CRITICAL REVIEW JURNAL THE LONDON CONGES

CRITICAL REVIEW EKONOMI KOTA |

1

Nama Jurnal

: The London Congestion Charge: A Tentative Economic Appraisal

Nama Penulis

: Re´my Prud’homme, dan Juan Pablo Bocarejo.

Jumlah Halaman

: 9 Halaman

Nomor Jurnal

: Transport Policy xx (2004) 1

–

9.

Asal Penulis

: University Paris XII, IUP, Avenue du Ge´ne´ral de Gaulle, 94009

Cre´reil Cedex, France

Tanggal Publikasi : Diterima 17 Desember 2004; Direvisi 22 Februari 2005; Dipublikasi 2

Maret 2005.

I. PENDAHULUAN

1.1 Latar Belakang

Gagasan adanya teori tentang biaya kemacetan perkotaan mulai diperkenalkan di

London pada tahun 1960 (Smeed, 1964; Walters, 1961). Secara substansi, gagasan ini

didukung oleh kebanyakan ekonom kota London. Namun, hanya sedikit kota-kota di dunia

yang menerapkan gagasan ini. Salah satu kota yang berhasil menerapkan gagasan ini

adalah Singapura. Hal ini menjadikan kebijakan mengenai biaya kemacetan

diperkenalkan London pada tahun 2003. David Banister, seorang ekonom transportasi

memberikan pandangan bahwa: Penerapan Biaya Kemacetan di Central London adalah

salah satu kebijakan transportasi radikal yang pernah diusulkan dalam 20 tahun terakhir

dan itu merupakan perubahan yang besar dalam penetapan kebijakan di London

(Banister, 2003, hal. 259).

Zona biaya kemacetan di London sekitar 22 km

2

(Lingkaran dengan radius 2,7 km)

yang didefinisikan dari pusat kota London, yang terdiri atas 370.000 penduduk yang

meng-cover 1,2 juta mobilitas pekerjan. Radius yang ditetapkan ini, presentasenya jauh

lebih kecil dari seluruh luas kota London yakni hanya 1,5% dari luas keseluruhan kota

London dan 5,2% populasi dari keseluruhan warga London. Namun, sampel ini sudah

dianggap mampu menggambarkan pertumbuhan ekonomi London yang signifikan.

Sejak Februari 2003, kendaraan yang melaju dalam zona ntara 7h30 dan 18h30 pada

hari kerja harus membayar biaya 5 pound atau 7,2 euros per hari. Adanya biaya

kemacetan secara general bisa dikatakan sukses dilihat dari pembayaran dan sistem

monitoring biaya kemacetan. Tujuan pengurangan lalu lintas telah tercapai. Jumlah

kendaraan km di zona menurun sekitar 15%, dan kecepatan mereka meningkat sekitar

17%. Penggunaan transportasi publik berupa bus juga turut meningkat. Secara politis,

adanya biaya kemacetan ini juga sukses karena kebanyakan masyarakat London puas

CRITICAL REVIEW EKONOMI KOTA |

2

dengan sistem yang diterapkan sehingga membawa Ken Livingstone selaku walikota

terpilih kembali sebagai walikota London pada tahun 2004.

1.2 Tujuan

Tujuan dari adanya critical review ini adalah :

1. Untuk mengetahui penerapan biaya kemacetan dapat merepresentasikan

keberhasilan secara ekonomi.

2. Mengetahui urgensitas adanya dari biaya kemacetan sehingga dapat menyimpulkan

kemungkinan kebijakan ini diterapkan di Indonesia.

II. DASAR TEORI

Congestion cost (biaya kemacetan) merupakan selisih antara marginal social cost

(biaya yang dikeluarkan masyarakat) dengan private cost (biaya yang dikeluarkan oleh

pengguna kendaraan pribadi) yang disebabkan oleh adanya tambahan kendaraan pada

ruas jalan yang sama. Perhitungan beban biaya kemacetan didasarkan kepada

perbedaan antara biaya marginal social cost dan marginal private cost dari suatu

perjalanan. Biaya kemacetan adalah biaya perjalanan akibat tundaan lalu lintas maupun

tambahan volume kendaraan yang mendekati ataupun melebihi kapasitas pelayanan

jalan. Biaya operasi kendaraan didefinisikan sebagai biaya yang secara ekonomi terjadi

dengan adanya pengoperasian satu jenis kendaraan pada kondisi normal untuk satu

tujuan tertentu. Biaya operasi kendaraan terdiri atas biaya tetap (standing cost) dan biaya

tidak tetap (running cost). (Miranda, 2010)

Biaya kemacetan timbul dari hubungan antara kecepatan dengan aliran di jalan dan

hubungan antara kecepatan dengan biaya kendaraan, seperti ditunjukkan pada Gambar

1. Jika batas aliran lalu lintas yang ada dilampaui, kecepatan rata-rata lalu lintas akan

turun. Pada saat kecepatan mulai turun, biaya operasi kendaraan akan meningkat dalam

kisaran (0

–

45) mil/jam dan waktu untuk melakukan perjalanan akan meningkat (Everall,

1968 dalam Stubs, 1980). Sementara itu, waktu berarti biaya, yang merupakan bagian

dari biaya perjalanan total yang ditimbulkan oleh menurunnya kecepatan akibat

meningkatnya aliran lalu lintas.

Persamaan estimasi biaya kemacetan dirumuskan sebagai berikut:

Dengan:

:

marginal social charge

; yaitu biaya yang dikeluarkan masyarakat untuk melakukan

perjalanan dari zona asal i ke zona tujuan j menggunakan moda jenis m.

:

marginal private charge

; yaitu biaya yang dikeluarkan pengguna kendaraan pribadi

dari zona asal i ke zona tujuan j menggunakan moda jenis m.

CRITICAL REVIEW EKONOMI KOTA |

3

III. REVIEW JURNAL

3.1 Model analisis

Gambar 1 dibawah mewakili kuantitas penggunaan jalan pada sumbu horizontal dan

unit (biaya perkilometer) pada sumbu vertikal.

Gambar 1. Diagram Biaya Kemacetan

Dalam hal ini, penggunaan jalan dapat diukur dalam kendaraan km. D (q) adalah

kurva permintaan, yang mewakili permintaan penggunaan jalan, sebagai fungsi dari biaya

unit menggunakan jalan. Unsur yang paling penting dari unit cost ini adalah biaya waktu,

biaya waktu yang dibutuhkan untuk bergerak sejauh 1 km.

I (q), yang bisa disebut kurva penawaran, adalah per km biaya ditanggung oleh

pengendara kendaraan ketika pengendara sendirian di jalan (ketika q=Z0). J adalah biaya

operasi mengemudi ditambah biaya dengan kecepatan maksimal waktu. Ketika ada

adalah kendaraan lebih (ketika q meningkat), kecepatan berkurang, waktu yang

dibutuhkan meningkat, dan I (q) meningkat. Keseimbangan akan dicapai pada A, di mana

I(q) dan D (q)

berpotongan, pada titik ini pengendara dikenakan biaya yang sama besar berasal dari

penggunaan jalan, bila pengemudi menanggung biaya yang lebih besar dibandingkan

manfaat yang ditrima maka pengemudi akan megurangi penggunaan jalan dan akan

beralih ke transportasi umum.

Sayangnya, keseimbangan ini bersifat sub optimal. Bila dilihat dari sisi social cost

S(q). Biaya sosial adalah sama dengan biaya masing-masing I (q), ditambah biaya dari

waktu tambahan dihabiskan oleh semua kendaraan lain karena satu kendaraan tambahan

di jalan. Titik B, di mana D (q) dan S (q) berpotongan dengan Y kendaraan, dan biaya

satuan M, adalah solusi optimal bagi masyarakat. Di luar titik, sebuah kendaraan

CRITICAL REVIEW EKONOMI KOTA |

4

tambahan menghasilkan biaya sosial yang lebih besar daripada manfaat sosial yang

diciptakan. Untuk mencapai keadaan yang optimal maka diperlukan adanya pajak/biaya

tambahan.

Kesimpulan dari adanya metode yang digunakan adalah yang pertama yaitu apabila

perpotongan kurva private cost relatif datar, maka bisa diketahui keseimbangan alami

kuantitas penggunaan jalan selalu lebih besar dari optimal (x>y) yakni jalan hampir selalu

padat atau kurang padat.

Kedua, jumlah optimal penggunaan jalan Y (dan tingkat optimal terkait kemacetan)

adalah fungsi dari permintaan untuk penggunaan jalan; jika permintaan meningkat, kurva

D (q) bergerak ke kanan, dan begitu juga dengan kuantitas yang optimal; jika kemiringan

kurva permintaan menurun, yaitu jika permintaan relatif terhadap harga menjadi harga

yang lebih elastis, kuantitas optimal penggunaan jalan menurun.

Ketiga, menunjukkan perbedaan utama antara pendekatan insinyur dan pendekatan

para ekonom: sementara insinyur mendefinisikan penggunaan jalan optimal dan

kemacetan sebagai fungsi dari karakteristik jalan saja, ekonom Pendekatan

mendefinisikan sebagai fungsi dari kedua karakteristik jalan dan permintaan jalan.

Keempat yakni pajak atau biaya optimal adalah eksternalitas dari kemacetan (perbedaaan

antara social cost dan individual cost) pada titik optimum, bukan pada keseimbangan

alami.

Pada dasarnya, akan dibutuhkan banyak referensi mengenai biaya kemacetan untuk

diterapkan di London. Mengingat pada keseimbangan antara social cost dan individual

cost. Referensi dibutuhkan untuk merumuskan metode biaya kemacetan yang sesuai

diterapkan di kota London.

