Making Loan Data Actionable: Transforming, translating, and assuring data quality and consistency across originators
Many pages could be devoted to the data idiosyncrasies between online loan originators, servicers, and other supporting data providers. Along with the rest of the data science team, and with the assistance of engineering, I’ve spent the better part of the last year researching the datasets of Orchard’s data partners and using this knowledge-base to create a process and infrastructure for standardizing the online lending industry’s data.
There’s no equivalent to the Financial Accounting Standards Board (FASB) and Generally Accepted Accounting Principles (GAAP) or Fannie Mae and Freddie Mac and the Uniform Closing Dataset (UCD) in the world of online lending. And rather than attempt to wrangle the entire industry into conforming to an Orchard Data Standard or promoting a Generally Accepted Loan-level Reporting Standard of some kind (GALRP?), we decided the best approach was to place the onus on ourselves (while doing our best to get people to conform to our standard!)
Much of our standardization process is spent trying to reverse-engineer originators’ reporting methodologies. We’re trying to understand the reporting methodology, checking that it fits our expectations, and, where it doesn’t, trying to figure out why. The quality of the results is largely dependent on the completeness of the raw data and our ability to understand exactly how the originator is reporting the data to us. A close relationship with originators and their data management teams helps.
It’s an ongoing process, and we’re always discovering and resolving nuances in the datasets of our data partners. We’ve managed to distill the basics of standardizing loan data into the roughly twelve pages of this new working paper, but as I mentioned, we could fill many more with examples.
Whether to power the tools investors need to conduct analysis across lenders or to allow originators to improve their businesses by providing better industry benchmarking, among other things, the creation of an industry-wide, standardized dataset will help advance the sector further. With the recent issues around data integrity and concerns about underwriting models and performance at some originators, the hope is that this paper helps advance the conversation around data. At this stage of the industry’s growth, it’s vital that we maintain focus on data quality, integrity, and transparency.
Director, Credit Analytics & Investment Strategy, Orchard Platform
Jonathan is a member of the data science team responsible for building out Orchard’s analytical offerings. He joined Orchard after spending over twelve years focused on consumer lending. First at Merrill Lynch, where he concentrated on agency pass-throughs and CMOs, he eventually focused his efforts on building out their non-agency platform before joining an internal MBS proprietary trading desk as a strategist. Most recently, he was the lead analyst for financial services clients at big data analytics firm 1010data. Jonathan received a BA from Cornell University in mathematics and economics.
We hope that you take the time to download & read the white paper below. Please feel free to contact us if you have any questions.