What are we learning from policy experiments to increase innovation and entrepreneurship?
What are we learning from policy experiments to increase innovation and entrepreneurship?
Everyone agrees that innovation is vital to accelerate productivity growth and solve the big challenges of our era. The problem is that we don’t really know what works and what doesn’t to increase innovation.
Across Europe governments spend around €150 billion every year supporting entrepreneurs and businesses to innovate and grow. Yet little effort is done to experiment with new approaches and create evidence about their effectiveness, particularly in OECD economies.
In order to change this, over the last four years at the Innovation Growth Lab (IGL) we’ve been working with governments, foundations and researchers to make innovation and entrepreneurship policy more experimental and evidence-based.
We’re still far from other fields such as medicine - where randomised controlled trials (RCTs) are standard; development - where organisations like J-PAL and IPA have run over 900 trials in the past 15 years; or education - where the EEF has supported over 150 trials to find out what works to improve educational attainment.
Our field is much younger, but over the last few years we have seen a growing number of RCTs. Many of these are (co)funded by the IGL Grants programme, which has supported over 30 trials with close to $3 million from the Kauffman Foundation, Nesta and the Argidius Foundation.
We are starting to learn valuable lessons about how we can encourage innovative ideas and support businesses, and many more lessons will emerge as the trials now in the field start to deliver results. Continue reading for a preview of what we’re learning, and join us at our Boston conference at HBS and MIT in mid-June to find out more about them and explore new ideas for policies and programmes to increase innovation and entrepreneurship.
How do we get more and better ideas?
There is no innovation without new ideas or the creative re-combination of old ones. But how can we ensure that ideas easily flow and thrive?
Are we as a society creating an environment that allows us to tap into all sources of new and interesting ideas? Unfortunately, the answer is no.
Whether we look at schools, universities, or businesses, we are missing out on many potential innovators and their ideas. Finding solutions to this is key for a better future.
Increasing exposure to innovation
For instance, Raj Chetty and his colleagues show that unless you are a top student from a high-income family, your chances of becoming an inventor or filing a patent are very low due to lack of exposure to innovation when young. This means we are missing out on entire generations of inventors and their good ideas - the so-called “lost Einsteins”.
Losing out on this group’s ideas is detrimental to all of us, so a new IGL trial led by the World Bank is testing out an online intervention to expose over 19,000 children in Latin America to STEM and entrepreneurship. This is only one of many ideas that could be trialled to cultivate innovative and entrepreneurial attitudes early on and address this important policy challenge.
Encouraging more people to participate
Encouraging individuals or groups who may not naturally consider themselves innovators or creative is another way to make innovation more inclusive, but does it pay off? Two IGL trials suggest that it does.
Both experiments were based around innovation contests, albeit in very different settings. The first trial was with engineering and computer science students at a US university, while the second took place at a large multinational in the Netherlands.
Despite the settings and research questions being slightly different, two of the findings were surprisingly similar. First, both studies showed it is possible to use messaging nudges and/or small financial incentives to encourage people to submit ideas into innovation contests. Secondly, and most importantly, there was no real difference in the average quality of the ideas submitted by someone who actively chose to participate in the innovation contest and someone who needed to be encouraged to join.
In other words, encouraging more people to participate can lead to more ideas without decreasing their quality, and a small tweak in the process can be sufficient to make it happen.
If we don’t do it and instead rely only on self-volunteered contributors, we are missing out valuable ideas.
The trial in the Netherlands also looked at other ways to influence the quality of the ideas submitted, such as trying to widen the horizons of participants by showcasing successful projects from prior internal innovation contests. It turned out that this was counterproductive, making people less creative rather than more. Whether there are simple ways to make people more creative is something that another IGL trial in the UK is exploring, in this case looking at whether creativity can be trained through habit creation.
Collaboration is an increasingly important component of any innovation process. The romantic idea of the sole inventor, with their ‘lightbulb moment’, is today generally considered to be a myth. Instead, most scientists agree that complex challenges benefit from the combination of different expertise, knowledge, and backgrounds.
Yet we have little evidence on what the best ways to encourage collaborations are, both within and between universities and businesses. For instance, how important is physical proximity between researchers to facilitate collaboration?
