Last week, IP Australia released the Australian Intellectual Property Report 2020 (‘IP Report’). As has become customary over the past few years, the launch of the IP Report was timed to coincide with World IP Day, which fell on Sunday 26 April 2020. As my own analysis, published back in January, indicated, the number of Australian standard patent applications fell in 2019, by just under 1%. The IP Report also confirms my expectation that Chinese patent applicants would surpass Germany and Japan, and push the UK out of the top five, to take third spot in the annual country rankings. Applications from China experienced year-on-year growth of over 45%!
I anticipated that IP Australia would have difficulty placing a positive spin on the 2019 patent filing numbers, given that the decline in 2019 follows two years of growth in patent applications. However, this has not stopped them from trying, by introducing a new metric which counts ‘original’ applications – i.e. those based on patent specifications being filed for the first time in Australia – as distinct from divisional applications – i.e. those based on the contents of earlier filings. The apparent implication of the use of this metric in the IP Report is that divisional applications should be ascribed a lower ‘value’ (in some sense) than ‘original’ applications, such that a decline in divisional filings is somehow less significant than a decline in ‘original’ applications.
Following this same approach, IP Australia has ranked the top applicants of 2019 according to the number of ‘original’ applications filed, ignoring divisional filings. On this measure, Australia’s biggest filer, Aristocrat Technologies, drops from fourth place, with a total of 238 standard patent applications, to tenth place, with just 96 of those applications being ‘original’ rather than divisional. An even bigger loser, however, is Apple Inc, which falls from eighth place, with 152 total applications by the conventional measure, to 41st place on the ‘original filings’ measure (45 qualifying applications), behind Australia’s CSIRO (46 ‘original’ applications, out of 51 total filings).
As I shall explain, I believe that there are a number of problems with the use of ‘original filings’ as a metric. Furthermore, the ‘Patents’ section of the IP Report unfortunately includes results that I have been unable to reproduce, some significant discrepancies between IP Australia’s data and my own analysis, and statements that do not make sense, even on their own terms. This is a pity, because the annual IP Report is generally a useful summary of activity during the previous year, and it is important that it should be a trustworthy source of information.
Problems with a Novel Measure of Filing Intensity
It is difficult to identify any intrinsic merit to IP Australia’s ‘original filings’ metric. I guess I can understand the theory they are working on, i.e. that a divisional application is, in a sense, not ‘new’, but rather it is a further application for an invention that has already been described in one or more earlier-filed applications. So the reasoning might be that if the applicant has done (i.e. invented) nothing new, why should it get full credit for the further patent application? In my view, however, this theory of divisional ‘value’ is deeply problematic, for at least a couple of reasons.Firstly, nobody counts patent applications this way. IP Australia has not used this measure in the past, and no other major jurisdiction (to my knowledge) provides separate numbers for ‘original’ and ‘follow-on’ applications. (I avoid using the term ‘divisional’ in this context, because some jurisdictions – most notably the US, via ‘continuation’ and ‘continuation-in-part’ applications – provide other follow-on filing options.) As a result, there are no points of comparison, either historically, or internationally, with the ‘original filings’ metric.
Secondly, not all divisional applications are created equal. Some really are filed primarily as a way to buy additional examination time, with the parent application being allowed to lapse. Such applications genuinely add nothing new to the roster of inventions represented by the patent records (even then, however, the fact that the applicant is willing to continue to spend money pursuing patent rights says something). At the other end of the spectrum there are divisional applications that are filed for the primary purpose for which they exist – to enable an applicant to obtain a patent for a distinct (second or subsequent) invention disclosed, but ultimately not claimed, in the original application. While such applications could have been filed independently in the first place, the applicant – whether intentionally, or otherwise – instead filed one application describing multiple inventions. There is no good reason not to treat these divisional applications as being fully equivalent in value to new, original, applications. And these are just the two extremes. In between, there are divisional applications filed to pursue claims with different and/or variously overlapping scope with those of the original application.
Taking Apple as an example, the company ranked seventh among recipients of US patents in 2019, roughly the same as its ranking in total Australian patent filings last year. We do not know, because it has not been reported, how many of Apple’s US patents resulted from divisional, continuation, or continuation-in-part (CIP) applications, or how the company’s use of these types of US follow-on applications compares with its use of divisional applications in Australia. Nor do we know why Apple files such a large proportion of divisional applications in Australia. It might simply be in the habit – either by accident or (more likely) by design – of filing patent specifications that disclose multiple inventions, and/or of seeking patents of varying scope for strategic reasons. Whatever the purpose, it seems very likely that Apple’s strategy in Australia is mirrored in other jurisdictions. And a metric that ranks a company like Apple near the bottom of the top 50 innovators, instead of in the top 10 where it clearly belongs, is inherently questionable.
A further point worth noting is that the 2020 IP Report also presents data on the number of applications filed overseas by Australian residents, noting that this figure increased by 3%, to 9023, between 2017 and 2018 (the most recent year for which data is available). To be consistent, the Report should also distinguish here between ‘original’ applications, and divisional (and other follow-on) applications, so that we can make a fair comparison on the basis of the ‘original filings’ metric. But it does not, not least because this data is not readily available.
Effect of ‘Original Filings’ Metric on Applicant Rankings
The table below lists the top 20 applicants for Australian standard patents in 2019, ranked according to the conventional measure of total filings. The table also shows the number of filings by each applicant that were ‘original’ (the difference between the two is the number of divisional filings).| Rank | Applicant Name | Country | Total Filings | ‘Original’ Filings |
|---|---|---|---|---|
| 1 | GUANGDONG OPPO MOBILE TELECOMMUNICATIONS LTD | CN | 314 | 313 |
| 2 | LG ELECTRONICS INC | KR | 245 | 217 |
| 3 | QUALCOMM INC | US | 244 | 221 |
| 4 | ARISTOCRAT TECH AUSTRALIA PTY LTD | AU | 238 | 96 |
| 5 | HUAWEI TECH LTD | CN | 187 | 169 |
| 6 | NOVARTIS AG | CH | 185 | 145 |
| 7 | ALIBABA GROUP HOLDING LTD | KY | 155 | 152 |
| 8 | APPLE INC | US | 152 | 44 |
| 9 | COVIDIEN LP | US | 136 | 99 |
| 10 | HALLIBURTON ENERGY SERVICES INC | US | 135 | 125 |
| 11 | MAGIC LEAP INC | US | 116 | 109 |
| 12 | ACCENTURE GLOBAL SOLUTIONS LTD | IE | 116 | 76 |
| 13 | ILLUMINA INC | US | 112 | 96 |
| 14 | COLGATE-PALMOLIVE CO | US | 109 | 83 |
| 15 | BIOSENSE WEBSTER ISRAEL LTD | IL | 108 | 89 |
| 16 | SAMSUNG ELECTRONICS LTD | KR | 106 | 89 |
| 17 | KIMBERLY-CLARK WORLDWIDE INC | US | 88 | 83 |
| 18 | BECTON DICKINSON & CO | US | 85 | 52 |
| 19 | BOEING CO | US | 83 | 76 |
| 20 | F HOFFMANN-LA ROCHE AG | CH | 80 | 63 |
The following table lists the top 20 applicants according to IP Australia’s ‘original filings’ metric (total filings are also again shown – I have swapped the column order). There are some significant differences from the conventional ranking, the most notable of which is Apple’s disappearance from the table – as I have already noted, it ranks 41st on this measure. Aristocrat is also a loser, slipping from fourth on the conventional metric to tenth on the ‘original filings’ metric. Beneficiaries include Magic Leap (leaping magically from 11th to eighth), Kimberly-Clark (up three places to 14th), and Electrolux (up from 26th to join the top 20 at 19th).
