This site is only a doorway

Everything here is gathered from things other people actually researched, thought through, and built: books, papers, historical records, and the people behind them. If something here catches your interest, go look at the original sources listed at the bottom of the page. They carry more depth than a page like this ever can.

Why Does Japan Check a Patent Application Only If the Applicant Asks?

Category: Technology

Imagine you invent something and send in a patent application. You might expect an expert at the patent office to start checking it straight away. In Japan, that is not what happens. Nothing gets examined until someone sends a second, separate request that says, in effect, “please check this one.” If nobody does it within three years, practitioner guides say the application is treated as withdrawn[5].

Why would an office make people ask? The short answer: it was a way to stop an overloaded line of applications from growing, and it hands the choice to the applicant. By the end you will also know how often applicants actually ask, and why a quick percentage can fool you.

In Japan, Filing a Patent Application Does Not Start the Checking

A patent is a legal right to stop other people from making or selling your invention for a limited time. To get one, you file an application with the patent office. In Japan, the office is the Japan Patent Office (JPO), and the step that follows has its own name, which translates roughly as “request for examination.”

Here is how the Patent Act puts it, in an English translation hosted by the World Intellectual Property Organization (WIPO, a United Nations agency): examination of a patent application “shall be initiated after the filing of a request for examination,” and any person may file that request within three years of the filing date[1]. “Any person” is not a typo. The request does not have to come from the inventor[1].

The other half of the rule is the part that surprises people. Practitioner guides say that if no request arrives by the deadline, the application is deemed withdrawn[5]. The statute also says a request, once filed, cannot be taken back[1]. So the applicant gets a choice, but it is a one-way door.

That leaves the question this article follows: why split “filing” from “checking” at all?

Before 1971, Japan Examined Every Application Without Being Asked

It was not always this way. A Japanese law-journal article by Nobuaki Kawai explains that, for applications filed before January 1, 1971, the examiner assigned to each one carried out the full examination on their own authority[2]. Nobody had to ask.

That sounds fair: every inventor gets a check. But it has a built-in weakness. If more applications arrive each year than examiners can finish, the unchecked ones pile up and the wait grows. Japanese accounts describe the old system as running into exactly this problem. I could not open a source that gives the size of the pile, so I am leaving the number out rather than guessing.

The 1970 Law Added Two Steps: Publish Early, Examine Only on Request

Law No. 91 was announced on May 22, 1970, and took effect on January 1, 1971[2]. It introduced two new ideas[2]:

  • Laying open (publication): showing an application to the public on a schedule, whether or not it has been examined yet.
  • Request for examination: checking only the applications whose owners ask for it.

Researchers who study the system describe it as giving a company time to sort out which of its inventions deserve patent protection[3]. Not every application is worth a full check. Maybe the product was cancelled. Maybe a competitor got there first. Maybe the idea simply did not turn out to be useful. If the owner loses interest, nobody spends examiner time on it.

An easy way to picture this is a library that only fetches a book once you hand in a slip. The shelves still hold every title, but the librarian does not run to every shelf every day. This is only a way of picturing it, not how the patent office describes itself.

The Time Allowed to Ask Was Cut From Seven Years to Three

The length of the waiting period has changed. Under the old rule, applicants could wait up to seven years to request examination. For applications filed on or after October 1, 2001, the limit became three years[3].

One study of the Japanese system looked at what happened after that change, and its authors argue, from a theoretical model and data, that the shorter period made it more likely that lower-quality applications would be sent for examination[3]. The reasoning is that when you have less time to decide, you are more likely to say “yes, just in case.” The same paper notes that requests for examination rose by 83% between 1997 and 2007, while the number of applications rose by only 1%[3]. Those authors treat this as the combined result of several reforms, not of the three-year limit alone.

So the deadline is not only a paperwork rule. It shapes how applicants decide.

