urban road congestion management - capacity investments and pricing policies

40
Urban Road Congestion Management: Capacity Investments and Pricing Policies Hugo E. Silva Pontificia Universidad Católica de Chile Leonardo J. Basso and Ignacio Riquelme Universidad de Chile 1

Upload: brtcoe

Post on 22-Jan-2017

113 views

Category:

Engineering


1 download

TRANSCRIPT

Page 1: Urban Road Congestion Management - Capacity Investments and Pricing Policies

UrbanRoadCongestionManagement:CapacityInvestmentsandPricingPolicies

HugoE.SilvaPontificia UniversidadCatólica deChile

LeonardoJ.BassoandIgnacioRiquelmeUniversidaddeChile

1

Page 2: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Introduction

• Congestionisamajorissue.

▫ Santiago2012:morethan30%ofworkersinSantiago

spendmorethan60minutesinbusesonlygettingtotheir

workplace

▫ US2013:eachpersonspends100minutesperdayintraffic

onaverage,valuedat$760billion(Winston,2013)

▫ London2015:AMPeakspeedincentralLondonis13,4

km/hr andininnerLondonis17,9km/hr

2

Page 3: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Introduction

• Congestionisamajorissue.

• Urbantransportpricesdonotreflectmgsocialcosts.

• Whatcanwedo?▫ Buildbigger(more)roads▫ Managecurrentcapacity� Publictransportpriority(e.g.subsidies,buscorridors)� Carcongestionpricing� Drivingrestrictions

3

Page 4: Urban Road Congestion Management - Capacity Investments and Pricing Policies

4

Page 5: Urban Road Congestion Management - Capacity Investments and Pricing Policies

5

$ 50,000 $ 55,000 $ 60,000 $ 65,000 $ 70,000 $ 75,000 $ 80,000 $ 85,000 $ 90,000 $ 95,000

2002 2004 2006 2008 2010 2012 2014 2016

Mill

ions

of d

olla

rs

Source: US. Bureau of the Census

US Total Construction Spending: Highway and street

Page 6: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Researchquestions

• Howefficientisaroadinvestmentpolicy?

• Howdoesitcompareandinteractwiththeother

possiblepoliciestoreducecongestion?

6

Page 7: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Literature

• Downs-Thomson paradox

▫ Downs (1962).Thomson(1977).Mogridge (1997)• Hypothesis:inequilibrium carandbusgeneralizedcost will beequal (perfect substitutes)

• Consequence:increasing road capacity makescongestion worse

• Graphically..

7

Page 8: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Downs-ThomsonhypothesisCost

Total demand

By car By public transport

Car

Public transport

Page 9: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Downs-Thomson paradox

What happens with road capacity expansion for cars?

Page 10: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Literature

• BassoandJara-Díaz(2012,TR-A)▫ Ifmodechoiceismorerealistic(imperfectsubstitutes)theDThypothesisdoesnotholdandcarsmaybebetteroffwithroadcapacityexpansion

• Zhang,Lindsey,Yang(2016,TR-B)▫ Relaxthe assumption offixed demand,perfectsubsitutability andinclude transit crowding.Differenttransit operation regimes

• The paradox may not hold

10

Page 11: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Literature• Duranton &Turner(2011.AER):empirical

▫ Increasinglanekilometers leadstoaproportionalincreaseofveh-kmtravelled.▫ Elasticityis1▫ Holdsforinterstatetravel

• Hsu&Zhang(2014.JUrbanE)▫ TheelasticityforJapanis1.2– 1.3

11

Page 12: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Literature• Issues with this approach

▫ Aggregate(spaceandtimeofday).▫ Nodistinctionbetweenincreasingcapacityofcurrentroadsorextendingthelengthoftheroads▫ Itdoesnotallowforcomparingpoliciesandassessingthesocialbenefit

12

Page 13: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Main features▫ Comparisonandinteractionofaroadinvestmentpolicy

withbuscorridors,congestionpricingandtransitsubsidization

▫ Detaileddemandmodel,MCPF,trafficengineering

▫ Cityeffects:� Congestionandnetworksize� Modalsplit� Income(GDP)

▫ Busstop&payment(BS&P)technology

13

Page 14: Urban Road Congestion Management - Capacity Investments and Pricing Policies

14

Page 15: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Low vs high BS&P

15

Page 16: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Methodology

1. (Simplified)theoreticalmodelforinsights

2. Morerealisticanddetailedmodelforsimulationsconsideringthreecitytypes:

� European

� American

� Developing

16

Page 17: Urban Road Congestion Management - Capacity Investments and Pricing Policies

