1. Home
2. / Policy & Industry
3. / Fair Housing

Revisiting Mortgage Lending in Detroit: New Insights From HMDA

By Aaron Terrazas on 20 Nov 2015

• Mortgage lending continued to pick up in Detroit in 2014, but remains well below the rates seen in comparable cities.
• Neighborhoods with stronger purchase originations in 2013 have experienced strong home value appreciation in the two years since.
• Mortgage application denial rates have come down, particularly for conventional purchase loans.

Detroit’s housing market is – slowly, seemingly perpetually – “getting there,” recovering in fits and starts and unevenly throughout the roughly 150-square-mile city that is home to more than 600,000 residents.

The number of purchase mortgages written is rising, though remains disconcertingly low for a city of its size. The mortgage denial rate is improving, but still staggeringly high. Home values are showing robust growth in a small handful of the city’s most popular neighborhoods, while in the rest of the sprawling city, home value growth has been far more subdued.

Detroit is turning a corner, but is not yet through a tricky three-point turn.

Zillow visited Detroit in May as part of our Housing Roadmap to 2016. Our analysis of 2013 data from the Home Mortgage Disclosure Act (HMDA) illustrated some of the challenges of mortgage lending in communities where homes often require substantial investment, leading to low appraisals. Armed with the most recent (2014) HMDA data, we’ve revisited that analysis, and discovered that while much has changed for the better in Motown, the city’s housing recovery still lags substantially behind similarly hard-hit American cities.[1]

Mortgage Originations Better, But Still Low

Purchase mortgage originations have continued to rebound in Detroit from historic lows recorded in 2011 and 2012, largely driven by growth in conventional mortgages (figure 1). But the overall number of originations remains well below origination rates for comparable, nearby cities. In 2014 there were only six conventional purchase mortgage originations per 10,000 Detroit residents, compared to 21 in Cleveland and 68 in Chicago (figure 2).

Purchase Originations, Appreciation Concentrated in a Few Neighborhoods…

Purchase mortgage originations also remain clustered in a handful of Detroit neighborhoods (figure 3). The top six census tracts for total purchase originations in 2013 – roughly including the city’s University District, Midtown, Indian Village, East English Village, Palmer Wood and Downtown neighborhoods –accounted for 33 percent of purchase mortgages originated citywide in 2013, up from 12 percent of originations in 2009 and 25 percent in in 2011. In 2014, the share of total purchase originations concentrated in these six areas retreated somewhat, to 26 percent, but remains well above levels of just a few years ago.

Those same six census tracts also feature both higher home values and stronger home value growth than the rest of the city. The median home value in these six census tracts was $104,600 as of September 2015, versus $60,100 elsewhere in the city. Since mid-2013, the median home in these six Detroit neighborhoods has appreciated 38 percent, compared to cumulative appreciation of 13 percent among homes elsewhere in the city (figure 4).

…But Home Improvement Loans are Increasingly Popular Citywide

Although much lower in number than purchase loans – no single census tract recorded more than six home improvement mortgage originations in 2014 – home improvement loan originations are much more spread out across Detroit (figure 5). Fifteen of Detroit’s roughly 300 census tracts had four or more home improvement loan originations in 2014, up from 11 in 2013, with the greatest numbers in the Parkland, Rosedale Park and Blackstone Park neighborhoods.

Mortgage Denial Rates Have Improved Remarkably

Perhaps the most encouraging sign of ongoing recovery in Detroit’s housing market is the precipitously declining mortgage denial rate, particularly for conventional purchase mortgages. In 2009, 61 percent of conventional purchase mortgage applications in Detroit were denied. By 2014, the denial rate had halved, to 31 percent, at least in the same ballpark as other hard-hit cities, if not exactly in line (figure 6a).

But while conventional mortgage denials have fallen precipitously, the decline has been less impressive for FHA and VA mortgages, and remains substantially higher in Detroit than in other cities (figure 6b). Last year, almost half (45 percent) of FHA/VA purchase mortgage applications were denied in Detroit, compared to 30 percent in similarly blighted Newark, and 14 percent in St. Louis.

Collateral Still a Barrier, Along With Credit History

Talking with Detroit mortgage lenders, realtors, city officials and other members of the community in May, many pointed to collateral as a key barrier to mortgage lending in Detroit. Given the overall low number of transactions, it is difficult to find meaningful comparable sales in the city, and many homes require substantial investment after purchase in order to be made habitable. As a result, some banks are wary of making loans against collateral with highly uncertain value.

As in 2013, insufficient collateral remained the most common reason cited for denying conventional purchase mortgage applications in Detroit last year, followed at a distance by insufficient credit history and too-high debt-to-income ratios (figure 7). Among FHA/VA purchase mortgage applications, credit history overtook collateral as the most common reason for application denials in 2014.

Two Steps Forward…

Detroit’s decline, culminating in recent years with the city’s 2013 bankruptcy, was decades in the making. In light of this long history, it’s unreasonable to expect its recovery to be quick, or even complete. Instead, the city is recovering in fits and starts. Some areas are prospering, while others still hope for a more robust turnaround. Compared to other nearby cities, Detroit’s housing market continues to confront enormous challenges. But green shoots are very clearly starting to emerge.

[1] Unless otherwise noted, data presented here are at the city level.