0 ~~~~~~~ti B L I C I' O 1, I C A F O R 'r 1f E SEP9TF.NfIBER 109'-
Privatesector
' D>:zm uSt z t| |1 8956
7 The Restructurin anc Pivaizain
of thet ILK Elecriit SU_______________
\NWIs It Vorthi It-
David IM.Neweiu a
Jlicbael G. Pollf
11 A Retrospectixv on tht Mexi.ln
Toll Road 1'rogr ( 1989 94) 1
icfRzits/er
19 The Private Sectir in Watcr and
Sanitation-I osP\ to Get Started L
Peteknefoe Brook, Cinena
23 Privatization and Restructuring in
Central and EalStern Europe
RKobert E AS ijderslim, Sim12 emi ~
Dj/zitkoi, (GerIird Pobl. (attc.
St/jil C'lesseos
27 The Drivers of the Information
Revolution-Cost, C(omputing
Poxver, and Coilnvergence
fumes 1ou
31 Telecoummunication-i Is I)ead.
Li ng Live Nettorking A
Jamies Boor!.
35 Tcleciomrnunicatiiins Refot im- r-
FlOx to SLCCeelrd
B/br,za We/eo izns [~6 *'557
39 Lilberalizing TeILreoFIIntWnicationns
andf the Roil of the NXorll Trade
Organiz.ation
Cr2rlos, - Primo Brguta
47 W1hat the TIrYansforimation ofn 4
Telecrom Mlarkets Means for-
R,gUilatitiOn
51 The Pr' ate S,ector and the r7?1TR ~
Itiiternet
(Cad/s -A Pi*iii rimofi.(Ca 611d
Carx/no Fink
The World Bank Group * Finianice, Private Sector, nmd Infrastructure Net,,\irlk
Private Sector is an open forum intended to
encourage dissemination of and debate on
ideas, innovations, and best practices for ex-
panding the private sector. The views pub-
lished are those of the authors and should not
be attributed to the World Bank or any of its
affiliated organizations. Nor do any of the con-
clusions represent official policy of the World
Bank or of its Executive Directors or the coun-
tries they represent.
Private Sectoris a quarterly publication distrib-
uted free of charge. To subscribe, please send
your name, mailing address, telephone num-
ber, and fax number to the editor (Suzanne
Smith, Room F6P-188, The World Bank, 1818 H
Street, NW, Washington, D.C. 20433, email:
ssmith7@worldbank.org, fax: 202 522 2961,
telephone: 202 458 1111).
Most Notes from Private Sector are also avail-
able on-line. There is full text in HTML format
for on-screen viewing or a downloadable file in
Adobe's PDF format (http://www.worldbank.
org/html/fpd/notes/notelist.html).
Mn,
Private sector
Quarterly No.1I1, September 1997
The Restructuring and Privatization of the U.K. Electricity Supply-Was It Worth It? 7
David M. Newbery and Michael G. Pollitt report on a studcy that assesses the costs and benefits of the
Frestrcturing and pri- atization of the electricity supplv sector in England and -\Vales. Since privatization, real unit
costs haxe fallen by ahout 50 percent and real pool prices by 20 percent, and a switch from coal to gas hals
ConltrihLuteCd to a dIop in polluting emilissions. But sorme of the positive changes can he attributed to external
cfators socIh as the development of high-efficiency combined cycle turhines and the fall in coal prices. So the
authors estimate the net henefits of privLtizLtion over the period 1995-2010 ly comparing the historical and
projected path of the privatized incdustry Nwitlh what miglht have happened had the assets stavecl uncler public
ow\ nership. They estimate the overall net henefits from privaltization at Al 1.9 billion. Costs fall. hut mucil of the
benefit goes to the generators-because prices in the wholesale market fall muchL less. The authors suggest that
| more competition may help to reduce these prices.
A Retrospective on the Mexican Toll Road Program (1989-94) 11
In 1989. Mexico initiatecl a private toll road program of fifty-three concessions inv olving an inivestment of al)out
5I bSS i 3illion in limited reCourse financing over the period 1989-94. The program more than doubled the size
of the national toll road network. but miscalculations of inv estment costs and overoptimistic forecasts of operating
0| Cincom1e uncdermiiiined the viahility of the toll ioads. An already bad situation xw as made wkorse b)y the Mexican
Currency criSis of Deeemlher 1994 and the private toll road initiative came to a \irtual standstill. Local commiler-
cial hianks \vere saddled w ith nonperforming loans estimzatecd at USS4.5 billion to USS5.5 billion. Concession-
airCs and their- affiliates have been faced with w,\r iting off large portions of their investmtients. Ancl users w-ere left
4i | i Nwitl som1e of the highoest tolls in the world. Jeff Ruster diagnoses the fla s in the clesign of the program and
showvs howN the failure was manifested in the implemenitation of clifferent phases of the projects.
The Private Sector in Water and Sanitation-How to Get Started 19
T'ihe moie risk and responsibility a government hands over to the private sector in water and sanitation, the
i:| more powNerful the incentives for better perforimanice-but also the moie clemands on the government in
cormimtitment and preparation. So a governmient about to enter into a long partnershiip foi a water concession or
b| uildl-operate-trainsfer arrangemient-typically for twenty-five to thirtv years-needs to he suLe that it does not
ov erlook cletails that \ill later land it in messy renegotiations. A lease is less demandling, but offers smaller gains
and wvill not fix sulIC problems as chroniC underinvestiment. It will, however, give the goverrnment time to
ipreplie Li longer-terim option. In this Note. hased on toolkits recently published by the Wo'rld Bank. Penelope
J. Brook Cowen sets out the range of options for involving the private sectoir in wxate and sanitation and
reviews the lessons oni -viat can make ot break a private participation process.
Privatization and Restructuring in Central and Eastern Europe 23
One of the Ilost imilportant policy questions in the transitioIn economilies is x hat go\ ernioments can clo to speed
the restruCturing of firms and thus hasten the transitioni to a mature market economiy. Robert E. Anderson,
Simeon Djankov, Gerhard Pohl, LincI Stijn Claessens report on) a study that pi-ovides some answers. l'rivat-
ization encouLrages restructuring if it is rapid andl comprelhensive and leads to concentraLted ownership. Irivat-
izLition also pi-omotes restruiCtLiring hecause privatized firmiis are more likely tball state-owned enterprises to
exercise waige restraint-and wLage restraint is vital to free up the necessary internal finance. Btit policies that
increase banik lencling to firiis, suLch as debt forgiveness aLind recapitalization. mav clo more harm than good.
The sLifest course is to recapitillize banks only as part of privLitization ancl to encourLige negotiLitions for finan-
- ciail restruCturinig only aifter the bLinks are privaLtizedl.
The Drivers of the Information Revolution-Cost, Computing Power, and Convergence 27
James Bond explains the drivers of the information revolution-the decline in the cost of transmitting informa-
tion, the increase in the power of computing. and the shift from analog to digital information teclnologies that
has joined the telecommunications and computing industries and merged segments of the information industry. F
Over the past twenty years. the cost of voice transmission circuits and the computing power per dollar invested .
have both fallen by a factor of 10.000. Prices have not fallen nearly as fast-they have been set hy a cartel-like I
system of international agreements hetveen incumbent monopolies. But as convergence restructures the tele-
commUnications industry. new operators are arbitraging the difference between costs and the old tariff strmc-
tures, putting pressure on incumbent telecommunications operators. 4
Telecommunications Is Dead, Long Live Networking 31
James Bond looks at hov the drivers of the information revolution are transforming the structure of the
telecommunications industry. The end of natural monopoly, the breakdo-wn of the old pricing mechanisms, the
increasing competition from new operators and new products, and the globalization of the industry are forcing
radical change. This spells trouble for incumbent telecommunications operators-most of wvhose assets are
holes in the ground. Many incumbents are responding by forming global alliances. But this trend imay have
more to do wvith their desire to recreate in international markets the oligarchies they are used to at home than _ - .
with the underlying market forces. Where is the industry heading? One view of the future sees transmission
capacity and bandwidth becoming tradable comm)odities, with the industry fragmenting into wvholesalers in-
Telecommunications Reform-How to Succeed 35 _ ;
Telecommunications reform-privatization and opening markets to competition-can he a positiVe-SUMII gamCe
in wlhich all stakeholders gain: customers, existing and new^- operators, employees, domestic and foreign inves-
tors, and government. But the extent and timing of benefits vary from one case to another. Bjbrn Wellenius . I r .
sets out some rules for reform that w-ill enhance those benefits: Get support at the higlhest level of political U
authority. Sort out conflicting objectives early-especially the conflict between maximizing revenue and delis -
ering more, better. and cheaper services. tIse market mechanismns rather than individual negotiations to select .
partners and determine the right sale price. Establish and follow clearly defined processes for sale and regul.l-
tion that are open to participation and reviev by all interested parties. And respect and trust the general public ! 4 ;
and keep it informeld. Although miajor transactions suclh as a privatization or the issuance of new licenses drive "A
the reform agenda, clhange continues wevll beyond them. Following the rules and hlonoring commiiiiitmcnts help X . I
consolidate an environment for sustainahle development of telecommunications.
Liberalizing Telecommunications and the Role of the World Trade Organization 39
In Febriary 1997, sixty-nine governments of high-income and developing countries agreed to liberalize their
basic telecommunications services under an agreement negotiated through the Norld Trade Organization. Most
participants in the agreement have subscribed to procompetitive regulatory principles, including independent
regulators, competitive safeguards, measures to ensure interconnection, universal service obligations, and trans-
parent and nondiscriminatory practices in licensing. The markets affected by the arrangement represent more
than 90 percent of the w-orld market for telecommunications. Carlos A. Primo Braga reviews the evolution of
the agreement and argues that the cr-itical issues novy are ensuring the quality of implementation and setting up .
a procompetitive regulatory environment. ..
-.. "0 -'"
What the Transformation of Telecom Markets Means for Regulation 47
Peter Smith looks at the impact of the changes in the telecommunications industry on regulation. As demand
changes, services converge, and new players emerge, the key issue for regulators is promoting competition.
The work of regulators is becolmiing increasingly complex at the same time that convergence and the common
principles estahlished in regional and international trade agreements are reducing their discretion. One way for
regulators to deal with the complexity is to privatize aspects of regulation-for example. hy creating property
rights to the spectrum and by outsourcing some regulatory tasks. But the specialized telecommunications
regulatory agency is probably a transitory entity that may eventually find itself merged into a multisectoral
antitrust agency.
The Private Sector and the Internet 51
Carlos A. Primo Braga and Carsten Fink look at the rise of the Internet as the main application behind the
emerging global information infrastructure. Many now believe that the Internet provides a window into a futLre
in whiclh access to information will be inclependent of geographic location and interactivity in a multimedia
environment will be ubiquitous. The authors review the need for a regulatory framework for the Internet in
three critical areas: provision of backbone access, Internet service providers, and information services. They
also explore the problem of the appropriability of content, discussing intellectual property rights in the digital
era and other remedies to the cost recovery problem. For developing countries, however, the critical bottleneck
is still their weak information infrastructure.
4.1
m C~~~
S; >~~~~~~~~
5
THREE WAYS TO ORDER
0,
For a free sabscription to Private sector
Fill oGlt and miail this card to SLtzanne Smith, Ec itor, or I x it to 202 522 29(i
FIRSI NANIE Nil iASS \NAMF Topics of
interest
JOB IITLE :IO'SPANS 3 Banking and
capitail markeE'
STREET ADDRESS 3 Priv,ttization
El CotQprct,tiun
CIII >TATlE POSTAL C()lk and regulation
El Thiccolrns and
COt \ITBSy ENLAIL techt lnaogy
=Encrgy
WVORK PHO\E FN
3 Transport
2 Call 202 458 1111 to record this information
LJNYater
3 Email this information to ssmith7@worldhank.org H1l
Netscape: 99Estach.pdf
% I
tGQt$i - �_ ()II Rt o a d ConcE ss n s
:vW*? VM.' ' ts *.1 Kaac*4 *1 ''ct ?*trnrg $$. t:0 are t; ';'
Lcssons fl rii Argentina
._ .,ut. 1. . . . 0 tSt'.1.,.1 ; = ;:
01 41>11. 0>0 1 4
WI I," A I
_ _ @ w ' J tz X v } v s T .1 .. ' la S F~~~~~~~~~~~'V"6 t'h
I I~~~~1 i., x IIII''II
Most Notes are now available on-line in full-text HTML format or in downloadable Adobe PDF
format (http://www.worldbank.org/html/fpd/notes/notelist.html).
g' 1 7 ~~~~~Editor:
A Suizanne Smith
Roomn F6P-188
The World flank
1818 H Street, NW
Washington, D.C. 20433
Telephone: 202 458 7281
Facsimile: 202 522 2961
Email: ssmith7@worldbank.org
Cover illustration by Ruth Sofair Ketler. Photos
on pages 1, 2. 3, 4, and 6 provided by FPG
International.
The entire contents of Private Sector �)1997
World Bank. You are authorized to reproduce,
duplicate, and disseminate all or part of this
publication so long as yotu include the name
of the publication and the name of the respec-
tive author. You may not, however, modify,
p e g I as alter, or otherwise change any part of this pub-
lication or sell, transfer, or otherwise dissemi-
nate any part of the publication for profit.
w � Printed on recycled paper.
7
The Restructuring and Privatization of the U.K.
Electricity Supply-Was It Worth It?
David MI. Xiubeilu andffi Mcbael G Pollitt
The electricity supply industry in England and Significant changes followed. In the first six years
Wales was under public ownership from 1948 after restructuring, labor productivity in the stuc-
to 1990. For most of this period, a single com- cessor companies more than doubled. There w as
pany, the Central Electricity Generating Board a marked shift away from coal and toward natu-
(CEGB), operated all generation and transmis- ral gas. At privatization, generation hased on
sion as a -ertically integrated statutory mo- fossil fuel used 92 percent coal, 7 percent oil.
nopoly, while twelve area boards acted as and only 1 percent gas. In the next fiv-e years.
regional distribution monopolies. The CEGB purchases of British coal fell from 74 million
during this time was a classic example of a cost- metric tons to 30 million, and by August 1996,
of-service regulated public utility-wvith exces- gas accounted for 23 percent of generation. In
siv-e capital costs, overdependence on high-cost the meantime, the price of coal delivered to
indigenous coal and nuclear power, low pro- power stations fell by 20 percent in real terms.
ductivity growth, and low return on assets. The switch from coal and the dash for gas
contributed to a substantial drop in emissions
In 1990, the CEGB vas restructuLred and priva- of sulfur dioxide and nitrogen oxides, hoth
tized. What xvere the costs and henefits? The sources of acid rain, and of carbon dioxide, the
question is important not only because the cause of global warming.
CEGB accounted for such a significant share
of economic activity-vith value added equal The pover generation sector added 9.5 mega-
to about 1 percent of GDP-but also because watts of capacity in combined cycle gas turbines
its restructuring was a key part of the British (nearly 20 percent of peak demand) in 1990-96,
electricity experiment,' wlhich has provided a while demand rose less than 6.5 percent: of the
model for power sector reform around the new capacity, half was installed by new entrants.
world. This Note reports the results of a social Fossil fuel costs per kilow att-hour (kWh) of elec-
cost-benefit analysis of the restructuring and tricity generated fell by -r percent in real terms
privatization of the CEGB. as a result of fuel switching and efficiency in-
creases, wvhile nuclear fuel costs per kWhI fell
The reform by 60 percent. Overall. real unit costs fell by
about 50 percent, while real pool prices fell by
The restructuring of the CEGB involved divid- a more modest 20 percent, \vith the difference
ing it into four successor companies on Mlarch betwN-een the two figures reflecting the lack of
31, 1990-three of which were soon sold to competition among generating companies.
the general public-creating a power pool, and
liberalizing entry into the generation market. Some of these positive changes could be at-
The three privatized companies are National tributed to external factors. The timely devel-
Power and PoxverGen, w hich took the thermal opment of high-efficiency combined cycle gas
generating plant, and National Grid, which was turbines, the lifting of the European Union (EU)
allocated the high-voltage transmission net- ban on burning gas to generate electricity, and
work. Nuclear Electric took the nuclear power tighter EU limits on sulfur emissions all encour-
stations. aged the sxvitch to gas, and the decline in
8 The Restructuring and Privatization of the U.K. Electricity Supply-Was It Worth It?
international coal and oil prices and in the the future, so the method involves projecting
domestic price of gas contributed to the re- into the future hoth the actual outcomes and
duction in unit costs. the counterfactual.
The restructuring and privatization were not The social cost-benefit analysis of the CEGB's
without costs. The 'dash for gas" greatly accel- restructturing and privatization proceeds in two
erated the decline of the coal industry. Em- stages: first the net benefits of the restructur-
ployment fell from nearly 250,000 miners at ing and privatization are calculated, then these
the time of the 1984-85 coal miners strike to benefits are apportioned among shareholders,
only 7,000 by 1994. The collapse of the British the government, and the power purchasers in
coal market was the subject of a Parliamentary the pool or wholesale market (the distribution
inquiry. Partly in response to that inquiry, policy companies and the supply businesses of the
toward the still publicly owned nuclear gen- generating companies) to see how the gains
eration industry was reviewed, plans to build are distributed. The first stage of the analysis
more nuclear power stations were abandoned values four areas of net benefits and costs sepa-
in early 1996, and the nuclear industry was re- rately: the efficiency savings, the investment
structured. The more modern nuclear power and fuel use effects, the costs of reorganiza-
stations were sold as British Energy in June tion, and the environmental benefits. In each
1996, leaving only the ruLmp of aging first- area, it establishes a set of counterfactuals with
generation Magnox stations and the fuel re- which data or projections for the actual indus-
processing facilities in the public sector. The try are compared. The start date from which
coal industry was privatized at the end of 1994. the effects of restructuLring and privatization are
evaluated is a weighted average of the years
A social cost-benefit analysis 1985-88. Actual data are available until March
1996, and projections are made to 2010.