The theory model

• 𝑌" = 𝐻 𝑔", 𝑔' 𝑌' = 𝑌 − 𝑌"

• 𝑔' = 𝐴𝐶𝐶 + 𝑃' + 𝑡' 𝑌", 𝑌', 𝑓, 𝐾 + 𝑡1 𝑌', 𝑓, 𝑘 + 𝑡3 𝑓

• 𝑔" = 𝐶𝑂" + 𝑃" + 𝑡" 𝑌", 𝑌', 𝑓, 𝐾 + 𝜀(𝑓) 8 𝑡1 𝑌', 𝑓, 𝑘

• 𝑊 = −∫∑ 𝑦=𝑑𝑔=�=

�� + 𝑃" 8 𝑌" + 𝑃' 8 𝑌' − 𝑐 8 𝑓 − 𝜌 8 𝐾

• Optimizeover𝑃", 𝑃', 𝑓,bussize(k)androadcapacity (𝐾).

• Constraintstomodeldifferentpolicies

17

Page 18: Urban Road Congestion Management - Capacity Investments and Pricing Policies

The theory model

𝑃" − 𝑃' = 𝑌"𝜕𝑡"𝜕𝑌"

+ 𝑌'𝜕𝑡'𝜕𝑌"

LMNOP

− 𝑌'𝜕𝑡'𝜕𝑌'

+𝜕𝑡1𝜕𝑌'

MNOQR"NOSQ

+ 𝑌"𝜕𝑡"𝜕𝑌'

+ 𝜀𝜕𝑡1𝜕𝑌'

LMNOQ

• Onlythepricedifference canbeidentifiedhere,butseparatelyinthesimulationmodel

• Interpretationasinseparatemodels

18

Page 19: Urban Road Congestion Management - Capacity Investments and Pricing Policies

The theory model

𝑃" − 𝑃' = 𝑌"𝜕𝑡"𝜕𝑌"

+ 𝑌'𝜕𝑡'𝜕𝑌"

LMNOP

− 𝑌'𝜕𝑡'𝜕𝑌'

+𝜕𝑡1𝜕𝑌'

MNOQR"NOSQ

+ 𝑌"𝜕𝑡"𝜕𝑌'

+ 𝜀𝜕𝑡1𝜕𝑌'

LMNOQ

• Onlythepricedifference canbeidentifiedhere,butseparatelyinthesimulationmodel

• Interpretationasinseparatemodels

19

Page 20: Urban Road Congestion Management - Capacity Investments and Pricing Policies

The theory model

𝑃" − 𝑃' = 𝑌"𝜕𝑡"𝜕𝑌"

+ 𝑌'𝜕𝑡'𝜕𝑌"

LMNOP

− 𝑌'𝜕𝑡'𝜕𝑌'

+𝜕𝑡1𝜕𝑌'

MNOQR"NOSQ

+ 𝑌"𝜕𝑡"𝜕𝑌'

+ 𝜀𝜕𝑡1𝜕𝑌'

LMNOQ

• Onlythepricedifference canbeidentifiedhere,butseparatelyinthesimulationmodel

• Cross-congestioneffects.

20

Page 21: Urban Road Congestion Management - Capacity Investments and Pricing Policies

The theory model

𝑃" − 𝑃' = 𝑌"𝜕𝑡"𝜕𝑌"

+ 𝑌'𝜕𝑡'𝜕𝑌"

LMNOP

− 𝑌'𝜕𝑡'𝜕𝑌'

+𝜕𝑡1𝜕𝑌'

MNOQR"NOSQ

+ 𝑌"𝜕𝑡"𝜕𝑌'

+ 𝜀𝜕𝑡1𝜕𝑌'

LMNOQ

• Onlythepricedifference canbeidentifiedhere,butseparatelyinthesimulationmodel

• Busstopandpaymenttechnologyeffects

21

Page 22: Urban Road Congestion Management - Capacity Investments and Pricing Policies

The theory model• Buscorridors

• 𝑃" − 𝑃' = 𝑌"TUPTVP

+ 𝑌'TUQTVP

LMNOP

− 𝑌'TUQTVQ

+ TUWTVQ

MNOQR"NOSQ

+ 𝑌"TUPTVQ

+ 𝜖 TUWTVQLMNOQ

• Timefunctions𝑡" and𝑡' aredifferentwithbuslanes

22

Page 23: Urban Road Congestion Management - Capacity Investments and Pricing Policies

