Article

The Remarkable Shift to a Point-to-Point Credit Ecosystem

Mangrove above and below water surface, half and half, with fish and a jellyfish underwater, Carribbean sea
Fred Kelly Photo
Founding Partner
OpFin

9 minutes

Credit unions have an opportunity to bring a new level of service and a wider product range to their members.

As I looked up at the enormous frame, it seemed to resemble the carcass of a humpback whale. I was standing on the floor of the Airbus aircraft manufacturing facility in Hamburg, Germany, witnessing one of the first Airbus A380s in production.

It was early 2007, and an ideological battle between Boeing and Airbus was taking place. On one side, Airbus was backing the hub-and-spoke transportation model with the largest passenger aircraft ever built, the A380. On the other side, Boeing—with its sleek B787—was betting on point-to-point travel, taking people directly from A to B on thinner routes. At stake was some $50 billion—the combined cost of both programs.   

About 13 years later, Boeing was announced the victor. On June 20, 2020, Airbus transported its last A380 fuselage to the final assembly line in Toulouse, France.

The Battle Over Credit Data

Credit unions currently have a ring-side seat for a similar battle over credit data. In the hub-and-spoke camp are TransUnion, Experian, and Equifax, the three major credit reporting agencies that receive and aggregate credit data from thousands of lenders across the U.S.

In the point-to-point corner, we have a rising tide of disruptive fintech companies that make it possible for lenders to receive key credit metrics directly from current and prospective borrowers in real-time.

The two main components to this are:

  • Credit data: how we responsibly source relevant and timely data
  • Credit scoring: how we make sense of this data so that we can better understand a borrower’s probability of default

Although many CRAs offer both these services, these are separate activities requiring different capabilities.

The Current Hub-and-Spoke Model

CRAs exist to address the problem of “information asymmetry,” in which a borrower has a better understanding of the probability of performing his or her obligations than the lender. To achieve this, the credit data market operates like a hub-and-spoke system in which CRAs collect and aggregate credit data from a variety of sources, and then sell credit searches back to those who need such data to make more informed decisions.

In the pre-digital age, the hub-and-spoke approach bridged an important gap. It was impractical for potential borrowers to source, organize and share their own credit data when applying for a loan. Now, with the advent of open banking and, by extension, open finance, data can be shared effortlessly in the background.

From Open Banking to Open Finance

Increased use of data and technology is changing how financial markets work. Open banking enables customers to consent to third-party providers accessing their account information.

Open finance extends open banking principles to give consumers and businesses more control over a wider range of their financial data, such as savings, insurance, mortgages, investments, pensions and consumer credit. It has the potential to deliver transformative benefits for consumers and open finance participants alike.

Drawbacks of the Current System

Overall, the current system has done a pretty good job of providing the market with the means to make credit decisions. However, it has some issues.

Lack of Differentiation. First, it’s hard for a credit union to differentiate its offering. If everyone is using the same scoring system, credit quality becomes commoditized in the same way loan pricing is.

Transparency. While FICO reveals the broad inputs into how the scores are calculated, the algorithms are, as you would expect, proprietary. In fact, FICO does not even produce the scores itself; FICO creates the software that is used by the three major CRAs. Those CRAs plug their own data into the FICO formula to produce proprietary results.

That means that lenders that rely on FICO scores as the principal driver of their credit decisions are relying on a black box, the workings of which are not shared with them. Credit reports contain a lot of useful information. But despite all this data, lenders are left to make guesstimates as to a member’s creditworthiness, because the probability of default is typically not disclosed. The whole point of a credit assessment is to ascertain the probability of default. This is a fundamental disconnect.

One of the great things about baseball is that the stats are there to see for everyone, and they are expressed in their natural form. A batting average is a percentage. It is not masked with an “MLB score” nobody understands.

Transparency is also an issue for consumers. Most seem to lack the basic knowledge of how this information is going to be used by lending institutions, employers, landlords and insurance companies.

Systemic Risk. According to FICO, 90% of top lenders use some version of the FICO Score to guide their lending decisions. Furthermore, most CRA models show a high degree of correlation with one another.  This homogeneity in approach by lenders, implies an elevated level of systemic risk, something we should be trying to avoid given what we learned in the 2008 financial crisis.

Narrow Focus. Despite the opacity of FICO scoring methodology, we do know that a borrower’s credit history and credit utilization rate are the two greatest factors. This narrow focus is troublesome. As the UK’s Financial Conduct Authority recently said in relation to current credit scoring offerings, “They are unlikely to be determinative given the range of other factors relevant to lending decisions.”

