Sabri Oncu wrote: >Revisions are always an issue. But there are also the >"reliability" and "meaninfulness" issues. I leave realibility >aside, too many lies are being told in these days so one doesn't >know what to believe and what not to believe anymore. But take a >look at this University of Michigan Consumer Sentiment and >Expectation Indices for example: they are some monthly numbers >based on a survey of 500 to 600 individuals who are supposed to >represent the 280+ million American consumers. There two releases >of these indices every month: one preliminary, one final. Mind >you, the priliminary numbers are based on roughly 250 to 280 >individuals. How "meaningful" measures can these indices be of >the feelings of the American consumers? Yet, they can move >markets.
There's a NY Fed paper on the confidence numbers <http://www.ny.frb.org/rmaghome/econ_pol/698bram.htm> - the Michigan ones aren't very useful, esp the early readings, but the Conference Board's is ok. What's best are measures of current household finances & people's sense of the job market. Someone posted a news story here on the Reserve Bank of Australia's study of Australian confidence numbers. They're pretty worthless, apparently <http://www.rba.gov.au/PublicationsAndResearch/RDP/RDP2001-09.html>. Yes they move markets, but sometimes for as little as 2 minutes. Anything moves markets. Markets move on their own, and people select reasons after the fact. Doug
