|
Synergisms Among Land Use, Transit, and Travel Pricing
Policies Robert
A. Johnston, Corresponding Author Dept.
of Environmental Science and Policy, University of California, Davis, CA
95616 Ph:
530 582-0700 Fx: 530 582-0707 rajohnston@ucdavis.edu Caroline
J. Rodier Institute
of Transportation Studies, University of California, Davis, CA 95616 Ph:
530 757-2791 Fx: 530 752-3350 cjrodier@ucdavis.edu For publication in the Transportation Research Record Abstract
We
review empirical studies of the effects of land use on travel behavior
and conclude that increasing density and mix can decrease vehicle
kilometers of travel (VKT) (vehicle miles of travel (VMT)). We outline
our research project for the USEPA and then describe the travel model
used in the first phase of this research. This is followed by
descriptions of the emissions model and the traveler welfare model also
used in this study. Then, we describe the six scenarios evaluated to
date and the results of the simulations for the year 2015 for travel,
emissions, and traveler welfare. Last, we present general conclusions. Key Words:
Transit.
Land use and travel. Travel pricing. Welfare evaluation. Travel demand
management. 4,600
words plus five tables = 5,850 words total. Synergisms Among Land Use, Transit, and Travel Pricing Policies Introduction
In this project, we simulated low-emission land use scenarios for
the USEPA, using the Sacramento, California travel models. Land use
measures were combined with transit and auto mode pricing policies, to
evaluate the effectiveness of various packages. The Sacramento models
are state-of-the-practice models with a walk and bike mode, an auto
ownership step, land use variables in some mode choice equations and in
the auto ownership equations, and full feedback among submodels. BACKGROUND
There is a great range of findings in the literature regarding
the effects of land use density and mix on auto ownership, mode choice,
and overall travel. In an earlier paper on this subject (1),
we concluded that the literature seemed to indicate that land use mix
(jobs-housing balance) was only slightly effective in reducing travel,
whereas higher densities seemed to reduce travel under some conditions.
A recent OECD conference determined that environmentally
sustainable transportation planning required one to distinguish weakly
unsustainable policies from strongly unsustainable ones and concluded
that the latter category included policies leading to: 1. emissions
resulting in climate change and 2. loss of biodiversity (2).
Higher urban densities were found to strongly reduce travel per person
and, therefore, emissions of greenhouse gases.
Holtzclaw found that
in several California urban regions a doubling of residential density
was associated with 16% lower auto ownership and 25-30% lower travel (VKT:
vehicle-kilometers traveled, or VMT: vehicle miles traveled) per
household (3). Ewing et al. studied six communities in Palm Beach
County, Florida and found that density, mix, and central location all
tend to reduce vehicular travel (4).
Nowlan and Stewart studied travel and development in the Toronto region,
and found that for each 100 dwellings built in the central city area,
about 120 inbound trips were eliminated in the morning peak period (5).
Frank performed a literature review and found two camps, those
who concluded that density and mix affect travel and those who admit
that density seems to affects travel, but mainly through higher parking
costs and self-selection of households who prefer transit and
nonmotorized modes (6).
Using Seattle region household survey
and census tract land use data, Frank found that density and mix
significantly explained amount of vehicle travel.
Using National Personal Transportation Survey (NPTS) data and
census data, Dunphy and Fisher found that greater density is associated
with lower travel (7).
Specifically, a doubling of density resulted in a 10-15% reduction in
travel per household. Some of this effect was thought to be due to
demographic differences between suburban and central city residents and
the authors note we cannot conclude that the creation of denser, infill
centers in the suburbs near to transit will bring about the travel
behavior seen in the cities.
Using 57 case studies from all over the U.S., Cervero found that
a mix of employment types in office areas reduced vehicle travel per
worker. Having residential land uses nearby also reduced travel (8).
Cervero also studied households living near to heavy rail and found that
of the households who recently moved to the area, 29% of those who
formerly drove to work now used rail transit (9).
Also, residents in these areas were about five times more likely to use
transit as the average household in the region.
TCRP Project H-1 used national housing survey data and detailed
data from three large urban regions and found that higher density
reduced auto travel for the worktrip and that greater land use mix often
strengthened this relationship (10).
