The Price Elasticity of Demand for Patents

By Dennis Crouch

In March 2013 the USPTO significantly raised the maintenance fee charges. Most pointedly, the 11 ½ year renewal fee has jumped 65% from $4,810 to $7,400. The 11 ½ year renewal is last of three fees due and also the most expensive. Market theory for ordinary goods suggests that an increase in the price of goods will result in a reduction of the quantity demanded. The relative proportionality of the reaction depends on a number of factors, including the price elasticity of demand. In the past, I have suggested that the price elasticities of many of the patent office fees are relatively inelastic at their current price point, but I wanted to see how patent owners are responding to the increased fees.

Initial Conclusion: The price elasticity of demand for the 11 ½ year renewal fee is quite inelastic.

Data: I created two groups of patents: One where the 11 ½ year renewal fee was due prior to the March 2013 fee-change and a second group where the 11 ½ renewal fee was due after that date. To simplify the analysis, the pre-fee-change group only included the patents whose 12-year date was prior to the fee change in order to also capture late payments. Similarly, the post-fee-change group only includes patents whose 11 ½ year date was more than 6-months after the fee date to avoid cases where the lower-fee was pre-paid. For each group, I pulled up data for 11-weeks of patents to see what percentage of 11 ½ renewal fees were paid (of those eligible). Those results are shown in the table and charts below.

Fee Due     Paid Second Fee Paid Third Fee Percentage
Pre-Fee-Change 4,810 23,686 17,709


Post-Fee-Change 7,400 23,014 16,592



This change was relatively constant for both large and small entities. [Small entities are less likely to pay renewal fees, but the relative drop at the time of the fee change was proportional.]

What does this mean?: First, the USPTO fee raise is a boon to USPTO revenue – at least tens of millions of additional dollars per year for the agency. The USPTO has monopoly power over patent rights in the US and, in that situation, the high inelasticity generally indicates 050514_1628_ThePriceEla1.gifthat raising prices will raise revenues. To even further maximize this profit, the PTO could hold an auction for each patent at the 11 ½ year mark – giving the patent to the highest bidder. Of course, this money must come from somewhere and many corporate patent budgets are relatively fixed – leading to a choice between new filings and maintenance fee payments. The USPTO is likely wise to avoid more substantial increases in fees – over the long run markets find ways to avoid monopoly power that is abused.

Big Caveats: I looked at this particular fee change because it was so large and so I suspected that a response would be fairly easy to measure. These fees are typically due at a point about 15-years after the invention date. Many folks would suggest that – by that point – the patentees should have a fairly good handle on the value of the underlying innovation and the patents covering that innovation. Further, at that point patent valuation will likely have a strong bi-modal distribution with a substantial number of patents having essentially no value because technology has moved-on (and is never coming back) and then a substantial number with value because the innovation (or its kin) is being used in products on the market. If an innovation covered by a patent is being used in the marketplace then the patent is likely worth significantly more than $10,000. The leading theory is that, by this 15-year-mark, there are not many patents in no-mans-land that are still not on the market but that have a good chance of being on the market within the next few years. Of course, this is a theory with little empirical basis except that it does help to explain the low elasticity shown above.

In other realms, I have shown more price elasticity due to USPTO fee changes. These are most likely where the patentee is not choosing between patenting and abandonment but instead is choosing whether or not to add a feature, such as additional patent claims or an extremely long disclosure. When the USPTO significantly raised the costs for both of those, patentees reacted by substantially reducing demand for those services.

31 thoughts on “The Price Elasticity of Demand for Patents

  1. 7


    The 72% rate for payment of the 11.5 year maintenance fee seems really high to me. It seems like you have posted other studies (see here: link to suggesting that the average rate of maintenance fee payment for the 11.5 year fee is between 40-50%. Is it possible that the samples you have chosen in the current post are unrepresentative of the larger population?

    1. 7.1

      I realize that the 40-50% figure in the post I cited refers to the percentage of all patents, while the current post limits the sample to those patents that are in force at least until the 12th year. Still, I wonder if the 72-75% rate in the current study is in-line with historical data.

    2. 7.2

      I think you largely answered your own question, but in general, the percentage of cases where fees are paid is increasing at each stage.

  2. 6

    Let’s take a hypothetical large-company like IBM who widely cross licenses their patents since that they have no resale value. The only question that one might ask in determining whether to pay the large maintenance fee at the 11 and one-half-year mark, would be whether the patent was licensed as a separate 1% royalty, whether the patent covered a current product, or whether the patent was perceived as necessary for the licensing program. If not, logic would dictate that the maintenance fee should not be paid.

