Invention Is Not Yet Scale
This chapter examines how innovative start-ups in the EU and the US move from invention to commercial scale, using patenting as a window into that translation rather than as a simple score of technological ability. The practical question is whether a firm can formalise and protect knowledge, navigate the institutional steps around it, and turn that knowledge into market expansion quickly enough. The evidence suggests that the EU–US difference is less about the inventive intensity of firms that patent than about who formalises innovation and how long the post-filing path takes.
A Patent Is an Institutional Decision, Not an Innovation Score
A patent is often treated as a convenient proxy for innovation. It is certainly evidence that a firm has chosen to seek formal legal protection for an invention. But that choice is not made in a vacuum. Filing creates costs, procedural work, decisions about jurisdictions, and later questions about enforcement and commercial use. A start-up may also rely on trade secrets, a platform position, rapid iteration, or another appropriation strategy rather than disclose an invention through the patent system. A lower patent rate therefore does not automatically mean a lower rate of invention.
The OECD study separates two margins of patenting. The extensive margin asks whether a start-up patents at all. The intensive margin asks how much inventive activity is visible among firms that do patent. Keeping the two margins apart prevents a common analytical mistake: inferring that a region has weak inventive capacity because fewer of its start-ups appear in the patent database. The first observation may instead reflect differences in the cost, expected value, or timing of formal protection.
The extensive-margin comparison is substantial. Among rising superstars, 18.44% of EU firms hold at least one patent, compared with 51.51% of US firms. Among growth-oriented firms, the corresponding shares are 26.56% and 45.30%. On the surface, this looks like a large innovation gap. But the comparison is between all firms in each category, not just the firms that decided to patent. It tells us that formal IP protection is less common among EU scale-ups, not that the non-patenting firms made fewer valuable discoveries.
The intensive margin points in a different direction. Among patenting rising superstars, the average pre-scaling invention stock is 11.31 in the EU and 10.42 in the US. Among patenting growth-oriented firms, the EU average is 11.14 compared with 7.18 in the US. These figures are descriptive averages among patenting firms, not estimates of the value of every invention. They nevertheless complicate the idea that Europe’s advanced firms are simply less inventive. The subset that formalises its inventions can show comparable or higher recorded invention intensity.
The combined pattern is important: fewer EU scale-ups patent, but EU firms that do patent are not visibly weaker on the study’s invention-stock measure. Innovation appears more concentrated among a smaller formalising subset. That concentration can matter for scaling even if the underlying technical capacity is present. A firm that does not protect an invention may have a sound strategic reason, but it may also face higher uncertainty about enforcement, licensing, financing, or cross-border commercialisation. The aggregate data cannot determine which explanation applies to a particular company.
Consider two start-ups with similarly promising technical discoveries. The first files early because it expects patents to help attract investment, defend a market, or license a technology. The second delays filing because disclosure is costly, its business advantage lies in data or network effects, or it is focused on reaching customers before competitors. A later difference in patent ownership would not by itself establish that the first firm is more innovative. It would show that the firms made different formalisation choices under different commercial conditions. Those choices may then affect how investors, partners, and competitors evaluate the firms.
The OECD measure of invention is designed to improve comparability. Patent applications are grouped into DOCDB families so that the same invention filed in several jurisdictions is not counted repeatedly. The earliest filing year dates the invention, and annual invention flows are accumulated into a stock that depreciates over time. This is a more disciplined measure than simply counting every application, but it remains a measure of recorded patent-family activity. It cannot capture all knowledge, and it does not turn a patent into proof of commercial value.
That is why patenting is used here as an institutional lens. The relevant environment includes the incentives and costs around formal protection, the predictability of procedures, the ability to enforce rights, and the friction of operating across jurisdictions. The measure is not a direct regulatory index. It is an observable behaviour through which regulatory and commercial conditions may become visible.
The timing of the decision can matter as much as the decision itself. Filing early may help a firm signal technical seriousness or preserve options for licensing and investment, but it can also divert scarce attention and disclose information before the product or market is ready. Filing later may protect a more mature invention, yet leave the company with less time to establish a defensible position. Neither choice is inherently superior. Its value depends on the firm’s technology, appropriation strategy, expected markets, and ability to carry the associated legal and commercial work.
The important point for the scale-up problem is that formal protection has to connect to a route for using the protected knowledge. A patent that cannot be enforced, licensed, integrated into a product, or translated into customer adoption does not by itself create scale. Conversely, a firm with a non-patent advantage may scale without making its knowledge visible in patent records. The extensive margin therefore tells us about the distribution of formalisation choices, while the intensive margin tells us about the inventive activity visible among the firms that made those choices. Together they expose a possible concentration of formal innovation, but they do not rank one firm’s underlying technology against another’s.
The sequence illustrates why the extensive and intensive margins belong together. A firm first decides whether formal protection is worth pursuing; after filing, it still has to use the protected knowledge in a product, market, licensing strategy, or financing story. A patent can support that process, but it cannot complete it. The main scaling question is what happens between formalised invention and a commercial position large enough to meet the study’s growth or valuation threshold.
The Delay After First Filing
The timing evidence makes that question concrete. EU rising superstars tend to file their first patent at younger ages than US counterparts, yet take longer to scale after that first filing. For growth-oriented start-ups, the age at first patent is broadly similar across the two regions, but US firms reach the scaling threshold more quickly after filing. Europe can therefore show early and intensive patenting among the firms that patent while still taking longer to turn that formalised activity into advanced growth.
The report estimates this interval with an ordinary least squares model. The dependent variable is the number of years between a start-up’s first patent filing and its scaling event. The model includes an EU indicator, a rising-superstar indicator, citation-weighted patent stock accumulated before scaling, and fixed effects for sector and founding cohort. In the reported sample of 3,524 scale-ups, the EU coefficient is approximately +1.494 years. Conditional on those controls, EU patenting scale-ups take longer to reach a scaling milestone after first patenting than comparable US firms.
The controls change the question being answered. The estimate is not merely comparing an EU firm with every US firm. It accounts for differences associated with the broad sector, the founding period, and the recorded quality-weighted stock of patents. That makes the remaining regional difference consistent with frictions along the path from formalisation to commercialisation. It does not identify which friction is responsible, and it does not show that a particular regulatory reform would remove the delay.
Several mechanisms could produce such a pattern. Protecting and enforcing rights across multiple jurisdictions can add cost and uncertainty. Administrative or legal procedures may be slower or less predictable. A fragmented market can make it harder to commercialise one protected invention through a single large customer base. A company may also need to adapt its product, find partners, raise capital, or build an organisation before the invention can generate large-scale revenue. These mechanisms can reinforce one another: a delayed market launch creates a stronger need for financing, while a financing gap can postpone the hiring and market work needed to exploit the invention.
Imagine the difference between two otherwise promising firms after filing. One can use the filing as part of a financing story, secure a lead customer, and expand through a market with predictable rules. The other may need to repeat legal work across jurisdictions, negotiate with several kinds of buyers, and raise another round before it can hire the people needed to commercialise the invention. The first firm’s shorter interval is not necessarily evidence of a better idea. It may reflect a setting in which protection, capital, customers, and organisational capacity arrive in a more coordinated way. The second firm’s delay is not proof that a particular rule caused failure; it is evidence that the path contains more work or uncertainty.
This distinction also prevents a false choice between “patent more” and “ignore patents.” A policy that increases filing without improving predictability, enforcement, licensing, market access, or financing may raise the visible number of patents without shortening the path to scale. A policy that focuses only on commercial speed may overlook the need to protect knowledge when imitation or cross-border bargaining matters. The relevant test is whether formalisation and commercialisation are connected for the kinds of firms and technologies being supported.
This is why the phrase “Europe is less innovative” would misread the evidence. The report finds that EU scale-ups are less likely to patent, but among those that patent they have comparable or higher invention stocks on average. It also finds a longer patent-to-scale interval after controlling for patent stock, sector, and cohort. The pattern is more compatible with a problem of formalisation and translation than with a simple absence of knowledge creation.
Still, compatibility is not proof. Firms in the EU and US may differ in business model, use of trade secrets, platform economics, financing strategy, or market structure in ways not fully captured by the model. Patenting may occur at different moments because investors encourage different disclosure strategies. Some firms may choose not to patent precisely because their strongest advantage is difficult to protect through formal IP. The observed delay could therefore include choices and selection as well as institutional friction.
The distinction matters for diagnosis. Before blaming regulation, a decision-maker should ask what kind of delay is being observed. Is the firm slow to file because protection has low expected value? Is it filing early but waiting for approvals, partners, or cross-border enforcement? Is the invention protected but the market too fragmented to support rapid expansion? Is the problem insufficient late-stage capital or a management team not yet equipped to commercialise? The same number of years can be consistent with different bottlenecks, and the appropriate remedy would differ accordingly.
The report’s regional aggregation creates another boundary. The EU coefficient describes a pattern common enough to appear in an EU–US comparison, but it does not show that every member state has the same IP procedures, market conditions, or commercialisation performance. Country-level data could change the relative importance of EU-wide institutions and national differences. Firm-level measures of revenue, employment, licensing, market entry, and enforcement would also help distinguish a regulatory delay from a deliberate business strategy.
What the evidence does establish is narrower and more useful. Patent ownership is not a complete measure of inventive capacity. Among firms that patent, EU scale-ups do not appear weaker on the study’s invention-stock measure. Yet EU patenting scale-ups take longer, on average, to reach the relevant scaling threshold after first filing. That delay is a signal that invention and commercial scale are different achievements, connected by institutional, financial, market, and organisational work.
Once that distinction is clear, finance becomes the next constraint in the sequence. A firm that spends longer translating an invention into market scale must fund more time under uncertainty. The question is not only whether it can obtain an initial investment, but whether its environment can assemble the much deeper and more coordinated capital required when commercial expansion finally accelerates.