Stale Data. A further problem is the timeliness of traditional credit scoring. Because data is normally only submitted monthly to CRA by lenders, credit scores always lag events. Given the reliance placed on such scores, we need data in real time.

Incorrect Data. To effectively predict a consumer’s relative risk of delinquency, credit scoring models depend on the credit information contained in consumers’ credit files to be accurate. Inaccurate credit information may cause credit scoring models to understate or overstate a consumer’s credit risk to lenders.

A 2012 report revealed that more than one in five consumers had a “potentially material error” in their credit file that makes them look riskier than they are. Presumably the industry has made progress since then, nevertheless much of the inaccuracy was down to the hub-and-spoke model in which CRAs need to match data from thousands of contributors. This inherent design weakness remains today.  

Conflicts of Interest. One of the curious flaws in the hub-and-spoke system is that your customers are not the customers of the CRAs. Your customers are their commodity. CRAs are in the business of selling data, not serving your membership. Credit reporting is the second most frequent source of consumer complaints handled by the Consumer Financial Protection Bureau, trailing only debt collection.

Risk of Being a Furnisher. The hub-and-spoke model of credit referencing relies on “furnishers” sharing their customers’ information with the CRA. There are inherent costs and risks in being a furnisher. Organizing and submitting credit data to a central third-party hub requires investment in qualified people and systems. There is also a real level of operational risk. For example, under the Fair Credit Reporting Act, when you provide information to a CRA, you have obligations under the act to ensure the accuracy of the information you furnish.

Key Questions for Credit Unions

Given these drawbacks, many credit unions are looking for better solutions and asking themselves some key questions:

  • How can we lend to more members without compromising credit quality?
  • Is there a way to better understand credit risk through more transparent scoring mechanisms?
  • Would broadening credit data inputs generate more accurate predictions?  
  • How can we monitor the credit quality of the portfolio in real time, without impacting the member’s credit score?
  • Is there a low-cost way to speed up the credit decisioning process without increasing risk?

Like many aspects of credit union management, there isn’t a silver bullet, but rather two layers in which we can find the best solutions. These reside in the process and technology.

The Process

Big transformational projects rarely deliver. What’s more, credit unions are resource-constrained, so it helps to bear in mind the words of Franklin D. Roosevelt: “Do what you can, with what you have, where you are.”

The best way to implement this is by looking at your existing processes before you make any decisions on technology. Before building a house, we design the blueprints. We then decide what tools we need. We don’t buy a hammer and then decide what kind of house we can build with it.

Technology is an incredible tool, but technology is there to serve us, not the other way around. Sometimes we see credit unions trying to adapt their processes to their technology rather than designing their processes for the outcomes they want to see.

Here are Bill Gates’ two rules: “The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.”

The Technology

So, having aligned the processes to meet your desired outcomes, it’s time to look at how we can leverage technology. As we have already seen, the emergence of the fintech sector has opened a range of options that were previously out of reach for credit unions.

There are essentially three components that have come together to provide workable solutions for credit unions:

  • Cloud-based SaaS services. Traditionally, accounting was executed using software hosted locally on a desktop computer’s hard drive. Now most businesses are using cloud accounting. Freshbooks, Sage, Wave and Xero are examples of this.
  • APIs. Application programming interfaces are software intermediaries that allow two applications to talk to each other. This means data can be more easily shared and utilized and that third-party services can be easily integrated with any bank core or credit platform.
  • Big data analytics. The ability to analyze large amounts of unstructured data allows us to learn from the above.

The ensemble of these three technologies has enabled the collection and analysis of new sources of data. So far in the U.S., this has manifested itself through a series of one-off bank agreements, such as the agreements entered into by Chase and Wells Fargo with Xero and Finicity.

Citibank, BBVA and Capital One, to name a few, have opened their APIs to third-party developers. This will continue to grow. The number of global API banking platforms grew nearly 19% between Q4 of 2019 and Q1 of 2020. In time, we will see the emergence of an ecosystem of apps and APIs that support broad synergies between banks, credit unions and consumers.

Credit unions have an opportunity to bring a new level of service and a wider product range to their members. This will ensure the viability of those credit unions that choose to embrace it.

Fred Kelly is founding partner of OpFin Process Optimization, Toronto.

Compass Subscription