1
25093
May 2002
Water and
Sanitation                           Vietnam
Program
An International      Evolving Management Models
partnership to help
the poor gain sustained  for Sma        Towns Water Supply
access to improved
water supply and
satnitationservices       in a Transitional Economy
Water and Sanitation Program                   J
for East Asia and the Pacific
J.~~~~~~~~~~~~~~~~~I
Prepared   by:                                                     .................
Caroline van den Berg
ADCOM pF
--~~~~~~~~~~~~~~J
May 2002          - 



ACKNOWLEDGMENTS
This report is based on site reports and community data records prepared by researchers from ADCOM.
Field team in Vietnam: Hoa Thi Hoang, Phung Xuan Bihn, Le Thi Thia, Bui Van Kiem, Nguyen Chau Loan,
Le Xuan Thuy, Nguyen Tranh Phuc.
Hoa Thi Hoang, Nguyen Cong Thanh, and Richard Pollard supervised the work in Vietnam. Technical
assistance for the mnethodology, training and documentation was provided from WSP-EAP's regional
office, and by Hydroconseil, the consulting group that is managing the global initiative for Small Town
Water Supply and Sanitation that was launched by the World Bank Thematic Group on Rural Water
Supply and Sanitation and the Water and Sanitation Program in 1 999.
Ms. Ann Thomas was instrumental in the drafting of this publication.
Peer reviewers: Nicholas Pilgrim and Richard Pollard provided valuable comments and questions for further
consideration.



Vietnam:
Evolving Management
Models for Small Towns
Water Supply in
a Transitional Economy
.
Contents
Executive Summary    .....................2
I Introduction ............4.........4
11 Methodology ..............6.......6
III Performance of
Management Models ......................9
IV Operational Performance ............. 1 1
V Financial Performance .................. 15
VI Social Performance ..................... 19
I    Vil Institutional Performance ............... 21
VIII Conclusions and
Recommendations ................   24
l ~~~~~~~~~~~~~~~~~~A               nnex   1  .............................. ..  28
t       i      '   *  '  §\
S*=SU____



Executive Summary
Access to water supply services in Vietnam, not to mention
sanitation services, is still rudimentary in small towns and
even more so in townlets and communes. According tow
the Ministry of Construction only 30 percent of small towns
have piped water systems. It is estimated that a mere 15
percent of the townlets have piped water systems. Most
piped water systems cover only a fraction of the                             III"
populations in small towns and townlets; it is estimated e77
that the connection rate ranges from 20 to 80 percent,
suggesting that only a very small part of the total     '
population in small towns and townlets is covered by
piped water systems.
It is not obvious from the research undertaken that any
one of the small towns water supply management models
is outperforming any of the others. Some management
models do better in some aspects than others, but this is
against the background of a sample that has flaws and
with a dataset that suffers from inconsistencies, especially
with regard to the financial data collected. What emerges
from  the data - with all these caveats - is a              _     i
2



multidimensional picture - looking into different aspects  business and proper tariff levels are better indicators
of sustainability results in different conclusions.  for success than are the management models per se.
Delivery of water supply services by companies that
The key discussion points are:                       are not exclusively concerned with water supply may
be less effective than those that are. In this review,
e Is there a 'best practice management model'? For   systems managed by communities, cooperatives and
Vietnam, due    to  varying  geography   and      private operators have much higher scores than
demographics, and a rapidly evolving economic policy  systems managed under other models.
and regulatory environment, the use of a variety of
management models is optimal. The use of more   e Economies of scale. Water supply systems in small
customer-oriented approaches results in better overall  towns are not necessarily performing better than
performance, which supports evidence elsewhere in  townlets. Level of demand measured in terms of water
the world that demand-responsive approaches have  sold is consistently a better indicator and guideline
a positive impact on systems' sustainability. Exclusive  than population size for determining success of a
focus on technical and financial efficiency does not  sustainable water supply system.
necessarily result in better service delivery. The
performancewithin categories of management models  e Serving the poor. Connection costs are a major
varies widely. The conclusion thus seems justified that  obstacle to achieving greater coverage of water supply
the more important aspect of management is not so  services. Cross subsidies may be useful in some cases
much the organizational model, but what "rules of  but more research is needed to understand how the
the game" are being applied. Some "rules of the   poor can benefit from the different types of cross
games" like autonomy in managing the water supply  subsidies currently in place.
3