3.2 Modifikasi Congestion Charge di Kota London

Dari model analisis secara general sebagaimana dijelaskan sebelumnya, London

mengembangkan sebuah model analisis berdasarkan anggapan alasan mengapa biaya

sebenarnya akan sama dengan biaya optimal. London mengasumsikan tidak seperti itu.

Penggunaan jalan didefinisikan sebagai jumlah roda empat kendaraan per km per hari di

zona yang telah ditentukan pada jam-jam dikenakan biaya.kemacetan Bus dikecualikan,

karena fungsi biaya mereka dan mereka kontribusi terhadap kemacetan yang sangat

berbeda dengan kendaraan lainnya. Bus hanya menyumbang 3,5% dari total kendaraan

km pada tahun 2002.

3.3 Kurva cost di London pada zona yang dikenakan biaya

Dalam perhitungan biaya kemacetan, terdapat 2 variabel yang mempeengaruhi, yakni

social cost dan individual cost. Pada Gambar 2 menjelaskan besaran dan penggunaan

CRITICAL REVIEW EKONOMI KOTA |

5

jalan di kota London sesuai dengan teori yang terdapat pada kurva biaya kemacetan di

kota London.

Gambar 2. Kemacetan Jalan dengan Biaya Kemacetan Kota London

Dalam perhitungan, mengandung bagian, yang mewakili amortisasi dan biaya

bahan bakar, dan bagian variabel, yang merupakan nilai dari waktu yang dihabiskan

mengemudi 1 km. Bagian tetap diperkirakan (Glaister 2003) menjadi 0,15 (Euro per km).

Bagian variabel adalah sama dengan waktu yang dihabiskan (T, dalam jam), yang

merupakan fungsi dari kecepatan (s, dalam km/ jam), yang itu sendiri merupakan fungsi

dari penggunaan jalan (q), dikalikan dengan nilai waktu (v, di euro per jam).

I(q)= 0,15+tv=0,15+[1/s(q)]v

Untuk melangkah lebih jauh, kita perlu perkiraan s (q), kecepatan sebagai fungsi

penggunaan jalan, dan v, nilai waktu. Untuk nilai waktu, ROCOL (2000) melaporkan

mengusulkan 15,6 euro per jam. Karena, ada rata-rata 1,34 orang per kendaraan, ini

menempatkan nilai waktu per kendaraan di € 20,9 per jam.

Kecepatan menurun, dan kebanyakan linear, fungsi jalan yang menggunakan q:

S= α

-

βq

dimana

α

adalah kecepatan saat jalan kosong yakni 31,6 km/jam. Dan diketahui

kecepatan rata-rata pada tahun 2002 di London adalah 14,3 km/jam dapat diasumsikan

β

adalah 00,01245, maka persamaan yang didapatkan adalah:

I(q)=0,15+20,9/(31,6-0,0124q)

Maka untuk menentukan social cost S(q), maka dapat ditentukan dari deduksi I(q). S(q)

sama dengan I(q) ditambah turunan I’(q) dan dikali dengan penggunaan jalan q:

S(q)

= I(q)+I’(q)q

S(q)= 0,15+20,9/(31,6-0,0124q)+0,26q/(31,6-0,0124q)

2

3.4 Kurva Demand pada zona yang dikenai biaya kemacetan

CRITICAL REVIEW EKONOMI KOTA |

6

Biaya rata-rata dibayar per kendaraan/km dapat ditentukan dengan membagi total

biaya yang dikumpulkan dengan jumlah kendaraan/km. Dari jumlah tersebut, dapat

diketahui biaya tahunan adalah 115 juta pound atau 165,6 euro, karena ada sekitar 255

hari dikenakan biaya per tahun, ini adalah 451.000

pound atau 649.000 € per

hari

dikenakan biaya-

dan € 0,56 per kendaraan/km.

Oleh karena itu, biaya unit ditanggung

oleh pengguna meningkat menjadi 1,93.

3.5 Besaran signifikan untuk daerah di London yang dikenai biaya

Tabel 1 menunjukkan angka pada situasi optimal, hal ini dijelaskan sebagaimana

pada jurnal, ketika kurva permintaan dan kurva biaya sosial berpotongan, yaitu ketika S

(q) ZD (q). Biaya kemacetan adalah nol dari situasi yang optimal. Pada prinsipnya, biaya

ini didefinisikan sebagai perbedaan antara integral dari biaya sosial kurva dan kurva

permintaan atas YX (atau Y0X) nilai q. Dalam prakteknya, perbedaan antara BCXY dan

BAXY.

Manfaat dari kebijakan biaya adalah pengurangan biaya kemacetan, relatif terhadap

situasi yang optimal. Biaya operasional pada tahun 2003-2004 adalah 138,8 juta euro.

Sebagian besar terdiri dari pembayaran yang dilakukan Modal, entitas swasta.

Disamping itu, pemerintah melakukan investasi dalam periode pra-charge biaya

investasi selama periode 2000-

2003 dilaporkan menjadi € 245.700.000. Dengan asumsi

biaya peluang 5% dengan tingkat depresiasi 10%, biaya investasi ini sebesar 36,9 M.

euro per tahun. Jumlah biaya yang dikumpulkan pada tahun 2003 sekitar 175,7 euro

per tahun, atau € 689.000 per hari dikenakan biaya.

CRITICAL REVIEW EKONOMI KOTA |

7

3.6 Pertanyaan mengenai skema biaya kemacetan kota London

Seberapa pentingkah biaya kemacetan? Tabel 1 memberitahu kita betapa

pentingnya kemacetan biaya (sebagaimana didefinisikan di sini) di zona yang

dibebankan sebelum pengenalan biaya, Biaya kemacetan pada tahun 2002 menjadi

sebesar € 296.000 set

iap harinya, bila diakumulasikan adalah sekitar 75 M. euro per

tahun. (Tidak termasuk kemacetan di akhir minggu dan hari-hari lain yang bebas dari

biaya kemacetan. Akumulasi biaya kemacetan prosentasenya cukup kecil dalam

menyumbang PDB London pada tahun 2001, PDB London, adalah 255.000 juta euro.

Sedangkan biaya kemacetan di zona yang dikenakan biaya mewakili hanya 0,03% dari

output ekonomi kota London.

Pada tahun 2001, sekitar 4,5 juta. pekerja di kota London dan 1,2 juta. pekerja di

zona zona biaya kemacetan. Bila diasumsikan bahwa produktivitas tenaga kerja sama

antara di dalam maupun diluar zona biaya kemacetan, kita dapat memperkirakan output

dari zona biaya kemacetan adalah 68.000 M. euro per tahun. Biaya kemacetan di

daerah ini mewakili sekitar 0,11% dari PDB luas area kota London. Hal ini sangat

sejalan dengan temuan Prud'homme (1999, 2000) untuk wilayah Paris.

3.7 Biaya dan keuntungan lain

Dengan adanya pengurangan kepadatan kendaraaan, memungkinkan adanya

penurunan tingkat polusi di kota London. Selain itu, adanya biaya kemacetan dapat

meminimalisir jumlah kendaraan menjadi 230 ribu perhari. Adanya peningkatan

kecepatan 17% berarti terjadi 34% penurunan emisi polusi. Hal ini dapat dikelompokkan

ke dalam manfaat tambahan yakni sebesar

€ 11.440 per h

ari atau 2,8 M. euro/tahun.

Selain itu, bagi pengguna kendaraan umum seperti Bus, kecepatannya meningkat

yakni meningkat sebesar 7%. Ini adalah keuntungan bagi orang-orang yang

bermobilitas dengan bus, dan itu adalah manfaat yang disebabkan oleh biaya

kemacetan. Pengguna bus, berjumlah 356.000, naik 1,34 menit per orang per hari, yang

menyumbang 124 ribu euro per hari, atau 31 M. euro per tahun. Ini hampir setara

dengan setengah manfaat dinikmati oleh pengguna mobil pribadi. Sehingga banyak

masyarakat yang mulai beralih ke transportasi massal seperti Bus.

IV. TINJAUAN KRITIS

Permasalahan transportasi erat kaitannya dengan kemacetan, banyaknya waktu

yang terbuang akibat kemacetan di jalan yang akan mempengaruhi produktivitas

masyarakat. Kemacetan bisa terjdi jika arus lalu lintas mendekati kapasitas, kemacetan

mulai terjadi. Kemacetan semakin meningkat apabila arus begitu besarnya sehingga

kendaraan sangat berdekatan satu sama lain. Kemacetan total apabila kendaraan harus

CRITICAL REVIEW EKONOMI KOTA |

8

berhenti atau bergerak lambat. Kemacetan adalah kondisi dimana arus lalu lintas yang

lewat pada ruas jalan melebihi kapasitas jalan tersebut.

Selama ini, di negara-negara di dunia, mengatasi permasalahan kemacetan itu hanya

berupa peningkatan kapasitas jalan saja, tanpa adanya disinsentif penggunaan

kendaraan pribadi, maka justru dapat menyebabkan peningkatan laju pertumbuhan lalu

lintas dan hasilnya tetap saja akan macet. Dari hal tersebut dapat muncul “induced

demand” yakni apabila supply ditambah maka konsumsi akan meningkat pula

(berbanding lurus), hal ini bisa dikarenakan konsumsi yang sangat tinggi dan tidak

terlayani secara optimal.

Jadi, pada intinya, dari permasalahan kemacetan ini, dicari suatu solusi agar

kemacetan dapat dikurangi dan meminimalisir kerugian masyarakat dan negara akan

kemacetan. Biaya kemcetan sendiri adalah selisih dari marginal cost yaitu biaya yang

dikeluarkan masyarakat dengan private cost yaitu biaya yang dikeluarkan kendaraan

pribadi. Terjadi apabila ada tambahan kendaraan pada ruas jalan yang sama. Biaya

kemacetan ini muncul karena adanya tundaan lalu lintas dan tambahan volume

kendaraan yang melebihi kapasaitas jalan. Adanya kemacetan ini menyebabkan

pengemudi kendaraan hmendapatkan manfaat yang lebih sedikit dari biaya yang harus

dikeluarkan.

congestion charge

diharapkan mampu mengurangi perjalanan dengan

kendaraan pribadi dan mengurangi perjalanan yang tidak perlu, terutama pada jam-jam

sibuk.

