Flying blind: no jobs report

The first Friday of October came and went without an Employment Situation report from the Bureau of Labor Statistics, one of the rare gaps of its kind in modern monetary history. With the government shut by a funding impasse in Congress, trading desks and Federal Reserve staff had to make do with noisy private figures.

The shutdown began on 1 October, after the Senate failed to pass a stopgap funding bill. Democrats are demanding an extension of enhanced health-insurance subsidies that expire at the end of the year, and Republicans want a clean extension first. The Labor Department's statistical staff were furloughed along with hundreds of thousands of other federal workers, and the department said it would not publish its data releases until funding is restored.

Why this report mattered more

The timing is poor. The last official report, for August, showed payrolls rising by only 22,000 and the unemployment rate edging up to 4.3 per cent. In September the bureau's preliminary benchmark revision indicated that the economy had created about 911,000 fewer jobs in the year to March than first reported. The Fed cut its policy rate by a quarter-point in September, to 4.00-4.25 per cent, citing the weakening labour market, and officials are divided over how much further to go. The September report was meant to tell them whether the slowdown was deepening.

The bureau's credibility was already under strain. In August the president dismissed its commissioner after a report that included large downward revisions, accusing the agency of manipulating the numbers. Economists defended the data, but a missing report now leaves a gap that partisans of every stripe can fill with their preferred story.

ADP fills the gap, badly

Without the official establishment survey, investors leaned heavily on ADP payroll reports, job postings data from Revelio Labs and high-frequency card spending. Private aggregators have serious sampling problems. ADP reflects payrolls at established corporate clients, so it undercounts small-business closures and public-sector cuts. The partial picture makes the labour market look more stable than it probably is.

Even so, the private data are not reassuring. ADP's report this week showed private employers cutting 32,000 jobs in September, and it revised August to a small decline. Postings data show hiring demand cooling steadily. If the picture from the private sources is flattering, the official one is likely to be weaker still once it appears.

How the official survey differs

The official payroll figure comes from a survey of about 120,000 businesses and government agencies covering a large share of all jobs, weighted to represent the whole economy, and it is benchmarked each year against tax records covering nearly every employer. The household survey that produces the unemployment rate is a separate sample of around 60,000 households. No private source replicates either. ADP covers only its own clients, postings measure intentions rather than hires, and card spending measures demand rather than employment.

That matters most at turning points. Private series built from existing clients tend to miss the effects of firms closing and new ones opening, which is exactly where a slowdown first bites. The official figures correct for that through statistical models and annual benchmarks. Without them, the market is reading a thermometer that is least accurate when the temperature changes.

The 2013 precedent

A shutdown in October 2013 also delayed the September jobs report, which eventually appeared more than two weeks late. The data, when they came, were unremarkable, and markets moved on quickly. The difference this time is that the Fed is in the middle of an easing cycle driven by labour-market weakness, so the missing data bear directly on a live policy decision. A delay of a few weeks now could mean the Fed's meeting at the end of October takes place without the September or October employment figures.

Paying up for protection

The lack of information pushed up the cost of hedging straight away. With no BLS numbers to anchor them, implied volatilities on interest-rate options jumped as traders priced in the risk of large revisions once the government reopens.

The mechanism is straightforward. A jobs report normally resolves a large share of the uncertainty about the Fed's next move in a single morning. When it is postponed, that uncertainty accumulates, and the eventual release, possibly combined with a second month of data, becomes a bigger event. Options on short-dated Treasury futures price that larger event. Market-makers also widen the prices at which they will trade, because they risk being caught on the wrong side of a surprise they cannot see coming.

What would change the picture

The strongest case against alarm is that the Fed has enough information to act. Weekly claims for unemployment benefits are compiled from state data, private surveys of purchasing managers include employment gauges, and the Fed's regional banks run their own business surveys. Officials also have their own contacts across industry. Those are real sources, but none carries the weight or precision of the payroll report, and a Fed that cuts on partial data risks being contradicted when the official figures return.

The signals to watch are the length of the shutdown, any statement from the Labor Department on rescheduled releases, whether the consumer price report due in the middle of the month is also delayed, and the tone of Fed officials' speeches. Without official jobs data, markets are trading on guesswork, and liquidity spreads in Treasury futures and corporate debt will probably stay wider until the BLS starts publishing again. If the shutdown lasts into late October, the Fed is likely to cut again on the strength of the private data, and the eventual official figures will probably show a labour market weaker than the stand-ins suggested.