07 Congestion Lewis

8/3/2019 07 Congestion Lewis

1/40

HAMILTONTHE

PROJECT

Advancing Opportunity,

Prosperity and Growth

The Brookings Institution

Americas TrafcCongestion Problem:Toward a Frameworkor Nationwide Reorm

D I S C U S S I O N P A P E R 2 0 0 8 - 0 6 J U L Y 2 0 0 8

David Lewis


2/40

The Hamilton Project seeks to advance Americas promise o

opportunity, prosperity, and growth. The Projects economic

strategy reects a judgment that long-term prosperity is best

achieved by making economic growth broad-based, by

enhancing individual economic security, and by embracing a

role or eective government in making needed public

investments. Our strategystrikingly dierent rom the

theories driving economic policy in recent yearscalls or fscal

discipline and or increased public investment in key growth-

enhancing areas. The Project will put orward innovative

policy ideas rom leading economic thinkers throughout the

United Statesideas based on experience and evidence, not

ideology and doctrineto introduce new, sometimes

controversial, policy options into the national debate with

the goal o improving our countrys economic policy.

The Project is named ater Alexander Hamilton, the

nations frst treasury secretary, who laid the oundation

or the modern American economy. Consistent with the

guiding principles o the Project, Hamilton stood or sound

fscal policy, believed that broad-based opportunity or

advancement would drive American economic growth, and

recognized that prudent aids and encouragements on the

part o government are necessary to enhance and guide

market orces.

HAMILTONTH E

PROJECT


3/40

JULY 2008

HAMILTONTHE

PROJECT

Advancing Opportunity,

Prosperity and Growth

Thi dicuion paper i a propoal rom the author. A emphaized in The Hamilton Project

oriinal tratey paper, the Project a deined in part to provide a orum or leadin thiner

acro the nation to put orard innovative and potentially important economic policy idea

that hare the Project road oal o promotin economic roth, road-aed participation

in roth, and economic ecurity. The author are invited to expre their on idea in dicuion

paper, hether or not the Project ta or adviory council aree ith the pecic propoal.

Thi dicuion paper i oered in that pirit.

Americas Trafc Congestion

Problem: Toward a Frameworkor Nationwide Reorm

David LeiHDR Corporation


4/40

AMERICAS TRAFFIC CONGESTION PROBLEM: TOWARD A FRAMEWORK FOR NATIONWIDE REFORM

2 THE HAMILTON PROJECT|

THE bROOkINgs INsTITUTION

Copyriht 2008 The brooin Intitution

Abstract

A large and growing burden on the nations economy, trac congestion arises or various

reasons, and more than one mechanism is needed to combat it. It is most unlikely, howev-

er, that serious inroads to address the problem will be made without undamental reorm

in the way consumers are charged or their use o congested highways. Congestion pricesare tolls that refect the economic costs o congestion, including productivity losses rom

trac delays, increased accidents, higher emissions, and more. Congestion prices would

help reduce these economic costs and guide transportation investment resources to their

highest and best usewhich would include a better balance between highway and transit

investment. In addition, such prices would generate revenues to help nance new invest-

ment and compensate low-income people and others or whom toll payments are especial-

ly burdensome. Requiring ederal, state, and local engagement, such reorm is a necessary

step in the development o an eective, ecient, and sustainable highway system or the

twenty-rst century.


5/40


www.HAMILTONPROJECT.ORg|

JULY 2008

Contents

1. Introduction 5

2. The Prolem 7

. A Propoal or Conetion Pricin 14

4. Eect, benet, and Cot o Reorm 19

5. Ditriutional Impact and Method o Redre 24

6. Quetion and Concern 0

7. Concluion 4

Reerence 5


6/40


7/40



JULY 2008 5

1.0. Introdction

M

obility, and the transportation inra-

structure needed to enable it, is ounda-

tional to American culture and econom-ic activity. In 1860, President Lincoln campaigned

on the importance o internal [inrastructure]

improvements. Almost a century later, President

Eisenhower spearheaded construction o the in-

terstate highway system. To nance it, he created

the Highway Trust Fund. The latter years o the

twentieth century witnessed a continuous series o

innovative legislative initiatives in Congress to a-

cilitate mobility as a vital national priority. In 2005,

President George W. Bush signed into law the Sae,

Accountable, Flexible, Ecient TransportationEquity Act: A Legacy or Users (SAFETEA-LU).

With guaranteed unding or highways, highway

saety, and public transportation totaling $244.1

billion, SAFETEA-LU represents the largest sur-

ace transportation investment in the nations his-

tory. Two landmark bills preceded SAFETEA-LU:

the Intermodal Surace Transportation Eciency

Act o 1991 (ISTEA) and the Transportation Eq-

uity Act or the 21st Century (TEA-21). Between

them, these bills introduced important new saety

programs, designated the national highway systemto extend ederal nancial assistance to principal

roads beyond interstate highways, extended the eli-

gibility o dedicated Highway Trust Fund revenues

to transit projects, and provided fexibility or states

and localities to employ innovative methods o -

nance and congestion management.

Despite the nations history o sustained investment

to create and maintain what is the most extensive

system o roads and bridges in the world, our mo-

bility and economic productivity is being eroded bytrac congestion at an alarming rate. Suburbaniza-

tion and urban sprawl continue apace (Lewis and

Williams 1999). Public transit attracts less than 10

percent o total passenger trips (Transportation

Research Board [TRB] 2006). Although the digital

revolution enables twenty-rst-century industry

to adopt just-in-time production, distribution, and

inventory management systems, the clogged twen-

tieth century transportation system is not up to the

task o enabling the ast and reliable just-in-timedeliveries on which such systems depend in order to

deliver enhanced productivity and competitiveness.

Although some congestion is a blessingan indica-

tor o vibrant economic and social activitybeyond

a certain point, the delay and uncertainty people

and goods endure in trac jams constitute net eco-

nomic and social burdens.

Compounding the congestion problem is the reality

that the nations primary source o unds or inra-

structure investmenttaxes on gasolineis dwin-dling. In the rst quarter o 2008, the average state

gasoline tax was $0.214 per gallon, plus $0.184 per

gallon o ederal tax, making the total tax $0.47 per

gallon. For diesel, the average state tax was $0.292

per gallon, plus $0.244 per gallon ederal tax, mak-

ing the total tax $0.536 per gallon. For much o the

twentieth century, gas taxes, with occasional modest

increases, were sucient to keep pace with inra-

structure costs. But the advent o more uel-e-

cient vehicles, infation in the cost o inrastructure

construction and maintenance, and resistance tohigher taxes is changing that situation. Unless there

is a sharp increase in gasoline tax rates, the Highway

Trust Fund is projected to be bankrupt by the end

o 2009. A recent ederal commission reports that

the gas tax would need to be increased by $0.25 to

$0.41 per gallon in order to close the gap between

projected inrastructure requirements and available

unds (National Surace Transportation Policy and

Revenue Study Commission 2007).

To ocus discussion o Americas congestion prob-lem on the shortage o gas tax revenues, however,

is to misdirect attention rom the more undamen-

tal issue. In the progressively more complex and

dynamic twenty-rst-century economy, America

needs not just more inrastructure investment, but

better investment, in the right amount, at the right

locations, and in the right balance between roads


8/40




and transit. The ocus o this paper is on the need to

reorm the way we charge or the use o congested

roads as a means o guiding inrastructure invest-

ment dollars to their highest and best use; this paper

will not examine the way we pay or inrastructure.

Instituting congestion prices (tolls that refect the

true economic cost o using congested roads) would

provide a powerul incentive to shit travel away

rom the peak hours, encourage greater use o mass

transit, reduce pressure to expand highway capacity,

and direct investment dollars to highway and public

transit projects o the highest value. In act, charg-

ing people the true cost o using congested roads

would help raise the money needed to pay or such

projects and could compensate low-income people

who are disadvantaged by the advent o road charg-es. It would be a win-win-win proposition: better

investment, the money to pay or it, and equitable

treatment o disadvantaged groups.

To augment the motivation or state and local gov-

ernments to implement congestion pricing, this pa-

per recommends that Congress (through legislation)

and the executive branch (through implementation

o regulations) redesign the way in which ederal

highway grants are established or projects on cer-

tain new and existing roads. Under one version o

this approach, designated projects to be undertaken

without the coincident introduction o congestion

pricing would be eligible or less than the high-est ederal nancial match that would otherwise

be allowable. Rather than diminish the allowable

match, another version o this approach would o-

er a higher match, a premium, or designated types

o projects implemented with congestion pricing.

