Making Housing More Affordable:
Correcting Misplaced Incentives
in the Lending System

Revision of 27 May 1996

by

David B. Goldstein, Ph.D.
Natural Resources Defense Council
San Francisco, California

I. Introduction

A set of generally accepted principles within the lending industry determines whether people can obtain financing to buy homes or apartments. These rules, which are applied equally to urban and suburban housing despite key differences, effectively force many American families to move to distant locations to own homes.

These lending practices exacerbate urban sprawl, while making home ownership more difficult for inner city families.

The Natural Resources Defense Council (NRDC) is working with the Center for Neighborhood Technology (CNT) and the Surface Transportation Policy Project (STPP) to revise loan qualification rules to provide greater affordability for locations that require less expenditures on driving. Our Location Efficient MortgageSM would use new, technically justified formulas to reflect more accurately the consequences of living in a home that is located where transportation costs are low. The proposed formulas are based on the expenses of an average household in a given location, providing greater allowed loan amounts for housing in locations with lower costs for transportation. They are described in Section IV. Examples of their application are provided in the Appendix.

II. The Problem: How Lending Affects Locational Decisions

Lending institutions provide mortgage loans up to a limit that depends almost solely on the income of the borrower. Typically, lenders allow a household to spend up to 28% of its gross income on housing expenses: loan principal and interest, taxes, and insurance (PITI). Only a small degree of flexibility is available to accommodate families with special circumstances, such as unusually good credit histories. This formula sets an upper limit on the amount a household can borrow, and thus on the amount a typical household can spend for home ownership.

The formula does not depend on the location of the house. Thus, a given household would qualify for the same mortgage amount whether their new house was located in a central city area or other location where the need for driving is relatively low, or in an area on the urban fringe where the need for driving is much greater.

In places where inner city housing costs more than outer suburban housing, current rules on mortgage qualification essentially force households of lower-to-moderate income that want to own their own home to move to locations that require excessive amounts of commuting.

Locational choices reduce or increase the need to drive significantly. Several studies (1) have shown that the total number of vehicle miles driven by a household depends on the density of its neighborhood and on transit usage. Doubling residential density reduces the need to drive by 20-25%. Thus, for example, an average household living in North Oakland, a relatively high-density neighborhood, drives only half as much as the average household in a low density outer suburban Bay Area location. The transportation cost savings from the more central location could be some $250 per month.

The $250 a month savings associated with living in North Oakland, in this example (2), is not considered when lenders decide how much money they will loan for housing. A family with a modest income can be priced out of the market for inner city locations and forced to move to locations that not only require more driving, but actually cost more overall. That is, the mortgage payments plus driving expenses are higher for the low density suburban location (that the lender thinks is more "affordable") than for the more central site. Ironically, this rule may not even be in the best interests of the lender: higher transit-related expenditures could make the suburban home loan more risky to the lender than a comparable loan in a central city location.

III. Environmental and Public Finance Implications

Current mortgage lending practices provide a strong economic incentive for suburban sprawl. Many homeowners borrow up to the limits of affordability imposed by lenders. These potential homeowners are being told that the only way they can "afford" a house is to move far away from town. Many households make that choice.

This situation exacerbates environmental and societal problems. By encouraging those suburban locations that require the most driving, the system creates more smog. It forces people to spend more time commuting than they would choose to do freely. It contributes to the decreasing investment in inner-city infrastructure, which in turn exacerbates urban flight.

A system that encourages urban sprawl also imposes immense new costs on society. Locating in currently undeveloped suburban areas requires the construction of new roads, new utility lines, new schools, increases in freeway capacity, and transit extensions. It paves over prime agricultural land and lowers the quality of life for everyone. Finally, it tends to isolate poor, elderly, and physically-challenged people.

IV. A Better Alternative

A. Location Efficient Mortgages

 

Lending qualification formulas should be based on the ability of a borrower to repay debt. A reasonable formula, both from the point of view of the borrower and the lender, would recognize both the income and the expenses of a household living in a given location. Such a formula should allow a dollar a month saved on transportation to be applied to a dollar a month higher loan payments. It should keep housing in the areas with the highest transportation costs equally affordable, but make housing in areas with lower transportation costs more affordable.