Studies have used several methods to assess the
economic effects of privatization on formerly To allow some sensitivity analysis for the more
state-owned companies. including financial debatable issues, two counterfactuals are used.
performance analysis, labor and total factor One is labeled proprivatization because the un-
productivity analysis, frontier efficiency measure- derlying assumptions about the industry under
ment. and social cost-benefit analysis. Although continued public ownership are more pessi-
all these methods are of interest, only a full so- mistic than under the other counterfacttual and
cial cost-benefit analysis identifies who gained, so it suggests greater net benefits from privati-
who lost, and by how much-by comparing the zation. The other counterfacttual is labeled pro-
historical and predicted future course of an in- CEGB because its more optimistic assumptions
dustry after privatization with a counterfactual about the industry under continued public
in which the industry remains unprivatized. ownership point to smaller net benefits from
Jones, Tandon, and Vogelsang (1990) set out privatization. The counterfactLals incorporate
this method, and Galal and others (1994) apply three key items: Productivity growth is lower
it to twelve privatizations, two of which involved in both counterfactuals compared with the ac-
Chilean electricity companies. tual, but slightly higher under pro-CEGB than
under proprivatization. Gas and coal prices are
In simple terms, the analysis reduces to a project the same as actual under pro-CEGB but higher
appraisal, in which restructuring and privatiza- under proprivatization. And under both
tion are an investment project that has associ- counterfactuals, the CEGB invests in uneco-
ated costs (redundancy payments, brokers fees) nomic nuclear capacity and retrofitting of some
and creates a stream of net benefits arising from coal plant with flue-gas desulfurization units,
the evaluated differences between the privatized but under proprivatization it builds more
industry and a counterfactual publicly owned nuclear and coal plant and does morc retrofit-
industry. The costs and benefits continue into ting. The counterfactuals are based on reports
The World Bank Group 9
TABLE 1 THE NET BENEFITS OF PRIVATIZING THE CEGB
RELATIVE TO TWO COUNTERFACTUALS, 1995-2010
(f billions at 1994-95 prices; at a 6 percent discount rate,
of the CEGB before privatization and an analy- discounted to April 1995)
sis of the CEGB's performance in the decade
before restructuring. The proprivatization Relative to Relative to
counterfactual is probably closer to what wooulcl proprivatization pro-CEGB
have happened in the absence of privatization. counterfactual counterfactual
The results Fuel and investment effects
End of nuclear expansion program 3.3 2.8
What does the analysis show abotut the net ben- Effect on price of French imports -2.6 -1.5
efits? The fuel and investment effects of the
privatization range from gains of �3.6 billion Netfossit fuel costs 2.9 -2.1
to losses of f0.7 billion (at the U.K. putblic sec- Total 3.6 4.7
tor's preferred 6 percent discount rate),' de-
pending on assurmptions about hov a utility Externality benefits
under continued public owvnership xxould have Reductions in sulfur dioxide emissions
invested in new capacity (table 1). The net gains (�125 per metric ton) 1.0 0.7
from privatization are higher relative to the Reductions in carbon dioxide emissions
proprivatization counterfactual; the oains conme
1~~~~~~~~~ (f�12 per metric ton of carbon) 1.4 1.2
from the ending of the expensive nuclear ex-
pansion program that mighlt have seen tw-o new Total 2.3 1.9
nuclear power stations built and the sharp
switch from expensive British coal to cheaper Restructuring
natural gas for electricity generation. Costs -2.8 -2.8
Cost savings 8.8 7.6
Regardless of the counterfactual Used, some of Total 6.0 4.8
the benefits of privatization are dissipated in
higher payments to Electricite de France (EdF), Total net benefits 11.9 6.0
the French utility, for its cheap electricity im-
ports. This happens because before privatiza- Total net benefits (pence per kWh) 0.21 8.09
tion EdF had received a price equal to the
average of the marginal costs of the two sys- Sorrce-Authiors'estimates.
tems (a price lower than the system nmarginal
cost in England and Wales), while since privati-
zation it has received the pool price (which is
at or above system marginal cost) plus a share nients in nitrogen oxide emissions and ben-
of the fossil fuel levy paid to non-fossil fuel efits from reduced coal burning.
generators (introduced at the time of privatiza-
tion to finance decommissioning at Nluclear Elec- The restructuring and privatization haxve high
tric). If privatization had not occurred in the clirect costs, �2.8 billion. This figtire includes all
United Kingdom, the paytnent terms would not the restructuring costs of the successor compa-
have changed to the benefit of EdF. nies. including substantial redundancy and early
retirement payments. But the restructuring and
Against both counterfactuals, privatization privatization deliver unambiguous benefits in
yields substantial environmental benefits as lower operating costs (08.8 billion relative to
cleaner gas generation replaces older coal-fired proprivatization. �07.6 billion relative to pro-
plant and thermal efficiencies rise at thie re- CEGB). The difference reflects the lower labor
maining fossil fuel plant, leading to sharply and materials andcservices costs that the restruc-
reduced emissions. The figures in table 1 are turing and privatization deliver-gains difficult
conservative estimates of the environmental to imagine under a counterfactual publicly
benefits, whichi include unmeasured inmprove- owned CEGB.
10 The Restructuring and Privatization of the U.K. Electricity Supply-Was It Worth It?
TABLE 2 DISTRIBUTION OF THE NET BENEFITS OF
PRIVATIZING THE CEGB, WITH PRICES CONVERGING
IN 2000 try and the counterfactual publicly owned firm
(� billions at 1994-95 prices; discounted to April 1995; converge in 2000. The study assumes that regu-
excludes externalities) lation would ensure long-run convergence of
precdicted and counterfactual prices. The results
Relative to Relative to of the calculation show the perverse nature of
Beneficiary proprivatization pro-CEGB the distributional effects of the privatization.
The government's substantial sales revenue
Powerpurchasers -1.3 4.4 (X9.7 billion) up to March 1996 is at least par-
Govemment (including sales proceeds) 1.2 0.4 tially offset by loss of flow revenue, because
Shareholders (less sales proceeds) 9.7 8.1 tax revenue from the successor companies falls
below the public sector dividend target. As a
Source: Authors' estimates. result, the government is �1.2 billion better off
relative to proprivatization if prices converge
in 2000. Relative to both counterfactuals, the
The overall net benefit of the privatization is shareholders benefit by more than the total net
substantially positive relative to both counter- benefit, even after the sales proceeds paid to
factuals: �11.9 billion and �6.0 billion. These acquire the assets are subtracted.
figures may be converted to permianent savings
in the unit cost of electricity of 0.21 and 0.09 Conclusion
pence per kWh at a time that electricity prices
were about 2.8 pence per kWh. Thus, privatiza- Was it worth it? Yes. but the analysis suggests
tion delivers a permanent cost reduction equiva- two major areas for improvement in the pro-
lent to about 3.2 to 7.5 percent of prices, or an cess of the restructuring and privatization. First.
extra 40 percent return on assets. about a quarter of the net gains were transferred
out of the country because of the change in
How has this net benefit been distributed among payment arrangements for French electricity. If
shareholders, purchasers in the wholesale mar- more attention had been paid to this possibility
ket, and the government? Examination of price at the time of restructuring, sonme arrangement
trends shows that wholesale prices have not could probably have been found to prevent it.
fallen nearly as fast as costs and that profits have Second. introducing more competition in gen-
risen sharply in the successor companies: com- eration (by creating more successor companies)
bined profits (before taxes and exceptionals) would have reduced excess entry and lowered
rose from �.2.0 billion in 1991-92 to �3.5 billion prices, improving the distribution of the net
in 1995-96. The share prices of National Power benefits and increasing social welfare.
and PowerGen have approximately tripled since
flotation, outperforming the stock market by Thi Note is ahort version of an aticle hv the authors dui to appear
more than 100 percent. Thus, the companies in Jolto'al oJ tInditu'stnl Ecotnomics (Siptletifner 199T).
seem to have unambiguously gained from the RcstrUctutring and pninatiz.ttion .are ., pUbliC sector project and so
privatization. Power purchasers seem to be pay- hutd le es alutted at an approprte pUbtiC sctot discount rate.
ing higher prices than they would have uncler References
continued public ownership (higher company
profit margins offset lower costs). And the gov- Galal Alimed, Lcov Jones, PFank.j Ta.ndon. and ingt Vogelsang. 1994.
ernment has gained from sales revenue, higher Co//nre (onu'qtuettce.s oSetltling IPliblic ELtetprnses: An Littpitlcal
taxes on profits, and dividend income, though Ana4tsis. Ness Ytrk: Oxford U nieisitv Ilr-ess.
Jones. Lerty, Pank.j Tatndon, and Ingo Vogelsang. 1990. Setling Putblic
it has lost the revenue associated with the pub- Entte)ptpsnc: A Cost-Benefit.Atethbdologt. Cambridge. Mass.: MIT P'ress.
lic sector dividend target for the CEGB.
David M. ,VNeubery anzd Micbael G. Pollitt,
Table 2 shows one possible calculation of the Department of Applied Economics and Faculty
distribution of the net benefits of privatizing of Economics, University of Cambridge,
the CEGB. In it prices in the privatized indus- Cambridge, UK.
11
A Retrospective on the Mexican Toll Road Program
(1989-94)
Jejf Raister
Mexico's private toll road program more than In retrospect, some industry observers lhave char-
doubled the national toll road network-from acterized the toll road program as a ruslhed and
4,500 kilometers in 1989 to 9.900 kilometers in poorly designed effort to develop the infrastruc-
1994. Fifty-three concessions were awrardcd for tture the country needed to compete effectively
the approximately 5.500 kilometers of roads, in an era of free trade. Others have simply la-
and by the first quarter of 1995 forty-four wTere beled it a mechanism to lift the construction in-
in full or partial operation, represenrtingp 5,120 dLustry Otut of the economic depression of the
kilometers. The investment of approximately 1980s. Whatever the diagnosis for the poor per-
US$13 billion in the program over the period fortnance of the sector. from a private investment
1989-94 was sourced from local commnercial perspective the impact wvas to shut off capital
bank debt, concessionaire equity, and federal flowvs to the sector anci to add to the Mexican
and state government grants ancI equity con- banking system's nonperforming loan portfolio.
tributions (figure 1).'
This Note presents a diagnostic of key policy
However, gross miscalculation of investment regulatory, and institutional gaps that under-
costs and operating inconme led to an unsus- mined the- financial equilibrium of the sector.
tainable set of operating conditions for these A checklist of recurrent problems illustrates
limited recourse financings. The financial eqcui- hoxy the failure to address these issues mani-
librium of the sector xvas fturtlher undermined fested itself in the course of implementation.
by the Mexican currency crisis of D)eceinher
1994. The combination of macroeconomic and
project-level factors brouglht new project de-
velopment to a virtual standstill, despite gov- FIGURE 1 SOURCES OF FUNDING FOR THE TOLL ROAD
ernment estimates that another 6(500 kilometers CONCESSIONS
of roads are needed by 2000. Restructuring of
both project debt and equity investments has Concessionaire
been the main focus of recent efforts.' equity Domestic
commercial banks
The financial and economic repercussions have
been widespread. Local cormmercial banks were
saddled with nonperforming toll road loans cs-
timated at US$4.5 billion to tS$5.5 hillion. Con-
cessionaires and their affiliates were facedith (US$2.5wt
writing off significant portions of their invest- billion)
ments. Moreover, the government has been un- Federal and state
able to unclog the road construction progrtm government grants
and has been under severe pressure to inject and equity
scarce financial resources to rescue investors. Source:Author'scompilation.
Users, in the meantime, were left with socme of
the most expensive road tolls in the xvorld.
12 A Retrospective on the Mexican Toll Road Program (1989-94)
BOX 1 THE TOLL ROAD CONCESSIONS
Legal framework. Under the program, the Secretary of Communica-
tions and Transport granted concessions to special-purpose entities, Major issues and sector performance
which in almost all cases were either directly owned by or were
affiliates of one or more local construction companies. The conces- Although the program attracted significant pri-
sion agreements were issued under the federal law of General vate investment, well-puLblicized problems
Means of Communication, which governs, among other things, roads negatively affected sector performance. These
revolved around the following issues:
that connect two or more Mexican states and bridges along any such * Inadequate tendering process and concession
road. Under this legal framework, concessions could not exceed design. The prequalification standard was not
fifteen years, though this was later extended to thirty years, and a rigorous enough (for example, bidders were
free, parallel alternative to each highway was required.3 not required to submit a detailed financing
plan). Also, the project award criteria lim-
Concession party responsibilities. The concessionaire was respon- ited the pool of potential candidates (and
sibleforfinancing, building, operating, and maintainingthetoll road thus haotential competition for themariket) to
a handful of local construiction compainies
subject to government regulation for a specified period of time in that were more interested in the construc-
exchangefortherighttoreceivetollrevenuesgeneratedbythe tion work than in the long-term financial vi-
project. The role of the Secretary of Communications and Transport ability of the projects. (See box 1 for an
centered on project definition, including highway path, location of outline of the concessions.)
interchanges and toll booths, and number of lanes, establishment of * Inadequate financial discipline in government-
design and construction standards; design and implementation of owned commercial banks. This led to large
tendering procedures; and supervision of the concessionaire. or nod iece unaking was not
or no due diligence undertaking. It wvas not
uncommon for lenders to waive important
Concession design. The concession specified the duration of the conditions precedent to initial and subsequent
agreement, the work to be undertaken, operational and maintenance funding (insurance and bonding requirements,
standards, govemment supervision, required maintenance reserve securing permits and rights of way, satisfac-
funds, concessionaire reporting requirements, certain fees payable to tory review of traffic studies, geotechnical and
the government, and the tolls to be charged. On termination of the environmental studies, and the like). As the
concession, the right to operate the highway and to collect toll story goes, such behavior was guided by an
implicit understanding that even if the projects
revenues reverted to the govemment The government was to remain proved commercially nonviable, ultimate re-
the owner of the project throughout the term of the concession. course was indeed to the government.
Underdeveloped local financial markets. Le-
Tariff and adjustment mechanism. Each concession set forth a gal and regulatory limitations combined with
schedule of tolls by category of vehicle. Tolls were allowed to poor macroeconomic fundamentals inhibited
increase semiannually in accordance with the consumer price index the capacity of local markets to provide long-
CPI) or whenever the CPI increased by 5 percent or more since the term fixed rate financing. Peso-denominated
debt featured very short maturities rarely ex-
previous adjustment. All other toll adjustments beyond the levels set tending beyond five years, with interest rates
forth in the concession required the govementfs written approval. often 1,000 to 1,500 basis points above those
paid by the government. This situation was
Guarantees. Each concession also contained guarantees of traffic exacerbated by the currency crisis of Decem-
volumes by category of vehicle. Most concessions provided that if the ber 1994, when all-in interest rates rose to
actual traffic volumes on the highway fell short of those specified in more than 100 percent a year for most
the concession, the concessionaire would be entitled to request an projects, which were already strapped to meet
their debt service obligations.
extension of the term of the concession to permit recovery of its Underdeveloped institutional capability. The
investment above three issues were aggravated by the
fact that the program's scope simply ex-
The World Bank Group 13
ceeded the technical and administrative ca- given concurrent plans to privatize the rail,
pacity of the local constrLction industry, the port, and airport sectors.
liquidity of domestic financial markets, the * Understaffing and limited institutional capa-
project finance experience of most financial bilities within the Secretariat of Communica-
intermediaries involved, and the institutional tions and Transport often led to problems in
capabilities of regulatory officials. Conse- obtaining permits or approvals for change
quently, the control mechanisms needed to orders on a timely basis and to inadequate
develop the roads within suchi a shlort time enforcement of the concession requirements
were never adequately acldressed. regardin(g construction and maintenance qual-
it- control standards.
A summary follows of how these four prob- u In addition to the problems relating to hid
lems manifested themselves in project imple- selection criteria, there was no efficient pre-
mentation-in the regulatory and institutionial selection process to screen out potential bid-
framework for the concessions, the operative ders that lackeed the capacity to assume the
period, and the financial and legal arrangements essential risks of construction design and man-
of the projects. agement, completion of large projects, and
commercial management of toll road opera-
The regulatory and institutional framework tions. While operating a toll road is fairly
simple (nostly consisting of collecting tolls
Problems relating to the regulatoryv and institu- from passing vehicles), managing a toll road
tional framework for the concessions included program is much more complex. It includes
vsaguie project selection criteria stemming in large estimating deniand in the face of competition
part from the lacki of an intermodal strategy and from toll-free roads or other forms of trans-
inadequate planning criteria at the fecleral and port, adjusting tolls to optimize revenues, plan-
state level, inadequate prequalification and ning maintenance to minimize long-term costs,
award criteria, uncertain tariff adjustment pro a'ndi managing short- and long-term financial
cedures. and lack of an independent regulatory commitments. Ideally, the hidding consortia
authority to supervise the contractual arrange- slhould be able to demonstrate that these sklills
ments. The major recurring issues includced the are available to them. The lack of a good
following: screening process led, for e\iample, to the
* This greenfield program sought to establish selection of mecdium-size concessionaires that
five main road corridors, three of which >were financed their equity contributions throtugh
to run between the main industrial centers in commercial loans. NXWhen projects began to suf-
Mexico and the principal border crossings into fer financial difficulties, these concessionaires
the United States.' Nonetheless, some high- wAere often unable to meet their equity infu-
priority segments were never concessioned, sion requirements. Others did not have the
while others that were constructecd lacked con- necessary technical c apabilities, inclucding spe-
tiguous sections that would integrate them into cialized machinery, skilled labor, and adequate
the network. This piecemeal pattern of coln- quality control procedures. Even some of the
tracting reduced the near-term attractiveness larger companies xxvere stretched too thin,
of the toll roads to long-distance traffic, par- given the speed at which different conces-
ticularly to truck-ers, xvho pay the higlhest tolls. sions were awarcled to the same firm.
* Somewlhat related is the lack of an intermodal * Formal mechanisms were never established
development strategy. Thus, project devel- for soliciting or channeling inquiries or re-
opment in the various transport sectors often quests from private sector participants before,
occurred without due coordination. Conse- dcuring, or aftcr the bidding process. Thus, the
quently, investors were unable to determine relationship that developed between the pub-
whether a project fit well into the long-term lic and private sectors often lacked transpar-
development plans of a region. especially ency and was at times adversarial.
14 A Retrospective on the Mexican Toll Road Program (1989-94)
Project cost structure at all. This, combined w ith the fact that lend-
ers only rarely hired an independent engi-
Cost overruns and delays frequently arose be- neer to assist them with their due diligence
cause of information deficiencies, problems investigation before financial closing or with
with securing rights of way, lack of effective supervision of the contractors' efforts, cre-
turnkey construction arrangements, unantici- ated a void in terms of monitoring quality
pated design changes, local community resis- control programs, permitting issues, and the
tance, and permitting issues. As a result, the progress of construction budgets, critical path
average cost per kilometer of new highway rose activities, and the like.
to about US$2.6 million to US$2.8 million, com- * In some projects, construction came to a vir-
pared with the original estimated cost of USSI.7 tual standstill because of poorly defined pro-
million. This figure does not reveal the full cedures and bureaucratic delays regarding
extent of the overruns associated with the "hard the issuance of permits for purchase and use
costs" of construction, that is, the costs associ- of chemicals or dangerous substances. In one
ated with required equipment, material, and project, time delays resulting from problems
labor, and as opposed to 'soft costs" (interest in securing permits for dynamite directly re-
payments during construction, cost escalation sulted in cost increases of nearly 30 percent.
due to inflation, advisory services, and the like).
The dramatic drop during 1990-94 in both in- Project revenue structure
flation and interest rates offset in part the real
increases in hard costs. Cash flow generated by the sector has been
far below base-case expectations as a result of
Reasons for the cost overruns included the traffic shortfalls and higher-than-expected
following: operations and maintenance expenditures. The
* Projects often broke ground with only very December 1994 currency crisis led to a sharp
preliminary engineering and design work. In decline in disposable income and thus road
the case of the 267 kilometer Cuernavaca- usage. along with a drop in economic activity
Acapulco toll road, for example, this led to that resulted in a marked decrease in commer-
cost overruns of 200 percent and time de- cial activity ancd freight transportation. As a
lays of thirty months. result, of the thirty-two projects for which op-
� Construction often began without first secur- crating data \vere available in March 1995, less
ing the right of way. This failure was often than five could meet their base-case revenue
exacerbated by mounting resistance from lo- projections. On average, actual project revenues
cal farmers and community groups, environ- were 30 percent below original projections.
mentalists, and historical conservationists,
and resulted in delays and even rerouting of Important factors leading to this situation in-
some projects. As problems occurred, ma- cluded the following:
chinery and material sat idle while mobiliza- * Shortcomings in the traffic studies reflected a
tion and interest costs mounted. general lack of expertise by the concession-
* One of the most frequently recurring problems aires, the lenders, and their consultants in
related to supervision and unilaterally man- developing adequate methodologies (box 2).
dated change orders by the Secretary of Com- On only five of the thirty-two toll roads for
munications and Transport. In a project in which traffic data are available has the aver-
which four pedestrian bridges were expected, age daily traffic been above base-case expec-
the final number reached almost sixty as a re- tations (table 1).
sult of government-mandated change orders, * In some projects, trucks were expected to ac-
often required to appease local interest groups. count for 20 to 45 percent of users. In reality,
* Many projects were financed under very loose trucks were less than 5 percent of the traffic
cost-plus construction arrangements or none on many roads, leading to a weighted aver-
The World Bank Group 15
BOX 2 TRAFFIC STUDIES
Traffic study methodologies often suffered from the following:
age tariff much lower than originally expectedl. Lack of analysis of specific traffic characteristics, including time
In some cases, the existence of a black mar-
ket for toll tickets contributed to this outcome. and seasonal variations by type of vehicle, trip origins and
Despite obvious time and cost advantages destinations, and purpose and frequency of trips.
of the new roads, many potentiual users were . Failure in projections to identify key economic parameters that
simply unw illing to pay the toll. Aside froim would affect road usage, such as population, employment, per
the extremely high tariff's, this unw illingness capita auto ownership, per capita and disposable income, and
was also duie to the fact that the conicession- performance of key industrial indicators.
aires did little to market the time ancd cost
savings of the roads (for examiiple. through Unrealisticgrowth rate assumptionsforextended periodsthat, if
monthly passes, volurme discounts, and cli- realized, would have exceeded the capacity of the road.
rect negotiations with high-volume uscrs such l Failure to include an end-user learning curve or differences in
as trucking or passenger bus companics). tariff elasticity between end users.
