Residential Mortgage Data Service- 05GA0A21Q0076- Solicitation Questions and Answers.pdf

Residential Mortgage Data Service Request for Quotation: 05GA0A21Q0076

QUESTIONS ANSWERS

1. Is this considered a new requirement or is there an incumbent? If there is an incumbent, would you be able to share the incumbent Vendor Name, Contract ID, Value, Award Date and Expiration Date?

1. This is a new requirement.

2. Please describe the type of data/records the GAO intends to upload to the cloud-based platform.

2. GAO would be interested in uploading and merging other data sets we typically work with (e.g., Census, HMDA) in addition to others we may obtain from a federal agency or other third party during the course of our work (e.g., consumer credit data).

This would be done in accordance with any technical and other requirements specified by the vendor and agreed upon by GAO.

3. Please clarify whether proposed exceptions to the contract terms should be redlined in the SF1449 or stated in a separate document.

3. Proposed exceptions should be stated in a separate document.

4. Does the GAO prefer the sample licensing agreement in word or .pdf format?

4. The sample licensing agreement should be presented in a .pdf format.

5. Please describe what is to be included under “Transactional Charge” CLIN item?

5. Transactional charges would include any products/offerings separate from the main data product (e.g., an existing module or extract from another related dataset).

Transactional charges are intended to cover any one-time data needs within a contract year. (See question 6.)

6. Is there a preference to include individual pricing for additional products/offerings such as;

Public/Property Record Data, mortgage performance modules, Home Price Index/indices, mortgage origination and performance indices, or GSE mortgage performance disclosure data (i.e. eMBS Data)?

6. (See question 5.) Yes, the vendor should list any other products/offerings and associated pricing in the transactional change lines. GAO may purchase one or more of these additional products/offerings on an as-needed basis.

7. When you state “current and historical data must account for a large enough share of the mortgage market to accurately assess trends at different levels of geography” What do you consider a large enough share by percentage?

7. The vendor must demonstrate the extent to which the dataset is representative at the national, state, Metropolitan Statistical Area (MSA), and metro- county levels. Percentage share of loans by geographic level (and range by state, MSA, and metro-county level) would be sufficient to illustrate coverage. GAO will consider how sample size/coverage compares to other publicly-available sample-based datasets, including HMDA and Census.

In general, GAO considers a sample with a 20 percent relative margin of error at the 95 percent confidence interval to be representative and reportable. However, this standard can vary depending on what GAO is reporting. Additionally, GAO recognizes that many county- and zip-code level samples do not meet this standard and will aggregate to higher geographic levels (such as MSA or state) in its reporting, as appropriate.

8. When you discuss linking loan level mortgage data to tax/deed and other data. This process will de-anonymize the loan level data. It is our understanding that de-anonymizing loan level data violates privacy laws.

Is our understanding incorrect or is GAO exempt from privacy violations?

8. GAO is seeking the technical capability to link vendor datasets (or third-party datasets) in an available cloud platform. This requirement, as written, is not intended to convey that the data come linked upon delivery.

In the event GAO were to purchase data under the transactional line (see questions 5 and 6) or request the vendor link a third-party dataset, we would work with the vendor to ensure personally-identifiable information are either sufficiently masked in the dataset or to ensure that GAO’s reporting of such information is aggregated in a way that does not allow it to be traced to individual loans/borrowers.

GAO regularly receives and works with datasets that contain personally-identifiable information (such as tax records). GAO will work with the vendor to ensure that our practices for receiving, storing, analyzing, and reporting of the data are consistent with data privacy and other related laws and practices and the vendor’s own requirements.