On 1 October, lawmakers in Congress, bitterly divided along partisan lines, failed to agree on a new budget. The US government closed. Roughly 800,000 civil servants, including thousands of scientists, were ordered to stay at home. Even accessing their work e-mail would constitute a federal crime, they were told. Now entering its second week, the shutdown is showing few signs of abating.

Non-government scientists must be imagining that this nightmare will pass. This is just a Washington DC thing, right?

It is true that the effects are most pernicious in greater Washington, where commuter buses and trains have plenty of empty seats. At the National Science Foundation (NSF) in Arlington, Virginia, 98.5% of the agency’s employees were sent home. But at the National Institutes of Health in Bethesda, Maryland, about one-quarter of the agency’s 19,000 employees have kept working, keeping mice fed and cell lines growing. And 45% of workers at the National Oceanic and Atmospheric Administration in Washington DC have kept at it, largely because data collected by the National Weather Service are so crucial.

There is a veneer of continuity. But it is an illusion. Clinical trials will not begin. Grant applications will not be evaluated. Even grants that have been awarded are in jeopardy if the cheque was not in the post.

And the consequences are not confined to Washington DC. Websites used routinely in research are not accessible. Conferences in which government scientists have vital roles are either being cancelled or going ahead as pale shadows of what they ought to have been.

The ripple effects will get worse as the government misses lump-sum payments to contractors. One casualty came on 4 October, when the National Radio Astronomy Observatory, based in Charlottesville, Virginia, closed its radio telescopes. This week, an NSF contractor is preparing to evacuate research stations in Antarctica, putting an entire summer season of research under threat. By the end of the month, the operator of NASA’s most famous observatory — the Hubble Space Telescope — will nearly have run out of money (see go.nature.com/smgwr1).

The public data collected at these facilities are used widely. Even scientists at financially secure institutions will soon find themselves missing a key tool or piece of data. In an interconnected world that relies on global collaboration, foreign scientists are not safe either.

Part of the frustration with the fiscal crisis is how manufactured it is. Republicans in the House of Representatives want to extract concessions on US President Barack Obama’s signature health-care law. Obama and the Democrat majority in the US Senate say that they will not be held hostage to these demands.

There are signs that the stand-off will persist until 17 October, when a new forcing factor would kick in: the government exceeding the amount of money that it is legally allowed to borrow. The economic consequences of not raising the debt limit are expected to be immediate and catastrophic, and so the warring parties in Congress might finally be forced to compromise.

That would be welcome. The damage being done to science — the slow business of meticulous data gathering — is not as immediately apparent as in other arenas. But it is insidious. A missed moment in a data campaign may not reveal its importance until much later. A talented scientist, fed up with budget vagaries, might seek greener pastures. And an experiment not performed might seem to be no worse than an unasked question. For these reasons, we must all ask the US Congress: why are you doing this?