    Certainly drug companies would renew their patents if they covered current products.

    Which takes us to the entity that does not cross license. Because these patents are not cross licensed, they could be of high value. If the patent does not cover a current product, sale of the patent should be considered, but in no case should the patent simply be abandoned. If the company goes belly up, the greatest asset the company may have are its patents precisely because they are not licensed.

    1. 6.1

      since that they have no resale value

      Where the H is this obviously false statement coming from?

    2. 6.3

      Ned, since this is after 15 or more years from their original filing dates, many patents are on technology that is simply obsolete by then and of no value to anyone.

      1. 6.3.1

        Paul, you’d be surprised. A lot of times the technology described in the spec has been superseded. But the inventive subject matter claimed still might be in wide use. It takes a patent attorney and an engineer combined to make the call on whether a patent is obsolete. That costs money all by itself.


          To supplement Ned’s comments, many people (mistakenly) believe that one technology replaces another technology. From my seat, what I see more of is that one technology builds upon another technology.

          Another point to consider, when big companies enter into cross-licensing deals, it is nearly impossible to go through each patent portfolio with any real analysis to identify what patents are infringed and what are not. In this situation, it becomes a measure of who has the biggest d__k. With that in mind, there is some benefit to be had in keeping alive patents that may never be infringed.

  3. 5

    “The USPTO is likely wise to avoid more substantial increases in fees – over the long run markets find ways to avoid monopoly power that is abused.”

    Maybe they should describe their fee payment schedule with functional language, that seems to work.

      1. 5.1.1

        Good. Here I was worried the joke would be over this lot’s collective heads.

    1. 5.2

      Actually, we all know that functional language is a terrible abuse. So, your suggestion would lead to more work arounds, not fewer.

  4. 4

    Dennis, this is interesting, but I have a few questions:
    1. Were many of these prior [lower fee] 11.5 year abandonments occurring during or near the 2008 era recession, when there was at lot of increased cost-cutting?
    2. In my limited experience, most of the abandonment decisions were already being done at the 7.5 year maintenance fee point, not waiting until the 11.5 year maintenance fee. Among other good reasons, that loosely coincides with some increased foreign equivalent fees.
    3. There are some patent owners that either do not want to be bothered with maintaining regular maintenance fee programs, and/or have difficulty finding an employee willing to make abandonment decisions for fear of being punished by a perceived mistake in hindsight.
    4. Some companies may want to maintain perceived stock price enhancement PR value in their “number of patents in force.”

    All of the above would reduce the elasticity of 11.5 year maintenance fee payments to fee increases. Furthermore, both the old and new 11.5 year fees are still small enough to be less than the total investment in a new patent.

    As to the sales suggestion, besides the obvious legal restraints on PTO compulsory sales of patents about to be abandoned that the PTO did not ever own, the last thing many companies would want is the worst kind of troll cheaply buying up and extending the terms of patents that could potentially be used to harass some their own customers. However, the opposite could be true in a few cases, namely, other companies paying patent owners to not pay maintenance fees on some patents so as to put them into the public domain.

    1. 4.1

      Thanks Paul – Your comments are helpful. In this post I am focusing primarily on the difference between two groups of patents where the 3rd-fee is due. Thus, to the extent that all of the factors your mentioned equally impact both groups then they should not skew any elasticity conclusions. I only used three controls to keep the groups parallel: (1) keeping the before and after groups as close to one another in terms of issuance date; (2) correcting for small/large entity differences; and (3) using the entire population of patents (for the given time periods) rather than a subset or sample.

  5. 3

    Dennis, this seems to argue that one should keep increasing the fee until one sees a significant reduction in renewals.

    The revenue could then be used to build better examination infrastructure, for example.

    1. 3.1

      I feel compelled to point out Ned that the ability of the Office to set fees, while greatly expanded with the AIA, still contains a restriction that seems to have been lost sight of.

      The Office does not have an open-checkbook.

      Fee setting still MUST balance in the aggregate and match costs. The implicit notion in your comment is a dangerous one. But (to me) it is a far less dangerous one that seems to continue to escape ANY oversight or commentary from the legal community (including academics): the overall limitation that all of the fees must still ONLY add up to the allotted budget. If some fees are greatly expanded, there MUST be somewhere else in the system a corresponding decrease – otherwise (and especially until there is a NO-Diversion law), what we really have is an open-ended innovation tax.

