Real Estate Market Forecasting & the Mortgage Servicing Industry

August 11, 2009

by Scott Sambucci


At the risk of over-generalizing, today’s conversation with a consultant working with rather large mortgage-servicing company proved, well, just a tad bit frustrating.  Here’s how the conversation went (I’m paraphrasing):

Him: “Scott, thanks for the data sets that you provided to me.  I ran correlations to the Case-Shiller Home Price Index, and it turns out that your data is highly correlated as a leading indicator of the index.  I got particularly good results by using the % Price Decreased stat.”

[In Altos-speak, % Price Decreased measures the number of active properties on the market that have had a price reducation in the last 90 days.  Here’s a more detailed definition.]

Me: “Great -that’s the results that you were looking for isn’t it?”

Him: “Well, yes, except that the the rest of the management here said they only care about forecasts two years out, and your data is more effective at forecasting the market 3-6 months.”

Me: “Huh?”

Him: “Well, they say they don’t care about where the market is going in the next six months, just where it’s going to be in 2 years.  They have a bunch of models that they built in 2001 that they’re using and they’d like to continue using those because they’re two-year modeling forecasts.”

Me: “But those models didn’t work in 2006 to predict the market crash. If you’re only looking two years down the road, you’re going to be highly inaccurate and you’re going to miss the turns in the market between now and then.  There’s already been inflection points in several key markets like Phoenix and Sacramento that turned up this Spring but might be headed back down as the year progresses.  I get that it’s important to know more six months out if you’re refinancing a mortgage, but suppose a market is forecasted to be higher or at the same level as today, but deteriorates further between now and then.  [see the graphs below] You could end up refinancing a mortgage based on the two-year forecast only to have the borrowers be underwater again in 6 months. Or what about Las Vegas?  It’s not pretty there, but the rate of decline is definitely slowing down there over the past two months.  There’s no other data that can provide this sort of real-time visibility.”

Him: “Yeah, I know, but that’s what they want to use. I love your data and was pushing hard, but I’m just a consultant here and they’re paying me.”

Me: “Didn’t they hire you to review their modeling platform and analysis tools?”

Him: “Uh-huh.  Like I said, I love your data – I just can’t these guys to get away from their view on the market.”

Not all mortgage servicers are this narrow-minded about modeling the market.  Some don’t even do it at all which makes it much easier to ignore what’s going on out there.

Sacramento, Phoenix, and Las Vegas Median Prices since 2007

Sacramento, Phoenix, and Las Vegas Median Prices since 2007

Previous post:

Next post: