Not all patents are created equal. While patent counts measure the quantity of innovation, researchers have developed a rich set of indicators to assess patent quality -- the technological significance, breadth, and downstream impact of individual inventions. This chapter draws on the framework established by Jaffe & de Rassenfosse (2017) to examine how multiple dimensions of patent quality have evolved over five decades of US patenting.
Claims & Complexity
Claims Per Patent Over Time
Average and median number of claims per utility patent, 1976-2025. Claims define the legally protected boundaries of an invention.
The doubling of average claims since the 1970s reflects both increasing invention complexity and strategic behavior by applicants seeking broader legal protection.
Citation Impact
Citation Impact Over Time
Average and median forward citations received within 5 years of grant, by grant year (limited to patents through 2020).
The decline in average citations partly reflects citation dilution — more patents means each individual patent receives a smaller share of total citations.
Forward citations -- how often a patent is cited by subsequent inventions -- are the most widely used indicator of patent quality in the economics literature. A patent that receives many forward citations has produced knowledge that others found valuable enough to build upon.
Backward Citations Over Time
Average and median backward citations (references to prior art) per utility patent, 1976-2025.
The growing body of backward citations reflects the expanding universe of prior art and more thorough examination and disclosure requirements over time.
Scope & Breadth
Patent Scope Over Time
Average and median number of distinct CPC subclasses per patent, measuring technological breadth.
Broadening patent scope reflects the convergence of once-separate technology domains, with modern inventions in areas like IoT, biotech, and AI spanning multiple classification categories.
Originality & Generality
The originality index (Trajtenberg, Henderson, & Jaffe 1997) measures how diverse the technology sources of a patent are -- a patent that cites prior art across many different CPC sections is more "original" than one citing only its own field. The generality index measures the reverse: how diverse the citing patents are, capturing whether an invention has broad applicability across fields.
Originality and Generality Indices
Average originality (1 - HHI of backward citation CPC sections) and generality (1 - HHI of forward citation CPC sections) by year. Higher = more diverse.
Rising originality scores indicate that modern inventions increasingly synthesize knowledge from diverse technology fields, consistent with growing interdisciplinary research.
Self-Citation
Self-Citation Rate Over Time
Average self-citation rate per patent (fraction of backward citations to patents held by the same assignee), by year.
Self-citation patterns reveal knowledge accumulation strategies within firms, with changes over time reflecting shifts between exploration of new domains and exploitation of established competencies.
Breakthrough Inventions
Breakthrough Patent Rate
Percentage of patents in the top 1% of forward citations within their year-technology cohort, 1976-2020.
Variation in the breakthrough rate over time reveals whether the right tail of the citation distribution is shifting, indicating changes in how concentrated inventive impact is among top patents.
Quality Across Sectors
Patent Complexity by Technology Sector
Average claims per patent by WIPO sector over 5-year periods. Shows how patent complexity varies across technology domains.
Biotech and pharma patents tend to have higher citation impact per patent, reflecting the slower but more impactful nature of pharmaceutical innovation.
Sleeping Beauty Patents
Some patents receive little attention for years, only to be "rediscovered" when technology catches up. These "sleeping beauties" received fewer than 2 citations per year for their first 10 years, then experienced a burst of 10+ citations in a 3-year window. They represent ideas that were ahead of their time.
Composite Quality Index
Individual quality metrics each capture one dimension of patent value. By combining forward citations, claims count, technology scope, and grant speed into a single Z-score normalized composite index, we can track overall patent quality trends across technology areas. Positive values indicate above-average quality; negative values indicate below-average.
Composite Patent Quality Index by Technology Area
Z-score normalized composite of forward citations (5yr), claims, scope, and grant speed. Values above 0 = above average.
Patents with high generality scores represent foundational innovations that influence many downstream technology areas, while the overall downward trend may reflect the system granting more patents of lower individual impact.
Quality vs. Quantity by Country
Do countries that patent more also patent better? Comparing average patent claims — a rough proxy for patent scope — across countries reveals that volume and quality do not always go hand in hand.
Average Claims by Country (Latest Decade)
Average number of claims per patent by primary assignee country, for the most recent decade. Higher claim counts generally indicate broader patent scope.
Countries with smaller patent portfolios sometimes achieve higher average claim counts, suggesting a quality-over-quantity approach. The US leads in both volume and average claims, while fast-growing patent origins like China show lower average claims — consistent with research on early-stage patent system development.
Self-Citation and Corporate Knowledge Recycling
Self-citations — where a company cites its own earlier patents — reveal the extent to which firms build on their own knowledge base. High self-citation rates can indicate deep, cumulative R&D programs, but may also reflect strategic behavior to build defensive patent thickets.
Self-Citation Rates by Top Assignees
Fraction of all backward citations that are self-citations (citing the same assignee's earlier patents), for the top 20 most-cited assignees.
Companies with deep, cumulative R&D programs — like IBM, Samsung, and semiconductor firms — show the highest self-citation rates, reflecting long-term knowledge building on internal prior art.
Having examined how we measure patent quality -- from citation impact to originality, generality, and composite indices -- the next chapter turns to the legal and policy framework that shapes the patent system itself. The rules governing what can be patented, how patents are examined, and how they are enforced have profound effects on the quality and quantity of innovation.
Quality indicators computed from PatentsView data following the framework of Jaffe & de Rassenfosse (2017). Forward citations use a 5-year window and are limited to patents granted through 2020 for citation accumulation. Originality and generality use the Herfindahl-based measures of Trajtenberg, Henderson, & Jaffe (1997). Breakthrough patents are defined as the top 1% of forward citations within each year-technology cohort. The composite quality index combines Z-score normalized forward citations (5-year window), claims count, technology scope, and grant speed (inverted). Sleeping beauty patents are identified as those with fewer than 2 citations per year in their first 10 years followed by a burst of 10+ citations in a 3-year window.