The Future of Trust: Research Event Recap

Credit unions must navigate the fast-paced world of fintech while staying true to their mission of serving members. What happens when a trusted community partner moves into a technological space that is fraught with mistrust, in which giving up control of personal data has become the price of online existence? And as technology advances, how can credit unions build and maintain a reputation for transparency and trustworthiness?

Presenters at The Future of Trust research event explored ideas and strategies around how to take advantage of our highly technological environment.

Data stewardship requires getting organized

Members trust their credit unions with an enormous amount of personal and financial data. The first step in taking care of this data – before it can be used to guide strategic planning or underpin decision making – is organization. At many institutions, data is fragmented, captured through disparate processes and saved in several locations. To combat fragmentation, credit unions should periodically conduct a data audit: what data do you have? Where is it stored? How can it be integrated to form a more complete picture of members’ financial lives? While researching the credit union-core provider relationship, Filene Fellow Bill Maurer and Filene’s Taylor C. Nelms observed the need for greater data organization. 

Taylor C. Nelms, Filene Research Institute 

"Data do not take care of themselves. So there's fundamental work to be done on data governance and data management."

One CIO of a mid-sized credit union explained: "First, I've got to fix all of my integration problems so I can automate, and then my people get to be creative and productive." 

Credit unions should have a data governance plan, Nelms observes – policies and procedures for managing data. This plan should include questions around transparency, accountability, and getting input from communities that are potentially affected by the data’s collection and use. Nelms suggests forming a team to take on this challenge and warns that a governance plan should be in place before moving forward with automation. 


A vision for the future: the credit union cooperative data trust

One exciting proposal for credit union data sharing is a cooperative data trust. As an independent organization, explains Maurer, the trust would allow credit unions and core providers to share data, with appropriate safeguards for privacy, that would be held in trust for the benefit of their members. Creating such a trust would mean navigating complex governance and organization challenges – how would data be submitted, integrated, secured, and accessed? How would the cooperative navigate regulatory hurdles around data privacy? What kind of consent would credit unions need to obtain from members before participating in the trust?

Bill Maurer, University of California Irvine

"Are there ways that credit unions can start to at least model or envision a kind of credit union data trust? Something that the credit unions and the cores together would collaborate in, leveraging the distinct advantages that credit unions have already around trust and around member ownership, to put that toward member ownership of data held in trust for members' benefit."

BILL MAURER

This idea is not without precedent, says the University of Colorado Boulder’s Nathan Schneider. For example, the Swiss data cooperative MIDATA offers patients a transparent way to submit their own data for medical research, while retaining control over how it can be used. And the organization digitalpublic.io helps start-ups plan for the possibility of their demise by holding their data in trust for them.

Nathan Schneider explains "the real sharing economy"

"People increasingly are recognizing a mismatch...that the investor-owned, venture-capital-driven platforms are misaligned with the interests of the users who increasingly depend on them." 

The vision of a cooperative data trust borrows from the platform cooperative movement. In an economy where value is increasingly created by connecting various players (e.g., buyers and sellers, drivers and riders), Schneider says that cooperatives are offering new ways of thinking about democratic decision making and value sharing.

Attendees explored Seattle's Living Computer Museum 

Click the image to the left to hear Bill Maurer explain why we need to think about data differently.

"At the core of it is really a shift in our understanding of what data is. We've been conditioned to think …that data is sort of the fuel of the new digital economy and that by freely and willingly giving over our data even without thinking about it …we're getting in return these great services that we can't live without! So we think of data as fuel. If you shift the perspective to data as property, and think of it as something that we own, well, then other things can start to happen. And that's at the heart of this movement toward thinking about collaborative or cooperative data trusts." 

Report No. 482

What you do with your data matters

When it comes to data security, there is bad news and good news, says CO-OP Financial Services’ Paul Love. The bad news is that hackers are becoming increasingly sophisticated, with specialized teams collaborating to take down huge corporations. Hacking tools are cheaply and widely available on the dark web, and the costs of an attack can be devastating.

Paul Love, CO-OP Financial Services 

"Being aware that there is a problem helps us to reduce the potential risk of a credit union or an organization that you care about... becoming a victim of it."

PAUL LOVE

 The good news? "Just do the basics,” Love says. “If you patch your systems... you're not the lowest common denominator anymore.” Credit unions can protect themselves from security incidents by training employees (don’t click on phishing emails!) and keeping systems and antivirus software updated.

Paul Love shares the easiest way to prevent a security breach

"One third of all of these [attacks] begin with a phishing email. You can spend millions and millions of dollars on your security program putting in great, cool technology, but if one person clicks the wrong email? Game over."


