Correlation, Causality and Common sense

WSJ looks at an updated metric of ‘low risk housing markets’.

Yeah, yeah, there we are… just as we have been for years at this point.  What makes this year’s iteration noteworthy is the WSJ’s summary for why Pittsburgh’s ranking is so low. Verbatim from the WSJ:  “Pittsburgh, where the natural gas industry is creating jobs and wealth“.

I bet the Marcellus Shale Coaltion makes a press release out of that.

So here is the deal.  I think the source of this metric, the PMI group, has been regularly ranking us as the single safest (i.e. least risky) real estate market for years.  Just one cite, but we were the single safest real estate market back in 2007.  I am pretty sure few were talking about, or noticing, any big boom in natural gas related jobs in the region in early 2007.  If that does not convince you, then lets go back further…  2004?  Ditto.  How about more recently?  2008: Check.   I am pretty sure with a a few minutes I could find the same story repeated each and every year going back some time.

So we have been the least riskiest market long before Marcellus Shale was rediscovered, and we have not budged much from the top ranking in recent years.   So the reason to gratuitously say natural gas is part of this year’s ranking is what exactly?  Maybe the local economic impact of shale has been positive… maybe not…  but the implied causality with mortgage and real estate risk is completely non-sensical as described.  Any undergraduate completing an introductory statistics class would hopefully not use such logic in any context especially in light of the observed history.

Thus the state of quantitative reasoning these days.  It’s just that it’s the WSJ no less in this case.  If we describe this so badly, it kind of explains how we missed all the problems with those slightly more complex credit default swaps.

Leave a Reply




You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>