Granularity

[Doug Smith SRA]

LSI introduced a new policy requiring two pending/listings on all forms except REO forms.

The LSI Chief Appraiser next distributed four FAQ statements, one of which discussed a specific requirement in the analysis of pending sales and listings.

Granularity: Several have written to argue that you track and report the supply and demand, price changes, etc. based upon a broader database. This might be regular MLS statistics, or an online database that breaks it down to single family or condo. Use of average LP to SP ratios offered by MLS statistics may tend to mislead. They are incorporating properties of different types, locations, pride ranges, and attractive to different buying groups. This averaging process is not granular enough. To better support the estimate LP to SP ratio, build your adjustments from the data you find which is similar, proximate, and most current in your subject's locale, in other words, from your closed comparable sales.

The recommendation comes down to sample size. In this squirrelly market, do you agree limiting the analysis to a small number of closed comparable sales will yield accurate information? If limits are set, what should the parameters be? Do you toss out or ignore nationally distributed data base information recommended by Fannie Mae?

 

[Pamela Crowley (Florida)]

So glad I never have agreed to jump to any tune they play no matter what.

 

[Randolph Kinney]

If you go on straight statistics, the larger the sample size, the more statistically significant any difference becomes between samples.

Of course, there are implicit assumptions about statistics and sampling. That is why he is making the argument of granularity. The sampling is not random and the population as a whole is not Gaussian.

I believe the term "sub market" is used to characterize the universe of the subject and its cohorts (salient property characteristics that define the sub market).

 

[John Hassler]

LSI introduced a new policy requiring two pending/listings on all forms except REO forms.

The LSI Chief Appraiser next distributed four FAQ statements, one of which discussed a specific requirement in the analysis of pending sales and listings.



The recommendation comes down to sample size. In this squirrelly market, do you agree limiting the analysis to a small number of closed comparable sales will yield accurate information? If limits are set, what should the parameters be? Do you toss out or ignore nationally distributed data base information recommended by Fannie Mae?

I'm in general agreement with the LSI Chief Appraiser about "granularity" here although I don't think the knee-jerk reaction to add two pending/listings automatically make for a better report (just a more costly one). Unless one uses the entire data set, in this case world housing prices, there is always some level of granularity (e.g. US, state, county, city, size, room count, subdivision, subject street, etc.). Choosing the appropriate parameters is akin to the skills necessary for comparable sales selection. Like the LSI Chief Appraiser, I believe using the marketing statistics of the comparables is an excellent place to start.

A Scope of Work which requires additional collecting, analyzing and reporting marketing data takes more time and it is only reasonable that a professional appraiser would expect a higher level of compensation. I, for one, am glad to hear LSI wants this increased SOW in light of the current mortgage mess as it should, hopefully, provide the lender with the information necessary to make better lending decisions. The next step would be to get LSI to put their money where their mouth is.

 

[Pamela Crowley (Florida)]

Choosing the appropriate parameters is akin to the skills necessary for comparable sales selection.

A Scope of Work which requires additional collecting, analyzing and reporting marketing data takes more time and it is only reasonable that a professional appraiser would expect a higher level of compensation. I, for one, am glad to hear LSI wants this increased SOW in light of the current mortgage mess as it should, hopefully, provide the lender with the information necessary to make better lending decisions. The next step would be to get LSI to put their money where their mouth is.

 

[Restrain]

Just a thought: LSI may be looking at a significant loss in business in about 6 months. Better than 50% probability that BA (very big LSI client) keeps Landsafe open and uses it for all its business. 'Course this is well off-topic.

To get back to the topic. We've all seen the NAR stats about how the market wasn't tanking, yet some areas show a 30% foreclosure rate. This is the result of using a broad brush on the market. I truly think the memo is the result of using too much techno-jargon, and not writing in plain English. If he had said, just use the data for similar properties within the neighborhood to make your determination, instead of rolling out "granular", well...

 

[Hall McClenahan]

What bothers me the most with the quote as written is how it can be misused by the uneducated, and I believe it was intentional.

It starts out talking about a general classification: Supply and demand.

It then transitions directly to a specific adjustment: List vs Sales ratio

Only it is written in such a way as to make the reader believe it is talking about the general. By applying the specific to the general it is easy to mislead the general public to believe the market is stable or increasing by using the four or five most recent sales/listings.

Just another well written "how to" attempting to keep appraisers playing ball and giving lenders an escape when it goes bad. Blame the stupid appraiser for providing a misleading appraisal.

The Mafia, hiring and training the FBI makes as much sense.

 

[Terrel L. Shields]

The sampling is not random and the population as a whole is not Gaussian.

good point. Statistics are a two edged sword.

The Mafia, hiring and training the FBI makes as much sense.

But who could tell you more about the Mafia than themselves? Better training right?

 

[Hall McClenahan]

But who could tell you more about the Mafia than themselves? Better training right?

Do you think so?

 

[Frederick]

They are directing the appraisers to disregard the larger market trend and identify the trend expressed in a smaller more specific and relevant submarket.

On the surface it would appear reasonable, and it is reasonable, however it also plays into the weakness that all statistics have, that is the selection of the population studied can lead to a predetermined result. Stable, declining, increasing can all be predetermined by selection of the data population.

It is giving a way out of checking the "declining box".

Rather that selecting appraisers who need to be told how to dress and conduct themselves and what needs to be included in an appraisal they should focus their attention on engaging qualified professionals who know the requirements each particular assignment requires to be credible and compensating them appropriately.

Did I miss LSI's directive on how they want the appraiser to brush their teeth?