Giving members control of their data

It may seem like internet users have relinquished any expectation of privacy. But that isn’t the case, says the University of Virginia’s Lana Swartz: internet users want to maintain their privacy, but feel overwhelmed and trapped by a lack of control – after all, what recourse do any of us have if we don’t agree with the terms of service?

In researching fintech apps, Filene Fellow Bill Maurer and the University of California Irvine’s Melissa Wrapp found that users were deeply concerned with privacy, withholding their social security numbers from apps and even uploading data manually to avoid linking bank accounts. Study participants were irritated when a budgeting app used personal data to advertise targeted credit cards. Instead of feeling like personalized service, this kind of advertising undermined participants’ sense of trust by raising questions about how their data was being used.

At the same time, Maurer and Wrapp found that users of fintech apps placed a great deal of trust in their credit union or bank with respect to their data. What if the credit union could be a trusted arbiter of members’ data and privacy, pioneering a new framework for consent that balances the needs of both parties? Credit unions can show members how their data is collected and used, seek their feedback, and report lessons learned, remaining transparent throughout the process.

Attendees explored Seattle's Living Computer Museum

"Trust is very fragile, and it seems to be particularly fragile right now, across a range of institutions in our society. And I think it's heightened - it's amplified - by the range of technologies that we engage in on a day to day basis, by all of the kinds of data that are being generated by these technologies, and all of the uses of that data, which we are, mostly, unaware of."

BILL MAURER

"We found a big credit union and bank advantage which is that in spite of everything, people still trust their banks and credit unions a lot. And a lot more than platform companies and fintechs."

Harnessing data to deliver value to members

Data analytics is one of the most impactful strategies credit unions can deploy in this age of digital transformation. Once you have organized data and a governance plan, it’s time to get creative on ways to generate more meaningful interactions with your members. 

Want to flag the early warning signs of attrition and deepen relationships with existing members? Analyze your membership to discover segments that offer untapped value? Automate processes to reduce costs and increase efficiency? 

Shazia Manus of AdvantEdge Analytics, part of CUNA Mutual Group, says the data analytics firm is piloting data-driven, predictive models that help credit unions achieve these objectives and more. She encourages credit unions to start with a business problem and an accompanying strategy, so as not to get too hung up on specific technology. Because data analytics maturity is a multi-year initiative, she warns, organizations should move slowly and be thorough. At its best, data analytics is not only valuable, but transformative – for both the credit union and the member experience it delivers.

Report No. 482

Your algorithms may be biased

With a broad, integrated data set, technology can deliver great value to your members – but it can also cause unintended harm. Because there is bias built into the structure of our institutions and society, algorithms that support decision-making can inadvertently discriminate against federally protected categories, says Filene Fellow Bill Maurer. 

Bill Maurer, University of California Irvine 

"We're deploying new systems now that frankly, few of us really, truly understand, that might be setting up the conditions for future kinds of structural bias that we need to guard against." 

BILL MAURER

Variables can serve as proxies for those protected categories, leading to harmful consequences. He points to a ProPublica exposé of the recidivism risk calculation for people arrested in Florida, where black defendants were far more likely than whites to be inaccurately flagged as high risk.

Monitoring your algorithms for bias can be surprisingly difficult, Maurer explains, because sock-puppet audits, in which developers test a process already out there in the market by using a series of fake profiles, run afoul of federal regulations. (You can always conduct such an audit with fake profiles before you release it into the market—but often bias only starts to show up once a process is released out into the wider world).

It is legal, however, to crowdsource an audit, paying users on a platform such as Amazon’s Mechanical Turk to make real loan applications and report the results. If credit unions and core providers were able to set up a cooperative data trust, then members might be able to audit each other’s algorithms and share lessons learned, Maurer notes, pointing to Canada’s Algorithmic Media Observatory as an example of what is possible.

Eric Bulmash, Vancity

Meeting community needs with technology

Are there groups in your local community that could benefit from a credit union-fintech partnership? Vancouver’s Vancity Credit Union discovered that the city’s growing population of independent workers were stressed by the precarity of their financial lives, with uneven income, a lack of social safety net, and an inability to access business loans. 

Based on their findings, Vancity’s Eric Bulmash says, the credit union piloted two programs: health benefits in conjunction with a local insurer and small dollar loans in partnership with a fintech company.

These pilots gave Vancity new revenue streams, improved the experience of both staff and members, and reduced precarity for independent workers.

Supporting other cooperative businesses

Young people at platform cooperative start-ups often seek credit union financing, notes Nathan Schneider of the University of Colorado Boulder, but credit unions can be hesitant when faced with unfamiliar business models. Given that cooperative businesses often struggle to access capital, can credit unions find better ways to support them? 