1. Introduction
In recent years, developing countries have focused most            "f
of their new water supply and sanitation investments on L
either urban or rural areas. Small towns often fall between                                  M|
these two settlement types in many ways - institutionally,                                      - 
financially, and legally. According to estimates from the  ,o 
Ministry of Construction 30 percent of small towns have
piped water, and only 15 percent of townlets have access
to that service. In most areas, the piped water service
extends onlyto a fraction ofthe population living in small                   |
towns and townlets, making the actual access to services
significantly lower than these data reflect. In contrast,                         .5
access to safe water in urban areas is 61 percent. In
rural communities an estimated 30 percent of residents
have access to water that meets basic domestic
requirements but only about 10 percent have access to  community management. TheVietnam Small Towns case
water that meets national quality standards for drinking  study identified, analyzed and highlighted problems,
water. This tendency to orient investment toward large  trends and opportunities in evaluating the performance
cities obviously has left the segment of small towns  of different small towns water supply management
seriously neglected in terms of access to water supply  models by looking into the institutional, financial, social,
services.                                         and technical performance of these systems.
Small towns do not completely fit within either the urban  In Vietnam, small towns are known as either small towns
or rural context. They are often considered too small to  (thi tran) or townlets (thi tu). Urban areas are classified
be managed institutionally, and too big for effective  according to 5 categories' . The total urban population
I For more detailed description, see Annex 7
4



is estimated at about 19 million people (24 percent of             Vietnamese administrative levels
thetotal population). Small towns are classified as "class
5 urban areas". They comprise more than 5 million
inhabitants (or about 7 percent of the total population).
These "class 5 urban areas" are defined as settlements
with:                                                                          National
e Population size ranges from      4,000 to 30,000                                 i
inhabitants (2,000 inhabitants in mountainous areas);             61 provinces (includes 4 cities)
e Sixty percent of the labor force is not engaged in                               i
agricultural activities;                                                         i
e Construction of public facilities and technical
infrastructure in early stages;                                490 district towns 40 small towns,
e Average population density of 6,000 inhabitants/km2                   and 8,850 Communes
(3,000 inhabitants/km2 in mountainous areas).                            (3000 Townlets)
A small town is an administrative unit that is equal to the
commune level of administration and authority, the lowest
level in the Vietnam administrative system. Small Towns
Source: Statistical Pub House Socia-Economic Stalistical
are under the jurisdiction of the District People's              Data of 61 Provinces and Cities in Vietnam (1999)
Committee (DPC). Most of them are district administrative
centers and have very limited autonomy for investment
management2. The Small Town People's Committee
(STPC) submits project proposals to the Provincial People's  e Minimum population size of 2,000 inhabitants (1,000
Committee (PPC) for approval and financial support. The    inhabitants in mountainous areas);
STPC can be the water supply project owner and           * At least forty percent of the labor force is engaged in
supervisor for the construction, operation, and            non-agricultural activities;
maintenance of water systems in their town.              a Initial construction of main public services and technical
infrastructure3;
The Government does not classify townlets as             a Average population density of 3,000 inhabitants/km2
administrative units, but considers them as residential    (1,000 inhabitants/km2 in mountainous areas).
areas under the management of the Commune People's
Committees (CPC). The townlets are typically "commune    No single organization is responsible for managing the
centers", the largest settlement within a commune. There  implementation and coordination of water supply services
are an estimated 3,000 townlets nationwide. The total    in small towns and townlets. Small towns fall under the
population residing in townlets is estimated at about 10  mandate or jurisdiction of the Ministry of Construction,
million (or 15 percent ofthetotal population). Temporary  as do water suplly services in all larger urban areas.
standards for planning and approving construction in     Townlets fall under the Ministry of Agriculture and Rural
townlets are:                                            Development, in which CERWASS is the lead agency.
2  Investment management depends on each province's decentralization policy
3  Examples of public services and technical infrastructure are transport, post office, water supply, sewerage and drainage, power supply and facilities for
daily services such as markets, shops, health care centers, schools, small industrial production units, sports and entertainment services, and cultural and
information facilities
5