A dierent ederal match would be in recognition

o the extra burden o highway maintenance and

congestion costs that untolled roads create by stim-

ulating excess demand, delays, and environmental

emissions during peak periods. States and localities

could opt out o the congestion pricing incentive,but would need to weigh the advantages o doing so

against incentivized ederal unding as well as the

loss o access to the signicant revenues available

rom congestion pricing. This proposal also in-

cludes mechanisms by which to mitigate the eects

o congestion pricing on low-income people.


9/40



JULY 2008 7

2.0. The Problem

R

esearch unded by the Federal Highway Ad-

ministration (2005) and conducted by the

Texas Transportation Institute (TTI) reportsthat urban trac congestion in 2005 caused the av-

erage peak-period traveler to spend an extra thirty-

eight hours o travel time and consume an addi-

tional twenty-six gallons o uel a year, amounting

to a cost o $710 per traveler (TTI 2007a). These

statistics are part o a worsening trend. In all such

places,

trips take longer,

congestion exists during more o the day,

congestion aects weekend travel and ruralareas,

congestion aects more personal trips and

reight shipments, and

trip travel times are increasingly unreliable.

I current policies and trends continue, analysts

oresee that by 2030 as many as eleven additional

urban areas could reach or exceed todays level o

congestion in Los Angeles, the nations most con-

gested urban area (Hartgen and Fields 2006). The

average traveler in Los Angeles spends an estimat-ed seventy-two hours a year stuck in trac (TTI

2007a).1

Although population growth, inexpensive down-

town parking, urban sprawl, and inadequate inra-

structure investment are correctly cited as causes o

congestion, the way we chargeor, rather, the way

we do not chargeor the use o roads and bridges

is in act a central cause. To appreciate the point,

consider the role that prices play in enabling the

economy in general to unction without persistentshortages and queues. The price o a good or a

service signals to prospective consumers the true

economic cost o using up the scarce resources

required in supplying it (the capital, the labor, andso on). Based on personal tastes and preerences,

and within the limits o their disposable income,

consumers establish their willingness, or unwilling-

ness, to pay the true cost by sizing up whether or

not the benets they will enjoy make paying the

price worthwhile. I people are not willing to pay

or all the washing machines available, or example,

producers resources are quickly shited into the

supply o other goods and services or which people

are willing to pay.

In this way, prices guide consumers to make mil-

lions o individual cost-benet decisions every day,

and thereby bring about the allocation o resources

that achieves, more or less, an eciently unction-

ing economy. In short, prices send cost signals to

consumers who, through the benet-cost choices

they make o what to buy and what not to buy, sig-

nal back to producers how to deploy resources and

avert persistent shortages, queues, and surpluses.

Apart rom a handul o places around the country,there are no roadway prices to signal consumers

about the real economic cost o their decisions to

travel during congested times o day. It should be

no surprise, thereore, that we witness an apparent

shortage o road space yet little use o public transit.

In deciding when and how to travel, people cer-

tainly take into account their private costs, such as

gas, oil, insurance, and so on. They also consider the

congestion they expect to encounter. Travelers do

not, however, consider the costs their trips impose

on otherswhen they add to the congestion (Mohring1999). These costs are external to peoples trip-mak-

1. By many measures, Los Angeles aces the most severe trac congestion in the nation. According to the most recent Urban MobilityReport rom the TTI, drivers in the greater Los Angeles region lost about 490 million hours due to congestion delays in 2005. About62 percent o the lane miles in greater Los Angeles were congested during the peak period, and about 86 percent o peak period traveloccurred in congested conditions. TTI estimates the cost o congestion in the greater Los Angeles regionin terms o wasted time andwasted uelat about $9.3 billion dollars annually. This represents a more-than-ourold increase rom 1985, when the annual cost (incurrent dollars) was estimated to be $2.2 billion.


10/40




ing decisions; economists thus call them external

costs.

They include the economic value o time wasted

in delayed and unreliable conditions, the extra gas

and other vehicle operating costs o stop-and-godriving, and the environmental damage and related

costs to human health. Although studies dier in re-

lation to denitions and methods, recent estimates

o external roadway costs vary rom $0.13 per ve-

hicle mile to $0.29 per vehicle mile (HDR|HLB

Decision Economics 2005; see also Small and Ve-

hoe 2007, Chapter 3). Compared with the private

costs o driving (about $0.52 per vehicle mile based

on AAA 2007), external costs thus exceed private

costs by some 25 to 56 percent. I the price o any

other good or service were set so ar below its cost,it would surprise no one to nd that its demand

routinely outstripped its supply and that there

would be very low demand or substitutes.

Time spent in trac jams is the maniestation o

roadway supply alling short o the demand or

travel. Delay is an economic cost because it means

less time available or productive work as well as

or nonwork activities that people value. Moreover,

unreliable conditionswide day-to-day variation

in the time needed to drive rom Point A to PointBlead people to guard against the risk o being

late or work and appointments by leaving early.

This time spent is at the expense o yet more time

or productive work, as well as more time at home

in the morning or amily or other personal busi-

ness.

For trucks, unreliable transit times are o special

signicance because o just-in-time penalties built

into many delivery contracts. A pattern o late de-

liveries or the receivers o goods can lead them tobear the cost o holding extra inventoriesshock

stocksto guard against the risk o material short-

ages in just-in-time production systems (Shirley

and Winston 2004).

In 2005, autos and trucks lost an estimated 4.2 bil-

lion hours to trac delays and to the eects o cush-

ioning against the risk o being late. The monetaryequivalent value o these losses, when combined

with the 2.9 billion gallons o uel wasted in stop-

and-go conditions, amounted to an estimated $78

billion lost during that year.2 Even with the exclu-

sion o environmental costs, $78 billion equates to

105 million weeks o vacation and 58 ully loaded

supertankers (TTI 2007b).

While statistics on the nationwide eects o

congestion are indicative o its importance as a

problem o national strategic signicance, the im-pacts o congestion on people and their well-being

are elt locally. A recent analysis o trac in New

York City nds that, even ater allowing or some

congestion as part and parcel o a vibrant economy,

congestion there has passed the tipping point

(Partnership or New York City 2006), stripping the

metropolitan economy o more than $13 billion a

year, including about $6 billion in wasted time and

workday productivity.

The study reports that shippers who rely on pre-dictable pickups and deliveries in order to maintain

low inventory costs (and to obtain value rom their

investments in just-in-time technologies and busi-

ness processes) hold costly shock stocks that reduce

productivity and competitiveness. Trucking rms,

which incur nancial penalties or late deliveries,

cushion against the risk o such penalties by leav-

ing earlier than they would under more reliable and

predictable travel time conditions, thereby reduc-

ing their productivity and competitiveness.

Congestion imposes an economic burden on a wide

range o industries. Those directly aected by con-

2. Economists assign monetary-equivalent value to time based on actors such as wage rates (which refect the value o productive work done)and peoples willingness to pay to save time or nonwork purposes, including commuting and leisure activities. Recent studies have revealedthat the rate at which road users value reliable andpredictable journey times actually exceeds the rate at which they value improvements inaverage journey times almost threeold (Small, Noland, Chu, and Lewis 1999).


11/40



JULY 2008 9

TABLE 1

The Cost Brden o Congestion on Indstry in the New York City Region (selected sectors)

Indstry and

type o congestion CBD Rest o

cost eect Manhattan Manhattan Bron Brooklyn Qeens Richmond Nassa

Retail indstry

Reduction in revenue,

Us$ million/year $99.5 $8.5 $1.2 $7.0 $7. $4.7 $20.9

Increae in operationalcot, Us$ million/year $66.5 $7.8 $4. $15.6 $15.9 $4.8 $17.0

Reduction in employment,

FTE/year 41 5 -5 29 0 19 87

Restarants


Us$ million/year $214.7 $7.4 $12. $22.0 $14.5 $1.0 $8.4

Increae in operational

cot, Us$ million/year $5.0 $0. $0.1 $0. $0.5 $0.1 $0.4

Reduction in employment,FTE/year 2,054 71 -117 -210 -19 -9 80

Arts & entertainment


Us$ million/year $181.7 $0.5 $11.6 $2. $21.7 $.0 $2.