A study by JHK and Associates performed a substantial literature
review and concluded that density and mix both can reduce travel (11).
A review of the empirical and modeling literatures by Breheny,
however, found no clear
evidence regarding whether centralized development patterns reduce
travel, emissions, energy use, and greenhouse gases (12).
He found only a weak preponderance of evidence that a
"decentralized concentration" of medium-sized cities, each of
which is fairly dense, has the lowest adverse environmental impacts.
Several authors caution, however, that such a land use pattern could
result in higher travel and energy use, unless accompanied by massive
transit investment in interurban heavy rail and intra-urban light rail
systems, accompanied by road tolls and parking pricing.
A recent study in the U.K. examined the empirical and modeling
literatures to determine the social, economic, and environmental costs
of different urban patterns and found that new towns of 5,000-30,000
population near to existing cities were weakly shown to be best on all
criteria, if high-quality public transport was developed (13).
A second U.K. study found that, in order to minimize travel and
greenhouse gas emissions, urban revitalization and medium-sized, compact
new towns were necessary, again with high levels of transit service (14).
This study found that many large nodes of employment throughout the
urban area were environmentally superior to concentrating jobs in the
central city.
A study by Cambridge Systematics examined 1,110 worksites in the
Los Angeles region and found that employer financial disincentives to
driving alone were highly related to mode choice, while local land use
characteristics were not at all (15).
A recent OECD panel of transport ministers found that land use
policies by themselves would probably not be effective, due to the low
cost of travel. This group recommended urban growth boundaries,
increased densities and land use mix, parking charges and limitations,
roadway congestion tolls, large investments in transit, traffic calming
and pedestrian streets, bikepaths, and a four-fold increase in
fuel taxes over 20 years (16).
Using NPTS data for travel and zip code-based data for
residential density, Schimek found that density has only a very small
effect on travel at normal, low densities (17).
Travel, however, was found to fall off rapidly above population
densities of 20
dwellings/ha (8 /acre) (22% of the households lived above this density).
At densities above 40 dwellings/ha (16/acre), travel was 60% lower than
the national average.
Many other papers are reviewed in the Great Sprawl Debate in the
Journal of the American Planning Association (18,19).
In our opinion, none of the dozens of papers is without methods flaws.
Also, as the debate papers make clear, we need to define density
carefully, at the residence and at the workplace. We also need to
specify the whole urban structure, for example the amount of open space
between centers and the distances between centers.
In earlier work, we simulated land use intensification near to
rail transit stations, together with peak-period road tolls and found
that, over 20 years, VKT (VMT) could be reduced about 7% and emissions
by 3-9% (1).
Using another earlier Sacramento travel model, we found similar results
for travel and emissions and found that these land use/transit policies
were economically efficient, for all travelers in the aggregate (20).
The land use plus transit scenario was economically positive for all
household income groups; however, the pricing plus transit scenario
affected low-income households negatively. These findings agreed broadly
with those of many other simulations reviewed by us and with theory.
We wished to refine this earlier work with the improved
Sacramento travel models and attempt to find pricing, land use, and
transit scenarios that were not regressive. In addition to light rail
and bus transit, we wanted to simulate paratransit in the rail station
districts and advanced traveler information systems (ATIS) for all
transit modes. METHODS Travel
Demand Model
This study used the 1996 Sacramento regional travel demand model
(SACMET96) (21).
The model was developed with a 1991 travel behavior survey
conducted in the Sacramento region. Some of the key features of this model include:
1. model feedback of
assigned travel impedances to the trip distribution step
2. auto ownership
and trip generation steps with accessibility variables
3. a joint
destination and mode choice model for work trips
4. a mode choice
model with separate walk and bike modes, walk and drive transit access
modes, and two carpool modes (two and three or more occupants)
5. land use, travel
time and monetary costs, and household attribute variables included in
the mode choice models
6. all mode choice
equations in logit form 7.
a trip assignment step that assigns separate A.M., P.M., and
off-peak periods Emissions
Model
The California Department of Transportation's Direct Travel
Impact Model 2 (DTIM2) and the California Air Resources Board's EMFAC7F
model were used in the emissions analysis.
