Most students who enrol in US science and engineering PhD programmes hope to pursue an academic career. However, the gulf between the supply of newly minted PhDs and the availability of faculty positions widens each year. Some 36,000 people earned science and engineering PhDs in the United States in 2011, but US universities create only around 3,000 tenure-track positions annually. And with about 70% of those graduates taking a postdoctorate (M. Schillebeeckx, B. Maricque and C. Lewis Nature Biotechnol. 31, 938–941; 2013), many trainees end up in a holding pattern, waiting for faculty jobs that are unlikely to materialize.

Credit: Neil Webb/Getty

This is because the higher-education sector has not delivered an essential component of an efficient market: current and precise information about job prospects, including the specific attributes and training that have enabled PhD holders to find success in and outside academia, and the differences in those job markets for science, technology, engineering and mathematics (STEM) subfields.

Although some trainees may be told that faculty positions are a long shot, self-delusion is often part of the decision to pursue lengthy postdoctoral fellowships. After all, those who go to graduate school undeterred were often at the top of their undergraduate class, and believe they can 'beat the odds' of the job market. They have never experienced what it is like to be average — surrounded by equally bright peers.

Of course, the research enterprise colludes in this self-delusion, because it would not survive without the cheap labour supplied by graduate students and postdocs. In addition, trainees may have the intellectual ability to excel in academia, but many are left unprepared for the financial, psychological and personal costs of being on the academic job market for several years. Nor do many of them fully grasp the cost of waiting for a job that might never materialize. Students may remain woefully uninformed about alternatives, because there is even less information about the job market for non-faculty careers, or about the combination of skills and experience that are required to land a job outside academia.

Without better information about the specific set of qualifications, skills and experience required for finding a tenure-track or non-academic job in individual STEM fields, the job market is unlikely to be self-adjusting. Institutional demand for cheap research labour will continue to pull in hordes of graduate students and postdocs, even though the demand for tenure-track faculty members is vanishingly small.

Those pursuing a PhD need a more accurate picture of the academic and non-academic job markets, and they need it well before they graduate. This would provide a smooth transition rather than a rude awakening upon graduation. Federal agencies such as the US National Institutes of Health (NIH) and the US National Science Foundation (NSF) provide support for most US graduate students and postdocs, and they should also step up as information brokers. They should track career outcomes for academics and beyond, and gather data to compare people with and without PhDs in similar industries or capacities so that potential candidates can assess whether a PhD programme is likely to pay off. This information will not only benefit trainees, but also help federal agencies to manage the investment of government money in training STEM graduates. When graduates have limited career options that match their training, the return on investment is suboptimal.

Trainees need to know what employers are looking for. For faculty jobs, publications, gender and institutional prestige play a part, but so do complementarity with potential colleagues' research and an institution's orientation towards research, teaching and diversity. The criteria for hiring in non-academic careers are different, with an emphasis on transferable skills such as leadership, communication and teamwork. Non-academic employers also value work experience — sometimes more than academic credentials — and trainees who are interested in non-academic careers should gain this experience early on to avoid the common PhD catch-22 of being labelled simultaneously over- and under-qualified.

For the sake of future scientists, information about the current and projected state of the job market should be regularly collected, analysed and disseminated. Universities should curate data about former trainees. For academics, that information should include their grant, publication and teaching records as well as outreach and mentoring activities, and the criteria that academic hiring committees use to evaluate candidates. For non-academics, data should include the training, internships and work experience that led to employment.

The US National Postdoctoral Association in Washington DC is collecting data about postdocs' career paths as well as about institutional compensation, benefits and career services (see page 122). The NSF this winter launched an Early Career Doctorates survey that will gather in-depth information about postdocs and others who have earned their doctorates within the past ten years. The NIH's newly created Division of Biomedical Research Workforce in Washington DC may become the ideal organization for gathering and disseminating data about PhD graduates in the biomedical sciences.

These data will help funding agencies to craft policies that encourage institutions to give people with PhDs options for careers in non-faculty positions. One way to do this is to provide diverse education and training to PhD trainees so they can pursue careers in industry, consulting, entrepreneurship, science policy, writing and editing, administration or management. Federal funding agencies must find ways to ease pressure on trainees to work day and night for publications and grants, and instead foster ways to gain work experience and explore non-academic career paths while still in training.

These long-term solutions will not help current graduate students and postdocs, who must seek professional-development counselling, develop transferable skills and network within and outside academia. Ultimately, the careers of hundreds of thousands of future PhD holders depend on access to career information that will help to better match supply with demand.