The market for rare earth elements is relatively small in value, but shortages can lead to disproportionate output losses. This column introduces a network model to quantify the output losses from a large and permanent shortage of rare earth elements. At short horizons less than one year, GDP losses range from 0.6% in the UK to 1.5% in the US and 1.8% in Japan. Supply chain risk assessments need to incorporate production networks to understand the vulnerabilities from such disruptions. Policymakers should build strategic reserves and develop domestic capacity to help withstand the short-run losses from rare earth shortages.
Recent events have reminded policymakers of the fragility of global industrial supply chains. A short-lived disruption to global rare earth supplies in April 2025 was followed by the disruption to the Strait of Hormuz, ongoing since March 2026, and, more recently, mounting concerns that Japan could run short of yttrium, a rare earth used in jet engines, as reported by the Financial Times (Dempsey et al. 2026). The market for rare earths is relatively small in value, but shortages can lead to disproportionate output losses: even the finest machinery can be brought to a halt by a single grain of sand, or lack thereof. In a recent paper (Bowman’s et al. 20265a), we introduce a network model to quantify output losses from a large and permanent shortage of rare earth elements (REEs) and REE magnets – a crucial input to strategic industries – in the US, the UK, Japan, Germany, India, and France. For horizons shorter than one year, when rare earth elements cannot be easily substituted, GDP losses can be substantial, ranging from 0.6% for the UK to 1.5% in the US and 1.8% in Japan. This amplification reflects not only the critical role of rare earth elements in strategic industries, but also the central position of these industries in broader production networks, causing the initial disruption to propagate through supply chains.
Why rare earths punch above their weight
Rare earth elements are 17 chemically similar metals with exceptional magnetic, catalytic, and luminescent properties that make them essential inputs to numerous goods including electric vehicles, wind turbines, magnetic resonance imaging (MRI) scanners, robotics, and defence systems. Yet their market value seems modest compared to oil, gas, or semiconductors: the global rare earth market reached roughly $6 billion in 2024, with permanent magnets, the largest application, valued at $25 billion. This understates their economic importance, however. Rare earth elements are used in tiny quantities across numerous and interrelated manufacturing sectors, unlocking trillions in downstream output. While alternatives exist for some applications, they typically entail significant penalties in performance, size, weight, efficiency, or cost.
Rare earth processing is the most important bottleneck
Technical expertise acquired over decades, established infrastructure, and farsighted policy support made China the world’s largest producer of rare earth elements during the 2000s. China’s market share varies by rare earth type and supply chain stage. For light rare earth elements (LREEs), the more widely produced group of rare earth elements, diversification in mining has reduced China’s share of global output from 97% at its peak in 2010 to 58% in 2024 (US Geological Survey 2026), but China maintains 88% of oxide separation capacity (Figure 1) and 93% of metal refining (Bedford 2025). For heavy rare earth elements (HREEs), by contrast, China retains a near-monopoly, with market shares of 90–98% across supply chain segments (Bedford 2025).
Figure 1 Estimated rest-of-world (RoW) import dependence on China by supply chain segment (in percentage)


Note: Import dependence is calculated for the rest of the world assuming China accounts for 60% of global consumption at each supply chain stage. A zero value indicates that aggregate RoW production is sufficient to meet RoW consumption, not that China has no role in global production. HREE = heavy rare earth element; LREE = light rare earth element; ROW = rest of world.
Source: Bedford (2025) and IMF staff calculations.
This concentration creates potential chokepoints, as geographically concentrated sourcing can amplify the output effects of input disruptions (Schwellnus et al. 2023). Separation and refining are the most critical stages, as nearly all rare earth concentrates, regardless of origin, flow through Chinese processing facilities.
The potential for disruption became evident on 4 April 2025, when China imposed export licensing requirements on seven rare earth metals, predominantly heavy rare earth elements, and related products, including permanent magnets. Introduced shortly after the US raised tariffs on most trading partners, the measures were justified on national security grounds.
The requirements caused a sharp drop in permanent magnet exports between April and June. In May, Chinese exports of permanent magnets stood roughly 70% lower year-on-year, before quickly resuming double-digit growth (Figure 2). Although short-lived, the disruption prompted import-dependent countries to ask a more pointed question: What would be the macroeconomic consequences of a major and sustained shortage of rare earth elements?
Figure 2 China’s exports of rare earth magnets and rare earth magnet components intended for permanent magnet production (in metric kilotons)


Note: Data are from the series labelled “CN: Export: HS8: Permanent Magnets or Articles Going to Be Permanent Magnets, of Rare-Earth Metals.” Code HS 850511.
Source: China Customs Statistics: General Administration of Customs of the People’s Republic of China (GACC), and IMF staff calculations.
Going beyond ‘value added at risk’
Industry analyses of rare earth elements’ vulnerability typically rely on summing up the value added of sectors that depend on these elements. This ‘value added at risk’ (VAAR) measure ignores medium-run substitution and, importantly, it ignores the network structure of modern economies, potentially missing indirect foreign exposure embedded in upstream input chains (Baldwin et al. 2023).
Consider the motor vehicle sector, a leading buyer of rare earth element permanent magnets and a central node in the production network, buying from and supplying many other industries. When a rare earth element shock curtails automotive output, it propagates upstream and downstream through the production network, including by constraining the supply of vehicles to transportation and logistics services, which are themselves inputs into many other industries, amplifying the aggregate loss. Electrical equipment and computer electronics similarly are important buyers and suppliers of intermediate goods.
Figure 3 Value added and gross output at risk


Note: Rare earth element-dependent sectors are identified at the OECD industry level. VAAR is their value added as a share of GDP; GOAR is their gross output as a share of economy-wide gross output.
Source: OECD Input–Output Tables; Nassar et al. (2025); and IMF staff calculations.
Value-added-at-risk analysis (Figure 3) suggests that Germany’s economy is the most exposed of the countries in our study, with 2.5% of GDP generated by sectors that rely directly on rare earth elements at some stage of production, largely reflecting the country’s sizeable automotive sector. The US, by contrast, shows a value added at risk of only 0.8%. Yet, as our results below show, the GDP impact tells a very different story once network effects and associated amplification mechanisms are incorporated.
A network model calibrated to country-specific production structures
To move beyond value added at risk, we construct a small open economy model with production network linkages, calibrated using a rare-earth-element-augmented input–output table built from US Geological Survey data and standard international input–output databases. We then analyse how an 80% reduction in available rare earth element imports, consistent with average import dependence on China across the economies studied, propagates through the economy under different substitution assumptions (see Bogmans et al. 2026a for more details).
Substitutability is key. Over horizons of less than one year, firms cannot readily replace rare earth inputs in production processes engineered around them. This means that even small shortfalls can cause large output losses (see also Panon et al. 2025). Over longer horizons, firms can redesign products, recycle rare earth elements, and build alternative supply chains. Our model captures both scenarios, with a substitution elasticity of 0.015 (near-Leontief) in the short run and 0.8 in the longer run.
Network amplification makes GDP losses surprisingly large and unequally distributed
Under limited substitutability, our model generates sizeable output losses. Japan, the US, and Germany are most exposed, with GDP declining by 1.8%, 1.5%, and 1.3% respectively in response to the 80% supply shock (Figure 4). France and India see losses of around 0.9% and 0.8%, while the UK experiences the smallest decline at 0.6%.
Strikingly, these results diverge sharply from value-added-at-risk analysis. Germany has the largest value-added-at-risk exposure in our sample, yet the US and Japan suffer larger GDP losses. Although Germany’s automotive sector is large relative to GDP, it is comparatively labour- and capital-intensive, which reduces its reliance on intermediate inputs. Moreover, the US and Japanese motor vehicle sectors have stronger production network linkages. These network effects more than offset Germany’s larger direct exposure.
Figure 4 Output losses by country from a major rare earth element disruption


Note: Scenarios assume an 80% disruption to imported rare earth element-related goods. The high-elasticity case uses a substitution elasticity of 0.8, following Alfaro et al. (2025); the low-elasticity case uses 0.015. REEs = rare earth elements.
Source: OECD tables; and IMF staff calculations.
Sectoral effects are highly concentrated. In the US, 19 sectors contract, with automotive output falling by around 9% and other transport equipment by 3.7%. German automotive output declines by only 5.8%, reflecting its greater reliance on domestic labour and capital. Almost all the differences between the short- and long-run scenarios come from firms’ ability to substitute domestic intermediates for rare earth elements. Macroeconomic damage thus hinges on a small number of upstream sectors that depend on tiny quantities of a highly specific, hard-to-substitute imported input.
Policy implications
Our findings carry two lessons for policymakers.
First, supply chain risk assessments should incorporate production networks. Value-added-at-risk analysis can understate vulnerability in highly networked economies such as the US and Japan, while overstating it in economies with large but less-connected rare earth element-intensive sectors.
Second, the critical policy horizon is the short run. Beyond five years, substitution, diversification, and technological adaptation reduce GDP losses close to zero in our model. The challenge is to bridge the period between a supply disruption and meaningful adjustment. Building strategic reserves, developing domestic capacity, and supply chain diversification can help reduce the macroeconomic costs and provide a roadmap to withstand potential future shortages (see also Bogmans et al. 2026b).
Source : VOXeu








































































