Eight years after the subprime mortgage crisis exploded, the name Fannie Mae still stirs unease. Although neither Fannie nor Freddie can be blamed for the meltdown, their misguided investments in mortgage-backed securities necessitated a government takeover that helped spark the biggest financial panic since 1929.
When Bruce Lee became Fannie Mae's CIO two years ago, the mortgage industry was still grappling with the consequences. “I think it’s fair to characterize the mortgage industry as a whole as dealing with the fallout of the crisis and implementing new regulations up until the middle of 2015,” he says. Lenders worked their way through implementing needed controls, leaving little bandwidth for innovation.
But Lee says the industry has made up for lost time in the past 18 months. Fannie Mae's customer base -- institutions that lend money to families acquiring homes and to developers financing rental housing -- has at last been able to address such tasks as automating manual processes, streamlining data exchange, and improving the customer experience. New regulations that mandated standardization helped make such advances possible.
According to Lee, the stage is now set for next-generation innovation involving AI, the API economy, and even the internet of things. In this interview, edited for length and clarity, I began by asking Lee about his mission as CIO.
InfoWorld: What did you see as your mandate when you were hired in 2014?
Lee: I was hired specifically to make the IT group better and higher quality, with more releases and so on -- faster but also at lower risk. In the past we assumed within IT that there's either a quick and risky way to do things or there’s a slow and methodical way to do things. So you either do your full waterfall with all your testing and everything else and you get high quality, low risk stuff -- or you have quick, clean, and new, where you cut corners and take more risks.
The housing industry of the United States is not a place to deliberately take a whole bunch of new risks. You really want to do better, be faster, and lower risk at the same time.
InfoWorld: Were you involved in implementing changes that resulted from the subprime crisis?
Lee: Part of my time here did overlap with finishing out all that work, but I think both the industry and we were ready to move on towards the end of 2015. Like so many industries, it was really a pivot towards the customer.
InfoWorld: How did you make that pivot?
Lee: Customer expectations had increased. Customers were experiencing different kinds of financial transactions in other areas of banking and lending and people were wondering: Why not the mortgage industry?
You have to both keep pace with regulation as well as meeting the customer need. Up until that point in 2015, it had been more of an either/or discussion with regulation taking the lion’s share of the investment. In the last year we’ve seen it more as being a “both together” thing.
Here’s a good little example: something called the universal closing document. Originally, it was imposed as a regulatory standard. When your mortgage closed, all the essential data about that closing would be captured in one fixed-format document. Now people see that by standardizing it we can actually automate the process. We can use that data for more certainty in the way a mortgage is originated. So people are now more ready to adopt it -- not just due to regulation, but because data and standards and so on are the keys to really making the experience better.
InfoWorld: It seems to me with the new administration you’re probably going to face a regulation rollback. Do you anticipate that?
Lee: I think it’s inappropriate for me to comment on that. Let’s wait and see, and we’ll do whatever we have to.
InfoWorld: Well, change is change, right? Were it to be a rollback, that’s yet another systemic change that needs to be implemented.
Lee: We’ve had a lot of change in this industry, whether regulation-driven or customer-driven, and we’re anticipating change into the future. One of the things we’re doing is moving to a more agile way of working -- with more devops, with more in partnership with other people in the industry -- so that the pace of change can go faster. At the heart of that you still want good data standards and so on if you’re going to really make the experience for the customer better. That’s where we’re focused at the moment: The standards help the process work better.
InfoWorld: So new regulations actually resulted in a better experience for your customers.
Lee: Automated compliance with these procedures leads to things like standards for data exchange. They lead to better outcomes for people. A good example would be collateral and the valuing of collateral.
What do I mean by that? The collateral is your house. When you’re borrowing money, the collateral you’re putting up is the house that you’re going to buy, so understanding the value of that house is very, very important. Up until the crisis, there were very few standards about how the value of the house was determined, such as how big it was, how many bedrooms it had, how big the plot was. None of that data was standardized.
Since the crisis, the data about the house and the characteristics that lead to its value have become standardized. With that standard appraisal data, we’re able then to do much more accurate modeling, much more accurate prediction of future value, and better understand the risk on your mortgage -- and provide access to credit for more people through that standardization. That's an example of how the crisis led to regulation; that standardization through regulation has led to a much better housing industry today.
InfoWorld: What other advances resulted from that kind of standardization?
Lee: We’ve also done things like standardize the addresses of places, if you can believe it. There used to be very few data standards around where you said the actual property was. Now everybody uses the same geocoding technique to name the location of the property.
Another example would be employment verification. Every time you applied for a mortgage, you had to produce a piece of paper to say that you actually were employed. Now we’re able to get that information electronically from a company called Equifax … we just do that in one call to an electronic interface.
