Skip to content
Site search

The cluster paradox

They bring together companies, people and skills to create a sum greater than their parts. But is there a right recipe to make cluster-led regeneration happen?

Hugo Cox
12 November 2019

South Yorkshire in the mid-1990s was a region still trying to recover from the ravaging effects of the decline in two once-proud industries: coal mining and steel. It was against this backdrop that two separate attempts were made to kick-start the local economy, each yielding starkly contrasting results.

Targeting the abandoned coalfields bordered by Barnsley, Rotherham and Doncaster, in 1995 Michael Heseltine, then president of the Board of Trade, set up the Dearne Valley Enterprise Zone, a 0.56 square mile (1.5 km2) area offering a mixture of tax breaks, grants and subsidised staff training designed to draw in new businesses and create thousands of skilled manufacturing jobs. When the zone's 10-year lifespan ended in 2005, the regional development agency Yorkshire Forward found that, despite the creation of 17,700 new jobs, most were low skilled and in the service sector, particularly call centres. Levels of employment, skills and aspirations were among the lowest in the country.

A few years later in Rotherham itself, a collaboration between the University of Sheffield and Technicut, a local specialist cutting tool company, was taking the first step in a regeneration initiative aimed at bootstrapping the fortunes of the local area. Today, the Advanced Manufacturing Research Centre (AMRC), to which this partnership gave birth, is the anchor of the Advanced Manufacturing Park (AMP), a cluster of world-class, knowledge-based manufacturing companies, drawing skilled, high-value workers to the area and providing a crucial spur to the growth of Rotherham, Sheffield and their surrounds.

"These are the type of high-skilled exporting businesses that create new innovations, boost local wages, and drive the local economy," says Anthony Breach of the Centre for Cities, a UK thinktank. "They are helping transform the Sheffield City Region's economy from a place for low-cost production to a place for high-value knowledge production."

The fact that a similar set of policies in two areas almost within walking distance of each other turned out so different is a lesson in the unpredictable alchemy of cluster-led regeneration. The practice, which aims to generate economic gains in a city or region by cultivating small, specialised business sectors, has become something of a cause celebre for economists and policymakers. The trouble is that although successful examples are easy to spot, and many people are ready identify the key ingredients, efforts to create them are as likely to fail as succeed.

"For all the compelling and widely accepted attention given to the concept of clusters over the past three decades, cluster initiatives in US regions have, for the most part, failed to live up to their expected potential," conclude Ryan Donanhue, Joseph Parilla and Brad McDearman in Rethinking Cluster Initiatives, a July 2018 paper for the US thinktank the Brookings Institution.

The theory is compelling. Productivity benefits of so-called "agglomeration" – where companies that form part of a cluster enjoy the economies of scale typically reserved for large ones – let high-value business clusters spawn long-term economic growth well beyond their specific specialisms (see panel). In the US alone, more than 1,350 academic articles have documented many benefits for the firms that are part of clusters and the regions that have them, since the term was first coined in 1990 by Harvard Business School professor Michael Porter.

"Clusters have become, in effect, simply the reality of how regional economies are organised. As such they have become a mainstay of local development policy," write Donanhue and colleagues in their paper.

Real steel

Among the classic successful examples is Sheffield's sister "steel" city, Pittsburgh, Pennsylvania. By 1983, with swathes of the US steel industry priced into abeyance by emerging world producers, unemployment in the city had hit 17.1%. Things would get worse: the wider region lost another 133,000 manufacturing jobs between 1979 and 1987. By 2003, Pittsburgh was designated as financially distressed under state law, triggering a broad programme of fiscal and employment reform, including provision for wide-scale layoffs.

But in its Oakland neighbourhood, the city had the perfect conditions for an innovation cluster. At its heart were two world-leading research institutes: the University of Pittsburgh and Carnegie Mellon University.

As start-up firms, and related support services like co-working spaces sprouted up across the district, three advanced industry clusters emerged in robotics and advanced manufacturing, technology, and healthcare and life sciences. In 2017, the Oakland district, a subsection of the wider Oakland area covering just 1.7 square miles, accounted for more than one-third of the university research output of the entire state, according to another Brookings Institution report: Capturing the next economy: Pittsburgh's rise as a global innovation city. Accounting for 3% of the city's land area, it boasted nearly 29% of its jobs.

