Innovation is widely viewed as the engine of economic growth. But not all innovations play the same role. Economists often distinguish between product innovations, which introduce new goods, and process innovations, which change how goods are produced (Cohen and Klepper 1996). While growth models emphasise innovation’s role in expanding the technological frontier (Romer 1990, Grossman and Helpman 1991, Aghion and Howitt 1992; see also the recent discussion on VoxEU by Klenow 2025), they rarely assign these different types distinct roles, and empirical research has only recently begun to distinguish between them (Bena and Simintzi 2025). As a result, we still know surprisingly little about how different types of innovation shape firm growth and technological progress, and what this means for innovation policy.

When we think about process innovation, cost-cutting often comes to mind – new assembly lines that reduce production time, supply chain optimisations that lower inventory costs, or automation that replaces expensive labour. These cost-reducing process innovations are vital for competitiveness. But our new research (Tham et al. 2025) reveals a different type of process innovation that has far more profound implications for sustained growth: foundational process innovations.

Unlike cost-reducing innovations that make existing production more efficient, foundational processes are game-changers that expand what firms can produce. Consider photolithography in semiconductor manufacturing, which enabled the miniaturisation of circuitry and catalysed an explosion of new electronic products: smartphones, wearables, and advanced computing devices that would have been impossible before. Or the polymerase chain reaction (PCR) method that revolutionised biotechnology by enabling rapid DNA replication, paving the way for countless new genetic tests, vaccines, and treatments.

These examples illustrate a crucial distinction: whilst cost-reducing innovations help firms do existing things better, foundational innovations help firms do entirely new things. This difference matters enormously for understanding firm growth and formulating innovation policy.

Measuring the unmeasured

The challenge in studying these different types of process innovation has been measurement. Unlike product innovations, which are relatively straightforward to identify, process innovations are notoriously difficult to classify and quantify over long time periods. We addressed this by developing a novel methodology that analyses the text of over 50 million patents from 1900 to 2020.

Our approach classifies patents into process and product innovations based on patent titles and claims. We use hypernym-based linguistic analysis, essentially teaching an algorithm to recognise whether a patent describes an ‘activity’ (like a method or process) or a ‘physical entity’ (like a device or composition). We validated this against expert classifications by patent examiners and intellectual property specialists, achieving over 93% agreement.

The more subtle distinction – identifying which process patents are foundational versus cost-reducing – required additional innovation. We calculated the textual similarity between each process patent and a firm’s past and future product patents. Foundational processes are those that look different from the firm’s existing products but highly similar to its future products. In other words, they open new technological possibilities. Cost-reducing processes, by contrast, align closely with existing products – they’re improvements to what the firm already makes.

What the data reveal

Our patent classifications reveal striking patterns. Process patents account for between 26% and 50% of innovation across industries, far too large a share to ignore in growth analysis (Figure 1, Panel A). Within process innovations, roughly 31% are foundational, whilst 69% are cost-reducing. The prevalence of foundational innovations has increased markedly since 1990, coinciding with transformative advances in digital technology, biotechnology, and microelectronics (Figure 1, Panel B).

Foundational processes have distinctive characteristics. They cite substantially more scientific publications (16-21% more non-patent literature) than other patents, indicating they draw heavily on basic research. This connection to fundamental science gives them deeper technological roots. We also find that firms with more foundational patents publish more academic papers themselves, creating a virtuous cycle between basic and applied research.

The market values foundational innovations accordingly: foundational process patents have 30% higher private economic value (Kogan et al. 2017) than cost-reducing patents, measured by stock market reactions to patent grants.

Figure 1

a) Historical evolution of heterogeneous innovations

b) Foundational process patents across industries 

Notes: Figure (a): the number of process and product patents (left axis) and the process-to-product ratio (right axis) for all USPTO granted patents from 1900 to 2020 (three-year moving averages). Figure (b): The figure presents the time-series variation of the number of U.S. foundational patents granted to publicly listed firms by Fama-French 12 industries for the sample period from 1930 to 2020.
Source: Tham et al. (2026).

From process to product: The growth mechanism

The most important finding concerns how these innovations affect firm growth differently. Using local projection methods to trace effects over one- to ten-year horizons, we find that all types of innovation boost firm profits, sales, capital, employment, and productivity. But the patterns diverge, as illustrated in Panel A of Figure 1. Cost-reducing innovations show immediate but short-lived effects, profits and sales increase in the first three years, then plateau. This makes sense: efficiency gains lower costs and boost margins quickly, but these are one-time improvements.

Foundational processes show sustained growth effects lasting seven years or more. The mechanism is indirect but powerful: foundational processes enable firms to introduce higher-quality products that better match customer needs. Products building on foundational processes have 34% more forward citations and 22.5% higher market value than other products. These products, in turn, drive the sustained growth in sales and profitability, as shown in Panel B of Figure 2 for profit growth.

