Some of Harvey spills into the data

Week Ending September 15 2017

Data Releases

Retail sales and industrial production data releases were so bad that markets wrote them off. The reports might have contained a downward signal for the outlook, but it was overwhelmed by some exogenous factor. Both headline retail sales and industrial production were the worst in years, on a monthly basis.

Another data highlight was the sustained (relatively) high level of Initial Claims for jobless benefits. Hurricane affected workers (all hourly workers in a populated region) drove rolls from 236k on August 26 to 298k on September 2, and still at 284k as of September 9. You have to go back to 2015 to see Initial claims this high, though presumably, this is a non-informative signal, and claims will fall back in a few months.

Interestingly, both gas stations and grocery stores sales rose smartly in the Retail Sales report, consistent with hurricane-induced panic buying. South Texas includes more than just metropolitan Houston. There are supply chain interlinkages that spread to the cities of the Texas Triangle and further across the US. This can explain both poor petrochemical and mining sub-indexes in the Industrial Production report, and perhaps the important Manufacturing Production index, which sank -0.3%, still only the worst number since May.

Some of the sub-indices in the industrial report rose in August, such as Durable Consumer Goods. In any case, it’s just one month, and monthly data are noisy. It’s telling that markets doubled down on their current take following both releases, with most major asset classes moving tiny steps in bullish directions on Friday. To spell out just what markets have been saying both this week and in general this summer:

The “Trump Bounce” means 4% NGDP growth, with a little upside risk on that. The initial jump in equity prices and bond yields following the election was enough to lead a 3.4% economy to an outlook of about 4%. Since April or May, stronger NGDP and NGDI reports have helped offset modest declines in yields and sideways equities (bearish signal). At present, the combination of hard NGDP data and live markets point to 4.1% NGDP growth, this week markets upheld that.

Market Moves

The S&P 500 rose overall for the week, up 1.6% Friday-to-Friday. US stocks are tending toward all-time highs, which should happen more often than not if expectations for aggregate spending are anchored in a narrow range.

The bond market was generally bullish for the week as well. The 5-year TIPS spread rose from 1.6% the week before, to 1.7% by Thursday, September 14. TIPS movements (basically a market CPI inflation forecast) can be both supply or demand driven, though our forecasting system says “higher means more NGDP”, this is a pretty reliable statistical signal. The 5-year Treasury yield, a good proxy for expected Fed funds in five or six years, rose from 1.61% to 1.7%, Thursday-to-Thursday, not a bad showing. The 2-year picked up 10 basis points too.

West Texas Intermediate is going for 5% more at the end of this week, than last week, copper futures were down, though that is an even noisier (international) signal than oil. The dollar was a bit stronger on our internal currency index that feeds the forecast models, as well as on headline DXY. In this way, the dollar and the copper price paid back the equity and bond markets, leaving the forecast rounding to 4.1%, even if the line is bowing downward on two digits precision.

NGDP Forecast

What does it all mean? Not a whole hell of a lot. The NGDP forecast is holding around 4.1%, which is fine with us, because this economy doesn’t feel like it’s going to show more than that. We’ll have a small acceleration in growth, from a mediocre growth rate to about 4% nominal, 2% to 2.5% real.  This is in line with what we’ve seen of late, even if the market composition of ‘heavy lifters’ has shifted relatively from the dollar and copper to yields, TIPS spreads and equities.

image_pdf
Share

Leave a Reply

Your email address will not be published. Required fields are marked *