NRDC has developed a model for such a formula and for a methodology to use the formula to evaluate transportation cost savings for any individual house. The formula is based on research that analyzes vehicle ownership and annual auto mileage data for households in twenty-eight California communities, and attempts to fit the results to four explanatory variables describing neighborhood characteristics as well as income and household size (3).

 

The results of that study show density of housing to be the primary determinant of transportation costs, with transit access a secondary determinant. The other variables, including income, have not been statistically significant when density and transit are considered.

Based on these equations, household density and transit accessibility can be used to predict automobile ownership per household and miles driven per household. Using cost estimates for the fixed and variable costs of automobiles allows the computation of average annual automobile costs from these results. Average costs for public transportation in a given community are much smaller, but can be determined from recorded revenue of transit agencies. The net costs are summarized in Table 1.

Note that these studies do not address behavioral parameters that can change over the life of a thirty-year mortgage. In particular, they do not look at the location of jobs compared to the location of the home (4).

 

Efforts to reform loan qualification rules will be coordinated with the efforts of the Consumer Home Energy Efficiency Rating System, Inc. (CHEERS) (5). As a first step, CHEERS has funded the study in noted in reference 1.

 

This discussion has focused only on qualifying potential homebuyers for the additional monthly payments associated with a more expensive house located in a more accessible neighborhood. For many prospective homebuyers, the key roadblock to home ownership is not the mortgage qualification, but rather the downpayment. To address this issue, we intend to find partners who have a direct stake in strengthening central cities and/or in providing increasing homeownership opportunities (i.e. government, redevelopment agencies, pension funds, credit unions, and employers) who will help subsidize the downpayment. We anticipate a 5% downpayment, which is in line with other homeownership programs such as those of Fannie Mae and large mortgage lenders.

B. How Location Efficient Mortgages Work

1. The Way It is Now

Mortgage qualification is determined by screening the applicant by the use of two ratios. The first ratio computes principal, interest, taxes, and insurance (PITI) and compares monthly payments to gross monthly income. An upper limit, typically 28% but occasionally "stretched" a few percentage points, is used as a primary screen. The secondary ratio looks at total recurring monthly payments, including both PITI and other debt, such as automobile loans, long term credit card debt, personal loans, and student loans. A typical limit for all debt payments is 36% of gross income.

 

2. Revised Formulas for Location Efficient Mortgages

Transportation costs can be folded into the equation by estimating transportation savings (TS) in dollars/month comparing transportation costs in the location efficient area to those of a representative non-location efficient area.

Thus, the primary criterion becomes:

 

PITI - TS £ .28 x income

For the secondary ratio, the formula would be:

PITI + other long term debt - net transportation savings £ .36 x income

Where net transportation savings is defined as transportation savings adjusted for monthly automobile payments.

The formula would be implemented very simply through the creation of a database for a metropolitan area or rural region in which the real estate agent, lender, or other party, would simply enter the address of the property and the database would provide a pre-calculated estimate of Transportation Savings (TS). The database would be created by measuring density and transit access for each property and calculating transportation cost savings using Table 1.

Adjustments to the formula would be made for an applicant's actual car ownership. For example, if the applicant has two cars and is applying for a mortgage in an area with an average car ownership of less than one car, the anticipated savings would be decreased by the corresponding amount.

V. Benefits to All Parties

The proposed system should be better than the current system for virtually all participants in the housing market. For lower- to moderate-income people who want to live in accessible central city neighborhoods (and denser suburban areas) the proposed formula would provide an additional margin of affordability--perhaps $40,000 or more (6)--that would allow them to qualify for homes or apartments that would otherwise be "unaffordable". Qualifying households at the lower end of income range would benefit the most, since the additional affordability, which is the same number of dollars for any house in a given location, is more critical to households with less financial resources. But an important benefit would also accrue to the middle class, who can play an important role in strengthening communities and for whom few programs exist.

This change would be a stimulant to affordable housing, because, simply by changing the rules of the game, currently existing housing stock would be made more affordable by a recognition of its transportation cost savings. New development of housing affordable to moderate income people would be spurred by such a system, since it could sell for a higher price without disqualifying its potential customers.