The outputs from the travel demand model used in the emissions
analysis included the results of assignment for each trip purpose by
each time period (A.M. peak, P.M. peak, and off-peak).
The Sacramento Area Council of Governments (SACOG) provided
regional coldstart and hotstart coefficients for each hour in a
twenty-four hour summer period. Consumer
Welfare Model
Kenneth Small and Harvey Rosen show how a consumer welfare
measure known as compensating variation (CV) can be obtained from
discrete choice models:
where
l
is the individual's marginal utility of income, Vm is the
individual's indirect utility of all m choices, p0 indicates
the initial point (i.e., before the policy change), and pf
indicates the final point (i.e, after the policy change) (22).
The change in indirect
utility is converted to dollars by the factor, 1/l, or the inverse of the individual's marginal utility of
income. Small and Rosen
show how marginal utility of income can be obtained from the coefficient
of the cost variable in discrete choice models.
The compensating variation formula (1) from above was adapted to
suit the specifications of the SACMET96 mode choice models. In these models, households are segmented into income/worker
categories and person trips are generated for those categories. To
obtain compensating variation for each income/worker category h the
following formula was applied for all modes m and for all trips Q
between all origins i and all destinations j:
where
l
is provided by the coefficient of the cost variable in the mode choice
equations. Total compensating variation was obtained by summing the
compensating variation obtained from each income/worker group. Scenarios
Base Case
A financially conservative expansion of the Sacramento region’s
transportation system that serves as a point of comparison for the other
scenarios. This scenario
includes a relatively modest number of road-widening projects, new major
roads, one freeway High-Occupancy Vehicle (HOV) lane segment, and a
limited extension of light rail (east to Mather Field Road).
High Occupancy
Vehicle Lanes (HOV)
The HOV lane scenario represents an extensive expansion of the
Sacramento region’s HOV lane system to encourage the use of carpools.
The HOV lane system is expanded east on SR-50 past Folsom, northeast on
I-80 to Douglas in Roseville, northwest on I-5 to the Sacramento
International Airport, and west on I-80 to Davis. HOV lanes are
increased from 42 lane-km (26 lane-mi) in the base case scenario to 288
lane-km (179 lane-mi). There is also an increase in mixed-flow freeway
lane-miles of 6% over the base case. Limited express bus service that
takes advantage of the HOV lanes is also added to the transit network. Light Rail
An extensive expansion of the region’s light rail system to
encourage the use of public transit. The light rail system is expanded
east to Folsom near SR-50, northeast to Roseville near I-80, northwest
to the Sacramento International Airport near I-5, west with a short line
from downtown Sacramento to West Sacramento, and south to Meadowview
road near SR-99. In total, these expansions include approximately 120 new
track km (75 mi) of light rail. Light
rail and bus headways are halved in this scenario. Pricing Policies
It is widely believed that road pricing policies may be some of
the most effective policies to reduce VKT (VMT) and emissions. In this
study, parking cash-out, VKT (VMT) tax, and peak-period toll policies
were examined in combination with some of the other scenarios described
above.
In regions that do not meet the California clean air standards
(such as Sacramento), California Health and Safety Code Section 43845
requires that employers (who rent parking spaces from a third party and
have 50 or more employees) offer commuters the option to choose cash in
lieu of any parking subsidy offered.
Recent changes in the federal Internal Revenue Code have removed
the tax barrier to enforcing California’s cash-out law. Thus, an
examination of this scenario in the Sacramento region is timely.
To
simulate the parking cash-out scenario in SACMET96, home-based work
trips were charged $5/day for workplace parking in the few zones
identified by SACOG staff as eligible under the policy. The income
benefits of the parking cash-out program are simulated in the consumer
welfare estimates by returning the parking charges to the travelers. Many
have advocated the imposition of a VKT (VMT) tax that captures the
monetary and nonmonetary external costs imposed on society by auto
travel. A 5 cent VMT tax (
3 cents/km) for the Sacramento region was obtained from the low end of
the average national estimates of the external costs of auto use (23).
