‘We have a winner-takes-all system’: VantageScore CEO Tavares

His childhood experience as an immigrant to the U.S. helped shape Silvio Tavares’s views on credit scores and financial inclusion. 

As VantageScore’s relatively new president and CEO (he took over in October for founder and former CEO Barrett Burns, who retired), he is evolving the company’s eponymous score, an alternative to the FICO credit score, to include more data gleaned from outside the credit bureaus and to try to more accurately predict the creditworthiness of lower-income and underrepresented populations.

VantageScore credit scores are used by more than 3,000 banks and nonbanks, including nine of the top ten banks, with use growing by approximately 18% as compared to 2019 levels, according to a forthcoming report prepared by Oliver Wyman.

Silvio Tavares, CEO, VantageScore
"Wealth inequality is one of the defining issues of our time," says Silvio Tavares, president and CEO of VantageScore. "Unfortunately, the pandemic has exacerbated wealth inequality in our society."

Fannie Mae and Freddie Mac, the government-sponsored entities that purchase mortgages, are considering buying mortgages made using the score.

Tavares, who formerly headed analytics business units at Visa and First Data, has authored over a dozen patents in the financial data space, and was chairman and CEO of the Digital Commerce Alliance, an association of large banks, tech companies and retailers, says he was drawn to VantageScore’s stated mission of using data and analytics to include more people into the credit system. In an interview, he explained his vision for VantageScore.

What was your family’s experience with financial services when you came to the United States?

SILVIO TAVARES: For most people, particularly people that are African American, Asian, American, Latino or Hispanic, there's this shroud around how credit works. It's not really clear. And we all feel that somehow we've been left out. 

My family moved here to the United States when I was about 10 years old. My family is originally from Angola and is Portuguese-speaking. 

Like any family, the first thing you do is you try to find a home in a nice neighborhood where your new family will be safe. Hopefully it has some good public schools. That's exactly what my parents did and they found the perfect home in the right neighborhood. 

And then my late father, who was a Ph.D. in economics and a professor at Harvard University, applied for a mortgage and he was denied. One of my earliest memories was him sitting at the dinner table and explaining to my mom that we were going to have to use all of our life savings to buy a home. And she was like, that's crazy. Why would we do that? If one of us gets sick, it could mean financial ruin for our family. And my dad explained to her, we're going to be able to get that credit score that we don't have and then we're going to be able to refinance the home and it's all going to work out. 

That's exactly what they did and that refinancing ultimately let them pay for my college and my siblings’ college and eventually grad school. It was a really formative experience for our family. When I go around the country and I speak about financial inclusion and I share the story I just shared with you, I've never met an immigrant that hasn't had a similar story where they were just excluded from the credit system. 

The reality is we have the technology, we have the data to include all creditworthy borrowers. And so for me, this was super personal. I've got two notches against me. I'm an immigrant and I'm African American. And so it was very, very important for me to use my arcane talents in data and analytics to enable VantageScore to go to the next level in terms of financial inclusion. 

This is about using data and analytics to include more creditworthy borrowers that have historically been excluded for unjustified reasons. That's why I'm here. We spend a lot of time looking at data analytics and figuring out how we can use that data to be more predictive and include more creditworthy borrowers so that they, like my family, can also realize their American dream. 

Did your family ever get a reason for being declined for that mortgage?

It was very simple. My parents had no credit score. They were new immigrants and they didn't have a credit file. 

From an outsider's point of view, FICO and VantageScore don't look that different. They are both pulling credit bureau data. They look at some of the same things. How do you look at the uniqueness of VantageScore today and where you might be taking it into the future? 

The reality is that we are very, very different. We look at some of the same data, but we treat that data in different ways. Let me give you an example. Until we innovated in this space, most Americans assumed that if they paid their rent, their rental payments would be included in making their credit score higher. They weren’t. So even though some of that data was available, most of the credit reporting agencies and specifically our competitor FICO wasn't including it in the credit score. And we said, that's wrong. For most people, your rental payment is probably your biggest financial obligation on a monthly basis, if you're not a homeowner. 

So we were the first to include that data in the credit score. And guess what? African Americans, Asian Americans, Latinos, new immigrants, they're all disproportionately renters. And so when you make the choice to look at that data and incorporate it as a predictive signal, it means that your credit score is going to be not just better and more predictive, but more inclusive. 

On the opposite side, we looked at things differently. For example, medical debt. It turns out again that African Americans and Latinos tend to have a lower proportion of medical insurance and health insurance. 

We looked specifically at medical debts that were paid and we realized that if you're African American, for example, you were more likely to have a paid medical debt on your record. And we looked at the data empirically and we realized that actually having paid off a medical debt had no predictive power at all. And the main reason for that is for most of us, when we get that bill from the hospital for an unexpected healthcare emergency, we're not sure: Is our insurance provider going to pay for it? Are we supposed to pay for it? It's just very unclear. And so you can imagine if you're an immigrant or a member of an underrepresented group, how much more difficult it would be to navigate that. But when you actually pay it off, even though it went into collection, it's really not predictive. 

