Why weak patents? Testing the examiner ignorance hypothesis

There is a widespread impression, reflected in recent legislation, that US Patent Office examiners issue many patents of dubious validity, and are insufficiently informed to distinguish these from other valid applications. We address this issue using related application outcomes at the European Patent Office as indicators for patent weakness. We create a proxy for potentially citable prior art using latent semantic analysis of US patent documents, and use this to construct a measure of examiner search effort. We find that US examiners tend to devote more search effort to weaker patents, implying that they can identify a substantial portion of the weak patents that they issue. Why the patent system fails to make better use of examiners' ability to identify weak patents is a question that merits further investigation.