Share

Let the chips fall where they may

Let the chips fall where they may: a story of subsidies and semiconductors

4 December 2019 | Jehan Sauvage

In December 1983, Tom Wolfe, the late American novelist, published in Esquire a portrait of Robert Noyce, the man who co-founded Intel in 1968, together with Gordon Moore of ‘Moore’s Law’ fame. What is so striking about Wolfe’s piece is that it depicts the invention of something as central to our economies as computer chips as largely a story of talent and luck.

Governments had surprisingly little to do with how, in 1947, three physicists working for AT&T’s Bell Laboratories invented the first semiconductor out of two gold wires and a piece of germanium. Governments also had little to do with the fact that one of those physicists, John Bardeen, was a former classmate of Grant Gale, a physics professor at a local college in Grinnel, Iowa. Hearing about Bardeen’s invention, Gale contacted him to obtain a semiconductor prototype and establish in 1948 the world’s first academic course in solid-state electronics (in Grinnel, of all places). Governments had little to do with the fact that one of Gale’s 18 physics majors was the young, Iowa-born Robert Noyce. And, finally, governments had little to do with the decision in 1955 by William Shockley, another of the three original inventors of the semiconductor at Bell Laboratories, to leave New Jersey and establish his own business, Shockley Semiconductor Laboratory, in his hometown of Palo Alto, California. Funding for this company, which would later hire Robert Noyce and Gordon Moore, came from Arnold Beckman, an entrepreneur and inventor of the first pH meter.

While the story of the early semiconductor industry is one of wits and happenstance, governments did play a role in what followed. By the early 1960s, NASA had become the major customer for integrated circuits, ensuring robust demand. Around the same time, Chinese authorities established Wuxi Factory No. 742 as the state training ground for semiconductor engineers. In 1967, France created the CEA-Leti, a public research centre specialised in microelectronics, followed, in the 1970s and 1980s, by authorities in Japan, Korea, and Chinese Taipei who all supported semiconductor research and development (R&D) through public institutes such as ETRI and ITRI. Likewise, the United States established the Very High Speed Integrated Circuit programme in 1980 and the Sematech research and development consortium in 1987.

Government involvement in semiconductors shows no sign of abating, and, indeed, is taking some new forms. Recent OECD work has identified more than USD 50 billion in government support for 21 large semiconductor firms over the period 2014-18. While a third of this support was for R&D – and arguably supporting innovation – the majority took the form of investment incentives, below-market loans, and below-market equity stemming in part from a flurry of government equity injections.

Such government support comes with costs beyond its immediate – already signficant -- price tag, not least because subsidies tend to beget more subsidies, as countries find themselves under pressure to match each other’s support.  

First is the cost to fair competition in the global economy, as government support escalates to the detriment of those that do not have deep pockets, undermining the competitive and complex supply chains that make smartphones and computers affordable. Concerns about unfair competition are already contributing to growing trade tensions and placing severe strain on the international trading system, which has been shown to have important gaps in its rule-book.

Second, it’s not clear that all this support is effective in promoting innovation in a semiconductor sector characterised by learning-by-doing, short product cycles, and risk taking. Subsidies and other forms of support can encourage complacency at recipient firms, paradoxically acting against technological upgrading. Indeed, those semiconductor firms in the OECD study that received the most support did not have a stellar track record on innovation. This risk is likely to get worse: after decades of downsizing semiconductor components so that a chip the size of a fingernail can contain billions of transistors, miniaturisation may soon be reaching its physical limits. As innovation moves off more predictable paths, governments may be even less well suited to navigating the terrain.

And last, but by no means least, government support comes with real opportunity costs for taxpayers and citizens as scarce public resources are spent on providing support to individual firms, rather than on providing the education and infrastructure from which all – including small as well as large firms – can benefit. Governments do have a critical role in ensuring the enabling environment for innovation and for competitive firms to grow and create jobs – and even in providing support in areas where the market cannot deliver.

Which brings us back to Tom Wolfe’s portrait of Robert Noyce and the role that talent and luck played in the early days of the semiconductor industry. Does this mean that current government efforts to support semiconductors will fail? Maybe, or maybe not, only time will tell. But even if they do succeed – and experience suggests that that’s a big if – it will come at significant cost, for people, for innovation and for the rules-based trading system that has underpinned global prosperity for over 50 years.  

Policy brief

Policy paper

Measuring Distortions in International Markets: The Semiconductor Value Chain

This report builds on the OECD’s longstanding work measuring government support in agriculture, fossil fuels, fisheries, and more recently in the aluminium value chain in order to estimate producer support and related market distortions in the semiconductor value chain. The report also discusses the implications that these findings have for trade rules, and in particular for subsidy disciplines in a context of growing government involvement in semiconductor production and poor transparency of support measures.

Read more...