UA Database Journalism Digs Into Home Mortgages, Race, Discovers...It's Complicated
News — the industry, the product, the hashtag — took a haymaker from a heavyweight Jan. 10 when then-President-elect Donald Trump called one of its largest purveyors, CNN, “fake news.”
If not the antidote to “fake news” then certainly its antipode is “data journalism,” darling of special projects desks (like the Boston Globe’s Spotlight team). Data journalism is the application of great sets of data to working hypotheses. To find out what’s killing Arkansans, for instance, the data journalist might begin with what. is. killing. Arkansans. — each deceased a data point, each point an Excel doc box, and each row and column an eventual pie slice or vector. Where a Capitol reporter is proud to acquire a working shorthand, the data journalist acquires a working knowledge of computer coding.
When the University of Arkansas’s journalism department went shopping for a new professor last year, it was looking for a data journalist. It found former Wall Street Journal and Dow Jones Newswire deputy bureau chief Rob Wells, by this time teaching at the University of Maryland’s estimable Philip Merrill College of Journalism.
Wells’ first class, last semester’s Database Journalism, dove into mortgage lending and race. Specifically, home loan applications and their rejection rates across five races — white, black, Asian, American Indian and Marshallese — and Hispanics (categorized as an “ethnicity” by the federal government).
To do it the class gathered roughly 25,000 Home Mortgage Disclosure Act records on file at the Federal Deposit Insurance Corporation and elsewhere. Together, they cleaned the data (filtered out reporting errors) and analyzed it in Excel. They shifted it over to a data visualization program called Tableau Public where they could create graphics. (All the while Wells graded them on practice reports.)
Their final report, an Arkansas Public Media exclusive, is published here.
Was race a factor in aspiring homebuyers’ ability to secure a mortgage? No, they discovered.
Mortgage loans jive with applicants’ incomes and credit history, the data show. Race is not a direct factor in who gets mortgages.
“We saw a pretty strong correlation between rejection rates and the median income. The poorer you were, the higher the rejection rates, which isn’t surprising and, in a way, might show that the bankers are doing their jobs,” Wells said.
Was the whole investigation a colossal bust, then?
“I mean, I’m not disappointed at non-racism,” byline reporter Ashton Eley said.
“You can’t twist the data to make the story you want,” said fellow byliner Isabel Dobrin. “You have to be a little more creative and find the newsworthy value in that data. People are denied home loans based on income; so then, why is the median income so low in these groups?”
Dobrin graduates in May, and with her degree and a couple of creditable bylines in her portfolio (the Wall Street Journal and Tulsa World), there’s no reason to believe she won’t be a paid professional reporter somewhere soon. When she thinks back on Wells’ class, she’ll remember the rigor with which she explored this very special skill.
“This is the first data-based reporting course we’ve ever had,” she said. “I could not graduate here and teach myself the things I’m learning in this course in a daily newsroom.”
It’s a dose of quantitative precision in a profession that’s often patchwork.
“I would add that we are professional across the board,” said Prof. Gerald Jordan. “I think that’s always been the strength of our department.”
IF YOU’RE CURIOUS
Good reporters are curious by vocation. Data journalism may produce a higher yield for that curiosity.
“It’s incredibly powerful because it allows us to ask original questions of business and government officials, and to penetrate spin and verify assertions,” Wells says.
So it’s an aid, not an alternative.
“There's been a bit of a misconception going around for a while that data journalism is somehow a replacement for shoe leather reporting, which I think is quite dangerous,” says Chad Day, a long-time special projects reporter at the Arkansas Democrat-Gazette now with the The Associated Press.
It “is more like traditional shoe leather reporting, even if it does at times involve coding. You still call people. You still gather documents. You still need to verify with other sources. And most importantly, you interview your data with the same skepticism you do with a person.”
Wells has seen his students take to the computer screen, the spreadsheets and the keyboard shortcuts — but not as quickly to the interpersonal investigation, the “shoe leather” stuff.
“On interviews,” he told his class late last semester, “really, phone calls, emails — you’re not going to have success [that way] like knocking on doors. Like what Quincy was doing,” he said, referring to Quincy Ward, the 29 year old and the class’s lone homeowner.
