Keystone Blog

This commentary focuses on the importance of an accurate appraisal whether it is for financing, buying or selling. I will tell you of one horror story that is not so uncommon and also provide a graph of market value that a friend of mine has completed, approximately every six months, using an AVM (Automated Valuation Model) based on assessment data.

Not so long ago a property owner called concerned about an appraisal she had done on her property. It was over $100,000 less than the mortgage amount. I reviewed the appraisal and saw no issues with it. I asked if I could see the previous appraisal. She said this was the only appraisal she ever had done. All previous property value was reliant on an AVM.

What a mess she is in. She will not be able to sell it until when and if the market catches up to what is owed as she has no other significant assets of which to transfer the balance to. How could the appraised value between the AVM and the local independent appraiser be so vastly different and put this property owner is such a burdensome and onerous situation?

What is an AVM? An AVM disguises itself as a property appraisal by providing an estimate of value determined by a computer program that uses a mathematical analysis, usually multiple regression analysis, combined with a database. The database if often tax assessment records of your property and the comparables used. In British Columbia, one of the most popular one is a private company that uses information from our provincial crown corporation BC Assessment records as its primary source of data.

There are some positives to an AVM. For the lender it is an additional valuation option and they do save time and money as they do not require a real human to inspect the property which results in increased lender profits. One of the claims is that they will also reduce fraud. However, this is easily disputed as systems can be easily manipulated and the data used could be incorrect.

The primary disadvantages of an AVM are:

• Unlike a physical inspection, a computer can’t tell you the current condition of the property, or even if the house is really there;
• If there are unique features of a property that might add or detract from market value, like does it overlook the river or the garbage dump;
• How long ago the property was assessed (The assessment notice that you received in January was based on a valuation date of July 01, 2010 – as of the writing of this blog, the valuation is already over 7 months old);
• What makes the comparables comparable (a computer might compare your property to one with similar square footage and sold three months ago and was only located a quarter of a mile away. However, that comparable could be flood-prone, fronting a highway, sold under duress and not determined yet, suffer from lack of maintenance, etc, etc., etc.)
• How long ago was the property and comparables inspected (when was the last time the assessment office had someone inspect the property, on the interior or even exterior – can you think of the last time you had an assessment officer through your property?);
• Whether the market is currently declining, stable or increasing (if the AVM is using transactional data from land registry, there is usually a lag time which can range anywhere from three to six months. Also, there could be concealed incentives);
• What qualifications, designations, experience and education the preparer of the value has.

Below is a graph of a property value completed by an AVM between December 2005 and November 2010. No changes have been made to the property and though I have not provided a graph of market conditions, I can tell you that there are times that when the market was actually decreasing, the AVM was indicating an increase in value. 

Right from the start there are obvious flaws with wide valuation swings. In one month, between December 05 and January 2006, the value increased $25,000 or almost 7%. In the next six months it decreased $32,000, or approximately 8%. The peak of the market, for the location of this property was actually in April 2009 and then decreased to the trough in September 2009 which is a decline of 6%. The AVM is indicating an increase of 4% over the same time period.