| 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 | 217 | 245 |
| 4 | HUAWEI TECH LTD | CN | 169 | 187 |
| 5 | ALIBABA GROUP HOLDING LTD | KY | 152 | 155 |
| 6 | NOVARTIS AG | CH | 145 | 185 |
| 7 | HALLIBURTON ENERGY SERVICES INC | US | 125 | 135 |
| 8 | MAGIC LEAP INC | US | 109 | 116 |
| 9 | COVIDIEN LP | US | 99 | 136 |
| 10 | ARISTOCRAT TECH AUSTRALIA PTY LTD | AU | 96 | 238 |
| 11 | ILLUMINA INC | US | 96 | 112 |
| 12 | BIOSENSE WEBSTER ISRAEL LTD | IL | 89 | 108 |
| 13 | SAMSUNG ELECTRONICS LTD | KR | 89 | 106 |
| 14 | KIMBERLY-CLARK WORLDWIDE INC | US | 83 | 88 |
| 15 | COLGATE-PALMOLIVE CO | US | 83 | 109 |
| 16 | ACCENTURE GLOBAL SOLUTIONS LTD | IE | 76 | 116 |
| 17 | BOEING CO | US | 76 | 83 |
| 18 | BASF SE | DE | 71 | 74 |
| 19 | ELECTROLUX APPLIANCES AB | SE | 66 | 67 |
| 20 | TELEFONAKTIEBOLAGET LM ERICSSON PUBL | SE | 65 | 72 |
I have used the same dataset for these rankings as in my article from the beginning of this year. It is important to understand, however, that patent records are not static – they ‘live and breathe’ as changes and corrections are made, rights-holders conduct and record transactions such as assignments, and other events occur. The data used for the above tables was obtained throughout the course of 2019, and largely reflects the initial information recorded when each application was filed (or, at least, within a month or so afterwards). By way of example, however, Novartis recently recorded assignments of a significant number of Australian patent applications to Alcon, reducing its ‘apparent’ number of 2019 applications (listed as 185 in the tables above), if current data were used, to 143.
Changes in recorded details may explain minor differences between the numbers above, and those in IP Australia’s report, which are based on data extracted from its databases more recently, and which thus reflect further changes that have occurred up until the extraction date. However, such changes probably cannot explain significant discrepancies in trends or filing statistics, which also appear to exist in the IP Report.
Should We Trust IP Australia’s Numbers?
It appears that one of IP Australia’s reasons for introducing the ‘original filings’ metric is an attempt to make the decline in overall filings seem less significant. Regarding total standard patent filings, the IP Report says:In 2019, a total of 29 758 applications for standard patents were filed, a 0.7 per cent decrease from their level in 2018…. The fall can be largely attributed to a decrease in divisional applications—patent applications which contain subject matter from previously filed applications. Divisional applications fell 2.9 per cent in 2019, while original applications fell 1.2 per cent.
On Australian resident filings in particular, it says:
Resident applications fell by 4.3 per cent in 2019 (from 2 756 to 2 637), but most (87 per cent) of the fall in resident applications is attributable to a decrease in divisional applications.
I have two problems with these numbers. Firstly, even on their own terms they do not make sense. It is impossible for an overall decline of 0.7% to result from a 2.9% decline in divisional applications and a 1.2% decline in ‘original’ applications. In recent years, divisional applications have made up between 20-25% of all standard patent applications, so without even looking at the specific numbers for 2019, a 2.9% fall in divisional applications and a 1.2% fall in ‘original’ applications should correspond roughly with a 1.5-1.6% decline overall. Furthermore, for divisional applications to be the main contributor to an overall decline, the percentage fall in divisional filings would need to be 3-4 times the percentage fall in ‘original’ filings.
Even more worryingly, however – either for me, or for IP Australia – I am unable to replicate the results in the 2020 IP Report. I have endeavoured to do so using data from the 2019 IPGOD release (which includes details of applications filed up until the end of 2018) in combination with 2019 data derived from IP Australia’s online AusPat service (see my earlier article for a brief discussion of the pros and cons of this data source). In contrast with the IP Report, I find an increase in the overall number of divisional applications in 2019, with ‘original’ filings thus showing a larger, rather than smaller, relative decline. For Australian residents, I find a 4.2% decline in ‘original’ filings, and a 4.6% decline in divisional filings, and that only 23% of the fall in resident applications is attributable to a decrease in divisional applications.
