsoft has thrown out. Microsoft's purchase of Yahoo is a twofer. If helps it compete with Google and since Yahoo owns the Open Source, Zimbra email client server that competes with Microsoft Outlook/Exchange, Microsoft can kill another Open Source competitor.Well, open source developers do not operate under the same incentives as proprietary software developers do. The later have strong incentives to serve their customers, while the former work to serve themselves while helping other developers like themselves at the same time. Open source is good for large collaborative projects between businesses, hobbyists, and anyone who needs to improve software for his own use. But its not good for serving the end consumer, so its rarely easy enough for someone's mom to use.
Microsoft can only create a barrier to the standards of free software. I'm not sure what license Zimbra is distributed under, but if its truly open source Microsoft can't do anything but disallow usage of the trademark. I agree with odograph that open standards are the future. I just don't feel that back when Microsoft got "locked in" we could have done a lot better (and I'm still not convinced the average user would be better served by more competition in some niches). I also believe that if Microsoft continues fighting against the open standard tide, the market will punish them in a very bad way. However, currently I don't think they are doing such a bad job competing against open standards. Yes they flex their networked muscles, but they also do try and innovate (Vista boasts numerous theoretical improvements over XP, for example).
Its all really pretty ironic, given that Microsoft and the IBM PC were the "open standard" compared to Apple. I'm sure once MSFT succumbs to open standards in OSes, we'll complain about some other standard which has gotten locked-in and needs to be opened up.
I certainly don't have any love for Microsoft, but their detractors do need to explain why their customer satisfaction numbers aren't abnormally low (or really even low at all, although they did take a dive with Vista).
Posted by: Grant at Feb 5, 2008 4:50:49 PM
"Why did MS fight so hard?"
Ah Slocum, you aren't evil enough ;-)
If MS had a real browser monopoly they could have forced a roll over to MS web servers (extending the monopoly again) and wiped out the UNIX market once and for all.
Posted by: odograph at Feb 5, 2008 5:07:09 PM
odograph,
I'd forgotten to mention, one of the instances where I think open standards are inferior to a "monopolized" standard is in graphics libraries. OpenGL and DirectX are both usable on modern 3D card. Both are generally installed with games or with video drivers. DirectX is updated often enough that MSFT's bundling it with windows doesn't seem to help. DirectX, however, is the preferred library for most graphics programming (although they taught us OpenGL in school). Some devs from EA told us they thought it was just that updating OpenGL standards was just too time-consuming. This fits with what I know of economics; I think transaction costs just impeded the process when compared to DirectX's autocratic development.
mike,
Sure, there are always better alternatives out there. But they have be significantly better to get a lot of people to change. A lot of people claim Dvorak keyboards are better than QWERTYs, but I don't think anyone has shown a really serious improvement (and any modern OS lets you switch between Dvorak and QWERTY with any keyboard).
Posted by: Grant at Feb 5, 2008 5:08:00 PM
(Beyond that though, you should focus on what the Windows-using computer companies could have delivered. They had options, and certainly would not have had to deliver "browserless" computers to their customers.)
Posted by: odograph at Feb 5, 2008 5:08:40 PM
Grant, I'm not a graphics guy so I'll accept all you say. I will point out an interaction though ... graphics cards are developed to excel at Windows games. And so they evolve within that framework. Truly different graphics frameworks are never going to get 1/1000th of the money for R&D.
Posted by: odograph at Feb 5, 2008 5:12:56 PM
OS/2 was vastly superior to Microsoft's primary operating systems at the time (DOS/win31, win95 and win98). It was Microsoft's monopoly power, Microsoft's profits and market network affects that prevented a much superior product from prevailiong.
No. OS/2 was originally a Microsoft product and then a Microsoft/IBM joint product. At the time Windows 3.0 was released, there were relatively few Windows applications, and Microsoft had nothing like monopoly power in the graphical interface computing market.
Windows 3.0 knocked out OS2 for the simple reason that it required far less memory to run at a time when computer memory was quite expensive. (For a short while, before the messy divorce, IBM and Microsoft jointly declared that OS2 was for computers with lots of memory and Windows 3.0 was for computers with limited memory).
Also, although there were not very many Windows applications, there were a few important ones (PageMaker, Excel). OS2 was different enough from Windows that it wasn't trivial to create OS2 versions -- so at the time Windows 3.0 was released, OS2 cost more to run and had fewer applications. OS2 also had worse device support and was initially tied to the 286 rather than 386 processor which, oddly enough, made it much worse for running DOS applications than Windows (which could run multiple virtual DOS sessions). Windows was certainly messier than OS/2, but it did more and ran on cheaper computers, and that was enough.
The "breakup" section of this wiki page is pretty good:
http://en.wikipedia.org/wiki/OS/2#Breakup
Posted by: Slocum at Feb 5, 2008 5:15:16 PM
"Also, although there were not very many Windows applications, there were a few important ones (PageMaker, Excel). OS2 was different enough from Windows that it wasn't trivial to create [alternate] versions"
That's the Network Effect in action. It was not all monopoly or anti-trust.