Reduction in employment,FTE/year 1,402 4 -89 -179 -167 -2 18

Health care & social services


Us$ million/year $152.5 $26. $14.1 $44.7 $22.8 $7.7 $2.5

Reduction in employment,FTE/year 1,626 280 151 477 24 82 250

Constrction

Reduction in revenue,Us$ million/year $280.6 $15.1 $28.8 $92.8 $20.9 $18.6 $78.5


cot, Us$ million/year $4.1 $1.8 $.5 $11. $24.8 $2. $9.5


Manactring


Us$ million/year $488. $.2 $24.6 $12.8 $159.2 $.6 $91.7


12/40





cot, Us$ million/year $59. $0.4 $.0 $16.1 $19. $0.4 $11.1


Wholesale

Increae in cot,Us$ million/year $688.1 $4.0 $17.0 $61.8 $70.0 $2.9 $52.4

Total revenue/loUs$ million/yeara $1,62.5 $61.7 $4.8 $26.4 $62.9 $2.0 $20.7

Increae in operatin

cot, Us$ million/yeara $852.9 $14.4 $27.8 $105.0 $10.5 $10.5 $90.4

Total jo lota 8,717 466 161 1,060 1,475 160 1,145

source: Partnerhip or Ne Yor City 2006.

Note: CbD = central uine ditrict; FTE = ull-time equivalent jo. Revenue and employment numer ith a neative in indicate ituation in hich conetion

caue revenue or employment to increae due to the reditriution o trac.a. Total refect more ector than hon in the tale.

gestion include the retail trades, restaurants, health

care and social services, construction, manuactur-

ing, wholesale trade, taxis, nancial and proes-

sional services, the services and repair industry, and

or-hire trucking. Table 1 summarizes the estimated

cost burdens borne by a selection o these sectors.

The impact o congestion on the retail, restau-

rant, entertainment, and other consumption-based

trades, or example, stems partly rom a reduction in

trips or consumption purposes. By increasing thecost o traveling to such destinations, congestion

deters some consumers rom using those services

and causes others to use them less oten than they

otherwise would. As a result, retailers earn less rev-

enue and employ ewer workers. Congestion also

adds to the logistics costs o retailers by reducing

the reliability o delivery times or merchandise

and supplies. This adds to costs by inhibiting the

adoption o inventory-saving and other productiv-

ity-enhancing strategies. Congestion imposes costs

on the nancial and proessional services industries,

due (inter alia) to the time spent by employees in

highly congested conditions when traveling to

business meetings. Frequently, proessional work-

ers will guard against the risk o being late or miss-

ing a meeting altogether by allowing extra time intheir travel schedules. Less congestion would make

additional time available or productive work in the

oce. In sum, the New York study nds that trac

jams in the region add millions o dollars to produc-

tion and distribution costs and erode the economy

o nearly ty-ve thousand jobs.


13/40



JULY 2008 11

2.1. Why Are Congestion Prices Rare onAmerican Roads?

Congestion pricestolls that vary by time o day

to refect the costs travelers impose on each other

when electing to use a congested roadwayarerare in America; they apply to less than 1 percent

o congested U.S. roadways. Congestion pricing is

not rare or lack o know-how or technology. In-

deed, dierent approaches and technologies have

been developed or a wide range o dierent cir-

cumstances. As shown in Box 1, tolls can be applied

over an entire geographic network, on particular

roads only, or on particular lanes.

Nor is congestion pricing rare or lack o successul

examples. Consider the SR-91 express lanes acilityin Caliornia. Opened in 1988, it is a our-lane, ten-

mile toll road built in the median o the Riverside

Freeway on the line between Orange and Riverside

counties, and the Costa Mesa Freeway (SR-55).

Users o SR-91 express lanes pay tolls rom pre-

paid accounts using a transpondera pocket-sized

radio transmission device mounted to the inside o

the vehicles windshield. This electronic toll col-

lection technology eliminates the need or travel-

ers to stop and pay tolls at traditional tollbooths,

thus helping acilitate the fow o trac on tolledlanes. One-way tolls or the ten-mile stretch vary

rom $1.20 during o-peak periods to as much as

$10.00 or travel during the busiest times o day.

As shown in Figure 1, SR-91 results are impres-

sive, with the priced lanes generating considerably

higher speeds than the ree lanes.

BOx 1

Types o Congestion Prices Available toStates and Localities

Areawide charges: Charges based on conges-tion level on all congested roads within a geo-

graphic area. Some believe this approach to be

the most eective means o reducing congestion

and vehicle emissions.

Variable charges on particlar roadways:

Tolls administered to both roads and bridges,

including rush hour ees on acilities that cur-

rently are toll ree. Examples include 407 ETR

(Toronto), Sanibel Bridge (Lee County, Florida),

New Jersey Turnpike, and the Port Authority oNew York and New Jersey interstate crossings.

Variably priced lanes and managed lanes:

Variable tolls implemented on separated lanes

within a highway, such as express toll lanes or

high-occupancy toll (HOT) lanes. Examples

include Interstate 15 (I-15), State Route 91 (SR-

91), and I-680 (under development) in south-

ern Caliornia; I-10 Katy Freeway (Houston);

I-394 (Minneapolis); I-25 (Denver); and SR-

167 (Washington state, our-year pilot).

Cordon charges:Charges to drive within or into

a congested area within a city. Examples include

London (England), Stockholm (Sweden), Sin-

gapore, Oslo, Trondheim, Bergen (Norway),

San Francisco (Caliornia; study), and New York

(New York; proposal).

Zonal charges: Similar to cordon charging,

but with adjacent charging zones. Examples in-

clude trials in Trondheim (Norway), Helsinki(Finland), and Copenhagen (Denmark).


14/40




London (England) provides another example. A

ee introduced in 2003 requires motorists to pay

8.00 (about $16.00) to drive into central London

on weekdays between 7:30 a.m. and 6:30 p.m. Some vehicles are exempt, including licensed taxis, mo-

torcycles, vehicles used by people with disabilities,

certain alternative uel vehicles, buses, and emer-

gency vehicles. Since introduction o this ee, au-

tomobile trac in central London has declined by

a reported 20 percent, average trac speeds have

increased 37 percent, and peak period congestion

delays are down 30 percent or autos and 50 per-

cent or buses. Importantly, net revenues (more

than $100 million a year) rom congestion prices in

London are used to nance increased public transitcapacity. Reports indicate that people who change

their travel patterns due to the ee have transerred

to public transit (Litman 2006).3 Other people

change their routes to avoid central London, travel

outside the charging period, or use taxis. Stockholm

(Sweden) has since introduced a similar ee, with

similar results.

Congestion pricing in the United States is not rare

or lack o unding to study it, experiment with it,

or implement it: Congress rst set up a program to

assist localities with local congestion pricing initia-

tives in 1991. In addition, under the Urban Part-

nerships Program established in 2006, the ederal

government makes more than $1 billion available

to localities or congestion management initiatives.

Although the number o lane miles on which con-

gestion pricing has been introduced remains pro-

portionately very small, especially on existing roadsas distinct rom newly constructed roads, more

than orty congestion pricing programs have been

undertaken since 1991 with the support o ederal

unds.

3. The share o total trips made by car ell rom 12 to 10 percent between 2003 and 2005.

source: Environmental Deene 2007.

Average Trafc Speed on SR-91, Peak Hours Eastbound, Friday Afternoons 2004

Congested General Purpose Lanes Toll Managed Lanes

40

50

60

70

30

20

10

0

FIGuRE 1

Recent Trafc Statistics or the SR-91 Epress Lane Facility


15/40



JULY 2008 1

2.2. A Political Impasse at the State andLocal Levels

So why is congestion pricing rare? An important

reason is political impasse at the state and local lev-

els, especially when it comes to introducing conges-tion pricing on existing roads as opposed to the con-

struction o new capacity. Voters dislike new taxes.

Congestion prices are not taxes; they are prices that

mirror real economic costs. But explaining the di-

erence to voters is challenging. Compounding this

conusion is the public perception that congestion

prices constitute a double tax since existing roads

have indeed been paid or with tax dollars. Impor-

tantly, there also is concern about the risk o mak-

ing poor households poorer, perhaps even driving

some low-income households into poverty due tothe economic burden congestion tolls could place

on them and the smaller market area over which

poor householders could aord the costs o travel

to jobs.

