Advancing IP Valuation: The Role of Econometric Approaches

When a company holds a patent, one of the most pressing questions it faces is: what is this patent actually worth? Traditional valuation methods – cost-based, income-based, or market-based – often fall short because patent value is notoriously difficult to pin down. A patent’s worth isn’t just about what it cost to develop or what royalties it might earn; it’s about how significant the underlying invention is in the broader technological landscape. This is precisely where econometric approaches step in. By using statistical models and large-scale empirical data, econometric methods offer a more objective, evidence-driven lens for understanding patent value – and two of the most powerful tools in this toolkit are citation analysis and patent renewal data.

What are econometric approaches to patent valuation? 🔗

Econometric approaches apply quantitative and statistical methods to estimate the value of patents by analysing observable data rather than relying on subjective projections. The core idea is straightforward: instead of asking “what do we think this patent is worth,” econometrics asks “what does the data tell us about this patent’s significance?” These methods draw on large patent datasets maintained by offices like the Indian Patent Office, the United States Patent and Trademark Office (USPTO), and the European Patent Office (EPO) to identify patterns between measurable patent characteristics and economic value.

The appeal of these approaches lies in their empirical foundation. They are based on what actually happened – how often a patent was cited, how long it was maintained, whether it was litigated – rather than on forecasts that could be skewed by optimism or strategic bias. Two indicators dominate this field: citation frequency and renewal behaviour.

Citation analysis: what citations reveal about patent quality 🔗

Every time a new patent is filed, the applicant (and sometimes the patent examiner) must identify prior art – earlier patents or technical documents that the new invention builds upon or relates to. These references create a citation network, and the frequency with which a patent is cited by subsequent filings has become one of the most studied proxies for patent value in the economics literature.

The underlying logic is intuitive: if many later inventors cite a particular patent, it likely represents a foundational or widely applicable technological contribution. Research on patent citation data consistently shows that citation counts are positively correlated with several dimensions of a patent’s importance – its technological impact, its private economic value to the holder, and even its social value in terms of knowledge diffusion.

Forward and backward citations 🔗

Citation analysis distinguishes between two types of citations. Forward citations are the number of times a given patent is subsequently cited by newer patents – a measure of how much later innovation builds on it. Backward citations refer to the prior art the patent itself cites. Both carry information. Studies by Harhoff, Scherer, and Vopel (2003) found that both the number of forward citations and backward references to patent literature were significantly correlated with patent value as self-reported by patent holders. Notably, backward citations are especially useful in practice, since they are available at the time of grant – unlike forward citations, which can only accumulate over time.

The landmark Harhoff et al. study 🔗

One of the most cited empirical validations of citation analysis comes from Harhoff, Narin, Scherer, and Vopel (1999), published in the Review of Economics and Statistics. The researchers surveyed holders of 962 inventions in the US and Germany – patents that had been renewed to full term – and obtained their own economic value estimates. They then cross-referenced these estimates with forward citation counts. The finding was clear: patents with higher economic value were significantly more heavily cited in subsequent patents. This was one of the first large-scale empirical confirmations that citation frequency carries real information about a patent’s market worth.

Similarly, a landmark study by Hall, Jaffe, and Trajtenberg (2005) in the RAND Journal of Economics demonstrated that citation-weighted patent counts were highly correlated with firm market value, providing evidence that citations carry information about the value of patented innovations well beyond simple patent counts.

Citation diversity and technological breadth 🔗

Beyond sheer volume, the diversity of citations matters. A patent cited across multiple technology classes signals broader applicability – an indicator that the invention has relevance in more than one field. Methodological reviews of patent valuation note that such cross-domain citations often correlate with higher strategic value, since the patent can potentially block or license across a wider range of commercial applications. A patent cited only within a narrow technical subfield may be technologically sound but commercially restricted in comparison.

Citations and trade probability 🔗

Studies have also explored the link between citation frequency and a patent’s probability of being traded or licensed in the market. Research across manufacturing, pharmaceutical, and semiconductor industries has confirmed that higher citation counts are associated with better firm performance metrics – including stock returns and market-to-book ratios – and that more highly cited patents are more likely to be the subject of licensing negotiations, acquisitions, and litigation. In short, a patent’s citation record is a signal that the market itself reads and responds to.

Patent renewal data: decisions that reveal value 🔗

The second major econometric tool is renewal data. In most patent jurisdictions, a patent does not remain in force automatically – the holder must pay periodic maintenance fees to keep it alive. In India, as mandated under Section 53, Rule 80 of the Patents Act, 1970, renewal fees are payable from the third year onwards on an annual basis. A patent that is not renewed lapses, entering the public domain – unless restored through an application under Section 60 within 18 months of cessation.

Renewal as an economic signal 🔗

The core insight of renewal-based econometric models is elegant: a rational patent holder will only continue paying fees if the expected future value of the patent exceeds those costs. When a holder chooses not to renew, they are making a revealed preference decision – signalling that the patent is no longer worth maintaining. When they continue paying, especially as fees escalate in later years, they are signalling continuing commercial relevance. Research by Bessen (2008) modelled patent renewal decisions as an ordered probit framework, combining renewal data with citation counts and patent holder characteristics to obtain dollar estimates of patent value increments attributable to each additional citation.