No one would question the legacy today. Employers range from tech stalwarts such as Google and Uber, to Bayer, one of the world's largest pharmaceutical companies, and banking group PNC, besides a host of other technology and healthcare firms – most notably the University of Pittsburgh Medical Center, a leading American healthcare provider. The city has also reprised its position in metal industries. Aluminium producer Alcoa has its headquarters there, as does Allegheny Technologies, which produces the specialist alloys necessary for many aerospace and defence applications.

"Growth in the city's start-up support systems – mentorship, flexible workspaces, capital and talent attraction – are fuelling a new generation of high-value firms as start-ups like NoWait are leveraging the full pipeline
of entrepreneurial services to attract investment and grow," says Scott Andes, a fellow at the Brookings Institution, who has studied the city's re-emergence.

Image of Pittsburgh
Pittsburgh's legacy as a global innovation city highlights the impact that clustering can have on a regional economy

A degree of certainty

What both Pittsburgh and Rotherham have in common – apart from their steel heritage – is that the leadership of and investment from higher education institutions has played a key role in instigating regeneration through agglomeration. "In Pittsburgh, the pre-existing 'meds and eds' element was crucial to sparking the growth of the surrounding cluster," says Neil Lee, associate professor of economic geography at the London School of Economics (LSE).

For its part, Rotherham's AMRC has been crucial in attracting specialist companies such as Boeing, Rolls-Royce and McLaren Automotive, on to the AMP. "The institutes provided the ready stream of skilled labour and a rich local knowledge environment required to attract specialist companies," says Breach. As if to affirm this approach, the UK Atomic Energy Authority announced in September 2019 that it would open a £22m nuclear fusion facility on the park in 2020.

This type of genuine knowledge advantage is instrumental to creating the virtuous circle – companies moving to an area for skilled labour, in turn attracting skilled labour in search of work – that is at the heart of what makes urban clusters work. Both the Advanced Manufacturing Park and the Oakland district provide an oasis of high-skilled labour in city regions generally characterised by low-skilled, low-paid work.

Genuine academic expertise puts both cities out ahead of those that today claim the status of start-up or biotech hubs. "So many cities claim specialised tech, health or innovation clusters. By definition they can't all be clusters. Clusters need a combination of expertise and specialisation to work in order to attract firms that wouldn't be better served elsewhere," says Breach.

“So many cities claim specialised tech, health or innovation clusters. By definition they can’t all be clusters. Clusters need a combination of expertise and specialisation to work in order to attract firms that wouldn’t be better served elsewhere.”

Anthony Breach
Centre for Cities

Lidl detail

Although it is tempting to assume that the high-level skills and expertise of universities are vital ingredients in creating a cluster, in some instances it's more a matter of private sector will and investment. Heilbronn is a small city about 50km north of Stuttgart, which forms part of the latter's automotive CARS (Clusterinitiative Automotive Region Stuttgart) supply chain and is also home to the headquarters of European discount supermarket chain Lidl.

Already there is a public institution in place – the recently opened campus of the Technical University of Munich School of Management, devoted to the study of how effective management and digitisation can grow small and medium-sized companies. But Frieder Mitsch, a PhD student at LSE's department of government who has been studying the city's development, says that in future the "Lidl campus" will play an even more important role. The brainchild of the chain's founder and owner Dieter Schwarz, the institute devoted to "the future of retail", aims to add several thousand jobs to the local market by 2025.

"On the one hand there is, in Lidl, a pioneering multinational company willing to make big investments. On the other are the R&D facilities – including the new Lidl campus – that will train and retain skilled people for the clusters of regional business sectors," says Mitsch.

Where they succeed, cluster-based development policies often piggyback off effective government services. Lee says that, where they have worked in Europe – besides Germany he points to the Danish capital Copenhagen and Malmö in Sweden – the local area has boasted a "basic hygiene infrastructure" comprising reliable public services and good connectivity – both physical and digital. The local authority's spending on cultural initiatives – including festival and arts programming – which broaden the city's appeal as a place to study, is providing an important plank of the Heilbronn development plan, says Mitsch.

