LandSafe checking appraisals with an AVM

[Ross (CO)]

Brad, ..... you did speak directly to the concept of valuation where we generally do go head-to-head with the client's reviewer, and I almost always feel this is going to be one of their in-house staff and not someone outside and independant. Just past middle of your post in one sentence you say "highest possible value" which is this valuation concept in conflict. "Possible" can be considerably far away from "probable", and so long as we appraise toward "possible" then our work is compatible with their modeling, at least in thought. In those times when I have been asked the "why did you not consider....?" questions about market data used or not used, it clearly shows that the client would have preferred, and accepted, my use of a sale which was further away in distance or date-of-sale or some other key selection parameter. These moments truly do reveal that these clients are cruising the same data sources that I have at my disposal, and that they have a different way of "appraising" than I might. Then again, I drove to the subject, went inside, looked around the neighborhood and looked at all my comps up-close-and-personal........and they did not. I did what any typical person might do if they were srcutinizing that same house to BUY it, and at that moment our client does not think about buying the house (back), just tweaking those numbers to some maximal amount. It really does show, as well, how many of the intermediary clients with whom we might work really do not have the full focus toward service for the clients who have hired them ! That has always been a mystery in understanding the appraiser-client-client-investor relationship. If investors talked directly to us......oh, well.

 

[bradellis]

Ross-

Interesting post. So, OK. I'm an investor talking directly to you now. We also sell to other investors, so I know what they want as well- solid collateral. I repeat, solid collateral.

Now, does that mean that your value should not be up at the top of the market? No. If that is what the market is doing, then have at it. We've been seeing this all over the country for some time now; however prices now appear to be mitigating- not necessarily going down but not going up as fast. So, in many markets, it is time to go back to the mid points, perhaps.

I guess what I am saying here is that many of us have been out there and know what you do- how buyers and sellers react. You saw the subject. You saw the comp exteriors. You formed an opinion.

But when I develop a ton of other data that looks solid, it means I have to ask questions. Why did you go .75 miles away when there were model matches .1 miles away. Please explain. Why did you use all superior comps that all had negative net adjustments? Are you trying to push this up to the upper end of the range- or beyond? if so, why?

My primary point is that the reviewers will have lots of "other" data. As an appraiser, one must understand this and lead the reader thru the thought process. You know the data is out there. Explain why it is not appropriate. Or, better yet, use those models matches that do not have the "view" and show us how they must be adjusted upward to accomodate this deal. Knowing that you considered the data and the reasons why you did not use it is VERY helpful.

I remarked to one of my colleagues last week that I had just seen the "best" "bad" appraisal ever. What do I mean by that? The appraiser knew our requirements- if it is suburban- 2 comps at least- within a mile. But the subject was on acreage. So, she gave us 3 comps with acreage- no location adjustments were appropriate- gave us land sales demonstrating the difference between a sale on acreage and one on a small lot- then proceeded to also give us 2 sales on the same block, but on small lots. 50% upward adjustments for the site- but it did not bother me a bit. Why? She gave me everything I needed to determine if her value opinion was appropriate. So, she turned what might have been a very bad appraisal into what was a very good one.

Whatever it takes. The loan seller complained to me that the appraisal charge was high. Why, I asked? Well, she had to do all this other work. Well, I replied, it happened to make the difference between my accepting this loan or rejecting it.

Have a great day.

Brad Ellis, IFA, RAA

 

[Pamela Crowley (Florida)]

Thanks for the explanation, Brad. I feel a little better now knowing that someone that knows what they are looking at and appreciates those that do what they are supposed to do is overseeing some of this. I will happily admit when I didn't explain something well enough and fix it and, I'd much rather talk directly to the underwriter. For those, I will happily fix whatever is needed. Unreasonable or absurd requests with no real explanation of 'why' will and do meet a brick wall, though.

As I finished reading your post, I was wishing that ALL appraisers could/would read it.

Would you please address the difference in explanations wanted/required for the 2055 in comparison to the 1004? Since it is a limited form, I do tend to do the minimum of explanations even though everything necessary is in my workfile - unless there is no way to make anything clear without it.

 

[Steve Owen]

Brad and Ross, I followed your line of comments with great interest. I was especially impressed with your first post in this thread:

Guess I have to ask the appraiser why he/she chose the comps they did, no? You do NOT want to hear most of the answers.

roflmao.

Last time I checked, you were supposed to use the most similar, proximate, and recent comps.

It's good to know you are out there. It's sad to know that the low-ballers are still getting away with it. Maybe someday that will begin to change.

 

[Caterina Platt]

Interesting thread!

We shouldn't have problems with any of our appraisals going under the microscope if we are truly doing our jobs. I don't have a problem answering questions if I haven't properly given the reader/user enough info to understand my reasoning and conclusions. If an AVM is one of their tools, so be it. If the reviewing party is asking legitimate questions, and has read the report prior to asking for more info, GOD BLESS THEM! Unfortunately, we all know this isn't usually the case. Usually we are sending a $7/hour clerk to school in the appraisal process. Or worse yet, explaining why we have to disclose what is there. Sorry if it's a negative factor, I'm not leaving it out!

