Lies, Damned lies, and IIPA/BSA/etc statistics.

One of the key tools used by lobby groups like the IIPA is statistics to try to "prove" that there is massive harm to them, and which must be remedied in the way they propose. An analysis of their statistics often indicates that their real target is competitors, not copyright infringers.

These studies need to be debunked as they have a large influence on governments who have outsourced this key policy tool to special interest groups. William Patry documents on his BLOG (Via Google Cache) how, lacking any investigative resources of its own, USTR uses figures given to it by IIPA. If you read transcripts from Canadian parliamentary committees studying "counterfeiting and piracy" or Copyright and patents, you will see that Canadian politicians are no better. It is frustrating to know how many amazing economists are working for Industry Canada and other departments, and yet Industry Committee never bothered to consult them when they are studying critical economic policy.

Those debunking these studies are not apologists for copyright infringement. I document how this infringement harms companies like my own more than the members BSA/ESA/etc. I am very interested to stop software copyright infringement, but believe that the policies promoted by the BSA/ESA/etc have an anti-competitive effect and do not reduce software copyright infringement.

[Read the rest of this entry on IT World Canada's BLOG, or continue below]

Software copyright infringement statistics

Since I am in the software sector, I will focus on the so-called "software piracy" study done by IDC for the Business Software Alliance (BSA) and their regional branches such as the self-called Canadian Alliance Against Software Theft (CAAST).

The BSA study uses a fairly simple methodology at its base, and then extrapolates other numbers from there. That methodology is documented within the study report (Start at page 13 of the 2007 report).

They calculate an estimate of consumer demand for software, and compare it to the various sources of software.

For demand they start with the number of computers shipped into a region, and use a multiplier to guess the per-machine estimate for different types of software.

There is then a number of sources of software:


  • BSA members: They have fairly accurate statistics of how many licenses of their own software are being used, either in boxed form or in various types of site licenses.
  • Non-BSA proprietary software vendors: IDC does other studies to estimate the percentage of the market which is BSA members vs non-BSA member.
  • Free/Libre and Open Source Software (FLOSS): IDC does studies to estimate the usage of Linux. They often bundle FLOSS, freeware and shareware which is flawed as the follow-on economic effects are different for each.
  • Infringing copies of BSA and non-BSA software: this is the 'unknown' that they are allegedly trying to determine.

Simple formula:
Infringement = (machines shipped) * (usage estimate multiplier) - (legal BSA) - (legal non-BSA) - (legal FLOSS)

The accuracy of their infringement number is very dependant on the accuracy of the input numbers. It turns out that the only numbers that are largely non-controversial are the number of machines shipped and the legal BSA number.

IDC is often critiqued in its Linux studies for only fully accounting for hardware which is shipped pre-installed with Linux, or boxed software sold by retailers, which represent the least popular ways to obtain Linux. All of the machines which I own, and most of the machines I manage for customers are running Linux, but the only one shipped with Linux was my XO laptop. Nothing I do is adequately accounted for by the IDC study, and it is similar for most of the FLOSS technical support people I know.

The IDC study claimed a 34% infringement number for Canada. If the IDC Linux study was off by 10% (which is highly probable), then that would mean the infringement rate was 24%. This is a massive error rate which puts the entire study into question. Without an independent study that looks at each of these factors objectively, something which the IDC has demonstrated it is not able to do, we simply don't know what the software copyright infringement numbers are.

The reality is that there is no objective way to know the number of people who are legally using FLOSS, freeware or shareware given there is no legal or moral obligation for end users to count or register this software. The fact that you do not have to count copies and keep expensive records is in fact one of the greatest advantages of using this type of software.

Derived numbers

The same study claimed that there was $784 million dollars lost in Canada in 2007 due to this infringement. That number is even more suspect than the inflated 34% number it is derived from.

One component of this number is royalties that would be leaving Canada and heading primarily to the United States as they are the recipient of approximately half of all worldwide royalties. While this is true of BSA and non-BSA member royalty-bearing proprietary software, this is not true for FLOSS where the pre-and-post-acquisition support is more likely to stay local where a Canadian support company will be servicing Canadian customers.

