That’s why we provide patents for good ideas (and, yes, some not-so-good ideas, too).

Except sometimes we don’t. By law, patents can be denied for ideas that may be good, but are not novel or are obvious. This sounds completely sensible, until you think through the consequences. It means that even potentially great drugs might not come to market because they were disclosed in the past — so they’re now not novel — or because they are a natural extension of existing knowledge — so they’re now obvious.

Prior disclosure can come in the form of an obscure research article; an old, expired patent; or inclusion of a chemical structure in a giant online database, for example. (When it comes to obviousness, there’s a tension. We don’t want to provide patent protection for profit-increasing activities that do not benefit patients, like “me too” drugs — drugs that are only trivially different from existing ones. But some things that are obvious also might be beneficial, and for those we’d want to encourage development.)

The crux of the problem is this: For pharmaceuticals, patent protection is used as a means for innovators to recoup the costly investments that drug development requires. But the patent system was not devised to solve this specific problem. It’s a broader system intended to encourage innovation, which it does, while at the same time permitting individuals and firms to exploit obvious and old ideas freely, which is generally a valuable protection.

When it comes to drugs, some of those obvious, old ideas that cannot be patented have not been clinically tested, as required for F.D.A. approval. That takes money, which nobody will invest without a patent. It’s a Catch-22 that ends up excluding potentially valuable drugs from ever even being considered for development.

There’s evidence that the provision of patents only for nonobvious and novel innovations affects drug development. In a paper published in the Texas Law Review, Mr. Roin documents many examples of patent invalidation on these grounds — importantly, even for ideas that had never been developed into drugs and that still required clinical trials for F.D.A. approval. For example, the patent of an anti-inflammatory drug was invalidated because it had been disclosed in a prior academic article. A patent for a hypertension drug was invalidated because it was deemed to have been created by a well-known process.

There are ways out of this bind. Kevin Outterson, a Boston University law professor, has called for greater government funding of clinical trials for public-domain drugs.

Mr. Roin, the M.I.T. professor, describes another approach: to provide a period of market exclusivity — long enough to motivate investment in clinical trials — to any organization addressing an unmet medical need with a drug that isn’t patentable. If it invested in securing F.D.A. approval for a drug based on active ingredients not found in existing drugs, an organization would be granted such a period of market exclusivity and the stream of profits that usually accompanies it, even if the drug was considered obvious and not novel. That idea has been included in congressional legislation, but has not made it into law.

If we want drug companies to invest more in drug development — as Mrs. Clinton has proposed — we might first consider reasons that, even with some good ideas, they don’t.