In all but a few concession agreements, the Overreliance on increased demand due to the opening of intercon-
concessionaire could adjust the tariffs only
withepriornappreoval byjt the S etarioff Coinl necting roads, the construction of which was often delayed or
wTith prior approval hy the Secretary of Comi-
munications and Transport (even for down- neverundertaken.
ward adjustments). This greatly reduced the Underestimation of the congestion relief thatthe opening of the
flexibility of the concessioniaire in efforts to new toll road would bring for the toll-free option, and thus
maximize cash flow. overestimation of the actual time savings of the new road.
Minimal attention was paid to the develop- * Insufficient attention to general conditions of alternative and
mnent of suich auxilkiarV SerxTiCeS as gas sta- feeder routes and the identification of factors influencing the
tions, rest stops, hospitals, tow truck services,
and restaurants. (For most projects, conces- traffic-carryingcapacityofkeysections.
sionaires were uranted the right to operate - Inadequate and at times not readily accessible data from the
these services for two years beyond the con- Secretary of Communications and Transport for traffic studies.
cession term.) A toll bridge expected to - Though investors sometimes employed their own independent
handle ',)00 trucks a day mov ing through a consultants, actual fieldworkwas limited toonetotwoweeks of
U.S. border crossing captured only 200 users traffic surveys. This was often the result of insufficient time
a dav. This shortfall wAas in large part clue to
inadequate attention to access roads and to allotted to bidders and financiers between the date of release of
installation of customs clearing facilities. the bid documents and the deadline for delivery of bids.
The government faced great resistance from
the trucking industry in implementing and
enforcing technical measurenment and axle-
weiglht standards. Truckers for the most part X Inadequate toll collection operations and sys-
continued to use the toll-free option, espe- teimis, poorly designecd fiduciary structures,
cially in light of the very hig,h tolls. and the inexperience of the trustees and com-
Operations and maintenance budgets often mercial banks responsible for super-vising the
were not heavily scrutinized by the conces- flowv of project funds led to less than strict
sionaire or its lenders, and in many cases controls over collection and proper applica-
extraordinary maintenance costs xx ere grosslv tion of road revenues.
underestimated. Though provisions for ma-
jor maintenance reserve funcls were incluclecl Project financial structure
in most concession agreements, enforcement
of these provisions by the Secretary of Com- Lack of liquidity in the local financial markets,
munications andl Transport and credlitors xv as use of short-term, high-cost. floating rate dlebt.
often lacking, especially' as concessionaires currencv risk (both devaluation and convert-
began to experience financial difficulties. ibility) faced by international investors, and the
16 A Retrospective on the Mexican Toll Road Program (1989-94)
high cost and limited availability of surety and * The only source of local debt financing was
insurance coverage severely hampered sector the commercial banking sector. But the ten-
performance. ors for such debt often extended only through
Concessionaires' financial contributions were the construction period, with the expecta-
in the form of 'sweat equity" provided tion that once the project was in operation,
through the retention of work from construc- cash flows would be securitized through lo-
tion affiliates. These contributions originally cal or international debt offerings. However,
amounted to 25 to 30 percent of investment as roads incurred cost overruns and the debt
costs, but as lenders demanded higher eq- servicing ability of the projects proved far
uity cushions and debt service coverage ra- less than had been expected, these construc-
tios, the contributions increased to about 50 tion lenders soon were forced to restructure
percent of project costs. This led to inflated and extend the terms of their bridge financ-
construction budgets (and hence toll levels), ing. In addition, the loans were character-
with some projects effectively financed with ized by high floating interest rates, often 1,000
100 percent or more leverage. Estimates of basis points higher than the local market ref-
the average gross margins in the road builcl- erence rate. This combination of high inter-
ing program range from 35 percent to 50 per- est rates and short maturities resulted in
cent of total costs. Like the distortions arising prohibitively high tariffs.
from the bid selection criteria, these exces- * As many projects became increasingly un-
sive margins in no small part were the result able to meet their debt service obligations,
of a lack of competition among the limited lenders' appetite for new toll road invest-
number of project bidders. ments declined. Consequently, many banks
The concession agreements contained an that had underwritten huge amounts for
adjustment clause to shorten the concession projects were later unable to syndicate or
term if traffic exceeded guaranteed levels. refinance the loans, and liquidity quickly
Because of the lack of any upside potential, dried up in the market. Once word spread
this clause led to significant disincentives to about the actual financial situation of many
apply true risk capital. projects, other, untapped sources of funding
(such as international institutional investors)
quickly turned their attention to other invest-
ment opportunities, both within and outside
the country. Likewise, in the few international
TABLE 1 DAILY TRAFFIC HAS NOT offerings, market liquidity and resulting pric-
MET EXPECTATIONS ing were adversely affected by the presence
of currency risk, in the form of both exchange
Average dailytraffic rate depreciation and convertibility or trans-
as a percentage of ferability concerns.'
guaranteed traffic Number of roadsfeaitycnrs.
* Local commercial banks were lacking in
credit analysis, loan documentation, internal
Above 100 5 controls, and risk and liquidity management.
75-100 2 Thus, the skills needed for limited recourse
50-74 8 financing-to analyze project credit, security
25-49 8 arrangements, and operative agreements-
6-24 9 simply were not adequate for the complex-
ity of the projects and the huge demand for
Note: As of December 1994. credit.
Source:Author's compilation. * Performance, advance payment, and hidden
defects bonds, as well as insurance for prop-
erty damage, third-party liability, force ma-
The World Bank Group 17
jeure, and delayed opening, were high cost ing power at the negotiating table with both
and very scarce. Where coverage was se- the borrower and the government.
cured, significant problems arose in collec- x Some concessionaires wxere not single-purpose
tion. These problems resulted as much from entities. In these cases, it was impossible for
lenders' inexperience in negotiating the terms lenders to isolate specific cash flowxs by project,
of such policies as fi-om cuinhbersome and and borrowersNwith multiple concessions were
vague collection procedures. able to apply the cash flow from some projects
to .support the financial needs of related but
Legal considerations separately financed ventures.
Under manv trust agreeiiients, local banks
Legal aspects of the projects that weakened allowed the concessionaire the final word in
financial discipline included issues associated technical decisions on such matters as change
with lender security and enforcement rights, orders, change of material subcontractors,
dispute resolution mechanisms, tax treatment, and toll collection procedures. This led to
and procedures for securing government capi- major problems relating to construction and
tal contributions. Key problems included the operating costs as wxell as quality control.
following: -Certain tax aspects affected the financial vi-
L Legal disputes in Mexico arising between a ability of the projects. Changes to the tax code
private party and the governmnent were to be wN ere required regarding the 2 percent tax
resolved within the constraints of the Mexi- on assets, application and calculation of cle-
can court system and were not subject to in- preciation and tax credits, and payment of
ternational arbitration. Being subject to the value added taxes. But these modifications
local court system represented a significant Nvere made only after nearly twenty-five
risk to international investors because of their projects had been concessioned, and in many
lack of familiarity with the legal system. cases they required annual approval and thus
� State governments were expected to providle subjected financiers to nonrenexwal risk.
graimLs or cash equity or to dedicate toll rev-
enues from existing roads for certain projects Policy conclusions
as part of the construction financing, as con-
tingent obligations to cover cost overruns, or Policies to address such issues will vary de-
to cover costs related to securing the right of pending on sector objectives, the current sta-
way. But there were often delays or actual tus of the legal and regulatory framework, and
defaults in the fulfillment of these financial the technical and financial capability of the
obligations, in part because the contributions public and private sector participants. Of the
were to be sourced from annual budget ap- many lessons to be learned from the Mexican
propriations, a process subject to tremendous program, however, perhaps the most impor-
uncertainty and discretion. As a result, state tant for governments developing a sector pro-
governments were often left without any gram based on private investment is the
means for meeting their obligations. Other necessity of devising systems of regulation and
problems arose because of the lack of a clear support that provide the encouragemilent and
registration process for public debt, wxThich left room for maneuver that the private sector
lenders witlh no clear understanding of where needs, while minimizing the government's ex-
they stood relative to other state creditors. posure to the host of commercial and financial
* Lenders were not allowed a collateral assign- risk-s surrounding projects. The sector strategy
ment of the concession agreement. Conse- must include sound and explicit incentives to
quently, they could neither secure revenue select worthwlbile projects. Prices should be
generated by the project nor exercise bor- set to ensure the viability of privatized enter-
rower substitution rights in the event of a prises without protecting private parties from
default. This greatly diminished their bargain- bankruptcy. Prices should also be allox-ed to
18 A Retrospective on the Mexican Toll Road Program (1989-94)
reflect actual demand-in this respect, the need stith an expected payotit in five and a half years and a targeted
to develop congestion pricing is of fundamen- average life nf tiree and a third Nears). The issue receised a
local CuLTrencv-hased "A" rating front Standard & Poor's. The
tal importance. The regulatory framework deal stas initially placed at abotit 350 basis points over US.
should check the abuse of market power and Treasuries.
ensure adequate services. Besides protecting A fourth financing of ahoit US5300 itillion for the Tepic Guadala;ara
ensure adequate ~~~~~~~~~~~~toll roadl wsta canceledi at the last minuite hecause of the onset of
investors, an appropriate regulatory and market the December 1994- cuirrency crisis.
structure protects the government and eventu-
ally taxpayers from bearing ultimate responsi- Jeff Ruster (jruster@avorldbank.org), Private
bility for the financial performance of privatized Sector Development Department
enterprises.
Feder.al funding also inicluded contributions by Petroleos Mslexicanos
(Peinex) and by Canminos V I'PlenLe Federales de Ingreso N Serticio
Conexo (Cipute). the federal higiwsays and bridge operator. for
niore than 1.100 kilometers (kmr) of public toll roads.
Recent government estimates shosN that 26 billion net pesos (US53
billion) will be recluired to restructure existing concessions. The
Wcirld BaRik's Operations Evaltiation Department reports that by
earlv 1997 nearly forty projects. accounting for LiSS11.5 billion of
debt and equitN investmients, haie submitted requests to the gov
etinment for finiancial restrcictciring.
Thie gosernments of several Mexican states also granted conces-
sions incler local l.iAs to build and operate higItays; these saNe
generally been modeled on tbose granted by the Secretary of Comr
mtinications and Transport cinder federal last
The fi e major axis links aire Nog.lles -Liliacan Tepic-Guadalajara-
TICtlcis.i-Mlexico Citn ('21 krn) Nuet o Laredo-Revrtosa-Monterres-
San LLiis Potosi Qcueretarc-Mexico Citn t480 kin): Citidacl Jciarez-
Mcxico Cinv-lt'cebla-Oaxaca (340 kin): Mexico Cinv-Veracruz Savula
Ocozocoaultla Arriaga-PuLerto Madero (428 km): and Tixpan-
liachuca-.Mexico Cin- (222 kin).
Thiee projects alrecdy in operation swere nevertheless able to re-
finance fiy tapping the international capital markets prior to the
December 1994 Ciurrencv crisis. This was duLe in large part to ex-
pectations that Mexico xx ould receis e an ins estment-gra.de rating.
These three projects are as follows:
* Tie ten veai. US5207 million placenisent ftr the Toluca toll road
in Jcine 1992. The deal initially s-as not *tell received despite
repr icing cif about 700 basis points over U.S. Treasciries. Inves
tor concerns centered on tight debt sen ice coserage r.itios of
1.25 tci 1. combined stwith the existence of cLirrency risk, a par-
tiCsIlarlv sensitive issuLe at the time because the peso ssas esti
matecd to le abouit 20 percent over.lcied.
* The 1SS1 10 imsillion placenment for the Fecatepec-Piramicles and
Mtlanzanillo-Airneria toll roads, which v-ere jointlv secciritized.
The proposecd financing and secciripn strcicture stas percei-ed as
mcich simpler than th.it in the Tolcica placeniemii. andl it pr'O-
vided for a clual amortization process. The target amortization is
based on a twelve year final matLiritv. Thie alternative schedule
reqLlires payNment (with a 1 percent added premicim for inives-
tors) in eighteen years if project revencies are impaired, whether
becatisc of insufficient traffic flos , ctirrencv fluctuations. or
similar risks. The deal s-as priced at abocit 500 basis points over
UIJS Treasuries at the time cif closing.
* The Mexico Citv Ciernataca toll road. ostned and ciperated bty
Capcife. The placenment of exchange rate-linked bonds in Ati-
gtist 1994 stas originally planned as a USS625 millionn stents-
tear final isiaturinv tiansaction. Bcit because of investor con
cerns afiotis ccirrency risk ancI long-terisi interest rate volatility.
the isscie swas cut back to a seven sear. I2ss265 million 144A
placenment (ttioLigli also featurinig a dLial atmortization scliedcile
19
.
The Private Sector in Water and Sanitation-
How to Get Started
PenelopeJ BTook Couien
The henefits from private participation in wa- itecd areas as building capacity and providing
ter and sanitation depend on the level of rislk plumlibing services, it is difficult to achieve in
and responsibility the government handcs over distribution and collection, core activities in
to the private sector. But strong government water and sanitation (table 1). So governments
commitment and careful preparation are re- wNanting to involve the private sector have
quired if the private sector is to take on signifi- been atble to rely little on competition to assure
cant risks and responsibilities. A g,ov-ernment g,ood outcomes for consumers and have in-
about to cnter into a concession contract for steadh ad to devise regulatory systems for this
twenty-five or thirty years, for example, needs purpose.
to be sure that it does not, in haste, gloss over
details that will later land it in messy renego-
tiations. wvith loss of face to all concerned and
reduced benefits to consumers. But govern- BOX 1 TOOLKITS FOR PRIVATE SECTOR PARTICIPATION
ments often worry that detailed preparations IN WATER AND SANITATION
take too much time. This Note, based on a set
of'KWorld Bank toolkits compiled from experi- The World Bank has recently published a set of toolkits to guide gov-
ences in involving the private sector in water ernments in designing and implementing private sector arrangements
and s.,nitation, reviewrs the essential factors In
anchosingtatn privatizaio optioessentian fargus that for water and sanitation. The toolkits focus on three sets of issues:
prepsing rat paysrdivatizatidends.tionandargiestt How to choose a private sector participation option. What are the
preparation pavs divildends.
options? What might you have to do to make your preferred option
What makes the business special? practicable? What are the risks?
* How to design the process for refining and implementing the
The activities of urban whater and sanitation chosen option. What might a critical path look like? How do you
utilities range from impounding and treating set up a government unit to run the privatization process? What
raw water, to distributing water and collecting can you expect from legal, financial, economic, and engineering
sewage, to treating sewage. In many w ays. advisers-and how do you go about hiring them? How do you
decisions about how to involve the private sec- design a bidding process? What can you do to keep the contract on
tor in these operations resemble decisions track once you've chosen a private partner?
about privatization in any other utility sector. * How to ensure that contracts cover all the issues. What should a
All such privatizations. for example. require management, build-operate-transfer (BOT), or concession contract
decisions on how to set up an independent cover? In writing or reviewing contractual documents for a con-
regulator anid how to set and enforce service
standards. Buit wuater ancl sanitation hlve special cession, a BOT arrangement, or a management contract, how do
features that governments must take into ac- you know if you've covered everything? What are your options for
count in choosing and clesigning a contract and allocating and managing the many risks that go with the contract?
in designing a supporting policy framework:
Vater and sanitation systems are clharacter- Infermation on how to obtain copies of the toolkits can be found at the World Bank's
ized by a high degree of natural monopoly. Website lhttp://www.worldbank.org).
Although competition is feasible in such lim-
20 The Private Sector in Water and Sanitation-How to Get Started
TABLE 1 COMPETITION AND MONOPOLY IN WATER
AND SANITATION
Activity Characteristics of competition ficient use. In India, for example, many cities
go short of water while farmers continue to
Allocation of water resources Natural monopoly in each receive subsidized water for irrigation.
and regulation of use hydrogeographical unit * Water and sanitation are well suited to local
(such as a river basin) management, and in many countries, respon-
Capacity construction Competitive (but may depend on sibility for service provision is decentralized
accessto water resources) to the provincial or municipal level. As a re-
sult, complex interjurisdictional issues often
Bulk supply generation Small number of possible need tombe rsledibeforth ivate sec
suppliers (often only one) nced to be resolved before the private sec-
tor can be brought in.
Water trea*ment* Local monopoly Many of the assets of water and sanitation
Localdisthibution* Local monopoly systems are buried, so obtaining accurate in-
Local sewerage network* Local monopoly formation about them is costly-increasing
the cost of preparing for private sector par-
SEwagpent and appliaence s LocalComonopolye ticipation, and the chance of surprises after
Equipment and appliance sales. Competitive the contract is signed.
plumbing services Broad access to water and sanitation yields
- Core activities of traditional water and sanitation utilities. important public health and environmental
benefits. Government interventions to pro-
mote these benefits are likely to remain after
privatization.
Water is essential to life, and access to it must
be ensured for all. Guaranteeing access for the None of these issues is a barrier to private sector
poor will sometimes require designing subsi- participation-all arise under both public and
dies or schemes for reducing the cost of deliv- private provision. But governments often sys-
ering services to the very poor. A complicating tematically confront their implications only
factor for reformers is that existing systems for when they begin to contemplate private sector
allocating scarce raw water resources among involvement. Failure to adequiately address
alternative uses-urban consumption, irriga- these issues increases the risk that a government
tion, industry-are often incompatible with ef- will be unable to find a partner for its pre-
TABLE 2 THE MAIN OPTIONS FOR PRIVATE SECTOR PARTICIPATION AND THEIR ALLOCATION OF RESPONSIBILITIES
Asset Operations and Capital Commercial
Option ownership maintenance investmnent risk Duration Examples
Service Public Public and private Public Public 1-2 years Chile (Santiago)
contract India (Madras)
Management Public Public Public 3-5 years Gaza
contract Trinidad and Tobago
Lease Public Public Shared 1-15 years Guinea (17 cities)
Poland (Gdansk)
Build-operate- P$$20-30 years Malaysia (Johor)
transfer Australia (Sydney)
Concession Pbi <>25-3 years Argentina (Buenos Aires)
~~4~~~< - - ~~~~Coite d Ivoire
Philippines (Manila)
Divestiture ~ sIndefinite England and Wales
The World Bank Group 21
ferred form of private sector participation or proving efficiency, constructing large-scale
that a private sector arr-angemiient will fall short projects, cutting the cost of public subsidies or
of its broad policy objectives. redirecting them to the poor. and making the
sector more responsive to customers. All forms
The main private sector options-and of private sector participation can be designed
who's doing what to improve technical and managerial capacity.
