A shutdown would leave the Fed flying blind

Congress is heading for a 1 October funding deadline with no compromise in sight, and the Federal Reserve is bracing for the consequences. A government shutdown would stop the collection and publication of official economic data at once, leaving policymakers without their instruments just when they need them.

The impasse is about health care more than spending levels. The House passed a stopgap bill to keep the government open until late November, but in the Senate, where it needs Democratic votes, Democrats are demanding an extension of the enhanced subsidies for insurance bought on the Affordable Care Act exchanges, which expire at the end of the year. Republicans want a clean extension first and talks later. The White House budget office has told agencies to prepare not only for furloughs but for permanent staff cuts, which hardens positions on both sides.

When the statisticians go home

If federal funding lapses, the Bureau of Labor Statistics, the Census Bureau and the Bureau of Economic Analysis are barred by law from working. Survey collection stops and data processing shuts down. Scheduled releases, including the September payrolls and consumer price reports, are postponed indefinitely. Central bankers in the middle of an easing cycle would lose their main gauges.

The calendar makes the timing awkward. The September employment report is due on Friday 3 October and the consumer price index for September in mid-month, both ahead of the Fed's meeting on 28 and 29 October. Retail sales, trade figures and the first estimate of third-quarter output would follow them into limbo. The Fed has just begun cutting rates on the strength of a weakening labour market, and it would have to decide on the next cut without the report that tells it whether that weakening has continued.

This has happened before, partly

The last two shutdowns offer a guide. In October 2013 the government closed for sixteen days, and the September jobs report, due on 4 October, was published only on 22 October, with price data similarly delayed. The longest shutdown, which ran for thirty-five days over the new year of 2018 and 2019, was less damaging for statistics than its length suggests: the Labor Department had already received its full-year funding, so the Bureau of Labor Statistics kept publishing, while the Commerce Department's agencies went dark and output figures were delayed for weeks.

This time the Labor Department has no such cover, so the core labour and inflation data would stop together. The delay compounds after reopening. Surveys that were not collected during the closure cannot always be recreated, and statisticians have to decide whether to publish figures based on thinner samples or skip a month altogether. A short closure mostly delays numbers; a long one can leave permanent holes in the record.

What still gets published

Not everything stops. The Fed funds itself from the income on its securities rather than through congressional appropriations, so it will keep operating, meeting and publishing its own data, including industrial production and the Beige Book survey of regional business conditions. The Treasury will keep auctioning debt and paying interest, since those operations are not affected by a lapse in appropriations. States will continue to process unemployment claims, even if the federal compilation of those figures is disrupted.

Private data, noisy data

Without official figures, markets and policymakers would have to lean on private high-frequency data such as payroll processors' figures, credit card spending and satellite imagery. These have shorter histories and patchier coverage than the official series, and they are not seasonally adjusted with the same care.

The weaknesses are specific. A payroll processor sees only its own clients, which tend to be larger and more stable firms than the economy as a whole. Online job postings measure employers' intentions, not hires. Card-spending data capture purchases but not prices, so they cannot separate stronger demand from higher costs, which is the question the Fed most needs answered with tariffs pushing up goods prices. None of these sources has a long enough history to show how it behaves at a turning point in the cycle.

The cost of flying blind

The economic damage of a short shutdown is usually modest. The Congressional Budget Office estimated that the 2018-19 closure reduced output by about $11 billion, of which about $3 billion was never recovered, as delayed federal spending and back pay for furloughed workers made up most of the loss. The greater cost this time lies in the decisions made without information, and in the threat of permanent job cuts that did not feature in past shutdowns.

For markets the consequence is wider uncertainty around each policy meeting. Traders would price the October decision on the Fed's own commentary and the private indicators, and then face a burst of delayed official data once the government reopens, possibly contradicting the view they had formed. Short-dated Treasuries and interest-rate options would carry more risk than usual, and any revisions to the labour data would land all at once.

The optimistic view is that shutdowns end, and that the Fed has enough information from the summer's weak jobs data to justify another cut regardless. That is true for a closure of a week or two. A longer one would leave the committee deciding in December with its main labour and inflation series weeks out of date. A shutdown would leave a data-dependent Fed without its data, and until the agencies reopen, Treasury prices would be set largely by guesswork built on noisy private indicators.