A matter of birth and death

Economic statistics go in and out of fashion like mullets and hoola-hoops. When I started in bond markets the obsession was with the monthly government borrowing number, the PSBR. This was at the time that the Conservative government was borrowing heavily in the middle of the last UK recession – the current numbers are just as high, but now the market doesn’t bat an eyelid. When Richard Woolnough started the market would eagerly await the publication of the shipping forecast, bringing news of the tides and winds, and the progress of the ships bearing spices and `baccy from the far Indies (actually to be fair it was the monthly Trade Deficit number). Nowadays, despite a fleeting interest in the CPI numbers, the focus of the markets is on the US employment data, known as the Non-Farm Payrolls, released on the first Friday of every month.

After four full years of positive monthly payrolls growth, 2008 has so far seen four negative prints. So far we’ve seen nothing like the -300,000 per month job loss prints that we saw in the tech recession of 2001, but it’s clear that we are seeing rising US unemployment. However, last Friday’s print at -20,000 was better than the markets had feared (-80,000 was forecast) and risky assets (equities, high yield, credit) rallied on the hope that things might be better in the economy than anticipated. BUT – here’s the problem. The reported number from the Bureau of Labour Statistics (BLS) is pretty much made up.

There’s an adjustment put in there to estimate new businesses that might have been set up (or indeed to have folded) since the BLS last did a proper survey. It’s called the birth/death model. In the case of April’s survey, the BLS added an additional 267,000 jobs to the US economy on the basis of their model – they implied that lots of jobs would have been created through new construction companies for example. In other words, without what they call their “birth/death” adjustment, the US economy would have lost a massive 287,000 jobs rather than the 20,000 reported loss. The BLS themselves realise that this guestimation is likely to be wrong, and we can expect subsequent revisions when better data is produced. Our economist, Steven Andrew, and I visited the Fed this time last year, and discussed the use of the birth/death model with them – the Fed official stated that people who look at the birth/death model in any detail are – to quote – “hobgoblins”. And perhaps we are, but I’m happy to take a bet that given the collapsing US housing market and stagnant commercial property market, new construction companies were not springing up throughout the US in April this year, and that come the revisions, things will look a bit bleaker than the statisticians implied.