Pada studi kasus di kota London, talah diberlakukan adanya

congestion charge

sejak

tahun 2003 namun dirasa masih belum cukup optimal dengan diterapkannya kebijakan

ini sehingga mulai dicoba dikaji untuk diterapkan pada kota-kota besar di Indonesia

seperti di Jakarta dan di Jogjakarta.

Congestion charge

ini diharapkan bisa membantu

mengurangi tingkat kemacetan, meningkatkan reliabilitas waktu perjalanan dan

mendorong masyarakat untuk beralih ke transportasi bila dirasa manfaatnya mendekati

menggunakan kendaraan pribadi. Di London sendiri, keberhasilan dapat dilihat dari

penurunan volume lalu lintas, penurunan tangka kemacetan, perjalanan lebih reliable

(antara waktu di perjalanan dengan waktu di tempat tujuan proporsional sehingga tidak

menimbulkan kerugian produktivitas masyarakat), adanya realibilitas jadwal bus kota yang

meningkat dengan signifikan, kecelakaan lalu lintas dapat menurun karena kurangnya

kemacetan dan dapat menjadi sumber pendapatan pemerintah dari Produk Domestik

Bruto yang dapat dipakai untuk perbaikan pelayanan angkutan umum.

Adanya penerapan

congestion charge

di London tidak lepas dari adanya dukungan

masyarakat akan sangat berpengaruh terhadap berhasil tidaknya aplikasi

congestion

charge

). Pemerintah harus melibatkan masyarakat dalam mengambil keputusan dan

CRITICAL REVIEW EKONOMI KOTA |

9

harus mampu menjelaskan tujuan dan maksud

congestion charge

dengan jelas dan

menyeluruh kepada masyarakat, penjelasan proses

congestion charge

dalam mengatasi

masalah-masalah kemacetan dan masalah transportasi lainnya, keuntungan dan kerugian

congestion charge

untuk masyarakat, dan menjaga rasa aman dan nyaman dalam

melakukan pergerakan dan transparansi pendapatan dari

congestion charge

akan

dikelola dan dipergunakan untuk kepentingan transportasi umum.

Dari jurnal

The London congestion charge: a tentative economic appraisal

terdapat

kelebihan dan kelemahan dalam penjelasan mengenai

congestion charge.

Kelebihan dari

jurnal ini adalah adanya penjelasan lain mengenai manfaat lain adanaya

congestion

charge

selain dari sisi ekonomi, yakni pada sisi lingkungan dan pemanfaatan PDB yang

didpatkan dari zona

congestion charge

. Selain itu, terdapat Appendix dari kalkulasi yang

terdapat pada jurnal sehingga memudahkan pembaca dalam memahami metode analisis

dan perhitungan yang digunakan dalam penentuan

congestion charge

di jurnal ini.

Namun, masih terdapat kekuranagan dari jurnal ini yakni belum adanya penjelasan

pembanding pelaksanaan

congestion charge

di daerah lain yang telah verhasil untuk

dijadikan referensi dalam analisis dan penyesuaiannya dengan kondisi kota London

sendiri. Sebagai tambahan informasi, terdapat beberapa kota yang telah berhasil

menerapkan konsep

congestion charge

ini yakni Singapura dan Stockholm.

Singapura memulai menerapkan

congestion charge

sejak tahun 1998 yang bertujuan

untuk membatasi lalu lintas kendaraan yang masuk CBD apalagi pada jam-jam sibuk

sehingga dapat mengurangi kemacetan. Dampak dari penerapan ini cukup signifikan

yakni prosentase penggunaan bus yang meningkat dari 41% menjadi 62% dan volume

lalu lintas zona

congestion charge

berkurang 44% (Sutarsono, 2008).

Di Stockholm, dimulai sejak tahun 2007 setelah proses uji coba sejak tahun 2006.

Tujuannya adalah untuk mengurangi kemacetan, peningkatan aksesibilitas, dan

perbaikan kualitas lingkungan. Hasil yang didapat adalah menurunnya prosentase lalu

lintas ke/dari pusat kota dari 20-25% menjadi 10-15%, meningkatnya aksesibilitas yang

ditandai dengan penurunan antrian di pusat kota dan daerah-daerah dekat pusat kota

sebesar 30-50%dan menurunnya total emisi kendaraan bermotor antara 10-14% di pusat

kota, dan antara 2-3% untuk total satu kota.

V. PENUTUP

5.1 Kesimpulan

CRITICAL REVIEW EKONOMI KOTA |

10



Berdasarkan penerapan

congestion charge

di kota London dapat meningkatkan

Produk Domestik Bruto yang dapat digunakan sebagai dana untuk pembangunan

transportasi umum yang lebih baik.



Kemacetan telah banyak dikurangi dan pengurangan ini merupakan keuntungan

pada bidang ekonomi.



Biaya ekonomi terkait dengan sistem lebih besar dari keuntungan ekonomi yang

dihasilkannya. Tabel 2 merangkum biaya-biaya dan manfaat. Kesenjangan antara

keduanya ternyata cukup besar. Manfaat secara ekonomi kurang dari 60% dari

biaya ekonomi.



Namun, dalam perjalannya hingga saat ini, implementasi

congestion charge

semakin membaik dan menjadi salah satu contoh kota yang berhasil

menerapkannya.

5.2 Lesson Learned

Sebagaimana kita mengetahui, di negara berkembang seperti di Indonesia

kemacetan menjadi salah satu permasalahan transportasi dan ekonomi yang terus

berlarut-larut. Dimana antara

supply

berupa kapasitas jalan tidak sebanding dengan

demand

pergerakan masyarakat. Diperlukan adanya sebuah metode yang mampu

mengurangi kemacetan tersebut. Dari critical review ini dapat memberikan pandangan

mengenai salah satu metodenya yakni

congestion charge

. Dari implementasinya

diharapkan dapat menyumbang peningkatan PDB sehingga dana tersebut dapat

dialokasikan untuk peningkatan kualitas angkutan umum.

Berdasarkan perbandigan dengan jurnal sejenis, telah dilakukan pengkajian

penggunaan metode ini di kota-kota besar di Indonesia seperti Jakarta dan Jogjakarta

untuk diterapkan. Namun, dalam prosesnya akan mebutuhkan regulasi yang kuat dan

dukungan penuh dari masyarakat untuk dapat memaksimalkan implementasi metode ini

di Indonesia.

CRITICAL REVIEW EKONOMI KOTA |

12

DAFTAR PUSTAKA

Sugiyanto, G. (2008). Biaya Kemacetan (Congestion Charging) Mobil Prbadi di

Cnentral Bussiness District (Studi Kasus Kawasan Malioboro Jogjakarta).

Media Teknik Sipil

, 59-65.

Sugiyanto, G., & Malkhamah, S. (2006). Model Pemilihan Moda Antara Mobil Pribadi

dan Bis Transjogja Akibat Penerapan Biaya Kemacetan.

Jurnal Transportasi

Vol. 9

, 97-106.

Susantono, B. (2008, September-Oktober).

Tata Ruang

. Dipetik Maret 15, 2015, dari

Online Bulletin ISSN: 1978-1571:

The London congestion charge: a tentative economic appraisal

*

,†

Re´my Prud’homme*,Juan Pablo Bocarejo

1 University Paris XII, IUP, Avenue du Ge´ne´ral de Gaulle, 94009 Cre´reil Cedex, France

Received 17 December 2004; revised 22 February 2005; accepted 2 March 2005

Abstract

Pre-charge and post-charge data (particularly on speed and road usage) in the London congestion charge zone is used to estimate demand and cost curves for road usage. Pre-charge congestion costs are estimated, and shown to be small (0.1% of the area GDP). They are largely (90%) eliminated by the charge, which produces an economic benefit. Charge proceeds are about three times larger than the value of the congestion. Unfortunately, the yearly amortisation and operation costs of the charge system appear to be significantly higher than the economic benefit produced by the system. The London congestion charge, which is a great technical and political success, seems to be an economic failure. It could be defined as mini Concorde.

q_{2005 Published by Elsevier Ltd.}

1. Introduction

The very notion of urban congestion pricing was introduced—in London— in the 1960s (Smeed, 1964; Walters, 1961). It was subsequently endorsed by all or most economists. However, very few cities (with the notable exception of Singapore) put the idea into practice. This is why the congestion charge experiment introduced—in London again—in 2003 is particularly interesting. David Banister’s view is widely shared by transport economists: ‘Congestion charging in Central London is the most radical transport policy to have been proposed in the last 20 years and it represents a watershed in policy action’ (Banister, 2003, p. 259). In addition, pre and after charge data gives a unique possibility to try and see how important are in

practice the theoretically large merits of a congestion charge.

The London congestion charge, its physical impacts, and its political acceptability have been described elsewhere in detail (Banister, 2003). A congestion charge zone of about 22 km2(a circle with a radius of 2.7 km) has been defined in downtown London, comprising about 370,000 inhabitants and 1.2 million jobs. This is a relatively small area, representing about 1.5% of the Greater London area and 5.2% of its population—and a much smaller proportion of the hard-to-define but economically significant London agglomeration. Since February 2003, vehicles driven in this zone between 7h30 and 18h30 on week-days must pay a charge of 5 pounds or 7.2 euros2 per day. The average charge paid is actually lower, because of exemptions3and reduced charges for some people4, not to mention charge evasion.

