Its got to do with elasticiti and responce times. Basically Price and # of sales are both products of demand. If Demand shifts up you expect both to go up, if shifts down you expect both to go down. When prices fall and number of sales increase then you do not necessarily have a shift in demand, but rather the market adjusting to find the new optimum.

Well, year-on-year, prices and sales are both down. But yeah, the "but" you quoted is nonsense. We should expect volume to go up when prices go down. Unless things are really fucked. Which they are :).

it looks like '08 is narrowing the Y/Y gap

It's very difficult to take single samples and draw such strong conclusions, so while it may look that way, variance suggests that it could in fact swing either way. If you looks at the broader trends, it looks likely that less homes will be sold in 2008 than in 2007.

That is a difficult one to say for certain, because not all days of the week are created equal. You end up needing to compare counts for each day of the week and weighing them according to typical activity levels for the day of the week. If this is closing data I would expect weekends to be lower then weekdays, if this is offers then I would expect Saturday to be the high day of the week. Once you take that all into account I am not sure that leap day would have a significant impact. (ALSO afterall we do tend to treat each of the months as equal, and feb with leap day is still shorter then any other month.)

I am curious what a graph of rolling 28 day blocks would look like, would it smoth out some of the monthly variation or not?

You're right, thanks.

They don't mention taking into account the leap day, though, and 3.6% extra days is big on the scale of the change they are talking about (+2.9%).

that makes sense.

Not for at least a couple years! Case-Shiller (quality-adjusted home price index) futures predict an 18% decline over the next year alone in the US for residential home prices. It would be a terrible idea to buy an investment property now.