Though understandable, these concerns, as this

paper goes on to show, are either inaccurate or re-

solvable through a range o policy approaches. And

despite the political impasse, there is signicant

evidence o emerging (latent) political support and

consensus with regard to congestion pricing. Latentsupport, on its own, appears insucient however to

overcome signicant grassroots opposition. In New

York City, a mayoral plan to introduce congestion

pricing or vehicles entering Manhattan won City

Council approval. It did however ace an uphill bat-

tle to gain grassroots political consensus and ulti-

mately ailed to obtain the necessary approval rom

the state legislature.4 An analysis o public opinion

surveys conducted in November 2007 nds that a

majority o surveys (56 percent) show support or

tolling and road pricing. Opposition was encoun-

tered in 31 percent o the surveys. Mixed results

(i.e., neither majority support nor opposition) oc-

curred in 13 percent o them (Zmud 2007). These

results show that in many parts o the U.S., a wide

gap exists between elected ocials perceptions o

what the public thinks about tolling and road pric-ing and what public opinion actually is (Peters

2008, 3). In London and Stockholm, broad-based

support has indeed been achieved, albeit ater the

politically unpopular decision to implement con-

gestion pricing had been taken and people saw the

results or themselves.

That the nations transportation inrastructure is in

trouble is not in doubt; nor is there much doubt

that ailure to charge congestion tolls is a signicant

source o the problem. We know, too, that charg-ing congestion prices is easible, that it would help

optimize the use o existing transportation acilities

in the short run, and that it would provide inorma-

tion vital to optimizing the characteristics o such

acilities in the long run (Mohring 1999). The op-

timization o investment would occur as the prices

o tolls signal which roadways have the highest level

o demand.

It seems almost certain that, with congestion pric-

ing in place, ewer trips would be made by car dur-ing busy periods, more people would use public

transit, and the allocation o investment resources

between highways and transit would better refect

the true transportation needs and preerences o

travelers relative to the costs o satisying those

needs. It is time or the ederal government to step

in to help break the state and local political impasses

that stand between congestion pricing and the re-

alization o an ecient, sustainable, and aordable

surace transportation system or the twenty-rst

century.

4. To understand the diculty in obtaining grassroots support or this proposal, see: Editorial, Reducing the cost o congestion,New YorkTimes, December 10, 2006; Editorial, Let NYC study pay-to-drive plan, Newsday.com, December 8, 2006; Clearing the air on tracproblem, Crains New York Business, December 10, 2006.


16/40




3.1. Federal Reorm

Notwithstanding the apparent political risk

taken to introduce congestion pricing in

London and Stockholm, making politically

unpopular decisions in the hope that broad-based

support will ollow is rare in the development o

transportation policy. Nevertheless, the evidence

and analysis outlined above indicates the latent po-

litical acceptability o congestion pricing. A key to

overcoming the political impasse at the state and

local levels o government lies in policy innovation

at the ederal level. To be sure, Congress and theexecutive branch have already displayed signi-

cant leadership with innovative programs designed

to promote and encourage local experimentation

with, and implementation o, congestion pricing.

The next step is or Congress to create powerul

incentives that make the adoption o congestion

pricing widely compelling at the local level.

Congresss ability to establish such incentives ex-

ists by virtue o the ederal-state relationship with

regard to highway inrastructure investment. Al-though the execution and administration o trans-

portation policy, planning, and investment in the

United States belongs with state and local govern-

ments, the ederal government plays a signicant

nancial role. The interstate highway system and

many other primary roads have been built and

maintained with 90 percent ederal and 10 percent

state and local unds. The signicant ederal nan-

cial role in transportation brings with it a great deal

o leverage over policy and planning at the state and

local levels. In return or ederal highway dollars,the ederal government mandates planning require-

ments, environmental impact analyses, saety stan-

dards, restrictions on the size and weight o trucks

that are allowed to use the roadways, and a range o

other conditions that Congress and the executive

branch deem to be in the national interest.

Against this background, the main policy attributes

o the existing ederal approach to congestion pric-ing are permissive and acilitating, but certainly not

mandatory. One option going orward would thus

be to add congestion pricing to the range o man-

datory requirements or state and local receipt o

ederal nancial assistance. A mandatory approach

would be clear and straightorward to administer

but would run counter to the trend in ederal policy

o seeking to grant fexibility to states and localities

to innovate and choose across the widest possible

array o technological, planning, nancial, and pro-

curement mechanisms.

3.1.1 Proposal: Congestion Pricing Financial

Incentive Program

An alternative to mandatory application o the con-

gestion pricing approach would be to redesign the

way ederal highway grants are established to create

a choice and an incentive or localities to introduce

congestion pricing in association with projects in

highly congested urban areas and congested inter-

city routes. Under this approach, construction and

major reconstruction projects with designated at-tributes would be eligible or less than the other-

wise highest allowable ederal match i they are to

be undertaken without the coincident introduction

o congestion pricing. The reduction in allowable

match would be determined by a method or or-

mula that recognizes the external congestion costs

that untolled roads create by stimulating excess

demand and corresponding increases in delay, ve-

hicle operating expenses, and environmental and

accident costs. One practical way in which to es-

tablish such a ormula would draw on analyticalevidence regarding the extent to which conges-

tion pricing, by diminishing demand, reduces the

cost o highway maintenance and expansion. For

example, according to model-based scenario analy-

ses by the Federal Highway Administration (2006),

applying congestion tolls to all congested roads in

3.0. A Proposal or Congestion Pricing


17/40



JULY 2008 15

the national highway system could reduce the cost

to maintain the system by about 27 percent (U.S.

Department o Transportation [DOT] 2006). Tak-

ing this percentage as a hypothetical estimate (the

DOT emphasizes the very preliminary nature o

the estimate) o the amount by which the absence ocongestion pricing increases the burden o highway

inrastructure investment on ederal resources, the

ederal matching ratio would be reduced propor-

tionately: an 80 percent match would be reduced to

about 58 percent, a 75 percent match to 55 percent,

and so on. States and localities would still be ree to

choose the lower match and the option o not insti-

tuting congestion pricing, but would need to weigh

the advantages o abstaining against the reduction

in ederal unds, as well as the loss o access to the

revenues rom congestion pricing.

An alternative to creating an incentive by reducing

the base ederal matching ratio or projects without

congestion pricing would be to institute a higher

match instead. This might be viewed as less disrup-

tive to the present equity characteristics o the ed-

eral program. An approach that lies between these

two alternatives is also possible, o course.

In addition to the incentive outlined above, this

proposal calls or Congress to direct the DOT andthe IRS cooperatively to design a model template

or a progressive reundable mobility tax credit

(PRMT) program or states and localities to adopt

in conjunction with congestion pricing. Since the

allocation o toll revenue between tax credits, direct

rebates, and inrastructure investment must be de-

termined and administered locally to align with local

circumstances and preerences, the model template

would be a guide to states and localities or them to

implement as they see t. As such, the PRMT tem-

plate design would provide wide local latitude with

regard to the choice o income thresholds, program

qualications, and other program parameters. At

the same time, however, the PRMT model would

provide states and localities sucient technical and

administrative specicity, and any necessary ederalauthorities, to acilitate ull-scale implementation.

In ollowing this recommendation, Congress can

help ensure that states and localities have the means

to put in place administrative machinery through

which to help protect low-income individuals and

other disadvantaged groups. The PRMT would be

nanced rom a portion o toll revenues and thus

would be scally neutral.

A considerable amount o work is needed to trans-

late the hypothetical approach given above into apractical basis or policy. As discussed next, Con-

gress should direct the DOT to establish the exact

method or ormula by which dierential matching

ratios are to be established. The DOT would also

determine the class o projects to which the incen-

tive plan will apply. The Internal Revenue Service

(IRS), in collaboration with the DOT, would be

directed to develop the model PRMT. Congress

would remove ederal prohibitions on the applica-

tion o congestion pricing to existing roadways.5

3.1.2 Enactment: Congestion Pricing Financial

Incentive Program

Due or reauthorization in 2009, SAFETEA-LU

(see 1 above) is the principal legislative mecha-

nism through which Congress establishes national

transportation law. As part o the reauthorization

process, Congress should establish that

by 2020, a ederal nancial congestion pricing

incentive program is in place or a designated

5. Importantly, such a prohibition presently applies to the interstate highway system. Some steps in the direction o removing such restric-tions have been taken already. The Interstate Highway System Construction Toll Pilot Program authorizes up to three acilities on theinterstate system to toll or the purpose o nancing the construction o new interstate highways. A state or an interstate compact o statesmay submit a single candidate project under this program. Each applicant must demonstrate that nancing the construction o the acilitywith the collection o tolls is the most ecient and economical way to advance the project. The state must agree not to enter into a non-compete agreement with a private party, under which the state would be prevented rom improving or expanding the capacity o publicroads in the vicinity o the toll acility to address conditions resulting rom trac diverted to nearby roads rom the toll acility. There is nospecial unding authorized or this program. Interstate maintenance unds may not be used on a acility or which tolls are being collectedunder this program.


