Help on a matched paired analysis

[George Hatch]

"I've only used the GLA "standards" per county, community, area etc. and the experience of my mentors in these areas."​

aka "the list". In doing this you're not extracting the adjustments from your sales data. You're working off a predetermined list and assuming those factors will work. Most likely your supervisor was trained the same way so that's why they're training you to do it this way. As well, all the other adjustments they're using are usually also coming off their "list".

Usually the way this is explained to new appraisers is based off the idea that "this is what the lender will accept", which when you think about it is not an appropriate way to appraise a property.

If you think about what it is you're actually supposed to be doing, it should be entirely market-centric, not lender-centric. It should be based entirely on what you think the market participants will do in these situations, not upon "what the lender will accept". The definition of value upon which these appraisals are based refers to the actions of the buyers and sellers. It doesn't refer to the actions of the lenders.

As for "what the lenders will accept", that's still a valid concern but the appropriate way to address that concern is to sell to them the credibility of your analysis. You are the appraiser, not them. You should be showing them how a legitimate appraisal is done, not the other way around.

 

[George Hatch]

In terms of this request to support your GLA adjustments, one way is to use "matched pairs", except that in real life actual 'matched pairs" are relatively rare and you will never have an assignment where the data lines up so well that you can derive all of your adjustments off of actual matched pairs.

There are several alternative methodologies you can use, including some off-the-shelf statistical modeling programs. Or you can learn how to do regression analysis and build your own in Excel or one of the other spreadsheet programs. Or you can use a sensitivity analysis. In fact, depending on what appraisalware program you're using you may already have access to one of these tools that's built right into the software.

The point is that you should be drawing these factors out of your comparable sales, not just pulling them off a prefabricated "list" or taking wild guesses out of thin air. Appraising is about trying to more/less "know" to whatever degree that's possible, and to refrain from doing unnecessary "assume"

 

[glenn walker]

For a newbie we always showed them how to do sensitivity analysis which is easiest because the Fannie-Freddie forms are like a mini-excel spread sheet fot the beginner. Often an- easy way to establish a starting point is to get a cost handbook and determine what costs new are in your market and establish three different sets of cost new based on ages and overall quality of construction. Example: lets start with a 1,200 Sq.Ft. CA bungalow located in Felony Flats and its and the comparables are between 55 to 65 years old- Average construction . Cost new handbook shows $125.00 Per Sq.Ft. - So my starting point is Half of $125.00 or $62.00 per Sq.Ft. I now insert that into all my comps grids and see what happens. What do my net & gross adjustments look like ? Now I will move up and down in $10-$20-$30 per Sq.Ft. increments from the $62.00 -Say maybe up to $80.00 and back down to $40.00. What I am looking for is the number that brings my comparables adjusted values into the tightest range and the one that results in giving me the lowest net & gross adjustments. Not always true but for general newbie advice the lower net & gross adjustments are usually results in not over-inflating a value and less challenges by reviewers because the more similar your comparables are the less adjustments are warranted.

Now the key is you never want to get caught in a corner and NEVER SAY you started at half cost new because that's not considered proper methodology but frankly for some folks they need a starting point to filter out all the other non-sense . Your statement to the not so nice Lady will be something like ( The GLA adjustments where extracted from the market based on buyer and seller reaction to price differences paid for physical size differences in the subject property's immediate and or general market area and these amounts are reconciled to the subject Property so that they are neither greater or less than the market called for and would support. Translation: I did not use a mechanical sytem or a GLA adjustment that someone told me to use or one that I made up or my mentor gave me. ( This will make that not so nice Lady not be able to pin you in a corner-as far as matched pairs with the same or similar GLA tell her you have none and neither does She and if you did there would be no need to make any GLA adjustments .

To Whom It May Concern -How I make my adjustments :
The appraiser employs a market-based system that relies upon the results of matched pairs when available and a percentage-based system accompanied by sensitivity analysis to assure the amounts that are reconciled the market to the Subject property were neither greater or less than the market called for and would support.

 

[George Hatch]

It's almost easier to just hand them a sensitivity analysis worksheet and let them play around with it a little. The process for using multiple iterations to see which factors most effectively prompt for convergence is pretty intuitive.

 

[Terrel L. Shields]

Who knows!! She only said I need to do it on GLA only. The thing is, she took my mentor's analysis free and clear. I did the same thing and nothing I do is working for her.

Use an Excel sensitivity analysis - this came from this web site year ago, I don't recall who posted it. You'll have to unzip it since it doesn't attach as an xls file

 

[Terrel L. Shields]

George beat me to it

 

[George Hatch]

The zip files don't seem to be attaching. Here's a couple of links to different worksheets that I have in my Box.Net account. I wrote the first one. someone else wrote the 2nd one, which is larger and more automated. The 2nd link has 4 pages and more equations so it won't show up in the preview, but it's there. you can still download it.

Box | Login

app.box.com

Box | Login

app.box.com

I should mention that although sensitivity analysis (which is basically comparing these comps to each other to see what works) is simple to do and simple to explain it nevertheless has its limitations, just like any other analytical method. So it's not a silver bullet app; it's just another tool for your toolbox.

 

[Terrel L. Shields]

The attachment issue is a problem. Pictures as well as other attachments. I think it's on the site or one of MS's innumerable updates...

 

[glenn walker]

I suggest a newbie just use their Fannie form software because working the numbers in the Fannie GLA grids is essentially a mini-excel without the learning curve and most newbies are not familiar with excel and spread sheets and the same can be accomplished on a smaller scale using comps for software GLA - Grids say using GLA from 6 to 9 comparables. Yes its slower with manual insertion but gives teh newbie more time to get focused on what is happening as they raise and lower their GLA adjustments. Got to keep a newbie focused on the end game the spread sheet analysis and other stuff can come after he/she gets their license.

 

[Sadie]

It's almost easier to just hand them a sensitivity analysis worksheet and let them play around with it a little. The process for using multiple iterations to see which factors most effectively prompt for convergence is pretty intuitive.

I had my trainee use this handy dandy spread sheet for support in her work files. I also used it when I was submitting my files for approval of my supervisory endorsement. Thank you it is a great tool. I have it saved and can share it if your upload isn't working.