In the early 90's, the software industry in the U.S. decried the rise of software patents, claiming that patenting would doom small companies and stifle growth in the industry. Even large companies, such as Oracle and Adobe voiced their opposition to software patents in testimony before the USPTO, stating that "existing copyright law and available trade secret protections . . . are better suited to protecting computer software developments."
Anti-patent activists, particularly in the EU, have used many of these comments as fodder for their quest to eliminate software patents altogether. They have even gone a step further in their rhetoric, suggesting that it will be impossible for software companies to operate once software patents are recognized, since larger sompanies will swamp the smaller sompanies with patent filings and litigation.
However, a current study being conducted by Robert P. Merges, demonstrates that this viewpoint is mostly wrong - patents have not killed the software industry, they have not led to a slowdown in entry, and they appear to have had a negligible impact, if any, on industry structure.
Professor Merges looks at several industry metrics and reaches the following conclusions:
Industry Structure and Concentration - anti-patent commentators previously alleged that, because of software patents, the software industry would become more concentrated over time. The idea was that strong patent protection over "backbone" software would tend to "lock in" a dominant product, and hence the company that owned it. However, this view ignores the dynamic sources of growth in the industry, and underestimates how real and effective this competition is in the fields where it operates. For every "backbone" product – such as an operating system program – there are many applications and ancillary products that connect to the backbone. For these products, ownership rights do not appear to create "lock-in" conditions on anything like the scale envisioned in the early 1990s:
Industry concentration statistics tell the story here. The conventional measure of concentration, the Herfindahl-Hirschman index (HHI) ranges from 0 to 10,000; the HHI for the software industry as a whole is less than 244 for software, compared to an average of 334 for U.S. manufacturing industries. What this means is that the top 20 sellers of pre-packaged software generate 61% of total industry revenues. This compares very favorably to other industries, many of which are considered quite competitive: autos, airlines, and personal computers, for example. There is evidence of significant turnover over time as well – a key indicator of a dynamic industry. Of the top ten software companies in 1990, five did not make the list in 2000, either because they went out of business or were acquired. This is remarkable turnover compared to some industries, such as pharmaceuticals, where similar comparisons from 1990 and 2000 show that eight of ten firms made both lists (and the ones that did not were acquired by others that did)Comparative Data on Patents and Innovation - to determine how patents relate to these trends, one area of study looked at comparative data concerning differences between the domestic software industries of various countries. Previous research has shown that as a foreign country moves up the learning curve in the software industry, inventors in the software industry from that country receive more patents. Accordingly, software firms in Israel, Ireland and India were studied to see if the patenting trends hold:
Software firms in India and Ireland are less innovative than elsewhere. Routine service-type programming is the norm in India, while in Ireland, the industry is dominated by subsidiaries of foreign firms who add minor value to the parent company software, and who are located in Ireland partly for tax reasons. IsraeliEntry to Market - since the software industry claims large sources of innovation from both large and small companies, a steady flow of new entrants would indicate a robust industry that seeks to establish innovative ideas or products. After reviewing the level of startup activity between 1970 and 1998, and considering the venture capital funding amounts between 1995 and 2005, Professor Merges findings indicate that software startups and financing activity have continued to do quite well.
firms, which are perceived as being the source of more innovative software, patent much more heavily than their Indian and Irish counterparts. While a number of reasons might explain this pattern, it is certainly consistent with the view that patents correlate closely with R&D and innovation – which would tend to refute the early 1990s argument that patents are anathema to software innovation. In addition, it can be said that the Israeli software industry is in no sense highly concentrated. So the comparative data once again supports the notion that predictions concerning the concentration-increasing effects of software patents have failed to materialize.
Other studies on Software Patents:
Ronald J. Mann, Do Patents Facilitate Financing in the Software Industry?, Texas Law Review, Vol. 83, p. 961, 2005
James E. Bessen and Robert M. Hunt, An Empirical Look at Software Patents, FRB of Philadelphia Working Paper No. 03-17
Ronald J. Mann and Thomas Sager, Patents, Venture Capital, and Software Start-Ups, U of Texas Law, Law and Econ Research Paper No. 57