It's the day before Christmas and all through the house...we're reading the NY Times editorial about China's growing thievery of intellectual property rights (IPR). It's rare for the world's "paper of record" to run an editorial about such a specific issue coupled with details about Siemens, railroads, etc.
As I've written before in this space and elsewhere, Indigenous Innovation policies are troubling, but mainly for their potential future harm, not for what has occurred so far. In the case of Siemens, I wonder why they haven't gone to court. Is it that they're afraid of upsetting the Chinese and being locked out of further business? That what the Chinese side did was hardball but legal?
The editorial notes that at the recent JCCT the Chinese promised to better protect software. On this point, it'd be instructive to look at some data. IPR theft data is extremely difficult to compile, and much of what is compiled is methodologically flawed (more in a second). But let's see what we've got. The most common measure is provided through an annual study released by the Business Software Alliance (BSA) and the International Data Group (IDG). The BSA/IDG studies have been released since 2004, and before that BSA released its own report for several years. But to be sure we compare apples to apples, let's just look at data from the seven BSA/IDG studies issued so far. I've compiled the data on China and the US in two graphs and a table:
Two things stand out from this table: (1) China's piracy rate, in terms of the percentage of pirated software to total software sold, is much higher than that in the US; but (2) the piracy rate in China has been consistently on the downward trend. Whereas it used to have the first or second-highest piracy rate in the world, in the 2010 study, China was tied for 26th.
Things look differently if we compare the China and American piracy rates in terms of dollar value of pirated software:
From this perspective, the United States has been and continues to be the world's leader in terms of software piracy. Just using the last seven years, total software piracy in the US is valued at almost $53 billion, compared to almost $38 billion for China. Earlier this month BSA lobbied Washington to take a new approach on IPR toward China, focusing on achieving actual results. Perhaps BSA should push Washington to change its approach toward protection software IPR in the United States.
Two wierd things about these tables stand out: (1) The reduction in China's piracy rate, in terms of percentages, appears to have accelerated in 2005 and 2006, just when China had launched the first stage of its Indigenous Innovation policies; and (2) Despite the decline in percentage terms, the dollar value of software piracy in China seems to have grown. That must mean the total hypothetical software market expanded dramatically during the past few years, and thus it is likely that actual software sales are also rising quite quickly.
What to make of these numbers? I think they're highly suspect guesses about what is going on.BSA/IDG explain that their data reflects the "value of pirated software if it had been sold in the market. It is calculated using a blended average price of software in an economy, as sold in retail stores, using volume licensing, and as bundled with hardware" (BSA/IDG, 2010 Survey, p. 8). This assumes that consumers would have paid for the software at the average price had the piracy option not been available. But we know that is not the case, and much less software would have been in use. This type of calculation may make sense when calculating the cost of stolen cars, but it doesn't transfer well to copyright infringement. Even the researchers at IDG who conducted the study for BSA disagree with terming the calculated figures as "lost sales" since it is likely only one in ten of those sales would have occurred at the retail price.
On the other hand, BSA and IDG are consistent, and although you don't want to be inaccurate when doing surveys, it is much more important to be consistent, that way your research maintains the same bias over time and space, and it is easier to make adjustments from that starting point.