18/40




category o highway projects associated with

both new and existing capacity; and that

by 2015, the DOT will have established regula-

tions and guidelines enabling states and locali-

ties to begin planning or 2020 implementation;and

by 2015, the IRS will have designed the pro-

gram template or a PRMT, and put in place

the necessary authorities to enable its imple-

mentation by states and localities.

3.2. Reglatory Direction to the DOT and the

IRS

In enacting the congestion pricing incentive program,

Congress would direct the DOT to promulgate bySeptember 2013 a Notice o Proposed Rulemaking

establishing the specic planning and implementa-

tion requirements or the Congestion Pricing Finan-

cial Incentive Program. As a means o indicating con-

gressional intent, ederal agencies would be directed

to address seven matters through regulatory and

administrative action.

1. The DOT would dene the method or ormula

by which dierential matching ratios are to be

established. The DOT would also dene themanner in which incentive program unding

relates to specic projects within the Statewide

Transportation Improvement Plans (STIP) that

provide the basis or ederal unding approvals.

It is not contemplated here, or example, that

matching unds or the entire STIP would hinge

on the treatment o projects that pertain only to

congested roads.

2. The DOT would dene the attributes o project

applications that would deem them to be subjectto the nancial incentive program. Attributes to

be considered would include the extent o exist-

ing congestion, and the extent to which the ap-

plicant road or (roads) would provide indepen-

dent utility as a congestion-priced acility.

3. The DOT would also establish principles and

guidelines regarding the level o congestiontolls with due regard or both the economic

cost o congestion and the eect o such tolls on

the diversion o trac to unpriced roads. The

DOT should enable and encourage project ap-

plicants to set balanced rates, with due regard

or evidence that tolls set to mirror the ull cost

o congestion can risk diverting so much trac

to unpriced routes that the aggregate economic

costs o travel over the entire network would be

greater than those roads with no tolls.6

4. The DOT would ensure both that states and lo-

calities provide reasonable alternatives to priced

roads, and that they apply a stipulated minimum

or reasoned percentage o the revenues rom

congestion pricing to monetary reimbursement

or disadvantaged groups and investment in pub-

lic transit. Others point out that the demand or

transit, which is likely to rise signicantly in some

localities with the advent o congestion pricing,

will automatically reveal the appropriate extent

o new investment and generate sucient rev-enues to nance it. The DOT rulemaking needs

to strike a balance between such approaches,

while leaving maximum easible fexibility or

local choice and innovation (see 5, this paper).

5. The DOT would also provide a ramework with-

in which states and localities are to adopt com-

mon technology platorms or toll collection to

ensure regional and national interoperability in

the use o congestion-priced highways. Similarly,

automobile and truck manuacturers should begiven rules by which to make provisions or all

new automobiles and trucks sold ater January

6. Mohring (1999), or example, reports that congestion tolls on expressways in the Twin Cities would need to be set at about 25 percent othe ull economic cost o congestion to ensure that spillover trac would not cancel the eciency gains o congestion pricing. He alsoreports that Singapore overdid it: Congestion outside the cordon was so great that, despite ree fow within it, travel times per bus orauto trip to central area destinations did not change (194).This problem has not been experienced in the case o Londons cordon pricingprogram (see 2.1, this paper).


19/40



JULY 2008 17

2015 to be equipped with onboard electronic de-

vices compatible with the common platorms to

be employed at the state and local levels.

6. For roads nanced with a combination o ederal

and private sector nancing (i.e., public-privatepartnerships), the DOT regulations would stipu-

late the requirement that sucient revenues be

reserved or compensation programs and that,

where revenues rom congestion pricing lead to

private sector prots that exceed economic rates

o return, such excess revenues are to be made

available or public reinvestment.

7. The IRS and the DOT would collaboratively

develop the detailed model template or PRMT,

including the means by which to set incomethreshold provisions, eligibility qualications,

administrative procedures, and ederal authori-

ties to enable states and localities to implement a

program with local discretion as to actual income

cut-os and other program criteria.

3.3. General Applicability o the Proposal

Although we cannot orecast the take-up o the

proposed congestion-pricing nancial-incentive

program, we can examine its scope o application

under book-end conditions. I the proposal were tolead to congestion pricing on all roads with conges-

tion above a 70 percent volume-to-capacity ratio,

pricing would apply to 15.3 percent o all road mile-

age (including interstates, other reeways, arterials,

and collectors) and cover 41.1 percent o all vehicle

miles traveled (at 2005 trac levels). The more-de-

tailed perspective on road mileage in Table 2 and

on travel in Table 3 indicates a similar pattern. Just

22.3 percent o the interstate highway systems road

mileage is seriously congested (with the volume-to-

capacity ratio exceeding 95 percent), but these roadshandle nearly 40 percent o vehicle miles traveled

on the interstate system (see nal column o Table

3).

3.4. State and Local Reorm

The ederal nancial incentive program outlined

above provides or a ederal policy ramework and

regulatory oundation, but leaves much to be done

at the state and local levels. Importantly, there is

no need to deer the implementation o congestionpricing programs to the 2020 deadline, especially

on congested roads that are not part o the ederal

system. In addition,

states with legislative prohibitions against the

implementation o tolls that might wish to take

advantage o the incentive program need to take

steps to remove such prohibitions;

states and localities need to begin now to evalu-

ate alternative congestion pricing mechanismsand to establish those o relevance and best value

to their various local and regional circumstanc-

es, taking ull advantage o ederal programs

designed to assist in that endeavor;

states and localities need to begin now to

engage the general public, stakeholder groups,

and community opinion leaders regarding the

nature o congestion pricing and the kind o

opportunities and issues entailed in the ederal

incentive program; and

states and localities need to begin now to assess

the range o ways and means by which to help

mitigate the negative eects o congestion pric-

ing on disadvantaged groups, partly through the

development o a PMRT (see 5).


20/40




TABLE 2

Road Mileage by Level o Congestion (as Measred by Volme-to-Capacity Ratio or urban Roads),2005

Road type Percentage o road length mile in road category by VC Percent ototal road

Le than 0.710.79 0.800.95 greater mile y Percent o0.71 than unctional coneted

0.95 cateory road mile

Intertate 48.1 10.8 18.8 22. 5. 17.1

Other reeay 6.0 8.0 1.5 15.5 .6 8.2

Other principal arterial 81. 6. 7.5 4.9 20.9 24.4

Minor arterial 86. 4.2 4.5 5.0 4.4 29.5

Collector 90.7 2. 2.8 4.1 5.9 20.8

source: Author calculation uin Federal Hihay Adminitration 2005.

TABLE 3

Share o Travel (Measred in Vehicle Miles) on urban Roadways, Categorized by Volme-to-CapacityRatio

Road type Percentage o travel volme, by volme-to-capacity ratio

greater

Le than 0.710.79 0.800 .95 than0.71 0.95

Intertate 27. 10.9 2.0 8.8

Other reeay 6.1 8.8 21.0 4.1

Other principal arterial 71.8 8.4 10.7 9.2

Minor arterial 77.9 4.7 8.6 8.9

Collector 79.8 .5 8.0 8.6

source: baed on FHwA run on hihay ection in the Hihay Perormance Monitorin sytem (HPMs) dataae.

Note: The volume o trac ued to calculate percentae hare in the lat column include local trac. Minor arterial and collector have een comined.

VC = volume to capacity.


21/40



JULY 2008 19

B

y making people aware o the ull economic

costs o their travel choices, the widespread

application o congestion pricing would en-courage roadway users to determine whether the

benets o using the road at busy times o the day

are worth the ull economic implications o doing

so. Many would continue to use the newly tolled

roads. Some would change to alternative routes,

change their schedule, switch to another mode such

as public transit, or ride a bicycle or walk. Some

would change their mind about making the trip or

perhaps combine it with another trip. Such changes

in behavior would help optimize the use o exist-

ing transportation acilities in the short run, andprovide inormation and revenue to help optimize

investment in the long run (Mohring 1999). As a

result, the nations highways and transit systems

would be more eective, more economically e-

cient, and more nancially sustainable.

