New FUD: Open Source Is "Economically Dangerous"

Friday, January 16 2004 @ 05:26 PM EST

Contributed by: PJ

Dr. Stefan Kooths, with the Department of Economics, the University of Muenster, Germany, argues that open source software is economically dangerous and governments shouldn't use it. He just gave a lecture on it yesterday, hosted by Microsoft Research in Cambridge, UK.


This is close enough to Darl's FUD that it just might be worth paying attention to this new argument. This likely represents the next FUD front in their war against Linux and particularly the GPL.

Here's the info on the talk, first, and then an email that brought it to my attention and which rebuts Kooths' concepts.

You can see the presentation, as PDF or Powerpoint, by going to the University of Muenster's web site to get the slides of yesterday'slecture (section "Presentations"). Dr. Kooths tells me that you can get the whole study itself from a link on the welcome page, but it doesn't work for me. Perhaps I am not using the "right" browser, but I don't see any such link.

UPDATE: Finally, the link to the study itself, thanks to Dirk.

Here is the info on the talk:

TITLE: The Economics of Open Source Software - Prospects, Pitfalls and Politics
SPEAKER: Dr Stefan Kooths
INSTITUTION: University of Muenster
HOST: Alexander Braendle, University Relations
DATE: 15 January 2004
TIME: 13:30 - 14:30
MEETING ROOM: Lecture Theatre
ADDRESS: Microsoft Research Ltd, 7 J J Thomson Avenue (Off Madingley Road), Cambridge

Open Source Software does not represent a suitable alternative to the commercial software market from an economic point of view, neither in terms of creating value-added nor in terms of economic efficiency. OSS does not create any new value-added potential, and offers only a fraction of the opportunities of the commercial market. The impact of OSS on sales and employment are therefore less than the effects of commercial software. Furthermore the de facto free availability ofGPL-licensed software, and hence the lack of a market price, have far-reaching economic consequences that are elaborated in the presentation. As far as packaged software is concerned its free availability very much limits the creation of profits, income, jobs or taxes. The loss of turnover in the area of software sales cannot be fully recovered with services linked to the software. So-called complementary OSS-business models work in the smaller customized software sector only. The incomes earned there are substitutive and not additional to those created in the commercial software sector. The lack of cost-reflecting prices for GPL-licensed standard software also hasconsequences for the market process as the pricing mechanism is associated with an important information and coordination function in a market economy. If there is no price, and hence no decisive guide figure for a market, it is, for example, more difficult to identify customerrequirements. Further problems can be identified when it comes to the allocation of resources, productivity-oriented factor compensation and incentives for innovations. The lower value-added potential and the reduced efficiency of coordination are weighty economic arguments. Theydemonstrate quite clearly that the promotion of open-source software cannot be an economically justifiable role for the state.

Here's the email I received bringing the lecture to my attention:

Hi PJ,

According to the abstract, Dr. Kooths will argue that promotion of open source software is economically dangerous, and "not economically justifiable for the state". The danger of Kooths'argument, in my view, is that not only does it suggest that governments should not actively promote the use of open source software, it suggests that they should go much further and oppose it and avoid using it, because of its *alleged* economic hazards. (Although Kooths does not draw this conclusion explicitly in his abstract, his ambiguous "the state should not promote OSS" conclusion sits oddly with his idea that OSS is some kind of economic cancer - to borrow thephrase of a certain Microsoft executive a few years ago.)

Secondly, this conclusion talks about "the state" as if _all_ states - irrespective of their circumstances and level of development - would overall benefit by supporting proprietary software (which quite often means supporting Microsoft, directly or indirectly).

Arguably, his logic is at its weakest when applied to nations for whom the largest long-term effect of excessive use of proprietary software on the national economy could be the diversion of large amounts of money offshore into the coffers of US companies such as Microsoft! It seems to bean all-too common weakness of a number of macro-economists (not all, of course, but quite a few) to draw conclusions based on abstract economic models without adequately considering the specific circumstances of a country - as, in another context, the former World Bank Chief Economist Joseph Stiglitz has cogently documented, in his book "Globalization and its Discontents" [1].