Anecdotal evidence suggests that distance matters, and many new science labs have been built under the assumption that locating researchers from different fields under the same roof will unlock interdisciplinary research and open original research avenues. An IGL trial in Eastern Europe is currently testing whether this is actually the case, by randomly distributing research groups within a large temporary research building.
But is physical proximity enough? A recent trial in the US by Karim Lakhani (professor at HBS and member of the IGL Scientific Committee) and his colleagues suggests it is not, although they show there is an easy fix. A simple low-cost intervention, like bringing together scientists working under the same roof to talk about their ideas for a couple of hours, creates new collaborations that otherwise would not exist. Both these trials are rare examples of applying the scientific method to science policy.
Collaborations between researchers and businesses are also a source of new ideas, but this too is an area where there is overwhelming agreement that we are missing out on many opportunities. A number of IGL trials are looking at different ways of addressing this challenge (from innovation vouchers to workshops), and we are planning further work in this space.
How can we support entrepreneurs and business to scale and adopt new ideas?
Good ideas are not of much use unless they are put into practice, scaled up and widely adopted. This is why we have a long list of programmes and policies to support this process, ranging from entrepreneurship training initiatives, accelerators and other startup support programmes at one end, to innovation grants, SME finance schemes, business support or tech adoption programmes at the other.
But the impact most of these programmes have is unclear, and we don’t know whether changing their design would make them more or less effective. A growing number of trials are trying to provide some answers.
Training entrepreneurs and supporting startups
Entrepreneurship has become increasingly popular, and this is reflected in the growing number of universities, private providers and governments offering entrepreneurship training today. What remains unclear is which type of training best fulfills the needs of entrepreneurs.
A recent trial by World Bank researchers in West Africa shows that a personal initiative training approach, which teaches a proactive mindset and focuses on entrepreneurial behaviours, can be much more effective than teaching them formal business skills, such as marketing or financial management.
A number of IGL trials are also considering similar questions. One trial in Jamaica is comparing traditional business skills classes with classes on personal initiative and persistence. Another, in Italy, is teaching entrepreneurs to become more experimental by teaching them to use hypothesis-based experiments to assess the viability of the business idea(s) and evaluate the effect of their strategies. Early results from their pilot study suggest that the training had a positive effect on startup performance.
Independence and not having a boss are often cited as reasons why entrepreneurs decide to start their own business. However, preliminary findings from another IGL trial taking place in a large accelerator programme in Latin America suggests that introducing some accountability structures is actually quite useful and helps to improve startup performance.
Accountability is important, but even simple feedback without strings attached can make a difference. Government agencies are often reluctant to share detailed feedback on the proposals that they review, for the fear of opening the door to lots of complaints. But are we losing out from that? A trial presented at the IGL2017 Conference showed that giving startups in the Startup Chile programme the feedback collected as part of the selection process increased both external fundraising and survival probability. One of our IGL partners is now trying to replicate this trial with one of its programmes, in order to decide whether it is worth sharing the detailed feedback that they are collecting.
Supporting SMEs innovation and productivity
Reversing the productivity slowdown requires getting more SMEs to innovate and/or adopt new technologies and production methods. But it's an open question what the best way to do this would be.
There is a broad spectrum of potential interventions, some very intensive and others much more light-touch. Two trials in India and China demonstrate that they both can work. Nick Bloom (a professor at Stanford University and member of the IGL Scientific Committee) and his colleagues found that deploying “high-grade” management consultants into poorly managed Indian textile firms improved their management practices and productivity. The intervention was not cheap, but the productivity gains more than offset the cost, and they persisted over many years.
At the other end of the spectrum, another recent trial shows that a simple low-cost intervention that gets businesses regularly meeting with each other in groups facilitated peer-learning and new collaborations, leading to increased sales and profits.
Instead of in-kind support, other programmes directly provide funding to SMEs, either small or large amounts. For instance, Nesta’s Creative Credits trial successfully used small vouchers to encourage SMEs to work more closely with creative suppliers (although these new relationships didn’t last in the long term). Many other vouchers schemes are based on a similar logic, and the jury is still out on what their ultimate impact is. Another IGL trial on innovation vouchers in the UK is trying to shed some more light on this question.