4.1. EectivenessAnticipating the quantitative eects o congestion

pricing on travel and trac behavior is both ana-

lytically dicult and dependent on the degree to

which states and localities would adopt it under theproposed ederal incentive program, yet evidence

rom actual experience and analytical models is

uniorm in suggesting that journey speeds, travel

times, and travel time reliability would improve.Based on a model that synthesizes various strands

o empirical evidence, the analysis reported in Table

4 indicates that average speeds on the nations most

severely congested roads might increase, on aver-

age, between 11 and 16 percent ater the introduc-

tion o congestion pricing. On the most congested

interstates and reeways, improvements in speed

would likely be even greater. These results depend

on estimates o toll rates and the elasticity o de-

mand or highway travel when those tolls cause the

cost o travel to rise.

Toll rates are calculated to refect the delay cost one

driver driving one mile imposes on all other drivers

on the same road at the same time. The delay costs

depend on the level o congestion on a given road at

a given time and the value o time lost, which I es-

timate at between $18 and $40 per hour.7 I assume

the elasticity o demand or highway travel is in the

range o 0.4 to 0.8.

While signicant in relation to current trac con-ditions, the eects suggested above are likely con-

4.0. Eects, Benefts, and Costs o Reorm

TABLE 4

Estimated Impact o Congestion Pricing on Trafc Volme and Speed on Interstates and Freeways

dring Peak Periods with Volme-to-Capacity Ratio above 0.95 (percent)

Trafc and speed Baseline elasticity and Alternative elasticity andvale o time assmption vale o time assmptions

Increae in peed +11 +16

Reduction in vehicle mile

o hihay travel 12 19

source: Adapted rom HDR|HLb Deciion Economic 2005.

Note: baeline aumption place the value o time at $18 per hour and elaticity o demand or hihay travel at -0.4. The alternative aumption place the value o

time at $40 per hour and the elaticity o demand or hihay travel at -0.8. Conetion pricin i aumed to e applied on Intertate and reeay ith volume-to-

capacity ratio aove 0.7. The etimated eect on trac peed on all Intertate and reeay ith volume-to-capacity ratio o 0.7 and aove i +7 percent and +10

percent repectively, or the to aumption cenario. The etimated eect o conetion pricin on vehicle mile o hihay travel on Intertate and reeay ith

volume-to-capacity ratio o 0.7 and aove i -10 percent and -16 percent repectively, or the to aumption cenario.

7. The wide range o values or time refects evidence that the empirically measured metric value o time is up to three times the prevail-ing wage rate when travel times are not only high but also widely variable rom day-to-day and thus especially hard or people to predict(Small, Noland, Chu and Lewis 1999).


22/40




servative, understating potential improvements in

speed and overstating the potential reduction in the

volume o travel. This is because the model does

not mirror the dynamics o trac volume and travel

time under the most severely congested conditions

(wherein at certain times tracin the absence ocongestion pricingcomes to a virtual standstill).8

This is indeed a problem with traditional analyti-

cal transportation models. Presently, a great deal o

research eort is going into improving the state o

the art o analytic models.

These results are similar to those obtained by

Mohring and Anderson (as cited in Mohring 1999)

in their simulations o congestion pricing in the

Twin Cities. They nd that putting tolls on all con-

gested roads would reduce expressway vehicle milesby 19 percent, and nonexpressway vehicle miles by

8 percent.

The discussion above pertains to the change in travel

on highways during busy periods: the change in the

total volume o travel will depend on the extent to

which those who reduce their use o highways dur-

ing peak times shit to other times o day, or to other

modesin particular to public transit. The extent

to which a decline in total travel is mitigated will

depend importantly on the use o congestion pric-ing revenues to invest in additional transit capacity.

A strong program o transit investment could go ar

toward minimizing the disruption o daily lie that

might otherwise arise with congestion pricing.

4.2. Economic Efciency Although the quantitative estimates vary widely,

economic theory, analysis, and evidence rom eld

applications point to the same thing: a more eco-

nomically ecient transportation system.

The immediate economic eciency benets o road

pricing arise in the orm o time savings to roadway

users, reductions in vehicle operating costs, ewer

collisions and related accident costs, and improved

environmental conditions. In the longer term, we

can expect less pressure to build highway capacity,

more cost-eective highway investment decisions

(due to the way prices help signal where investment

is most needed and worthwhile), and a level playingeld or transit, resulting in a better balance o in-

vestment between highway construction and public

transit.

The costs o congestion pricing include the capital

and lie-cycle expenses o toll collection and admin-

istration; and the loss o economic and social value

incurred by highway users in various categories.

Such groups include

highway users who cut back the total number ojourneys they make,

highway users who adopt new activity schedules

they nd less convenient,

highway users who switch to transit or other

modes o travel (like walking) that they preer

less than driving,

highway users who make shorter journeys than

beore,

highway users who divert to auto routes they

preer less because they are more circuitous or

inconvenient because o intersections and traclights, and

highway users who experience increased conges-

tion on roads to which people divert in order to

avoid tolls.

Even a partial analysis o the benets o conges-

tion pricing in relation to a more comprehensive

examination o costs indicates a strong likelihood

o a quantitatively signicant gain in economic e-

ciency. Employing the same model and assumptions

as those used above in assessing trac impacts inTable 4, Table 5 compares the estimated economic

benets due to time savings and reduced accidents

to the loss o value to highway users who divert to

8. This point pertains to the diculty modelers have in representing the backward-bending relationship between trac volume and traveltime during periods o hypercongestion. Indeed, the model underlying Table 4 assumes a monotonically increasing relationship betweentravel volume and travel time per trip.


23/40



JULY 2008 21

other times o day, other routes, or other modes

and the costs o toll collection and administration.

The analysis indicates net benets, over a twenty-

year period, o $113 billion. By the twentieth year

o the program, the ratio o benets to costs is an

estimated 2.7:1.0 ($2.70 in benets or each $1.00in cost).

Though not dissimilar rom other benet-cost as-

sessments o congestion pricing, the estimated e-

ciency gain (net benet) given above almost cer-

tainly understates the gain likely to ollow rom the

introduction o nationwide congestion pricing or

our reasons.9 First (and as stated above) the model

does not mirror the dynamics o trac volume and

travel time under the most severely congested con-

ditions.

Second, the gures in Table 5 do not account or

reduced uel costs, reductions in environmental

emissions, and savings that might arise rom a re-

duction in pressure to build new highway capacity

9. See, or example, Mohring and Andersons analysis o congestion pricing in the Twin Cities (in Mohring 1999).

TABLE 5

Time Savings and Accident Cost Savings rom Congestion Pricing Relative to the Loss o EconomicVale (Consmer Srpls) or those Priced O Roads (Interstates and Freeways with Volme-to-Ca-

pacity Ratio above 0.7)

Economic benefts and costs Year 1 o Year 20 ocongestion pricing congestion pricing

Benefts (u.S. billions, 2002 dollars)

1 Travel Time savin to Road Uer ho stayon the Road and Accident Cot savin $1.68 $26.84

Costs (u.S. billions, 2002 dollars)

2 Lo in Economic Value to Road Uer ho Reduce theNumer o Trip They Tae or Divert to Other Time

o Day, Other Road or Other Mode $0.70 $2.

Cot o Toll Collection $6.20 $7.60

4 Total Cot: (2)+() $6.90 $9.9

5 Net benet: (1)-(4) $6.78 $16.91

Net beneft (net present vale) over 20 years (7 percent discont rate) $113 billion

source: Adapted rom HDR|HLb Deciion Economic 2005.

Note: Time avin repreent $4.8 illion (5 percent) o total enet in Year 1 hile the ocial enet rom accident reduction account or $8.9 illion (65 percent).

by the tentieth year, time avin and accident cot avin account or 44 percent and 56 percent, repectively, o total enet. Time avin are proaly

undertated due to the implied verion o the underlyin peed-fo relation ued. Reearch i onoin in relation to thi iue.


24/40




due to the reduced peak period demand. Analy-

sis presented in the DOTs (2006) report on the

conditions and perormance o the nations high-

ways and transit systems reports that universal

congestion pricing, by improving the perormance

o current highway system, could signicantlyreduce the level o uture highway investment that

would be required to maintain or improve the

condition o our highways. The DOT report in-

dicates that applying congestion tolls to all o the

congested roads in the system could reduce the cost

to maintain the system by $21.6 billion per year, or

27.5 percent, leaving capital needs at $57.2 billion,

well below the current level o capital spending.

