Imitation is a first-order concern for enterprises operating global production processes. Firms routinely source knowledge-intensive inputs from many countries, including those with weak intellectual property laws. This column argues that firms pursue a particular strategy to protect their knowhow in such an environment: slicing production across multiple suppliers to ensure that no single supplier learns enough to imitate. In global micro data on the auto industry, the concentration of sourcing across suppliers is U-shaped in the strength of intellectual property rights – a pattern driven by the protection of high-tech components.
The battle for technological leadership among the world’s major economic powers is ongoing. The US-China trade conflict, starting in 2018, was in part motivated by allegations of unfair trade practices, forced technology transfer, and intellectual property theft. And the underlying concern was real: China has long forced Western multinational corporations to transfer their technology in exchange for market access (Holmes et al. 2015), and joint ventures did diffuse technology to Chinese partner firms and beyond (Jiang et al. 2018). By 2025, China had evolved into the world’s leading producer of electric-vehicle batteries and placed eight key battery technologies under export licensing, seeking to keep advanced knowhow in China. Also, in Europe policymakers are debating whether Chinese battery plants should be welcomed only under the condition of sharing technological knowhow with local partners (Tagliapietra et al. 2025).
These policy battles rest on a simple premise: whoever produces components for a product learns about the underlying technology. For firms, this is a daily concern rather than a grand geopolitical game. Firms are demonstrably wary of exposing their technology in global value chains: US firms holding patents are significantly more likely to export to countries with stronger intellectual property rights (IPR) protection, especially when selling to unrelated parties (Lin and Lincoln 2018). Yet avoiding countries with weak IPR protection is rarely an option. Cost advantages, market access, and local-content requirements compel firms to produce in emerging economies despite imperfect legal protection (Antràs and Chor 2022). But when institutions are incomplete, firms require alternative protection strategies.
In recent research (Eppinger et al. 2026), we propose that firms ‘slice to protect’ their knowhow: they divide the production of their inputs across many suppliers, sharing with each only the subset of information that is strictly necessary. The underlying idea is that technical knowhow behind a production process can be partitioned into separate, non-overlapping bodies of knowledge (Baldwin and Henkel 2015). By slicing production into more and smaller puzzle pieces, a firm ensures that no individual supplier – nor a rival the supplier might inform – can put together the whole puzzle. Apple, as a case in point, is known for assigning narrowly defined tasks to its suppliers, which receive minimal information and operate under tight security provisions (Pomfret and Soh 2010).
Slicing up production does not come for free, however. Dealing with more suppliers requires more searching, contracting, and coordination. Firms therefore face a trade-off between protecting their knowledge and saving on fragmentation costs – and they should slice most where the risk of imitation is greatest.
We examine this strategy in the automotive industry, using global micro data from IHS Markit at the highly granular level of car producers, their suppliers, and individual car components. The data record which of more than 100 distinct car components each car manufacturer – or ‘original equipment manufacturer’ (OEM) – sources from each of its first-tier suppliers around the world. Our sample comprises 401 original equipment manufacturers buying 104 components from 3,641 suppliers in 59 countries. The car industry is a natural testing ground: its value chains are deeply fragmented – the average original equipment manufacturer sources from 91 suppliers in 12 countries – and many components embody sensitive technology. China hosts more than 1,000 suppliers in the data, which lends the question particular relevance.
We measure concentration of production – the inverse of slicing – by the number of distinct components an original equipment manufacturer sources from the same supplier. We relate this measure to the quality of intellectual property rights protection in the supplier’s country, as perceived by business executives surveyed for the World Economic Forum’s Global Competitiveness Report (WEF 2019). Importantly, we compare sourcing decisions within the same original equipment manufacturer, and we control for characteristics of the supplier’s corporate group as well as a host of country-level, bilateral, and firm-level factors.
The data reveal a robust U-shaped relationship between concentration and intellectual property rights protection (Figure 1). The number of components per supplier is relatively high where intellectual property protection is weakest, reaches its minimum at intermediate levels, and is highest where protection is strongest. The differences are moderate in size but precisely estimated: compared to sourcing in South Korea, the median country by intellectual property protection, original equipment manufacturers source about 0.2 more components per supplier in the UK, at the 75th percentile. This U-shaped pattern survives an extensive battery of robustness checks.
