As Business Insider puts it, the National Association of Realtors just made the mother of all data corrections. 

Basically, the National Association of Realtors (NAR) retroactively adjusted its accounting of the number of home sales that took place in 2007 (down 11%), 2008 (down 16%), 2009 (down 15%) and 2010 (down 15%).  Put simply, NAR had thought and had broadcasted quite widely that home sales were much stronger over the past four years than they really were -- and bear in mind, even the inaccurate, inflated numbers were quite depressing!

Why did it happen?

• FSBO Sales Declined:  NAR estimates the home sales based on a variety of data sources.  One reason for the inaccuracies is that the proportional number of FSBO sales declined significantly over the past four years.  Back in 2000, 16% of home sales were by owner, but in 2010 only 9% of home sales were by owner.  NAR didn't realize that this ratio had adjusted, and thus projected total home sales assuming that plenty (16% +/-) of sales were still by owner. 
• Builder Sales Declined:  Just as many more owners used Realtors over the past several years to sell their homes (instead of going FSBO), many more builders also used Realtors to sell their new homes.  This change in proportions also lead to over-reporting of home sales by NAR -- they were assuming the same number of sales were taking place directly from builders.
• Marketing Properties Through Multiple MLS's:  As housing markets worsened, more and more owners marketed their properties through multiple MLS's to broaden exposure.  This sometimes results in home sales being reported through more than one MLS, again over-inflating NAR home sales data.

Does it matter nationally?
It sort of matters retroactively -- everybody needs to go back and think that the housing market was worse than they thought before, and make their past decisions differently.  Oh wait, that's somewhat difficult to do -- so the national impact is simply that we need to readjust our historical understanding of the pace of home sales from 2007-2010 and be a bit less dependent on NAR home sales estimates, as they are ultimately, just estimates.

Does it matter locally?
Since the national housing market (if one could be said to exist) doesn't impact us too much at all, these adjustments to national housing sales figures don't impact us that much either.  Of much more importance is the pace of sales (and median price of those sales) on a local level.

Could it happen locally?
NAR and yours truly make different attempts as far as reporting home sales.  NAR attempts to report on total home sales (MLS, by owner, by builder, etc) -- and thus there can be inaccuracies in their data.  My home sales reporting is based solely on data reported through the HRAR MLS.  Thus, I'm not making any guesses -- I'm just performing analysis and providing commentary on home sales reported through our local MLS.

NAR's adjustments are rather significant, so let me know if you any questions or concerns beyond those addressed above.