The study claims that for every dollar of software sold there is $1.25 in services. This is obviously a number that only applies to proprietary software given the royalty charged for FLOSS is $0, and yet the costs to evaluate (pre-acquisition), install, support and customize the software is not $0. One of the legitimate complaints about FLOSS is that these support costs are sometimes higher for FLOSS than for non-FLOSS given the often higher skilled support person demands higher wages. In this case money that would otherwise leave Canada in the form of royalties and other overhead would instead stay in Canada in the form of wages for skilled technical support people. This should obviously be considered a good thing for Canada, but largely shows up as a loss in the IDC study.

The study claims that there is an additional dollar in "channel" revenue, but this also depends on the distribution mechanism. More and more legitimate software is being sold through online retailers which do not have similar costs as brick-and-mortar retailers. Whether for legal software in all its various forms or infringing software, this trend away from traditional retail for software has the identical effect and thus this number should not be included in the study at all. FLOSS also tends to transfer lower-paying retail jobs into higher paying pre-and-post acquisition services, which is better for these skilled workers but not so great for retail channels.

What do the vendors really think?

While these studies are good for lobbying purposes, with recent studies highlighting WIPO treaty ratification and strong anti-circumvention (AKA: anti-competition) legislation, some key people have disclosed what they believe the real economic impact to be.

In March of last year then Microsoft business group president Jeff Raikes had the following to say at a Morgan Stanley Technology conference in San Francisco:

"If they're going to pirate somebody, we want it to be us rather than somebody else ...
We understand that in the long run the fundamental asset is the installed base of people who are using our products, ...
What you hope to do over time is convert them to licensing the software."

This means that there is a hierarchy of possibilities in Mr. Reikes mind:


  1. People use and pay for Microsoft software (a subset of BSA members in the IDC study)
  2. People use but do not pay for Microsoft software (a subset of infringement from the IDC study)
  3. People use competing software

(Note: Jeff Raikes plans to leave Microsoft at the end of September of this year. There is considerable movement at Microsoft recently. See The 25 Most Influential People at Microsoft . )

Microsoft is by far the most influential member of the BSA, and thus what benefits they see should be understood as key to the purpose of the multi-year IDC commissioned study.

While not explicitly mentioned, the worst case scenario for Microsoft is that people use FLOSS. FLOSS has a large philosophical barrier for people to get past. People instinctively value something that 'costs more' than something that costs less, suggesting that they consider a free and illegal copy of proprietary software to be more valuable than a free and legal FLOSS, without ever actually comparing features to find out what software would better serve their needs (provides more real value). Once people get over that philosophical barrier, however, they are going to evaluate software on a more level playing field and thus are far less likely to over estimate the value of or use Microsoft software.

This also exposes another issue, which is that infringement of Microsoft software copyright has a greater negative impact on competitors than on Microsoft. Software has a zero marginal cost, meaning that there is no cost to Microsoft of people unlawfully using their software other than potential lost revenue. From the above it should be obvious that the real long-term costs to Microsoft of people switching to competing software is far greater than people infringing Microsoft software.

If the only way that people could use BSA member software is if they paid full price for it, then this would drive cost conscious people over that philosophical barrier to investigate legally royalty-free FLOSS. This means that FLOSS vendors have a much greater interest to be "Against Software Theft" than the members of the almost Orwellian double-speak named "Canadian Alliance Against Software Theft".

Summary

When all these factors are combined I believe it suggests that the primary goal of these studies is to aid in lobbying for changes to the law that would benefit the lobby groups over their competitors.

A similar breakdown can be made of similar studies done by the entertainment industry members of the IIPA where they take the revenues from some bubble peak, subtract their current lower revenue, and declare the difference to be caused by their favorite lobbying scapegoat. They ignore changing market conditions and the massive entry of legal competition from creators who use competing methods of production, distribution, and funding than the incumbent industry associations. (See: Yochai Benkler's book The Wealth Of Networks).

This is a typical market dynamic discussed in books such as The Innovators Dilemma by Clayton M. Christensen. This book even predicts attempts by incumbents to manipulate governments into changing laws to privilege the incumbents over the new competition, causing a reduction in innovation and associated harm to the economy. This is something which governments need to become more aware of and avoid. This clearly requires governments ignore industry sponsored studies, such as the USTR fully replacing the the IIPA as a data source for their report.

See also: After 20 Years, Critics Question the BSA`s Real Motives.

(Note: Some links updated Aug 19, 2012)