patrick.net
An Antidote to Corporate Media
1,295,308 comments by 15,448 users - WookieMan online now
Why do home appraisers not use regression statistics to determine home values?
invite response
2012 May 6, 7:49am 5,610 views 8 comments
Comments 1 - 8 of 8 Search these comments
ELC
2012 May 9, 12:25am
According to the Dept. of Revenue, "Fair market value or true value is the amount that a willing and unobligated buyer is willing to pay a willing and unobligated seller."
What gives an appraiser or even a lender the right to get in the way of that axiom!
Keep movin', movin', movin',
Though they're disapprovin',
Keep them doggies movin' Rawhide!
Don't try to understand 'em,
Just rope and throw and grab 'em,
Matt McKissock developed and marketed a multiple regression analysis software package for appraisers several years back. McKissock is one of the largest appraiser educators in the US. Matt is a chemical engineering graduate of Carnegie Mellon, one of the brightest people I've ever met. I took a couple of courses in multiple regression analysis. The software appears to work well, although I'm not sure if it has been widely adopted. Multiple regression analysis was made possible in everyday problem solving by the evolution of the superfast computers we take for granted today. Not so song ago, a simple query using multiple parameters would have taken hours to process. Multiple regression analysis is used to calculate spacecraft trajectories (literally rocket science!) and to solve complex scientific and engineering problems, so I don't see why it couldn't work in appraisal. Several years back, there was fear that such programs operated by clerks sitting in remote offices would displace appraisers.
If the purchase was for cash and no third parties were involved (lender, title insurer, mortgage insurer, etc.) there would be no need for an appraiser.
Unfortunately, subjectivity will always enter into a home purchase. Example, take 2 condos identical in all respects: size, age, furnishings, land occupied, etc. located opposite each other: one faces an almost vertical hillside, the other with a panoramic ocean view. Do you really think they are equal in value?
Homeboy
2012 May 9, 3:50am
This process would have restrained the housing bubble by eliminating skewed appraisals from opinionated, self-interested individuals.
You are assuming they would have wanted to restrain the housing bubble. That is an incorrect assumption. They don't want accurate appraisals. They want appraisals that will make the deal happen. If they could start another housing bubble right now, they would not hesitate to do so.
APOCALYPSEFUCK is Tony Manero says
The Realtor® and mortgage broker tell the appraiser what the appraisal needs to say to get the mark lined up for a note.
Mortgage brokers are not allowed to speak to the appraisers. I had the appraiser ask me, the homeowner - sorry, piemann, houseowner - sorry, Patrick, fool, what the loan balance was. Code word question for, 'what number do I need to add 20% to make this come in.' Sure enough, it came in.
I always thought that appraisals may be more accurate if appraisers didn't have any of the offer/deal price or loan balance info., since they so often seem to come in right around there, and not entirely by coincidence, I suspect. But pietmann introduces an opposite idea, that if the parties have agreed to a price, the appraiser ought to run with it - but in that case, appraisers would only be for... refis?
Currently I have one in escrow for $445k. The 1st appraiser calculates the value to be $447k. ($2k higher!). The lender is not happy. They order a second appraisal, at the expense of the buyer. Sorry for that. The second comes in at $448k. I am not making this up.
Is this also a case where the buyer wants to pay the smallest down payment possible for the loan type in question... I mean if the buyer had put a few $k more down would the bank still order more appraisals...?
Here we already have 3 variables
- age
- sqft (living space)
- type of ownershipand I have not even started..
I could easily come up with 300 parameters, most of them are not quantifiable.
My experience has been that humans are very bad at processing multivariate data, whereas computers do rather well. The other piece of experience I have is that humans usually think they are much better at analyzing multivariate data than computers.
This would generally require appraisers to have brains, or at least have an interest in an accurate home appraisal. With today’s technology and a simple knowledge of a statistical package, anyone can compare 15 or 20 variables to get a more scientific measurement of local home values. It is very easy, once the program is set up; all the appraiser would need to do is input the home’s specifications into the program to get a more objective home estimate. This process would have restrained the housing bubble by eliminating skewed appraisals from opinionated, self-interested individuals.
The statistical data base should be gathered from previously sold home prices and their specifications over the previous 5 years. Once a sale price goes beyond 5 years its data in the statistical base would be eliminated. A longer look-back period would result in better long-run price stability. Currently the appraiser has a 5 home look-back, which is equivalent to about 2 days, and is not based on the sale price but instead home estimates. Then once a home is sold its estimated home value will no longer resemble the data, its sale price will become part of the data. This would create a revolving, up-to-date estimate of home values based on current trends.
There may still be some opinion bias based on the estimated value and quality of updates and specific indicators/features. But compared to the complete estimation process currently used, this bias will be greatly limited through the new structured data entry process. No longer will the homes magically begin to escalate in a snowball effect due to previous high home valuations, as they did in the bubble years. This adjustment in procedure would help reduce 60-80% of the subjectivity in estimating home values, which is good for price accuracy.
Better accuracy in home pricing would naturally result in more price stability. The only problem is that this stability may not be where current homes are valued…. The results may prove that homes are overvalued in many local markets. This may be the reason why this style of home estimation has not already been implemented. It would be too good. It too accurately shows what home values should be which would expose the fact that our biggest banks are essentially bankrupt. Again, when the people who are self-interested are in on the decision making process, it is fairly improbable that they, the bankers, are going to support or require a program that would essentially put themselves out of business!
#housing