Chris Hopen, Switch Inc., Meg Young, University of Washington, Brian Knollenberg, BECU

Case Study: Cultivating a reputation for fairness

When public transit advocates in Seattle found themselves stymied by public records constraints, they created the University of Washington Transportation Data Collective, a public-private data-sharing effort that enables a variety of stakeholders to compile and use a large data set, with clear policies governing ownership, use, privacy, and ethics.

UW’s Meg Young challenges credit unions to define themselves by their adherence to fairness, accountability, and transparency – qualities that are conspicuously undervalued in the tech sector.

Organizations like Start.coop, Coop Exchange, and Purpose Ventures have emerged to help fund platform cooperatives, and credit unions would be a welcome partner in that space. Beyond funding, credit unions can build relationships with local platform coops and help to share their story more widely.

Taylor C. Nelms shares wisdom about the future of credit unions

"Credit unions' historical strength has always been what makes them credit unions: their cooperative structure."  


Report No. 482

Fintech as a gateway to the branch

Filene Fellow Bill Maurer and the University of California Irvine’s Melissa Wrapp were surprised to learn that the budgeting, saving, and investing apps they studied were driving users toward in-person financial advice and support. Rather than viewing fintech as the competition, Maurer advises, credit unions should think of it as a gateway to the branch. This is essential because, while technology can reduce costs for financial institutions, it doesn’t replicate the relationship that people build with their financial provider. 

Melissa Wrapp and Bill Maurer, University of California Irvine

"How are we going to think about taking the advantage that credit unions already have as trustworthy partners, as incredibly trusted by their members, as one of the few institutions in society right now that people still invest a great deal of trust in, as credit unions start engaging with some of these new technologies?"

BILL MAURER

Even in an era when consumers expect 24/7 access to financial tools and transactions, they are still more inclined to trust a person than a new product, or their ability to use it.

A recent participant explaining why they trust their financial institution: 

"I know this is [Big Bank] and people are like [Big Bank] is trash, they'll take your money. And I'm like, I know. But like, the person who would always help me was really cool. He's a great guy who like, helped me with all my cards and bank account and working everything out. He was a great person so I trust him." 


Partnering with fintech: watch for red flags

Lana Swartz challenges entrepreneurs to design better fintech

“Most of these actors have just built kludgy systems that are built on top of either the card networks or the ACH. And then have only imagined terms of service and freezes as how they're going to manage them. And are looking for their next exit or acquisition. So I want to see passionate entrepreneurs who really care about these issues imagine better. Because it can be." 

Lana Swartz, University of Virginia

Click the image to the right to hear Lana Swartz explain why money needs to move (much) faster.

When it comes to payment technologies, fintech companies have inherent governance problems that may impact your members, says the University of Virginia’s Lana Swartz. Traditional legacy systems, she explains, have failed to enable peer-to-peer payments at the speed and scope of our internet-connected lives. But fintech, while solving issues of speed and connectivity, operates on a Silicon Valley governance model: the terms of service agreement.

Where legacy systems have a marketplace for handling risk, fintech companies seek to minimize risk with a single tool: account freezes, handed out indiscriminately and inconsistently to users who are flagged for even the slightest suggestion of risky behavior – say, tagging a payment “Cuban” to describe a sandwich. Under this model, users can be hurt, while the company remains protected, accountable to funders rather than users.

When vetting a potential fintech partner, it is important for a credit union to evaluate the systems in place for handling risk and to ask questions about how users are treated if they are suspected of violating policies. What remediation is available if a member’s account is frozen? When it comes to payments, Swartz points out, losing access to your money can be just as painful as not having it in the first place.

The evolving credit union-core provider relationship

The relationship between credit unions and core providers seems to be moving in one of two directions, notes Filene’s Taylor C. Nelms. Some cores are billing themselves as a one-stop-shop, offering every capability a credit union might need. Others have shifted to a model in which they serve as a trusted reviewer of third-party apps, offering case-by-case integrations. The key to a smooth relationship, Nelms suggests, will be the ability to communicate about how to best integrate data without compromising integrity, security, or speed.

Credit unions and core providers need to communicate what they want and need.

When it comes to trust, design matters!

Technology design is critical for building trust. Maurer and Wrapp found that to users, good design signals underlying competence. Just as banks in the nineteenth century used neoclassical architecture to signify permanence and trust, today, financial service providers offering apps need to ensure their interface design meets the highest standards. No weird fonts or UIs that feel like the 1990s. Additionally, apps shouldn’t offer mixed messages—like, “you should be saving more” together with “get this credit card!”. 

Apps should also tread cautiously when offering personalized services based on user data – one wrong step and people find it creepy that the app seems to know too much about their personal lives.

Report No. 482