Market Adjustments

[phoebe33764]

I have read close to 100 posts and reply's about market adjustments in our forum. Joyce Botts gives excellent responses to the subject. Being an appraiser in Florida as well, market adjustments are necessary. I do understand the concept however, what do you do with regard to making a market adjustment when there are more varibles than just market?

For example, the subject has 1850 sq feet and does conform to the subject in all aspects including size however, the closest sale to the subject is 2.35 miles away and 7 months old. Recent sales in the immediate area are present yet are not similar to the subject in size. There is a sale within the mi, not in the same subdivision as the subject, that has sold 8 months ago slightly larger than the subject.

Finding a matched pair is like looking for a needle in a haystack and to try and attempt a market adjustment when everything is "all over the place" seems sketchy and difficult to explain.

There is only one active listing in the neighborhood similar to the subject.

In the county the subject is located in, there has been an overall decline in sfrs as a whole of 10%. Because of the lack of sales in the immediate neighborhood, how can one make a market adjustment if the data is lacking?

I'm sure that I could find similar homes that have sold recently but I have to go at least 5 miles out (to avoid the market adjustment) but this is an FHA vs conventional appraisal.

Any suggestions? m2:

 

[Terrel L. Shields]

Use multi- linear regression and try to find what forces are driving the market place. Use of a MLR, with say data from the past 3 years or so, would tend to smooth out the trend lines. Time, size, etc. ought to either 'fall out" in the regression or suggest that the data are too sparse to make reasonable conclusions about it.

I don't believe a technique so often employed by appraisers where a single or at most 2 paired sale is used to justify the contribution of a pool, fireplace, location, time, etc. is particularly useful. Often however, you have less than ideal data to apply.

In applying time adjustments, I would prefer to use a large region and large data set. Yes, the markets may be somewhat different but in a typical MSA the same factors are affecting the sales. Even the entire MLS (if not an overly large regional one) will reflect a better estimate of the time adjustment over a few isolated sales in a smaller market area.

 

[Mztk1]

I look at the market from three points of view. The first is a running count of the median price and the average price per square foot on a month by month basis that goes back 12 months and shows year over year for the entire county that I provide on page 3 of 6 in the URAR. This data I collect myself.

The second is the MLS statistical sheet that MLS makes available to members on line, that shows the average sale price, average days on market, the number of sales, and the total inventory count on a month by month basis for a four+ year period for all data in the GTAR MLS database.

These two groups give me a good indication of market direction and when I am appraising in an area with few sales to work with they serve as my backbone for a market adjustment.

My third category is the immediate market that the subject competes in. I search for all houses that sold for the past year and for the year prior to that, that would compete with the subject, and then I compare the change in median price, average price and average price per square foot. I also look at all sales in the market on a year-to-date basis compared to all sales in the same market on a year-to-date basis one year ago. Finally, because in many markets in my area the values have stabilized since September, I look at sales between February 1, 2008 and March 31, 2008 and compared their values to April 1, 2008 to May 31, 2008 and see what the changes are.

When I am in a market where the number of sales is limited, then I am usually in a rural market and I can expand the distances. When doing a 8 acre house of 2,000sf built in 1990 in rural Wimauma, for example, I will search for all houses east of I75 and south, north and west of the Hillsborough County border that are on 5 to 10 acres that are between 1600 and 2400sf and were built from 1975 to 2008. I don't know how many square miles that is, but it is a huge area. So while Wimauma may have 2 sales or maybe three that fit these parameters, by expanding my boundaries I may get 10 or 12.

Finally, I also analyze the listings in my market area to show market direction. I will provide a table of how many houses are listed in the market, what the average list price is, the average LP:SP ratio, and a projection of the average price for that pool; I will do the same with the pending sales data. I can figure on getting 2X to 3X the amount of listings in this particular market compared to the number of sales. I can use the average projected price of the listings and pendings against the actual average price of the closed sales to show the continuation of the decline.

Ultimately, using my most competitive active listings and pending sales, though, is what shows whether the market is declining or not, and by how much.

It is not an easy process. It literally takes an hour or two to get through on every report, and since every market changes with new data, I have to redo it for each neighborhood each time, even if appraising the same project for the 2nd time in 3 weeks. But the market adjustment using these techiniques has always been measurable for me.

 

[phoebe33764]

I agree, Jim. I follow the same procedure as you stated in paragraphs one and two. I feel blessed that our MLS board gives us such great data to utilize.

My subject happens to be in a waterfront area with more water sales than non. I will analyze the competing market areas to determine their decline..that is a great suggestion. In this particular area, there is one active listing, I provided this in my report, very similar to subject. I would have preferred more listings to get better idea of a trend however listing and pendings are few and far between.

