Comment by tripletao | Hacker News Reader

Comment by tripletao | Hacker News Reader

Parent Full thread

If I worked for a regulator and wanted to maximize my revolving-door compensation, then I'm not sure whether I'd want to be harsh or lenient. As you say, there are forces in both directions. I'd probably aim for the middle of the pack, since I (a) don't want to make my future employer hate me, and (b) also don't want them to think I'm an idiot lackey more valuable to them in government than on their payroll.

But, I think I have a pretty straightforward incentive to create maximum procedural complexity--that is, to draft and enforce regulations that require lots of administrative effort to comply with, but in the end have little economic effect beyond the legal fees. If I'm really clever, then I might even get whoever thought regulation was needed to perceive the paperwork burden as satisfying that--so the company might not even mind the legal fees, if they get an offsetting PR benefit from the procedurally strict (but substantively meaningless) rules. This is the legal equivalent of a computer programmer's job security through poorly-documented code, just with the creation and maintenance phases split across two different employers.

A lot of what patent attorneys do sure feels like it meets that description. If the regulatory capture mentioned in the paper exists, then I'd guess it works more through that procedural complexity than through substance.

Unless I misunderstand, the parent's hypothesis doesn't try to explain the data, which apparently shows a strong correlation between:

A) volume of patents granted by the examiner, and

B) applicant hiring that patent examiner & applicant hiring the same year as the grant of patent.

The parent seems to be an argument that they would expect the data to be otherwise, but it's not. Do I misunderstand?

I don't think natosaichek's comment above attempts to explain the paper's observed correlations. Possibilities include that:

1. Examiners who want to get hired in private industry go easy on their future employers' applications. That would explain the linked paper's correlations.

2. Examiners who want to get hired in private industry help create a quagmire of complex procedures that they will then be uniquely qualified to help their employer follow. That also would explain the linked paper's correlations.

3. Examiners who want to get hired in private industry go hard on their future employers' applications. That would not explain the linked paper's result, though it could still be true if some confounding variable (e.g., different approval rates for different fields and types of client, as patentatt speculates) explains the correlation.

People worried about the revolving door are usually thinking (1). I suspect that it's actually (2), since that achieves the same effect in a way that superficially seems less corrupt, and that requires less trust between the revolving-door employer and employee.

Run of the mill examiners don’t set policy, so I’m not sure how that would work

An employee who knows the policy has revolving door value, even if they didn't personally create it. You'd know better than I would how much of the USPTO's policy exists as law vs. published guidance vs. internal documents vs. shared habits that never actually get written down; but it often seems as if our lawyers act based on "policy" that's on the fuzzier end of that continuum. That's what's hardest to learn without hiring/being a former examiner.

Like, where do you learn to say "a plurality of" instead of "two or more"? "A first X" and "a second X"? When to express algorithms as flow charts? Those particular examples are well-known, but I presume there's a lot more culture/style like that. That seems as important to me as the formal policy, not necessarily anything substantive but a sign that you're in the club.

“not necessarily anything substantive but a sign that you're in the club.”

That’s most of what this stuff is, you hit the nail on the head. Normal people don’t talk or write in ‘patentese.’ Some of it is just an optimal way to be extremely precise in written English, but a bunch of it is just accumulated cultural nonsense. Good point.

> 2. Examiners who want to get hired in private industry help create a quagmire of complex procedures that they will then be uniquely qualified to help their employer follow. That also would explain the linked paper's correlations.

From my amateur point of view, that would seem to cause a decrease, not an increase in granted patents. At best the number would be unchanged and the process would take longer. So how would (2) explain the observation of an increase?

The paper studied the relative benefit to firms that hire former examiners. That relative benefit is the same regardless of whether examiners go easy on revolving-door employers, or hard on non-revolving-door employers.

The problem is that a lot of people think that way, but "middle of the road" actually means average where the average is people who are middle of the road to people trying to be nice to a future employer.

The result is a ratchet effect making patent examiners more and more friendly to patent applicants over time.

That kind of ratchet can happen, but I still think the "be nice" effect is much less important than the "add procedural complexity" effect. I don't see how a study could distinguish the two without re-examining every patent, but:

1. Anecdotally over my interactions with patent examiners (using both large firms studied in the linked article and small firms), that's how it felt. I never felt the big firms were getting directly special treatment. I often felt they had an oddly perfect understanding of a mostly-arbitrary set of rules not necessarily grounded in any statute or case law that they nonetheless somehow shared with the examiners.

2. The "complexity" effect doesn't require trust. The former examiner genuinely has specialized knowledge that will benefit the new employer, so the employer has an incentive (beyond some vague hope of goodwill from examiners still in government) to keep up the bargain.

Examiners don’t set policy or create procedures, they just examine patents. There’s thousands of them, it’s not like every patent examiner has different procedures.

In the UK the patent examiner's manual - the Manual of Patent Practice - is publicly available; it explains standard precedence cases, tests used (eg to tell if an invention is software as such).

Patents aren't that complex that you can tie them up to make them only processable by the original examiner. UK examiner's have about 1-2 days total processing time per search/examination (less in some fields).

The technical aspects are set by the applicant, as is the language of the claims.

There are definitely some hacks to force through a patent, but it will be of low worth and easily defeated in court (and obviously so to a good patent attorney).

Reply on news.ycombinator.com