Are you adjusting your comps for AGE?

[Greg Parker]

Why not alternate weeks doing it one way this week and the other way next week just to screw with their data mining.

Brilliant.

 

[Mile High Trout]

If you operate without a fixed notion of adjustment, you're always screwing with data mining.

You don't need to try and screw up data mining.

It happens naturally when a human appraiser is not behaving and form filling like a computer.

Automate yourself - support your demise.

Who in their right mind uses ancillary services for report completion, even in small part?

Nobody touches my reports start to end. If you've fallen for the outsourced data entry deals, well there truly is a succer born every minute.

 

[HF Mudd]

The markets I cover are mostly mature neighborhoods where age adjustments generally make little sense, especially since the differences between older and newer dwellings can be accounted for with adjustments other than age. For example, in some neighborhoods, older 2 family dwellings would typically have 1 bathroom per apartment, whereas newer 2 family dwellings would typically 2 or 2.1 bathrooms per apartment. The market reaction would be adjusted for by applying the appropriate adjustment for bathroom count and no age adjustment would be required. In other cases, layouts might be dissimilar for older vs. newer dwellings (i.e.: "railroad" flats which require walking through a bedroom to get to the main living area), in which case a functional utility adjustment makes more sense IMNSHO than an age adjustment.

In some neighborhoods, older buildings sell for a premium compared to newer buildings (usually neighborhoods with dwellings of brownstone or limestone exteriors). Try explaining a negative age adjustment to "reviewers" or "underwriters" for newer buildings. IMHO, it makes more sense to utilize quality of construction adjustments rather than age adjustments in such cases.

 

[George Hatch]

I try to stick with the same age ranges as much as possible to begin with.

One thing I think a lot of people overlook with respect to mixing and matching comparables built during difference economic cycles is that designs, features and floorplans changed. Condition notwithstanding, a 1950s tract home has a lot of differences from a home built in the 1920s or 1930s. Remodeling and upgrades usually don't touch the floorplans and room sizes. Just as one example.

 

[CANative]

I try to stick with the same age ranges as much as possible to begin with.

One thing I think a lot of people overlook with respect to mixing and matching comparables built during difference economic cycles is that designs, features and floorplans changed. Condition notwithstanding, a 1950s tract home has a lot of differences from a home built in the 1920s or 1930s. Remodeling and upgrades usually don't touch the floorplans and room sizes. Just as one example.

Take a gander at the CA assessors residential cost handbook. The dividing line is 1990.

 

[George Hatch]

One reason why comparing the results of an analysis of the raw data from one period to the next can sometimes create a problem is because you have to use a couple assumptions in order to make that comparison. The failure of any of those assumptions can be of effect on the results.

Firstly, you have to assume there are enough recent sales for the comparison to be statistically meaningful. Many of the automated valuation models include sales that date back quite a while, much more dated than what an appraiser or a would-be buyer would consider as being indicative of the current sale.

You have to assume that the data in both datasets actually includes arms-length sales and do not include foreclosure transfers, interfamily transfers, divorce sales or other forced sale transactions, etc.

You have to assume the sales consist of properties that are at least somewhat comparable in terms of the physical attributes.

You have to assume that the data distribution in both datasets is generally similar - that means that if the system was looking for homes built between 1950 and 1970 and ranging in sized between 1000 sf to 2000 sf that both datasets include a similar number of the superior condition homes as well as the inferior condition homes.


If my first dataset is comprised with 70% of the inferior homes and 30% of the superior homes, it's going to show a much lower value indication for your home (at that time) than the newer dataset where those percentages are reversed.


In the example you cited, if Trulia was including sales from 2 years back, both of those dataruns would include some really dated sales. Or, they could include non-arms-length transfers. Or, they might include datasets with different mixes. ar any combination of the above.

You especially would want to be leery of really small datasets where a single oddball transaction can skew the medians and averages for the group. One oddball in a 30-record dataset is almost invisible, but that same oddball in a 4-record dataset could really skew the results of the analysis.

As far as big gains in the last 6 months, let me say this: If your market was really showing those types of gains for the exact same homes then you'd have agents knocking on your door 3 times a day soliciting you to sell, because those kinds of gains only occur in a red hot market. Everyone in town would be talking about living the high life and flipping houses for a living instead of working a job. If you're not seeing these other indications of a real estate boom then the chances are that there is no boom to see, and what you're getting from Trulia is just another "failure".

 

[TJSum]

I try to stick with the same age ranges as much as possible to begin with.

One thing I think a lot of people overlook with respect to mixing and matching comparables built during difference economic cycles is that designs, features and floorplans changed. Condition notwithstanding, a 1950s tract home has a lot of differences from a home built in the 1920s or 1930s. Remodeling and upgrades usually don't touch the floorplans and room sizes. Just as one example.

This is what I normally see for market differences, a 70's home might be newly renovated, but if the closets, bathrooms and kitchen are still smallish compared to newer homes, there is a huge difference in market appeal.

 

[Terrel L. Shields]

Quote:
Originally Posted by George Hatch
I try to stick with the same age ranges as much as possible to begin with.

One thing I think a lot of people overlook with respect to mixing and matching comparables built during difference economic cycles is that designs, features and floorplans changed. Condition notwithstanding, a 1950s tract home has a lot of differences from a home built in the 1920s or 1930s. Remodeling and upgrades usually don't touch the floorplans and room sizes. Just as one example.

Take a gander at the CA assessors residential cost handbook. The dividing line is 1990.
__________________

Nice point. I think in the future, the McMansions will stand out as a sore thumb imho. Either they get serious modifications or suffer a long-term obsolescence due to that style/market appeal problem.

There is a difference in pre-1973ish homes here because the old homes rarely had any wall insulation, and most were retro insulated in the ceiling...also was the time when most homes begin installing CHA.

However, in some rural areas, particularly E. Oklahoma I have appraised new homes that had no central heat and air and relied upon window air and space heat - often wood stove.

Pre- WWII homes are almost invariably remodeled and the larger ones are money pits. Again, no one in these parts insulated a house until the 70s.

 

[Punalava]

Pre-UAD, I use to adjust based on effective age (i.e. 30act/20eff in grid). Since UAD doesn't like this approach, these adjustments are now made in "condition" field, with a big fat ZERO for actual age differences.

I agree with Mile High Trout. Sometimes with relatively younger homes I will apply an age adjustment when comps are in similar condition, but differ in age by 5 years or so.

 

[Baxter]

There are appraiser defined fields at the bottom of the grid. Effective age is more relevant than actual age in 99% of my appraisals. Actual age gets a zero, and Effective age gets an adjustment... easy. Functional obsolescence for older design styles should be dealt with separately. I haven't had any problems adjusting only for effective age.