Scientists in the United States can find plenty of good news as they page through President Barack Obama's 2013 budget proposal. Despite substantial cuts elsewhere — and fierce pressure from Republicans to cut more — Obama called for healthy overall increases in both fundamental research and science education (see page 283).

But the good news, of course, is tempered by reality. Obama's budget document is one long struggle to balance two contradictory goals: to stimulate the lagging US economy and to curb the annual budget deficit, which is more than US$1 trillion. Science and science education are widely viewed as helping with the first, and will doubtless continue to be seen as such no matter who wins November's presidential election. The idea that science is a driver of prosperity is one of the few things on which the United States' bitterly divided political parties still agree. But the science funding agencies themselves are by no means immune to the second goal. The harder the cuts bite, the more those agencies will have to streamline their operations and merge or terminate programmes.

Researchers, societies and funding agencies can no longer afford to be purely reactive.

This week's budget proposal, which contains many references to “tough choices”, shows that this process is already well under way. The Department of Energy (DOE), for example, wants to discontinue funding of several dozen projects that have not met their research milestones, or that seem otherwise unpromising. The National Science Foundation (NSF) is likewise cutting back on some $66 million in lower-priority education, outreach and research programmes. The National Institutes of Health (NIH) has been ordered to pursue “new grant management policies” to increase the number of new grants by 7%. And NASA is being obliged to make drastic cuts to its Mars exploration programme so as to finish building its flagship James Webb Space Telescope.

Conceivably, this process could get even more drastic. Last month, Obama asked Congress to give him the authority to consolidate and streamline agencies on his own initiative — and suggested that one early application would be to transfer the National Oceanic and Atmospheric Administration from the Department of Commerce to the Department of the Interior. If Congress were to give Obama that power, it is possible to imagine him — or some future Republican president — sending all of the NSF's science-education programmes to the Department of Education, or merging the DOE's particle and nuclear physics research into the NSF, under the guise of making management of science more efficient.

White House officials insist that no one in the administration is even contemplating such a wholesale restructuring. But the arithmetic of the deficit is unavoidable. Individual researchers, scientific societies and science funding agencies can no longer afford to be purely reactive, responding to each cut as it comes along. They need to be part of the debate, thinking systematically about how programmes and even whole agencies could be restructured to make them more efficient at using the scarce funds available, and more effective at promoting the best science.

To do that, and to address the increasing demands from politicians and voters for evidence that fundamental research is useful, scientists must also find better ways to measure the effectiveness of the nation's investments in science. The usual technique is to insist that principal investigators produce more and more reports, which tends to be a waste of everyone's time. A consortium of six universities called Star Metrics, launched in 2010 and headquartered at the NIH, has shown that it is possible to do better by using natural language processing and other tools to mine the data and reports that the agencies already collect. But even that is just a beginning. Researchers and research institutions need to help to devise still better measures — because if they don't do it themselves, politicians and others who know much less about science may very well do it for them. And who knows where that would end.