Figure 1 Components per supplier and the strength of intellectual property rights protection
A U-shape may seem puzzling at first sight: if slicing protects knowledge, should firms not slice the most where intellectual property protection is weakest? The key to the puzzle is that intellectual property rights institutions play a dual role. On the one hand, better enforcement directly reduces the risk that shared know-how is exploited – a protective effect. Taken by itself, this effect lets firms concentrate purchases on fewer suppliers as institutions improve, which explains the upward-sloping part of the curve. On the other hand, intellectual property institutions shape which components are traded in the first place – a compositional effect. Low-tech components embody little proprietary knowledge, so their production can be entrusted to any supplier. High-tech components are different: they require sharing sensitive knowledge, and firms source them relatively more from countries with better intellectual property protection, in line with a large body of evidence on IP and technology transfer (Branstetter et al. 2006, Ivus 2010).
Together, the two effects can explain the U-shape. Where intellectual property protection is very weak, firms procure only low-tech components, imitation is of little concern, and they concentrate purchases to save fragmentation costs. As protection improves, firms start sourcing knowledge-intensive components while enforcement remains imperfect – so that is where they slice to protect. Beyond a certain threshold, the protective effect dominates and firms increasingly concentrate production again. Slicing thus peaks at intermediate levels of protection, precisely where valuable knowledge is exposed but the legal shield remains weak.
This explanation has a testable implication: the relationship between sourcing concentration and intellectual property protection should be present for knowledge-intensive components but absent for simple ones. We classify car components as high-tech or low-tech, building on an engineering cost index originally constructed for the auto industry by Monteverde and Teece (1982). We update this index to obtain a simple binary classification of high-tech components, such as the instrument panel or the AC compressor, as opposed to low-tech components such as the floor carpet or the headrest. The data confirm our prediction (Figure 2): the number of high-tech components per supplier rises strongly and significantly with IP protection, while there is no significant relationship for low-tech components.
Figure 2 High-tech versus low-tech components and intellectual property protection
The same contrast emerges from our most demanding test, which compares high-tech and low-tech components within the same supplier country, within the same original equipment manufacturer, and even within the same supplier firm. The differential relationship for high-tech components is positive and statistically significant throughout. Where intellectual property institutions are weak, firms slice up the production of their technology-intensive components to protect their know-how.
Three decades after the signing of the Trade-Related Aspects of Intellectual Property Rights (TRIPS) agreement, whether and how stronger intellectual property protection benefits developing countries remains a controversial question. Our analysis suggests that the effects are unlikely to be linear. Strengthening intellectual property institutions from a very low level may change what suppliers produce: it encourages specialisation in inputs with higher technology content. Only once protection is sufficiently strong do foreign firms entrust more knowledge to individual suppliers. This concentration of knowledge can eventually help suppliers upgrade and may facilitate their transition into final goods.
The findings also speak to today’s contests over technology. Governments can mandate joint ventures, restrict exports of sensitive technologies, or attach technology-transfer conditions to inward investment. But how much knowledge actually flows through a value chain is, to an important degree, a firm-level choice. By deciding how thinly to slice production, firms influence how much knowledge any individual supplier can absorb. Policymakers seeking to attract knowledge-intensive production, or to obtain technology from foreign investors, should expect firms to respond along this quiet margin. Whether the slice-to-protect strategy shapes other industries, and how firms react when intellectual property protection changes over time, remain open questions for future research.
Source : VOXeu
The labour force participation rate in the euro area has risen in recent years while…
Economic historians have long placed culture at the centre of the rise of the West,…
Many public institutions are now experimenting with AI tools to support surveillance, supervision, and policy…
Sharp increases in interest rates can reduce the market value of banks’ securities holdings and…
When the US imposed sweeping tariff increases in 2025, most economic forecasters predicted a sharp…
The European Union's AI Act – the world’s most ambitious regulatory framework for artificial intelligence…