II. Methodology
Vietnam is in a dynamic period of economic growth and
liberalization. Regulations are changing, and new
opportunities are opening (and others closing). That is
one reason why there is what on the surface appears to
be an odd mix of old and new management models.
Most of the new water supply systems are in management
arrangements that have only recently become possible
with changes in regulations and decentralization. For
example, the Enterprise Law made it possible for the
private sector to enter the market and compete against
state-owned enterprises (SOEs). That is one reason for
the experimentation bythe private sector in water supply,  i 
and SOEs attempting to become more competitive.
The research team investigated the water supply and
sanitation situation in 22 small towns and townlets.
A variety of management models was looked into:
(i) direct management by Small Town People's               '1
Committee (STPC) in small towns and Commune                         4
People's Committee (CPC) in townlets;        .x 
(6i) community management;
(iii ) cooperatives;
(iv) provincial water supply companies (PWSCs). This is
6



a special form of state-owned enterprises. They are   1. Variation in models between smoll towns and towniets:
responsible for providing water supply services to      In the sample, small towns mostly resorted to PWSCs,
provincial capitals (Class IV towns), but recently have  other SOEs or private water companies, while in the
seen their mandate extended to also provide their        7 townlets (one of which is served by 3 water supply
services to small towns (Class V towns);                 systems4), the variation in management models is
(v) other state-owned enterprises. In the sample, they       substantially larger -with cooperatives, communities,
mainly consist of district water supply companies and/   private operators and STPC/CPC managing water
or environmental service companies; and                  supply systems.
(vi) private water companies.
Figure 1 Selected provinces for Small Towns Study
In addition to these models, the research team identified          and their location within Vietnam
two other management models, which are much less
common. The first is management by the Provincial
Center for Rural Water Supply and Environment Sanitation
(PCERWASS). PCERWASS coordinates the management                                                 Northern Uplands
of water supply systems in townlets and in all rural
communes. Another model refers to the District People's                          -                Red Rver Delta
Committee (DPC) coordinating water supply through
district administrative units such as the District Economic
Office, the Division of Public Works and Transport, or
directly managing water management boards in the
district.          North Centrol
district.                                                                         \ 
Initially five provinces were selected for the study: three
provinces in the Red River Delta (Nam Dinh, Thai Binh
and Ha Nam) and two provinces in the Mekong River                                                      Central
Delta (Long An and Tien Giang). The study was then               Central Highlands                      Coast
extended to two more provinces to improve the regional
representation of the sample of case studies: Thua Thien              Soutk East
Hue in the North-Central region and Quang Ninh in the
North-East. The selected small towns and townlets are
/    w   j    ~~~Ho Chi Minh City
described in Annex 1.                                                                         Mo            a
(?  *  r'    ~Mekong River Delia
The following criteria were taken into account when
evaluating the performance of the different management
models:
4  In the Mekong River Delta, it is very common that more than one water supply systems exist in one small town or one commune, the district town Beni Luc
In Tien Giang Province has 9 systems Usually, a maoor public system was built in a populated center area and some small systems owned by individuals
or communities were built in surrounding villages. In Le Loi, one of the townlets in the sample, the research team looked into three existing systems, two
monaged by communities and the third by a private operator
7



2. Years of operation: Different systems have been in  common in the Mekong Delta but almost absent
existence for different time periods. The more  in the North East region.
"traditional" management models have been in
operation much longer than the recently approved  4. Exclusion of two small towns. Data analysis excluded
management models (such as cooperatives, private  two small towns, Sia and Mao Khe due to their
water operators and community managed systems).  particular setups. Sia's water supply system is an
Age of the water supply system is likely to have an  extension of two other water supply systems. As it is
impact on performance indicators such as coverage  not an independent water supply system, it provided
and per capita consumption, and subsequently affect  insufficient information on how the system actually
financial and technical performance.            works, and as a result it gave few clues about its
sustainability. The water supply system for Mao Khe
3. Geography. The study was conducted in four      was originally designed to supply a cement factory.
regions which are economically and hydrologically  Because of the importance of non-residential water
diverse. Per capita income and access to        use in overall water use, and the lack of information
alternative water supply sources (rain water, well  on actual non-residential consumption, it is difficult
water, etc.) vary widely among the towns       to judge the performance of this system in anyway
examined. Moreover, cultural differences may be  comparable to that of the other small towns and
reflected in patterns of sharing water, which is quite  townlets studied.
8