Fannie Mae itself has innovated as well. We use artificial intelligence through a product called Kira to harmonize all the contracts we have with lenders. Lenders have been brought onto the platform over the last 30 to 40 years and all their contracts are different, so we’re using AI to find all the different contractual terms and harmonize them. We’re using Narrative Science to produce some of our standard economic research reports about our borrowers or our customers. The big equity houses are using this kind of stuff.
InfoWorld: What about advances going forward?
Lee: The harder areas are those where we see an opportunity to change difficult, longstanding practices with technology.
One of them is the title. There’s a system for determining that the person selling the house owns the house. They have the title and when they transfer it to you, you become the owner of the title of that house -- and it’s registered in all sorts of local systems at the county level all throughout the country.
The big question there is this one: Is that an opportunity for a blockchain structure? Again, we’re just at the beginning of those kinds of discussions and the title companies of America are also thinking about this kind of stuff. When you think industrywide, and you think of this unique trusted proof of ownership, you can’t really go there without thinking blockchain. So I think that’s an interesting one.
On the security side, in the mortgage process I don’t know if you’ve experienced this, but you might notice that typically when you’re filling out 100 pages of mortgage application the bankers conveniently put your social security number on every single page.
InfoWorld: Why do they do that?
Lee: Because if the pages of the paper application get mixed up, they can find all the pages that belong to the right person. It’s very much a hangover from the physical paper world. But really we only need your social security number at two points in the process: just as you’re applying to verify you’re who you are, and maybe if you get into difficulty and we need to renegotiate your mortgage. But every person in the mortgage value chain is storing those social security numbers for 30 years.
So one area we’re exploring at the moment is how we could create a central authority that holds the social security numbers and we only refer to them when we need them. Instead of having social security numbers in millions of mortgage applications across tens of systems in thousands of institutions, maybe as an industry we can largely banish the use of social security numbers and enact a central authority. That would be another example of industrywide innovation that might change the risk profile.
InfoWorld: Who do you think might provide that central authority?
Lee: We have no idea yet.
InfoWorld: The Social Security Administration?
Lee: Exactly. We’ll rely on the secure government for that … I’m kidding! We don’t know. Other people contest it because they say it creates central risk. I would argue we have that already.
InfoWorld: You’ve mentioned Quicken Loans Rocket Mortgage as an example of innovation. Why did you call that out, and how did you work with them?
Lee: When we’re doing innovation in our industry we often reach out for partner organizations to work with, because ultimately we want the benefits to roll to the consumer, and we don’t touch the consumer. We partnered with Quicken during the development of Rocket Mortgage. That was to try and agree on the electronic connections in the collection of data behind that process, which would enable them to speed up confirmation for their borrower that the mortgage was eligible or was approved.
That was an example of trying something in partnership that is now available to everyone. We’ve released all the technology that was developed in that partnership to everybody in the mortgage industry.
InfoWorld: The Fintech space is a happening area right now. What do you find exciting out there?
Lee: Two things spring to mind. One is just increasing the use of data visualization for insights. We see that in lots of areas, such as in operational intelligence for the ecosystem of this new, more complicated mortgage market, but also for things like credit quality, who’s borrowing, and those kinds of things.
I think data visualization that will eventually combine with AI to give us real insights. The other thing that’s emerged from all this innovation is a focus on the API economy.
InfoWorld: Is there a specific collection of APIs that you have your sights on? APIs for government data? Integration among financial services businesses?
Lee: It’s a mixture of things based on what trusted role you can perform in the broader economy through an API. I spoke earlier about ensuring we understand the value of a home. There are people who would like to understand the value of a home outside the mortgage market. Maybe we would produce an API for that.
InfoWorld: What about analytics? Are their opportunities for the internet of things?
Lee: This, for me, is a fascinating area. In a company like Fannie Mae, analytics have always played a big part. We’ve always had lots of models, lots of looking at where the housing industry is going to go, and so on. The difference between analytics and the internet of things, I think, is that the analytics have to guess about far fewer things if you have an internet of things.
I’ll give you an example. Today, we would need a model to estimate how many people aren’t going to properly heat their houses -- and if they don’t, they’ll have burst pipes that will devalue their houses. Do they have appropriate insurance? And if they don’t, will the value of the asset go down?
At a very high level there’s a simplification. How many securities on mortgages will lose value because houses don’t get heated when it gets cold? In the future, with the internet of things, we won’t have to model that. We would be able to know because every Alexa or Nest thermostat that’s sitting in those houses would be able to tell us whether they’re heated or not.
With the internet of things, we’ll be able to know things that we have to model today. It will have a profound impact on the housing industry. There are so many cases. People have to say on their mortgage application whether it’s a primary residence or not. One of the ways people commit fraud is to say it’s a primary residence and they’re going to live there and get around restrictions about the number of second homes and rental properties that can be covered under a mortgage. But if somebody said it was their primary home and their Nest thermometer tells us nobody ever goes in it ...
InfoWorld: I can see that working for the home insurance industry. You could verify that smoke alarms or carbon monoxide alarms were working.