The 2020 Australian IP Report: Revisited

Last month, I wrote about IP Australia’s release of the Australian Intellectual Property Report 2020 (‘IP Report’).  I focused in particular on data included in the ‘Patents’ section of the report, and raised some questions about the methodology and the accuracy of some of the results presented.  I am pleased to say that IP Australia (more specifically, Benjamin Mitra-Kahn, who is the General Manager & Chief Economist in the Policy & Governance Group) reached out to me regarding the issues raised in my article, and we have spent the past couple of weeks working through our respective data sets to identify the origins of the discrepancies between our numbers.  IP Australia will be updating the online version of the IP Report to incorporate a number of corrections.  In this article, I will run briefly through the issues that we identified, and present some updated numbers of my own.

Users of the 2019 release of the IP Government Open Data (IPGOD) will want to take note of a problem with the ‘related patent information’ table (a.k.a. ‘ipgod128’), which records data relating to divisional and additional applications, whereby a number of entries are missing corresponding with applications filed during the final weeks of 2018.  This error resulted in my undercounting of divisional applications in 2018, such that I found an increase in 2019 where the IP Report identified a decline. 

Australian organisations, including a number of universities, that might have felt short-changed by the ‘top-five’ list of applicants presented in the IP Report, should be pleased by corrections that will reflect the true numbers of filings they made during 2019.

To some extent, IP Australia and I have agreed to disagree regarding the merits of distinguishing between ‘original’ and divisional applications, and of ranking applicants on the basis of ‘original’ filings rather than total filings (i.e. including divisional applications).  We do agree, however, that not all divisional applications are created equal, and that a more nuanced approach is necessary to develop a better understanding of how applicants use divisional applications, and what implications this may have for the operation of the patent system, as reflected in this tweet from IP Australia’s Office of the Chief Economist:

Hi Mark, thanks for the collaboration offline, and the ongoing conversation. As discussed, divs aren’t 2nd class filings, or perceived as such, but an important part of the current patent application story. Keen to explore how to better categorise them.

— IPAustraliaEconomist (@IPAustralia_OCE) May 8, 2020

While there may still be some minor differences between my numbers and those of IP Australia, due to changes in the live data between the times that we respectively extracted our information, the numbers in this article should be in substantial agreement with those in the revised version of the 2020 IP Report.

Divisional Filings – In Decline?

The IP Report identified a decline in filings of divisional applications in 2019 as compared with the previous year, whereas I found there to have been a slight increase.  As already noted above, my data was wrong because, although I had the correct numbers for 2019, I relied upon IPGOD data for 2018 and earlier years.  As the chart below illustrates, this was a mistake – the IPGOD data is missing information on most of the divisional applications filed in December 2018.

Including the missing data results in the following corrected chart showing the total number of Australian standard patent applications filed each year, from 2000 to 2019, along with the corresponding numbers of ‘original’ and divisional applications.

This chart indeed shows that divisional applications falling from 5,745 to 5,578 made a greater contribution to the overall decline in Australian standard patent applications in 2019 – from 29,960 to 29,787 – than ‘original’ applications, which fell negligibly from 24,215 to 24,209.  This is the second consecutive year in which divisional application numbers have fallen, following a peak in 2017, however it is the first time since 2016 that there has not been growth in the number of ‘original’ applications filed in Australia.

Unpacking the Stats on Divisional and ‘Original’ Applications

This does not, however, explain the statement in the IP Report that ‘[d]ivisional applications fell 2.9 per cent in 2019, while original applications fell 1.2 per cent,’ which does not match up with the above data, and which I previously criticised as being mathematically inconsistent with an overall decline of just 0.7% in total applications.  The explanation from IP Australia is that they computed these numbers by separating ‘original’ applications into PCT national phase filings (of which there were 20,914 in 2018 and 20,940 in 2019) from ‘direct’ filings (9,046 in 2018 versus 8,847 in 2019), and then further dividing the ‘direct’ filings into ‘original’ and divisional applications.  There were 3,301 ‘direct-original’, and 5,745 divisional applications in 2018, versus 3,269 ‘direct-original’ and 5,578 divisional applications in 2019.  Considering only this subset of applications, the decline in divisional applications was 2.9%, while the decline in ‘direct-original’ applications was 1.0%.  (The 0.2% discrepancy between my numbers and IP Australia’s results from a counting difference of only five applications over the year, due to differences in the live data at extraction, and is not significant.)