Ill. Performance of Management Models
The objective of the study was to determine how the  It is obvious that many of the indicators are interrelated.
different management models affected       the  For example, per capita investments affect tariff levels,
performance of water supply systems. In this respect,  and hence profit margins, and hence the operations of
performance is measured in terms of sustainability.  the system, while they also affect the affordability of the
services, especially to poor consumers. Design standards
Sustainability has a number of different key aspects:  affect technical performance, and also environmental and
social, financial, institutional, technical and  financial performance. Influences to one aspect of
environmental. Environmental aspects of sustainability  performance may impact overall performance and
were also studied but as these data were more   therefore this indicator is not only a sum of individual
qualitative than quantitative, they were not included  features but provides an overall image of the system's
in the final performance analysis. The research team
compiled a set of ten indicators grouped under the  performance.
categories of institutional, financial, operational and  For each indicator, the models were ranked relative to
social sustainability. The ten variables were selected
to, as a whole, represent a comprehensive picture of  each other and scored based on this ranking. The
the sustainable performance of the water supply  rankings in each categorywere converted to a maximum
systems, taking into account the multi-faceted character  score for each indicator. If the data were continuous, a
of sustainability. Six of the indicators refer to on-going  maximum score of 22 could be achieved. In case of non-
performance, the other four relate to the design of the  continuous data, the actual score was lower as several
system and financing rules (government subsidies,  observations would share the same ranking. The
community contributions, investment per capita and  maximum achievable score - in relation to the sample -
design standards).                              was 202.
9



Table I    Performance Indicators
1. Operational   Per capita investment cost                           10        The lower per capita investment costs,
(66)                                                                         the higher the ranking
Access rate: A measure of the number of households that  10     The higher the access rate,
have access to piped water (without necessarily owning          the higher the ranking
the connection) as percentage of the total number of
households in the settlement.
Per capita billed consumption                        10         The higher the per capita
consumption, the higher the ranking
11. Financials   Profit margin:The difference between revenues and    10        The higher the profit margin,
(59)          operating expenses plus depreciation per cubic meter of        the higher the ranking
water distributed. The indicator is not adjusted for the
payment of interest and capital repayments. As such, it may
overestimate the actual profit margin.
Government contribution in investment funding        10         The lower the government subsidies,
the higher the ranking
Community/customer contributions in investment funding  10      The higher the contribution of
customers and/or communities,
the higher the ranking
111. Social -    Ability to Pay (i.e., water bill divided by household income)  10  The lower the ability to pay, the higher
proxy for                                                                    the ranking
customer      Maximum Connection Charges                           10         The higher the connection charges,
satisfaction                                                                 the lower the ranking
(55)          Satisfaction with pressure as indicator for quality of services  10  The higher the satisfaction, the higher
the ranking
IV. Institutional  Adherence to technical standards set by Government:  10      The wider the divergence between
(22)          The maximum standard for supply is 50 lcd for townlets         standards and actual design capacity,
and 120 lcd for small towns. Hence, the difference is          the higher the ranking if the
calculated as actual design capacity minus standard             divergence was negative (i.e. design
dependents on the type of settlement.                           was below government technical
design standards), the ranking was
lower than if the divergence was
positive.
5  In the case of community contributions, for example, there were a number of systems that registered a zero contribution All these observations were given
a similor rank, and hence instead of 22 being the highest ranking, the highest ranking would be lower, because of the shared ranking of a number of
observations
10



IV. Operational Performance
Operational performance was measured using three proxies.  Figure 2 Investment per copita (US$)
(i) per capita investment costs, (ii) coverage, and (iii)
consumption rates.
40
Per Capita Investment Costs. Per capita investment cost
varies widely over the different small towns and townlets,  35
and between manogement models. The most obvious outlier
in terms of per capita investment costs is Tu Ha, where they  30
reached more than $1,000 per capita. The average per
capita investment costs are substantially lower, and if  25                     U
corrections are made for the most obvious outliers, the
average per capita investment cost is lower than $40. Figure  20             I
1 shows that cooperatives and community-managed systems
show the lowest per capita investment costs, and SOEs the  15 
highest.
1 0
More interesting is the relationship between the investment
cost per capita and the different regions - when major    5         -
outliers are excluded. It is interesting to note that the outliers  X                .i.
are concentrated in the North Central Region and to a lesser  0  ,                'h
extent in the North East Region. Per capita investment costs 
are low in the Mekong Delta, slightly higher in the Red River  9C 
Delta and, depending on the inclusion of outliers,
substantially higher in the North East and North Central
regions.



