How to read the US economic growth data with scepticism

Stay informed with free updates

Simply sign up to the Global Economy myFT Digest — delivered directly to your inbox.

Financial markets tend to approach US economic growth data releases with a simple formula. if growth > forecast that means buy, but if growth < forecast, then sell. Simply reverse the formula for bonds.

Indeed, America’s 2.8 per cent annualised growth in the second quarter was instantly compared to economists’ 2 per cent forecast, with predictable results. It’s a handy heuristic. But it’s far too simplistic.

Here are some steps to help put “higher than expected” numbers into context

1) Work out what is driving it: Some basic digging shows the 2.8 per cent growth includes chunky contributions from: healthcare services (0.45 percentage point), inventories (0.82 percentage point), and government spending (0.53 percentage point).

That should take some of the gloss away. FTAV has written about how high healthcare spending is perhaps not the best indicator of a healthy economy. Inventories tend to be volatile, and sometimes reflect weakening demand. Lastly, government spending is perhaps not the best measure of underlying growth.

2) Understand the timeframe: Economic data releases come with a lag. Thursday’s data is for the US economy from April to June. But it is fair to say that since at least May, momentum indicators of the US economy have started to look weaker (delinquencies, cooling job market). Where the economy is headed is what matters:

3) Put it in historic context: Things look a bit less spectacular when taking the data back to 2010. “Real final sales to private domestic purchasers” — a broad measure of US economic demand, which excludes trade, inventories, and government spending — was the same as the quarter before.

4) Any one-offs? Capital Economics noted “a 50 per cent annualised surge in transportation equipment, thanks to gains in both motor vehicle and aircraft investment”. Those gains mostly reflected rebounds from disruption in the first quarter:

They are unlikely to be repeated, particularly with the durable goods orders data released today . . . showing a slump in aircraft orders and lower motor vehicle orders.

Right. You’ll probably say FTAV is just cherry-picking. And, anything looks good when you exclude what you don’t like. That’s fair.

But, that does not negate the findings: Forward economic indicators (including rising unemployment) don’t look great. Some elements of the economy — which arguably do not really reflect underlying growth — seem to be supporting the numbers.

This is all before one even factors in the prospect of revisions. Thursday’s data is the advance estimate. The BEA itself says that the second reading of the GDP rate on average climbs or falls 0.5 percentage point, and 0.2 percentage points in the third reading.

On average, the first flash numbers tend to be too pessimistic. But it is probably better, in the long run, to cherry-pick data points that go against your inclinations, than to use basic heuristics. Healthy scepticism keeps your feet on the ground.