Geoff Mulgan

10 min

When should states be creative, innovative or entrepreneurial - and when should they not?

Updated: Sep 4, 2023

The summer brought several festivals about innovation in government. Berlin hosted the sixth Creative Bureaucracy festival with an extraordinary range of people and presentations, from Brazil to Sierra Leone, Colombia and Canada to Germany and Estonia. And UCL’s IIPP hosted a festival around the idea of the ‘entrepreneurial state’, also with some excellent speakers from around the world.

Having worked in this space for several decades I'm often asked about the desirability and viability of states becoming creative, entrepreneurial and innovative. I'm generally an enthusiast for anything which brings life and imagination to the work of governments. But as I will show, the answers are not as obvious as they might seem.

There is no doubt that public sectors can innovate – the Internet and the Worldwide Web, dramatic reductions in child mortality and the fast responses to the pandemic are all counters to the lazy prejudice which assumes that the state cannot innovate (I've written various pieces over the last decade trying to document examples, methods, barriers and enablers, some summarised here. The picture below - from Reykjavik - is a nice summary of the traditional view).

Much of the history and many of the recent examples aren't widely known and when my former boss Tony Blair commented recently that ‘in the private sector, if you don’t innovate, you go out of business, but it doesn’t happen in the public sector’ he was just repeating a widely held view. As I will argue, however, the paradox of the public sector is subtler: it often does innovate, but it doesn’t usually manage its own innovation well, with the kinds of funds, teams, accountabilities, and investment in capabilities, that are normal in much of business or science.

But just how far should governments go? How creative, innovative and entrepreneurial do we want them to be? Here precision helps. Creativity, entrepreneurialism, innovation and invention are often used with overlapping meanings but they are distinct. You can be creative but not entrepreneurial, entrepreneurial but not creative, and inventive without ever getting to innovation. And none of these is inherently benign.


Let's start with creativity. It can be wonderful to see surges of creativity in governments, often led by new leaders or teams. Sometimes necessity forces it as in Ukraine where the government has leaped ahead in its use of digital technologies pushed by the needs of war, and as in some natural disasters.

History is full of examples of officials - and ministers - thinking and acting creatively (the film Oppenheim documents one such example), and indeed it's only fairly recently that creativity has been more associated with business than the state.

But we wouldn’t want governments to be creative all the time, experimenting with different coloured traffic lights or changing the names of agencies every year. Just as we don't want accountants to be too creative, there are also good reasons why much of the world of government be relatively uncreative, focused more on disciplined, rigorous implementation than novelty.

I love creative bureaucracy festivals in part because governments greatly underestimate how useful creativity can be, and how much latent creativity there is in their employees (and Charles Landry deserves a lot of credit for making his Creative Bureaucracy festival such a success). But I wouldn’t want every bureaucracy to strive for creativity every minute of the day.

States supporting innovation

It’s slightly ironic that the core argument of the influential book titled ‘the entrepreneurial state’ (by my UCL colleague Mariana Mazzucato) was not that states should be entrepreneurial but rather that they should support innovation, and indeed had supported innovation for a long time.

The book – and its follow ups – steered a long way from the textbook definitions of the word ‘entrepreneurial’ (which usually emphasise profit and financial risk) but instead provided telling stories of public funding for key technologies and showed how vital government engagement had been in the past and would be in the future. Support for innovation was key to the development models pursued by Germany, France and Japan since the 19th century, and by countries like south Korea and Taiwan more recently. For nearly two centuries the US was a trailblazer, from government support for agricultural innovation in the 19th century to semiconductors in the 21st, while China’s vast ‘Government Guidance Funds’, launched nearly 20 years ago, are probably the biggest contemporary example (though rarely mentioned in the West).

But theory, and in particular, much of mainstream economic theory, had lost sight of this, especially in the US and UK, and it was useful to return them to the older common sense, and the lessons of history.