There are many explanations forthese higher costs. The  The tendency to share piped water connections is
North Central and Northeast regions are characterized  regionally very different. In the Mekong Delta and North
by (i) scarcity of water and hence no or few alternative  Central Region, no less than 25 percent of the population
sources of water supply; (ii) low per capita income; (iii)  shares connections, while the number is substantially
price of supplies may be higher due to the high     less in the other regions. As a result in the Mekong Delta,
transportation6 cost as they are located less favorably  access is set at 67 percent, which is much higher than
than the provinces in the Red River and Mekong Deltas;  that in the other regions (with the exception of the North
and (iv) the source of water being used (In the Mekong  East Region)7 where access does not exceed 40 percent.
Delta, almost all systems investigated used ground water,
while the systems in the North almost all relied exclusively  Figure 3 Coverage and Access by Management Model
on surface water. In general, water supply systems using
surface water tend to be more expensive due to the
higher water treatment cost involved).
Coverage and Access to Water Supply Connections.       60%
Coverage and access have distinct meanings. Coverage
refers to those households that are connected to a piped
network, while access includes those households that
not necessarily are connected, but nevertheless use piped
water connections. In general, access exceeds coverage  40%
significantly in Vietnam due to the practice of sharing
connections.                                                             I
Coverage is relatively low in Vietnam. Only 38 percent  30%
of the population in the 22 small towns under review   20%
owns a piped water connection. Even in systems that
have been in existence for many years, coverage is far
from universal. Coverage differs substantially among the  l0%
different management models - with SOEs having the
highest coverage rates and community-managed systems    0%
showing the lowest coverage. However, access to piped       G         0  ' \
water is significantly higher; on average 51 percent of   1o C°°   C0
the population has access to services. For PWSCs, for
instance, sharing is very significant, with average
coverage being 37 percent, while access is no less than
60 percent.
6  The latter could not be verified due to lack of data on the inputs and cost of the maor cost items
7 In the North East Region, coverage is high at 67 percent, which is not the result of sharing that is not a common practice here However, this region is
characterized by a relative lack of alternative water sources.
12



Figure 4 Coverage Rate by Income/Wealth Category
100%
90%
80%
70% _____
60%
50% -
40% _-   i
30%                                                                           _____-____
20% - t-
10%-                                              P                            __
Trung Lao      Que          Nhi Quy       Cai Be      Quang Yen      Khe Tie
Rich U Middle-income * Poor  |
This phenomenon of sharing may be linked to the lack  tradition of sharing connections (sharing connections is
of alternative water supply sources, but it is not likely to  rather atypical in the small towns and townlets in the Red
be the only reason. Even though the North East region  River Delta). It may also be due to a specific pro-poor pricing
has less alternative sources, sharing seems to be a non-  policy in place in this town, which exempts the very poor
existent practice.                                  from paying connections fees thereby facilitating access to
services.
Most households that own a piped water connection are
not poor. The households that are not connected are often  In general, coverage is low but access is significantly higher
poor, with no resources to pay for the connection cost. In  - attributed largely to "sharing". The incidence of sharing
addition, they may live far from the main trunk line, making  connections varies regionally. High access rates are also
the connection cost 2 to 3 times higher than for households  related to wealth distribution and pricing policies favoring
that live near the trunk line.                      the poor.
With the exception of Que, poor households do not own  Billed Water Consumption Water consumption in the
piped water connections. The higher proportion of poor in  PWSCs is much higher than for any other types of
Que owning water connections may be due to the lack of  management model The higher consumption is mainly
13