But even for countries that reap significant benefits - in at least a crude sense, in terms of jobs created, tax paid and such things - from having a large proprietary software industry, I don't think the argument is quite as one-sided as Dr. Kooths makes out.

Actually, at this juncture it should be pointed out that Microsoft paid no tax at all in 1999, despite reporting $12.3 *billion* in profits, and apparently this was all quite legal [2]. And as for jobs, the open source community has long argued that a vibrant IT economy can be sustainedaround open source, in terms of training, technical support, custom development, and other such services.

Dr. Kooths seems to be aware of this argument, because he refers to "packaged software" in his abstract - presumably meaning shrinkwrapped software such as Microsoft Office that you can buy in a store, as distinct from software developed from scratch or customised for a particularorganisation or group of organisations. But brittle, packaged, closed-source software has formed the minority in the world of software since the early days. The great majority of programmers in the world are employed maintaining or developing custom code, not writing code for Microsoft Office or the like. This would continue in a world with a healthier balance of open source and proprietary software.

Likewise, for the vast majority of companies and organisations - and governments - out there, software is predominantly a cost centre, not a profit centre. It is therefore a powerful argument that the overall economic effect of more use of open source code could be the freeingup of cash previously spent on proprietary software licenses (and the time and money frustratingly wasted on trying to cope with the hideous bug and interoperability problems of many commonproprietary apps) into lowering costs and increasing productivity elsewhere in the economy. Actual productivity, that is - rather than paying Microsoft for the ephemeral activity of issuingyou one more license.

Dr. Kooths' insinuation that open source software is by its nature as a development model, unresponsive to user requirements because of the "lack of a market price", is simply silly. There is no other word for it. Firstly, to the extent that developers are also users - which they are in cases like gcc - it works really quite well. Secondly, the responsiveness of developers to user requests is variable, and in many projects can be quite good - better even than paid commercial support from certain companies! Thirdly, as two counterexamples, I put forward:

- the Kroupware project - originally commissioned by a branch of the German government to build upon existing open source software to produce a groupware system to meet their precise needs;- and, also concerning groupware, the Open Source Application Foundation's recent announcement that it had obtained $2.75m additional funding from a consortium of US universities and a non-profit foundation to extend the OSAF's Chandler groupware project to meet the needs of large organisations like universities [3]. This is a promising and interesting model for funding (and thus speeding up) large and complex open source project developments where individual users or organisations might be unable to fund/create a project - but putting their heads together, they can co-fund it and all win. It reminds me of the scientists who've decided they don't like journal publishers charging monopoly rents for the privilege of being published in a journal, and then charging again for copies of the journal, and have decided to set up their own online journals. It's quite a similar case in some ways.

Granted, there are many cases where open source projects have not yet got around to meeting the needs of absolutely _all_ users - but this also holds true with Microsoft applications and special needs.

The difference, which favours open source, is that not-for-profit organisations, generous individuals etc. can fund or give their time to alteration projects which would otherwise be almost or completely uneconomic (especially to a company like Microsoft, for whom even amillion dollars is less than 0.1% of its bank balance, let alone its total assets) due to their smaller market size.

Finally, whether as *many* programmer jobs will be needed in future as open source software improves and matures is a debatable point, and I personally think it is a point that could be conceded. Certainly programmers will be needed for the forseeable future, but perhaps notso many.

But even if were conceded, it does not really make sense to ask all the other sectors of a national economy - and the government itself - to somehow forgoe the huge benefits of open source software - just to protect programmer jobs[4] (and, incidentally, the profits of companies likeMicrosoft, who would be the biggest beneficiaries of such an implausible move, let's not forget). It seems to me that Microsoft, and their old-guard supporters such as Forbes and Dr. Kooths, are playing King Canute.

Robin Green

[4] The tendency of technology to shrink employment in one sector - and often create employment in another, but not always - is a whole 'nother issue that I'm not going to get into here.