Third, eciency gains are probably underesti-

mated in Table 5 due to exclusion o the possibleeect o congestion pricing on the quality o u-

ture investment decisions. The proound linkage

between economically correct prices and the qual-

ity o resource allocation and investment decisions

means that congestion pricing could well give rise

to better highway and transit investment decisions

going orward. Prices send better signals to trans-

portation planners as to where capacity expansion

is most critical.

The ourth reason why the estimated twenty-yeareciency gain o $113 billion given in Table 5

might be understated is that it ignores the impact o

longer-run changes in consumer behavior, such as

the possibility that travelers would lobby employ-

ers or widespread introduction o staggered work-

hours in order to avoid congestion prices during

peak periods, and would alter the pattern o their

residential location choices in avor o shorter work

trips and higher density (less automobile-intensive)

living patterns. A study by Langer and Winston

(2008) reports as ollows:

Based on a sample o the ninety-eight largest Met-

ropolitan Statistical Areas (MSAs) in the nation, we

nd that ecient road pricing would generate $120

billion in annual revenues (2000 dollars), while re-

ducing the value o the annual fow o services rom

housing $80 billion (2000) dollars, thus generating

an annual net benet o $40 billion. Our estimate

o the benets o congestion pricing is considerably

greater than previous estimates that do not account

or adjustments in land use and represents a rst

step toward accounting ully or road pricings ben-

ets. We conclude that policymakers should rec-ognize that road pricing mitigates congestion and

improves the quality in lie in a metropolitan area

by improving land use.

Langer and Winstons estimate o $40 billion

annually in net benets rom congestion pricing

radically exceeds the estimated twenty-year net

benet o $113 billion reported in Table 5.

4.3. Eqity

While the discussion above indicates that con-gestion pricing is likely to generate a gain in the

eciency o the road system, it does not account

or the increase in driving costs that tolls impose

and the implications or individuals. In the cost-

benet analysis in Table 5, toll payments are treated

as a transer o resources rather than as a cost to

society. However, this social cost-benet analysis

does not take into account the eects o toll pay-

ments on individuals. For some the value o time

savings rom reduced congestion is greater than the

amount o the toll, but or others congestion pric-ing would leave them worse o. Proessor Robin

Lindsey states it thus:

Tolling raises drivers private costs, as indeed

it must i travel is curtailed. The revenue rom the

toll accrues to the toll-road operator, which is usu-

ally assumed to be a government agency. Unless

the government uses the revenue to expand road

capacity, to improve an alternative orm o trans-

port, to reduce other user charges, or to provide

rebates to drivers in some lump-sum ashion, driv-ers end up worse o (Lindsey 2006).

But we also know the ollowing rom Mohring:

That tolls would eliminate the deadweight [e-

ciency] losses rom unpriced congestion and lower

the time cost o still-made trips guarantees that


25/40



JULY 2008 2

increased toll revenues would exceed consumer

losses. Hence, in principle, a compensation system

or losers could be ound that would not only leave

them better o, but also provide unds or the high-

way authority to perorm good works (Mohring

1999, 186187).

In other words, while congestion pricing dispro-

portionately hurts certain drivers with low time

savings, especially low-income drivers, the govern-

ment could use the revenue rom the tolls, and oth-

er transers i necessary, to ully oset any negative

distributional impact and leave everyone better o.

I toll revenue alone is not enough to ully com-

pensate or consumer losses, the government can

use the toll revenue to at least partially oset these

eects. I discuss using revenues rom congestionpricing to compensate or consumer losses in 5.

4.4. Financial ContetProjected toll revenues rom congestion pricing

can be placed in a nancial context by compar-

ing them with DOTs estimated requirements or

highway investment over the next two decades. One

preliminary estimate o toll revenues rom apply-

ing tolls on all congested interstates and reeways

places them at about $105 billion annually (in con-

stant 2002 dollars; HDR|HLB Decision Econom-ics 2005).According to the DOTs report (2006) on

conditions and perormance, the average annual

cost to maintain highways (and bridges) or the

twenty-year period 200524 is an estimated $78.8

billion (in 2004 dollars). These gures represent

the estimated level o investment by all levels o

government required both to maintain the existing

level o bridge deciencies in constant dollar terms

and to keep the physical condition and operational

perormance o the highway system at a level su-

cient to prevent average highway user costs (in-

cluding travel time costs, vehicle operating costs,and collision costs) rom rising above the existing

level in constant dollar terms. Congestion prices

would not only provide signicant revenues to -

nance these requirements but would also reduce

the cost o these requirements by lowering demand

or highways and reducing their wear and tear. For

transit, the National Surace Transportation Policy

and Revenue Commission (2007) puts average an-

nual investment requirements in the range o $21

billion to $32 billion annually. Estimated revenues

rom congestion prices could help nance these re-quirements as well.

The DOT (2006) report also gives the average

annual maximum economic investment level or

highways and bridges or the twenty-year period

200524. This value, estimated to be $131.7 billion

(in 2004 dollars), represents the level o investment

by all levels o government required to implement

all the highway and bridge improvements judged in

the DOT model to be cost-benecial improvements

on highways and bridges. As indicated earlier, theDOT nds that congestion pricing would, by im-

proving the perormance o highways and bridges,

reduce total investment requirements. Thus even

under a maximum economic investment scenario,

the revenues rom congestion pricing are relatively

signicant.


26/40




=$100,00

0

10

0

1

2

3

4

5

6

7

8

9

AverageAnn

ualExpenditurePerHouseholdon

C

ongestionTolls,

dollars

Household Income

AverageAnnualExpenditurePerHouseholdon

Conges

tionTollsasPercentofIncome

2,500

2,000

1,500

1,000

500

0

Average Annual Expenditure Per Household on Congestion Tolls as % of IncomeAverage Annual Expenditure Per Household on Congestion Tolls, dollars

I

ntroducing congestion pricing would increase

the eciency o the road network, but additional

steps need to be taken to ensure that benets arebroadly shared. The immediate eect o congestion

pricing would be to penalize those who pay the tolls

or take ewer road trips to avoid the tolls. This e-

ect would, moreover, be regressivenamely, it

would be inversely related to the incomes o those

aected, as can be seen in Figure 2 and Table 6. 10

I states and localities eventually adopt congestion

pricing on all congested interstates and reeways(those where the ratio o volume to capacity exceeds

70 percent), Figure 2 indicates that additional out-

lays would be proportionately higher or lower-in-

come households.

10. A correlation between income and the requency o use o roads with congestion prices is evident is surveys o users o the SR-91 expresslane acility. They show that commuters in the high-income group (those earning more than $100,000 a year) are slightly more than twiceas likely as commuters in the low-income group (earning less than $25,000 a year) to be requent toll lane users (23 versus 10 percent):high-income users are about hal as likely as low-income users to be nonusers (37 versus 73 percent).

FIGuRE 2

Income Distribtional Conseqences o Congestion Pricing on Congested Roads in urbanized Areas

source: Author etimate aed on 2001 National Houehold Travel survey and an averae conetion chare o $0.25 per mile. Fiure plotted in thi ure are hon

in Tale 6.

5.0. Distribtional Impacts and Methods o Redress


27/40



JULY 2008 25

TABLE 6

Income Distribtional Impact o Congestion Charges on Interstates and Freeways with Volme-to-Capacity Ratio above 0.7

Gross annal Annal hosehold Annal hosehold Annal cost ohosehold income ependitre on ependitre on compensation or

congestion charges congestion congestion charges

per hosehold charges as a to hoseholds

percentage o

annal income

(Billions o 2007 dollars)

< $5,000 $428.55 8.6% $1.0

$5,000$9,999 $655.28 8.7% $.4

$10,000$14,999 $622.0 5.0% $2.9

$15,000$19,999 $90.49 5.% $5.

$20,000$24,999 $1,061.29 4.7% $5.1

$25,000$29,999 $1,198.90 4.4% $7.9

$0,000$4,999 $1,4.4 4.1% $5.8

$5,000$9,999 $1,50.65 4.1% $10.0

$40,000$44,999 $1,540.55 .6% $4.94

$45,000$49,999 $1,70.82 .6% $9.61

$50,000$54,999 $1,858.29 .5% $4.80

$55,000$59,999 $1,890.05 .% $8.47

$60,000$64,999 $1,871.12 .0% $.57

$65,000$69,999 $1,950.9 2.9% $6.2

$70,000$74,999 $2,24.0 2.9% $.58

$75,000$79,999 $2,146.45 2.8% $5.98

$80,000$99,999 $2,248.22 2.5% $11.