He took the initiative to visit an Asian market to try and drum up homeowners’ stories of the mortgage process.
“At least then you got an … answer. Right? But you know, that’s what happens. I’d get shut out all the time when I was at the Wall Street Journal, the AP … although at the AP you got your phone calls returned faster than any place else I’ve worked!”
TEACHER AND STUDENTS
One of Wells’ charms is that he channels the stress of a working journalist. He sweats this stuff.
“I’m really enjoying it a lot,” said Grant Messina, a public relations/business major taking the class for a necessary credit. “This isn’t anything I’m going to do with my experience, but it’s awesome.”
His career opposite may be Ginny Monk, editor of the student paper, The Traveler, who last year reported from the Southern Baptist Convention as an intern for the Arkansas Democrat-Gazette.
This specific reporting project was “something I was really excited about with this class, when Prof. Wells told me about this project. I thought we’d see explicit racism from the banks, and I think we’ve really let the data speak for itself in that it’s really more a poverty issue, not explicitly race.”
“I hate that that’s the perception of the media because I’m glad to see that there’s not people saying, ‘I’m not giving you a loan because you’re black or Marshallese.’”
One pitfall in data reporting is small data sets and uneven data samples. For instance, in the Northwest Arkansas metropolitan data, at 32 percent, Native Americans suffer the highest mortgage loan rejection rates. White applicants, by contrast, are rejected at a rate little more than half that. But white applicants rejected in 2015 for all reasons number nearly 1,200, whereas the total number of Native American mortgage loan rejections is 30.
Still, the most fascinating initial finding of the project, Wells said, is the extreme disparity between whites and American Indians in mortgage acceptances in Northwest Arkansas. “Some of our neighbors due to their heritage like Native Americans haven’t built up the wealth to get a mortgage.”
Sometimes data journalism projects are predicated on a hypothesis, but sometimes the project generates them.
Back in his office, Wells and Eley spoke by phone with Joanna Donohoe, a partner at Seven Sisters Community Development, LLC, a nationwide community-based economic development consultant and a self-described American Indian lending expert.
WHERE FROM HERE?
Wells is currently reprising the role of player-coach in his spring J-405V course, but he’s not planning the same big project.
This semester students Quincy Ward and Taylor Pray will work with Wells to investigate mortgage interest rates across races. (Home loan approval is one thing; the price of borrowing is another, and loan underwriters have some discretion about, say, mortgage insurance.) He’d also like to further explore socio-economic differences, such as the debt-to-income ratio across races. Maybe look at gender and lending, too.
For these students and any future data journalism assignments, the future is — unclear, says department chair Larry Foley, an Emmy-winning documentarian.
“It takes a commitment from a journalistic employer to allow somebody who’s qualified to cover that story and to dig it out, knowing that there might not be a story.”
“Right,” Wells jumps in. “Happened to me at the AP.”
“In the climate we’re in today, would a Woodward and Bernstein exist? For a long time they didn’t have a story.”
Foley is the one who personally wooed Wells away from Maryland, precisely for the data journalism curriculum, so his apparent cynicism belies some very hopeful actions.
“I think the important thing for us in our journalism school, and why we’ve invested in Rob Wells to come in and teach this, it’s a tool that young enterprising investigative journalists need to learn. Maybe in the future there’ll be more investigative journalism than there ever has been. If someone figures out how to make a profit on it you can bet that it’s going to happen.”
(Rob Wells’ student journalists' report on mortgage rejection rates and race in Northwest Arkansas, an Arkansas Public Media exclusive, is Housing Loans Difficult For Most Northwest Arkansas Minorities.)
This story is produced by Arkansas Public Media, a statewide journalism collaboration among public media organizations. Arkansas Public Media reporting is funded in part through a grant from the Corporation for Public Broadcasting, with the support of partner stations KUAR, KUAF, KASU and KTXK and from members of the public. You can learn more and support Arkansas Public Media’s reporting at arkansaspublicmedia.org. Arkansas Public Media is Natural State News with Context.