March 15, 2008
Contrary to propaganda, patent damages are not burgeoning; rather, remaining fairly constant. On average, awarded damages in software cases are less than for mining patents, and almost a third less than telecom patent cases. Juries award larger damages, but that may be a statistical anomaly for the types of cases tried before jury rather than at the bench. As purely economic patentees (PEPs) participate more in the patent market in this decade, reasonable royalties becomes more the basis for awards than lost profits.
Contrary to the noise level, software patents are no big deal. While the big boys in IT are screaming bloody murder over patents, it's a Chicken Little shtick by oversized infants. Rick Frenkel of Cisco, please take a bow.
In constant dollars, median damages awards have remained fairly constant. But there's a little devil in the details: the Consumer Price Index (CPI) was used to adjust for inflation. Any econometrician worth his salt can tell you that, historically, the CPI both overstates inflation and skews it, owing to the basket of goods and services it includes and excludes. So treat all the $ numbers in these charts with a pile of salt. Also keep in mind that the numbers are spun from aggregate available data, ignoring that the value of different technologies varies widely, and that patents come in waves by technology, so different years may reflect widely different cases by technology, making this an apples-and-oranges comparison more than anything. Hence, the numbers actually tell very little story. Damn deceiving statistics.
The median for jury damage awards has in the past five years been multiples of bench awards. It is impossible to make much sense of the numbers. It would be interesting to see a breakdown by technology type on this, though the sample size then would be tiny. Again, the numbers say little.
35 U.S.C. §284, the damages statute, sets reasonable royalty as the minimum threshold for infringement compensation. Royalty may be a % of revenue or profit, per unit sold, or a lump sum. PEPs, which own patents, but don't manufacture, rely upon royalty for compensation. Lost profits require fiendishly complex economic analysis, and are thus more difficult to prove. That has become increasing true, as market characterization itself is sometimes problematic. Price erosion too is an econometric exercise, but more readily demonstrable for competitors than lost profits, which require more revelation of internal figures than a company may be willing to tolerate.
The latest on big awards. Microsoft really knows how to splash out.
Here are the seeming Big Kahuna in historical damages, though not really properly so, as they are not in constant dollars.
Figuring what patents really are worth, all the damages numbers are heavily skewed by the fact that only 3% of patent assertions make it through trial; 97% settle. Given that lower numbers more easily result in settlement, and settlements are easier for clear-cut cases, damages awards as much as anything represent stubbornness, unreasonableness, and denial. Reliable numbers upon which conclusions may be surely based are very hard to come by for patents.
Posted by Patent Hawk at March 15, 2008 2:43 PM | Damages
The problem with this analysis is that it fails to say that the data set is puny. You get about 100 patent case per year going to trial. All it takes is only blockbuster case to throw off the stats.
Posted by: Steve Sereboff at March 15, 2008 8:37 PM