While small vouchers schemes are popular, much more budget is allocated to funding large R&D and innovation grants. But do these types of large grants replace existing investment that firms would have made in any case, or do they mostly lead to new activity? An IGL trial led by the World Bank in Latin America is trying to answer this question.
Randomising innovation grants as large as $250.000 would not be a very popular policy decision, so all the funding applications that are scored highly by all reviewers will get the grant, while those that everyone scores poorly will not. Funding for applications in which there are disagreements between the different reviewers will be randomised, and the impact tracked. The implicit, yet untested, assumption is that value for money is higher for the applications with the top scores, although it could well be that these are precisely the ones that companies or investors would have funded in any case.
This trial in Latin America is also trying to understand who is better at making decisions about which companies to support. This fits into a much wider question, namely how do we run selection processes to allocate public research and innovation funding. This is an area that is ripe for experimentation in which at IGL we are planning additional work.
The time has come for governments to embrace experimentation
Governments have been slower to set up trials in this space. One exception is the UK’s Department for Business, Innovation and Skills (BIS). Our IGL colleague James Phipps was much involved in the BIS experimentation journey, and describes his experience here.
BIS started small, with simple messaging trials that among others demonstrated the power of an Adam Smith quote (on the virtues of volunteering) to recruit more mentors. These messaging trials demonstrated that simple tweaks could have a substantial impact - the difference between delivering or missing a ministerial target. And, crucially, they demonstrated that despite their best efforts they didn’t have all the answers, often encountering surprising results that showed that their initial assumptions were wrong. This led BIS towards much larger-scale trials, including the largest RCT undertaken in business support (the Growth Vouchers scheme with 30,000 participants).
This is a journey that many of the other government agencies and ministries that are partners in IGL are currently going through. At IGL we’ve been working with them to identify where experiments could increase the effectiveness of their programmes, and we are supporting them to take some of these trial ideas forward.
While some are “simple” messaging trials, we are also helping set up trials looking at other issues, such as improving SME-university collaboration, providing management and coaching support on top of R&D and innovation funding, and helping SMEs to adopt digital technologies.
Governments are not the only source of ideas for new schemes, so we are also advising a few of them on how to design bottom-up experimentation funds. These are a useful mechanism to encourage new programme ideas from those working most closely with innovators and entrepreneurs, and to test them in order to find out which programmes actually work, with the aim of scaling those up.
Through our work advising, training and supporting over 12 governments around the world we’ve learnt some useful lessons about how to embed policy experimentation in the delivery of support programmes.
It can be done. It doesn’t need to be expensive. It helps to develop more effective interventions. It can be incorporated into existing programmes. There are lots of missed opportunities and the results may surprise you.
Ultimately, it can make the difference between success and failure.
So now is the time to look for new ideas, whether small or large, and test them out.
Learn about new ideas at the IGL2018 conference in Boston
Building on the success of last year’s IGL2017 conference, when we brought together 250 senior policymakers, practitioners and researchers from over 30 countries to discuss new policy ideas and the latest policy experiments, this year the IGL conference is heading to Boston, in partnership with Nesta, the Laboratory for Innovation Science at Harvard, J-PAL, the Kauffman Foundation, the World Bank, and the Sloan Foundation.
Taking place on 12-14 June at HBS and MIT, IGL2018 will explore future innovation, entrepreneurship and small business policies with over 50 world-leading experts.
With 10 keynotes and panel discussions and over 13 practical sessions and workshops, IGL2018 will be an opportunity to:
- Join a global community of delegates at the forefront of innovation, with confirmed participants from 20 countries;
- Learn about the next generation of innovation, entrepreneurship and small business policies;
- Find out about the latest evidence and policy experiments; and
- Take part in capacity-building workshops on new policy tools and methods, led by organisations such as NASA, Harvard Catalyst, the US Small Business Administration and the World Bank, among others.
We hope to see you there, but if you cannot make it, don’t forget to sign up to the IGL newsletter to find out the results of these trials, and others to come.