But whether the other objectives can be met
Different countries have adopted different op- depends on whiclh option is chosen and whether
tions for private sector participation. Trinidad the government can do a good job on the en-
and Tobago is using a management contract abling and regulatory environment. A poor job
for water and sewerage services and plans to can lead to dissatisfied customers and difficult
replace it with a concession. Guinea has a lease renegotiations with the private partner. Under-
arrangement for water treatment and supply in the Guinea lease, for example, consumers have
seventeen cities. Buenos Aires and several other lost out because disputes over the division of
Argentine provinces have concessions for va- responsibilities between the government and the
ter and sewerage. And England and Wales hax-e operator have hampered new connections and
divested their water and sanitation utilities. service improvements. The Buenos Aires con-
cession has led to better service, but there have
The main options can be clearly distinguished been costly disputes over the definition of the
by how they allocate responsibility for such func- regulator s role (for example, in determining in-
tions as asset ownership and capital investment vestment requirements) and the handling of ad-
between the public and private sectors (table 2). justments in tariff levels and structure.
The more risk and responsibility are passed to
the private sector, the more powerful are its in- To determine wvhich private sector options are
centives to improve services. Service contracts, feasible-or what must be done to make a pre-
which confer little risk and responsibility on the ferred option possible-a government needs
private sector, offer commensurately small gains to undertake a range of analyses:
-and are simply not designed to address mana- * An analysis of the state of the utility-looking
gerial inefficiency or chronic underinvestment. at the current level and standard of service,
Concessions and divestitures are well suited to the condition ancd serviceability of assets, the
tackling these problems-but demand more from human resources, and the financial perfor-
government in commitment and preparation. mance. Is information about the utility's assets
good enough to serve as a base for long-
In practice, private sector arrangements are of- term contracts? If not, can better informnation
ten hybrids of these models. For example. be produced rapidly? Where information
leases may pass some responsibility for small- about the quality of underground pipes, for
scale investment to the private sector, and example, is partial or inaccurate, revelations
management contracts may, like leases, have about the true state of the system that come
revenue-sharing provisions that pass on some after a concession contract has been signed
commercial risk. Options can also be used in may lead to costly renegotiations.
combination-for example. a build-operate- i An analysis of the existing regulatory frame-
transfer contract for bulk water supply might work-both general laws that might affect
be combined with a management or lease private participation in the sector and sector-
contract for operating the distribution system. specific laws and institutions focusing on pric-
ing and quality standards. Does the existing
Key factors in choosing an option regulatory framework provide sufficient sup-
port for the private sector so that it will take
Governments seeking to involve the private sec- on commercial risk? If not, can the necessary
tor in water and sanitation may have a range of changes be made fairly easily'? And if not, can
objectives-introducing greater technical and parts of the regulatory function be simplified
managerial expertise and new technology, im- or contracted out in the short term? Where
22 The Private Sector in Water and Sanitation-How to Get Started
TABLE 3 PREREQUISITES FOR SUCCESSFUL IMPLEMENTATION OF DIFFERENT PRIVATE SECTOR OPTIONS
Stakeholder sup- Potential
port and political Cost-recovering Good information Developed regulatory Good country benefits
Option commitment tariffs about the system framework credit rating of the option
Service Unimportant Not necessary Possible to proceed Minimal monitoring Not necessary Low
contract in the short term with only limited capacity needed
information
Management Low to moderate Preferred but not Sufficient infornation Moderate monitoring Not necessary
contract levels needed necessary in the required to set capacity needed
short term incentives
Lease Not necessary
Build-operate- fllitsn mlSnl
transferC! G
Concession
Divestiture It!]l W'" =
- 1 Slmfu'iatisi.ws_ ~~~~~~~~~~~~~High
Note: The shading signals the degree of importance: 0 not significant * moderate
El low * high
regulatory capacity is weak, for example, col- analysis can sometimes lead to redefinition of
lection of information on the utility's techni- a private sector project-for example, re-
cal and financial performance could be balancing planned investment expenditures
contracted out to a private auditing company. between new production capacity and the re-
* An analysis of which stakeholders (employees. habilitation of existing distribution systems.
consumers, environmentalists, government
agencies) support private participation and As table 3 shows, in a very simplified way, the
which oppose it. Can processes and policies results of these analyses can point the govern-
be put in place to meet stakeholder concerns? ment to an appropriate choice of private sec-
Can the risk of political interference be mini- tor option. If regulatory capacity is weak and
mized? Often, a key factor in the success of a political commitment is low, for example, a con-
private sector project is identifying the con- cession will be difficult to implement. Even with
cerns of employees early on and finding con- strong political commitment to a concession
structive ways of addressing them-rather than or divestiture, however, countries that lack a
allowing those concerns to derail the reform good business climate or a strong track record
process later. of successful private investment may not im-
* An analysis of the financial viability of alter- mediately be able to attract large-scale private
native options. Do current tariffs cover costs? financing for infrastructure projects. These
Can the private sector reasonably be expected couLntries may need to start out with a man-
to boost efficiency enough to meet the pro- agement contract and work up to options that
posed service objectives without increasing demancl more of the private sector.
tariffs? If not, will consumers be willing to pay
higher tariffs? And if not, can grant finance Penelopej Brook Couwen (pbrookcoweng
(or subsidies to needy households) support worldhank.org), Private SectorDevelopment
service improvements? This kind of financial Specialist, Private SectorDevelopment Department
23
Privatization and Restructuring in Central
and Eastern Europe
Robert E. Anderson, Siunzeon Djankov, Gecbcv cl Pobl aind Stdn ClaGessens
This Note reports on the first comprehensive analysis of the industrial restructuring that has
taken place since 1992 in Central and Eastern Europe. The study, covering more than 6,000
industrial firms in seven countries, looks at which government policies have been most effective
in speeding enterprise restructuring. The results show that privatization is the single most
important factor in restructuring. The method of privatization has been less important:
management buyouts and massive giveaways of firms through voucher privatization have led to
results similar to those of case-by-case sales to foreign or domestic investors. The study also
shows that privatizing industrial and commercial firms is the most effective way to improve the
solvency of the banking sector-more effective than bank recapitalization or debt forgiveness.
The approach policies from the effects of such other factors as
size, sector, and initial productivity levels.
The study compares the extent of restructuring
by firnis in seveni CenLral and Eastern European The measures of restructuring used in the study
countries: Bulgaria, the Czech Republic, Hun- include profitability, proportion of firms with a
gary, Poland, Romania, the Slovak Republic, and positive operating cash flow, average operating
Slovenia. The number of firms in the sample for cash flowv as a percentage of revenue, grow"th
each country ranges from 700 to 1,000, and the
firms account for 40 to 90 percent of employ-
ment in manufacturincg (table 1).' The govern-
ments in the seven countries have used different TABLE 1 FEATURES OF THE DATABASE
policies to encourage restructuring, and com-
paring enterprise performance among the coun- Employees as a
tries should shed light on which have been most percentage of total
effective. The study tests the restructuring data Employees manufacturing
for the effect of such policies as rapid privatiza- Country Firms 1992 employment
tion, concentrated outside ownership (for bet-
ter governance), wage growth restraint (to Bulgaria 828 314,042 48
improve cash flow and fund restructuring), fi- Czech Republic 706 829,312 64
nancial discipline (firms may be more likely to Hungary 1,044 428,645 41
restructure if neither the government nor banks Poland 1,066 1,338,645 45
finance their losses), and maintaining debt obli- Romania 1,092 2,121,102 91
gations (firms may have a greater incentive to SlovakRepublic 905 578,737 93
restructure if banks do not forgive or reduce
debts). The data are subjected to econometric
analysis to separate the impact of government
24 Privatization and Restructuring in Central and Eastern Europe
TABLE 2 PROGRESS IN PRIVATIZATION, 1995
(percentage privatized)
in labor productivity, growth in total factor pro-
Mnfcuinfirms ductivity, and growth in exports. The data show
that for each firm these measures tend to be
Country firms weighted by output highly correlated. The econometric analysis fo-
cuses on the two most reliable indicators of re-
Bulgaria 8 7 structuring: growth in labor productivity and
Czech Republic 89 93 growth in total factor productivity.2
Hungary 67 65
Poland 61 60 The study defines a 'privatized' firm as one
Romania 15 12 for which more than a third of shares have
been transferred to private investors. It mea-
Slovak Republic 79 83sures the extent of privatization in a country
Slovenia 41 41 by using both a simple count of the firms clas-
sified as privatized and a count weighted by
output to reflect differences in size. On both
measures, the Czech Republic, Hungary, and
the Slovak Republic come out ahead, while Bul-
TABLE 3 ANNUAL LABOR PRODUCTIVITY GROWTH, 1992-95 gai Sloa Rbein hain m ade litle rgrs
garia lags behind, having made little progress
(percent) in privatization (table 2).
Country Privatizedfirmso State-owned firms All firms Impact of privatization
Bulgaria 12.4 -1.4 -1.4 The data show that labor productivity growth
Czech Republic 8.6 -2.6 6.8 across the seven countries averaged 7.3 percent
Hungary 6.0 3.2 4.8 a year for privatized firms during 1992-95, but
Poland 7.5 1.4 5.4 -0.2 percent for state-owned firms (table 3). The
econometric analysis indicates that privatization
*omania 1.0 accounts for almost all this productivity growth.
Slovak Republic 7.8 -4.1 5.1 The only exception to this pattern is Hungary,
Slovenia 7.2 1.8 3.6 where state-owned firms achieved half the pro-
ductivity gains of privatized firms. In Bulgaria
Average 7.3 -0.2 3.6 and Romania, where privatization has been in-
significant, productivity in state-owned firms is
a. Firms privatized by 1995. declining, pulling down labor productivity for
Source: Autors' estimates.
the manufacturing sector as a whole.
FIGURE 1 TOTAL FACTOR PRODUCTIVITY GROWTH Results are similar for the productivity of all
factors of production. The cumulative gains in
(percent. cumulative) total factor productivity for privatized firms far
15 exceed those for state-owned firms in the
12 sample (figure 1). The analysis shows that pri-
vatization has increased total factor productiv-
9 ivatizedfirns ity growth by about 4 percentage points a year.
6 The data also show that even a credible threat
of privatization promotes restructuring. For ex-
3 ample, in Poland, where the government's com-
0, State-ownedfirms mitment to privatization was perceived as
0 1 2 3 credible, the firms included in the mass privatiza-
Years since privatization tion program began to show rapid improvement
Source: Authors estimates. in profitability in 1994 and 1995-long before
they were formally privatized in November 1995.
The World Bank Group 25
FIGURE 2 ANNUAL GROWTH IN LABOR PRODUCTIVITY
AND REAL WAGES FOR PRIVATIZED AND
One possible explanation for this is that man- STATE-OWNED FIRMS, 1992-95
agers, improved their performance because they (percent)
expected to be held accountable by the ftiture
LSabor productivity * Privatized fiitms
new owners. Government plans for large-scale 15 D State-owned firms
privatization programs appear to boost produc- 12 Bulgaria *
tivity in state-owned firms, probably because of
similar anticipation and signaling effects. 9 Slovak Republic z R
6 ~~~~~~Poland-111 *Si-ovenia
6 Hungary *
Method of privatization 3 Hungary
O Rormania a Slovenia Poland Real
Many foreign advisers to the governments of Bulgaria Ro ania wages
transition economies initially believed that mass -3 Slovak Republic Czech Republic
privatization and insider buvouts wTould lead to -6
wveak pressures to restructure and that the pre- X -2 0 2 4 6 8 10
ferred strategcy should be sale to strategic inves- Source:Authors' estimates.
tors. But the study findls no significant differences
in the effectiveness of privatization methlods. TABLE4 AVERAGE ANNUAL INVESTMENT PER WORKER,
Productivity growth for privatized firms in the 1992-95
Czech and Slovak Republics, which chose mass
privatization. is similar to that in Hungary and
Poland, which have relied more on case-by-case
privatization (table 3). Also in the Czech Re- Country Privatizedfirms State-ownedfirns
public, where data on the results of different
privatization methods are available for a suffi- Bulgaria 2790 90
ciently long period, the study finds onlv mintor Czech Republic 3,290 470
differences amocng the methods. But it finds Hungary 2,990 460
strong effects of ownership concentration on Poland 1,880 410
the speed of restructuring. Romania 590 110
The role of wage restraint Slovak Republic 3,340 230
Slovenia 1,690 310
Restructuring is likely to be encouraged if the Source:Authors estimates.
workforce does not initially absorh all the pro-
ductivity gains through higher wages. Firms tmuhst FIGURE 3 SHARE OF NONPERFORMING LOANS TO
finance much of their investment with retained INDUSTRIAL FIRMS, 1992 AND 1995
earnings from current cash flow, especially when 9
the financial system is weak. But this demands Bulgaria _ 1995
new habits. While firms in industrial countries CzechRepublic
have relied mostly on internal cash flow to fi-
nance working capital or new investment, firms Hungary
in the formerly socialist economies have relied
heavily on loans from state-owned banks. Poland
Romania
The study finds that privatized firms have re-
tained most of their large productivity gains to Slovak Repubiic
finance productivity-enhancing investments. In
all the countries, labor productivity has grown Slovenia
faster than real wages in privatized firns (as _
shown by their position above the diagonal in 0 10 20 30 40 50 60 10
figure 2). That does not mean that real wages in Percent
these firms did not also grow rapidly. Blut sincc Soorce:Authors'estimates.
the firms maintained a large margin between
26 Privatization and Restructuring in Central and Eastern Europe
labor productivity and wages, they were able to than expected, making government interven-
sustain high levels of investment per worker tion in the banks unnecessary. By contrast, in
(table 4). By contrast, real wage growth in state- countries that have done little or no privatiza-
owned enterprises has exceeded labor produc- tion, firms' financial conditions did not improve
tivity gains, eroding internal financing. and the banks' bad-loan problems are worse
than the pessimists expected (figure 3). The
The analysis shows that privatization has had a speed of privatization of the industrial and com-
greater effect on wage restraint than have gov- mercial sectors has proved to be the most im-
ernment wage policies. Most of the seven coun- portant policy issue for the financial sector.
tries had a policy of limiting wage increases
(though by 1995, all countries had market- Conclusion
determined wages). For example, the Czech Re-
public, Hungary, and Poland each introduced an One of the most important policy questions in
excess wage tax during 1991-94. But even though the transition economies is what governments
government-led wage restraint applied primarily can do to speed the restructuring of firms and
(or exclusively) to state-owned firms, wages grew thus hasten the transition to a mature market
faster in the state sector than in the private sector economy. The study provides some answers.
in both Hungary and Poland. And in Bulgaria Rapid and comprehensive privatization lead-
and the Slovak Republic, which pursued more ing to concentrated ownership encourages
vigorous wage restraint in the state sector, real restructuring. Privatization also promotes re-
wages still outstripped productivity in state-owned structuring because privatized firms are more
firms-but not in the private sector. likely than state-owned enterprises to exercise
wage restraint-and wage restraint is vital to
Financial restructuring and the role of free up internal finance. Policies that increase
banks bank lending to firms, such as debt forgive-
ness and recapitalization, may do more harm
What actions, if any, should governments in the than good. The safest course is to recapitalize
region take to encourage the financial restruc- banks only as part of privatization and to en-
turing of overindebted firms? In industrial coun- courage negotiations for financial restructur-
tries, most financial restructuring takes place ing only after the banks are privatized.
through private negotiations between private
lenders (mostly banks) and private firms. But in This Note is based on a paper bv the authors of the same title (World
transition economies, most banks and many Bank Tecihital Paper 368, Washington, D.C., 1997)
firms are still under state ownership, so the in- To ensore conitparability. the stoidy adjustecd the data to reflect
centives to negotiate are different. Many banks ditferences in accotinting standards both over time and among
cenive bas coLuntries and excluded otilit, banking, and agricultural firms and
in the region inherited large portfolios of nco private companies. The distribution of firms among
nonperforming loans when state enterprises, suhsectors-mosttv food, textiles, chemtials, nsetals, machinery.
suddenly exposed to competition, started run- tnd transport equtipment-is similar across the cotintries.
-Lahor prt)durttvttv (valoce added per nosan-lhoort does not take
ning big losses. Audits done in accordance with into account depreciation. debt service, and taxes, which are more
international accounting standards showed that liket to cliffer frotTi countrv- to countrv because of differences in
up to 60 percent of the banks' loans were con- Set Sticn Ciaessens, Simeon Dan k or accoting standards.
sidered irrecoverable. The usual advice at this shiip and Corporate Governminent-Evidence fromii the Czech Re-
point was for the government to take over the public (Pnvatte Sector. June 1997)
bad loans and recapitalize the banks (usually
through an asset swap). Robert E. Anderson (anderson9@worldbank.
orgy), Simeon Djankov (sdjankov@worldbank.
The study's analysis shows that this course was org), and Gerhard Pohl, Europe and Central
premature. In countries that pursued large and Asia, and Middle East and North Africa
rapid privatization programs, privatized firms Technical Department, and Stijn Claessens,
have improved their profitability much more East Asia and Pacific Vice Presidency
27
The Drivers of the Information Revolution-
Cost, Computing Power, and Convergence
alatm(es Bon(id
In the past few years, there has been a technological phase-shift as computers have become
ubiquitous, communications technologies have multiplied, and the Internet has become a widely
used means of doing business. The three most powerful trends driving these developments are
the decline in the cost of transmitting information, the increase in the power of computing, and
the shift from analog to digital information technologies that has joined the telecommunications
and computing industries and merged market segments of the information industry. This Note
explains these three trends.
Cost of communicating chmnging the industryvs cost structure, moving
it awN ay from existing tariff setting mechanisms.
The cost of communicating has declined dra-
matically in the past twenty years. The cost of Cheap electronics. A key part of the telephone
a voice transmission circuit, for example, has infrastructure is network exchanges, made up
fallen by a factor of 10,000 as a result of the of switching equipment. Automatic switches
development of fiber optics, cheap electron- were originally electromechanical, but the
ics, and smart wireless (figure 1). sw,itches installed today are electronic-
essentially specialized computers. The advent
Fiber optics. First produced commercially by of cheap, pow erfiul, microprocessor-based com-
Corning Glass in 1970. fiber-optic cahle has puting has alter-ed the economics of switching.
become the increasingly dominant means of reducing costs and increasing reliability while
signal transmission in telephonysince the mid- also delivering new value added services for
1980s. replacing copper cables, microwave the user (such as call wT-aiting and caller ID).
transmission, and satellite. Optical fiber has ecx- Cheap electronics are also at the heart of celltn-
tremely high capacity (bandwidth) because of lar telephony and personal communications
the light it transports. The high frequency of services, which use electronmagnetic spectrum
light allows higher information density than more efficiently than conventional wrireless. And
conventional cable: a fiber thinner than a single cotmputing power now, makes it possible to run
hair can carry a laser signal combining many existing telecommunications infrastructure as an
thousands of telephone conversations, so that intelligent network,' improving capacity utili-
the cost per voice circuit becomes almost in- zation, lowering the cost of maintaining sw>;itches,
finitesimal. Because fiber optics reduces the and creating new services, such as virtual pri-
cost of signal transmission so much. ancd be- vate networks.
cause this cost is increasingly fixed (with main-
tenance costs much lower than for conventional Smart wireless. Wireless technology is evolving
cable, the cost is mostlv in installation), the toward I higher frequencies (inherently more
generalization of fiber optics is profoundlv information dense), vwith a range of clever
28 The Drivers of the Information Revolution-Cost, Computing Power, and Convergence
FIGURE 1 COST TRENDS IN OPTICAL FIBER TRANSMISSION
Index of transmission cost
per bits per second per kilometer
(log scale) * - compression algorithms to squeeze many con-
1000 bransmision osts have fallen versations into a given frequency (such as Time
45Mbps Division Multiple Access, or TDMA, and Code
100 e 90135MIPS Division Multiple Access, or CDMA). This de-
0 40D OMbpvelopment, coupled with cheap electronics,
^ 12-1J Gbps permits mobility for the user-in some situa-
1 2anwidps tions, wireless has become an alternative to
10 lGbps conventional wireline technology for basic ser-
0.1 - bps vices. Cellular telephony is growing rapidly,
0.01 more than doubling worldvvide every two years,
1975 1980 1985 1990 1995 2000 while fixed wireless is increasingly being de-
Note: Mbpsis megabits persecond- Gbps is gigabits per second. ployed for the "local loop," or local access net-
Source:AT&T data. work-that part of the network providing
FIGURE 2 PRICE TRENDS IN INFORMATION PROCESSING access to the end user. So, besides bringing
overall costs down, fixed wireless has also in-
Index of investment cost troduced real opportunities for competing pro-
per instruction per second viders of local services an area earlier deemed
(100 scale) Computing power per to be a natural monopoly. Finally, the cost of
IBM mainframe dollar invested has isen bya deploying wireless is much less sensitive to
factor of 10,OOOin 20 years subscriber density (the number of customers
10 . Digital VAX in a given area) than that of installing wireline,
Cray1 I so wireless is of great interest to developing
I countries, particularly for rural areas.