(It is arguable that a less evil MS would have achieved the same things, or better. Perhaps that is the Google perspective.)
Posted by: odograph at Feb 5, 2008 5:26:31 PM
odograph,
Certain cards (ATI's, if I remember correctly) tend to run better on OpenGL, while others tend to work more efficiently on DirectX. Many of the games I've seen use OpenGL (such as any using id's Doom 3 engine) run just as well as would be expected of a DirectX game. A few years ago, I know Linux OpenGL performance was behind Windows, but that may have changed. I don't think there is any evidence of Windows hampering the OpenGL standard, although you'd probably have to talk to graphics programmers to really know. The guys from EA I talked to were XBox 360 devs, so they may have been biased (incidentally, they also remarked on how much they hated developing for the Playstation 3).
It would be a massive mistake for MSFT to sabotage OpenGL. If they did, some gamers (who tend to be computer-savvy) might abandon Windows altogether. Although I really don't know if they could. Graphic languages seem to be collections of matrix operations. I don't think there are many ways to optimize that sort of thing for different OSes (beyond thread handling).
Posted by: Grant at Feb 5, 2008 5:29:52 PM
Grant, I wasn't speaking to that. MS doesn't need to worry about anything now because all card makers are targeting them as a primary market.
This is a different idea. If Kid X at University Y conceived of a whole new framework for graphics, but one that was incompatible with both OpenGL and DirectX (they both have network effects going for them) ... how much money could he get for doing silicon?
Certainly not enough to be competitive. He faces a catch-22 (similar to the OS writer's catch-22) in that he needs massive funding to prove his concept, but he can't have the funding until he proves his concept.
Posted by: odograph at Feb 5, 2008 5:56:04 PM
...once can easily make the case that the alternative is better on purely objective grounds. But that doesn't mean anything to an economist because we don't know how objective measures map into preferences, short of simply observing whether a consumer chooses one good or the other.
You are correct. However, the U.S. v. Microsoft fiasco was based exactly on that conceit; i.e. the government--relying on dubious scholarship--WAS ABLE to make that judgment.
What Liebowitz and Margolis showed was that the claims of the Paul David-Brian Arthur-Gary Reback axis were wrong. There were no examples of superior products 'locked out' due to network effects. And, that possibility, is the entirety of the theoretical case against Microsoft.
There is also an extensive professional literature documenting the superiority of the Dvorak keyboard, Betamax, etc. Whoop dee do.
If you'll bother to read the papers I linked to, you'll see that those papers have been exposed as exceptionally sloppy scholarship.
Posted by: Patrick R. Sullivan at Feb 5, 2008 6:01:31 PM
If Kid X at University Y conceived of a whole new framework for graphics, but one that was incompatible with both OpenGL and DirectX (they both have network effects going for them) ... how much money could he get for doing silicon?
How do you explain how we went from black and white television to color, or from vinyl 33-1/3 rpm records to CDs, from the VCR to DVD? There were network effects at work there too.
Posted by: Patrick R. Sullivan at Feb 5, 2008 6:08:22 PM
"How do you explain how we went from black and white television to color, or from vinyl 33-1/3 rpm records to CDs, from the VCR to DVD? There were network effects at work there too."
There seem to be some feature-barriers (probably a better way to say it) that force a change. Sure. Nothing is absolute.
As it happens I said to my peers a few years ago that Windows had the future until such time as we hit something that they couldn't do, some feature they could not import. So far they have been doing ok. The most recent hurtle has been video integration, but everyone has done that "good enough."
Posted by: odograph at Feb 5, 2008 6:18:49 PM
odograph, you're a long way from: 'History is full of "lock in" of course.'
Posted by: Patrick R. Sullivan at Feb 5, 2008 8:05:25 PM
Said the guy who did a http post over tcp/ip ... heck, powered by 110v.
There are no absolutes, and dominant designs can lose favor, but they enjoy a strong position. As long as they can do their job adequately they may not be displaced by something that is merely better.
Posted by: odograph at Feb 5, 2008 9:33:59 PM
odograph,
I get your meaning. But a truly different hardware architecture on PCs would be a hard sell for many reasons other than Microsoft. Typically things like that happen on console platforms (although it made for a rough start for the PS3).
I agree with him that history is full of lock-ins... I just don't think they are so bad.
Posted by: Grant at Feb 6, 2008 12:29:19 AM
I'm no comp-sci geek (I type with two fingers), but I think of Windows the way I think of electrical outlets and cable jacks. The benefit to the consumer comes from the uniformity that a monopoly provides. A million great software programs on the market does nothing for me if they are incompatible with my computer. I would argue that a uniform OS benefits most software developers by providing a stationary target and easy mass distribution.
Posted by: nick at Feb 8, 2008 1:03:33 PM
Posted by: 深圳翻译公司 at Feb 23, 2008 10:30:55 AM
