Late last year the European Commission released a report calling for a new industrial policy to ignite report advocated government programs (a la Airbus) based on Free/Libre/Open Source Software (FLOSS). ACT analyzed the report with a series of blog posts reviewing each section. We found that the report’s analysis is not adequately supported by the data. Moreover, by advocating interventionist public policies the report’s recommendations may harm, not help, Europe’s overall ICT sector.

Overview of the FLOSS Report’s Major Claims

First, we discussed how the report overstated the breadth of popularity and depth of penetration FLOSS has attained. While FLOSS use is unquestionably growing, the report inflated claims of corporate adoption by conflating survey terms (“use” versus “adoption”), and by lumping together all software under any of the many different FLOSS licenses, even though there are important distinctions among these licenses.

Next, we described how the report wrongly assumes that a vast army of FLOSS developers are already in Europe, yearning to show the world what they can do. The report argues that more than three-fifths of the worldwide FLOSS developer community lives in the EU. But not all FLOSS developers are the same—most FLOSS code is generated by a handful of paid developers, not altruistic programmers working on their own time.

Do FLOSS programmers work faster than proprietary software developers? Maybe, as we said in our next posting—and there’s good reason to think that FLOSS programmers should work faster. FLOSS developers have focused much of their effort on writing software that imitates features and functionality in existing commercial software products. By doing so, FLOSS projects avoid important risks–limits in technology and non-acceptance by users–meaning less time is spent on market tests, dead-ends, and do-overs. Moreover, one has to question the long-term viability of an industrial policy based more on imitation than on genuine innovation.

The FLOSS report says that firms working on FLOSS are more financially successful than non-FLOSS firms. The report generously counts as “FLOSS contributors” any company that employed any of its resources on projects with FLOSS code. AT&T, for example, represented 12.5 percent of the revenue and 28 percent of the jobs the authors attribute to FLOSS contributors. But what’s the percentage of AT&T’s $47B revenue and 160,000 employees that are dedicated to open source projects? Probably close to nil, yet as we described in our blog post, these entire sums were included in the study’s analysis. When asked, the study authors declined to provide a list of the 158 firms that provided the data for their claim that FLOSSers are richer.

While the report claims that firms who FLOSS get richer, it goes on to say that developers who FLOSS will get poorer since they cost less than their non-FLOSS brethren. The report contends that FLOSS developers should be less expensive because almost any teenager can become a FLOSS programmer through informal learning on open source projects. However, as we analyzed, FLOSS really doesn’t promise a new and better way to learn how to write software. Programmers have always learned their craft through a combination of education, on-the-job training, and self-study, frequently guided by others with more experience.

How one learns to code is beside the point. The real issue is whether an emphasis on cheaper software developers will help grow Europe’s ICT sector. We doubt it. An emphasis on building an ICT sector around inexpensive labor will drive wages down, creating a race to the bottom for jobs that developers from other countries would rush to fill. The result is likely to be an increase in offshoring to China and India —not more ICT jobs in Europe.

Analysis of the Report’s Public Policy Strategies

The report makes no attempt to disguise its purpose: to convince European policymakers to favor FLOSS in their procurement decisions and other programs. Section 9 of the study—Trends, Scenarios and Public Policy Strategies—suggests a number of public policy programs to promote FLOSS:

  • Publicly-funded FLOSS research. All ICT-related research using public funds should generate software that’s openly and freely licensed under FLOSS licenses.
  • Treat FLOSS as a charity donation. Companies and individuals should be allowed to take a tax deduction for the value of any software released under a FLOSS license.
  • Educate students about FLOSS. Fund and facilitate measures that reward and facilitate FLOSS use by schools, teachers and students.
  • Force unbundling of hardware and software. Use European competition laws to stop firms from bundling software features and linking software with specific devices. This is no idle threat, as seen by previous European competition actions: Several European nations are saying that it’s illegal for Apple to restrict some iTunes content to be played only on Apple’s iPod device. And the EC ordered Microsoft to sell a version of Windows in Europe with Windows Media Player software code stripped out.
  • Governments should fund “non-commercial information exchanges” to create FLOSS-licensed open technology infrastructures that cater to perceived “prosumer” needs and without the immediate requirement of economic sustainability. Translation: Government-funded groups should decide what kind of ICT products consumers need, and these needs should be met by FLOSS producers.

These programs amount to industrial policy—along the lines of the Airbus program—for favoring FLOSS in the ICT sector. Although this strategy relies heavily on government procurement preferences, the above wish-list shows that it also includes direct subsidies and antitrust intervention.

The FLOSS authors anticipated criticism for advocating government preferences, and attempted to answer with this:

A criticism might be that “if FLOSS is indeed better and cheaper, no active policy in favor of FLOSS would be necessary.” In fact, this criticism is quite wrong, in that it naively supposes there are no network externalities and installed base effects.

But is FLOSS so disadvantaged that it requires government regulatory intervention and preferences? Not if you believe Section 6 of the report, which showed that FLOSS is widely adopted by business and wildly popular with programmers. Even if there were network externalities, recall that personal computers managed to overcome a firmly installed base of minicomputer and mainframes in the 1980s—without government assistance or intervention.

Furthermore, a policy favoring preferences presumes that any industrial policy will make a difference in the unpredictable and fast-changing world of information technology. But as we’ve said before, while markets are sometimes messy, it’s a far bigger mess when governments manipulate supply and demand to achieve a predetermined result. Moreover, the ICT industry has been a remarkable success in its own right and has boosted productivity in nearly all sectors, without the guiding hand of industrial policy like that being called for in the FLOSS study.

The FLOSS plan calls for more than just government and institutional preferences, going so far as to invoke antitrust law to stop bundling and product integration. But this strategy ignores the efficiencies and pro-consumer benefits of bundling and integration.

Web2.0, for instance, changes all the rules for software competition and network effects. A centralized service like Google’s on-demand desktop can instantly and continually deploy new features, easily integrate these features on the server side, and enable unprecedented collaboration among connected users. Indeed, Google seems to pursuing competitive differentiation through integration-based innovation. That makes sense because integration innovation reduces the customer’s cost of complexity by integrating disparate elements into a single easily manageable system. So it doesn’t make sense for Europe to wage war on firms who innovate by integrating.

As much as we think the FLOSS report is based on incoherent analysis of incomplete data, we readily acknowledge that there are many useful and truly innovative FLOSS projects. But we just can’t agree that governments should implement FLOSS preference policies, or use antitrust law as an industrial policy weapon against firms who innovate by integrating (see ACT’s paper on software feature integration & innovation). And we’d take the same stance if someone proposed an industrial policy to favor proprietary software over FLOSS.

Simply put, the marketplace—not governments—should pick winners and losers, especially in a dynamic and global market like information technology.