Latin America is styled by the 2016 Global Innovation Index as "a region with untapped innovation potential [but] with important risks to innovation efforts in the near-term". A crucial element to avoiding the risks, and realising the potential, is the capability and capacity of the innovation system in the countries and region to nurture and support effective innovation policies.

Part of the issue is simply understanding the current state of play of the innovation system. For example, what areas of capability are being focused on which innovation challenges? How do parts of the system work together? What skills and authority are they deploying to achieve their goals, and why? For this we need stronger analyses of both the individual institutions which develop and deliver innovation policy, and the ways in which they interact to create an (effective) system for innovation.

Nesta has recently led pioneering work on analysing innovation agencies, examining their role, strategies and development - which will shortly be built on by major contributions from the World Bank and the Inter-American Development Bank.

This work can be complemented by analyses of how individual agencies contribute to the capability of the system - in 2015 Nesta looked at the structures of the public systems of the Pacific Alliance countries: Mexico, Chile, Peru and Colombia. Many of those countries are seeking to actively co-ordinate across the public sector institutions using 'co-ordinating councils' or overarching commissions of ministers to help bridge traditional divides between, say, the strategies of research funding agencies and entrepreneurship agencies, to varying degrees of success.

So, how can Latin American countries build on their emerging understanding of their innovation systems in order to better develop their innovation policy capabilities to compete and thrive in the 21st century? Here are five recommendations:

1. Think big - look to tackle the global challenges

Many of the individual institutions of Latin American innovation systems are focused primarily on private business development in the economy. Although the Latin American framing of some challenges has traditionally had more of a combined social and economic perspective than its developed economy counterparts, the primary strategies of the main innovation-responsible agencies tend to focus on private sector, cluster or business development.

The excellent work of Johan Schot and others at SPRU suggests that Latin American countries and others need to position innovation policy as a mechanism for tackling broader economic and social challenges.

As Professor Schot says, the future of innovation policy must be a more ambitious set of goals, focusing "on mobilising the power of innovation to address a wide range of societal challenges including inequality, unemployment and climate change."

2. Engage across the whole ecosystem

It would be an oversimplification to suggest that Latin American governments do not consult or work with external groups - universities, business sector representatives etc - when they plan innovation policies. But many of their policies are initiated from government and, largely, the development and enactment of policy is through public sector ministries and agencies, particularly so when fresh legislation is required to execute them.

Innovation policy, however, is increasingly formulated and delivered by the close collaboration of different parts of the innovation ecosystem, at all stages of development.

The pioneering MIT development initiative, the 'Regional Entrepreneurship Acceleration Programme' is deliberately structured to require senior representatives of government agencies, university-based entrepreneurship programme managers, private sector innovation promotion professionals, and sector and industry representatives to work together on shared-risk-and-reward projects to develop their innovation systems.

These types of programmes offer a model for the more collaborative ecosystem-wide policy making for innovation.

3. Use the most effective data and evidence (which might not be the most robust)

The quality of 'classical' economic data for innovation policy making across Latin America could probably best be described as highly varied. But even in the UK, there are sharp critiques of those traditional economic measures and indicators when it comes to assessing innovative sectors, industries and practices; that they can't keep pace with change, that they don't describe the reality on the ground, and that they can't easily apply to complex service organisations.

Newer approaches to the sourcing, gathering and analysing of data for innovation policy will be crucial for both developed and developing countries in the future.

Nesta's work with the Welsh government on 'Arloesiadur', a dashboard of analysed new data types for innovation policy, points to a potential solution. For example, forecasting job skills needs has traditionally relied on periodic economic surveys. By contrast, constantly scraping online job boards for a city or region and using machine learning to process that data can provide rolling updates on, say, programming languages required for different jobs in different locations or sectors.

This kind of data usage may be particularly powerful in Latin American and other emerging economies, where the traditional data sources are already less reliable than developed counterparts, allowing them to 'leapfrog' the quality of data without relying on hugely expensive exercises to increase traditional economic data quality.

4. Be able to better evaluate what works

This is also a necessity for innovation policy in the future for both developed and emerging economies: a very large number of innovation policies and strategies proceed without robust evidence that they work. Latin American countries should join the growing global movement for more serious evaluation of innovation policy initiatives, through a variety of means, up to and including randomised control trial approaches.

Having a much stronger handle on which initiatives are the 'difference that makes the difference' in a particular country context will be a crucial step in increasing the effectiveness of an innovation system.

Nesta's Innovation Growth Lab provides a forum and expertise on how to improve the quality of innovation policy evaluations, and a repository of knowledge on 'what works' in this field.

5. Use international collaboration to accelerate capability development

The 2016 Global Innovation Index also notes: "much underused potential for innovation collaboration exists at the regional level— within Africa, within Asia, within Latin America and other regions.”

The majority of government innovation agencies seem prone to view their counterparts in, particularly, neighbouring countries as rivals rather than as potential collaborators. And yet the vast majority of innovation goals for a country are not zero-sum games: in discovering what works, and for learning new approaches, techniques and strategies on innovation policy, collaboration between agencies and nations is likely to be highly beneficial.

Country groupings in Latin America, such as Mercosur and the Pacific Alliance, have strong possibilities for effective collaboration on innovation policy initiatives.

In 2016-17, the Nesta-led Global Innovation Policy Accelerator provided the UK expert support for teams of innovation policy makers from Pacific Alliance countries to share experiences, understandings and to consider how to develop a shared agenda. It offers a model for future work on regional collaboration for innovation policy.

This is a summmary version of a talk given to the University of Edinburgh Latin American Forum, 15th February 2017. Many thanks to the Forum team and Forum VP Carolina Isenberg for inviting me to speak.