Using Open Data for Credit Risk
Understanding credit risk for a client whether it be personal or commercial is an exercise in understanding everything you can possibly know about a customer, and then using that information to quantify their risk… Open Data is one of the many data sources credit risk assessors should be using to gain and maintain a deeper insight into their customers.
Currently, there is a breathtaking amount of open data available from the government and other public data sources, but the problem is it’s siloed and fragmented so it’s very difficult to join all the different sources to one another, making it impenetrable as a cohesive data source.
However that situation is now changing, and open data is becoming accessible to every insurer, lender and other risk-sensitive business. This article is a guide on how you can use open data for credit risk assessment.
How is Open Data Relevant to Credit Risk?
As we mentioned credit risk is all about understanding your customer, who they are, how they run their business and all the metadata that surrounds their business.
An example; A business trying to understanding the risk of insuring a chain of stores needs to know a lot of information such as:
- The credit score of the company
- The companies financial status
- Assets the company owns
This is all standard information. But open data can augment that with additional data points to understand a customer more deeply without you ever having to ask them a question or rely on their telling the truth.
Some examples of these open data points are:
- The location of all the properties which the business owns and operates from
- What the crime statistics are in those locations
- The average income of households in the surrounding areas
- If that company’s directors have current, ongoing, or completed:
All of this information is available as open data and is incredibly important to build the best risk models.
Monitoring Credit Risk Beyond the Quote
The role of open data in credit risk doesn’t end with the initial assessment of risk. It’s extremely important that credit risk is an ongoing process where risk is monitored. Wouldn’t you want to know if a company you are insuring or lending to has just been delivered a regulatory notice or changed its business structure significantly?
A good example is from open property and business data. If a business buys a new property or takes on a new director with a history of bankrupting companies, those data points alone could significantly increase their risk. Again… all available as open data.
Understanding Open Data
A common misconception about open data is that people believe it’s all government-produced data. While it’s true that the UK government does produce a lot of open data there are many other sources of information that can be of use to insurers.
An important thing to note is that open data is generally defined as: “Any data that can be freely used, re-used and redistributed by anyone – subject only, at most, to the requirement to attribute and share-alike.”
So using that definition it’s easy to see that open data isn’t just limited to government or local government sources, but it can also include regulatory data and other items such as information published by the Health and Safety Executive.
But How Do you Access Open Data?
There are a few potential ways you could access open data by scraping, collecting, cleaning, storing, de-personalizing, and matching all the related data sources but the reality is that they are all extremely expensive, difficult, distracting and are not core to most business.
To make a long story short; data scientists and scraping infrastructure are both expensive and you shouldn’t waste valuable resources on data acquisition and cleaning. To get good results requires a dedicated company spending years building algorithms and tools to get the job done. That’s exactly what open data providers like Doorda have been doing for years.
“There’s clay in the fields but you still buy plates!”
– Clifford McDowell, CEO at Doorda
Buying Open Data
There are several important benefits, and points to consider when buying open data:
- Return On Investment (ROI)
- Integration and deployment costs
- Data Quality and Freshness
The business case for buying open data is quite simple. All you need to know is:
“Does having the data provide you with a sufficient uplift in credit risk accuracy and monitoring to justify the costs?”
The only way to really know is to buy some data, benchmark a new proof of concept risk model against your old models in their ability to more accurately predict claims made.
This might sound simple but does lead nicely onto…
Integration and Deployment Costs
The ease of integrating an open dataset into your technology stack is so important it’s hard to overestimate how important a good integration system is. Merging and enhancing existing with new data sources can be easy, or can potentially cost you a lot of money.
When setting up a proof of concept, FTP might be good enough to share sample data, but a good data provider like Doorda and provide an API and Hosted Data environment which developers can use to quickly and easily deploy new data into their systems.
Data Quality and Freshness
The final and probably the most important thing to consider is the quality and freshness of the data you buy.
Not all data is equal.
Data quality can vary massively from provider to provider.
Some providers will give you:
- Smaller databases made up of a small number of sources, missing data.
- Databases where data is not linked together by indexes
- Data from propriety and/or questionable sources
When buying open data, speak to your provider and ask questions like:
- “How is the data you provide sourced, and what quality assurance is there?”
- “How often is the data refreshed?”
- “How many sources do you use for data?”
- “How do you ensure your data is clean and fresh?”
- “How accurate is your record matching?”
The answers should be things like:
- “We get data directly from over 400 official and primary data sources.”
- “It’s refreshed daily.”
- “It’s cleaned and matched by robust heavily used algorithms, and automated testing is done on every data source and data point to ensure complete accuracy.”
If you get answers like… We scrape company websites to create insights… or we update our database every week or two, then run for the hills. Data quality is paramount, and if you put data into your risk assessments they will be wrong, exposing you to risk yourself.
Overview and Checklist
So now you know:
- What open data is
- Why it’s useful
- Where it comes from
- Why it’s hard to get
- How you should and shouldn’t access it
- What to look out for when buying it.
At Doorda we’re proud to say we’ve built unique technology to collect, analyse, clean, match, and deliver open data, so now any business can enhance their credit risk assessments and monitoring systems quickly, safely and at minimal cost.
We believe data providers should act as guides for customers, giving you a better understanding of what is out there and how it can be used. If you do want to know how open data can help you specifically, get in touch for a free consultation and we will discuss your use case and show you our data catalogue.