Science, technology, engineering and maths training needs a boost, say Fanuel Muindi and Moytrayee Guha.
In today's global economy, a workforce trained in science, technology, engineering and maths (STEM) is recognised as a primary driver of growth.
Around the world, STEM education initiatives vary in scope, size, type, target populations and funding sources. What's missing is a unified global mechanism for STEM education.
Creating a Global STEM Fund would help support and implement effective and innovative STEM programmes in developing countries. 
The NGO Cosmos Education, the STEM Innovation Camp in South Africa, the African Institute for Mathematical Sciences and the Bunengi STEM Africa are but a few examples of organisations and programmes that could benefit.
The new fund would aim to improve the accessibility and quality of STEM education, particularly in the developing world, where the number of STEM programmes is likely to increase in future years. To be successful, they will need support - and that means not just funding, but also knowledge about best practices.
Is a fund needed?
Why create a STEM fund and what activities should it support?
Despite recent progress, many parts of the developing world still face shortages of highly trained scientists and engineers. According to the UNESCO science report 2010, African countries had an average of 164 researchers per million people in 2007, more than six times lower than the world average of 1,081 researchers per million. 
Poor funding limits universities' ability to set up and maintain well-equipped laboratories; offer high enough salaries to attract science lecturers; and sponsor a greater number of low-income students interested in science and engineering subjects.
The fund we propose would unify existing national and international funding. And it would need initial investment from governments, firms, NGOs and academic institutions.
Private sector backing will be critical - for financial and infrastructure support, the fund would need to work with organisations such as the Global Business Coalition for Education.
Quality and accessibility
But increasing the number of trained scientists and engineers isn't purely a matter of money. In Africa, for example, the general sentiment is that the problems are systemic and include poor accessibility and quality of science education.  The UNESCO report also paints a similar picture for many developing countries in Latin America and the Arab world.
The lack of good quality, centralised data on existing STEM programmes worldwide is also problematic. The data that do exist are patchy and do not capture indicators such as the types of educational strategies being used, collaborations with industry and universities or how STEM programmes affect students in the long term.
Without continuous and systematic evaluation, it is difficult to understand what is and isn't working. The proposed fund should initially devote significant resources to collecting and continuously updating such data so that informed decisions can be made on improving current programmes and designing or funding future ones.
The fragmented landscape of STEM programmes also leads to duplication. A number of STEM programmes with similar mission statements and goals have come together to form networks around a central mission.
One example is the 100Kin10 initiative in the United States, which hosts more than 150 organisations that share the goal of training 100,000 new STEM teachers by 2021.
The network supports member organisations by providing tools, resources and grants for the programmes they run. By working together on a central mission, these organisations can share knowledge about best practice.
A network hosted by the non-profit organisation Sasol Inzalo Foundation similarly facilitates the implementation and spread of good practice in STEM education in South Africa.
It is unclear what level of coordination exists between STEM programmes in developing countries. Creating a central database to assess this would be a necessary initial step - and this should be the fund's first goal. 
This database could be modelled after the STEMconnector database in the United States, which provides a comprehensive directory and analysis of more than 6,200 STEM education programmes across the country. One type of analysis for a future global database may include assessing the impact of STEM programmes on students' careers.
Which programmes to fund?
The success of the proposed fund will depend on its ability to bring various stakeholders together to improve the accessibility and quality of STEM education in developing countries. It will also depend on making sound decisions on which programmes to fund.
To be successful, funding proposals for each country would need to fit in with both the goals of the fund and national science policies.
And they should be evaluated based on four criteria: impact (does the proposal address a critical problem in STEM education?); approach (is the strategy and analysis appropriate, measurable and feasible?); sustainability (will the interventions be sustainable and have a measurable long-term impact?); and innovation (does the proposal challenge or shift current STEM education paradigms by using novel interventions?).
With the world discussing the targets that will succeed the Millennium Development Goals after 2015 - which should include access to quality science education - it is the right time to create this fund.
Fanuel Muindi is a postdoctoral fellow in the Department of Brain and Cognitive Sciences at the Massachusetts Institute of Technology and Moytrayee Guha is senior health analyst at the Brigham and Women's Hospital, Harvard Medical School, both in the United States. Fanuel Muindi can be contacted at firstname.lastname@example.org and on Twitter @FanuelMuindi
 Fanuel Muindi and Moytrayee Guha Developing world: Global fund needed for STEM education (Nature, 27 February 2014)
 UNESCO science report 2010 (UNESCO, 2010)
 Vivienne Irikefe and others Science in Africa: The view from the front line (Nature, 29 June 2011)
 Fanuel Muindi and Moytrayee Guha Training: African database for education schemes (Nature, 12 December 2013)