So again, we looked at that data differently than the other credit scoring models and we stripped that out of our credit score all the way back in 2013. The others are now doing the same thing because they realize it's not predictive. So we look at the data in fundamentally different ways and include new data in our models. 

Right now, we're focusing on incorporating additional alternative data sets into our credit scores so that we can include more people. It's not just altruism. It's good business practice. I live in San Francisco and in California, the largest state in the country by population and economy, 58% of the population is African American, Asian American or Hispanic. So including everybody is good business sense. 

I have heard that more people can get mortgages and other types of credit when lenders use rental data. I've also heard concerns expressed about rental data being used against people. So when they fail to pay their rent, that can be reported, and yet not all landlords do report the consistent repayment of rent. Are you doing anything to ensure that the rental data is used in a positive way? 

Yes. The way we do that is by looking at the data and regression-testing it. So we look back at history to see whether the addition or deletion of certain data sets will have an impact. We especially focus on disparate impact, so would it negatively impact one group of people, like African Americans or Asian Americans versus the overall population. We publish that analysis. 

But I also want to point out that our job as a credit score is not to enable everybody to get credit, regardless of whether they repay loans or not. The core value proposition that we deliver to banks and other stakeholders is to be able to predict the credit repayment behavior. So if someone hasn't paid their obligations in the past, that's an important aspect of their credit score. We want to reward people who have exhibited the right behaviors and then we want to have a different ranking for people that don't exhibit the right behaviors. But we always look at that in aggregate. And so for the specific example that you talked about, which is rental payment data, what we've found is that when you include rental payment data, on the whole, it actually benefits many more people than it hurts. 

We're finding the same thing when we add alternative data sets to the VantageScore credit score model. So actually when you add consumer-permissioned data from a bank account, as just one example, a larger proportion of people at lower income and middle income get a score bump. But obviously there are going to be some people where you look at that data and it shows that they’re not timely paying their debts. Overall, it benefits people. 

My understanding is a lot of landlords don't report rental payment data. So are you taking the rental payment data from the bank account data?

There are lots of mom and pop landlords who don't have the wherewithal to report data into a credit reporting agency. When it's available, we use [landlord- supplied] data in the VantageScore algorithm. When it's not, we look for other sources. Bank account data is a good example. Our newest model does not yet incorporate that. That's something that's coming up and we're working on it. 

You also talked about medical debt. My understanding is that the credit bureaus are eliminating records of medical debt of up to $500. So this is mainly affecting co-payments rather than the kinds of large, uninsured medical emergencies a low income family might encounter. How does VantageScore approach this? 

Unfortunately, two-thirds of all collection accounts are medical debt. So the biggest driver for collections are these unexpected emergencies. And there's no transparency into what the cost is.

It turns out that these types of debts are not predictive of how a consumer is going to repay any other debts, like a credit card or an auto loan, because it's a completely different type of circumstance where there's no transparency. That's the key principle here.

The pandemic, which is after all a medical emergency, made medical debt data even less predictive than it was before. Many hospitals and care providers have stopped reporting medical debt data because of the chaos that's been brought about by the pandemic. So there's less data available. For all of those reasons, this data is not reliable, it's not predictive. The industry as a whole is very much moving away from that. 

Do you have any mandate or mission for the coming year?

Wealth inequality is one of the defining issues of our time. Unfortunately, the pandemic has exacerbated wealth inequality in our society. In the United States, we have very much a winner-takes-all system. And when it comes to wealth inequality, access to credit is one of the most important things. And a very large proportion of our society has been locked out of access to credit. The biggest way Americans build wealth is by having access to a mortgage. VantageScore competes in auto loans, credit card loans, fintech loans. In all of those categories, we are often the leader or number two. 

In mortgage, we are not allowed to compete, and that is because the government has a monopoly around FICO. So approximately 80% of the mortgages in the U.S. are what are called conforming mortgages that are guaranteed by the two government-sponsored entities, Fannie Mae and Freddie Mac, and VantageScore is not allowed to compete. We think that's unfair and it hurts consumers, too. 

Many consumers who are African American, Asian American, Latino and in rural areas have been shut out of access to a mortgage because of the use of the FICO score. 

So one thing that we're super focused on as a company is gaining access to that market so that more Americans will have a chance at home ownership. We'll be able to shrink that gap, which, by the way, has actually been getting worse, believe it or not.

For reprint and licensing requests for this article, click here.
Credit scores Data management Technology NMN Underwriting 2022
MORE FROM AMERICAN BANKER