City of Copenhagen
LSE associate professor of economic geography Neil Lee points to Copenhagen as a city where cluster-based development policies can piggyback off effective government services

But Bangalore bucks IT

A successful cluster, then, is one that can draw on the expertise of academic institutions, and is supported by local government policy. So how does that explain the world-leading tech-based clusters that have formed in Bangalore over the last 40 years, which have emerged in spite of, rather than because of, the support of reliable local governance?

Following state elections in 2005, where a poor rural majority showed little interest in the prospects for the urban elite, the city's infrastructure was neglected. It had long been characterised by water shortages, poor sewers and unreliable power supply, features that stood in stark contrast to its breakneck growth, which by then was adding nearly 900 cars per day to the streets.

A new 13.5% "entry tax" raised on goods entering the state imposed a heavy burden on local IT firms that relied on imported computers and components. "Few in the IT industry have much good to say about the new government," observed an Economist article at the time.

But the impediments did little to stem the rapid growth of the city's technology clusters. The big IT firms – many based on huge purpose-built campuses in Electronic City, a few kilometres outside the city centre – were building their own infrastructure to secure reliable power and a harmonious working environment. Nasscom, the industry's lobby, forecast IT service growth of 25-28% per year in 2005.

More than a decade later, the gap between the demand for infrastructure spending in India and its supply is still a gulf. "India's infrastructure deficit is simply too large to eliminate any time soon," declared Standard & Poor's Global Ratings credit analyst, Abhishek Dangra, in the 2018 report India's Infrastructure Marathon: Why Steady Growth Can't Close The Supply Gap. Dangra notes the $4.5tr of investment that the government estimates will be needed by 2040.

Tech leaders remain unconcerned. An Economist Intelligence Unit survey of 2,620 executives in 45 cities, commissioned by telecoms company Telstra in 2017, revealed that executives in Bangalore were the most optimistic in the world that they had the right environment around them to grow their companies. Two other Indian cities – Mumbai and New Delhi – make the top five, with San Francisco and Beijing completing the line-up (see infographic).

"Digital ecosystems can thrive without the municipal government playing a proactive role," the authors of the report, Connecting Commerce: business confidence in the digital environment, conclude. As hard as they are to seed, when business clusters take root, they will thrive under even the harshest environments. No wonder policymakers are still scratching their heads.

  • We will explore successful city clusters at our next World Built Environment Forum Summit, to be held in Shenzhen, China, on 18-19 May 2020. Register your interest at

What is a cluster, anyway?

Business clusters are not a new phenomenon. Around the world and throughout history, towns and cities have attracted labour and capital by specialising in specific industries. In the UK, Nottingham hosted lace-makers, Luton hatters, and the cluster of towns that form the city of Stoke-on-Trent are so associated with pottery that the region's nickname, "the Potteries", endures today.

With the disbanding of the US movie studio system in the 1930s, specialist companies and individuals sprung up in its place across Hollywood, California. Co-locating in the same area, sharing knowledge, a labour pool and – sometimes – resources allowed them many of the economic benefits that had allowed the studios to grow so mighty before them, not to mention sidestepping some of the disadvantages – such as unionised labour – that had helped the studios fail.

The productivity benefits of co-locating in this way – which economists call "agglomeration" – come in three areas: sharing, matching and learning. In the first case, companies can share the facilities, infrastructure and suppliers that meet their particular needs. Second, they are able to find workers quickly and efficiently by virtue of a deep, well-suited local labour market. Third, the dense, knowledge-rich environments facilitate the exchange of ideas and innovation between companies.

"Probably the most important dimension is matching," says Neil Lee, associate professor of economic geography at the London School of Economics. A company seeking specialist skills such as coding would rather locate somewhere it can find skilled workers than fund the expense of training staff to code. Workers, meanwhile, are drawn to specialist clusters by the prospects of finding the biggest range of jobs and the highest wages for their skills.

Where a number of inter-related specialist industries locate near each other – like the finance, technology and legal skills required by a fintech firm in Shoreditch, adjacent to the City of London – clusters foster particularly speedy benefits.

In an era of global labour markets, rapid transport and high-speed telecommunications, many forecasted the waning benefits of clusters as business broke free of the constraints of geography. The death of clustering turned out to have been greatly exaggerated, and it remains a central plank of current economic development.