The per mile cost of auto travel (5 cents) included in the mode choice
models of SACMET96 was increased by 5 cents to represent the VMT tax (3
cents/km).
In addition, a peak-period pricing policy to reduce congestion
during commute hours was also examined.
To simulate a peak-period charge, an extra 10 cents per mile (16
cents/km) was added to the per-mile auto operating cost of travel in the
home-based work mode in the mode choice model. Transit Oriented
Development (TOD)
Transit oriented developments or TODs describe rail station
centers with a relatively high-intensity mix of shopping, other jobs,
and housing located around light rail stations.
In TODs, many activities are within walk or bike distances and
high-quality transit service is readily available for activities that
are near the TOD. The TODs in this study include zones within a one
fourth-mile radius of light rail stations. In
this study, 79 TODs were located around light rail stations and have an
average density of 15 households per acre (6/ha), 10 retail employees
per acre (4/ha), and 20 non-retail employees per acre (8/ha).
These density levels were developed based on a review of current
land use densities in Sacramento areas that are considered to be TOD
prototypes. To achieve the
TOD densities, growth in households (147,917), retail employment
(40,505) , and non-retail employment (135,768) from 1995 to 2015 in the
outer zones (farther than 1 mile (1.6 km) from the light rail lines) are
moved to the zones in the TODs. The
ratios of the household classifications are held constant in all zones,
and thus only the total number of households is changed in zones. School
enrollments are also adjusted to correspond to the changes in
households. To reflect the
improved walk and bike environment of the TODs, the pedestrian
environment factors are increased.
Two types of advanced transit are included in the 2015 TOD
scenarios, Advanced
Traveler Information Systems (ATIS) for transit
and Demand Responsive Transit (DRT).
Many believe that the faster transit travel times provided by
these advanced transit technologies may allow transit to compete more
effectively with the auto for riders and thus reduce VMT and emissions.
ATIS for transit takes the form of pre-trip transit service
information. Transit users are assumed to access real-time transit
scheduling information through 100 kiosks located at transit stations
and workplaces, the telephone, the Internet, and cable television.
To simulate the travel effects of this pre-trip information, the
maximum initial wait times for all transit service in the model are
reduced to five minutes.
DRT service is added to the light rail networks described above
to provide an additional transit option to travelers in nine suburban
zones and to connect travelers to light rail transit.
DRT service is assumed to make use of computers to satisfy
real-time transit trip requests, providing transit service to travelers
when and where they need it. To simulate DRT service, it is coded in the
transit network as 246 km (153 mi) of new transit-only routes with short
direct routes between zones in the identified suburban areas and to
light rail station locations. Headways
for DRT service range from fifteen to thirty minutes. Initial boarding fares for DRT service are $1.25 and
transfers to light rail and regional transit bus service are $0.75.
Limited express bus service is also provided in this scenario. DRT
service is also provided in TODs outside the downtown for a one mile
(1.6 km) radius to expand access to light rail transit stations. DRT
feeder service was not added to downtown TODs because those areas have
adequate feeder bus service to light rail stations.
ResultS
Travel Light
Rail by itself had very little effect on VMT, but the TOD/LRT/Advanced
Transit scenario decreased VKT (VMT) by 6.5% (Table 1). Pricing by
itself was somewhat less effective, while Pricing plus TOD/LRT/Advanced
Transit was the most effective policy set, reducing VKT (VMT) almost 9%.
Pricing plus TOD/LRT/Advanced Transit reduced delay less than Pricing by
itself, due to congestion in the TODs. The HOV scenario increased trips
and VKT (VMT), because of the new capacity.
