Most of us have seen comparisons between appraised values and the valuation models accessible via the web. I hadn’t seen an actual comparison to recent closed sales and these models, so thought it would be an interesting exercise to see how they stack up with properties that closed within the past couple of weeks in my local MLS.
The data that follows is a compilation of 21 recent sales in the Ann Arbor school district that closed in mid to late December 2013. These houses were exposed through the local MLS and picked up by a number of different valuation models. Since these houses were exposed on the open market, my thinking was that these models would be fairly accurate, but nothing seems to be further from the truth! The value estimates are really all over the board as far as too high, or too low, with only a few exceptions.
The sales above are sorted by sales price. To help you read this chart, the layout is as follows: zip code and some comments about the property. The next column is the actual sales price, followed by a valuation model, followed by the percent difference between valuation model and sales price. For example, on the first sale, the sales price was $80,000 and the first estimate was $104,654 or 30.82% over-estimated compared to the sales price. There are a total of four different models used as well as the State Equalized Value from the tax roll (multiplied times two as is practice in Michigan).
In general, the lower priced sales were over-estimated and the higher priced sales were under-estimated, but even standard subdivision houses that were exposed through the MLS had discrepancies in the variance. This was unexpected as they were on the market, in homogeneous areas, and not in need of work, so why did the models fail?
The models fail in my opinion, because the market is simply not perfect and can’t easily be measured by pure statistics. Value really relates to a specific market, buyer preferences, condition (can’t be measured by computer model) and locational nuances.
Instead of either overpricing a property based on potentially erroneous information in these models, or leaving money on the table by underpricing; hire a local expert to provide you an analysis of the market and how your property fits within it.
Comments(2)