Innovation underpins sustained long-run economic growth. From rapidly advancing artificial intelligence to green technologies, governments in developed countries seek to stay close to the technological frontier and avoid falling behind. But how can technological success be tracked? A central challenge is measurement. Comparing technological trends across countries over long periods is difficult. The standard macroeconomic proxy of total factor productivity (TFP) has many problems such as shifting production structures and data availability, especially over centuries (Bergeaud et al. 2016). And research and development (R&D) spending measures inputs to innovation, not outputs.

Patent data, a direct record of inventive activity, is an attractive alternative. Yet patents also raise well-known issues. In particular, intellectual property offices differ across countries and over time, with a ‘home bias’ towards filing in an inventor’s own country (De Rassenfosse et al. 2013).

In a recent paper (Bergeaud et al. 2026), we address these difficulties by shifting the focus away from patenting in one country and toward a country’s inventive performance across multiple foreign patent offices. This exploits the fact that an important invention will be filed not only in the country where the potential patent holder (the ‘assignee’) lives, but also in other jurisdictions where she wants intellectual property protection. Consider some examples. If American inventive activity truly surged in the early 20th century, this boom should be visible not only in the US Patent Office, but also in patent offices in countries like the UK, France, and Germany. Similarly, a sustained decline in British inventive strength should be evident not just to UK examiners, but also those overseas.

To implement this idea, we extend the PatentCity database (Bergeaud and Verluise 2024) and construct a dataset covering domestic and foreign assignees between 1836 and 2016 across four major patent offices – in the US, UK, Germany, and France. We derive measures of innovative activity by country of origin and study their evolution over time. We call these patterns technological trajectories, defined as countries’ relative rates of innovation over time as revealed by international patenting trends.

How have technological trajectories evolved?

Several findings stand out. First, looking over the long-term, the American technological surge in the late 19th century following the second industrial revolution, and the acceleration after WWII, shows up simultaneously in the US Patent Office and in all three European patent offices. This broad-based rise suggests a genuine transformation in innovative capability rather than a purely domestic institutional shift. Germany also displays a strong twentieth-century technological expansion, consistent with the development of science-based industries.

Japan’s rapid catch-up beginning in the 1960s is also visible, confirming its emergence as a leading innovative economy. More recently, China’s innovative presence has become increasingly evident internationally. A notable feature of China’s trajectory is its connection to manufacturing expansion – consistent with the idea that becoming a major industrial hub can support capability accumulation, learning-by-doing, and diffusion of production knowhow.

We focus in more detail on innovation trajectories across 40 economies since 1960. We control statistically for permanent differences in patenting across countries, institutional changes in patent offices, home bias, population trends and other confounding influences. Figure 1 shows the results, where for example, the “DEU” bar indicates that Germany had about 200 more patents per year than the UK (the omitted baseline).  Other nations in this innovation ‘premier league’ including the US, Japan and South Korea. At the other end, a group of roughly a dozen countries, ranging from Norway to Bulgaria, followed markedly slower trajectories. Between these extremes lies a large middle group of countries such as Spain, Australia and Denmark whose technological performance broadly tracked those of Britain.

Figure 1 Technological trajectories in the post-1960 period across multiple countries

Note: These are the 1960-2016 technological trajectories of 40 countries (all relative to the UK). The vertical axis indicates the technological trajectory coefficients estimated as additional number of annual patents by the indicated country across multiple patent offices. The height of the bar is the point estimate, and the vertical bars are confidence intervals. For more details see Bergeaud et al. (2026).

Technological trajectories and economic performance

We matched our estimated technological trajectories with other economic data and documented that countries with stronger trajectories had significantly faster rates of TFP growth. Nonetheless, there remained plenty of differences between trajectories and different productivity measures, which is unsurprising as TFP reflects catch-up growth more than innovation at the technology frontier.

We then asked what were the initial policy choices that drive our estimated technological trajectories? We documented that countries who started with (i) higher R&D intensities, (ii) a more educated workforce, and/or (iii) greater defence investments subsequently enjoyed faster technological trajectories over the next half-century. The importance of intellectual and human capital is consistent with modern growth theory (e.g. Aghion and Howitt 1992) and the importance of defence chimes with the many civilian spinoffs from military R&D (e.g. Moretti et al. 2025, Howell et al. 2025).

Conclusions

We detail a new way of measuring the trends in long-run technological capability of a country leveraging filings in multiple patent offices and compiled two centuries of patent data to document the rise and decline of technological leadership. These trajectories are not random manna from heaven, but rooted in social and political choices, and can be fostered through long-run investments in R&D, education, and national security.

References

Aghion, P and P Howitt (1992), “A model of growth through creative destruction”, Econometrica 60(2): 323–351.

Bergeaud, A, R N Gozen, and J Van Reenen (2026), “Mapping Technological Trajectories: Evidence from Two Centuries of Patent Data”, CEPR Discussion Paper 21066.

Bergeaud, A, G Cette and R Lecat (2016), “Productivity trends in advanced countries between 1890 and 2012”, Review of Income and Wealth 62(3): 420-444.

Bergeaud, A and C Verluise (2024), “A new dataset to study a century of innovation in Europe and in the US”, Research Policy 53(1): 104903.

De Rassenfosse, G, H Dernis, D Guellec, L Picci and B V P De La Potterie (2013), “The worldwide count of priority patents: A new indicator of inventive activity”, Research Policy 42(3): 720-737.

Howell, S, J Rathje, J Van Reenen and J Wong (2025), “Opening up Military Innovation: Causal Effects of Reforms to US Defense Research”, Journal of Political Economy 113(11): 3605–3651

Moretti, E, C Steinwender and J Van Reenen (2025), “The Intellectual Spoils of War: Defence R&D, Productivity and Spillovers”, Review of Economics and Statistics 107(1): 14–27.