Apples and Oranges in the IP5 Statistics, or how to make your patent filing statistics look more fruitful

Patent filing statistics: a fruitful topic to pursue. Peter Arrowsmith (partner in London-based patent attorneys Cleveland) has been subjecting patent statistics to some careful scrutiny of late. If readers' comments are anything to go by, Peter's two-part post on the Great British Vanishing Act (here and here), on whether British inventors are really a disappearing breed, was very much appreciated. This time round, Peter looks at the presentation of patent filing figures and makes some acute observations. This is what he says:

I was recently reviewing the key IP5 statistical data for 2013, which is released jointly by the five largest intellectual property offices in the world (IP5): the European Patent Office (EPO), the Japan Patent Office (JPO), the Korean Intellectual Property Office (KIPO), the State Intellectual Property Office of the People’s Republic of China (SIPO), and the United States Patent and Trademark Office (USPTO). In reviewing the report I noticed an inconsistency in the basis on which patent filing data are calculated. The net result is that EPO figures are presented in a more favourable light than they should be. It seems that this is down to the surprising way in which the EPO calculate their filing statistics.

Summary of the IP5 Statistical Data

An extract from the IP5 report reads as follows:

“The IP5 Offices here report a large growth in patent filings in 2013, with altogether 2.2 million patent applications filed. This is an increase of 10.6% compared to 2012. The growth over recent years is shown in the following graph, which demonstrates the need for continued efforts to improve the efficiency of the system”

Patent filings at IP5 Offices

“EPO, USPTO, KIPO, and SIPO each had an increase in patent filings ranging from 2.8% at EPO to 26.4% at SIPO, while JPO had a decrease in filings of 4.2%".

The underlying data from the graph are shown in the table below, also showing the percentage change on 2012 figures: 

Origin
Office	EPC states	Japan	R. Korea	P.R. China	United States	Other	Total
EPO	93 905	52 437	16 857	22 292	64 967	15 232	265 690
	- 0.1%	+ 1.2%	+ 14.0%	+ 16.2%	+ 2.8%	- 1.9%	+ 2.8%
JPO	20 604	271 731	6 134	2 064	23 481	4 422	328 436
	- 1.4%	- 5.3%	+ 7.5%	+ 2.1%	+ 2.4%	+ 4.5%	- 4.2%
KIPO	11 734	16 300	159 976	1 144	12 977	2 458	204 589
	+ 15.1%	+ 1.8%	+ 8.0%	+ 16.5%	+ 14.4%	+ 9.1%	+ 8.3%
SIPO	33 287	41 193	10 866	704 936	29 992	4 862	825 136
	+ 4.7%	- 2.6%	+ 20.9%	+ 31.7%	+ 1.6%	- 1.0%	+ 26.4%
USPTO	87 314	84 429	33 039	14 702	293 588	58 540	571 612
	+ 2.5%	- 4.8%	+ 12.1%	+ 10.8%	+ 9.2%	+ 2.0%	+ 5.3%

I was immediately suspicious of these data because they quote an increase in the number of patent filings at the EPO in 2013, whereas some of my recent research indicates that there has actually been a decrease.

Filing data from the EPO

It is apparent from the EPO’s annual report that the office bases its filing statistics on the number of direct European patent applications plus the number of international-phase PCT applications (EPs + PCTs). This is the case whether or not these PCT applications subsequently enter the European regional phase. Technically, of course, a PCT application that designates the EPO is equivalent to a regular European application (according to Article 153(2) EPC). However, it is also clear that many PCT applications are filed by applicants who ultimately have no interest in entering the European regional phase. Therefore, this seems a perverse basis on which to calculate the total number of European filings.

A more meaningful basis would be the number of direct European patent applications plus the number of PCT applications that actually enter the European regional phase (EPs + Euro-PCTs). These data are available for download from the EPO’s website and, by this measure, it is apparent that the number of European patent applications actually decreased by around 0.5% in 2013. Also, the total number of applications is more than 100,000 less than the number quoted by the EPO. The figure below demonstrates the difference in the trend of applications, depending on whether you use (PCTs + EPs) or (Euro-PCTs + EPs).

The President of the EPO blogged about the statistics on 12 March 2014, saying: “The filing figures for 2013 confirm the trends of the past few years, with an overall increase (266,000 filings, +2.8% versus 2012) resulting once more in an all-time high”. Although not technically incorrect, this statement from the President is, at best, misleading.

Apples from the EPO and Oranges from other offices

The EPO use PCTs + EPs when comparing their filing data with the other IP5 offices. This is a bizarre method of comparison. If other offices were calculating filing data on the same basis then all of the data would be strongly correlated, rendering a comparison pointless. On closer inspection of the underlying data, however, it is clear that the other IP5 offices do not calculate filing data on the same basis as the EPO. This is apparent by looking at data from WIPO regarding the number of PCT applications filed by applicants from different origins. From these data it is apparent that the following PCT applications were filed by applicants from IP5 offices, excluding the EPO: 

Country	PCT applications filed in 2013
United States	57,239
China	21,516
Japan	43,918
Korea	12,386

Comparing these data with the table from IP5, it is clear that the other IP5 offices do not include international-phase PCT applications. Taking just one example, at KIPO the number of applications filed by Chinese applicants is quoted as 1,144; far less than the 21,516 international-phase PCT applications filed by Chinese applicants.

Thus, the IP5 report compares apples from the EPO with oranges from the other IP5 members. Consequently the IP5 report falsely suggests that EPO filings have increased, whereas they have actually decreased.

Reflections

It seems that great care is required in interpreting any statistics from the EPO. It is important to check the underlying data to determine precisely what is being represented. In my view it is misleading of the EPO to use PCTs + EPs to represent the number of filings, as this is not a true gauge for actual desire for European patent protection.

The problem is that, once released, the EPO statistics are divorced from their original context. I suspect that this is how they found their way into the IP5 report, with the collators simply assuming that the statistics were accurate representations of the number of European filings. It is troubling that statistics such as these may be relied upon by stakeholders when it comes to important decisions on the patent system.

As usual, readers' thoughts are most welcome!