, IBM PC > Sun Microsystems 2
These three developments have necessarily been
0.1 accompanied by a move away from analog to
digital technology, in which signals are trans-
0.01 Pentium mitted as binary code. Digital telephone net-
1975 1980 1985 1990 1995 works ensure better quality and allow the use
of packing protocols for data transmission, such
Source: World Bank compilafton based on industry data. o
as frame relay, Asynchronous Transfer Mode
FIGURE 3 TRENDS IN TRANSISTORS PER MICROPROCESSOR (ATM), and the Internet protocol TCP/IP.
(log scale) Pentium II Power of computing
The number of transistors (7.5inillion)
1,000,000 on a microchip doubles Pentium Pro The second important driver of the information
every 2 years (f5 millon revolution has been the relentless increase in
1,000,000 48fi / Pentium the power of computing. Computing power per
4,000 (3.1 million) dollar invested has risen by a factor of 10,000 in
386iu twenty years (figure 2). Power has increased
100,000 2 / and costs have fallen because of the develop-
Motorola 68000 a ment of integrated circuits and microchips, be-
Intel * a Sun SPARC cause of increasing transistor density on
/0,000 microchips, and because of economies of scale
10,000 / in production.
Intel 8080
, Intel4004 Integrated circuits, miniaturization, and micro-
1,001' (23 ) chips. The modern electronics era began with
1970 1975 1980 1985 1990 1995 2000 the invention of the integrated circuit in a Texas
Source:WorldBankcompilationbasedonindustrydata. Instruments laboratory in Dallas in 1958. The
integrated circuit, which groups transistors and
The World Bank Group 29
other electronic circuits on a tiny piece of semi- began penetrating the business environment
conductor, is a breakthrough in product design aroun(l 1981, they were used mainly as stand-
because of its enormous potential for miniatur- alone workstations. Independent mainframe
ization and for reducing unit costs. The micro- computers, accessed by "dumb" terminals, con-
chip, essentially an entire computer on a chip. tinued to handle much of the heavy process-
was developed by Intel (as the four-bit 4004 ing. Today's business computers are connected
processor) in 1971. Its 2,300 transistors provide to one another in local area networks (LANs),
all the essential functions of a computer. and increasingly, these private netvworks are
interconnected through the Internet, the inter-
Increasing transistor density. The density of national "network of networks," which is dou-
transistors has been rising exponentially-a hling in size every year. Because of the growing
phenomenon sometimes characterized as interconnection, PCs' primary fuLnction has
Moore's law. In the 1960s. Gordon Moore. an shifted from document and spreadsheet man-
electrical engineer and a cofounder of Intel, agement to communication and information
observed that the number of transistors on a processing. 'T he growth of the Internet illus-
microchip doubles every one to two years. Be- trates Metcalfe's law (Metcalfe was the co-
cause computing power is roughly proportional founder of modern computer networking),
to the number of transistors, or "gates," on the which states that the value of a netwvork equals
microprocessor, this would translate into a the square of the number of interconnected
doubling of computing power per microchip nodes. As new users join the Internet, its value
every eighteen months or so. And because the for all users increases geometrically. Metcalfe's
cost of a microchip rises only slowlIy from one law illustrates how, networking PCs radically
generation to the next and represents onlv increases their value as a knowlIedge tool.
about 5 to 15 percent of the cost of the com-
puter, Moore's law woould translate into a near Convergence
doubling of computing power for a given in
vestment every eighteen months. In fact, the As costs have f'allen and digitalization has re-
growth in power over twenty-four years-from placed analog technologies in telecommunica-
the Intel 4004 (2,300 transistors) of 1971 to the tions, the telecommunications, information
Pentium II (7.5 million) of 1997-averaged technology, and media industries are merging
nearly 40 percent a 'ear, corresponding to a into a ibit industry" that manipulates voice,
doubling every twenty-five and a half months- image, video, and computer data in binary form.
close enough to Moore's estimate (figure 3). This convergence has profound implications for
the industries involved.
Economies of scale. Computing has also become n Communications and information services are
far cheaper and more powerful because of econo- being delinked from their underlying deliv-
mies of scale in production, not only of micro- elr infrastructure: telephone services can be
chips but of such essential ancillary equipment delivered through coaxial cable, data services
as mass storage (disk drives), removable stor- and Internet access through telephone lines,
age, and computer network equipment. The and cahle TV through direct broadcast satellite.
emergence of a set of de facto industry standards u Accompanying the delinking is increasing
-based on the first IBM personal computer (PC) overlap between the two primary components
of 1981. the Intel microprocessor instruction set, of the communications industry, which have
and the Microsoft operating systems (DOS and traditionally been segregated: common car-
Windows)-has also enabled producers to stan- rier conduit systems and networks designed
dardize equipment and sofm' are and encouraged to transmit signals anonymously (telephony)
price competition. and content-based information sources and
technologies (broadcasting).
One result of the increase in the powver of com-
puting and the decline in the cost of commu- Thus. it is nowx possible to receive radio
nicating is the rise of networks. When PCs broadcasts over the Internet (using telephone
30 The Drivers of the Information Revolution-Cost, Computing Power, and Convergence
FIGURE 4 THE CHANGING INFORMATION INFRASTRUCTURE UNDER CONVERGENCE
Before convergence After convergence
_ Ei 1 R_
networks), and telephone services can be pro- legal sanctions (throuigh the courts). Telecom-
vided by companies in cable TV (a broadcast- mnunications content has been largely tinregui-
ing medium). Broadcasting (from one to many) lated, because it is not technically possible to
now shadies fromn narrow-casting (custom- do so uising existing content regulation mecha-
tailored infor-mation) to one-to-one communi- nisms. But, convergence opens new r-ealms of
cation like telephony (figure 4). communications where traditional content regu-
lation cannot be applied. In this context, what
Convergence has important implications for do) policymakers do about decency, privacy,
policymakers. First, it has made existing models and intellectual property righits?
for the telecommunications industry obsolete.
Those models have assumed that telecommu- Convergence also opens up huige opportuni-
nications is a puiblic service, delivered through ties for developing couintries to accelerate the
a network that is a natural monopoly. Buit these rolOLout of connectivity to their populationS uISing
models are negated by the competition now innovative technologies and private sector-led
possible between segments of the delivery investment in a competitiv e mode. Cheaper
infrastructure (intermodal competition) and, commuLnications are offering new possibilities
increasingly, within segments (intram-odal for couintries to be internationally competitive
competition). Convergence thus means that and to "plug in" to the global economy-and
governments must lower barriers to entry and providing much more cost-effective ways to
overhaul telecommujnications regulatory sys- dleliver essential puiblic services to the poor.
tem--s to promote competition, moving away
from utility-type regulation.' See Petty >miitl. XWhUt the Tratnsformtation of Telecom Marrkets
M\ie.n'I for Regulatito (page -4 to tii', 'sir
Second, convergence raises serious issues re-
lating to content reguilation. In broadcasting, J/ames Bond (/bond@worldbank.org), Divisioni
countries have applied standards of decency. Chief, Telecommunications a/id DInbrmatics
privacy, and protection of intellectual property
rights using different mnechanisms, buit usually
relying on a combination of self-reguilation and
;31
Telecommunications Is Dead,
Long Live Networking
The effect of the information revolution on the telecom industry
Jamzes Bondl
Economic history teaches us that no industrv structures, and the increasing competition and
is imrnuine to chanige. Canals gave way to rail- globalization in the induLstry are forcing radi-
roads, which in turn bowed to road transport cal change.
when the cost of motor vehicles fell enough to
make it more cost-effective. Banks are scram- The end of scale and natural monopoly
bling to adjust to a world in which debt can be
raised on bond markets and consumers can In a conventional wired network, most of the
obtain many traditional banking services on- investment goes to establish the local loop, par-
line. And the telecommunications giants, which ticularly the civil works needed to extend the
have reigned supreme for the past fifty years, network to the end user. About two-thirds of
are being besieged in their turn as the infor- the assets on the balance sheet of telecommu-
mation revolution overturns the certainties on nications operators are `holes in the ground"-
which their strength is based. the trenches needed to lay the cables. Thus, in
a traditional netwTork, 70 to 85 percent of the
The threat to the giants arises from the huge de- cost of a call, even an international one. con-
cline in the cost of communicating. the increase sists of the cost of the low-technology link cov-
in the power of computing, and the shift to digi- ering only the last couple of miles.
tal technology. These forces have led to indus-
trial convergence as communications and
information services (such as basic telephone
service and cable TV) have been delinked from BOX 1 WHAT IS THE TELECOMMUNICATIONS INDUSTRY?
their delivery infrastructures (the telecommuni-
cations and cable networks). With the delinking ' TELECOMMUNIcATIONS INFRASTRUCTURE has three main components:
has come increasing overlap between the two terminal equipment, such as telephones and fax machines in
main components of the communications indus- users' homes and businesses, the local loop, generally a pair of
try: common carrier conduit systems (delivering copper wires connecting the tenninal equipment to switching
telephonyv) and content-based information
sources and technologies (broadcasting). As a equipment in the local exchange; and long-distance or interna-
result of these changes. new competitors are tional transmission networks, made up of fiber-optic cables,
emerging from unexpected directions, and the microwave links, and satellites.
market domination on which telecommunications * TELECOMMUNICATIONS SERvicES have focused on calling services: local
companies base their strength is melting away. calls (within the local exchange network) and long-distance or
The economics of telecommunications international calls. Increasingly, however, these basic services have
been augmented by the transmission of data in binary form and by
New technology has profoundly altered the value added services (such as call waiting and Internet access),
industry's cost structure, and as a result, the which increasefunctionalityforthe end userand generate
structure of the industry and its pricing meth- supplementary income for the telecommunications operator.
ods have become incoherent. The end of natu-
ral monopoly, the trend toward new pricing
32 Telecommunications Is Dead, Long Live Networking
FIGURE 1 BURIED COPPER WIRE GIVES WAY
TO WIRELESS IN LOCAL LOOP
AS TECHNOLOGY DRIVES COST DOWN
Because in the conventional local loop based
Annual lifetime cost (U.S. dollars) on copper wires the marginal cost of each new
1200 Buried copperwire subscriber declines no matter how many
1,000 existing subscribers there are, the telecommu-
800 \ nications sector-or at least its local loop
600 \ portion-has been considered a natural mo-
I1E I nopoly. Economic theory suggests that the best
400 111i_ 8 f ) fway to manage a natural monopoly is to cre-
200 ate a regulated utility, granting it a franchise to
O Wireless deliver the service in exchange for certain ob-
0.1 1 10 100 1,000 ligations (such as nondiscriminatory treatment
Subscriber density (lines per square kilometer) of consumers) and regulating the prices it can
(log scale) charge for the end product. This explains the
Source: Coopers & Lybrand data; European Bank for Reconstruction and Development nearly universal model for the telecommuni-
data. cations sector: a local monopoly company, of-
ten a public enterprise, with regulated prices.
FIGURE 2 WORLDWIDE EXPLOSION OF NEW SERVICES But wireless, cable TV, and other technologies
Index (1990 100) are now challenging the conventional local loop
1,500 based on wireline technology and buried cop-
Internet subscribers (94^20/) per (figure 1). In many cases, wireless is already
1,000 I cheaper per new subscriber than wireline. And
the much flatter cost curves of wireless show
that size no longer brings any real cost advan-
500 Cellularphones(51.5%) tage. It is possible to have several competing
providers of local service without raising the
Tele hone 5. n8%(5.8%) network's overall costs much. The implications
0 are considerable: the best way to deliver service
to customers is no longer through a utility but
Note:Figures in parentheses are annual growth rates overthe period shown. through competing providers of local telecom-
Source: World Bank and International Telcommunication Union data. munications services. The prevailing model for
the telecommunications sector in most of the
FIGURE 3 EXPLOSION OF NEW TECHNOLOGIES world is simply wrong. Furthermore, the tele-
communications reforms sweeping the world
19974Voicepager should focus more on the structure of the sec-
1996 Networked PC
199 Fixed wireless tor-providing as much potential for competi-
Digital TVina osbl-hno h
Asynchronous Transfe 1980 Pesonal computer tion as possible-than on the transfer of the
1993 /Mode 1918 Compact disc monopoly telecommunications company from
9 system 1971 Chip the public to the private sector. In many cases,
1987 Internet exposion 1 Laser however, the reverse is happening.
1986 Local area networks 1969 Computer network
1985 J Cellular telephone 1960 Satellite
_190 nTransistor The move to bandwidth-based pricing
1t45 <Computer
t95 Television
19 2 -Facsimele Almost universally, how much you pay for a
1906 Radio telephone call depends on how long you talk
1876-0 _ Telephone and how far away your correspondent is. If you
____ ___ ___ ___ ___ ___ ___ ___ ___ ___ ___ ___ ___ ___ ___ Telegraph
185D 190D 1950 2DOD are in Paris, it is much more expensive to call
Source:World Bank compilation. New York than Toulouse, and the price you pay
(beyond the monthly rental fee for the line) is
proportional to the time you talk. But that is not
The World Bank Group 33
BOX 2 THE IMPENDING COLLAPSE OF THE INTERNATIONAL
ACCOUNTING RATE SYSTEM
how costs aLre huilt up for telecomniunicationis
operators. For example, tariffs for international International tariffs fortelecommunications services are based on the
calls, based on the outdated accounting rate sv s- accounting rate system, which was developed as part of a regulatory
temn, nearlv alwiays far exceed the cost of pro- tradition holding that international services are supplied through a
viding the service (boX 2). bilateral correspondent relationship between national monopoly
A second problem relates to new services. The carriers. An accounting rate is the price the two national carriers (or
new communications services that custolmers their governments) negotiate for handling one minute of international
increasingly demancl send varying amnounts of telephone service. Revenues are shared between the two carriers. The
information per second down the transmission accounting rate system was originally intended to allow each carrier
line. Paging, for examople. reqtuires narrow to recover its costs for handling an international call.
bandwidthi (a small am-iount of information per The main problem with the accounting rate system is that for
second), while new multimcdia services (such
as teleconferencing ) require a huge amount of nearly every country the cost of transmitting a call has fallen
bxandwidth because transmitting video sends dramatically over the past twenty years, but the fall in price has
much more information dowx n the line than lagged this decline. As a result, the rate greatly exceeds the cost of
does transmitting sound alone. Btit most tele- providing the service, so accounting rates, which still assume that
communications operators dlo not offer chioice the sector is a monopoly, generate huge economic rents for tele-
in bandwvidth: cuistomiers get a standard tele- phone companies handling international calls. In some cases,
phone line, accommodating (64 kilobits per internationalcallsaccountlortheentireprofitofthesectorandeven
second (kbps) in Europe and 56 khps in the
lnited States. generate foreign exchange for the government
This system is showing signs of imminent collapse. New possibili-
To an increasing degree, the costs borne by ties for competition in international services make the sector a fertile
telecommunications operators are mtade up of area for arbitrage-and so we are seeing significant activity in call-
thiree elemients: ai fixed monthly7 aMoUnt, X\ 'hich back services, calling cards, Internet telephony, and the like. These
corresponds to the capital costs of the local
loop) a one-timc cost for ach connectin, cor- new sources of competition undercut the hugely inflated accounting
responding to the cost of swvitching that call; rates and eat into the income of telecommunications operators-
and a transmission cost, which is proportional especially those offering the lowest prices. As a result, the U.S.
to the bandw idth. But actual ctistomir charges regulator, the Federal Communications Commission, is seeking to
are quite different. And because it is increas- replace international accounting rates with a new benchmark
inglv possible to compete for customers, nese
agile operators arc emerging that take advan- system based more closely on actual costs.
tage of the possibilities for arbitrage between
tariffs and actual costs. Much of the new, activ -
ity is in international service. But nexw plaVers e\en for international calls (in addition to the
are also emerging in other areas, such as Inter- montlily rental for local access).
net setAvice providers, w hich enable customers
to place long-distance calls for the price of the Intermodal competition, globalization,
local loop connection (figure 2). and the WTO negotiations
It is safe to predict that competition andn mar Policvnmakers increasingly accept that competi-
ket forces will drive tariffs closer to long run tion in the local loop is both possible and desir
incremental costs. both in level and in struc able. But competition is also bringing changes
ture. And within a couple of years. consumers that policvmiiakers are much less knowvledgeable
will probably be able to but the bandwidth ahbout-in the explosion iof new techniologies.
capacity thev need for a given connection, products, and services competing with one an
wxhich, because of declining transmission costs, other to deliver connectivitx to the end user (fig-
wvill cost no more than a few cents per hour utre 3). NeN- technologies increase arbitrage
34 Telecommunications Is Dead, Long Live Networking
possibilities for new operators, and they compli- thing from bulk transmission to value added
cate the work of regulators. But above all, they services, in such a world. Instead, the world
challenge existing operators, which are often slow would segment into wholesalers, which would
in responding to new customer demands. invest in and sell capacity; retailers, which
would be in contact with the final consumer;
Also changing is the geographic service area in and traders and brokers, which would inter-
which end users are interested-increasingly not mediate supply and demand for capacity.
only national but international. International traf-
fic is growing by 12.4 percent a year, compared Globalization goes beyond the actions of na-
with 5.9 percent for domestic calls. Thus, as com- tional telecommunications operators. As the
petition in domestic markets becomes the norm, World Trade Organization (WTO) focuses in-
ternational attention on liberalizing trade in ser-
vices, policymakers are beginning to realize that
The trend toward global alliances may opening telecommunications markets to foreign
the desire investors and operators has important benefits
stem morefjrom the desire offormer both locally and globally. The WTO has placed
telecommunications at the top of its agenda
monopoly players to re-create at the for multilateral trade liberalization (page 39).
As markets open to foreign participation, and
internati'onal level the oligarchaes to as technology creates new markets, we can
expect to see an entirely new cast of players
which they are accustomed in their investing in markets that are new not only tech-
domesti'c markets thanfrom the nologically but also geographically.
underlying marketforces in the industry.