The congestion charge is generally seen as a great success. It is a technical success. The payment and monitoring system, after some initial difficulties, functions well. The zone traffic reduction objectives have been reached. The number of vehicle km in the zone declined by about 15%, and their speed increased by about 17%. Bus

Transport Policy xx (2004) 1–9

www.elsevier.com/locate/tranpol

0967-070X/$ - see front matterq2005 Published by Elsevier Ltd. doi:10.1016/j.tranpol.2005.03.001

*

This paper benefited from a grant of PREDIT, a research program funded by the French ministry of Equipment. A draft was presented and discussed at a meeting of an ECMT/OECD working group on congestion, Paris, Nov. 29, 2004. The authors wish to thank two anonymous referees for their thoughtful comments.

† A response to this paper is included in the Topical Issues section of this issue of the journal. The Editors invite further comments and will consider them for publication is subsequent Topical Issues sections.

* Corresponding author. Tel.:C1 48 87 48 38; fax:C1 48 87 66 62.

E-mail addresses:prudhomme@univ-paris12.fr (R.P.’ homme), juan. bocarejo@wanadoo.fr (J.P. Bocarejo).

1_{Respectively Professor Emeritus and Ph. D. student, University Paris}

XII.

2_{A 1 pound}

Z1.44 euros, the average exchange rate in 2003, is used throughout this paper.

3_{Motorbikes, taxis, handicaped persons, buses, power-fueled motor}

vehicles, public utility vehicles are charge exempt.

4_{Residents of the zone pay only 50 cents of a pound perd ayK10U` of the}

patronage in the charged zone increased. Politically, the charge is a also a great success. Most Londoners are satisfied with the system, and Ken Livingstone, the mayor who introduced it, was widely applauded for it, and was re-elected in 2004, in part because of the congestion charge.

But, is it an economic success? This question is the focus of this paper. How important were the potential benefits of a congestion charge? Have they been reaped? Has the level of the charge been correctly defined? How do the actual benefits compare with the costs of operating the system?

Such an economic appraisal is difficult, and necessarily tentative, for several reasons. First, the congestion charge is relatively recent. Short-term behavioural reactions may not hold in the medium-term. Some of the changes induced by the charge (for instance changes in business location) will require several years to materialise. Second, some of the recorded changes in transport patterns (which are often uncritically attributed to the congestion charge) may in reality be caused by exogeneous events. The most glaring example is the drop in subway patronage in 2003, which is mostly the consequence of the temporary closure of a subway line. Third, very little information is known about what happens outside the charged zone, in the ‘rest of London’, as a consequence of the charge. Fourth, the congestion charge is the most important element, but not the only element, of the policy changes introduced in 2003. Bus supply, in particular, was significantly increased. There is, therefore, an ambiguity in all evaluation: are we interested in the impacts of the congestion charge only, or in the impacts of the package that included the congestion charge? For all these reasons, any pronouncement about the London congestion charge must be prudent, and seen as tentative.

This paper is an attempt to provide a quantitative analysis of the scheme. It is based on earlier work byPrud’homme

(1999, 2000) and Prud’homme and Yue-Ming (2000) on

congestion in the Paris area. Most of the data utilized comes from Transport for London (subsequently TfL) website (www.tfl.gov.uk). The paper begins with a simple model of congestion. It continues with a modified version for the case of London, and proceeds to establish the cost and demand curves that make it possible to provide quantitative estimates of the main characteristics of the London system. These estimates in turn make it possible to answer some of the important questions raised by the scheme.

2. A simple model of analysis

Let us consider a diagram representing the quantity of road usage on one axis and the unit (i.e. per km) costs of road usage on the other, as inFig. 1. This can be applied to a given road, or to a given area—such as the London congestion zone. In this case, road usage can be measured in vehicle km.

D(q) is a demand curve, that represents the demand for the use of the road, as a function of the unit cost of using

the road. The most important element of this unit cost is a time cost, the cost of the time needed to drive 1 km.

I(q), which could be called a supply curve, is the per km cost borne by a motorist. When the motorist is alone on the road (when qZ0), this cost is J, the operating cost of driving, plus the time cost at the maximal speed. When there are more vehicles (whenqincreases), the speed is reduced, the time needed is increased, andI(q) increases.

An equilibrium will be reached atA, whereI(q) andD(q) intersect, withXvehicles km driven in the zone, and a unit cost of L. At this point, the marginal driver bears a cost equal to the benefit he/she derives from road usage. Beyond, he/she would bear a cost greater that the benefit derived, and would not use the road.

This natural equilibrium is unfortunately suboptimal. This is easy to see when we consider S(q), the unit social cost created by a vehicle as a function road usage. This social cost is equal to the individual costI(q), plus the cost of the additional time spent by all other vehicles because one extra vehicle is on the road. Point B, whereD(q) and S(q) intersect, withYvehicles km, and a unit costM, is the optimal solution for society. Beyond that point, an additional vehicle generates a social cost greater than the social benefit it creates. This optimal situation can be reached by the imposition of a tax equal to EB—a congestion charge—that will reconcile the private cost and the social cost.

Several interesting conclusions can be derived from this analysis.

First, except when the demand curve intersects the private cost curve in its flat part, the natural equilibrium quantity of road usage is always greater than the optimal quantity of road usage:Xis greater thanY. In other words, roads are nearly always congested; they are only more or less congested.

Second, the notion of an optimal quantity of road usage implies the notion of an optimal level of congestion. The objective of policies, therefore, should not be to ‘eliminate’

Unit costs

Social cost: S(q)

Individual cost: I(q)

Demand: D(q) R N M L P J

O Y X q

C A B G E _F H

Fig. 1. Road congestion.

R. Prud’homme, J.P. Bocarejo / Transport Policy xx (2004) 1–9

congestion—an objective that does not make much sense, since there is always some congestion—but to make sure that the optimal level of congestion prevails.

Third, the optimal quantity of road usage Y (and the associated optimal level of congestion) are a function of the demand for road usage; if the demand increases, the curve D(q) moves rightward, and so does the optimal quantity; similarly, if the slope of the demand curve decreases, that is if the demand relative to price becomes more price elastic, the optimal quantity of road usage decreases.

Fourth, this points to the main difference between the engineer’s approach and the economist’s approach: while the engineer defines the optimal road usage and congestion as a function of road characteristics only, the economist approach defines it as a function of both road characteristics and road demand.

Fifth, the optimal tax or charge is the congestion externality (the difference between the social cost and the individual cost) at the optimum, not at the ‘natural’ equilibrium. It is EB and not AC, contrary to what is often suggested. A congestion charge equal to AC would overshoot, and reduce road usage to a point (not indicated onFig. 1) much to the left ofY, that would be suboptimal. Sixth, congestion costs should be defined as what is lost by society for not being at the optimum, for being atArather than atB, for havingXrather thanYvehicle km. Congestion costs are, therefore, equal to BCA. They are also equal to the increase in welfare associated with the move fromAtoB, that is to PRBE (the utility after) minus LRA (the consumer’s surplus before), which is equal to LGEP-BAG. They are the benefits of introducing a congestion charge. This point is not necessarily obvious, and requires some elaboration.

To define congestion costs, one needs a reference situation, to which the present congestion situation can be compared. This reference situation cannot be the empty road. Roads are not built to be empty. There is nothing optimal or desirable in an empty or quasi empty road with a few vehicles driving at a free-flow speed. Saying that the difference between time actually spent and time that would be spent at free-flow speed is ‘time lost’ does not make much sense. If free-flow speed were the implicit norm, why not apply it to public transport? We would then compare time actually spent in public transport to that norm, call the difference ‘time lost’, value it and present it as a ‘social cost’ of public transport. Fortunately, nobody engages in this futile exercise. Congestion costs are, therefore, not equal to the difference between the unit costs withXvehicle km and the unit costs with zero vehicle km, multiplied byX, that is to LAHJ—although this naive and erroneous view is often held. The reference situation cannot be either the maximum flow situation, as the engineers are tempted to suggest, because it ignores completely variations in the demand for road usage. The reference situation cannot be one in which there are no external costs, which is what is implied in the not uncommon definition of congestion costs as equal to

NCAL, the product of the present marginal unit cost CA by the quantity of vehicle km. This would correspond to the empty road situation. The only meaningful reference situation is the optimal situation, and congestion costs have to be defined as the difference between the present welfare and the potential—and higher—welfare associated with this optimal situation.

Seven, the amount of the congestion charge paid, MBEP, is larger, often much larger, than the economic benefits brought by the congestion charge. To an economist, this is not a problem, because the charge is a transfer, not an economic cost. Drivers may of course have a somewhat different view.

Finally, transaction costs (collection costs in the case of a charge) should be deducted from the benefits of congestion reduction. Economists have a tendency to ignore transaction costs. As we shall see in the case of London, this tendency might be misleading.

3. A slightly modified model for London

The data available on the London congestion charge experiment makes it possible to implement this model, or a slightly modified version of it, in order to throw some light on this experiment. The modification is the following. There is no a priori reason why the actual charge would be exactly the optimal charge EB. Let us assume it is not, and that it is E0B0. Such a charge moves the equilibrium point from A toB0, and road usage fromXtoY0.

