Monday, November 5, 2018

9 Issues that Make an AVM Inefficient, Often Ineffective

Inefficient or Ineffective AVM

1.Time Adjustment: Some practitioners use Number of Months since Sale (NMSS) as an independent variable to ascertain the rate of growth (+/-) in the targeted price level (aka, time coefficient). While this is acceptable as a lead-up regression to generate the time-adjusted sale price (primary dependent variable), it is however unacceptable as an independent variable going into the valuation regression. The reason is simple: since NMSS would be missing in the unsold population (the model would be applied to), the application would fail. Ineffective modeling.

2. Sales GIS: Testing the representativeness of the Sales GIS is not an easy proposition, forcing many practitioners to skip this sampling test. Sales GIS is often a function of market dynamics, deviating from the Pop GIS. Therefore, an untested Sales GIS paves the way for an inefficient AVM. That is why many practitioners tend to use “fixed neighborhoods” in the modeling process as they are more stable (do not succumb to short-term market swings) and established (generally liquid enough to test sample’s representativeness). Inefficient modeling.

3. Chasing Trophies: An AVM is not meant for the entire population. Trying to achieve a two sigma solution (95%) is more meaningful, leaving out the admixture of the unattainable 5% including the limited trophy properties. It is therefore advisable to leave out those properties from the modeling spectrum. Inefficient modeling.

4. Chasing Tiny Bungalows: The flip-side of trophies. Especially the waterfront ones (primarily land value). Inefficient modeling.

5. Combining 2, 3 and 4-Family with SFRs: Modeling 2+ (mostly income-producing) with SFRs is not prudent, even if they are part of the same tax class or mortgage category. Mother-daughter set-up is not a technical 2-family so they could be modeled with SFRs. Inefficient modeling.

6. Synthetic Variables: Synthetic variables like (X * Y) ^ Z may enhance model’s favorable stats but reduces its explainability and decomposability, and therefore overall utility! Inefficient modeling.

7. Untested Models: It's always a good practice to test the model on to a mutually exclusive holdout sample before being applied on to the population. Holdout test must produce very similar results, both before and after the outliers. Inefficient modeling.

8. Sales Complex: Sales complex, directly or indirectly (SP/SF, etc.), must not be used as regressors as it violates the basic regression assumptions. Ineffective modeling.

9. Lack of Value Optimization: Since the error-baed (generally COD-based) regression models do not test the true optimality of the solution (is COD of 8 better than COD of 10?), final regression values need to be optimized via Linear/Non-linear Programming. Inefficient modeling.


- Sid Som, MBA, MIM
President, Homequant, Inc.

No comments:

Post a Comment