Gazing into the future of real estate valuation

February 7, 2011

by Mike Simonsen

9 comments

I don’t practice Santeria / I ain’t got no crystal ball / Well I had a million dollars but I / I’d spend it all…

- Santeria, Sublime

If you’ve followed Altos Research for any length of time, you’ll have noticed that we’re a bit pedantic about the orientation of housing market data. For years we’ve been talking about the power of the insights in the active market. Understanding real estate is not just about which transactions closed, it’s about how much the sellers are asking, changes in those  ask prices; it’s about inventory and demand metrics. The active market of homes for sale is rich with information and insight about future transaction prices, home values, mortgage defaults.

Today’s Real Estate Market Yields Tomorrow’s Home Prices

AltosEvaluate screen shot

The at-a-glance view of a mortgage portfolio or RMBS security in AltosEvaluate Forward Valuation Model.

We’ve noticed more and more lately that our institutional clients understand this active market informational wealth, and they keep asking us the same questions, “How much should we adjust our bids? How much are home values going to change in [random zip code]?” You only have to smack me in the face so many times before I wake up and realize the world has a need. That brings us to today.

Announcing AltosEvaluate: Forward Valuation Model

Today, Altos Research launches a technology we call Forward Valuation Modeling (FVM). Part of our AltosEvaluate Mortgage Analytics Suite, the FVM looks at everything thing we know about the current, local housing market–all the pricing and price change trends, all the supply and demand metrics, plus everything we know from years of tracking this data– and investigates how those characteristics have impacted local transaction prices previously. The FVM uses that insight to forecast price changes in each local market over the next 3, 6, 12 months.

In the capital markets, investors buy residential mortgage bonds or whole loan pools to trade or keep as investments. When you make these investments, you have to investigate every asset in the deal, maybe 1000 or 5000 loans in random neighborhoods around the country. You look at the loan, LTV, the borrower, FICO score, payment history, everything you can find to apply a value to that loan. AltosEvaluate FVM brings the ability to look at the local housing markets underlying every mortgage in the deal and make decisions based on the trend in housing values.

For example: You might realize that the 85% average LTV in the pool today is looking more like 95% LTV as the FVM illustrates how property prices are likely to fall in most of the markets over the next twelve months. These loans are MUCH more likely to default and you’re much less likely to get a return on your investment. Bid accordingly.

There’s so much to tell about this new technology, I’ve struggled to keep this post at a manageable length. If you’re interested in more, just contact us. Here’s a quick FAQ to wrap up. AltosEvaluate Mortgage Analytics Suite brochure is embedded below.

AltosEvaluate FVM FAQ:

What does the FVM do? The FVM forecasts the home price changes coming in local housing markets in the next 3, 6, and 12 months.

How does it work? The FVM platform is what’s known as a machine learning technology, a system where we have a vast amount of data, some of which is useful for forecasting the future, the system “trains” itself on the data to build models on how to process the current market. In the case of real estate, this is critical because you have to have local models (what’s interesting in the San Jose housing market is not useful in the Miami Beach condo market, for example). Altos Research tracks 20,000 zip codes, so the AltosEvaluate FVM literally trains tens of thousands of local forecasting models.

Who uses it? Mortgage and RMBS bond investors, companies with exposure to real estate and who need to understand where the local housing values are heading so they can adjust their bids, or make buying and selling decisions in their portfolio.

Is it an AVM for the future? No. AVMs (automated valuation models) take the recent closed transactions and estimate a price for a given property.  The FVM says, based on everything known about this local market, real estate prices are headed down (or up).

How do I get more info? Email our housing market analytics team for more info, white papers, etc.

Forward Valuation Model (FVM) Brochure

{ 8 comments }

Greg Fleischaker February 7, 2011 at 8:42 am

Very cool. Looks like this might pack more muscle than I need as a local real estate agent, but always curious as to what you guys are up to next! I may not need it, but might end up taking it out for a test drive anyway!

Ralph Bredahl February 15, 2011 at 4:00 pm

Exciting, I believe staying a head of the curve in technology can make or break a realtor.

ArrowCreek Homes February 16, 2011 at 2:01 pm

This deserves to be re-posted:

"Understanding real estate is not just about which transactions closed, it’s about how much the sellers are asking, changes in those ask prices; it’s about inventory and demand metrics"

Mike, we may never have a crystal ball but studying the above metrics puts you the closest.

"…what I really want to know my ba-by, what I really want to say…."

Sean Pearson June 10, 2011 at 7:45 pm

Sounds like a great tool to help investors and buyers make solid decisions.

Michael Shaun July 11, 2011 at 12:14 am

This will help and save the day.

men's suits March 12, 2013 at 11:42 pm

I think real estate yields tomorrow because institutional clients understand this active market information.

michelle March 15, 2013 at 12:58 am

Anyone involved in a real transaction can benefit from gaining a basic understanding of the different methods of real estate valuation.
http://listingpros.com/Content.aspx?pid=66158

pdu training courses March 18, 2013 at 1:19 am

Look steadily and intently, especially in admiration, surprise, your thought as well.

Comments on this entry are closed.

{ 1 trackback }

Additional comments powered by BackType

Previous post:

Next post: