For the cost estimation, we understand that each parcel has its own physical characters that may add up or lower the cost of construction. Our estimate is based on appendix 1 of the Environmental Statement Impact (EIS) report on ADUs from the City of Seattle (2018).
For the mortgage calculation, we assumed the homeowner will borrow from home equity loans (HELOC) for 15 years. This assumption is based on the experience of Oregon where most ADU constructors use either cash/savings or home equity loans. The 6.9% APR is the prevailing market rate accessed from Wells Fargo website for a 15-year fixed-rate loan. Both interest rates and rental income are subjected to future market fluctuations, our analysis doesn’t take this into consideration.
For rental calculation, we used the Zillow median rent listing price ($ per square feet) for one bed room for the current month if the zipcode exist in the data. However, the median rent listing covers only a small range of zipcodes in Seattle. We compare it with the Zillow Rent Index (ZRI) which underestimate the per square feet value since it includes many single family residence(SFR). SFRs are larger and tend to have lower per square feet value. We impute the value of rent for the missing zipcodes based on the ZRI values if they are not in rent listing database. For value-added, we used the Zillow Home Value Index (ZHVI). From the past experience on remodelling, the new constrution tends to add around 50% value to investment.
The approach taken with the application is dependent on the database being hosted locally. Also, as a prototype, the current structure is not readily suitable for hosting as a functioning web application. A potential improvement to that end would be to find a service to web-host our database to ensure all potential users are referencing the same data. The size of our database eliminates some options as solutions. In consideration of Microsoft Azure as a hosting option, changes would be required to shift to that technology.
Our process for generating data and building out our database is not a streamlined process. This limitation curtails any potential future attempts to redo our analysis from scratch. Without access to our revised database, attempts to reproduce our environment could prove quite difficult for a secondary party. Further, there are potential challenges in the event new data surfaces to be added for analysis. Putting a data pipeline in place could go a long way towards supporting the tool’s long-term viability.
Some people construct ADU with the intention to sell them in the future. There are a few instances where existing houses with ADU are sold in the market. We didn’t provide any information about the potential change in valuations. It would be great if additional information can be obtained about these transactions, then inference on house value changes would be possible.