Road usage is defined as the number of four-wheel vehicles km per day in the charged zone at charged hours. Buses are excluded, because their cost function and their contribution to congestion are very different from those of other vehicles. Buses km, however, accounted for only 3.5% of total vehicles km in 2002. In 2002, before the introduction of the charge, road usage thus defined was 1,390 thousand vehicles km per day, according to TfL. In 2003, after the charge (and, assuming all other things equal, because of it) it was 1,160 thousands, a 16.5% decline which is the main achievement of the charge5. In other words, we have (in 1,000 vehicles km):

XZ1;390 Y0Z1;160

4. Cost curves for London charged zone

The next step is to write the equation of the cost curve I(q), expressed in euros per vehicle km. It consists of a fixed 5_{This decline is slightly larger than the figure generally given (15%)}

because this figure is calculated on the total number of vehicles km, including buses km.

part, representing amortisation and fuel costs, and of a variable part, which is the value of the time spent driving 1 km. The fixed part is estimated (Glaister, 2003) to be 0.15 (euros per km)6. The variable part is equal to the time spent (t, in hours), which is a function of speed (s, in km/hour), which is itself a function of road usage (q), multiplied by the value of time (v, in euros per hour):

IðqÞZ0:15CtvZ0:15C½1=sðqÞv (1) To go further, we need an estimate ofs(q), the speed as a function of road usage, and ofv, the value of time. For the value of time, the ROCOL (2000) report proposes 15.6 euros per hour. This is high number. The values used in the Paris region, particularly to justify transport investments, are significantly below, and they are considered high by many. We will nevertheless keep this 15.6 euros per hour estimate in this paper. Since, there are on average 1.34 persons per vehicle, this puts the value of time per vehicle at 20.9 euros per hour.

Speedsis a declining, and largely linear, function of road usageq:

sZaKbq (2)

a, the speed on empty roads (when road usage qZ0) is given by TfL as 31.6 km/h. Since we know the average speed in 2002 (whenqwas equal to 1,390) which was equal to 14.3 km/h, we can calculate b, which turns out to be 0.01245. We therefore have:

IðqÞZ0:15C20:9=ð31:6K0:0124qÞ (3) The social cost curveS(q) can easily be deduced from I(q). It is equal to the individual cost curve I(q), plus

the derivativeI0(q) multiplied by road usageq:

SðqÞZIðqÞCI0ðqÞq (4) SðqÞZ0:15C20:9=ð31:6K0:0124qÞ

C0:26q=ð31:6K0:0124qÞ2 (5)

5. Demand curve for London charged zone

The following step is to determine the equation of the demand curveD(q). We know one point of this curve, the equilibrium point A in 2002, because we know the speed at the time. Its coordinates are 1.61 (euros per vehicle km and 1,390 (thousand vehicles km per day). We can also figure out the coordinates of pointB0, the equilibrium point in 2003 after the charge, for which we already know the number of thousands vehicles km per day, Y0Z1,160. The individual unit cost for this point is equal to the fixed cost plus the time cost plus the charge paid.

The first two elements are given by equationI(q). With qZ1,160 we haveI(q)Z1.37 euro. This is a measure ofE0Y0

or P0 in Fig. 2. The average charge paid per vehicle km

driven can be determined by dividing the total charge collected by the number of vehicles km. The amount of the yearly charge is 115 million pounds, or 165.6 euros. Since there are about 255 chargeable days per year, this is 451,000 pounds or 649,000 euros per chargeable day—and 0.56 euro per vehicle km. The unit cost borne by users is therefore increased to 1.93. This, by the way, indicates aK0.83 price elasticity of demand for road usage in the chargeable zone. A andB0are both on the demand curveD(q). With the coordinates ofAandB0, it is easy to calculate the equation of the demand curve:

DðqÞZ3:54K0:00139q (6)

Unit costs

Social cost S(q)

3.38 N C

2,39 Q B’’

2.09 M B Individual 1.93 M’ B’ cost I(q)

1.61 L G A 1.37 P’ E’ F’

1.28 P E F D(q)

0.81 J H

Road usage (q) Y Y’ X

1055 1160 1390 Fig. 2. Road congestion with a congestion charge.

6_{This is an approximation; fuel consumption is also in part influenced by}

speed, which is also influenced by road usage; but the estimates provided do not vary much (from 0.14 to 0.16); we retained 0.15 for the sake of simplicity.

R. Prud’homme, J.P. Bocarejo / Transport Policy xx (2004) 1–9

6. Significant magnitudes for London charged zone

Equipped with these equations, we can determine the coordinates of all the points represented in Fig. 2, and produce the numbers inTable 1.

The optimal situation (qZY) is obtained, as mentioned before, when the demand curve and the social cost curve intersect, that is whenS(q)ZD(q).

Congestion costs are defined as BCA in the case of the pre-charge situation, and as BB00B0 in case of the present situation. By definition, congestion costs are zero in the case of the optimal situation. In principle, these costs are defined as the difference between the integrals of the social cost curve and the demand curve over the YX (orY0X) values of q. In practice, the difference between the BCXY and BAXY (or between BB00Y0Y and BB0Y0Y) quadrangles is an acceptable approximation, although it certainly overesti-mates the true value of congestion costs.

Alternatively, congestion costs can be defined as the difference between the consumer’s surplus plus the govern-ment’s gain after and the consumer’s surplus before the introduction of an optimal charge. In practice, there is a slight difference (20 thousands euros per day) that may reflect the overestimate just mentioned.

A third approach to congestion costs is to look at the changes introduced by a move from the pre-charge situation (A) to the optimal situation (B). They are equal to the difference between the benefits of the move to the X remaining road users, LGEP, and the losses it inflicts upon theY–Xexcluded road users, BAG. This approach produces numbers approximately similar to the numbers produced by the other two approaches.

The benefits of the charge policy are the reduction in congestion costs, relative to the optimal situation, the reference situation.

Collection costs can be estimated with TfL data (www. tfl.gov.uk). Operating costs in 2003–2004 were 138.8 million euros. They consist mostly of the payment made to Capital, the private entity to which the operation of the scheme was contracted out. Such payments obviously include the depreciation and the opportunity cost of capital of the investments made by Capital. But in addition, the UK government made investments in the pre-charge period. Investment costs over the 2000–2003 period are reported to be 245.7 million euros. Assuming a 5% opportunity cost of capital7and a (rather conservative) 10% depreciation rate, these investment costs amounts to 36.9 M. euros per year. Total collection costs in 2003 were therefore about 175.7 M. euros per year, or 689 thousand euros per chargeable day.

7. Questions about the London congestion charge scheme

Table 1, which is the heart of our work, allows a number of interesting conclusions.

How important are (were) congestion costs?—First, Table 1tells us how important congestion costs (as defined here) in the charged zone were before the introduction of the charge, or, in other words, what was at stake. Congestion costs amounted in 2002 to about 296 thousand euros per chargeable day8. This is about 75 M. euros per year (excluding congestion on week-ends and other days excluded from the congestion charge). This is what a congestion charge is expected to eliminate, and this elimination is the main raison d’eˆtre and the main benefit of such a system. How important is it?

It is a very small part of the GDP of London, and even of the GDP generated in the chargeable zone. In 2001, the GDP of Greater London, or more precisely generated in the Greater London area, was 255,000 million euros. Conges-tion costs in the chargeable zone represented a mere 0.03% of the economic output of Greater London.

There were in 2001, about 4.5 M. workers in the Greater London area, and 1.2 M. workers in the chargeable zone. Assuming that labour productivity was the same in the chargeable zone and in the Greater London area—a very conservative estimate, because this productivity is likely to substantially higher—we can estimate the output of the chargeable zone at 68,000 M. euros per year. Congestion costs in this area represented about 0.11% of the GDP of the area. This is very much in line with the findings of

Prud’homme (1999, 2000)for the Paris area.

Table 1

Motor vehicle transport in the london congestion charge zone Pre-charge situation Present situation Optimal situation Road usageq(1000 veh km) 1390 1160 1055

Speeds(km/h) 14.3 16.3 18.5

Time for 1 km (min) 4.2 3.6 3.2 Individual costI

(euro/veh km)

1.61 1.36 1.28

Social costS(idem) 3.38 2.39 2.09

Charge (idem) – 0.56 0.81

Marginal congestion cost (idem)

1.77 0.46 –

Congestion costs (1000 euros/day)

296 24 –

Benefitsa_{(idem) – 272 296}

Charge proceeds (idem) – 650 854 Collection costs (idem) – 689 689 Benefits net of costs – K417 K393 Source: See Annex A.

a _{Benefits for bus users, for increased reliability, and environmental}

improvements are not included.

7_{The UK Government’s test discount rate is reported to be 3.5%; using}

this low value (can the UK Government really finance all public investments with an internal rate of return higher than 3.5%?) would decrease our estimate of collection costs by 3.7 M. euros.

8_{This is 4.3 times less than the number produced by the naive and}

frequently used method of comparing the effective cost (1.61) with the zero road usage cost (0.81) and multiplying by the effective road usage (1390).

Congestion costs can also be related to the utility derived from motor vehicle usage. This utility is equal to what users pay, plus the consumer surplus they obtain, that is to area RAXO in Fig. 1. In 2002, this can be estimated to 3,579 thousands euros per day, to be compared with the 296 thousands euros per day of congestion costs. This is a ratio of about 8%. Thus, in 2002, traffic congestion costs represented about 8% of the utility generated by traffic.

Is the present charge optimal?—Second,Table 1tells us whether the present level of the congestion charge (5£ per day) is optimal or not. On the one hand, it can be said that the charge level is too low. The optimal road usage would be require a move fromY0toY, that is a further 9% reduction in traffic. This would be obtained with an increase in the charge level of 0.56 to 0.81 euro/vehicle km, a 45% increase. Because of a rough proportionality between the charge per day and the charge per vehicle km, this means that the charge should be increased from 5 to 7.2£ per day. On the other hand, it must be observed that the economic benefits associated with such an increase would be very small. This increase would reduce congestion costs, but would reduce them by only 24 thousand euros per day. The present charge already captures nearly 90% of the potential benefits of a charge. Increasing the charge by 45% to increase benefits by 10% would meet with some resistance. This finding is also dependent upon the value of time. With the Paris value of time, a different result is obtained. The present charge level appears very close to the optimal level (B0_{andBbecome very close).}

It can also be observed that a reduction in the relatively high level of fraud, which would result in an increase in the effective charge (as calculated), would also contribute to make the formal 5£ per day charge closer to optimal.

