Is there any in those 97% Fannie's or those 100% VA or the 95% FHA with 6% seller paid concessions ?Where will they find the V portion of the LTV equation if there aren’t appraisals?
I expect this to be the best side gig ever in the coming years.
Wait till a refi boom finally hits. If you can wait.
From what little I know about mortgage loans offered without the comfort of an int/ext appraisal, it seems that the relative value will be determined only by comparing accuracy of an appraisal versus loans based on AVMs, waivers, etc., if necessary by default rates or other tangible benchmarks--altough how many periods will be required to obtain enough data upon which meaningful results can be based, or who would conduct the research if the entities that favor alternatives presumably wouldn't be objective. The comments above about the availability of data is compelling although very rarely is the critical "condition" factor mentioned that is missing from values that are not based upon an appraisal; a vast majority of my residential reports reveal that condition is that primary adjustment factor. I still think there must be a champion for the appraisal industry who's waiting to be crowned.
I'm thinking thaat movie star Jimmy Stewart would do a good job, or maybe The Arnold who is stumping for President Trump about climate change and immigration. Just needs to determine the negative potential impact of automated valuations.... or the positive impact of real-life appraisals...on the environment, or the economy, or the livelihood of us blue collar, working class citizens--which is as Real as any other contemporary political topic, e.g., "Stop the Creep of Automation: The Impact of AI on the Rights of .... something or the other....He died in 2023 but you can take his place just declare your the new guy to fall on a sword. I will give you 10 minutes air time on our midnight propaganda station.
Is it, tho? If a particular tranche in an MBS pool has, say, 500 loans, and the value of the collateral for those loans range from ~ $300k to $800k (even though, technically, they're not pooled by value). Is it a sound hypothesis that exactly 49.99% of the loans will be below the mean and 49.99% of the loans will be above the mean (only possible under a normal distribution)? If you assume a somewhat normal distribution, the lower loans are cancelled by the higher ones, right? If that's the case, then what percentage of accuracy is acceptable? 5%? 10%? OTOH - if there is skew in the distribution, then your hypothesis will not hold and you'll either have more loans with value above the mean or more loans with value below the mean. This is one reason skewness in any distribution is really important. Back to the initial point, though, accuracy at the transactional level is not nearly as important (IMO) as the shape of the distribution of the data points.That is way too high for prudent lending decisions.
Curious: How is "error rate" defined? Who publlished that statistic? And, what is the corresponding rate for appraisal-based values? [Really interesting post]The 'error rates' of off-market property for AVMs is 7% (Zestimate & Redfin, Realtor.com and Movoto don't give their error rate and are probably even higher than 7%). That is way too high for prudent lending decisions. Appraisers know the immediate neighborhood market better and recognize upgrades better than AVMs. Reports by appraisers are more 'relevant' than using AVMs.