Shostack + Friends Blog Archive


Housing Bubble?

Kip Esquire discusses “Housing Bubble: The Non-Lessons of the Past:”

Today, we get some unhelpful noise from TCS Overlord James “Always Wrong” Glassman. (Remember “Dow 36,000”? The only thing dumber than the book was his half-hearted non-apology for it.)
Now he’s fanning the flames of “What, us worry?” for the housing market:

Since 1950, according to data gathered by Freddie Mac, which provides financing for mortgage lenders, U.S. home prices overall have never declined over the course of any year.

Of course, they never went up 20%, 30% or even 50% in a single year either.

But far more importantly, the period 1950-2005 is a strange selection. Why start there? Because Freddie Mac makes the data easily available? Why not start in say, 1900? Because 1900-2000 includes 2 world wars and the great depression, and I bet there are several years in there when US home house prices have declined.

Actually, whenever I see data selection bias like this, I assume that I’m being lied to.

The graph, by the way, is a sample from Tufte. It’s his choice for a classic ‘select your data to support your claim’ lie. For the full graph with notes, and a second graph which displays full context, see the end of Pitfalls of Data Analysis, or purchase the Visual Display of Quantitative Information from your favorite bookstore.