Are charge proceeds greater than economic benefits?— Table 1makes it possible to compare the charges proceeds, the amount of money that is collected, with the economic benefits of the system. It appears that presently, the ratio is 2.4. With an optimal charge, it would be 2.9. In other words, what users pay in charges is two or three times larger than what they get in congestion reduction. Similar or even higher ratios are common in congestion charge schemes. This does not worry economists. They note that charges, unlike congestion, are not economic costs. Charges are transfers, and the product of the charge can be put to useful, welfare producing, uses. This (correct) view is not always easily accepted by the general public.

Is the congestion scheme economically justifiable?—The standard economic theory of congestion ignores manage-ment and collection costs, and assumes them to be zero. In Table 1, it looks only at the line ‘benefits’, sees a positive number, and concludes that the scheme is justified. In reality, operating a system like the one that has been introduced in London is costly. It involves the use of economic resources, and the expenditures made to that effect are indeed economic costs. We considered the investments, which have been made for the system, to

determine an investment component of the yearly cost (equal to 5% for the opportunity cost of capital, plus 10% for the depreciation), and added it to the operation component. The result is very high indeed. It is presently roughly equivalent to the charge proceeds. It would be lower than the charges proceeds with a higher, optimal, charge. In any case, it is much higher than the economic benefits of the scheme. The benefits net of costs of the scheme appear to be negative.

8. Other costs and benefits

Several other associated issues can be discussed. Environmental benefits—Less vehicles km at a lower speed means less pollutants produced, and lower pollution costs. Curiously, this benefit seems not to be appraised by TfL, probably because no improvement in air quality has been recorded in 2003. This is because vehicles km driven in the charged zone represent a small fraction (about 1%) of total vehicles km driven in the London agglomeration. Air quality in London depends upon total emissions, so that even if driving was completely eliminated in the charged zone, total emissions would only decrease by about 1%, and the improvement in air quality would hardly be noticeable. The benefits can nevertheless be estimated—and valued.

Vehicles km decreased by 230 thousands (1,390K_1,160) per day. Taking the official French value of pollution costs in dense urban areas of 29 euros per 1,000 vehicle km (Boiteux, 2001), this translates into 6,670 euros per day or 1.7 M. euros per year.

The remaining vehicles are driven at an increased speed. The elasticity of pollution to speed is a least equal toK2 at urban speeds9. A 17% increase in speed means a 34% decrease in pollution emissions. This translates into an additional benefit of 11,440 euros per day or 2.8 M. euros per year.

A similar calculation can be made for the reduction in CO2 emissions. Taking again the official French value of 7 euros per 1,000 vehicles km, the benefit associated with a reduction in traffic of 230 thousands vehicle km can be estimated at 0.4 M. euros per year.

Total environmental benefits generated by the congestion charge (ignoring additional emissions by additional buses) can be estimated at 4.9 M. euros per year. This is not negligible, but it does not substantially change the economics of the scheme.

Benefits to former bus users–The speed of buses is reported to have increased by 7%. This is a benefit for the people who were travelling by bus, and it is a benefit caused by the congestion charge. Bus users, numbering 356,000, 9_{This rough estimate is based on unpublished data communicated by}

UTAC, an independent organization that measures emissions of pollutants by new vehicles.

R. Prud’homme, J.P. Bocarejo / Transport Policy xx (2004) 1–9

part, representing amortisation and fuel costs, and of a variable part, which is the value of the time spent driving 1 km. The fixed part is estimated (Glaister, 2003) to be 0.15 (euros per km)6. The variable part is equal to the time spent (t, in hours), which is a function of speed (s, in km/hour), which is itself a function of road usage (q), multiplied by the value of time (v, in euros per hour):

IðqÞZ0:15CtvZ0:15C½1=sðqÞv (1) To go further, we need an estimate ofs(q), the speed as a function of road usage, and ofv, the value of time. For the value of time, the ROCOL (2000) report proposes 15.6 euros per hour. This is high number. The values used in the Paris region, particularly to justify transport investments, are significantly below, and they are considered high by many. We will nevertheless keep this 15.6 euros per hour estimate in this paper. Since, there are on average 1.34 persons per vehicle, this puts the value of time per vehicle at 20.9 euros per hour.

Speedsis a declining, and largely linear, function of road usageq:

sZaKbq (2)

a, the speed on empty roads (when road usage qZ0) is given by TfL as 31.6 km/h. Since we know the average speed in 2002 (whenqwas equal to 1,390) which was equal to 14.3 km/h, we can calculate b, which turns out to be 0.01245. We therefore have:

IðqÞZ0:15C20:9=ð31:6K0:0124qÞ (3) The social cost curveS(q) can easily be deduced from

I(q). It is equal to the individual cost curve I(q), plus

the derivativeI0(q) multiplied by road usageq:

SðqÞZIðqÞCI0ðqÞq (4)

SðqÞZ0:15C20:9=ð31:6K0:0124qÞ

C0:26q=ð31:6K0:0124qÞ2 (5)

5. Demand curve for London charged zone

The following step is to determine the equation of the demand curveD(q). We know one point of this curve, the equilibrium point A in 2002, because we know the speed at the time. Its coordinates are 1.61 (euros per vehicle km and 1,390 (thousand vehicles km per day). We can also figure out the coordinates of pointB0, the equilibrium point in 2003 after the charge, for which we already know the number of thousands vehicles km per day, Y0Z1,160. The individual unit cost for this point is equal to the fixed cost plus the time cost plus the charge paid.

The first two elements are given by equationI(q). With

qZ1,160 we haveI(q)Z1.37 euro. This is a measure ofE0_Y0 or P0 in Fig. 2. The average charge paid per vehicle km driven can be determined by dividing the total charge collected by the number of vehicles km. The amount of the yearly charge is 115 million pounds, or 165.6 euros. Since there are about 255 chargeable days per year, this is 451,000 pounds or 649,000 euros per chargeable day—and 0.56 euro per vehicle km. The unit cost borne by users is therefore increased to 1.93. This, by the way, indicates aK0.83 price elasticity of demand for road usage in the chargeable zone.

A andB0are both on the demand curveD(q). With the coordinates ofAandB0, it is easy to calculate the equation of the demand curve:

DðqÞZ3:54K0:00139q (6)

Unit costs

Social cost S(q)

3.38 N C

2,39 Q B’’

2.09 M B Individual 1.93 M’ B’ cost I(q)

1.61 L G A 1.37 P’ E’ F’

1.28 P E F D(q)

0.81 J H

Road usage (q) Y Y’ X

1055 1160 1390

Fig. 2. Road congestion with a congestion charge.

6_{This is an approximation; fuel consumption is also in part influenced by} speed, which is also influenced by road usage; but the estimates provided do not vary much (from 0.14 to 0.16); we retained 0.15 for the sake of simplicity.

6. Significant magnitudes for London charged zone

Equipped with these equations, we can determine the coordinates of all the points represented in Fig. 2, and produce the numbers inTable 1.

The optimal situation (qZY) is obtained, as mentioned before, when the demand curve and the social cost curve intersect, that is whenS(q)ZD(q).

Congestion costs are defined as BCA in the case of the pre-charge situation, and as BB00B0 in case of the present situation. By definition, congestion costs are zero in the case of the optimal situation. In principle, these costs are defined as the difference between the integrals of the social cost curve and the demand curve over the YX (orY0X) values of

q. In practice, the difference between the BCXY and BAXY (or between BB00Y0Y and BB0Y0Y) quadrangles is an acceptable approximation, although it certainly overesti-mates the true value of congestion costs.

Alternatively, congestion costs can be defined as the difference between the consumer’s surplus plus the govern-ment’s gain after and the consumer’s surplus before the introduction of an optimal charge. In practice, there is a slight difference (20 thousands euros per day) that may reflect the overestimate just mentioned.

A third approach to congestion costs is to look at the changes introduced by a move from the pre-charge situation (A) to the optimal situation (B). They are equal to the difference between the benefits of the move to the X

remaining road users, LGEP, and the losses it inflicts upon theY–Xexcluded road users, BAG. This approach produces numbers approximately similar to the numbers produced by the other two approaches.

The benefits of the charge policy are the reduction in congestion costs, relative to the optimal situation, the reference situation.

Collection costs can be estimated with TfL data (www. tfl.gov.uk). Operating costs in 2003–2004 were 138.8 million euros. They consist mostly of the payment made to Capital, the private entity to which the operation of the scheme was contracted out. Such payments obviously include the depreciation and the opportunity cost of capital of the investments made by Capital. But in addition, the UK government made investments in the pre-charge period. Investment costs over the 2000–2003 period are reported to be 245.7 million euros. Assuming a 5% opportunity cost of capital7and a (rather conservative) 10% depreciation rate, these investment costs amounts to 36.9 M. euros per year. Total collection costs in 2003 were therefore about 175.7 M. euros per year, or 689 thousand euros per chargeable day.

7. Questions about the London congestion charge scheme

Table 1, which is the heart of our work, allows a number of interesting conclusions.

How important are (were) congestion costs?—First,

Table 1tells us how important congestion costs (as defined here) in the charged zone were before the introduction of the charge, or, in other words, what was at stake. Congestion costs amounted in 2002 to about 296 thousand euros per chargeable day8. This is about 75 M. euros per year (excluding congestion on week-ends and other days excluded from the congestion charge). This is what a congestion charge is expected to eliminate, and this elimination is the main raison d’eˆtre and the main benefit of such a system. How important is it?