The theory model• Increasingroadcapacitymakessometermslessimportant.Ifroadcapacityisverylarge:

• 𝑃" − 𝑃' = 𝑌"TUPTVP

+ 𝑌'TUQTVP

LMNOP

− 𝑌'TUQTVQ

+ TUWTVQ

MNOQR"NOSQ

+ 𝑌"TUPTVQ

+ 𝜖 TUWTVQ

LMNOQ

• 𝑃" ≈ 𝑃' −TUWTVQ

𝑌' + 𝜖𝑌"

• Hence,foranygivenPB,theoptimalcongestiontollmaybeverysmall.Evenhitzero.Itdependsheavily on𝑡1

• =>importanceBS&Ptechnology

23

Page 24: Urban Road Congestion Management - Capacity Investments and Pricing Policies

The simulation model• BasedonBassoandSilva(2014,AEJ– EconomicPolicy)

• Totalelasticdemand.Twoperiods(peakandoff-peak)withintertemporal elasticity.Marginalcostofpublicfunds

• Weusethebestpossibleengineeringfunctionsavailablefor𝑡",𝑡',𝑡1,ε andbuscosts.Twobusstoptechnologies(Tirachini,2014TR-A).

▫ Partialeq.(nohousing.nolabor)=>fixedtraveldistance.shorterrun(thane.g.Duranton andTurner).

▫ Possiblymorefavorable forroadexpansionasinduceddemandcouldbehigher

24

Page 25: Urban Road Congestion Management - Capacity Investments and Pricing Policies

• Nestedlogitdemandmodel

25

The simulation model

Page 26: Urban Road Congestion Management - Capacity Investments and Pricing Policies

• Realdata

• Congestedbaselinescenario:13-18km/hr

• TwoBS&PTech:frontdoorboardingonly.Similarinitial

speeds.

▫ Inefficient:(A)paymentincash.1berth

▫ Efficient:(B)contactlesscard.2berths.

• Buscorridor:atmostonelane.

26

Data european city

Page 27: Urban Road Congestion Management - Capacity Investments and Pricing Policies

27

Efficiency Results

Scenario REF SUBref100 CONref DLref CAP SUB100 CON DLSocialBenefit 0 3550 6725 5179 5033 5998 6746 7376CSchange 0 37130 28672 33954 42207 44624 29808 42899

Busfarepeak 0.19 0.00 0.16 0.16 0.17 0.00 0.16 0.16Busfareoff-peak 0.19 0.00 0.16 0.16 0.17 0.00 0.16 0.16Cartollpeak 0.00 0.00 0.62 0.00 0.00 0.00 0.58 0.00

Cartolloff-peak 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00Frequencypeak 25 51 50 63 37 47 50 62

Frequencyoff-peak 25 30 24 44 24 28 24 39Bussize 39 35 51 47 37 40 50 39

Carpeakspeed 13 15 22 13 21 21 22 18Buspeakspeed 11 12 15 19 16 15 15 21

Caroff-peakspeed 41 41 41 36 47 45 42 50Busoff-peakspeed 26 25 24 26 28 26 25 28

Peakshare 41 43 41 43 45 45 42 45Off-peakshare 49 48 49 48 46 46 49 46NoTravelshare 9 9 9 9 8 8 9 8Carshare|peak 85 74 61 57 81 74 63 66Busshare|peak 15 26 39 43 19 26 37 34

Carshare|offpeak 84 76 83 79 83 77 83 81Busshare|offpeak 16 24 17 21 17 23 17 19Busstopsperkm 2 2 2 3 2 2 2 3

Numberofbuslanes 0 0 0 1 0 0 0 1RoadCapacity 3600 3600 3600 3600 4662 4355 3706 4446

3Lanes– BS&PTechA

Page 28: Urban Road Congestion Management - Capacity Investments and Pricing Policies

28

Efficiency Results• CaseA(lowBS&P)and3lanes

0

1000

2000

3000

4000

5000

6000

7000

8000

SUBref100 CONref DLref CAP SUB100 CON DL

Social Benefit

Without capacity investment With capacity investment

+0,9 lanes

+0,6 lanes

+0,1 lanes

+0,7 lanes

Page 29: Urban Road Congestion Management - Capacity Investments and Pricing Policies

29

Efficiency Results• Investinginroadcapacityasauniquepolicyincreaseswelfare• =>TheDowns-Thomsonparadoxdoesnothold