> = $100,000 $2,277.41 1.8% $18.92

Total $41.4

source: Author etimate, aed on data rom 2001 National Houehold Travel survey and aumption in the note.Note: Aumption are (1) the percentae o vehicle mile o travel (VMT) in uran area i equal to 55 percent o total VMT (aumption aed on 2001 Hihay

Perormance Monitorin sytem (HPMs) run otained rom FHwA); (2) the percentae o VMT in conetion condition exceedin VC o 0.7 i equal to 41 percent(aumption aed on 2001 HPMs run otained rom FHwA); and () the expenditure on conetion cot include a chare o $0.25 per mile or all mile driven in

coneted condition. Annual compenation i calculated a the averae expenditure on conetion chare in Column 2, multiplied y the numer o houehold in

income racet cateory. Income roupin hon refect 2001 condition, herea toll paid refect 2007 price.


28/40




One way in which to help redress the regressivity

o congestion pricing would be to reimburse house-

holds below a designated income level or some or

all o the nancial losses they incur. The third (last)

column o Table 6 estimates the annual cost o such

an approach. Since the gures in Table 6 assumeexisting travel rates (i.e., no economic losses rom

ewer trips, diversion to more circuitous routes and

so on), the costs shown in column three might be

sucient to compensate or both tolls paid by low

income people who pay them and economic losses

or those priced o. A major problem with this ap-

proach, however, is that it would destroy the incen-

tives to drive less or at less busy times. One way in

which to minimize such incentive-blunting eects

is to oer a degree o tax relie to lower income

households that pay tolls rather than compensatingthem or the ull amount o the tolls they pay. This

approach would also allow a share o the available

revenues rom congestion tolls to be invested in

transportation improvements.

Indeed, investing congestion pricing revenues in

improved alternatives to highway transportation is

an important acet o any plan to help alleviate the

distributional consequences o congestion pricing.

Options include

directing toll revenues or particular transit

projects,

establishing minimum quality standards or

alternative ree routes as a precondition to

imposing tolls, and

establishing programs to encourage rms to

permit fexible working hours, nancing the

administration o ride-matching, ride-sharing

programs, and the like.

Whereas total revenues rom tolls are likely to besucient to mitigate the most egregious social

costs, there is no economic rule by which to ascer-

tain the allocation o such revenues best suited to

alleviating problems or disadvantaged groups. The

literature oers various approaches. Small (1992)

proposes an allocation o one-third in monetary

reimbursement to trip makers, one-third to oset

general taxes presently used to und transporta-

tion services, and one-third or new transportation

services (transit, roads, or both). DeCorla-Souza

(2004) proposes a scheme wherein tolls would be

exempted or designated user groups (such as high-

occupancy vehicles [HOVs]). In his plan, 70 percento net toll revenues would be employed or roadway

and transit capital improvements, and 20 percent

o net revenue would go to cash reimbursements

against tolls and transit ares incurred by low-in-

come travelers. Goodwin (1989) proposes a division

o a one-third reduction in existing taxes; one-third

in the construction o new roads, the improvement

o existing roads, or the improvement o roadway

maintenance standards; and one-third to improve

public transport services. Goodwins suggestion is

thus that about one-third o toll revenues wouldgo to direct reimbursement while ully two-thirds

would be steered to indirect investment in mitigat-

ing social costs through improved roads and tran-

sit. Both Small and Goodwin recommend their ap-

proach as being easy or the public to understand,

and being a reasonable basis or widespread public

support and consensus. These two attributes are

also attributes o a compensation and mitigation

program to which states and localities should pay

particular attention.

Under the proposal or reorm presented in this

paper, Congress would stipulate that steps be taken

to alleviate problems or disadvantaged groups, and

would direct the DOT and the IRS to establish a

model program or state or local implementation

o a PRMT, along with specic guidelines, admin-

istrative mechanisms, and authorities or doing so.

The principal aim o the proposed congestion pric-

ing policy ramework is to introduce the incentives

needed to acilitate an eciently operating high-

way system while ensuring that congestion pricingprogram is as equitable as possible. In particular,

the policy ramework seeks to avoid leaving disad-

vantaged groups worse o, such as groups with low

incomes and people without access to alternative

means o getting about. Congress would direct the

DOT and IRS cooperatively to design the PRMT

model program or states and localities to adopt


29/40



JULY 2008 27

in conjunction with congestion pricing. The allo-

cation o toll revenue between tax credits, direct

rebates, and inrastructure investment must be de-

termined locally so as to align with local circum-

stances and preerences. As such, the PRMT model

should provide wide local latitude with regard toincome thresholds, program qualications and

other parameters. At the same time however, the

PRMT model should provide states and localities a

template o sucient technical and administrative

specicity, and any necessary ederal authorities, to

guide ull-scale implementation.

To help illustrate the intent o this recommenda-

tion, Table 7 gives a broad, hypothetical example o

such a model template. Figures shown or average

expenditures on tolls as a percent o income (second

column) are based on the national data in Table 6. In

this example, a 100 percent tax credit is occasionedby households within a designated income bracket

with two or more wage earners. While progres-

sively smaller credits are occasioned by households

with ewer workers, households with no workers

(such as retirees and people with disabilities who are

not working) would be eligible. The intent o the

model must be to balance the provision o nan-

TABLE 7

Progressive Rendable Mobility Ta Credit Program: Eample o A Model Template

Annal Average

income ependitre

(dollars) on tolls as a

percent o 0 wage 1 wage 2+ wage

income earners earner earners

09,999 8 2 4 810,00019,999 5 1.25 2.5 5

20,00049,999 4 0.75 1.5 50,000+ 2 0 0 0

Qalifcations

Live and/or or ithin a conetion pricin area

Live and/or or more than one-ourth mile rom tranit or eliile or ADA paratranit

Other requirement pertainin to children, ae, diaility, invetment income, reidency, citizenhip, etc.

Eample Hoseholds

Rendable ta credit as a percent

o earned income

Hosehold #1

Income: $15,000/year:

To ae earner

Livin in conetion pricin zone:

Not livin ithin one-ourth mile o

tranit or eliile or ADA paratranit

Rendable credit; 5.0 percent o

gross annal income = $750.00

Hosehold #2

Income: $25,500/year:

One ae earner

worin in conetion pricin zone:

Not orin ithin one-ourth mile o



gross annal income, $382.50

Hosehold #3

Income: $5,000

No ae earner (retired couple)

Reidin in conetion pricin zone

Not reidin ithin one-ourth mile o



gross annal income, $262.50

source: Author calculation and tatitic rom Tale 6.


30/40




cial assistance to those who require it while avoid-

ing the granting o windalls to those who would

not incur losses rom congestion pricing. Limiting

the maximum tax credit to households with two or

more workers, as in the Table 7 example, might be

appropriate in light o relevant commuting statis-tics. Mohring (1999), or example, nds that in the

Twin Cities ewer than about one-th o house-

holds, on average, have one or more members who

travel during the morning peak period.

In application, the PRMT model would provide

guidance to states and localities as to how best to

construct column two rom local survey analysis.

Based broadly on the national data in Table 6, the

cost o the example in Table 7 would be in the order

o $20 to $25 billion annually. Administration o theprogram would add a urther 5 to 10 percent. These

costs would be lower to the extent that the model

program is designed to urther minimize windall

gains (to those who do not drive at all, or exam-

ple). The complete elimination o such windalls

is probably impossible, however, without targeting

the program so precisely as to blunt the incentive

characteristics o congestion pricing.

The model program illustrated above refects the

evidence given earlier that, beore taking into ac-count the way in which the revenues rom tolls

might be put to use, many people could be made

worse o by congestion pricing. Mohring (1999)

orecasts that, under areawide congestion pricing in

the Twin Cities, the time-plus-money costs o travel

or those people in the lowest quartile o house-

hold income would almost double as they seek less-

congested (thus lower-tolled), but more circuitous,

routes. Mohring nds, moreover, that even higher-

income groups would not be ully compensated

or their tolls by the travel-time savings they gain.Sarova, Gillingham, Parry, Nelson, Harrington,

and Mason (2004) draw similar conclusions in a

2003 model simulation o congestion pricing in

metropolitan Washington, DC.

Equity depends importantly, thereore, on the way

in which congestion pricing is implemented in a

given local area, on the volume o revenue generated

rom tolls, and on the way in which su

07 Congestion Lewis

Documents