Would you adjust the comps?

[Wendy]

Doing a bear of a report on a manufactured in the middle of nowhere.

Home +2,000SF, couple of years old, and on 10.00 ac. NO sales that are worth spit in the past 6 months. Four OK sales (or as good as it gets out there) that are 1yr+.

Overall market indicates a +/-9% decline comparing 1st quarter '07 to present quarter. Sample size was about 40 sales each quarter w/ consistent average SF living area.

Graphing just the manufactured sales results in the attached pdf. (quick and dirty graphs using $/sf). Sample size for charts: #1 - 5 sales, #2 - 12 sales, #3 - 16 sales.

Question - would you time adjust the sales used in the report? If so, what data would you use to support your adjustment?

 

[Couch Potato]

What about current listings in the area?

With the small number of sales and the wide range on either side of your linear approximation, I don't think I put much confidence in that approximation but it may be the best you can do.

 

[Wendy]

I have a total of three similar listings for the area. Two support the idea of stability with higher list prices than the previous sales charted.

The third is a listing of the oldest large ac manufactured sale in my data. That one sold for $170,000 in 3/2007 and is active for $179,900 since 4/2008. No change to the property. Listing agent hasn't returned my calls.

PS - You're right - the data quality does not inspire confidence.

 

[Mike Boyd]

What about current listings in the area?

With the small number of sales and the wide range on either side of your linear approximation, I don't think I put much confidence in that approximation but it may be the best you can do.

I agree with Couch. You can only do what you can do. However, you might want to check on values of stick built homes on similar parcel sizes. It is not realistic to think that manufactured home values are stable if stick built homes are declining

 

[Lawrence R.]

In the name of what your peers would do.

I can't see myself coming up with anything more than what you did with what you have...based on what you have it seems the larger sites with larger homes are trying to hold steady...

Listings, though limited, seem to say the same thing...

Call a new dealer and ask him how hard it is to get a new one to "appraise out" If he says he can't get value, there may be a decline hiding in there somewhere that you can't see, as new homes are usually land/home, and don't get reported in MLS, and you can't use them for comps, anyway...

They can, however, reflect(or predict) what is down the pike in the resale market....

 

[Wendy]

..... However, you might want to check on values of stick built homes on similar parcel sizes. It is not realistic to think that manufactured home values are stable if stick built homes are declining

That is what is bugging me. The 9% decline for the market as a whole is realistic and in-line with other similar areas.

However & on the other hand, the market as a whole is not my subject's specific market. If I apply the logic of using the general market rate - then should I not use it on all reports? I know that's a bit extreme, but you get my drift.

edit to add - off topic, but interesting: I'm being forced to use a stick built as a comp (don't ask). I did get pretty good data that supported a 1/3 whack for being manufactured vs. site built. Always surprises me when I get good data out in the sticks.

 

[VolcanoLvr]

Wendy..........

You are absolutely correct that the entire universe of sales activity reported to the various reporting agencies does not reflect your specific assignment.

I like Mike's suggestion. People who are looking for a home generally 'need' a certain amount of s/f due to furnishings, kidlets, lifestyle, etc. So comparing site-built w/ MFH of similar GLA & acreage should yield approximately the same trend % results, despite any potential negative bias toward the MFH pricing. It's the trend % you're trying to establish, from which your time adjustment can be applied if warranted.

 

[Mztk1]

Using $20 increments in your chart makes it look flat. In the $/sf for 1,000sf+ units you went from about $100/sf to about $70/sf, which is 30% down and should look like a cliff because a 30% drop in real estate values over a year would be steep indeed. But you've made it look like contour leveling.

On the first chart with 2+ acres, I appraise in similar markets and know how misleading that can be. The data might be very small, which leads to less reliable results, more lots might be subdividable in the current year than the last because people are looking for "deals" and that is a way to get them. It could be you have a sale on 20 acres in one set of data, etc. Just something small can through it off.

Then the last set shows a drop from what? 90 to 70? That too is 28.5% and should look much, much steeper.

To me 9% is steep enough to label a decline and consider market adjustments. But if I am reading your graphs correctly, your drop is a bit higher than that.

 

[Wendy]

For giggles - I charted all sales of site built +1500 and +2ac (same as chart #1 'cept the building type). Sample size was 9 and the properties varied greatly.

Results are attached.

I hesitate to use this as there was an extraordinary run-up on this type of housing during the boom. Mainly due to market participants being idiots and the wide variance on this type of property.

The results do add another wrinkle to this though.......

 

[Wendy]

Using $20 increments in your chart makes it look flat. In the $/sf for 1,000sf+ units you went from about $100/sf to about $70/sf, which is 30% down and should look like a cliff because a 30% drop in real estate values over a year would be steep indeed. But you've made it look like contour leveling. Good point. I just let it default on the increment - told you it was quick and dirty.

On the first chart with 2+ acres, I appraise in similar markets and know how misleading that can be. The data might be very small, which leads to less reliable results, more lots might be subdividable in the current year than the last because people are looking for "deals" and that is a way to get them. It could be you have a sale on 20 acres in one set of data, etc. Just something small can through it off. I culled the data for ringers.

Then the last set shows a drop from what? 90 to 70? That too is 28.5% and should look much, much steeper. ditto above

To me 9% is steep enough to label a decline and consider market adjustments. But if I am reading your graphs correctly, your drop is a bit higher than that. My problem is identifying a large enough sample to be reliable. What is the cutoff?

Good points - my comments in bold

Edit to add - FTR chart #1 of the manufactureds on +2ac is worthless. 4 of the 5 sales are +1 yr old and the 5th sale, while recent, had some crazy big outbuildings that really skew it. I just included it so you could feel my pain.