We demonstrate this foundational process-based product channel using FDA drug approval data. Pharmaceutical products linked to foundational patents command higher market valuations at approval, are more likely to receive priority review status (indicating significant therapeutic advance), and generate substantially higher patient spending through Medicare and Medicaid. The quality premium is real and translates to commercial success.

Foundational processes also expand firms’ technological scope. Firms with more foundational patents innovate across more technology classes and are significantly more likely to enter entirely new technological domains. A 10% increase in foundational patents increases the breadth of technology classes by 4.3%. Cost-reducing innovations, by contrast, are negatively associated with technological diversification, efficiency gains in existing processes don’t encourage exploration of new product categories.

Figure 2

a) Heterogenous innovations and firm growth (profits)

b) Profit growth from product innovations: The role of foundational processes

Notes: Panel (a) shows the estimated effects of product, cost-reducing process, and foundational process innovations on firm profit growth. Panel (b) shows the effects of product innovations that build on foundational versus cost-reducing processes. Regressions control for firm characteristics, industry and year fixed effects. For more details, see Tham et al. (2025).

Aggregate implications

These firm-level patterns have economy-wide implications. Using our international patent classifications across 60 countries, we find a strong positive correlation between the ratio of foundational to non-foundational innovation and subsequent GDP and total factor productivity growth. Countries that generate more foundational process innovations experience higher growth rates in the following four years.

This relationship helps explain productivity dynamics. Foundational innovations create technological possibilities that propagate through the economy as firms build complementary products and services. The photolithography example is instructive: the initial process innovation enabled not just better semiconductors but entire new industries, smartphones, IoT devices, artificial intelligence hardware. Each wave of foundational innovation triggers cascades of product innovation.

Policy implications

Our findings carry important implications for innovation policy. First, the emphasis on R&D intensity may be too narrow. What matters is not just how much firms invest in innovation, but what type of innovation they pursue. Policies should distinguish between innovations that improve existing products and processes versus those that expand technological frontiers.

Second, the tight link between foundational innovations and basic science argues for sustained public investment in fundamental research. Foundational corporate innovations build on scientific foundations. When governments cut research funding or when firms focus narrowly on near-term commercialisation, they may undermine the scientific base that future foundational innovations require.

Third, the results speak to debates about manufacturing and industrial policy. Because foundational process innovations are so intimately connected to manufacturing know-how, policies affecting the location of production have innovation consequences. Offshoring advanced manufacturing doesn’t just shift production, it may shift the locus of foundational innovation to other countries, with long-term growth implications. This argues for policies that maintain domestic capabilities in advanced manufacturing sectors.

Fourth, competition policy should consider innovation type. Our framework suggests that firms with advantages in foundational innovation can sustain these advantages through the superior products they enable. This creates natural barriers to entry that competition authorities should account for when evaluating market power and merger activity in technology-intensive sectors.

The distinction between foundational and cost-reducing process innovations matters. Both contribute to growth, but through different channels and over different time horizons. Cost-reducing innovations deliver quick wins through efficiency gains. Foundational innovations deliver sustained growth by expanding what firms can produce and enabling waves of complementary innovation. Understanding this distinction is essential for designing effective innovation and industrial policies. The firms and countries that master foundational process innovation will shape the technological frontier and reap the growth rewards that follow.

Authors’ note: The views expressed here are those of the authors and do not necessarily reflect the views of the Federal Reserve Bank of Atlanta, the Federal Reserve System, or its staff.

References

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

Bena, J and E Simintzi (2025), “Machines could not compete with Chinese labor: Evidence from U.S. firms’ innovation”, Review of Finance.

Cohen, W M and S Klepper (1996), “Firm size and the nature of innovation within industries: The case of process and product R&D”, Review of Economics and Statistics 78(2): 232–243.

Grossman, G M and E Helpman (1991), “Quality ladders in the theory of growth”, Review of Economic Studies 58(1): 43–61.

Klenow, P (2025), “Sustained growth through creative destruction: Nobel laureates Philippe Aghion and Simon Johnson”, VoxEU.org, 22 October.

Kogan, L, D Papanikolaou, A Seru and N Stoffman (2017), “Technological Innovation, Resource Allocation, and Growth”, The Quarterly Journal of Economics 132(2): 665–712.

Romer, P M (1990), “Endogenous technological change”, Journal of Political Economy 98(5, Part 2): S71–S102.

Tham, W W. S Baslandze, E Sojli, and L Liu (2025), “Foundational Processes and Growth”, CEPR Discussion Paper No. 19858.