Manufactures, from Victor Niederhoffer
The % of manufactures that saw increasing orders went from 53 % to 54.8 % this month, and it caused a break in the round and a 1% up move. Let's say there are 1000 or n manufactures who report in the survey. The standard error of a proportion is 1/2 divided by the square root of n, i.e. 1/60. Thus the actual proportion is less than 1 standard error away from expectation, a 35% shot by randomness to say nothing for the quantum increases in randomness caused by faulty seasonal adjustments. When you add in that manufacturing these days represents 10 or 20% of the economy, it's pretty iffy all around. That's what makes the markets run.
Steve Ellison writes:
If we are generous and estimate "more than 300" as 399 respondents, the margin of error for a 54.8% result is 4.9%, if Manta's listing of 45,000 US manufacturing companies is a rough approximation of the population.
Victor Niederhoffer replies:
As Sholem Alechem would say, "we are both right".
Anton Johnson adds:
An off topic anecdote. For those don't who use Gmail, they mine email text to provide targeted ads. From time to time I get a chuckle at what AdWords elicits for me. Today, from Vic's "sholem alechem" comment, the algo determined that soon I will be travelling to Israel and require lodging in Tel Aviv.