174,000 Commandments

One of my favorite Cato publications back when I was on staff there was Wayne Crews’ Ten Thousand Commandments. Published every year, the report documented the mountains of burdensome regulations that businesses had to comply with that year. He has since moved to the Competitive Enterprise Institute, where he has continued to produce the report. Here is the 2006 version:

The 2005 Federal Register, the daily depository of all proposed and final federal rules and regulations, contained 73,870 pages. This is a 2.4 percent decrease from 2004’s 75,675 pages, which had been an all-time record. In 2005, 3,943 final rules were issued by agencies. This is a 3.8 percent decline from 2004’s 4,101 rules. Whereas regulatory agencies issued 3,943 fi nal rules, Congress passed and the President signed into law a comparatively low 161 bills in 2005. In the 2005 Unifi ed Agenda of Federal Regulatory and Deregulatory Actions, which details rules recently completed as well as those anticipated within the upcoming 12 months, agencies reported on 4,062 regulations that were at various stages of implementation throughout the 50-plus federal departments, agencies, and commissions.

Crews is doing an important service by highlighting the severe burdens the regulatory state places on private industry. It occurs to me that a similar analysis would be illuminating for the patent system.

A quick search at the USPTO’s website suggests that approximately 174,000 patents were granted in 2006. If we assume (conservatively, I think) that the average patent is 10 pages long, that means that if you published all of 2006’s patents in a single publication, it would be about 1.7 million pages long—or about 20 times the size of the Federal Register for the year. Indeed, there appear to be about 2.3 million patents in effect at the moment, which means that the text of all patents currently in force would fill about 25 million pages, or about 400 times the size of the tax code that libertarians (justifiably) complain so much about.

Most of the standard arguments libertarians make to criticize the regulatory state and the tax code apply with equal force to the patent system. As I’ve tried to document, many patents are every bit as vague as the worst government regulations. And there are so many patents that it’s effectively impossible to be sure your new product isn’t infringing on someone’s patent.

Libertarians emphasize that a complex tax and regulatory regime hurts small companies for whom the compliance costs are a disproportionate burden. If anything, that critique applies with more force to the patent system, which doesn’t just subject small companies to the same rules as large companies, but actually allows the large companies to use the patent system as a weapon against their smaller competitors.

Another frequent critique of the regulatory state is its cost. Critics estimate that complying with all the regulations currently in effect costs upwards of a trillion dollars. I don’t think the patent system can match that figure, but the costs are clearly significant. For starters, the cost of simply obtaining the patent appears to be about $10,000, suggesting that just the paperwork of filing for patents costs companies nearly $2 billion per year.

But that’s likely to be just the tip of the iceberg, because the costs of patent litigation likely dwarf the costs of filing applications. Last year’s BlackBerry case cost Research in Motion $612 million. The Eolas patent dispute cost Microsoft $521 million. And there are dozens of patent lawsuits each year, many of which net their plaintiffs tens of millions of dollars. And these eye-poppiung payouts don’t count the millions of dollars it costs to litigate each case. Put them all together, and I’d be shocked if those costs didn’t significantly exceed $2 billion/year.

But even those costs are likely to understate the full cost of the patent system. With almost any system of regulations, the largest costs are the costs of changes in behavior people make to avoid patent lawsuits. A major cost of the tax code, for example, is the cost of setting up tax shelters and then changing their business practices in order to take advantage of them. Most likely, companies play similar games to avoid being sued by patent holders. These costs are always very difficult to estimate, but here’s one data point that hints at the scope of the costs. I was recently reading a post (I forget where-if someone knows please comment and I’ll add a link) about an interview with a senior IBM official, in which the official states that IBM earns about a billion dollars in licensing revenues from its patent portfolio, but that the value it gets from being able to get access to others’ patent portfolios dwarfs that figure. Now, to say that the value of access to patent portfolios dwarfs a billion dollars is equivalent to saying that the costs of not having access to those patents would be a lot larger than a billion dollars. That suggests that the thousands of smaller companies that don’t have IBM’s extensive patent portfolio are collectively inconvenienced to the tune of tens of billions of dollars by the costs of accessing patents covering their products.

Now obviously, the patent system has some benefits as well, and it’s important to weigh those benefits against the costs. It’s quite possible that when we do that analysis, we’ll find that on net, patents are beneficial to most sectors of the economy. But I’m relatively certain that the analysis won’t come out that way for software and business method patents, for which there’s hardly any upside. And I think as libertarians we should at least give some careful thought to the way patents affect other sectors of the economy. Patents are a variety of government regulation, and so as libertarians it’s important for us to give them careful scrutiny.

SHARE: