The state of aid
Image credit: Cover Images
The UK government claims to want to do a lot more post-Brexit, but how well does state aid work internationally?
One of the UK government’s vaunted benefits for a hard Brexit is the greater freedom it will have to boost its use of state aid. It might have to do a lot more beyond the grants it has already provided in the wake of the Covid-19 pandemic.
In an interview for the political online magazine Project Syndicate, Willem Buiter, visiting professor at Columbia University and former Citigroup chief economist, said the end of the Brexit transition will “likely bring a significant and persistent negative aggregate supply shock... the supply shock will come mostly from supply-chain disruptions, a significant decline in foreign direct investment and the end of the UK’s infinitely elastic supply of labour from the rest of the EU”.
A 2016 study by Swati Dhingra and colleagues at the London School of Economics’ Centre for Economic Performance concluded foreign direct investment would suffer a 22 per cent fall over a decade from Brexit. A big problem that comes with a drop in foreign investment, economists have long argued, is that the arrival of these operations tend to help indigenous companies through the establishment of clusters that promote local suppliers. The government may find itself pumping money into other industries that are affected by a drop in inward investment. A rich irony of this recently adopted stance is that the Conservative administration of the 1980s, together with The Netherlands, pushed for the rules that now permeate the EU state-aid structure in order to prevent national governments from doing just that.
Though it took a couple of decades to solidify, the outcome of that historic lobbying is that the EU is unusual in the way it has defined and implemented rules on state aid. Even in the US, individual states have the latitude to provide funding and tax breaks to companies without an overarching set of rules set by the federal government. The outcome of this could be seen with the interstate bidding war over where Amazon would build its second headquarters, offering more than 10,000 jobs. Some 200 cities submitted offers of incentives rising to as much as $9bn at one stage: all for the third largest company by market capitalisation in the US. Amazon finally settled on a more modest half a billion dollars from the state of Virginia. Such bidding wars obscure the fact there is some coordination between states and the federal government. Arizona took advantage of the Trump administration’s willingness to pull more advanced manufacturing into the country for Taiwan Semiconductor Manufacturing Corporation to build a semiconductor fab there.
Though there are reasons to employ state aid during economic shocks, economists who are often heavily in favour of free-market approaches frequently warn against it. A major risk is a tendency for aid to focus on declining industries. A 2002 paper by Richard Baldwin and Frédéric Robert-Nicoud for the US National Bureau of Economic Research argued owners and managers of established companies facing losses tend to lobby far harder for aid than those expecting success in the future. A further issue is that ailing industries are more visible because they have more employees in each location than the many start-ups spread around a region.
Aid focused on individual companies is not necessarily focused on the past. Previous UK governments have tried to favour sunrise industries. Into the first half of the 1980s, though it had an ideological resistance to state aid, the Thatcher administration decided to continue the funding of Inmos, conceived under Callaghan’s Labour government. Overall, the Bristol-based company received approximately £100m as it first established a presence and market success in specialist memories for graphics before turning to processors in anticipation of a shift to parallel processing, albeit this time way too early for the market.
Although Inmos itself faded from the scene, the legacy of that investment did live on. Before Inmos was set up, there was a small electronics-design presence in the area courtesy of a design office for Fairchild Semiconductor. However, it is easy to trace the heritage of many of the start-ups established since the 1990s, a number of which were acquired to form the UK design centres for a variety of larger overseas electronics companies, to Inmos.
The clustering effect from the government investment is mirrored in France, Germany and Benelux, where the kernel is typically a research institute such as CEA-Leti, Imec or one of the Fraunhofer Institutes. The giant Silicon Valley cluster itself was originally the recipient of significant state aid for R&D under a system that remains peculiarly American, and may well become the target again.
In effect, the Atlantic Ocean divides two models for government intervention in private R&D. On the eastern side, there is the system of pre-competitive grants awarded to multi-organisation groups operated by the EU under a series of initiatives such as the current Horizon 2020 programme and its Horizon Europe successor. On the western side, a group of organisations that began with the Advanced Research Projects Agency (ARPA), formed as a response to Soviet Russia’s Sputnik launch, is the key provider of applied R&D funding.
As part of its ambition to break with European procedures, the Johnson administration has made no secret of its plans to transplant the ARPA structure to the UK. Although adviser Dominic Cummings has argued strongly for such an agency in recent years, the international fascination with the US approach to R&D funding is hardly new. Russell Group universities chief executive Tim Bradshaw, who contributed to a white paper on the subject published by Policy Exchange at the start of 2020, penned a brief that recommended emulating the model for the CBI in 2006.
