====== Part I
Most homes start off as rather conforming subdivisions, where there are plateaus of models with about the same features and quality. Homes in such plateaus have about the same price, perhaps varying a bit on location. But as time passes, some homes are taken better care of than others, some get upgrades or remodeling every so many years, some don't, and then they start to diverge from specific models into an almost endless variety. Appraisal keeps getting more complex - and that is where we are at today - as no time in the past. To get away from subjective, fly by the seat of pants estimations of market values on specific features you do need to get into non-parametric regression such as MARS. But it produces complex models that are difficult to communicate to laymen or even real estate attorneys and judges. That kind of communication, to be effective, requires a tool set founded in R or Python graphics. You shouldn't be surprised to find MAIs and SRAs very good at generating graphics with R.
- But now you are getting into the very high end of appraisers - and even they have their limits. Many can do the graphics but tend to stick to parametric and simple statistical methods such as cluster analysis and at best multi-linear regression.
A good appraiser capable of being an Expert Witness should know:
1. Data-mining, e.g. non-parametric statistics.
2. Skill at communicating relationships and patterns through graphics.
3. Skill at communicating with professionals such as real estate attorneys and judges, - and possibly, heaven forbid, juries.
4. Plus many other things - such as simply being able to explain the reason appraisers are needed over real estate brokers.
....
So, there is that. And another big problem is that good appraisal requires more time and money than a CMA or BPO.
Part of the reason we have Appraisal Organizations such as the AI is to explain these things to the public and market Appraisal so that appraisers get paid reasonable fees for their time.
And it seems all these organizations do is try to find less qualified individuals to work at lower fees.
So, why should appraisers pay into these organizations? It should be the mortgage brokers and lenders paying their fees and they should rename their organizations to something else, I had better not
===== Part II: "Financial Engineering". (or more specifically for appraisers "Valuation Engineering")
1. Financial Engineering (of which the newer field of Valuation Engineering is a specialty of) will become more dominant in the future. Anybody who wants to become an advanced appraiser or "Valuation Engineer" should think about getting a degree in Financial Engineering (MFE - Master of Financial Engineering) is a highly sought after degree:
2. The equivalent certification is 'CFA' which I am sure you have heard of.
3. Besides Python, most Financial Engineers also know C++ because they write complex programs in fast C++. Rust is also used. R is still good as well - and R uses a lot of C++ libraries as well. Some would be surprised to know that R is often faster than Python, primarily because of a plethora of C++ packages written for R.
4. There are Financial Engineers that do high level valuation that uses very advanced statistics, who are not licensed appraisers. What they do is often way beyond what any appraiser would understand. Their end product is typically high quality, and they likely consider USPAP superfluous and unimportant.
5. That's the way things are going. You can certainly find differences of opinion. But, some organizations are more concerned about "the real thing" - rather than some, for them, meaningless stamp of approval, that in the end guarantees very little of what they need.