The old monopoly telecommunications sector
is fast disappearing. National telecommunica-
consumers will find operators that can offer tions markets are fragmenting into a multiplicity
packaged services on an international scale in- of niche markets at the same time that trade
creasingly attractive. The growing demand for barriers are falling. Many new operators are
international services helps explain the trend emerging, each targeting the segment that best
toward global alliances among telecommunica- corresponds to its comparative advantage. In
tions operators. But this trend may stem more this new networked bit industry, offering a huge
from the desire of former monopoly players to range of competing technologies and services,
re-create at the international level the oligar- the future for the incumbent telecommunica-
chies to which they are accustomed in their do- tions companies looks increasingly bleak. Over
mestic markets than from the underlying market the next decade, as the market shifts from un-
forces in the industry. der their feet and as new, more nimble actors
emerge, we can expect the dominance of the
In an alternative future, transmission capacity telecommunications operators over their tradi-
and bandwidth might become tradable com- tional markets to erode spectacularly. Some
modities, with a spot market on which capac- countries are likely to see their major telecom-
ity is bought and sold in half-hour slices and munications operators fail as new players ap-
megabit-per-second tranches. The spot market pear that quickly become household names
could be associated with a ftitures market on around the world.
which contracts would allow operators to hedge
future positions. It is difficult to imagine today's James Bond (fbond@worldbank.org), Division
mastodon operators, which now cover every- Chief Telecommunications and Informatics
35
Telecommunications Reform-How to Succeed
B/dirn W1lZenius
Today, more than sixty emerging economies- services. Pressures from interest groups-incum-
some twenty-five in Sub-Saharan Africa alone bcnts wanting continuLed protection. new en-
-are at some stage of transformation fromn a trants seeking special deals, treasury officials
state telecommunications monopoly to private- expecting to use sale revenues to reduce budget
led, competitive markets. WX7hen wvell done, such deficits, financial advisers earning success fees
reforms can be positive sum games in wrhich all tied to transaction prices-can steer reform off
stakeholders gain-customers, existing and new this track. In particular, sale strategies that drive
operators, employees, domestic and foreign in- up the prices for existingJ companies or new
vestors, and the government. Faster market licenses can repress growth, recluce the fund-
growath, new and better services, loxver costs, ing availahle to invest in the companies, or re-
and, eventually. lower prices followv. This Note suit in high tariffs. The Mcxican government,
outlines the key factors in successful telecom- for example, concluded that the six-year long-
munications reform. distance monopoly granted to Telmex in 1990
had led to higher consumer prices and slower
Get support at the top growth thani would have resulted under coin-
petition. So, in 1996. it chose to forgo the high
Reform is most likely to succeed if it is led at fees it could have obtained by tendering one or
the highest level of political authority. That is two new licenses and instead opted for unre-
usually the head of government, wvho then allo- stricted entry. In India, the exorbitant prices bid
cates responsibility for the reform to a single for second fixed operator licenses in 1996. com-
person with direct access to senior government bined with modest revenue projections (based
officials, freedom to cut red tape, and resources on the low per capita income), are making it
to assemble a small support team and hire the difficult to raise debt financing for investment.
necessary experts. Such was the case in the BY contrast, the government of Bolivia priva-
privatization of Mexico's telecommunications tized Entel in 1996 by issuing new shares for
company, Telmex. The president announced the which the xvinning bidder paid US$600 million-
reform in August 1989, appointed the minister flunds immediatelv available for investment in
of finance chairman of the board and gave him the company. In Brazil, the consortium that won
overall responsibility for the privatization, and the cellular license in Sao Paulo in 1997 with a
handed over the chief executive's job to an ex- US$2.5 billion bid-fotir timnes the government's
perienced public administrator syith a clear re- asking price and 60 percent more tlhan the sec-
form mandate. Privatization wTas cormpleted in ond-highest bid-is likely to pass on the cost to
December 1990. By contrast. Brazil's attempts customaiers through munch higher tariffs than those
at reform begyinning in the early 1980s did not proposed by rival bidders.
muster the necessary political muscle. Not until
1997 has real progress been made. Sale strategies that place less emphasis on cash
up front can, moreover, yield substantially more
Sort out conflicting objectives early cash to the government later. For example,
axvarding a cellular license to the bidder that
The primary purpose of reform is to get con- offers the largest build-out plan-rather than
sumers more, better, new, andc less costly the one offering the highest license fee-can
36 Telecommunications Reform-How to Succeed
increase tax revenue for years to come by cre- The interconnection olligation of the dominant
ating more business. And initially selling only operators, the principles under whichi terms of
the minimuLm number of government shares interconnection will be negotiated, and the pro-
needed to effectively transfer control of the state cess and timetable for a regulatory decision if
company to the new owners (usually 20 to 30 the parties fail to reach agreement must be
percent) allows the government to float the hal- clearly spelled out. A new operator's ability to
ance later and obtain much higher prices, once reach (and be reached by) customers of the
the company appreciates uncler private man- existing operator and to use parts of existing
agement. In the Telmex privatization, for ex- networks on reasonable technical and price
ample, the government initiaLly sold 20 percent terms (rather than building complete new fa-
of the shares to a strategic investor in 1990 for cilities) plays a big part in determining not only
US$1.8 billion, then sold 31 percent more its own viability but also the economic efficiency
through public offerings in 1991 and 1992 for of the sector. In Ploland, failure to sort out inter-
US$4.5 billion-70 percent more per share. connection with the incumbent meant that of
some 200 licenses issued to independent op-
Set clear policies, rules, and erators since 1990, only about twelve were in
procedures use in 1996. Licensees cited the main iinpedi-
ments as unfavorable terms for sharing revenues
T he business offered to investors must be with the dominant state operator. limited access
clearly defined in the laws, regulations, and to its network, slow negotiation of interconnec-
main transaction documents (licenses, contracts tion agreements. and a prohibition on setting
of sale). The most critical policy issues relate up their own transmission facilities.
to pricing, competition, and interconnection.
In pricing, governments most bite the bullet Reforms should follow clearly definec processes
early and rebalance tariffs. The price an op- that are open to participation and review by all
erator is allowed to charge its customers is the interested parties. The public should be kept
most important determinant of profitability and informed. Market mechanisms. not individual
ability to finance growth. Existing tariffs are negotiations, should be usedc to select partners
often way out of line witlh costs-far too high and cletermine the right sale prices. And the
for international calls, too low for local calls award of licenses and contracts should strictly
and fixed charges, and somewhere in between adhere to the evaluation criteria announced at
for domestic long-distance calls and initial con- the outset. Once a window of political opportu-
nection (sometimes tariffs are also too low nitv for reform opens, time is of the essence-
overall). Including rebalancing plans in licenses but that shotuld not be used as an excuse to cut
or contracts often delays further reforms; new corners or strike deals behind closed doors.
owners tend to defer raising some prices to
avoid the public fallout, yet later expect the Clear rules and processes must also be applied
licensing of competitors to be delayed because to the regulatory function. The loCcus and func-
tariffs remain unbalanced. In setting new tariff tions of regulatory authority and the basic pro-
structures, calculating the actual costs of each ceclures that will govern its relationships with
operator is seldoni a viable method. Rather, operators and customers must he defined, pref-
tariffs observed in competitive markets proh- erably by law. That does not mean that a ftull
ably offer the best guidance on efficient prices. regulatory capability must be in place before
Some cost elements (land, labor, taxes) vary major reform steps can be undertaken. Initial
considerably among countries, but the main regulatory decisions can be written into licenses
costs (equipment, capital) are determined in and contracts of sale. A core decisionmaking
global markets and international benchmarks capability in the form of a commission, say.
are thus relevant. As the market becomes more and a secretariat with processing capability,
competitive, pricing can increasingly be left to supported by outsourcing of expertise, can
the operators. handle essential tasks in the first two or three
The World Bank Group 37
TABLE 1 FASTER GROWTH IN OPEN, PRIVATIZED MARKETS
(annual percentage growth in main telephone lines)
years, such as issuing licenses, marnaginyg coni 1984-89 1989-94
flicting dcemands on the radlio spectrum, and
resolving interconnection disagreements. Other Brazil, Colombia,7.0 7.8
areas of competence can he gradually devel Ecuador,Peru,Uruguay
oped as needed. Chances are that successive Argentina, Mexico, Venezuela 6.7 t1.3
problems will arise, peak. and then decline to
a low simmer, so that a permanent. comilpre- Chile
hensive in-house capability may never he
needed. Moreover, in most emerging econo-
mies, anvthi regulatory ~~~~~~~state monopolies
mies, anvthing heyondl a minimalist reglatory ivatized monopolies
institution* is not feasihle. open, privatized market Source: Pyramid Research data; World Bank data.
Open all markets to competition
Without competition, the benefits from increased tions companies. failed shortly before closing
private participationi will not he fully realized. when tlhe winning hidder realized that the gov-
In Latin Anmerica, for example. countries that eirnent did not intend to grant Svyazinvest a
granted monopoly privileges of six to ten years license to builcl its ov,-n long distance netwxork.
to privatized state enterprises saw connections
grow at 1.5 times the rate under st.te monopolies Enhance credibility and stability
-but at onlv lhalf the rate in Chile. wxhere the
government retained the right to issue comrPt- Even if a government gets all the policies, rules
ing licenses at any timc (table 1). Rural areas, and procedures right, operators and investors
too. can become an attractive business undler vxill c)rmne and remaiin only if they believe that
liheral entry and pricing policies. In Chile, gov- the government will stay the course. Govern-
ernment subsidies equivalent to less than 0.5 iments can do several things to enhance cred-
percent of total telecorninunication.s rev enue. al- ibilitv and stability. To safCguard rcforms against
locatedl through competitive h-idding in 1995, political changes. governmnents slhould develop
mobilized twenty times as much private invest- the reforms wvith the support of maior stake-
ment to extend hasic telephone access to rural holders-various branches of government,
areas. The program bromught service to ahout a public and private sector users, chambers of
thircl of the rural population lackinig it. commerce, consumcr groups, largc cnterprises
(inclutcling state-ow ned firms) that could become
Contrary to views often expressed by financial alternative network providers. local banks and
advisers, investors are not opposCd to com- investors, and the staft and management of ex-
petition-as long as they are inot also burdened isting operating comnlpaniies.
with regulatory uncertainity, unrealistic service
obligations. and rigid tariffs and employvment In cmerging economies, most with strong growth
rules. This is true even in small, low'-income potential. the concerns of labor can be readily
markets. Ghana Telecotmv was stuccessfullv accommodated. MIost workers stand to gain from
privatized in late 1996 at the saiiie time that a higher salaries, improved career prospects, and
license was awvarded for a second full-service new opportunities as employees or entrepre-
national operator and three other cellular comIl- neurs in a rapidly expanding market. Groxxth
panies xvere already in place-and the price alloiws major gains in labor productivity with
per line was similar to that paid for the mo little reduction in personnel. As Ghana Telecom
nopoly in neighboring C6te cl'Ivoire. Btut lack prepared to privatize in 1996, some 500 wsork-
of clarity regarding competition policy does ers (14 percent of the telecomiimunications and
drive ixrestors away. Partial privatization in postal xxworkforce) agreed to leave xxith sever-
1996 of Svyazinvest, the Russian holding com- ance packages that cost the government less
pany of eighty-five regional telecormmnunica than 3 percent of the initial proceeds from pri-
Telecommunications Reform-How to Succeed
vatization. After privatization, potential labor the country's own reform targets, abide by a
problems largely disappeared as a result of common set of regulatory principles, and rec-
managements promises of no forced redundan- ognize the WTO as an instance of intergovern-
cies, the introduction of training programs, and mental appeal (see page 39). All this is likely
expectations of growth. By contrast, labor unions to provide comfort to investors worried about
whose concerns-and political clout-had been regulatory risk. Similarly, loans, credits, and
ignored brought Sri Lanka's reform program to guarantees from multilateral agencies such as
a halt in the mid-1980s. In the restructuring of the World Bank Group involve government ob-
state telecommunications enterprises in Latin ligations that can be tailored to help offset such
America, an additional enticement has been of- risks as failure of the government to abide by
fered-employee stock option plans that trans- the terms of licenses (on pricing, for example)
fer about 5 percent of shares to employees on or ensure access to foreign exchange for debt
favorable terms. service or dividend payments. A 1993 invest-
ment of USS90 million by the International Fi-
Essential for reducing investor risk is limiting nance Corporation and the European Bank for
the opportunity for discretionary government Reconstruction and Development in the Hun-
or regulatory intervention in business, espe- garian state telecommunications company mo-
cially in the early years. In Uganda, initial de- bilized US$1.2 billion in foreign funds at the
cisions on tariffs, service obligations, and time of privatization.
default interconnection terms are being writ-
ten into licenses and contracts (as was also done Investors, operators, and customers will be re-
in Ghana). Numbers that will remain firm for, assured by a telecommunications law that es-
say, five years-subject if necessary to auto- tablishes broad principles and rules governing
matic adjustment, based on simple formulas, the sector. But a law with a narrower objec-
for inflation, foreign exchange, or other fac- tive, such as establishing a regulatory author-
tors-are more effective at reducing risk than ity, may suffice. The timing of amending or
are rules for calculating these numbers. replacing a dated law must weigh the poten-
tial delays and political cost. Telmex was priva-
Telecommunications reforms gain credibility tized in 1990 in the framework of a 1938
when coupled with broader programs in which transport and communications law-but pas-
the government has a large stake. The privati- sage of a new law in 1995 was essential to
zation of ENTel in Argentina was the flagship of open the market for competition in 1996.
President Menem's multisectoral public enter-
prise reform program in the early 1990s, and Conclusion
everyone knew that a failure by the govern-
ment to stick to the rules it had set for telecom- Major transactions such as a privatization or the
munications would have undermined the whole issuance of new licenses tend to drive the re-
program. More generally, telecommunications form agenda, but change continues well beyond
reforms benefit from a healthy overall business these transactions. Following the rules and hon-
climate-political stability, sound macro- oring commitments help consolidate an envi-
economic management, and policies favoring a ronment for sustainable growth. Also critical are
private-led, competitive, open economy. to build a regulatory capability to suit changing
needs, take every opportunity to enhance com-
Anchoring key elements of reform in interna- petition, and address any persistent gaps be-
tional frameworks also adds credibility. Every tween development and commercial objectives.
World Trade Organization (WTO) member
country that subscribes to the telecommunica- Bjorn Wellenius (bu.ellenius@n'orldbanpk.org),
tions agreement of 1997 enters a binding inter- TelecommunicationsAdOiser, Telecommunica-
national commitment to implement aspects of tions and Informatics Division
39
Liberalizing Telecommunications and the Role of
the World Trade Organization
Carlos A. Porino Braga
Februarv 15, 1997, will be remembered as a land- on the best way to proceed. Developing coun-
mark date in the history of the multilateral trade tries, for example, opposed the negotiations
system. On that day, sixuv-nine governments for- either because they believed that they did not
malized commitments to liberalize their basic enjoy comparative advantage in the relevant
telecommunications services under the General industries or because they feared that these
Agreement on Trade in Services (GATS). The negotiations would intrude into other areas,
resulting Decision on Commitments in Basic such as foreign direct investment policies and
Telecommunications has both symbolic and national regulatory regimes. Needless to say.
practical meaning. The final act of the Uruguay other topics-for example, agriculture and
Round in Marrakech in 1994-giving birth to textiles-were much more prominent on the
the World Trade Organization (WTO 1-was de- negotiating agenda of developing countries.
scribed by many as the beginning of a new era
in international trade rtules. The agreement on For telecommunications, resistance to trade
basic telecommunications can be characterized negotiations also came from major players in
as the first major accomplishment of this new the industry. After all, state-owned enterprises
era. Moreover, it will foster the liberalization of were the suppliers of telecommunications ser-
telecommunications, bringing significant benefits vices in all but a handful of countries, and inter-
for industrial and developing countries alike. national telephony was conducted like a cartel,
with transactions closely regulated under rules
This Note reviews the evolution of the services negotiated under the International Telecommu-
and telecommunications negotiations, the scope nication Union (ITU). Using trade negotiations
of the new set of multilateral disciplines, and to promote the liberalization of telecommuni-
the implications of the agreement for WTO cations was an alien concept to most of this
meimbers, particularly developing countries. community. It was also perceived as a threat
The basic message is that the agreemenit is a to national regulators, and in some countries,
major accomplishment. but implementing the it was even portrayed as a threat to national
commitments it contains will pose a significant sovereignty.
challenge for many of the developing country
WTO members. But the potential benefits of Despite the opposition, the services negotia-
following through with these commitments are tions progressed more smoothly than most
not trivial, and policymakers should maintain analysts had predicted at the beginning of the
the focus on the liberalization agenda. build- Uruguay Round. The internationalization of
ing on the commitments already made. services is at the very core of the process of
economic globalization. Service industries (for
A clash of worlds example, telecommunications, transport, finan-
cial services) provide critical links among
In the early 1980s, when the trade community, geographically dispersed markets. Efficient,
under the leadership of the United States, began high-quality links are fundamental for trans-
to discuss the inclusion of selvices in the multi- national corporations-the most dynamic actors
lateral trade regime. there was no consensus in globalization-and this critical need explains
40 Liberalizing Telecommunications and the Role of the World Trade Organization
their strong interest in the establishment of to publish anci make available to the public
multilateral disciplines in services trade. Pushed the laws and regulations that affect trade in
by these powerful interests anci the growing services.
recognition of the potential benefits of liberal-
izing services, the negotiating agenda evolved Market access ancd national treatment are spe-
gradually, and by 1993, the basic architecture cific obligations under the GATS. They apply
of the GATS had been agreed on. only to the sen vice industries and activities listed
by a country in its schedule of commitments.
The GATS in a nutshell' These obligations are specified at the level of
eaclh of the four modes of supply and subject
The GATS comprises the framework agreement to the limits made explicit in the offer. The
(with its twenty-nine articles and eight annexes) GATS adopts a positive list" approach with
as well as the schedules of specific commit- respect to sectoral coverage of service indus-
ments and the lists of exemptions to most tries-that is, only the industries and activities
favored-nation (MFN) treatment submitted by scheduled in the commitments are subject to
member countries. It covers four modes of the GATS's specific obligations.
international delivery of services: cross-border
supply (for example, international telephony), The treatment of telecommunications
consumption abroad (tourism), commercial in the GATS
presence (provision of services abroad throughl
a branch, agency, or subsidiary), and the pres- The definition of telecommunications services
ence of natural persons (entry and temporary for GATS purposes is comprehensive, encom-
stay of foreign individuals in order to supply a passing both basic services-those that involve
service). simiiply end-to-end transmiiission of voice or
data-and value addled services-those that
It broadly follows the tradition of its counter- modify the form or content of the messages
part for trade in goods-the General Agreement relayed through the networks.' By the end of
on Tariffs and Trade (GATT)-emphasizing the tJruguay Round, forty-eight schedules
nondiscrimination and imposing limits on the (representing 59 of the 125 governments parti-
tise of quantitative restrictions on trade. But it cipating in the negotiations) contained com-
introduces innovations, covering transactions mitments in telecommunications. BIut almost all
associated with commercial presence and intro- of these commitments covered only value
ducing a concept of market access that goes added services, reflecting the resistance still
beyond border restrictions (for example, in facing this novel approach to telecommuinica-
principle, it proscribes restrictions on the type tions negotiations. In short, most of the rel-
of organization under which foreign providers evant markets for communications continued
can establish commercial presence). to operate otitsicde multilateral disciplines.