Personally, I am of the view that this comparison of only ‘direct-original’ applications with divisional applications is essentially meaningless.  The reasons why applicants choose to file directly in Australia, rather than filing via the PCT system, are completely unrelated to divisional filings, and I cannot see any logical basis for separating and comparing only the ‘direct-original’ applications.  IP Australia’s analysts obviously do not agree with me on this, which is entirely their prerogative!  Readers of this blog, and of the IP Report, can make up their own minds.

Top Applicants of 2019

I was unable to reproduce IP Australia’s numbers for the top five resident (Australian) and non-resident patent applications for 2019.  IP Australia has conceded that there were problems with the way its numbers were derived, and we have worked together to reconcile discrepancies between our datasets.  The online version of the IP Report will be updated to include the correct numbers.

For the record, the tables below list the overall top-20 and the Australian resident top-10 applicants, ranked according to IP Australia’s criterion of ‘original’ applications.  These table are based on my data, which I have been able to improve slightly as I worked through the differences with IP Australia.  As a result, there are a few applicants whose numbers have been slightly adjusted (mostly by just one or two applications) from those in my earlier article.  The top five in each of these lists should agree with the corrected lists in the IP Report.

Overall Top-20 Applicants by ‘Original’ Filings 2019

Rank	Applicant Name	Country	‘Original’ Filings	Total Filings
1	GUANGDONG OPPO MOBILE TELECOMMUNICATIONS LTD	CN	313	314
2	QUALCOMM INC	US	221	244
3	LG ELECTRONICS INC	KR	218	246
4	HUAWEI TECH LTD	CN	170	188
5	ALIBABA GROUP HOLDING LTD	KY	156	159
6	NOVARTIS AG	CH	145	185
7	HALLIBURTON ENERGY SERVICES INC	US	126	136
8	MAGIC LEAP INC	US	109	116
9	COVIDIEN LP	US	101	138
10	ILLUMINA INC	US	99	115
11	ARISTOCRAT TECH AUSTRALIA PTY LTD	AU	96	239
12	BIOSENSE WEBSTER ISRAEL LTD	IL	89	108
13	SAMSUNG ELECTRONICS LTD	KR	89	106
14	COLGATE-PALMOLIVE CO	US	83	109
15	KIMBERLY-CLARK WORLDWIDE INC	US	83	88
16	BOEING CO	US	78	85
17	ACCENTURE GLOBAL SOLUTIONS LTD	IE	77	117
18	BASF SE	DE	71	74
19	CHINA UNIVERSITY OF MINING & TECH	CN	67	67
20	ELECTROLUX APPLIANCES AB	SE	66	67

Australian Resident Top-10 Applicants by ‘Original’ Filings 2019

Rank	Applicant Name	‘Original’ Filings	Total Filings
1	ARISTOCRAT TECH AUSTRALIA PTY LTD	96	239
2	COMMONWEALTH SCIENTIFIC & INDUSTRIAL RESEARCH ORGANISATION	47	52
3	UNIVERSITY OF SYDNEY	19	21
4	NEWSOUTH INNOVATIONS PTY LTD	17	18
5	UNIVERSITY OF QUEENSLAND	14	16
6	MONASH UNIVERSITY	13	15
7	UNIVERSITY OF MELBOURNE	12	12
8	FASTBRICK IP PTY LTD	10	13
9	CCL SECURE PTY LTD	10	10
10	BREVILLE PTY LTD	9	15

Tags: Australia, IPGOD, Patent analytics, Patent Office