The world has many problems and it will take a lot of bright, educated people to solve them. So, on the face of it, it seems like a good thing that more and more people are earning PhDs in science, technology and engineering. Most countries, convinced that higher education and scientific research are key to economic growth and prosperity, are expanding doctoral education in science. The thought, as one researcher who has studied doctoral-education trends puts it, is that you can “grow PhDs like mushrooms”.

The consequence of that mushrooming depends on where it is taking place, and in which discipline, as our overview of PhD systems around the world shows (see page 276). Clearly, such expansion results in an extraordinary amount of good research (see page 283). And in the rapidly growing tiger economies, for example, most of those with PhDs quickly find good jobs.

But there are reasons for caution. Unlimited growth could dilute the quality of PhDs by pulling less-able individuals into the system. And casual chats with biomedical researchers in the United States or Japan suggest a gloomy picture. Exceptionally bright science PhD holders from elite academic institutions are slogging through five or ten years of poorly paid postdoctoral studies, slowly becoming disillusioned by the ruthless and often fruitless fight for a permanent academic position. That is because increased government research funding from the US National Institutes of Health and Japan's science and education ministry has driven expansion of doctoral and postdoctoral education — without giving enough thought to how the labour market will accommodate those who emerge. The system is driven by the supply of research funding, not the demand of the job market.

Widening concerns about dismal job prospects are dissuading the brightest candidates from the PhD route.

The problem is widely discussed, yet many PhD programmes remain firmly in the traditional mould — offering an apprenticeship for academic research, even as numbers of academic positions stagnate or decline. Yes, there are many worthwhile careers outside academia for science PhD holders (Nature would be down to a skeleton staff without them). And most people with science PhDs eventually find satisfying jobs. But they probably feel that spending years performing minipreps was not the most appropriate way to become a banker or a teacher. Widening concerns about dismal job prospects are dissuading some of the brightest candidates from taking the PhD route.

Something needs to change — but what? Ideally, the system would produce high-quality PhD holders well matched to the attractive careers on offer. Yet many academics are reluctant to rock the boat as long as they are rewarded with grants (which pay for cheap PhD students) and publications (produced by their cheap PhD students). So are universities, which often receive government subsidies to fill their PhD spots.

One way in which governments can bring about change is to better match educational supply with occupational demand. They should get smart, independent labour economists to comb through wage and employment data that reveal which types of science-related job are in short supply, and talk to stakeholders on the ground to confirm the findings. Governments should then open the doors to more PhDs only where they are most needed. Such analyses are already under way, and should be encouraged.

A second route is to reform the PhD itself (see page 261), and reset the expectations of those in the system. Imagine bright young things entering a new kind of science PhD, in which both they and their supervisors embrace from the start the idea that graduates will go on to an array of demanding careers — government, business, non-profit and education — and work towards that goal (see page 381). The students meet supervisors from a range of disciplines; they acquire management, communication, leadership and other transferable skills alongside traditional academic development of critical thinking and analysis; and they spend six months to a year abroad.

Some such efforts have already begun: for example, US institutions vie to win prestigious grants from the Integrative Graduate Education and Research Traineeship (IGERT) programme run by the National Science Foundation, which promotes highly interdisciplinary PhDs (see page 280)

The IGERT scheme shows how appropriate reward structures can drive change. Governments and funding agencies should require educational institutions to release figures showing how many of their PhD students complete the course, and how many go on to find employment and where, and should award some proportion of funding accordingly. This would also help prospective students to select a good course, and force worse-performing programmes to shape up or close.

Until any of this becomes commonplace, it is up to prospective graduate students to enter a science PhD with their eyes open to the opportunities — or lack of them — at the end. Not all mushrooms grow best in the dark.