One of the more idiotic* things I kept seeing in the MSM last year is quotations from various "experts" that while it's possible that certain "frothy" housing markets will see a correction, home prices never decline (in nominal terms) year over year nationwide, and thus will not do so in the future. This was said with great certainty, as if past performance (though conveniently the "past performance" guideposts started after the Great Depression) is a guarantee of future results, even though the nationwide housing bubble of the past several years is absolutely unprecedented in American history.
Guess what? According to data from the National Association of Realtors [Excel file], the median price for a used home sold last June was $229,000; for March 2006 it was $218,000. The mean price for a used home sold last June was $275,000; for March 2006 it was $266,000. True, this isn't a year over year decline yet, but with inventories reaching record levels in formerly "hot" markets around the country (and up 39% nationwide, year over year), speculators heading for the exits (e.g., the D.C. area), and interest rates rising, it's hard to see where a sudden price spurt is going to come from.
I still remember reading in 1999 that "long-term" investors should hold their NASDAQ stocks, because while a small correction is possible, long-term holders can expect to achieve price appreciation, just at a lower rate than in the past. I was skeptical, but I heard the same message so many times from so many places that it affected my investment decisions, obviously for the worse. And that's when I learned never to listen to the conventional wisdom in the MSM about financial matters.
* Why idiotic? (a) there's no inherent economic reason why home price can't/won't fall nationwide; (b) they did in the U.S. in the 1930s; (c) they did regionally between 1989 and 1995; (d) they recently did, sharply, in Japan; and (e) the nationwide runup in prices is unprecedented, rendering past behavior of the market less useful in predicting future returns.