about this blog
- Cody Willard is an anchor on the FOX Business Network. Willard is also the principal of an investment management company.
He was a long-time featured columnist for the Financial Times and TheStreet.com as well as a regular featured economist and stock picker on CNBC's ''Kudlow & Company."
most popular posts
-
- There are no viewed posts at this time.
SD
More grist for the Dept of Poorly Constructed Economic Data ... "After the Census Bureau releases the preliminary New Home sales estimate, they revise the estimate three times (over the next three months) as more data becomes available. Historically the revisions can be in either direction - higher or lower - but beginning in 2005 I noticed that most of the revisions were down. All of the revisions (between preliminary and final) were down in 2006 and for most of 2007. It appeared that the Census Bureau had a systemic error in estimating preliminary sales during periods of rapidly declining sales(...)" http://calculatedrisk.blogspot.com/2008/03/new-home-sales-revisions.html This is a governmental report that has a statistical confidence interval of over 10%. In other words, a 10% rise or decline in new home sales from one month to another isn't even a statistically significant change!
SD
As far as GDP there was an article in this week's Economist on how an overall stat like that is misrepresentative since population growth is built into it so we should really be looking at things per capita. (Per capita income apparently rose more quickly over the last 5 years in Japan than it did in the US.) Pollution and stuff like that (estimated annually at 10% of output in China) is also overlooked. I think the average revision from the 1st report to the last revision is something like between 0.5% and 1.0% on average.
SD
Interestingly private research groups have recently replicated 2 of the major US economic releases that the government had control over. There's the ADP report (computed by Macroeconomic Advisers and based upon data from ADP payroll services covering hundreds of thousands of firms and millions of employees) that proxies non-farm payrolls and is released 48 hours before it. With less media attention but a still comprehensive scope MasterCard SpendingPulse proxies retail sales. IIRC the report is based upon credit card sales that cover 20% of the country's retail spending. So we have checks on government reports in those areas. As far as the much maligned inflation reports published by the BLS, I guess a libertarian economist would say we have commodity prices and interest rate markets to look at as a check.