Trends in ‘Original’ and Divisional Filings
The chart below shows 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. It is interesting to note that the number, and proportion, of divisional filings have generally grown throughout this period, and that divisional filing numbers were significantly less affected by the global financial crisis (GFC) of 2007-8 than ‘original’ filings.The ‘bringing forward’ of filings into 2013 in anticipation of the commencement of the Raising the Bar IP law reforms – which caused a spike in filings, followed by a dip in 2014 – affected both ‘original’ and divisional filing numbers. Stronger performance of divisional filings in 2016 and 2017 may have been due to a ‘follow-on’ effect, as a result of divisional applications being based on applications filed during the 2013 peak. It is also possible that the more recent drop in divisional filings is, similarly, at least partly a result of the 2014 dip, rather than indicating any underlying decline in demand for divisional applications.
Contrary to the statement in the 2020 IP Report, my data does not indicate a decrease in divisional applications in 2019. Rather, the above data shows a 4.7% increase in divisional filings, along with a 2.2% decline in ‘original’ filings relative to 2018.
The following chart shows the corresponding numbers of applications filed by Australian residents annually, between 2000 and 2019. (Note that I count each application only once, regardless of number of applicants, and define a ‘resident application’ as ‘any application that names at least one Australian-resident applicant’.)
While there was growth in ‘original’ resident applications, relative to divisional applications, in 2018, on these figures both ‘original’ and divisional applications by Australian residents declined in 2019, by 4.2% and 4.6% respectively. Again, this result conflicts with the 2020 IP Report, which claims that the decline in resident applications was dominated by a fall in divisional filings.
Leading Resident Applicants
In addition to the issues discussed above, there also appears to be a problem with IP Australia’s tally of the top Australian resident applicants. Again using ‘original filings’ as a metric for ranking, the five leading domestic applicants are identified in the IP Report as Aristocrat (96 ‘original’ applications), CSIRO (42), NewSouth Innovations (15), University of Queensland (13), and University of Sydney (11). The table below shows the top 10 domestic applicants according to my data, ranked according to the ‘original filings’ measure, and also showing total filings (i.e. including divisional applications).| Rank | Applicant Name | ‘Original’ Filings | Total Filings |
|---|---|---|---|
| 1 | ARISTOCRAT TECH AUSTRALIA PTY LTD | 96 | 238 |
| 2 | CSIRO | 46 | 51 |
| 3 | UNIVERSITY OF SYDNEY | 18 | 20 |
| 4 | NEWSOUTH INNOVATIONS PTY LTD | 16 | 17 |
| 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 |
While small discrepancies in application counts might be expected, due to the timing of extraction of data as discussed above, the results in the IP Report somehow miss seven ‘original’ applications filed during 2019 by the University of Sydney. Applications by Monash University and the University of Melbourne also must have been missed, because both exceed the 11 filings attributed to fifth-placed University of Sydney in the IP Report.
Conclusion
The IP Report has become a useful annual summary of activity relating to Australian IP rights during the previous year, and it is therefore important for it to be a reliable and accurate source of information. This year, however, I have some concerns about the ‘Patents’ section of the report.- Using ‘original filings’ – i.e. excluding divisional filings – as a metric for filing intensity is unconventional, lacks historical and international points of comparison, and is of questionable merit without further information regarding scope, and reasons for filing, of divisional applications.
- IP Australia’s motivation for introducing this new metric appears to be, at least partly, in an attempt to present a decline in Australian patent filings in 2019 in a more favourable light.
- I have been unable to independently reproduce a number of IP Australia’s results, with my findings contradicting statements made in the 2020 IP Report.
I cannot be sure that results in the IP Report are wrong, or that mine are right. However, the fact that the Report contains mathematically impossible assertions – i.e. that a 0.7% overall decline in filings corresponds with a 2.9% decline in divisional applications and a 1.2% decline in ‘original’ applications – does not inspire confidence. Hopefully IP Australia can revisit its analysis, and address or refute these concerns.
Tags: Australia, Patent analytics, Patent Office
0 comments:
Post a Comment