About One in Four Applications Never Gets an Examination Request

What share of applicants do ask? A patent-attorney firm that rewrote the Japan Patent Office’s own annual statistics reports 74.7% for applications filed in 2019, and 72.7% for those filed in 2016[4]. These come from a secondary source, so treat them as approximate. Subtract: 100 − 74.7 = 25.3. So roughly one application in four never gets that request. “Roughly a quarter” is the safe way to say it, not “most.”

There is a trap in this number. You might try to find the rate for a year by dividing the number of requests that year by the number of applications that year. That gives a different thing, because requests made in a given year include ones for applications filed one, two or three years earlier. The honest rate follows one filing year as a group, and counts how many from that group were requested within three years.

Once a request is in, the same summary of JPO statistics puts the wait for the first reply from an examiner at 10.1 months in 2022[4]. Applications put on a fast track took less. That is about the length of one school year.

Try It: Test the Percentage With a Pretend Pile of 1,000 Applications

With a calculator or a piece of paper, pretend 1,000 applications were filed in one year. At 74.7%, about 747 would get a request and about 253 would not. Now try the 2016 figure of 72.7%. How many fewer applications get examined?

Then try the trap. Suppose 900 requests came in during one year, but 300 of them were for applications filed in earlier years. Is the rate 900 divided by that year’s applications? Why or why not?

If you want the original wording, the WIPO page in the list below has the Patent Act text in English. The Japan Patent Office also publishes yearly statistics that the numbers above come from.

Sources

What I could read: [1] full text of the relevant articles (the page was long, so I read the part with Articles 48-2 and 48-3). [2] The title page, abstract and opening of the journal article; it is in Japanese and I could not open the whole PDF, so I used only what the opening states. [3] The abstract and introduction of a Hitotsubashi University working paper; the point that the request system gives time to sort out which inventions deserve protection comes from search-result summaries of its text, not a line-by-line reading. [4] A law firm’s blog post that restates JPO figures (secondary source, in Japanese). [5] A law-firm overview, which I read through a search summary only. The ratio figures were not checked against the JPO report itself. Another Japanese-language account says there was a pile of roughly 750,000 waiting applications before the reform; I could not confirm that number in any source I opened, so it is not used here. I did not look at how other countries handle examination, so the article makes no comparison.

  1. WIPO, “Legislative Implementation of Flexibilities – Japan: Articles 47-49 of the Patent Act (Act No. 121 of 1959, as last amended by Act No. 109 of 2006).” https://www.wipo.int/ip-development/en/agenda/flexibilities/details.jsp?id=9383 (examination starts after a request; any person; three years; no withdrawal of a request; the Act has been amended since 2006, so check the current text)
  2. Nobuaki Kawai, “A case study of the changes in management from invention notification to request for examination” (Japanese title), Intellectual Property Journal 2012, Nihon University. https://www.publication.law.nihon-u.ac.jp/pdf/property/property_5/each/05.pdf (Law No. 91 of May 22, 1970, in force January 1, 1971; examiners examined all applications before; laying open and examination request adopted; in Japanese, opening only)
  3. Isamu Yamauchi and Sadao Nagaoka, “Complementary Reforms of Patent Examination Request System in Japan,” Hitotsubashi University Institute of Innovation Research. https://pubs-stg.iir.hit-u.ac.jp/admin/en/pdfs/file/932 (period cut from seven to three years on October 1, 2001; 83% rise in requests vs. 1% rise in applications, 1997–2007; effect on quality; purpose of the system)
  4. Koyama Patent Office blog, summary of the JPO Annual Report 2023 (Japanese). https://www.koyamapat.jp/2023/07/29/report_patent_2023/ (74.7% for 2019 filings; 72.7% for 2016 filings; first-action period 10.1 months in 2022; secondary source)
  5. Nakamura & Partners, “Outline of Patent Prosecution in Japan,” Mondaq. https://www.mondaq.com/patent/562966/outline-of-patent-prosecution-in-japan (practitioner overview; application deemed withdrawn if no request by the deadline)

Update history

  • First published.