Light Rail by itself had almost no effect on the Drive Alone mode
share and Pricing by itself was ineffective (Table 2). The TOD/LRT/Advanced
Transit scenario by itself was quite effective, reducing Drive Alone by
3.7%. Pricing plus TOD/LRT/Advanced Transit was most effective. The
Transit share and Walk and Bike share were increased most by Pricing
plus TOD/LRT/Advanced Transit. In both cases, the TOD/LRT/Advanced
Transit scenario by itself was more effective than Pricing by itself,
underlining the importance of coupled land use/transit
strategies in fostering non-auto modes. Emissions
HOV lanes increased emissions (Table 3). All emissions rank the
same as VKT (VMT) for the scenarios, with Pricing plus TOD/LRT/Advanced
Transit the best, followed by TOD/LRT/Advanced Transit, and then
Pricing. Again, we see the importance of combining land use, transit,
and pricing. Welfare
In policy evaluation, economic effects may be evaluated using
several consumer welfare measures. We have found the CV user welfare
measure easy to implement and it produces very useful measures of
aggregate change in traveler welfare and change in welfare by household
income class. The latter measure permits a vertical equity evaluation.
Traveler welfare measures are also useful for public review, because
they aggregate all trips and modes into one measure.
In the aggregate, the HOV scenario produced a small loss to the
region, because of the additional travel induced by the added road
capacity (Table 4). The travelers "feel" 5 cents per mile (3
cents/km) in making travel choices, but actually pay 40 cents (25
cents/km), in the long run, and so the additional value of the added
trips is smaller than their additional cost. (Additional trips have, on
average, half the utility as trips made without the capacity expansion.)
The highest welfare gain was with the Pricing plus TOD/LRT/Advanced
Transit scenario, followed by the TOD/LRT/Advanced Transit scenario. In
both of these scenarios, there is a large decrease in VMT and increase
in the Walk and Bike mode share. It is notable that the Pricing
scenario was not nearly as advantageous as the TOD/LRT/Advanced Transit
scenario.
It is useful to calculate the welfare changes by income class on
a per trip basis, to put them into units the public can relate to (Table
5). The HOV scenario disadvantages the middle income group slightly,
which, along with the upper income group, travels more, but their
savings in time costs is not sufficient to compensate for the added
direct (distance-based) costs. Light rail benefits all income groups,
the lower one of which gains better transit access to more destinations,
and the other groups gain somewhat better transit access and faster
travel times on roads.
Pricing by itself hurts the lower-income group, as we expect.
Because of their low value of time, their time savings do not offset
their higher money costs. The TOD/LRT/Advanced Transit scenario benefits
all income groups, however, because more lower-income households can now
walk and bike and more households gain local transit service and light
rail access to many zones in the region. It is notable that Pricing
plus TOD/LRT/Advanced Transit eliminates the regressive effect of
Pricing by itself. This is an important indicative finding that
needs to be more thoroughly researched, since pricing seems important to
regional reductions in emissions and the regressive effects are
difficult to compensate. Free transit passes would also help in this
regard, in most regions. Conclusions
We found that the TOD/LRT/Advanced Transit scenario was much more
effective in reducing emissions than was Light Rail by itself. Pricing
made the package even more effective. The TOD/LRT/Advanced Transit
scenario by itself decreased emissions and increased regional traveler
welfare more than did our Pricing policy by itself. Pricing plus TOD/LRT/Advanced
Transit was the most effective scenario for reducing emissions and
increasing welfare. Most interesting was the finding that this scenario
was not regressive. Acknowledgements:
We
thank the US EPA for supporting this project. This paper covers only the
first phase of this work, the travel demand modeling. References
1.
Johnston, Robert A. and Raju Ceerla. Land Use and Transportation
Alternatives. Transportation and
Energy: Strategies for a Sustainable Transportation System, ed. by
D. Sperling and S. Shaheen. ACEEE, Washington, D.C, 1995. 2.
Towards Sustainable
Transportation: The Vancouver Conference. OECD, 1997. 3.
Holtzclaw, John. Using
Residential Patterns and Transit to Decrease Auto Dependence and Costs.
Natural Resources Defense Council testimony before the California Energy
Commission, June 1994. 4.
Ewing, Reid, et al. Getting Around A Traditional City, A Suburban PUD, and
Everything Else In-Between. Preprint 940218. Transportation Research
Board, Annual Meeting, Washington, D.C., 1994. 5.
Nowlan, David M. and Greg Stewart. Downtown Population Growth and
Commuting Trips: Recent Experience in Toronto. Journal
of the American Planning Association, Vol. 57, 1991, pp. 165-182. 6.
Frank, Lawrence D.