My biggest beef with Landsafe is how they developed their AVM. Through the data mining of appraisals via Lighthouse. They did not purchase this data, they have whittled the fees down and forced the appraisers over a barrel to use their system to develop such. Even this wouldn't be so bad, if I didn't have the strong suspicion that the AVM was being used as an appraisal in many cases. The only orders I ever got from Landsafe were oddball and difficult properties. If it were merely a tool to use along side my report and analyze my data, bring it on! I'm game.

My hope is that through the repeated use of their AVM as a review tool that it shatters their confidence in it's capabilities as a replacement for a live body. :roll: And off I go wearing my rose colored glasses to have another puff of 'dream on smoke'.

 

[Ross (CO)]

Brad, .... Thanks for your reply. While I may have commented to a desire to discuss a report's details and my analysis with the investor if necessary, I truly meant to tilt that more to the underwriter. I know full well that readers of our reports have heaps of data at their disposal and can match and compare one's value results with other scenarios. My posting above to you discussing "possible" vs. "probable" relates to one experience back in the winter months when I was pressed by a client to consider a $50K hike to a $225K value I had for a mountain property. After submitting report the client calls and leaves recorder message, and then faxes about 6 MLS pages from local database of 3 sales and 3 current listings and asked (demanded ! ) I re-grid my report "using these" and re-submit it, and "we think $275K is reasonable". I called back to say I had already considered 2 of their 3 sales and sale dates (long pre-Sept. 11) and further distance from subject were my prime reasons to summarily reject them for use in my report. Three current listings were considerably larger and inappropriate to use when I had recent sales of similarly sized homes in subject's immediate area ! This info sent to me was provided to my client by an area Realtor. This was in late-January, in middle of slowest time for sales of mountain property, and compounded by still-present aftermath of Sept. 11th. I expressed no need to re-work my entire analysis with their data.....suggested that they get my report reviewed or consider a second appraisal be performed. This time it should NOT be just a 2055 DRIVE-BY but a full appraisal. Let them go inside the house and see it ! I had previously asked this client on two other assignments to upgrade to interior/exterior but they were confident and assured that exterior-only satisfied their client and my addendum comments stated that quite clearly. There they sit, about 1800 miles away, and assuring me I am missing something. ( Duh, how 'bout an interior inspection ? That might be worthwhile. Most people buying a home will generally want to go inside for a little look, won't they ? ) After a 7-month run of about 6 orders each month, this was the proverbial straw.... Keeping a client for more than a year is a definite challenge.

 

[Pamela Crowley (Florida)]

Ross,

Don't feel too bad about that client. It was obviously a mis-match of Unethical Lender/Ethical Appraiser. The only lender/appraiser relationships that last is where they both have very similar ethics. Finding Ethical lenders is getting very difficult!!!!

 

[bradellis]

Pam,

Thanks for your kind words. I'll try to explain what I look for in an appraisal, but it is not the form itself that causes me any concern. It is the content.

First is the data. How similar are the comps? If you gave me, say, 3 models matches in the grid, I'd likely instatntly see what your train of thought is. Over and out- we are both done- no guesswork. But life is not usually that simple. How many assignments have this kind of data available? Not many.

So, if your comps are not so good (best available are not always good), I'd also look to the cost approach. Remember that the form does not dictate the scope of your work. You must do whatever it takes to competently complete the assignment. So, now you have done a 2055 and have lousy comps. Next question- would the cost approach shed any light on the value- if yes, then do it. If not, then perhaps extra market data- more, evern if they are less similar, comps.

Then you simply explain what is going on- not very good comps, but here is the true range of value for this subject. I'm at the upper end because it is has been recently remodeled, or it has a larger than typical lot in an area where that matters- or whatever.

Just lead me through your thought process. it does not take much explanation. But, it will convince me that you are not just blowing smoke.

Brad Ellis, IFA, RAA

 

[Austin]

Bradellis: I have discovered of late that if you use the proper sequence of adjustments, which are market derived, so called bad comps can reveal more information about the subject and market than conventional good comps. For example, if you have three comps that you are suspicious of, adjusting for physical features first with a properly executed size adjustment in the correct sequence and then observe the graph of the results. In this market, most of the time any remaining differences can be reconciled at this stage. In my experience, bad comps can be better than good comps if the sequence of adjustments is market derived. There are two ways to derive a price estimate: You can prove it with data showing what the subject is worth, or you can prove it with data showing what the subject is not worth. Or so it would seem to me. Wouldn’t you agree?

 

[bradellis]

Austin,

No, sorry. I still believe that good comps are better than bad comps. That is not to say that you cannot derive a credible opinion using bad comps- as you have described. However, it would generally work better with good comps.

I'd be careful about the sequence of adjustments. What works well in general in your market may not work well in other markets. Using your statistical analysis you can isolate the variables; however, you still need the skill and knowledge to know what variables you seek to quantify.

Brad Ellis, IFA, RAA