Semiconductors are the world's fourth-most-traded product after crude oil, refined oil, and cars. The trade value of the finished product (taking into account the whole value chain needed to produce them) reached $1.7 trillion in 2019, four times the value of their global semiconductor sales. Over 60% of the world’s countries are involved, either as exporters or importers, in this highly specialized value chain, given the complexity of semiconductors' design and manufacture. No other industry has the same intensity in R&D: about $90 billion, equal to 22% of annual sales, compared to the 21% of the pharmaceutical sector.
As a result, no region can have the entire semiconductors' production on its territory and be "self-sufficient". As such, deep interdependencies characterize the resulting value chain: the United States, Taiwan, South Korea, Japan, Europe, and China specialize in different process steps and rely on specific inputs from suppliers. Thanks to the World Trade Organization's Information Technology Agreement (in operation since 1997), all semiconductor-related products, materials and tools are subjected to one of the lowest tariffs in global trade. Therefore, they are free to move around the world to reach the optimal location for performing each phase such that self-sufficient local supply chains would result in a 35% to 65% overall increase in semiconductor prices.
While this geographic specialization has created economic benefits, it also brings about vulnerabilities. There are over 50 technology or process steps along the value chain wherein one region holds over 65% of the global market share. Consequently, if one of these "choke points" was to be affected by a natural disaster, an international conflict, or export control measures, the whole supply of chips could be fundamentally disrupted. Where are these choke points, and how have they been used? What are the economic consequences of the current semiconductors' scarcity, and what future scenarios can we expect?
The main choke points
The production process for semiconductors comprises three main steps: design, fabrication, assembly, and test, each characterized by regional monopolies.
Developing a modern chip requires advanced design tools, called electronic design automation (EDA) software, on which leading-edge chip design depends. 85% share in EDA tools is concentrated among three US-based companies: Cadence Design Systems, Synopsys, and Mentor. Given this context, the US was able to curb Huawei's chip design capabilities by imposing export control measures to block the company's access to US-origin EDA software.
The finished chip design is then sent to a wafer fabrication plant (fab) for production. There are 400 to 1,400 steps in the overall manufacturing process, requiring between 12 to 20 weeks and 300 different inputs, mainly chemical ones. Japan's Shin-Etsu and Sumco are the major suppliers of these chemicals, playing a critical role in the semiconductors’ value chain. The Japanese government exploited that dependency by introducing export restrictions on such products to Korea in July 2019.
These necessary inputs are processed by over 50 types of highly engineered precision equipment. For 5 of these 50, US firms account for more than 50% of the global market share. While the EU's ASML (based in the Netherlands) has a 100% global market share in the Extreme Ultra-Violet (EUV) scanners, which are essential to manufacturing chips below 7 nanometers.
The race for smaller transistors and more powerful chips has increasingly raised the prices of these manufacturing tools, which account for 80% of most modern fabrication plants' costs. As a result, a semiconductor fab requires roughly $5 to $20 billion of capital expenditure. Unsurprisingly, today's foundry market is intensely concentrated, with about 75% of semiconductor manufacturing capacity in China and East Asia and five companies accounting for 53% of production. Furthermore, the world's entire semiconductor manufacturing capacity below 10 nanometers is currently located in two countries alone: South Korea (8%) and Taiwan (92%).
Finally, given its lower capital intensity, assembly and test firms are mainly in East and Southeast Asian countries, where wages for skilled labour are up to 80% below US levels. As such, 9 of the 10 largest firms by revenue for these processes are headquartered in China, Taiwan and Singapore.
Explanations and consequences of the global semiconductors’ scarcity
Semiconductors are getting increasingly smaller in size: such technological developments have paralleled the growing relevance of the global manufacturing industry. A major portion of the chips’ global market (around 3/5) consists of laptops and smartphones, while cars and other vehicles make up for 10% of the global demand for such components. Although this does not seem like a crucial part of the demand side, one should consider that semiconductors have quickly become key components for today’s (and in particular for tomorrow’s) cars when it comes to a vehicle’s electrification, safety and driver assistance, and connectivity items. The need for semiconductors is even higher for electric cars: it transpires that, in view of the energy transition that will increasingly skew the market towards the production and sale of electric vehicles, car manufacturers’ appetite will get bigger and bigger.
As a consequence of the lockdowns imposed by the Covid-19 pandemic, global supply of semiconductors has become progressively scarce. This can be explained by a variety of factors: the rising demand for electronic items (particularly personal computers for homeworking), the fast recovery of the car manufacturing sector that determined a peak in the demand for components, the power outages in key production facilities in Texas which disrupted whole segments of the chips supply chain last winter, as well as other (temporary) bottlenecks such as the blockage of Suez Canal after the Ever Given container ship got stuck at the end of March. The combined effect of a surge in the demand for semiconductors and physical constraints on the supply side might have a considerable impact on the production of automobiles, with the US and Western Europe mostly affected as they might see production shrink it from 2% to 6%.
This contraction in the car manufacturing industry would produce effects at the macro level: for instance, Goldman Sachs estimated US GDP might suffer from a 0.5% reduction as a consequence of the shortage in “semis”. Given the interdependence and complexity of the global chips supply chain, the negative impact on growth could have cross-country effects and thus hinder global recovery. However, the countries most badly affected could be those ones that rely on semiconductors for the lower ends of their supply chains, characterized by higher value added.
In parallel with constraining economic activity and growth, scarcity in semiconductors could also increase general price inflation, which is already on the rise (particularly in the US and China) as a result of the ongoing economic recovery. Although it is unlikely to foresee a sharp rise in price indexes in the near future (given cautious monetary policies and appropriate surveillance put in place by key central banks), mismatches between demand and supply might contribute to creating economic instability during the ongoing recovery phase.
What’s next: how to address current problems and the future global “race for chips”
The strategic importance of semiconductors, as shown by their wide range of industrial applications, their high technology intensity, and their highly concentrated market power, is likely to produce a geopolitical race for global leadership in the production of such microscopic components. The European Union has already made clear it intends to establish a “European” chips industry by 2030 within the framework of the so-called “digital compass”. This will entail adopting a multi-layered strategy that encompasses both the support and development of technologies as well as an increase in production capacity. The United States is going to pursue similar objectives by striking deals at the international level with key industrial partners such as South Korea with a view to containing China.
However, given the complexity of this supply chain (which starts from the collection of raw materials and requires many intermediate stages), fragmentation doesn’t seem like a feasible answer. Over the next decade, the global chip industry will need an additional $3tn investment in R&D alone to meet the increasing demand for semiconductors. In order to supply the appropriate level of financial resources, technology, and know-how, international cooperation will remain key, to some extent at the very least. We might see new alliances and partnerships built around the global quest for something that is becoming almost invisible but also extremely tangible in our daily lives.