Industrial Production
By Spencer
Industrial production only increased 0.1% in October, from a previosly estimated 98.5 to 98.6.
But I suspect this number was biased downward and will be revised higher.
As the chart shows, this changes the impression the previous reports had given that this was a normal to strong recovery in industrial output to one that it is a weak recovery.
A primary reason industrial production appears so weak is that October auto and light truck output fell to 6.83 million vehicles as compared to 7.15 million in September and a cyclical low of 4.05 million in June. However, as the production of all items excluding autos and high technology was unchanged in October the weakness appears to be widespread.
However, I am suspicious that this report overstates the weakness and will be revised up as other information is included and revised data is reported.
Much of the initial estimates of monthly industrial production data is based on electricity consumption data. However, the national average temperature days for October 2009 were extremely low at only 50.8 degrees Fahrenheit. In a quick check of my data base this is the second lowest October reading on record going back to 1921. The lowest was 49.4 degrees in 1925 and the only one I saw below 50 — the highest was 60.0 in 1963. The norm is around 55 degrees so the October temperature days was some 10% below normal. This strongly implies that the electricity usage would have been significantly below normal in October and consequently the industrial production data estimates are undoubtedly biased downward.
A little more info [see http://www.eia.doe.gov/cneaf/electricity/epm/fig1.jpg%5D which js-ket won’t embed because it “has the wrong format” That’s the EIA chart of US electrical usage by month Sep2008 thru Aug2009. Salient points:
1. Peak electrical usage is in August which is a warm weather month with lots of daylight
2. There is a secondary peak in January, a cold weather month with comparatively little daylight
3. September — a warm weather month — used substantially more electricity (about 5%) than March — a cold weather month. The daylight hours for the two months should be quite close.
My conclusion: Both temperature and daylight hours/cold weather affect electricity usage, but warm temperatures have a greater effect. — maybe 5% of total usage
Which month would you expect to have a great lighting effect, January or December? It should be December, because that month contains the shortest day of the year and you also get the bump in usage from Christmas lights displays. January, however, is colder, and since many residential users do use electric for heating you get larger load.
Spencer, this may seem like a disjointed comment, but do you support a financial transaction tax in order to raise revenue and curb the massive bulk of speculative trading in US financial markets?
I work for the Center for Economic and Policy Research in Washington D.C. and we are currently asking fellow economists to sign on to a letter in support of such a policy measure. A modest set of financial transaction taxes could raise a substantial amount of needed revenue while having little impact on trades that have a positive economic impact. To sign on, please follow the link below:
http://salsa.democracyinaction.org/o/967/t/9788/petition.jsp?petition_KEY=2167