Economists were surprised again today, this time over industrial production.

The Econoday Consensus Estimate was for a flat reading of 0.0 percent but production fell 0.3 percent.

Moreover, the Fed revised December production lower, to +0.6 percent from +0.8 percent. Effectively, the consensus was off by 0.5 percent.

Blame the Weather

A swing in utility output skewed what is, however, no better than a modest industrial production report for January. Industrial production, reflecting a 5.7 percent weather-related drop for utilities, fell 0.3 percent which is below Econoday’s no-change consensus.

But the real disappointment in the report is the manufacturing component which could muster no better than a consensus gain of 0.2 percent. This reading hasn’t been able to build any momentum to speak of and was held down in January by a sharp 2.9 percent monthly downswing in vehicles. Excluding motor vehicles, manufacturing volumes rose 0.5 percent which is really the highlight of today’s report. Also a highlight though is mining which is the report’s third and smallest component. Mining continues to show new life with a very sharp 2.8 percent jump in January.

Overall capacity utilization reflects the general softness of the industrial sector, at 75.3 percent for a 3 tenths decline in the month and 4.6 percentage points below its long run average. Manufacturing utilization is likewise soft at 75.1 percent.

The industrial economy, held down by weak global demand, has been running below average the past 2-1/2 years, when energy prices first collapsed in mid-2014. But advance indicators, including this morning’s Empire State report, are almost uniformly pointing to a rebound ahead, a rebound however that has yet to appear in the government’s definitive data.

Easy to Predict

This was too easy to predict. On January 18, 2017 I noted Industrial Production Jumps Due to Weather: Good News Stops There.

Due to the jump in December industrial production, The FRBNY upped its first quarter GDP Nowcast estimate up from 2.1% to 2.7% on January 20.

On January 24, in Formulas Don’t Think: Investigating Weather-Related GDP, I commented: “Temperatures have been warmer than normal for most of January. Temperatures in Chicago hit a remarkable 60 degrees last Saturday.

Here are the weather-related charts I posted.

January Chicago Temperatures – Actual Highs vs. Average Highs


December Chicago Temperatures – Actual Highs vs. Average Highs


Temperatures from Accuweather.

Unthinking Formulas

I calculate a net of -67 degrees for the month of December. That’s an average of only -2.16 degrees per day.

I calculate a net value of +132 [for the month of January]. That’s an above average performance of +4.26 degrees per day. It would have been much greater except for a nasty five-day stretch from January 4th through January 8th.

The problem with unthinking formulas is they extrapolate going forward. December capacity utilization affected projections for the entire quarter going forward.

If we attribute at least some of the capacity utilization increase in December to unseasonably cold weather, then that portion and more will be taken back in January.

Other data can impact the results as well, but all things being equal, the New York Fed capacity utilization projection for 1st quarter GDP will be unwound.

Emphasis added.

Here we are. Economists were surprised again.

GDPNow Forecast Dives

Check today’s update to the GDPNow Forecast.


Latest forecast: 2.2 percent — February 15, 2017

The GDPNow model forecast for real GDP growth (seasonally adjusted annual rate) in the first quarter of 2017 is 2.2 percent on February 15, down from 2.7 percent on February 9. The forecast for first-quarter real consumer spending growth declined from 3.1 percent to 2.8 percent after the retail sales report from the U.S. Census Bureau and the Consumer Price Index report from the U.S. Bureau of Labor Statistics were released this morning. The forecast for the contribution of inventory investment to growth fell from -0.20 percentage points to -0.39 percentage points after the industrial production report from the Federal Reserve Board of Governors and the Business Inventories report from the U.S. Census Bureau were released this morning.

Seasonal Adjustment Silliness

The models took unusually cold December weather for a genuine spike in industrial activity. The models then projected that spike forward for the entire first quarter!

The New York Fed Nowcast comes out on Friday. Don’t be surprised by a serious dive.

By the way, please note this Econoday comment on industrial production: “This reading hasn’t been able to build any momentum to speak of and was held down in January by a sharp 2.9 percent monthly downswing in vehicles.

Auto inventories are building, the manufacturers cut production, yet according to CPI data today, auto prices rose a whopping 0.9 percent in January.

This was one of the screwiest CPI reports that I can recall. For discussion, please see CPI Jumps Most Since February 2013 on Energy: Did Gasoline Prices Really Rise 7.8% in January?

Mike “Mish” Shedlock