Calling Out the BSA on Its BS
The Business Software Alliance (BSA) is trying to pull wool over government officials' eyes by equating software piracy with tax losses. Pranesh Prakash points out how that argument lacks cogency, and that tax losses would be better averted if BSA's constituent companies just decided to pay full taxes in India.
In the past we have covered the Business Software Alliance's lack of rigour in their piracy statistics, and disconnect from their constituent members when it comes to opposing free and open source software. In reaction to the criticism they have received over the years, BSA has finally stopped equating lack of sales with losses. But now, they have started equating software piracy with tax losses.
How IDC thinks tax works
In a report prepared by International Data Corporation (IDC) for the Business Software Alliance (BSA), they note:
Substantial value in form of potential industry and tax revenues is lost to software piracy: The situation in India is not healthy with a software piracy rate of 65% in 2009 (more than six out of ten PC software programs installed in 2009 were not paid for). Only one-third of the overall PC software revenues are captured by the industry incumbents and the rest are lost to software piracy. Most of the unlicensed software use occurs in otherwise legal businesses installing the programs on more PCs than allowed by the licenses they have paid for. Consequently, in 2009, the state exchequer tax receipts loss was roughly US$866 million at the current piracy and employment levels, as the industry lost its otherwise legitimate share of revenues to piracy.
For this to be true, there must be two assumptions that are satisfied. First, those who are pirating software must not spend the money that they save by doing so on any other taxable activity. Second, the companies that would get the money if the software weren't pirated must pay the Indian government taxes. As we'll see, neither of these two assumptions are warranted.
The BSA-IDC report reasons as follows: Pirates don't pay taxes on the illegal software that they sell, so that is tax evasion and consequently a tax loss. It states:
Higher demand for legal software will result in higher flow of license volume through the supply chain, resulting in increase in volume of business transactions. Each transaction adds a certain percentage of the deal or value added to the state exchequer's coffers in the form of indirect tax revenue[...] Increase in demand will also result in increased employment. Consequently, revenues from direct taxes will be increased for the government, as employees join newly created high-paying jobs.
How tax actually works
That reasoning is flawed. The majority of software piracy in India happens through two methods: violation of software licence terms by using the software on more computers than it is licensed for; and pre-loading of illegal software by computer sellers. Those 'computer seller' pirates do not sell the software separately, but bundle it with the computer as an additional service. In other words, they don't charge for it in the first place. So, quite clearly, there is no tax evasion.
Despite there being no tax evasion, there is the possibility of tax loss for the state. That would happen when instead of doing taxable activity A with with their money, they do non-taxable activity B. Putting money in special government bonds instead of spending it on software, for instance, is one such instance. However, that is a strange, unwarranted assumption. People don't always put the money that they don't spend on software into government bonds. It is a much more reasonable assumption that people would spend that money on other consumables, like food or other such tangible commodities.
Lastly, there is the unwarranted assumption that increase in demand for legal software increases employment. In fact, it is a much more reasonable assumption that increase in piracy increases employment in case of developing countries. Printing ("DTP") shops use pirated versions of Photoshop, CorelDraw and InDesign, computer education centres use pirated versions of Microsoft Windows, offices use pirated versions of Microsoft Word and Excel. If these didn't teach their employees the use of pirated software, millions of people would lose their jobs. All of these employees pay direct taxes. There is no analysis in the BSA-IDC report that accounts for this, treating all these millions of people as non-existent for purposes of their analysis.
Increasing tax: Make MNC software companies pay full taxes
Thus, there is no real tax loss to the government if the money that would have been spent on commercial software was instead spent on some other commodity. Indeed, there might even be an increase in tax collection because software companies, including leading ones such as Microsoft, are much more likely to avoid taxes than companies that deal in tangible commodities. There are well-known routes of decreasing tax liability for intangible goods such as software. Software companies normally state that they license software instead of selling it (as this suits them on issues such as customs duties), but when it comes to income tax, they try to paint the transaction as a sale of a product. (Microsoft, for instance claims that its earnings in India are 'business income' and not 'royalties' and hence is exempt under the Double Taxation Avoidance Agreement between India and the USA.) A company that deals with tangible commodities has no such 'licensing vs. sale' loop-hole that they can try to exploit. Further, many software companies are located in special economic zones that are "software exporting zones", and hence get large tax deductions.
In India, for instance, Microsoft is resisting payment of income tax for by routing all licensing to distributors in India through a shell company in Singapore and holding that Microsoft India had no income tax liabilities. Microsoft has been fined Rs. 2 crore because it tried to separate the importing of software into India from the (more valuable) granting of licences to customers and pay only nominal customs duties on the former and under-declaring the value of the latter as zero. From nine Microsoft dealers a total of Rs 255 crore was collected as tax. Of the roughly Rs. 4000 crores loss that the BSA-IDC report claims, around 6% is realizable from just a single tax (customs duties) from 9 companies dealing in the products of one company. If we multiply this by all taxes (income tax included) amongst all the dealers of all the constituent companies of BSA, then the Indian government might recover more from taxes than is supposedly lost to piracy!
Elsewhere around the globe, the 'Double Irish' arrangement, the 'Dutch Sandwich' route and other such are used by MNC software companies to evade taxes. Just as there are tax havens, there are some IPR havens that cater to companies selling/licensing software and other such intangible commodities.
If only these software companies were to stop evading taxes in the countries in which they sell software, then the government's tax collections would automatically increase.
Final idiocies, and conclusion
In the BSA-IDC report, they write: "Assessing the relationship between software piracy rates and UN Human Development Index (a measure of average achievements in a country in three basic dimensions of human development) suggests that countries with greater rates of software piracy tend to have lower levels of economic development. This further strengthens the hypothesis that IP rights (IPR) enforcement increases economic activity.".
This is as sensible as saying "countries with greater rates of industrial espionage (such as France, Germany, and USA) tend to have higher levels of economic development" strengthens the hypothesis that industrial espionage increases economic development. While it is empirically true that most countries with greater rates of software piracy have lower levels of economic development, it is equally true that countries with lower levels of economic development (being countries with poorer populations) have more software piracy. It is equally true that software piracy decreases if the cost of software decreases, as shown by the more carefully-conducted analysis in the Media Piracy in Emerging Economies report.
To use greater software piracy and lower economic development as evidence of the causal link between IPR enforcement and economic activity is to betray absolute ignorance about both economics and logic.
The startlingly poor level of analysis of the BSA-IDC report leaves no question that the conclusions were arrived at independently of the analysis. Such misleading analysis is worse than trash: it is downright dangerous as an instrument of policy setting.
To increase tax receipts, the government may as well start by making BSA's constituent companies pay all the taxes they owe.