4. Frequently Asked Questions

a. What is a Vehicle Ownership Model?

Traditional vehicle ownership models (VOMs) are undertaken by two sets of groups for two broad purposes.  First, regional transportation planning agencies, such as the Metropolitan Transportation Commission, develop VOMs in order to predict growth in vehicle ownership and travel for the ultimate purpose of planning transportation investments such as highway widenings.  The second group of researchers who develop VOMs are primary researchers interested in exploring the causes of travel behavior.  Frequently, the research is related to exploring whether factors such as land use, density, and transit can impact travel. 

Back to Top

b. Who developed the model?

The underlying Bay Area vehicle ownership model (VOM) was developed by John Holtzclaw, Robert Clear, Hank Dittmar, David Goldstein and Peter Haas for a project called the Location Efficient Mortgage (LEM).  The LEM results from a three year long research program led by three non-profit organizations: the  the Center for Neighborhood Technology, the Natural Resources Defense Council, and the Surface Transportation Policy Project. Together they have formed a new non-profit organization called the Institute for Location Efficiency (ILE). The U.S. Department of Energy, the Federal Transit Administration, and the U.S. Environmental Protection Agency, as well as Nathan Cummings Foundation, the Joyce Foundation, the John D. and Catherine T. MacArthur Foundation, and the Surdna Foundation funded ILEís research. The market test of the LEM is sponsored by Fannie Mae. A Location Efficient Mortgage allows homebuyers to borrow more for the purchase of a home based on predicted household savings from living in a neighborhood that has quality transit service and or neighborhood services. More information on the LEM can be found out www.locationefficiency.com.

The Location Efficient Mortgage uses neighborhood (or "zonal") data to predict the behavior of individual households.  Because this can be problematic, the LEM program includes an education element for homebuyers to make them aware of the cost savings from reducing vehicle ownership and driving trips.  The Housing Development Vehicle Ownership Model, which applies the LEM's model to a whole development as opposed to an individual household, was developed by NPH.

Back to Top

c. Where does the underlying data for the model come from?

The data used for the Bay Area VOM come from the 1990 MTC Household Travel Survey, The 1995 Nationwide Personal Transportation Survey, and the 1990 US Census.

Back to Top

d. What did the LEM study on Vehicle Ownership show?

The statistical analysis indicated that four variables were best at predicting vehicle ownership: household size, household income, residential density, and transit accessibility (a measure of the quality of transit service).  Because these variables are basically known when developing housing, particularly affordable housing, the vehicle ownership model for the LEM can be used to predict the number of vehicles owned at a particular development.

Back to Top

e. How does the Housing Development Vehicle Ownership Model work?

The LEM VOM uses four variables to predict vehicle ownership rates:

• Household Size - Larger households tend to own more vehicles.
  Income per Household Member - Higher income households will generally own more vehicles than lower income households.
  Households per Residential Acre (Density) ñ Neighborhood density is generally a proxy for a number of factors that would lead to fewer vehicles per household.  Those factors include, pedestrian friendliness, neighborhood shopping, slow vehicle speeds, and closeness to jobs. 
  Transit Accessibility Index (TAI) - This measure of the quality of transit service.  Specifically the TAI represents the daily average number of buses or trains per hour times the fraction of the neighborhood/zone within 1/4 mi of each bus stop (or 1/2 mi of each rail or ferry stop or station), summed for all transit routes in or near the neighborhood/zone.

The Housing Development Vehicle Ownership Model takes the proposed details of an affordable, multi-family housing project and converts them into these variables.  Household Size is based on unit size.  For example, since the average bedroom occupancy is 1.5 persons, a 3-bedroom unit is assumed to be a 4.5 person household.  Household income is determined by taking the proposed rent levels and dividing by 30% (the standard rent burden).  This assumption is valid given the income restrictions on affordable units.  Household income is then divided by the persons per household to obtain the income per household member.  The neighborhood density and Transit Accessibility Index is based on the location of the project. The MTC divides the Bay Area into 1099 Traffic Analysis Zones, each of which has its own Density and TAI.  After converting the project details into the VOM inputs, a predicted number of vehicles owned can be generated. See How the Model Works for more. 

Back to Top

f. Does the model work only for affordable housing?

Since income is one of the key inputs into a vehicle ownership model, the Housing Development Vehicle Ownership Model is most effective for affordable housing developments for which the income level of the residents is generally known.  For market rate housing, the income of residents may be above what is inputed from rent levels (by assuming rent is 30% of income).  However, given the generally high rent burdens in the Bay Area, using rent to impute income for market rate housing is acceptable.  Thus market rate projects can benefit from the model as well. 

Back to Top

g. What will impact the accuracy of the model ?

A number of factors specific to a development or neighborhood can impact vehicle ownership so that the model may over predict or under predict vehicle ownership.  Some of those factors include:

• Improvements in transit: If an area is or will be receiving transit improvements, vehicle ownership levels may be lower than predicted by the model.  The model uses a current Transit Accessibility Index (TAI) which could be updated to reflect the neighborhood transit improvements.  
  Changes in neighborhood land use: Since density is a determinant of vehicle ownership and an input into the model, a neighborhood undergoing significant land use change that will increase neighborhood density is likely to have lower vehicle ownership than predicted.  An example is the Mission Bay neighborhood which is adding density and transit service.  Using existing conditions for the model may lead to overprediction.
  Parking costs and availability:  Since data on parking costs is generally unavailable and residents almost always have parking costs bundled into their rent, parking costs can not be modeled.  However, if households were able to choose how many parking spaces they would like to pay for, the vehicle ownership decision would be significantly impacted. Thus, if a development plans on unbundling parking from rent, the vehicle ownership model results should be adjusted lower. If there is plentiful, free parking, the model will under predict vehicle ownership. For more on unbundling, go the the Unbundling section of this website. 

Back to Top

h. My results seem too high, or too low, why?

The model is just that, only a model.  Unique conditions will impact its accuracy.  A dense, transit full neighborhood amidst many neighborhoods that are not dense or transit rich may cause the model to under predict vehicle ownership.  Conversely, a lower density, transit weak neighborhood in a city which is high density and has good transit may cause over prediction of vehicle ownership.  In general, common sense about a local developments conditions in conjunction with the model and its inputs: income, household size, density, and transit accessibility.

Back to Top

i. Where can I get more information on the underlying vehicle ownership model?

An article explaining the research methodology is expected to be published shortly.  For more general information on location efficiency, browse the following websites:

Back to Top