The Impacts of Mixed-Use and Density on the Utilization of Three
Modes of Travel: The Single-Occupant Vehicle, Transit, and Walking.
Preprint 940425. Transportation Research Board, Annual Meeting.
Washington, D.C., January 1994. 7.
Dunphy, Robert T. and Kimberly Fisher. Transportation,
Congestion, and Density: New Insights. In Transportation Research Record 1552, TRB, National Research Council,
Washington, D.C., 1996, pp. 89-96. 8.
Cervero, Robert. Land-Use
Mixing and Suburban Mobility. Transportation
Quarterly, Vol. 42, 1988, pp. 429-446. 9.
Cervero, Robert.
Transit-Based Housing in California: Evidence on Ridership Impacts. Transport
Policy, Vol. 1, 1994, pp.174-183. 10.
Parsons Brinckerhoff Quade and Douglas. TCRP
Report 16. Transit and Urban Form. U.S. Department of
Transportation, 1996. 11.
JHK and Associates. Transportation-Related
Land Use Strategies to Minimize Motor Vehicle Emissions: An Indirect
Source Research Study. California Air Resources Board, Sacramento,
June 1995. 12.
Breheny, M.J. The Contradictions of the Compact City: A Review. Sustainable
Development and Urban Form, ed. by M.J. Breheny. Pion, London, 1992. 13.
Alternative Development Patterns: New Settlements.
. U.K. Department of the Environment. HMSO, London, 1993. 14.
Reducing Transport Emissions Through Planning.
U.K. Department of the Environment. HMSO, London, 1993. 15.
Cambridge Systematics, Inc. The
Effects of Land Use and Travel Demand Management Strategies on Commuting
Behavior. U.S. Department of Transportation, Washington, D.C., 1994.
16.
Urban Travel and Sustainable Development.
OECD, 1997. 17.
Schimek, Paul. Household
Motor Vehicle Ownership and Use: How Much Does Residential Density
Matter? Preprint 969472. Transportation Research Board, Annual Meeting,
Washington, D.C., January 1996. 18.
Gordon, Peter, Harry W. Richardson. Are Compact Cities a Desirable
Planning Goal? Journal of the
American Planning Association, Vol. 63, No. 1, 1997, pp. 95-106. 19.
Ewing, Reid. Is Los Angeles-Style Sprawl Desirable? Journal
of the American Planning Association, Vol. 63, No. 1, 1997, pp.
107-126. 20.
Rodier, Caroline J. and Robert A. Johnston. . Travel, Emissions,
and Welfare Effects of Travel Demand Management Measures. In Transportation
Research Record 1598, TRB, National Research Council, Washington,
D.C., 1997, pp. 18-24. 21.
DKS Associates. Model
Development and User Reference Report. Sacramento Area Council of
Governments, Sacramento, CA, October 1994. 22.
Small, Kenneth A. and Harvey
S. Rosen. Applied Welfare Economics with Discrete Choice. Econometrica,
Vol. 49, 1981, pp. 105-130. 23.
Delucchi, Mark A. The Annualized
Social Cost of Motor Vehicle Use in the U.S., 1990-1991: Summary.
Institute of Transportation Studies, University of California, Davis,
June 1997. Table 1
2015 Scenarios for the Sacramento Region: Daily Vehicle Travel
Projections
a vehicle hours of delay are vehicle hours traveled under
congested speeds minus
vehicle hours of travel under free flow speeds on the same
facility. b figures in parentheses are percentage change from the
base case scenario. Table 2
2015 Scenarios for the Sacramento Region: Daily Mode Share
Projections
a figures in parentheses are percentage change from the
base case scenario. Table 3
2015 Scenarios for the Sacramento Region: Daily Emissions
Projections
a figures in parentheses are percentage change from the
base case scenario. Table 4
2015 Scenarios for the Sacramento Region: 1995 Present Value of
the Change in Consumer Welfare (with Preliminary Capital and O&M
Costs), from the Base Case Scenario
Table 5
2015 Scenarios for the Sacramento Region: 1995 Present Value of
the Change in Consumer Welfare by Income Class (with Preliminary Capital
and O&M Costs), from the Base Base Case Scenario
|