One key to its relative popularity is that ARPA can provide a way for research-oriented start-ups to gain access to the huge US military procurement machine, which spends almost $150bn a year. The deployment of government procurement is one of the features that backers of the idea writing for the Policy Exchange white paper want to see replicated in a UK ARPA.
On mainland Europe, there have been similar calls. Though industrial giant Siemens has been busily buying companies to support its software operation with a focus on ‘digital twin’ technologies, researchers are increasingly concerned that European countries risk falling behind in information technologies. Some, such as Dalia Marin, professor of international economics at the Technical University of Munich’s School of Management, argue that the European Union needs its own version of DARPA, the current defence-focused research agency, in addition to the Horizon programmes. She claims that for countries such as Germany, such an agency would help drive R&D more likely to be useful to start-ups and help meet the Nato 2 per cent of GDP target for defence spending.
An open question is the degree to which a defence focus is required for an ARPA-like institution. When the Trump administration took power a decade after the establishment of the energy-focused ARPA-E, some officials were keen to disband the agency on the basis that it was doing work the private sector would just as readily take on whereas the government itself had no direct need for its output.
A significant factor in favour of the defence-oriented ARPAs lies in the dual-use nature of many technologies. As with the space programme that spawned the original ARPA in the late 1950s, many fans of the agency’s approach point to its role in creating the basis for the modern internet and in launching the constellation of satellites that form the heart of the Global Positioning System. Both are responsible for trillion-dollar markets thanks to an army of companies embracing the technologies. Despite lacking the ready customer of the DoD, ARPA-E has helped some companies get to market. ARPA-E provided initial funding of just under $4m in mid-2011 to start-up Smart Wires to fund the construction of prototypes that lowered the risk for venture capitalists who were expected to step in after successful demonstrations. Since then, the company has raised more than $130m from the private-equity market and the company is now providing its power-flow control technology to a number of grid operators, including the UK’s National Grid Electricity Transmission company.
A number of concepts have gone outside energy, including Agrivida and Ceres, which used biotechnology to improve crop production. However, a report compiled by the National Academy of Sciences found it difficult to assess ARPA-E’s overall success since its creation.
A criticism of the ARPA approach typically used by economists who back free-market policies is that the programme managers have, like politicians or anyone else, limited visibility of what a particular project is worth long-term because there is no market that might put a price on the innovations. However, arguments against the supposed value of ‘market price’ can be made against venture capital investments as they are subject themselves to imperfect information, which leads to bidding wars long before the viability of a start-up’s approach can be demonstrated.
Though the verdict for R&D-focused state aid is more positive than for corporate subsidies aimed at sustaining existing business, the question of what works best in practice remains hard to answer. As with venture capital, getting any outputs relies on providing cash to many potential duds in the hope something good will emerge.
What is DARPA?
What is DARPA? Even people who have studied the organisation for years cannot give you an answer. Defence analyst Richard Van Atta wrote more than a decade ago: “There is not and should not be a singular answer on ‘what is DARPA’ – and if someone tells you that [there is], they don’t understand DARPA.”
DARPA is famous for its big-picture outcomes, such as the ARPAnet, which ultimately morphed into the public internet we now know so well. The reality is that the agency makes numerous relatively small investments that add up to around $3.5bn (£2.7bn) a year, which amounts to less than 3 per cent of federal R&D funding and less than 1 per cent of the total that US public and private institutions spend. The annual budget of UK Research and Innovation (UKRI), which collected together the work of the nation’s old research councils, is close to three times larger than DARPA’s budget at £8bn.
Another element that seems to be attractive to the UK government is a more individual approach to decision-making. Rather than operating formal grant committees, DARPA revolves around project managers who are recruited from industry to work for terms of no longer than five years. These project managers have significant freedom to select focus areas and award grants, though the approach risks a lack of continuity over a long period.
One argument in favour of DARPA-style projects is that they are frequently aimed at off-roadmap ideas that industry investors dislike, reducing the potential for these projects to duplicate effort. At the same time, because numerous technologies only achieve success through economies of scale, off-roadmap projects do not pay back as easily.