Potential homebuyers would have greater freedom of choice to select housing where they want to live, rather than where lending policies determine they can "afford" to live. Institutional obstacles to making an environmentally preferable choice of housing location would be removed, and market forces would be enhanced.

Lenders would incur less risk under the proposed system than under the existing system. Currently, lenders focus new housing loans on remote new communities where large tracts of housing are being built at the margins of affordability. When the next gasoline-related economic disruption occurs--as has already happened three times in the last twenty years--these borrowers could, all at once, be at risk of default, either because they couldn't get enough gas to commute to their jobs, or because the cost of gasoline goes up with no alternatives to driving.

The broad range of parties that benefit from this proposal is significant--it allows for a coalition of interests that could support the plan. NRDC, CNT and STPP will work with other interested organizations to assemble this coalition and work with lenders to change home loan qualification rules in a workable fashion. This effort will focus first on the secondary market. The Location Efficient Mortgage concept was presented to senior officials of the Federal National Mortgage Association ("Fannie Mae") in December 1994. Fannie Mae indicated its interest in this effort, supported further research into how Location Efficient Mortgages could be developed and implemented, and expressed a willingness to work with these non-profit organizations to see how the secondary market could participate in the program.

We plan to work more broadly with the lending industry and with interested non-profit and business organizations in developing a successful program for Location Efficient Mortgages.

APPENDIX

Examples of the application of Location Efficient MortgageSM Formulas

The examples below are provided in order to clarify the concepts described in Section IV.B above. We provide illustrative cases of families at different income levels attempting to qualify for a mortgage under the current rates and with Location Efficient Mortgages.

For all of these examples, we assume a mortgage rate of 7%, slightly higher than variable rate mortgages available as of February 1995, and a downpayment of 20%. Examples of qualification for a given level of monthly payments are provided in Section A. Adjustments in downpayments are discussed in Section B. We assume, for convenience, a property tax rate of 2% for Chicago and 1-1/4% for California (7). We also select arbitrary estimates of savings from location efficiency that are consistent with Table 1 but are not calculated precisely.

A. Changes in Monthly Payment Requirement for Location Efficient Mortgages.

Chicago example:

1) Moderate-low income urban infill:
Assumed family income: $28,000
Cost of house (3br family condo): $110,000
PITI: $769/month
Ratio of PITI to income: 0.33: FAILS
Cost of house to meet 28% cutoff: $93,500
Transportation savings credit (TS): $200/month
Ratio of PITI-TS to income: 0.243:

PASSES

San Francisco Bay Area examples:

1) Middle income inner suburban buyer (house near BART):
Assumed family income: $47,000
Cost of house (3br detached): $200,000
PITI: $1273/month
Ratio of PITI to income: 0.325: FAILS
Cost of house to meet 28% cutoff: $172,300
Transportation savings credit (TS): $180/month
Ratio of PITI-TS to income: 0.279: PASSES
2) Low-moderate income single person:
Assumed family income: $20,000
Cost of house (small condo): $95,000
PITI: $605/month
Ratio of PITI to income: 0.363: FAILS
Cost of house to meet 28% cutoff: $73,300
Transportation savings credit (TS): $150/month
Ratio of PITI-TS to income: 0.273:

PASSES

3) Upper middle class urban family (house in central SF neighborhood):
Assumed family income: $95,000
Cost of house (detached): $400,000
PITI: $2546/month
Ratio of PITI to income: 0.322: FAILS
Cost of house to meet 28% cutoff: $348,300
Transportation savings credit (TS): $390/month
Ratio of PITI-TS to income: 0.272:

PASSES

As seen from these examples, Location Efficient Mortgages allow a variety of families who would not currently qualify for a loan to obtain a mortgage.

B. Other Methods of Decreasing Cost to Borrower

A variety of techniques can be used to decrease standard downpayment requirements, closing costs, or total monthly payments. Possible methods of assistance include:

• EMPLOYER SUBSIDIES OF DOWNPAYMENT &/or CLOSING COSTS
  
• GOVERNMENTAL OR QUASI-GOVERNMENTAL SUBSIDIES
  
• SILENT SECONDS
  
• MORTGAGE CREDIT CERTIFICATES
  

We will be working with lending institutions and local government entities to identify these and other subsidies that would best ensure that the Location Efficient Mortgage offers the central city homebuyer a fair chance at affordable homeownership.