Unconditional MFN treatment is a basic obli- The Uruguay Round accomplished some impor-
gation of signatories that applies to all services, tant results for the sector, however. First, it raised
an obligation that bars a WTO member from awarencss about the potential role of trade ne-
treating other members less favorably than any g,otiations in fostering the liberalization of tele-
other country. BLit the GATS allows MFN ex- communications. Second, it helpecl to diminish
emptions as long as the member country icden- the gap in understanding between the trade and
tifies them explicitly. The list of exemptions is teleconmmunications communities by promoting
supposed to be time-bound and, in principle, a dialogue on their distinct approaches to regu-
shoulci not last more than ten years. Another lation. Thircl, it established that access to tele-
basic obligation of members is a commitment communications services was critical for trade
to transparency, which requires governments in services and that users were entitled to fair
The World Bank Group 41
terms of access (as outlined in the Annex on cess commitments would be greatly reduced
Telecommunications of the GATS). unless a procompetitive rcgulatory framew ork
was also put in place. A draft reference paper
But the limited progress in effective liberaliza- describing regulatory disciplines supportive of
tion of basic telecommunications led WTO market entry was negotiated. and most coun-
members to agree to continue the negotiations tries became signatories to this text (partially
beyond the date of the Round's completion or in its entirety) in the context of additional
(April 15, 1994). Basic telecommunications commitments made in their offers (expanding
joined maritime transport, financial services, on their market access and national treatment
and the movement of natural persons as topics
for sectoral negotiations. The Negotiating
Group on Basic Telecommunications (NGBT) 7ThiS Ccn be charactePlzedt as
was created in Mav 1994, with a deadline of
April 30. 1996, for completing the talks. thefirst undltilateral eJfort to deal
From the NGBT to the GBT explicitly with su bstantive aspects of
Participation in the NGBT was voluntary. Ini- coimpetition policy. Even though limitect
tially, fifty-three WTO members decided to par-
ticipate in the negotiations, with twenty-four to teleconmmunications, it w as a major
other governments attending the meetings as
observers. The attitude of most participating achiev)emWenJt, and it paves th wae Ivyjor
countries about the usefulness of engaging in
these negotiations had shifted significantly by fiatture multilateral disc'Alines a d
then. In part, this simply reflected a better
understanding of the potential benefits of lib- international harmonization.
eralizing telecommunications. More funda-
mentally, however, it reflected the growing
recognition that the industry faces a paradigm commitments). This can be characterized as the
shift. Technological progress is rapidly erod- first multilateral effort to deal explicitly with
ing the sustainability of old practices based on substantive aspects of competition policy. Even
monopolistic behavior, state control, and thouglh limited to telecommunications, it was
protected markets for local providers. Callback a major achievement, and it paves the way for
systems, virtual private networks, the Internet. future multilateral disciplines and international
and the growing promise of modern satellite harmonization.
communications are multiplying the opportu-
nities for bypassing telecommunications mo- By April 1996, thirty-four offers ('encompass-
nopolies. At the same time, the increasing ing forty-eight governments. with the European
information intensiveness of transnational cor- Union's submission counting as one) were on
porations and the dramatic reductions in the the table. Still, some countries-particularly the
cost of communications create additional in- United States-- were dissatisfied with the qual
centives for customers to actively explore by- ity and coverage of the offers. Moreover, in
passing alternatives. the final phase of the talks, the issue of satel-
lite services-that is, to wlhat extent explicit
In the NGBT, the focus of the debate rapidly provision for these services needed to be made
mio\ved oIn froii "wlhy to liberalize" to '-how to in the offers-added noise" to the negotia-
liberalize." Important conceptual progress was tions. As a result, no deal was attained by the
made as participants recognized that for tele- deadline of April 30. 1996. Given the progress
communications, the value of the market ac- already achieved. however, there was broad
42 Liberalizing Telecommunications and the Role of the World Trade Organization
FIGURE 1 COUNTRIES MAKING BASIC TELECOMMUNICATIONS COMMITMENTS
Voice telephony
Local
Domestic long distance
International
Resale
Data transmission
Private leased circuit services
Terrestre mobile
Other terrestre mobile services
Mobile satellite services
Fixed satellite services
Trunked radio
Reference paper (additional commitments)
0 10 20 30 40 50 s0 70
* Developing countries * High-income countries
Source: World Trade Organization data.
support for continuing the negotiations. Seiz- give rise to anticompetitive practices by for-
ing this opportunity, Renato Ruggiero, Director- eign monopolistic carriers), on satellite services,
General of the WTO, suggested that countries and on what constituted an adequate "critical
should be given a chance to improve on their mass" for a deal. Other controversial issues
offers, and February 15, 1997, was established included how to avoid discriminatory practices
as the new deadline for the negotiations. A new in the allocation of spectrum and how to draw
body-the Group on Basic Telecommunica- the line between telecommunications and au-
tions (GBT)-was created to carry on with the diovisual services given the growing techno-
negotiations, replacing the NGBT, and the rules logical convergence in these areas.
of participation were changed to make all WTO
members full participants. Gradually, however, technical and political
solutions began to emerge. With respect to
The negotiations restarted in July 1996, and by international services, the United States unilat-
the WTO Ministerial Conference in Singapore erally announced a new policy toward inter-
in December 1996, several countries had al- national settlement rates in December 1996,
ready tabled improved offers, signaling support creating a mechanism for addressing the con-
for a successful conclusion of the negotiations. cerns of its own carriers about the distortions
Still, some thorny issues remained. Lively dis- of the accounting rates system outside the WTO
cussions continued on international services framework. This helped deflate opposition to
(for example, countries with more liberal the agreement based on concerns that it could
regimes were concerned that an MFN commit- foster anticompetitive practices (for example,
ment to liberalize international services could through one-way accounting rate bypass). Also
The World Bank Group 43
FIGURE 2 DISTRIBUTION OF WORLD TELECOMMUNICATIONS REVENUES BETWEEN PARTICIPATING AND
NONPARTICIPATING COUNTRIES
United States (29.70)
European Union (28.27)
Japan (15.59)
Australia (1.89)
Canada (1.78)
Switzerland (1.48)
Republic of Korea (1.45)
Brazil (1.43)
Mexico (1.08)
Argentina (1.00)
Other countries that made commitments (7.44)
Countries not participating in the telecommunications agreement (8.56)
0 5 10 15 20 25 30
Percentage share
Note:Warld revenues from telecommunications services amounted to US$602 billion in 1995.
Source: World Trade Organization data.
helping to pave the way to the final agreement the final stages of the negotiations.' By early
were the adoption of a technologically neutral 1997, it became clear that a "critical mass' of
approach to scheduling (that is, unless other- offer,s would be achieved. On February 15,
wise noted, the commitments would cover all 1997, the telecommunications talks were suc-
transmission possibilities including satellite ser- cessftilly concluded.
vices) and the acceptance of the concepts that
frequency and spectrum management should The scope of the agreement
not be used to undermine market access cotim-
mitments and that MFN exemptions cotild temn- Sixty-nine WXrTO members tabled commitments
porarily be used to address the differences in by FebrtLary 1z, 1997f These schedtiles will
treatment of audiovisual services. become formally binding by January 1, 1998.
Not only were several new offers added to
In a parallel effort, governments, the WTO, and those available in April 1996, but thirty-two of
several other multilateral organizations worked the thirty-four original offers wxere revised, typi-
to raise awareness of the importance of the cally leading to more substantive commitments.
negotiations for developing cotintries and to Commitments were made in all basic telecom-
help these cotuntries prepare their own offers. mtnications services by both high-income and
The World Bank. for example, through its In- developing countries (figure 1). Moreover, most
formation for Development (injh)Dev) program participants made commitments either to all or
and in close cooperation with the WTO, spon- to parts of the reference paper, subscribing to
sored a project to provide technical assistance procompetitive regulatory principles (for ex-
to more than twenty developing cotintries in ample, the establishment of independent regu-
44 Liberalizing Telecommunications and the Role of the World Trade Organization
FIGURE 3 COUNTRIES WITH LOWEST TELEDENSITY LESS ACTIVE IN WORLD TRADE
ORGANIZATION NEGOTIATIONS
Teledensity in 1995
Main telephone lines
per 100 inhabitants
*>25
* 10-25
0] 1-10x
WTO offers
(February 1997) _
* Countries making offers
The boundaries, colors, denominations, and any other information shown on this map do not imply, on the part of the World Bank
Group, any judgment on the legal status of any territory or any endorsement or acceptance of such boundaries.
Source: international Telecommunication UnijO and World Trade Organization data.
lators, the adoption of competitive safeguards, The exact implications of the agreement for a
measures to ensure interconnection, transpar- particular country can only be assessed by a
ent and nondiscriminatory practices with careful analysis of its schedule of commitments,
respect to licensing, and universal service including phasing considerations, list of quali-
obligations). fications by activity and mode of delivery, and
The World Bank Group 45
eventual recourse to MFN exemiiptions (nine driver of the paradigm shift in telecommunica-
governments claimed such exemaiptions for cer- tions, and it explains to a large extent the chang-
tain activities). It is fair to say, hoV ever, that in ing attitude in the industry on the desirability
contrast with the GATS's results in 1994, xvhen (and inevitability) of competition. But credible
most schedules were characterized by status rules relating to market access, constraints on
quo commitments (that is, governments basi- discrimination, and a procompetitive regula-
cally bound themselves not to adopt more tory environment play an important part in
restrictive policies). the outcome of the basic shaping the outcome of this revolution." par-
telecommunications negotiations w-ill foster ticularly in influencing the distribution of its
significant additional liberalization.' In this benefits.
sense, the agreement proved wrong those ana-
lysts who were skeptical of the role of sectoral Private capital is expected to take the lead in
negotiations in fostering liberalization at the funding investments in telecommunications in
multilateral level. the developing world. In the early 1990s. 65
percent of the financing for basic wireline tele-
The markets affected by the agreement repre- comnmunications in the developing world came
sent more than 90 percent of the world market
for telecommunications (figure 2). Dev eloping
countries account for less than 20 percent of 7The utaikets affected by the agreement
the global revenues from telecommunications
services, but they are the fastest-growing markets represet wore than 90 percent of the
for these sernvices. Their participation in the WTO
process is thus important not only for develop- w'orld zarket Ibr telecommu inications.
mental reasons, but also because these markets
are bound to increase in relative importance.
The countries with the weakest telecommuni- from internal sources (,profits), with commer-
cations infrastructure (such as in Sub-Saharan cial funds (20 percent) and official sources (15
Africa) are also those that participated less active- percent) playing a more limited role. By the
ly in the WTO negotiations (figure 3). Assistance end of this decade, it is estimated that private
to help bring these countries into the system capital flows will finance a much greater share
should be a priority for the donor community. of the investments in the sector (55 percent),
with internal ( O percent) and official sources
The road ahead (5 percent) reducing their relative participation.
Analyses of the importance of the basic tele- But private investors wxill be willing to invest
communications agreement tend to cluster in modernizing the teleconmmunications infra-
around two extreme positions. Most analysts structure of developing countries only if they
have been extremnely enthusiastic and present can count on fair and stable rules of the game.
the agreement as delivering swx ift liberalization Accordingly, developing countries able to sig-
of participating markets. Others are more skep- nal their commitment to liberalization and to
tical, pointing out that the multilateral regime adopt a procompetitive regulatory environment
and the regulatory authorities are being over- will he in a better position to attract the capital
taken by the velocity of technological change flows required for these investments. WTO
in the industry. According to these skeptics, commitments can play an important part in this.
the agreement plays at best a secondary role
in this process of change. Benefits for developing countries are not lim-
ited to attracting foreign direct investment. Lib-
Reality is somewhere in between. It is true that eralization will also improve local firms access
fast teclnological change has been the main to efficient telecommunications service provid-
46 Liberalizing Telecommunications and the Role of the World Trade Organization
ers. This will increase their competitiveness-
and thus their ability to explore the dynamism
of international trade in information-intensive
products and services. Last but not least, com-
petition will improve the price-quality mix of
the telecommunications services available to
consumers. The critical remaining issue is the
quality of the implementation of the commit-
ments. Many developing countries are entering
uncharted territory, particularly with respect to
procompetitive regulatory disciplines. Those able
to meet these challenges will be much better
positioned to benefit from the information age."
This section relies on lPrinio Blraga (1996). For fuirther details on the
-.ition.ile for liheralizing services and on the ,Irrhitectore of the
GAIS, see tI'NCTAD and World Bank (1994) and Hoekmaln (1996).
IBasic sen ices coser noice telephony, telex. telegraph, facsimile.
dalta transmission, pr ivate leased circuit ser ices. fixe(d and mttohile
satellite cysterns and sets ites, cellular telephons mobile data ser
'iices. paging. atnd pei sonal commusnic.tion selts iC cV altLie added
ser ices incluide email, oice mail. on line dat.i processing. on line
dat.uihse storige and retries al. and electonic d.itai interctange.
hFor tirtiher details on this project, fhich sas executed hN the
Internaitional Institcite of Cotninunic ations. isit the in/heDs \ c-bsitse
at http: \s'N'. Ntorldh)ank.org/html intfdes inriodlex lhtul.
Antignia and Barbocla, Argentina. AUstralia. Brngladesh. Belize,
Bolisxia. Brazil. Brujiei I)artissalar. Brilgaria. Canadcl. (thile,
Colotbhia, Cote dIsoircr Czech Republic. Doiitinica. Doittinican
Rephhlic. Ecuadiior. El Salt aidor, Etniopean COmntltUnitics and its
Meriher States. Ghana. Gienada. Gciatemiala, Hong Kong. Hcingars,
Iceland, India.z Incdonesia. Israel. Janmaica. Japan, the Repcublic of
Korea. Malaxsia, Matiritiris. Nlexicr, M,orocco. Nes Zealand,
N or.ss, Pitkistan, PapcLi Ness Gcuinea. Perci, Philippines. Tioland.
Romania, Senegil, Singapore, Sri Lbaitka, Six itzerlandc Slosak
Repuhlic, SoLitih Africa. Thailand. Trinidad and Tolvigr. 'i Tinisia.
Turkev. tJnited States, and Nenezecela.
Foir a clisctisssion of the rescilts of the GATS in pronitoting sel ices
tiber.lization. see Hoekm.n and Pirinmo Braga (1996).
References
Hoekinan. Bernird. 1996. "Assessing the General Agreement on Trade
in Ser ices." In Will Martin and L. Alan Winters. eds.. The tlugitorl,
Ronit ii aotd the Deetlopintg Coii nries. Cambridge: ('ambridge t'ni-
s ersit Pi- ess.
Hoekinan. Bernard. and C.A. Primo Braga. 1996. 'Irade in Sernices.:
The GATS and Asi.` Asia Pac'u/ic L-cotwntittic Ret iei' 2 (Apt il ): 5-20.
Inrino Braga C. A. 1996. "The Impact of the Intetnation.alization of
Set' ices on Des eloping Cointries.'' Fi'naoce no!d Detvelopment(
(March): 34-37
UNCTAD and WXorld Bank, 1994. Liber/alizing IJnterniationial Trans-
actionts itt Ser ics e: A Hatndbook. Ness York: I Inite(i Nations.
Carlos A. PRino Braga (cbraga@n orldbank.org)
Principal Economist Telecommunications and
InJormatics Division
47
What the Transformation of Telecom Markets
Means for Regulation
Peter Smith
In most countries, telecommunications regulators no longer regulate a static, monopolistic
industry that provides essentially a single product, telephone service, but a dynamic, multi-
product, multioperator industry. In this respect, the telecommunications regulator is way ahead
of its peers in other utility sectors in moving from a monopolistic to a competitive market. This
environment is a fast-changing and increasingly complex one in which regulators face reduced
scope for discretionary decisions. This Note explores the implications for the regulatory agenda.
The transformation of telecommunications lular mobile telephony is now a substitute
markets is occurring in several dimensionis- for conventional local telephone service for
in the changing structure of demand, in the many customers; the distinction between lo-
convergence of services, and in the changing cal and long-clistance calling or, with the
structure of the industry. The transformation is pending introduction of global personal nio-
driven mainly by technological developments.
But competitive pressures resulting from the long
globalization of the world economy and the
ideology and results of reform policies in the distance calling or, with thepending
sector are also important forces.
� Changing demand structure. Only ten years i ' " ' "
ago, conventional 'fixed" voice telephony v onale
dominated the revenues of all telephone
companies. Today. unprecedented growth in stellite service, between domestic and
demand for nevw= services-facsimile. mobile i
telephony, and Internet-is fundamentally interntional senke iS becoming less a) d
changing the overall structure of demand in less relevant
the sector. For example, in Thailand recently,
24 percent of all telephones wvere mobile. The
exponential growth in the number of Internet bile satellite service, between domestic and
servers and users, and consequently in the international service is becoming less and less
demand for bandwidth to carry graphics-rich relevant; and paging and cellular telephony
data files, is intensifying the demanld for na- are nov sometimes bundled as a single ser-
tional and international transmission links. vice delivered through the same handset.
� Convergence of services. Convergence is oc- * Changing industry structure. There has been
curring not only between telecommunica- a fundamental shift in the industry structure
tions, broadcasting, cable television, and the in many countries toward a multioperator
Internet, but also within segments of the tele- environment. Several factors are driving this
communications market. For example, cel- shift. New operators are entering the market
48 What the Transformation of Telecom Markets Means for Regulation
from other utility sectors (in the United setting the terms of entry-and thus creating
Kingdom, for example, electric utilities and market forces.
cable TV companies both provide telephony Adopting processes for the award of licenses
services). Service suppliers are going inter- to service providers. (These may include bid-
national as the era of national monopolies ding processes in which the evaluation cri-
passes. And the resale of network services is teria are clear and easily measured, as in price
becoming an increasingly important business bids, or "beauty contests," in which the bid
as separating network ownership from ser- evaluation criteria are subjective and the se-
vice delivery becomes operationally and lection process is less transparent.)
commercially viable. * Resolving network interconnection issues and
managing numbering plans to promote the
The future of telecommunications emergence of a multioperator environment.
regulation * Authorizing rate rebalancing (whereby prices
are moved closer to costs by reducing prices
These trends in telecommunications markets for international and long-distance services
mean that regulators will operate in a rapidly and increasing them for local and network
changing environment characterized by in- access service) in order to reduce economic
creased complexity, reduced scope for discre- rents and cross-subsidies.
Applying new approaches to cross-subsidies,
The resale of network services is becoming such as improved targeting of beneficiaries,
bidding for minimum subsidies, and the ad-
an increasingly important business as ministration of subsidies in a way that does
not favor one operator over another.
s,eparating network ownershipfrom
separatig networ n hp fromA task of new signifigance is spectrum manage-
service delivery becomes operationally ment. The wireless revolution, reflected in the
rapid growth of cellular telephony, the increasing
and commercially viable. significance of wireless local loop systems, and
the planned deployment of several new-gen-
eration global personal mobile satellite systems,
tionary decisions, increasing privatization of demands that regulators respond to the increased
some aspects of regulation, and the conver- need to manage radio spectrum. Typically, this
gence of regulation of different sectors and of task involves allocating portions of the radio
regulation within the sector. spectrum to different uses, assigning frequen-
cies and authorizing transmission power levels
Managing the transition to transmitters at specified locations, maintain-
ing standards to ensure that transmitters make
In contrast to such utilities as power and wa- optimum use of the radio spectrum, and imple-
ter, telecommunications is now clearly a menting measures to control unauthorized use.
multiproduct sector with several alternative ser-
vice delivery mechanisms that permit compe- While the emphasis shifts to managing the tran-
tition in service provision. Thus, the regulatory sition to competition, the fundamental reasons
agenda has shifted from minimizing the price for regulation of the sector-the need to pro-
of subscribing to local telephone service or tect customers from potential monopoly abuses
maintaining cross-subsidy to managing multiple and to allocate the scarce radio spectrum-have
issues related to competition, entry, pricing, not gone away. Voice telephony remains widely
and cross-subsidies: regarded as an essential public service, and the
* Determining whether entry in different mar- sector is still a potential monopoly in which
ket segments should be limited or open and operators could adopt strategic behavior with
The World Bank Group 49
respect to network interconnection, numbering tion with a major operator on nondiscrimina-
plans, allocation of radio spectrum, and the use tory terms, and allocating scarce resources such
of cross-subsidies. Regulators are still gate- as radio frequencies and telephone numbers.'
keepers of the transformation of the telecom-
munications market-they influence the speed, Privatizing regulation
conditions, and areas of change, and they arbi-
trate conflicts that arise between winners and With the increased complexity, the option of
losers in that change. Consequently. regulators privatizing some aspects of telecommunications
often must keep an eye on politically accept- regulation is increasingly attractive. Two main
able limits to change (for example, to the struc- approaches are possible. One is to create pri-
ture of cross-subsidies in the sector) wxhile vate property rights over the radio spectrum,
steering a course toward regulatory reform. which has been implemented to some extent in
the United States and other countries through
Increased complexity and reduced discretion radio spectrum auctions. Once such property
rights are established, the new owners of the
The transformation of telecommunications mar- spectrum tnay wish to take an increased role in
kets has made the job of regulating the sector sublicensing to other users and in policing the
much more complex. At the same time, it has use of the spectrum. Establishing property rights
reduced the scope for discretionary decisions.