The Electronics Resurgence Initiative (ERI), for example, ostensibly has a mission to make the design of integrated circuits much cheaper than it is today, but it is doing so away from mainstream semiconductor design. The underlying reason for that is to provide access to custom high-performance silicon for low-volume defence-sector projects that have been more or less priced out of getting chips made by conventional means. Many have shifted to field programmable gate arrays (FPGAs) over the past two decades, but contractors are increasingly looking to only use the programmable silicon on a portion of the die with much of the rest dedicated to cheaper hardwired circuitry if they can.
The ERI has taken on a geopolitical dimension in the past couple of years amid a wider shifting of emphasis in government attitudes to technological development. As well as reducing design costs, a number of projects focus on ways to build chips that cannot be compromised by Trojan horse circuits inserted in foreign fabs.
Hey, big spender
Although the UK government has claimed the rules imposed by the EU on state aid act as a problematic straitjacket, the country turned out to be the largest single provider of cash to business in a 2013 study of state aid across the union carried out by Miguel Ángel Bolsa Ferruz and Professor Phedon Nicolaides of the College of Europe in Bruges. The reason? The sheer size of the nationalisation of Channel Tunnel rail link developer London and Continental Railways. Across Europe, the Bruges researchers found fewer than a third of cases led to action by the EU though close to two-thirds triggered investigations.
Those numbers pale into insignificance compared to the level of support the Chinese government poured into its solar industry in the mid-2000s. In 2009, the government made some $30bn available in low-cost loans to manufacturers in a move that helped expand production capacity by a factor of four. By the end of 2013, according to the Ren21 renewables community group, China accounted for two-thirds of solar PV production worldwide. The top-ten suppliers had outstanding borrowings of more than $16bn by mid-2013 and Suntech became the first to default on publicly traded debt. It took several years before China promoted local installation of solar through subsidies. The US and Europe launched anti-dumping investigations, which resulted in heavy tariffs being levied on Chinese imports. Despite having no treaty obligations to comply with other than World Trade Organisation, the state still had to face the penalties that come with international trade.
Another factor is that employment turned out to be a secondary driver. Instead, China used cheap loans and direct grants to focus effort on industries it saw as strategic but which are not huge employers.
According to Bart van Hezewijk, innovations officer of the Dutch consulate in Shanghai, in 2018 alone China pumped $8bn of state money into semiconductor R&D. That took the total spent in the country close to that of South Korea, where state support accounted for just $1bn, though it was still just a quarter of the total spent by US-based companies. The attempt by the Trump administration to decouple from China on trade has focused more attention on building up a domestic industry while the US attempts to reshore its own electronics industry with a series of DARPA projects.
Far from the end of history
It was President Dwight Eisenhower who warned of too close a link between defence and commerce in his 1961 farewell address with the famous phrase “military-industrial complex”.
In the following half-century, globalisation greatly weakened the strength of this complex as manufacturers sought the lowest-cost option. When the Iron Curtain fell, the idea in Francis Fukuyama’s 1992 book ‘The End of History and the Last Man’ was that liberal democracy and trade were becoming global norms. Reflecting that, the peacetime applications of DARPA’s projects became increasingly important.
Times are changing. Fear of a China that has spent billions to get ahead in quantum computing, mobile communications, artificial intelligence and electronics manufacturing has seen US politicians and government agencies argue they similarly need to bolster their own industry.
Mike Brown, director of the US DoD’s Defense Innovation Unit, said in an autumn seminar organised by Stanford University on the issues posed by China: “In the 20 to 30 years to today you could argue China was catching up. That’s changing as they take the lead in a number of technologies. The multiplier effect of technology means we could be facing a power as strong as the US even before it overtakes ours in terms of dollar GDP.”
Ellen Lord, US undersecretary for defence who handles procurement issues, said in DARPA’s Electronic Resurgence Initiative conference: “US dependency on offshore suppliers is a national-security risk. Our US industrial base is the nexus of industrial security and economic security. I am proposing a step-by-step process for reconstituting a domestic electronics supply chain.”
It is not just the Trump administration. In May, with support from both sides, Democratic Party senator Charles Schumer proposed the Endless Frontier Bill, which would expand the National Science Foundation’s remit to fund technology R&D in up to ten target sectors.
Noting that Schumer’s act is one possible step towards the goal, Brown says the US should boost federal R&D spending by at least another $200bn, more than doubling the total. “The Chinese have made industrial policy work for them. Many folks do not see us as in a technology race with China. But China has made this a reality ... whether we like it or not.”
Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.