It is a very small part of the GDP of London, and even of the GDP generated in the chargeable zone. In 2001, the GDP of Greater London, or more precisely generated in the Greater London area, was 255,000 million euros. Conges-tion costs in the chargeable zone represented a mere 0.03% of the economic output of Greater London.

There were in 2001, about 4.5 M. workers in the Greater London area, and 1.2 M. workers in the chargeable zone. Assuming that labour productivity was the same in the chargeable zone and in the Greater London area—a very conservative estimate, because this productivity is likely to substantially higher—we can estimate the output of the chargeable zone at 68,000 M. euros per year. Congestion costs in this area represented about 0.11% of the GDP of the area. This is very much in line with the findings of

Prud’homme (1999, 2000)for the Paris area. Table 1

Motor vehicle transport in the london congestion charge zone Pre-charge

situation

Present situation

Optimal situation Road usageq(1000 veh km) 1390 1160 1055

Speeds(km/h) 14.3 16.3 18.5

Time for 1 km (min) 4.2 3.6 3.2

Individual costI

(euro/veh km)

1.61 1.36 1.28

Social costS(idem) 3.38 2.39 2.09

Charge (idem) – 0.56 0.81

Marginal congestion cost (idem)

1.77 0.46 –

Congestion costs (1000 euros/day)

296 24 –

Benefitsa_{(idem) – 272 296}

Charge proceeds (idem) – 650 854

Collection costs (idem) – 689 689

Benefits net of costs – K417 K393

Source: See Annex A.

a _{Benefits for bus users, for increased reliability, and environmental} improvements are not included.

7_{The UK Government’s test discount rate is reported to be 3.5%; using} this low value (can the UK Government really finance all public investments with an internal rate of return higher than 3.5%?) would decrease our estimate of collection costs by 3.7 M. euros.

8_{This is 4.3 times less than the number produced by the naive and} frequently used method of comparing the effective cost (1.61) with the zero road usage cost (0.81) and multiplying by the effective road usage (1390).

Congestion costs can also be related to the utility derived from motor vehicle usage. This utility is equal to what users pay, plus the consumer surplus they obtain, that is to area RAXO in Fig. 1. In 2002, this can be estimated to 3,579 thousands euros per day, to be compared with the 296 thousands euros per day of congestion costs. This is a ratio of about 8%. Thus, in 2002, traffic congestion costs represented about 8% of the utility generated by traffic.

Is the present charge optimal?—Second,Table 1tells us whether the present level of the congestion charge (5£ per day) is optimal or not. On the one hand, it can be said that the charge level is too low. The optimal road usage would be require a move fromY0toY, that is a further 9% reduction in traffic. This would be obtained with an increase in the charge level of 0.56 to 0.81 euro/vehicle km, a 45% increase. Because of a rough proportionality between the charge per day and the charge per vehicle km, this means that the charge should be increased from 5 to 7.2£ per day. On the other hand, it must be observed that the economic benefits associated with such an increase would be very small. This increase would reduce congestion costs, but would reduce them by only 24 thousand euros per day. The present charge already captures nearly 90% of the potential benefits of a charge. Increasing the charge by 45% to increase benefits by 10% would meet with some resistance. This finding is also dependent upon the value of time. With the Paris value of time, a different result is obtained. The present charge level appears very close to the optimal level (B0andBbecome very close).

It can also be observed that a reduction in the relatively high level of fraud, which would result in an increase in the effective charge (as calculated), would also contribute to make the formal 5£ per day charge closer to optimal.

Are charge proceeds greater than economic benefits?—

Table 1makes it possible to compare the charges proceeds, the amount of money that is collected, with the economic benefits of the system. It appears that presently, the ratio is 2.4. With an optimal charge, it would be 2.9. In other words, what users pay in charges is two or three times larger than what they get in congestion reduction. Similar or even higher ratios are common in congestion charge schemes. This does not worry economists. They note that charges, unlike congestion, are not economic costs. Charges are transfers, and the product of the charge can be put to useful, welfare producing, uses. This (correct) view is not always easily accepted by the general public.

Is the congestion scheme economically justifiable?—The standard economic theory of congestion ignores manage-ment and collection costs, and assumes them to be zero. In

Table 1, it looks only at the line ‘benefits’, sees a positive number, and concludes that the scheme is justified. In reality, operating a system like the one that has been introduced in London is costly. It involves the use of economic resources, and the expenditures made to that effect are indeed economic costs. We considered the investments, which have been made for the system, to

determine an investment component of the yearly cost (equal to 5% for the opportunity cost of capital, plus 10% for the depreciation), and added it to the operation component. The result is very high indeed. It is presently roughly equivalent to the charge proceeds. It would be lower than the charges proceeds with a higher, optimal, charge. In any case, it is much higher than the economic benefits of the scheme. The benefits net of costs of the scheme appear to be negative.

8. Other costs and benefits

Several other associated issues can be discussed. Environmental benefits—Less vehicles km at a lower speed means less pollutants produced, and lower pollution costs. Curiously, this benefit seems not to be appraised by TfL, probably because no improvement in air quality has been recorded in 2003. This is because vehicles km driven in the charged zone represent a small fraction (about 1%) of total vehicles km driven in the London agglomeration. Air quality in London depends upon total emissions, so that even if driving was completely eliminated in the charged zone, total emissions would only decrease by about 1%, and the improvement in air quality would hardly be noticeable. The benefits can nevertheless be estimated—and valued.

Vehicles km decreased by 230 thousands (1,390K_1,160)

per day. Taking the official French value of pollution costs in dense urban areas of 29 euros per 1,000 vehicle km (Boiteux, 2001), this translates into 6,670 euros per day or 1.7 M. euros per year.

The remaining vehicles are driven at an increased speed. The elasticity of pollution to speed is a least equal toK2 at urban speeds9. A 17% increase in speed means a 34% decrease in pollution emissions. This translates into an additional benefit of 11,440 euros per day or 2.8 M. euros per year.

A similar calculation can be made for the reduction in CO2 emissions. Taking again the official French value of 7 euros per 1,000 vehicles km, the benefit associated with a reduction in traffic of 230 thousands vehicle km can be estimated at 0.4 M. euros per year.

Total environmental benefits generated by the congestion charge (ignoring additional emissions by additional buses) can be estimated at 4.9 M. euros per year. This is not negligible, but it does not substantially change the economics of the scheme.

Benefits to former bus users–The speed of buses is reported to have increased by 7%. This is a benefit for the people who were travelling by bus, and it is a benefit caused by the congestion charge. Bus users, numbering 356,000,

9_{This rough estimate is based on unpublished data communicated by} UTAC, an independent organization that measures emissions of pollutants by new vehicles.

gained 1.34 min per person per day, which represent 124 thousand euros per day, or 31 M. euros per year. This is almost equivalent to half the benefit enjoyed by car users.

Benefits associated with charge-financed expenditures. It is often claimed that ear-marking of the proceeds of the charge can produce benefits that should be taken into account. If this income is invested in public transport, or in road extension, it is argued, this will produce social and economic benefits, which should be added to the main benefits of the scheme. This is not correct. Such expendi-tures will indeed produce social benefits (it is extremely difficult to spend public money without producing benefits) but these benefits should not be added to the benefits of the scheme. They are nothing but the counterpart of the social cost of the charge payment. Either we ignore both this social cost and this social benefit (this is what economists suggest when they say that the charge is a ‘transfer’), or we value both of them. But counting the benefit and ignoring the cost (or vice-versa) is not a reasonable option. Ear-marking (or hypothecation, as it is called nowadays) does not add anything at all. Spending the charge proceeds on transport expenditures might create utility, but spending it on health or education would also be useful, and presumably equally useful.

Ear-marking may be politically very expedient because it makes it easier to sell a congestion charge. But by itself, it does not produce additional economic benefits.

A qualification may be added. Most taxes are distortive. They embody incentives that tend to discourage work or savings or investment, and to decrease output. This is why taxes with expenditures are, in general, not merely transfers, but imply a welfare loss, also called the opportunity cost of public funds. The magnitude of this welfare loss varies with the nature of the tax and the economy considered, and is not well known. It could be in the 10–30% range. A congestion tax, by contrast, is not distortive. More precisely, it distorts behaviours in a desired direction. If a congestion tax replaces an ordinary tax, the distortive effect of the ordinary tax will be saved, and the welfare loss decreased. In this case, it would be justified to consider 10–30% of the congestion tax proceeds as a benefit of the tax. Note that this would apply to non marked taxes as well as to ear-marked taxes. As a matter of fact, it would apply more convincingly to non ear-marked taxes, because ear-marking suggests addition rather than substitution.

In the case of the London congestion charge, this discussion is largely rhetorical. It obviously applies to tax proceeds net of collection costs and we have seen that collection costs are nearly equal to charge proceeds.

Increase in bus supply–The congestion charge has been introduced jointly with another measure: a significant increase in bus supply. It is reported that some 250 new buses were purchased and are operated. Bus ridership in the zone increased. The two policy measures were obviously complementary. Without new buses, bus crowding would

have increased, and the quality of bus trips declined for all bus users—a typical congestion phenomenon.

If bus transportation were an ordinary, market-driven, good, this would not cause any additional cost or benefit. More bus transportation would be supplied as a result of an increase in the demand for bus transportation. But bus transportation is not an ordinary good. In London, as in most other cities of the developed world, it is heavily subsidised, and users pay about half the economic cost of it. This implies a welfare loss—which is much lower than the amount of subsidy.

Bus transportation costs are independent of bus trans-portation quantity: marginal costs are equal to average cost in this industry. The additional bus supply and demand can therefore be treated as the initial bus supply and demand. This makes it possible to provide a gross estimate of the welfare cost associated with the recorded increase in supply and patronage.

