Market Conditions: Am I on the right track?

[Kevin Keck]

Interesting work. I am with the too technical crowd. Typically, I am finding that in declining markets the data is so strong that I really don't need that much of a presentation to prove the obvious. Mean sale price is declining, average sale price is declining, and average and mean sale price per square foot is declining. Further, every media outlet within an 100 hundred mile radius is reporting declining home values. Case closed. For a period of time I use to include several graphs in each of my reports until I realized that no one was going to pay me extra or even send me more work as a result of my statistical and graphical presentation. That, however, might be more of a commentary on my clients and not yours.

Specifically, I think your presentation is strong but you should consider your reader. It is unlikely that many readers are going to understand standard deviation or margin of error. Although your explanation of margin of error is so well done that even a dope like me can almost understand. Still, I would consider not including that portion of your analysis in your presentation. Like others, I think the 5 year trends seem like overkill. Also, I would also consider losing the table for non-arms length sales. In my mind, it really serves no purpose and your presentation is just as strong if you present data on all sales and then present the filtered data that excludes those sales that are not arms-length.

 

[zhencai]

The attached PDF file is a sample of what I've been doing for my market analysis. I would appreciate comments.

Just out of curiosity, how much do you charge for a report with such detailed info? I hope you get compensated well for your time and effort.

 

[Ariba]

Also, breaking data down to smaller increments of time helps to prove a stable market which is what most of my area actually has and no underwriter wants to believe.

How do you factor/account for seasonal adjustments, spring versus winter?

 

[Rufus]

For a period of time I use to include several graphs in each of my reports until I realized that no one was going to pay me extra or even send me more work as a result of my statistical and graphical presentation.

I think times are changing and the more thorough report will garner a larger share of the work. It is already true in areas such as litigation. Compare one report with a few lines of boilerplate and another with an in depth analysis and then decide which report will be given more weight.

It is unlikely that many readers are going to understand standard deviation or margin of error. <snip> I would consider not including that portion of your analysis [standard deviation] in your presentation.

Good point. I will include an explanation of standard deviation so the reader will understand that it shows the variability in the data and how that corresponds to the price range of the neighborhood.

Like others, I think the 5 year trends seem like overkill.

Metamorphic explained that in post # 8. A snippet is included here.

All that old data speaks to the credibility of the current data.

Also, I would also consider losing the table for non-arms length sales. In my mind, it really serves no purpose and your presentation is just as strong if you present data on all sales and then present the filtered data that excludes those sales that are not arms-length.

Here I simply disagree. Especially for this particular assignment where over two-thirds of the sales were REO.

Just out of curiosity, how much do you charge for a report with such detailed info? I hope you get compensated well for your time and effort.

Less than a thousand beaver pelts for most assignments.

Thanks again to all as your input will help me improve my analysis and more importantly the presentation of same.

 

[Metamorphic]

Great data, but probably overkill. Here's what I supplied lately.

(( BTW _ Would love to know how you get the DOM data onto same chart w/right hand scale. ))

http://www.screencast.com/users/Qaila/folders/Jing/media/0af755a0-5dc0-4aca-a514-2ef573230c19

There you go. That's a short video screen cap showing how to put up data on a second y axis.

 

[Daystar]

meta,
i am impressed.

 

[Tim Hicks (Texas)]

How do you factor/account for seasonal adjustments, spring versus winter?

The only areas that have seasonal adjustments are lake front properties and not really there. We have only one season here. Summer.

 

[KD247]

It's always a challenge to avoid misleading the intended readers of a report. This is evidenced in the "Ask an Appraiser" forum, where questions are often based on confusion regarding the implications of trends, the impact of one or two comparable sales, or the labeling of specific property characteristics. e.g. "If I divide an existing space into two bedrooms, than shouldn't my house appraise for the price of a two bedroom home?" or "Since the average sales price has increased by 6% from last year, can't I count on my appraisal coming in at exactly 6% over last year's appraisal?"

Showing the wide range of the data points, and not implying unrealistic precision in the computed trend line, can sometimes paint a clearer picture of what appraisers are dealing with.

 

[PropertyEconomics]

Great data, but probably overkill. Here's what I supplied lately.

(( BTW _ Would love to know how you get the DOM data onto same chart w/right hand scale. ))

Rick .. what does your analysis show other than the market is all over the place? I would say your method is a bit under analized. IMHO.
If you look at your data, your 1/07 data appears to be skewed to the high side by a few higher sales while the vast majority of the sales fall below the line. Its why I think averages should never be used do analize a market.

 

[Metamorphic]

Riick's graph is a good example of why you want to go back longer than a year. It also points out why you want to adjust the y-axis scale so that the mass of your data is at least three times as wide as it is tall. You're trying to convince the reader than the sales in the market area are are constrained within some range of prices, and that what we're calling a increase or decrease in the market is represented by that range rising or falling. Having the data be wider than it is tall helps your eye make that connection. Interestingly though, it doesn't really matter for the for the algorithms that excel uses to create trend lines. You just tend not to believe the interpretation on a visceral level if you cant see it.

Of course you can take it too far. The way Riick has the y-scale you cant really see the slope of the range. If he jacked the y-scale up till it covered, say, 0 to 500,000,000 his data set would be a line and you wouldn't be able to perceive the trend because it would only be a pixel or two high. Somewhere there is a happy medium.

Also, by taking data at least 100% older than you really care about, you make sure that a couple of unusually high or unusually low data points at the beginning of the range cant skew the graph. All that extra data provides a mathematical "center of gravity" that dampens some of the variability.

You have to be careful with these techniques because they are often used to spin data. In my former career we used to occasionally do public meetings where you go up in front of a bunch of citizens concerned about the contaminated groundwater under their town, and explain to them what's what. In preparing materials for the meetings we spent a lot of time tweaking these kinds of details to fine tune the message. For instance, if I wanted to make the case that Riiks market was not decreasing, I'd tweek the start and end dates and maybe do some selective filtering of the data to flatten that red line out, then I'd take the line off, and I'd step by step increase the y-scale till I was just shy of the point where you could start to see a decreasing trend. Sneaky sneaky.