This approach was pioneered in the economics literature by Pakes (1986), who treated patents as financial options – the holder has the right, but not the obligation, to maintain protection for another year by paying the renewal fee. Each year’s renewal decision is essentially a fresh calculation of whether the patent’s continuing value exceeds its cost. This model has since been extended to Indian patent data, with studies using full-length renewal information to estimate the average commercial life of Indian patents across technology sectors, comparing results with US, European, and Chinese patent renewal patterns.

Skewed value distributions 🔗

One of the most consistent findings from renewal-based econometric research is that patent values are highly skewed. The vast majority of patents have modest or negligible commercial value, while a tiny fraction – often described as the “long tail” – account for a disproportionate share of total patent value. Harhoff, Scherer, and Vopel’s analysis of German patent value distributions found that the log-normal distribution best describes this skewness, with the top patents contributing outsized value. This has direct implications for IP valuation practice: assigning an average value to patents in a portfolio is likely to be misleading, since a few high-value patents dominate the portfolio’s worth.

Combining citation and renewal approaches 🔗

The most robust econometric valuations combine both citation data and renewal behaviour rather than treating them as separate tools. Work by Lanjouw and Schankerman (2003) demonstrated that a composite indicator – incorporating citation counts, number of patent claims, and number of countries in which the invention is protected – has strong predictive power for identifying which patents will be renewed and which will be litigated. This composite approach addresses a weakness of relying on either indicator alone: citation counts are noisy signals, and renewal decisions reflect the holder’s private information which may not be publicly observable.

For IP professionals and business managers in India, this has practical consequences. When evaluating a patent portfolio – whether for acquisition, licensing, litigation strategy, or balance sheet reporting – combining the patent’s citation history with its renewal track record provides a more complete picture of value than either metric alone.

Limitations and emerging directions 🔗

Econometric approaches are not without limitations. Citation counts can be inflated by differences in institutional practices – the USPTO requires applicants to disclose all known prior art under a duty of candour, while the EPO does not impose this obligation, meaning citation volumes are not directly comparable across jurisdictions. Furthermore, as patent data research highlights, citation frequency measures technological diffusion but may miss foundational innovations in systemic technologies where influence is structural rather than directly cited. Additionally, a patent’s strategic value – its capacity to block competitors – is difficult to capture in citation or renewal data alone.

The field is evolving to address these gaps. Machine learning algorithms are increasingly applied to patent datasets, identifying non-obvious patterns across larger samples than traditional regression models can handle. Network analysis goes beyond counting citations to map the structural position of a patent within the citation graph – patents that bridge different technology clusters often represent uniquely valuable innovations. Text mining of patent claims themselves is another frontier, allowing analysts to extract value-relevant signals from the language of the patent document.

Relevance for Indian IP practice 🔗

For students and practitioners dealing with Indian IP management, understanding these econometric tools is increasingly important. India’s patent landscape is maturing, with growing volumes of domestic patent filings across pharmaceuticals, information technology, and manufacturing. As the Indian Patent Office continues to digitise and expand its database, the availability of renewal and citation data for Indian patents is improving – making these econometric approaches progressively more applicable to domestic portfolio analysis. Whether advising a startup on the commercial strength of its technology, or helping a large enterprise determine which patents in a portfolio are worth maintaining into their later, higher-fee years, citation and renewal analysis provide a rigorous and data-backed framework for decision-making.

What do you think? If a patent consistently receives forward citations from multiple unrelated technology sectors but is allowed to lapse in its tenth year, what does that combination of signals tell you about its value – and about the patent holder’s strategy? And should Indian IP valuation practice invest more systematically in building citation-based quality indicators for domestic patents, similar to what the USPTO and EPO datasets enable globally?

References

1. https://ipindia.gov.in/
2. https://www.nber.org/system/files/working_papers/w21868/w21868.pdf
3. https://asistdl.onlinelibrary.wiley.com/doi/10.1002/asi.23731
4. https://www.researchgate.net/publication/24095603_Citation_Frequency_And_The_Value_Of_Patented_Inventions
5. https://eml.berkeley.edu/~bhhall/papers/HallJaffeTrajtenberg_RJEjan04.pdf
6. https://journals.uniurb.it/index.php/ijmeb/article/download/4797/4497
7. https://www.sciencedirect.com/science/article/abs/pii/S1751157711000782
8. https://ssrana.in/ip-laws/patents/patents-annuity-payment-india/
9. https://www.puthrans.com/restoring-a-lapsed-patent-in-india-process-deadlines-and-best-practices/
10. https://scholarship.law.bu.edu/cgi/viewcontent.cgi?article=4180&context=faculty_scholarship
11. https://arxiv.org/pdf/2208.06157
12. https://link.springer.com/chapter/10.1007/978-1-4757-3750-9_13
13. https://www.nber.org/system/files/working_papers/w17773/w17773.pdf
14. https://www.intechopen.com/chapters/54656