Inventory cycles - a forecasting nightmare
US GDP has been dragged down by a sharper-than-expected inventory cycle, but this should reverse in the second half of the year.
One of my main jobs is to try and work out where we are in the economic cycle. I’m not too fussed about precisely nailing GDP forecasts and neither are financial markets, but a failure to anticipate major turning points would be embarrassing. More importantly, it would seriously undermine our macro-based investment process.
So what do I make of the latest US growth numbers and should we be worried?
I have to confess that the 1.2% growth rate fell woefully short of what I was expecting a year ago and represented a decent loss of momentum. Since, by definition, all recessions are preceded by a slowdown in GDP growth, a good kicking of our economic models has been required. However, they refuse to budge, so I think it would be unwise to extrapolate the current weakness.
Between 2010 and this year, US real GDP growth has oscillated between a 1-3% growth rate over the previous year. A return to the underlying 2% growth pace seems the most likely outcome and is the message from recent high frequency data and the Beige Book. There has been some underlying slowdown in US growth to 2% from above 3% at the start of 2015. But part of this reflects a sharp reduction in energy-related investment. With oil prices stabilising and rig counts beginning to recover, this drag should now be over. In addition, the inventory cycle and the impact of dollar strength on exports have made this slowdown appear even sharper. It is really hard to forecast fluctuations in inventory levels and they can often cause big swings in GDP, although the effects usually wash out after a few quarters.
Surprisingly, Q216 GDP showed a reduction in the level of inventories. This is unusual in the middle of an expansion phase. It is the change in inventories that matters for growth, so even if inventories continue to decline at the same pace, the negative impact on growth will not be repeated.
More likely, companies will increase production to prevent inventories from becoming too lean. This will add to growth in the second part of the year. July payrolls and in particular the recovery in manufacturing jobs suggests this recovery might already be underway. The main source of underlying growth remains the consumer, with support from housing and government spending. The latter two categories were surprisingly weak in Q216 and should also bounce back. Indeed, the initial estimates from the Atlanta Fed GDPNow forecast indicates growth of 3.7% in Q316, although this will be prone to heavy revision as more source data becomes available.
If the data remains favourable and financial conditions benign, the Fed will increasingly be thinking they need to resume their gradual rate normalisation path. August is expected to be a relatively quiet month for Fed speakers. The market is currently pricing in a low probability of a rate hike in September and still less than 50% chance of an increase by year end. Fed Chair Yellen is scheduled to speak at Jackson Hole at the end of the month. That might be a good opportunity to attempt to steer market expectations.
US risk assets have rightly in my view shrugged off the inventory cycle, but might be more worried about the threat of Fed hikes.