    2. 3.2

      What I’m suggesting is that the PTO could likely further increase its fees with the result of further increasing revenue. I am not arguing that the PTO should take that action.

      1. 3.2.1

        I am not arguing that the PTO should take that action

        Regardless of the “should,” I would point out that the “could” contains a bit of fallacy, Prof. Crouch.

        There remains the important AIA budgetary balancing restriction.

  6. 2

    First, the USPTO fee raise is a boon to innovation tax revenue


    How much to date has been skimmed from innovators paying into the USPTO, but the moneys moved to the general ledgers?

    1. 2.1

      “First, the USPTO fee raise is a boon to innovation tax revenue”

      A renewal fee is simply an offer to extend the exclusionary monopoly. It’s neither a tax, as is it not mandatory, nor is it on innovation as society has already extracted the innovation-producing information by that point. It’s an optional fee on commercial activity.

      Further, that argument assumes that there exist people who invent, don’t disclose purely due to a couple thousand in fees, don’t bring a reverse-engineerable product to market within what would be the renewal window, and whom nobody else provides similar innovation that they either disclose or simply give away within a short period of time.

      There are no such people.

      A couple thousand in fees 11 years after the invention has zero effect on innovation.

      1. 2.1.1

        As with the laws itself, you seem oblivious to the points in these threads.

        The tax comes about through diversion – when monies paid by innovators into the USPTO are shifted away from what those monies were paid for and end up in the general ledgers – which are supposed to be filled with normal tax money.

        Further, your “the argument assumes” is an errant assumption, as no such assumption has been made. Nice strawman – might as well just close the patent office doors.


          monies paid by innovators into the USPTO

          Happy to be corrected on this point, but doesn’t the vast majority of that money come from foreign entities or so called “Big Corp”, the latter of which (we are constantly being told by the self-styled “David’s” of the world) doesn’t do any innovating?


          “The tax comes about through diversion – when monies paid by innovators into the USPTO are shifted away from what those monies were paid for and end up in the general ledgers – which are supposed to be filled with normal tax money.”

          Again, it’s not a tax, and again it’s not related to innovation. You’re hung up on the fact the money is paid to the PTO to begin with. The only reason the PTO collects maintenance fees is because it’s the most efficient arm of the government to do so.

          If Congress wanted to imbue another agency with the fee collection (for example, via the IRS on a yearly basis) or somewhere else in the DoC generally it would make just as much sense. The IRS would collect fees and upon failure to do so would order the PTO to hold the patent abandoned. There’s no patent or innovation-related nexus to the payment of maintenance fees – it is a fee to continue to enjoy a grant like they have with public broadcasting or anything else.

          Maintenance fees are not an “innovation” “tax” because money is “shifted” out of the PTO – it’s no different than paying $50 for a hunting license or for registering your car.

          If you were talking about search or examination fees going outside you would have a point, but the PTO makes a ton of revenue off of maintenance fees and frankly the fact that it goes to the PTO at all is a corrupting thing. It costs virtually nothing to maintain the roster of active patents, certainly not what is being charged for it. Instead it encourages the PTO to grant patents that shouldn’t be granted for it’s own pecuniary interest.


            Whether maintenance fees are a ‘tax’ likely depends upon context because the meaning ‘tax’ can vary according to the circumstances (and potential statutory definitions).

            I think personal property tax provides a very good analogy to maintenance fee dues. In both cases, there is a property right and the rights-holder owes money to the government in order to continue having full rights. There are a couple of key differences. Personal property taxes are generally owed as a consequence of past ownership while maintenance fees are owed as a prerequisite to future ownership. Also, the remedies are different – but those differences can be explained by the ex anti / ex post difference just highlighted.


              Whether maintenance fees are a ‘tax’ likely depends upon context because the meaning ‘tax’ can vary according to the circumstances (and potential statutory definitions).

              And sometimes, whether or not a fee is a “tax” can be answered multiple different ways. Even in the same Supreme Court decision.


            Random Examiner,

            You clearly do not understand the operating system that the Patent Office exists in.

            Come back when you have a clue as to what you are talking about.

  7. 1

    The Belgian economist Bruno van Pottelsberghe has done research on the price elasticiy of patents in Europe. I recall that he did observe some elasticity.

Comments are closed.