0

1000

2000

3000

4000

5000

6000

7000

8000

SUBref100 CONref DLref CAP SUB100 CON DL

Social Benefit

Without capacity investment With capacity investment

+0,9 lanes

+0,6 lanes

+0,1 lanes

+0,7 lanes

Page 30: Urban Road Congestion Management - Capacity Investments and Pricing Policies

30

Efficiency Results• CaseA(lowBS&P)and3lanes

0

1000

2000

3000

4000

5000

6000

7000

8000

SUBref100 CONref DLref CAP SUB100 CON DL

Social Benefit

Without capacity investment With capacity investment

+0,9 lanes

+0,6 lanes

+0,1 lanes

+0,7 lanes

Page 31: Urban Road Congestion Management - Capacity Investments and Pricing Policies

31

02000400060008000

100001200014000

+0,9+0,6 +0,1 +0,7

02000400060008000

100001200014000

+0,1+0,0

+0,6+0,9

Case A Case B

• WhathappensifBS&Pisimproved:CaseB• Inallcases,busandcarspeedsarelowandquitesimilaracrossreferencecases

Efficiency Results

Page 32: Urban Road Congestion Management - Capacity Investments and Pricing Policies

32

02000400060008000

100001200014000

+0,9+0,6 +0,1 +0,7

02000400060008000

100001200014000

+0,1+0,0

+0,6+0,9

3

Case A Case B

• IfBS&Ptechisimproved,allotherpoliciesworkbetter(this,inadditiontothedirectbenefitsofimprovingBS&Ptech)

Efficiency Results

Page 33: Urban Road Congestion Management - Capacity Investments and Pricing Policies

33

02000400060008000

100001200014000

+0,9+0,6 +0,1 +0,7

02000400060008000

100001200014000

+0,1+0,0

+0,6+0,9

3

Case A Case B

• IfKisfixed:Orderofpoliciesdoesnotchange

Efficiency Results

Page 34: Urban Road Congestion Management - Capacity Investments and Pricing Policies

34

02000400060008000

100001200014000

+0,9+0,6 +0,1 +0,7

02000400060008000

100001200014000

+0,1+0,0

+0,6+0,9

3

Case A Case B

• Investinginroadcapacityasauniquepolicyalwaysincreaseswelfare=>TheDowns-Thomsonparadoxdoesnothold

Efficiency Results

Page 35: Urban Road Congestion Management - Capacity Investments and Pricing Policies

35

02000400060008000

100001200014000

+0,9+0,6 +0,1 +0,7

02000400060008000

100001200014000

+0,1+0,0

+0,6+0,9

3

Case A Case B

• CaseA:Capacityinvestmentisdominated.Buildcapacitybutforbuses(corridor).

• CaseB:improvebusstops,implementapolicyandDONOTexpandroadcapacity

Efficiency Results

Page 36: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Conclusions• Increasingroadcapacityisefficientbyitselfbutitisnotthebestpolicy

• BS&Ptechiskey:itimprovesthebenefitsofallotherpolicies▫ LowBS&Ptech=>incentivestoinvestinroads,butFORBUSESdosomethingelseadditionally▫ HighBS&Ptech=>bettertoimplementcongestionmanagementpolicies,donotinvestincapacity

• Whatmattersisnotcongestionbut“congestionability”• BS&Pisneverseenasastrategictool.Butitaffectsstrategicdecisions

36

Page 37: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Conclusions• Resultsarerobustwithrespecttonetworksize

• Buildcapacityforbuses(corridor)

• WithoptimalK,optimal“congestion”tolliszero.

• “Optimal”subsidyisnegative.

37

0

5000

10000

15000

20000

25000

30000

+1,5

+1,2+1,5+1,2

02000400060008000

100001200014000

+0,9+0,6 +0,1 +0,7

Case A

Page 38: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Conclusions• Resultsarerobustwithrespecttonetworksize

• Manageroadcapacityanddonotbuild.

• Before:congestionpricing.

• After:buscorridors

02000400060008000

100001200014000

+0,1+0,0

+0,6+0,9

05000

1000015000200002500030000

+1,4+1,0

+0,0

+0,0

Case B

Page 39: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Conclusions

• PreliminaryresultsforU.S.

▫ Buildingcapacityimprovesasastand-alonepolicy

▫ Itcansurpassmanagementpolicies(atref.capacity)

▫ LowBS&P:expandcapacitylessandbuildacorridor

forbuses

▫ HighBS&P:sameconclusion,implementcongestion

management

39

Page 40: Urban Road Congestion Management - Capacity Investments and Pricing Policies

Thanks! Questions? Comments?

40