However, these are all examples of governments supporting innovation rather than government being innovative itself. And they leave open the purpose of innovation – is it military prowess, glory, economic growth or public wellbeing? As I show in my forthcoming book on science, state-driven innovation has tended to serve the interests of states and business more than the public. Moreover, even if economic growth is the only goal states want to promote through innovation (rather than, for example, improvements in care), the logics of support can be tricky. Most would agree that state support can be good both for the quantity and the direction of economic growth. But it’s harder to specify exactly how it does so. One recent survey summed up the difficulty of knowing how it works: ‘despite the fact that governments expend significant resources on R&D every year – over $158 billion in the OECD in 2020 – the answer remains unclear’ and another recent study argues that although 85 per cent of the US’s economic growth derives from research and development, precisely which R&D achieves the impact is less clear, with many high profile and very expensive programmes delivering meagre results, from nuclear fusion and graphene to the $100 billion War on Cancer.

I’m one of the people who has been arguing for a decade that governments needed to be much more strategic about AI, including its uses for public services. In this respect I'm in sympathy with a recent comment from OpenAI’s Sam Altman, who complained that governments should be running the kind of big R&D programmes on LLMs that Open AI is doing, but just lack the ambition to do so. He's right; but doing this well requires a lot of complementary skills, capabilities and methods that are largely missing.

The same applies to decarbonisation. As governments vastly expand spending on innovations related to net zero we should expect to see both the virtues and the vices amplified. The upside will be an acceleration of development in fields ranging from batteries to wind power, EVs to solar. The downside will be lots of duplication and waste.

States supporting entrepreneurship

What about states supporting entrepreneurship as opposed to innovation? Many do: through subsidies for start-ups, venture capital, accelerators, training programmes and much more. Again, support for entrepreneurship is often seen as a good thing. But, again too, it isn’t always in practice. The Innovation Growth Lab at Nesta was a unique attempt to bring more rigour to understanding all of these programmes which cost hundreds of billions in public money each year - and to find out, using rigorous experiments, which if any were successful.

It turned out that many probably weren’t and had very high deadweight costs. But it’s easy to see why governments like to be seen to back entrepreneurship, and in cultures that were suspicious of start-ups, more support for accelerators and early-stage investment of all kinds was a useful correction.

Innovation within government

A rather different set of issues arise around innovation in government itself – from welfare systems and tax collection to schools and police forces. I have written about this over many years and worked with dozens of governments on the practical ways they can embed innovation in everything from schools and healthcare to welfare and tax collection.

A multitude of methods are available for finance, for organising innovation, whether involving users, frontline staff, civil society and business; for organising experiments and testbeds of all kinds; for evaluating innovations and much more (and Nesta published many guides and toolkits, and compendiums, as well as gathering hundreds of innovation labs together to share experiences and make sense of the patterns). There's a lively literature on how to innovate, mainly written by practitioners such as Christian Bason, Jennifer Pahlka, Bill Eggers and Hilary Cottam, and there are also now growing subfields, with their own methods and literatures, often linking public and private sectors, with labels such as GovTech, Edtech or Healthtech, or involving spinoffs (like the big cybersecurity sectors that have grown up around the NSA and GCHQ).

My main complaint is that not that public sectors aren’t innovative – they often are – but how little they manage innovation. I’m still disappointed how poorly organised this function still is in most governments, without well designed teams, processes, budgets or metrics (of the kind that are normal in business), despite many detailed prescriptions and proposals from Nesta, the OECD, Bloomberg and others (it’s also quite poorly organised in academia – with many writings on the subject apparently unaware of even the recent history, so that it often feels as if the field is going backwards rather than forwards, forgetting faster than it learns).

In the UK government for example, as far as I'm aware, noone has responsibility for public sector innovation, and could tell you how much is spent on it, or whether it's working well or badly.

But innovation shouldn’t dominate public organisations. As in business, its vital to have some teams and units focused on innovation, and vital that all staff have at least some opportunities to be creative. But most public organisations, most of the time, should emphasise adoption and use of the best available evidence rather than novelty.