Right again....it is not an easy process. What stumped me is that I couldn't even find one sale similar in size that had sold in the past 6 months within the 1 mi range. Heck, I only found one sale, 7 month old, much larger than the subject at .80 miles.

 

[CANative]

Don't make GLA adjustments. Bracket the subjec't GLA using comps smaller and larger. Reconcile towards the sales that make most sense.

 

[Mike Boyd]

Don't make GLA adjustments. Bracket the subjec't GLA using comps smaller and larger. Reconcile towards the sales that make most sense.

Sorry, Greg, but that's not good advice. Not making GLA adjustments invites too much criticizm and stips from underwriting that will drive you mad. I don't think all appraisers understand a matched pairs analysis. They do not have to match the subject....just each other except for the specific amenity or feature in question. Although not the most accurate for resales homes, an easy practice is in new home developments. After adjusting out upgrades and/or options and maybe lot premiums and additional baths, pricing differences can then be attributed to the difference in GLA. In a new tract, ALL are equal in quality, location, age and condition. That removes any guess work. Do not try to estimate market reaction, use the builder add on lists to even things out. Note: I do not mean for this to be a matched pairs analysis if the subject is one of the new homes in the tract......only as a way to understand the reasoning and methods of arriving at a MP analysis.

 

[CANative]

I usually manage to get them through when necessary. For some property types where minor differences in GLA can't be extracted (1,800 sf home on an oceanfront with 3 acres and the comps are 1500 sf on 5 acres, 2100 sf on 1/2 acre, 1650 on 4 acres...the most recent sale is 6 months ago and there have only been 5 sales in 3 years) I don't make adjustments. Two comps to bracket and one comp is reasonably similar.

But I don't disagree with you because in some cases reviewer from New Jersey or North Carolina who just go hired on and no longer works in a mall kiosk told me I had to adjust on of my comps because the GLA difference was 118 square feet (1618 for the subject (measured) versus about 1500 for one of the comps (who the heck knows?... it was what the stupid agent put in MLS and there are no records). This was for a property in Redwood Valley... what do you think?

There's no law for adjustments.

 

[Terrel L. Shields]

I feel blessed that our MLS board gives us such great data to utilize.

Lucky indeed. Our current system lacks any sophisticated way to analyze data and lacks exportability into excel...at least I haven't figured out how. Regardless the grid you create, it posts what it dang well pleases and there is not one blessed thing you can do about it... MLX sucks.

 

[Ted Markow]

Jim Klos,
I read your post with full attention. When I got to the end of it all I came up with a couple of questions.

1. After all the research and numbers crunching where do you find time to write a report?
2. I can't but wonder with all the fraud that is popping up all over Florida and California (because they seem to have such high occurances of it) how accurate and reliable do you think your data is. I wonder how much that fraud skews the data.
3. Do you realy think that the underwriters that gets all that data really understands its validity, can verify it or Frankly gives a damn?

Remember my intent is not to rain on your parade.

Don't get me wrong I am thoroghly impressed with the depth of your research and for you it helps you personaly understand the market. But at the end of the day does it realy help the reader of your report to understand and be informed of what you are conveying to them? Maybe you dela with a different breed of underwriters and I realy am not attempting to put them down (gods knows how many read this forum) reality is real!!!

 

[Metamorphic]

Here's what I've been doing.

I define an area in the MLS large enough to have a statically valid number of transactions. I prefer to use "the neighborhood" for consistency sake, but if there's not enough transactions, I broaden out. I pull 3 years worth of ALL data for the defined area, I dont filter at all in MLS. I export it into excell and graph the data. Typically I do sales price and DOM. Then I apply a trend line. I'll play with different order polynomials till I find one that I think best represents the market without any foolishness. At that point I'll go back into the data and do some filtering. If I want to get rid of all the shotgun shacks and mega mansions, I'll sort by GLA, and "HIDE" (not delete) all the ones less than 1000 feet GLA and all the ones larger than 4500 GLA. We have some areas with 100+ year old Victorians I'll usually filter out anything really old cause they're their own animal. I'll get rid of the stuff on big acreage if I'm doing normal residential lot size stuff. etc etc. I'll end up with something like this.

I like this because it provides back up for a lot of the URAR data. Things like DOM, high, low, predom property values, market activity, rate of decline in values. I can also play with the data in the spread sheet to calculate sp/lp and then take averages at various points in time. At least for me is quicker and easier to grab a big wad of MLS data and play with in excell rather than deal with the MLS filters and search engines.

I've got to the point where I can pull the data, graph it, and have it ready to stuff in the report in about 30 minutes.