Market transformation is reducing the scopc for Witb the inicr-eased conmplexity the
regulators to maintain cross-subsidy, for -J
example, as a result of the convergence of ser- ohtiou ofpritizi rig some aspects
vices within the sector. Traditionally, the clear 17e Of
segmentation of the market enabled regulators telecomn'umications regulation is
to treat different categories of customers and
service providers differently, influencing the iticreasingly altracti2ve
profitability of services and the flow of cross-
subsidies. Thus, mobile telephone service prices
were typically unregulated, wxhile fixed tele- increases the commercial value of the spectrum
phone service prices were regulated. And mo- and tlhus provides incentives for more efficient
nopoly international telephone service prices use. An important extension of this approach,
could be set very high in order to generate a proposed by advisers to the government of El
pool of funds for cross-subsidy. But the conver- Salvador in 1996, calls for creating rights over
gence of services-reflected in the increasing designated commercial radio spectrum bands
substitutability of mobile and fixed services, the not just for specific uses, but for any use. This
increasing ease with which high-priced interna- gives owners of designated commercial bands
tional telephone service can be bypassed an incentive to assign frequencv to the most
through private networks, the introduction of profitable (or highest-value) use.
callback services, and the pending introduction
of global personal mobile satellite services-cre- Creating property rights thus substitutes a mar-
ates pressures to reduce differential regulatory ket process for the government role in assign-
treatment and to push prices closer to costs. ing radio spectrum for specific uses such as
broadcasting, cellular telephony, or private tele-
In addition. the involvement of the World Trade communications networks. The government
Organization (NTO) in setting rules for regulat- role could he limited to managing the initial
ing basic telecommunications services further sale of spectrum, ensuring compliance with in-
reduces regulators' scope for discretionary ternational agreements on spectrum use, and
decisions relating to. for example, preventing ensuring that ownership of the radio spectrum
anticompetitive practices, providing interconnec- is not monopolized.
50 What the Transformation of Telecom Markets Means for Regulation
The second approach for privatizing regula- of procompetition policies in telecommunica-
tion is outsourcing. While regulatory authority tions is sector-specific (or in some cases, spe-
would remain with a government agency, many cific to network industries) in important ways,
functions could be contracted out, such as the policy itself is essentially competition policy.
auditing the performance of operators, pre- This pressure for regulatory convergence is an
paring public consultation documents, or outcome not only of the changing technology
implementing alternative dispute resolution and market structures, but also of the increas-
mechanisms. The multioperator environment ing role of international agreements on tele-
emerging in most countries promises a heavy communications regulation. And as a result,
workload for regulators in adjudicating billing, telecommunications regulatory agencies will in-
numbering plan, and interconnection issues. creasingly become specialized competition
Alternative dispute resolution mechanisms and policy agencies.
other forms of outsourcing are important
options for reducing that workload as well as Fourth, the high level of insularity or compart-
the budgetary burden on telecommunications mentalization that has been possible at the na-
regulators. tional and international level as well as the
sectoral level is being eroded. Until recently,
Convergence of regulation for example, France could have a completely
different regulatory approach than the United
Multisector public utility boards have been around Kingdom. But the recent completion of the
for many years in the United States. These utility WTO agreements setting out commitments for
boards often have a mandate over telecommuni- regulating basic telecommunications services
cations, natural gas, and electric power supply. is a step toward international harmonization
And in some jurisdictions, communications regu- of regulation in the sector. And in the Euro-
lators have a mandate over transport or broad- pean Union, the application of European
casting as well as telecommunications. competition policy has played a key role in
liberalizing basic telecommunications. These
Now, new pressures for convergence in regu- recent EU and WTO initiatives, though not com-
lation are arising from four main sources. First, prehensive, are important steps in harmoniz-
the overlap between regulation of carriage (tele- ing national approaches to telecommunications
communications) and regulation of content regulation.
(broadcasting) will increase as both telephone
companies and cable TV operators begin to Conclusion
provide services previously provided only by
the other and as the Internet's capability to Regulation is profoundly changing the telecom-
deliver video improves. munications sector. But change in the sector is
also driving the agenda for regulation. In time,
Second, the substitutability of services across it is not inconceivable that telecommunications
subsectors or market segments, particularly regulatory agencies will eventually disappear,
between telecommunications and cable TV, absorbed into multisector antitrust agencies.
broadcasting, satellite broadcasting, or Inter-
net, also creates pressures for harmonizing SeeCar1osA,PrinoBrg, Ller.ilizinglTelcommuniAtiorsand
regulation across communications subsectors. tfhc Role of the World Trade Organization' (page 39 in this issue)
Third, the critical issues that are emerging in PeterSmith 6psmith2fworldbank.org), Principal
telecommunications relate to promoting com- Telecommuinications Policy Specialist, Telecom-
petition: interconnection arrangements, revenue munications and Informatics Division
settlement, numbering plans, number portabil-
ity, and the like. Although the implementation
51
The Private Sector and the Internet
Carlos A. Prniom Braga mul Carsten Fink?
Advances in telecommunications and infor- archiitecture typically used for voice telephony.
matics have transformed the Internet from an it does not require an opcn, or point-to-point,
academic experiment into a householdl n ame connection. Packet-switching allows many users
in most industrial countries. The number of to share a circuit, with no particular connection
computers connected to the Internet grew from dedicated for a given communication session.
535,000 in Tuly 1991 to close to 16 million by MIoreov7er it increases the network's reliability,
January 1997 (figure 1), and it is estimated that allowing it to operate even under catastrophic
the numlnber of Internet users has already ex- conditions (for example, amid a nuclear war).
ceeded 50 million. Althoughi still concentrated
in industrial countries, this network of net- The resulting network, Arpanet. began opera-
works'" is rapidly expanding in the developing tion in 1969. linking four sites. In the 1970s.
world. Many now be]ieve that it provides a win- other governme nt-supported networks emerged
dow into a future in vwhich access to informa- in the tUnited States, but access remained
tion wvill he independlent of geographic location restricted to the research community connected
and interactivity in a multimedia environment with the Department of Defense and other
will be ubiquitous.
This Note briefly reviews the history of the In-
ternet and its evolution from an academic ex-
perimcnt into the main application behind the IGUre 1 GRoW T E
emerging global inform ation infrastructure. It I t
16
discusses the role of the private sector in these 14
developments ancl the regulatory envTironment A Domain names
required for the Internet to fulfill its proimiise. 12 Other
It concludes with some lessons for expancling 10 edu
the Internet in developing countries. 8 com
6 neMot
The rise of the Internet 4
2
The origins of the Internet can he tracecl to the o-
1960s, when the U.S. Departmcnt of Defense Jul Jul Jul Jul Jul Jul Jan
decided to fund the development of a packet- 91 92 93 94 95 96 97
switching data network that wvould allow net-
w c r s a Ilfote:An Internet host is a computer that acts as an information and communications
nTorlSel coptr ltpCs server and has a direct connectionto the Internet. Top-level domain names identify
to comiimunicate efficiently. In packet-switching hosts by type or origin-for example, .com Icommercial) .edu (educational), and .net
netwrorkis cdata files are br-oken into small pack- (network), the three most frequently used. Others include .org (organization, .gov (U.S.
government), and country domains such as .uk (United Kingdom) and de (Germany).
ages that are sent independently over the But top-level domain names can be misleading: a host might not be located in the
network and then reassembled at the final desti- country indicated or might be either a public or a private institution.
nation. This permits efficient use of commutnica- Source:NetworkWizards(http://wwnwcom).
tions lines because, unlike circuit-switching, an
52 The Private Sector and the Internet
government agencies. This changed in 1986 with been the rapidly evolving network architecture
the creation of the NSFNet, also subsidized by and user interfaces. This technology has ben-
the U.S. government. The idea of the NSFNet efited from the decline in computing costs rela-
was to provide high-speed backbone services tive to transmission costs. On the user side, the
connecting regional networks as well as cam- growth of the Internet has been promoted by
puses and research centers. The network of net- the appearance of powerful programming lan-
works communicating through the Internet guages, new network "tools," and user-friendly
protocol began to expand rapidly. interfaces. The World Wide Web, a sophisticated
application that allows users to access any kind
In the early 1990s, the management of the of digitized information (text, picture, sound,
NSFNet backbone was subcontracted to private vileo) and configure it for display with a mouse
firms, which were allowed to route commer- click, has given multimedia capabilities to the
cial traffic through the Internet. The explosive Internet. The growth of the Web has been as-
demand for network service in the 1990s- tounding: betweenJune 1993 andJanuary 1997,
mainly from the private sector-led to the the number of Websites leapt from 130 to
emergence of several commercial Internet back- roughly 200,000. Fostered by the improving
bone networks (such as Alternet, PSNet, and multimedia capabilities, commercial use of the
SprintLink), and in October 1995, the NSFNet Internet overtook research and educational use
backbone was shut down. U.S. government and has been growing exponentially in the 1990s
subsidies for the Internet have fallen to an in- (see figure 1). ByJanuary 1997, there were close
significant amount, and almost all the costs of to 4 million hosts in the .com domain.' The pri-
the Internet are now borne by its users. vate sector has clearly taken the driver's seat in
providing both the Internet's infrastructure and
Supporting the explosive growth of the Internet its content in the United States.
and of the demand for Internet services have
Regulation
The Internet has blossomed in a relatively
regulation-free environment. Most regulatory
FIGURE 2 MONOPOLY VERSUS COMPETITION- activity has concentrated on defining standards
AND THE GROWTH OF THE INTERNET for the formats and protocols necessary to
operate the network. But as the commercial
Inteet hostsper1,000inhabitants presence on the Net increases, regulatory is-
sues relating to the provision of the network's
14 1 OECD markets with infrastructure monopoly infrastructure and services become increasingly
1Z * OECD average important. The development of a regulatory
12 *OECDmarkets withinfrastructur competition framework is critical in three areas: provision
10 of Internet backbone access; Internet service
8 - : - providers (ISPs); and information services.
6
4 | * - The Internet backbone servers are the highest-
level network servers-those to which ISPs pay
2 connection charges. The basic regulatory options
0 are to provide public support for backbone
Jul 91 Jul92 Jul 93 Jul:94 Jul 95 Jan 96 access to promote connectivity or to leave back-
Soure: OECD, "nformation infrastructure Convergence and Pricing: Tme Internet bone service provision to the market. As men-
(Paris, 1996). tioned, the original backbone in the United
States, NSFNet, was government-funded until
rapid growth in networking demand led to the
The World Bank Group 53
emergence of commercial backbones. A similar
pattern can be found in other industrial coun-
tries. In Germany, for example, the first Internet BOX 1 INTELLECTUAL PROPERTY IN CYBERSPACE
backbone-UNIDO, for Universitat Dortmund-
was run by the university and later replaced by The rise of the Internet gives new relevance to the issue of extra-
commercial backbones.
territorialiy and increases the demand for convergence among
ISPs provide Internet services to the end tisers. national intellectual property rights regimes. The Intemet not only
In the United States, ISPs are competing private opens new possibilities for dissemination of information; it also
firms. In other countries, the major ISP is the expands the scope for activities that may infringe on someone's
state-owned telecommunications operator. of- intellectual property rights. With a few keystrokes, an Intemet user
ten a monopoly. Regulatory options for ISPs can anonymously download copyrighted material in bulletin boards
depend on the market structure in telecommu- around the world. Prosecuting Internet service providers can
nications. In many cases, the telecommunica-
nicationsoperators mare wsell pitined tcom vide discourage infringement, but it may inhibit the expansion of the value
tions operators are wTell positioned to provide
Internet services. But it is worth pointing out addedservicesthatmakethelnternetsopowerful.
that OECD countries with more competitive tele- The most important international treaty on copyright protection is
communications sectors tend to have greater theBemeConventionfortheProtectionofLiteraryandArtisticWorks
Internet connectivity than countries with a mo- of 1886, which provides for national treatment of domestic and
nopoly (figure 2). foreign copyright holders and sets minimum standards for copyright
Policymakers have to decide w,vhether telecom- protection. The World Intellectual Property Organization (WlPO)-a
munications network operators should be per- specialized United Nations agency that administers the Brne
mitted to offer information services in direct Convention-held a diplomatic conference in December 1996to
competition with independent information ser- revise the convention and to clarify the scope of copyright protection
vice providers. As a rule of thumib, if a telecom- in the digital environment. The main outcome of this conference was
munications operator has market power in the the WIPO Copyright Treaty. This treaty makes clear that the reproduc-
transport network, structural or at least account-
iangseparatione hor, .uldub r to avoi nt- tion rights of copyright owners encompass the right to make digital
ing separation should be required to avoid anti-
competitive cross-subsidization. In other words, copies. But its language is broad enough to allow national legislation
the Internet services unit of the operator should to limit (or remove) liability at the level of network providers with
be required to buy access to the transport net- respect to, for example, temporary digital storage. The treaty thus
work on an arm's-length" basis. Another, very achieves a balance between the concerns of content providers and
sensitive issue is voice telephony over the In- those of contentcarriers.
ternet. which may become a serious threat to
the traditional circuit-switched network.
Note: For discussion of the economics of intellectual propert riglas in cyberspace. see
Appropriability of contentGCA Primo Braa and C. Fink, "The Economic Justification for the Grant of Intellectual
Property Rights: Patems of Convergence and Coafficf," in E Abbt and D. Gerber. ads.
Public Policy and Global Technological lnregrdtoo (London: Kluwer, lU9117.
Digitized information can be easily reproduced
and redistributed on the Internet, and providers
of information find it difficult to charge users
directly. Most private content providers recover
costs indirectly-by providing information to But the conventional remedy for the cost recov-
potential customers about other goods and ser- ery problem is intellectual property rights pro-
vices. The indirect incentive structure is strong tection. Copyright, for example, protects an
given the low cost of disseminating information author's work-whether a book, a performance,
on the Internet relative to the number of users a recording, or a computer program-from unli-
and the bright future prospects of the Internet. censed copying. In principle, traditional copyright
54 The Private Sector and the Internet
law applies to the Internet environment. But such provide digital protection across national
major industrial economies as the United States boundaries. Incrcasingly sophisticated digital
and those in the European Union have revised rights management technologies are becoming
or are now revising their intellectual property available on the World Wide Web. But national
rights laws to address specific needs of electronic security reasons have been invoked to limit
networks. Moreover, the global character of the the dissemination of cryptographic capabilities.
Internet demands international legal governance The United States has been trying to address
(box 1). the national security issue by promoting data
encryption standards that can be broken by
An important problem in legal protection for intelligence agencies and by controlling the ex-
copyright holders on the Internet is enforce- port of encryption technology, restrictive mea-
ment, given the speed and magnitude of data sures that may inhibit widespread commercial
transmission. This is an area where digital rights use of these technologies.
management technologies can be of help.
These hardware and software devices control A third way to recover costs is through the sale
access to information and the ability to use and of advertising space on information pages.
further distribute it. In principle, these "encryp- Commercial advertisements first appeared on
tion" technologies are attractive because, un- the Web in 1994. Although this step toward
like intellectual property rights, they can pure commercial use of the Internet initially
FIGURE 3 GLOBAL INTERNET DENSITY, 1996
Internet hosts per
10,000 inhabitants
* More than 100
E 1-10-l
E Less than 1
The boundaries, colors, denominations, and any other information shown on this map do not imply, on the part of the World Bank Group, any judgment on the legal
status of any territory or any endorsement or acceptance of such boundaries.
Source: Network Wizards (http:/www.nw.conVzone/WWW/top.html).
The World Bank Group 55
met with strong resistance from the researclh telecommunications operator. Recognizing the
and education communities, growth has been economic benefits of electronic networking, the
rapid, and advertising revenues are estimated Brazilian government supports an Internet
to have exceeded US$250 million in 1996. backbone open to commercial connectivity and
traffic while limiting the dominant carrier's ac-
Internationalization tivity in the direct provision of Internet ser-
vices to the public. The number of Internet
Although the Internet is still most widely used
in the United States, the 1990s have been
marked by its rapid internationalizationi. The By the niid-1990s, Br-azil had a hiuher
share of non-U.S. hosts increased from 20 per-
cent in July 1991 to 36 percent in July 1996. ratio of Internet hosts to PCs than such
But most non-U.S. hosts still reside in indus-
trial countries, and in July 1996, roughly 96 econo/zies as France, Germany, Hong
percent of all Internet hosts were in OECD
countries. Kon?7g, and Sitngapore.
Most developing countries are connected to the
Internet-if only through email-though pen- hosts in Brazil (br domain) grew from 300 in
etration is still low7 (figure 3). In 1996. there January 1992 to more than 50,000 in July 1996,
was on average only 0.5 Internet host per of which some 20.000 are commercial ('.com.br
10,000 inhabitants in developing countries, domain). By the niid-1990s, Brazil had a higher
compared with 101 in industrial countries. The ratio of Internet hosts to PCs than such econo-
low pcnctration is due mainly to the poor in- llies as Francc, Germany, Hong Kong, and
formation infrastructure in developing coun- Singapore.
tries-the roads and ports that carry and process
digitized information. Average teledensity (tele- There are, of course, many obstacles to the
phone lines per person) is thirteen times lower, diffusion of the Internet in developing coun-
and average PC density thirty-eight times lowTer. tries. National laws regarding privacy and in-
than in industrial countries. tellectual property rights protection must be
refined. And the predominance of English-
For developing countries, establishing the right language content may deter local researchers
regulatory environment is as critical as it is for or local firms that could use the Internet to
industrial countries-though the relevance of add value to their goods and services.
the regulatory experience of industrial econo-
mies discussed above is open to debate for coun- The critical bottleneck, however, continues to
tries with poor telecommunications networks, be the weak information infiastructure of de-
low computer penetration, and incfficient, state- vcloping countries. Government activism to
owned telecommunications monopolies. But a promote Internet connectivity at the level of
few developing countries have managed to rap- the research and education communities may
idly expand Internet connectivity despite weeak help jump-start the national information infra-
information infrastructure, such as Brazil, Chile, structure. And governments should support
the Czech Republic, Malaysia, Mexico, and community access in public libraries and com-
South Africa. munity centers. But an increasingly important
role for governments in fostering the Internet
Brazil, for example, successfully adopted a revolution is that in the regulatory arena. Most
model of public-private partnership to diffuse important here is to promote a competitive en-
the Internet, and it has increased Internet vironment for Internet service providers, es-
connectivity despite its dominant state-ovwned tablish adequate rules of the game for electronic
56 The Private Sector and the Internet
commerce, and ensure effective incentives for
the provision of content-essential measures
for attracting private investment in the infra-
structure and in content generation. Those
countries able to attract such investment will
be better positioned to benefit from the emerg-
ing global information infrastructure.
This understates the nuLmber of comniercial hosts, since the net
domain (with more than 1.5 million hosts) and some of the country
domains also include commercial hosts.
Carlos A. Primo Braga (cbraga@worldbank.
org, Principal Economist, and Carsten Fink,
Consultant, Telecommunications and
Informatics Division