Over a longer run, the buyer of the more location efficient house would be a better credit risk because housing costs consist primarily of debt service and are mostly fixed, whereas transportation costs will rise with inflation.

Table 1    Predicted Annual Household Auto Expenses -- Ownership & VMT  Dollars
Density	Public Transit Service 50 Seat Vehicles Per Hour Within 1/4 Mi (1/2 Mi for Rail & Ferries); 24 Hr Avg
Census Tract
HH/Res Ac
	1000	500	100	50	30	20	10	5	3	2	1	.5
1000	1,517	1,542	1,605	1,634	1,657	1,676	1,709	1,744	1,771	1,794	1,833	1,875
500	1,804	1,834	1,908	1,943	1,970	1,993	2,032	2,074	2,106	2,133	2,180	2,230
100	2,698	2,742	2,854	2,906	2,947	2,980	3,039	3,102	3,150	3,157	3,260	3,334
50	3,206	3,261	3,394	3,456	3,504	3,544	3,614	3,688	3,746	3,793	3,877	3,965
30	3,646	3,705	3,856	3,927	3,981	4,026	4,106	4,191	4,256	4,310	4,382	4,506
20	4,034	4,100	4,267	4,346	4,406	4,456	4,545	4,638	4,710	4,769	4,875	4,986
10	4,798	4,876	5,075	5,168	5,240	5,299	5,404	5,516	5,601	5,672	5,797	5,928
5	5,705	5,799	6,035	6,146	6,231	6,302	6,427	6,559	6,661	6,745	6,894	7,052
3	6,483	6,588	6,857	6,983	7,080	7,160	7,302	7,453	7,568	7,664	7,833	8,012
2	7,174	7,291	7,588	7,728	7,835	7,924	8,081	8,248	8,376	8,481	8,669	8,867
Auto Ownership = 2.704 x (Density)-.25
Annual VMT/HH = 34,270 x (Density)-.25 x (TAI)-.076
Average auto costs = $2,203/auto + $0.127/mile, based on keeping a new car for 12 years and driving it 128,500 miles,
Cost of Owning and Operating Automobiles, Vans and Light Trucks, 1991, Federal Highway Administration
The communities studied fall within the cells blocked off with dotted lines.
Prepared by John Holtzclaw

References

1. See J. Holtzclaw, "Using Residential Patterns to Decrease Auto Dependence and Costa," Natural Resources Defense Council, June 1994, which derives independent results from data on 28 California cities and analyzes other studies on the topic.

2. The cost savings figure is derived from differences in automobile ownership per household, miles traveled per household and the average fixed and variable costs of automobiles. See J. Holtzclaw "Using Residential Patterns to Decrease Auto Dependence and Costs," Natural Resources Defense Council, June 1994.

3. The number of miles traveled per year was determined from California Bureau of Automotive Repairs smog checks, by correlating changes in odometer readings with ZIP codes of the vehicle owner, and from U.S. Census data on vehicle ownership. The average distance driven per year is strongly correlated to residential density. The study does not measure commute distances, only total driving. See ref. 1 and D. Goldstein and J. Holtclaw "Efficient Cars in Efficient Cities," Natural Resources Defense Council, San Francisco, 1991.

4. The one attempt to incorporate distance to job sources that was evaluated in the study, an index of neighborhood shopping and thus retail employment, proved to be statistically insignificant.

5. CHEERS is a non-profit consortium of utility, business, and state government interests that is developing formulas and methods to calculate the utility costs of a house. This rating system could be used to qualify homeowners for greater loans, with a one-dollar-a-month savings in lower utility energy costs providing a one-dollar-a-month increase in the maximum allowable mortgage payments.

6. There would also per some absolute maximum ratio of PITI to income.

7. We realize that tax rates are sometimes higher than this, this rate is for example only.

Copyright 1996 David B. Goldstein, Ph.D. Natural Resources Defense Council, San Francisco, California. This work is used with the permission of the copyright owner for publication on the Smart Growth Network web site. Any copies of this work shall include this copyright notice.