Indeed, I think they could be far more systematic in their use of evidence and in their attention to adoption (the former is a large part of my work in IPPO, the International Public Policy Observatory). If other places have tried things that work, and there is strong supporting evidence, then it's hardly wise to spend scarce resources on novelty. I'd also love to see more governments become as serious about adopting and adapting others' ideas as firms like Apple (who were brilliant at synthesising new products out of the innovations of firms like Rank Xerox) or Samsung (which did much the same to Apple).

Entrepreneurial states

So what about governments being entrepreneurial as opposed to creative and innovative? Here things get even trickier. The classic commentary on the subject was written by the great Jane Jacobs (in her book ‘Systems of Survival’). She pointed out the differences between what she called the ‘guardian syndrome’ and the ‘trader syndrome’. The first is common in governments, the second in business. She argued that all societies have to find a balance between these very different views of the world. The first is concerned with looking after things and protection, originally of land, and can be found in governments, ecological movements as well as aristocracies. The second is concerned with exchange and profit, and is the world of commerce and trade.

These each see the world in very different ways. But in practice they complement each other – indeed their complementarity is what helps societies to function.

In her view, however, fusions of the two tended to be malign pathologies, for example when businesses became like governments, running large areas of territory, or when governments start thinking like traders. Donald Trump was a classic example – who saw the government machine rather as an entrepreneur would see his own business. Silvio Berlusconi was another – a remarkable proportion of his initiatives were essentially designed to promote his businesses, or protect him from prosecution.

Jane Jacobs’ points become very obvious in some industries, like the contemporary digital industries that have become de facto utilities on which we depend every day. It remains far from clear that companies like Meta or Google appreciate that they risk becoming pathological fusions of business and government, without the mindsets appropriate to their new-found power.

The pathologies are also very visible in many parts of the world where the state runs a lot of industry, often with the military playing a leading role. Examples include Pakistan, Myanmar, China and Russia. In these cases public servants really have become entrepreneurs. In some cases – like Huawei – great businesses have been grown. But most of the time such fusions of government and entrepreneurialism tend towards corruption, and predatory extraction of value, because when the state’s monopoly of coercion connects to the power to make money abuses are inevitable.

There may be occasional examples where states should be entrepreneurial at least in mindset – spinning off a function or using some of the ethos of a start-up, for example to create a new digital service. But in such cases very tight rules are vital to avoid abuse, so that if, for example, a part of the state is spun out it doesn’t do so with advantages or easy money or legally guaranteed monopolies or inflated salaries. Much depends on whether you use the word entrepreneurial in a precise sense (the first definition that comes up on Google is: ‘characterized by the taking of financial risks in the hope of profit’) or as a much looser synonym for being innovative or problem-solving.

Many other philosophers and social thinkers have made similar points to Jane Jacobs about the need for different cultures, ethos and values in different parts of society: why we want cautious auditors and bold entrepreneurs, tough brave soldiers and sensitive, creative artists; why we want to avoid accountants becoming too creative, the military running factories or doctors doing unnecessary operations to maximize their income.

This point – that all healthy societies rest on a plurality of cultures and moral syndromes – is often under-appreciated. Students can learn about economics, or business, or even public administration, without ever having this pointed out to them. Perhaps this is why there is such scope for confusion when words like creative, entrepreneurial and innovative are mixed with words like bureaucracy, state and government.

So yes: states should support innovation and generally have, for more than a century. Only in orthodox economics did this ever look strange. Yes states should encourage innovation in their public services, and celebrate the creative bureaucrats, but alongside serious attention to evidence, adoption and learning from others. Yes they should be entrepreneurial (in the broader sense of the word). And yes they could be far more strategic in shaping, and making use, of powerful emerging technologies such as generative AI.

But no, they shouldn't try to innovate and be creative all the time and no, states and bureaucrats should not seek to be entrepreneurs, or if they do, should operate within very strict limits. Otherwise you risk finding yourself on a fast track to corruption, predation and abuse of power.