Let us consider the demand for bus transportation AB, the unit cost CC0, and the price paid PP0.Cis much higher thanP, and PC is the unit amount of the subsidy (Fig. 3).

In the absence of subsidy, the equilibrium would be inA, with Qa unit of bus transportation consumed, at a priceC. With a unit subsidy equal to PC, the price paid by users isP, and the quantity consumed becomes Qb. The total cost of providing the service is CEQbO. The total amount of the subsidy is CEBP. The additional welfare gain generated by the subsidy, and the increase in demand, is the increase in consumer surplus: CABP. The additional economic cost generated by the subsidy is AEQbQa. The variation in welfareD_{Wgenerated by the subsidy is, therefore:} D_WZCABPAEQbQa

This can be simplified with a few not unreasonable hypotheses. Let us assume that the demand elasticity isK_1,

and that the subsidy equals 50% of costs. It is easy to see that in such a case, CADPZDBQbQa, and that:

DWZAEBZ1=8CEQbO

In other words, the welfare change, which is negative, is equal to one-eighth of the total cost of bus transportation.

Unit cost

R

C A E C’

P D B P’

O Qa Qb Quantities

It is reported that some 250 additional buses, purchased at a cost of 100 M. euros and operated at a yearly cost of 38 M. euros, have been introduced to accommodate the increased demand in bus transportation. Assuming an opportunity cost of capital of 5% and an amortisation rate of 10%, the yearly economic cost is 53 (15C_{38) M. euros.}

The welfare loss associated with an increase in subsidised bus transportation can therefore be estimated at about 7 M. euros per year.Value of time–As mentioned earlier estimates of congestion costs are obtained with an official value of time of 15.6 euros per hour. In France, the official value of time for the Paris region, as indicated in

Boiteux (2001), is only 8.8 euros per hour. If Paris values of time were used in London, many of the numbers estimated in this paper would be substantially changed.

Congestion costs (which are proportional to the value of time) would be decreased by 45%. Yearly congestion costs would be valued at 36 million euros per year. The benefits of congestion reduction would be reduced similarly. So would the benefits to (former) bus users. The present level of the congestion charge (5£) would probably appear too high. Since the costs would not be altered by this change, the estimated gap between costs and benefits would greatly increase.

Selecting an appropriate value of time is a delicate task. The difference between London and Paris is a priori hard to justify—although we cannot exclude the possibility that the value of time be grossly underestimated in France. It has been argued that the value selected in London is particularly high because the share of business trips (with a high value of time attached) is particularly high in the car trips made in the charge zone. This is a meaningful argument. It would probably imply the selection of a lower value of time to estimate the benefits for bus users. And it would suggest that benefits of a congestion charge would be even lower in less business-oriented areas, and/or in less developed countries or cities.

9. Conclusions

This quantitative—and tentative—exercise has produced some preliminary findings. First, the supposedly high and

unbearable congestion costs that motivated the introduction of a congestion charge were in reality relatively modest: about 0.1% of the GDP produced in the charged zone. Second—as predicted by theory—these congestion costs have been largely eliminated by the congestion charge, and this elimination represents an economic gain. Third, the proceeds of the charge are about two-and-a-half times larger than this economic gain.

Fourth, and this might be the most important finding of this study, the economic costs associated with the system are larger than the economic gains it generates. Table 2

summarises these costs and benefits.

The gap between the two appears substantial. The economic benefits represent less than 60% of the economic costs.

These findings are preliminary. They are based on published Transport for London data on speed and road usage before and one year after the scheme. They use a generous value of time. They assume that recorded changes were caused by the scheme. They are focused on the charged zone only, and ignore what might have happened outside the zone as a result of the charge system. We do not even know whether congestion in the rest of London decreased (because of complementarity) or increased (because of substitution). It could be that congestion decreased because the number of trips to the charged zone decreased, but it could also have increased because some drivers are now going around the zone in order to avoid paying the charge. Our findings also ignore a likely gain in transportation reliability experienced by both car and bus users, which is hard to measure and harder to value. The very high operation and collection costs, which are the main flaw of the charge, may well decrease substantially in the future. Finally, it does not take into account changes in business or residential location that could be induced by the congestion charge over the course of time.

Additional studies are required to get a better under-standing of the economic consequences of this important policy experiment. However, the gap between costs and benefits appears so large at this stage that it is difficult to see how additional information could eliminate it, and turn it into a substantial net gain10.

Appendix A: Calculations forTable 1

Eqs. (3), (5) and (6) are the equations of the the curves

I(q),S(q) andD(q) ofFig. 2. They are given in the text. They make it possible to calculate the coordinates of all the points Table 2

Benefits and costs of the london congestion charge

Per day (1,000 e) Per year (million e) Benefits

Reduction in congestion costs 272 68 Increased speed for bus users 124 31

Environmental benefits 20 5

Total, recorded benefits 414 104

Costs

Implementation costs 172

Subsidy to buses 18 5

Total, recorded costs 707 177

10_{Londoners who might not like this critical appraisal by Parisians will} perhaps be relieved to know that we are much more critical of the policy conducted in Paris. It has consisted in reducing driving space, in order to increase congestion, in the hope of inducing car drivers to shift to buses. Congestion has indeed increased, but bus patronage has actually declined. This, however, is another story.

and areas inFig. 2, i.e. of the magnitudes of interest reported inTable 1.

Eq. (6), D(q)Z3.5K_0.0013q, yields the coordinates of

A(1390; 1.61), of B(1055; 2.09) and of B0(1160; 1.93), as indicated in the texte. This impliesLZ1.61,MZ2.09 and

M0_Z1.93.

Eq. (3), I(q)Z0.15C20.9/(31.6K0.1245q), yields the coordinates of E(1055; 1.28) and E0(1160; 1.37), which impliesPZ1.28 andP0Z1.37.

Eq. (5), S(q)Z0.15C20.9/(31.6K0.0124q) C0.26q/ (31.6K_0.0124q)2, yields the coordinates of C(1390; 3.38) andB00(1160; 2.39), which impliesNZ3.38 andQZ2.39.

We, therefore, have as reported inTable 1, the individual cost I for the pre-charge situation (LZ1.61), the present situation (P0Z1.37 and the optimal situation (PZ1.28). We also have the social costSfor the pre-charge situation (NZ3.38), the present situation (QZ2.39), and the optimal situation (2.09).

The present chargeE0B0has been calculated directly to 0.56 (we can verify that is equal toM0K_P0_Z1.93K_1.37).

The optimal charge would be EBZ2.09K1.28Z0.81. The marginal congestion cost in the pre-charge situation is ACZ3.38K1.61Z1.77, and in the present situation

B0B00Z2.39K1.93Z0.46.

The congestion cost CC in the pre-charge situation is defined as the integral of S(q) between Y and X minus the integral ofD(q) between the same values ofq. This can be approximated as CCZBCAZYBCXKYBAX. YBCXZ0.5(YBCXC)YXZ0.5(2.09K3.38)(1390K 1055)Z916. YBAXZ0.5(YBC_{XA)YXZ0.5(2.09}C

1.61)(1390K1055)Z920. We, therefore, have CCZ 916K620Z296. The congestion cost in the present situation is BB00_B0 _{and can similarly be calculated to be} 24. By definition, congestion costs in the optimal situation are zero.

Two other approaches to the calculation of congestion costs can be utilised. The first one compares the consumer’ surplus (including the charge paid) in the optimal situation (PRBE) with the consumer’s surplus in the pre-charge situation (LRA): CCZPRBEKLRA. Noting that

MBRK_LRAZK_{LMBA, CC becomes: CCZPMBE}K

LMBAZ(2.09K_1.28)1055K_0.5(1055C_1390)(2.09K

1.61)Z268.

The second approach to congestion benefits (again defined as the benefits of moving to the optimal situation) compares the time gain of the Y road users in the optimal situation (PLGE) with the loss to the XKY users excluded by the charge (GBA): CCZPLGEKGBA. PLGZ(LKP)YZ(1.61K1.28)1055Z348. GBAZGB! GA!0.5Z0.5(2.09K1.61)(1390K1055)Z80. Hence CCZ348K80Z268.

These two complementary approaches give identical results. These results are somewhat lower (by 28, or 9%) than the result obtained by the first method. This is because the straight line BC is an approximation of the curve BC that should in principle be considered to estimate BCA. We nevertheless retained the higher figure, 296 rather than 268, to be on the safe side.

References

Banister, D., 2003. Critical pragmatism and congestion charging in London. International Social Science Journal 176, 249–264.

Boiteux, M., 2001. (Report of a Commissariat Ge´ne´ral du Plan committee chaired by). Transports: Choix Des Investissements Et Couˆts Des Nuisances. La documention franc¸aise, Paris.

Glaister, S., et al., 2003. Transport Pricing and Investment in England. Imperial college (Department of civil Engineering).

Prud’homme, R., 1999. Les couˆts de la congestion dans la re´gion parisienne. Revue d’Economie Politique 109 (4), 425–441.

Prud’homme, R., 2000. La congestion et ses couˆts. Annales des Ponts et Chausse´es. 94, 13–19.

Prud’homme, R., Yue-Ming, S., 2000. Le couˆt e´conomique de la congestion du pe´riphe´rique parisien: une approche de´sagre´ge´e. Les Cahiers Scientifiques du Transport 37, 59–73.

ROCOL, 2000.Anon., 2000. Road Charging Options for London: A Technical Assessment. Report for the Government Office foir London. HMSO, London.

Smeed, R.J., 1964. Road Pricing: The Economic and Technical Possibilities. Her Majesty’s Stationary Service, London (report of a Commission chaired by).

Walters, A., 1961. The theory and measurement of private and social cost of highway congestion. Econometrica 29, 676–699.