Figure 5 Consumption Patterns by Management Model     level of consumption of the privately operated systems, which
is mainly due to the very young age of these systems. Two of
140%                                                the three systems are less than 18 months in operation,
while all private operators in the study work in townlets
120%                                                 (instead of small towns)
100%                                                 Water consumption can be measured in different ways.
Using the number of actual users of the connection, the
80%                                                figure is likely to be different than the figure that is related
to the consumption of the households that are connected.
In general, the numbers are affected by (i) the years a
60%                                                 system has been operating - older systems have higher
consumption mainly due to the maturation of the system,
40%   -        :                                    as people's water behavior changes only slowly even if
more water is available; (ii) coverage - per capita
20%             [1                                  consumption decreases as more households connect to
jf 1  11  |     11     11          the system. This confirms the usual pattern of connection,
0%/ !|11  _111 I-1!I. _|1,    _q       -1in which first rich, and then middle-income and only then
0%    40S    \ess  4Ox0    A9,S   C           poor households connect; (iii) tarff structure - block rates
c\    e09°  '/      ,re     (,v   <gwJtend to have a more positive impact on per capita water
consumption than flat rates9; (iv) per capita income - the
higher per capita income, the higher water consumption;
|7n Consumption by HC  * Consumption by users  and (v) type of settlement - small town systems are designed
for 80 to 120 liter per capita per day (Icd), compared to
25 to 50 lcd for townlets and rural areas. Analysis shows
related to the higher per capita income in the small towns  that there is a correlation between per capita income and
served by PWSCs8, the fact that most of these systems are  type of settlement, with small towns having higher per
operating for much longer than any of the other managed  capita incomes leading to higher consumption. The final
systems, and sharing is much more common under these   factor affecting consumption is (vi) supply factors - the
management models than in any of the other management  existence of alternative sources lowers piped water
models, resulting in higher access. Also striking is the low  consumption.
8 Only one PWSC served o towniet, all others were providing services to smoll towns
9  Block rates, even though often designed otherwise, tend to have a positive impact on consumption ond hence sales, instead of resulting in a decline of
consumption and therefore reduce the stress on water resources
14



V. Financial Performance
Financial performance was measured using three
proxies: (i) profit margin, (ii) government funding and                         JT
(iii) community contributions. Financial autonomy of
water supply stations in townlets and small towns is still
limited because water supply stations are not allowed  -
to set the water tariff. It was found that while provinces
control the water tariff in large urban areas; in small
towns, townlets and rural areas the tariff is mainly set
by the people's committees (STPC or CPC). The water
tariff of community-managed systems is usually set by
the community and endorsed by the People's
Committee.
Tariff Structures. National Urban Water Supply policy
states that tariffs overall should be adequate to cover
the full cost of water supply operation and maintenance,
but it is not enforced due to the political constraints.  block rate tariffs in some systems different tariff levels
Some provinces (such as Tien Giang, Hue, Quang Ninh)  are applied to different categories of users, which allows
have applied block rates, i.e. low tariff rates applied to  for cross subsidies for different categories of users, e.g.
a block of low consumption volume up to approximately  residential users.
15 m3/month to serve essential household consumption
demand per month. The tariff for the subsequent blocks  Block rates are usually promoted to help the poor gain
is then calculated progressively in accordance with  access to services, yet as most poor customers do not
increasing water consumption volumes. In addition to  have access to piped water supply, the actual benefits
15



of a cross-subsidy system    do not accrue to poor       The profit margin data collected by the research team
customers, but to customers with house connections,      show some interesting features. The data show that there
most likely relatively well-off customers. The most      is a large discrepancy between water tariffs per m3 and
common tariff structure is a flat rate structure, in which  water revenues per m3. In many cases, the actual tariff
the operator does not apply different rates for different  revenues per m3 are much higher than the water tariffs
blocks of consumption10. Tariff structures are distinct  per m3 would suggest. This phenomenon occurs among
between different management models. Block rates are     all management models (except for community-managed
especially common in PWSCs, and less so in privately     systems). The explanation for this may be improper
operated systems.                                        accounting that does not distinguish tariff income from
connection fees, and other income. Another explanation
Figure 6 shows that the actual tariff levels vary between  may be insufficient "ringfencing", which is a specific form
the different management models, with SOEs charging      of imprecise accounting. Many PWSCs and/or SOEs are
the highest and community managed systems charging       allowed to conduct multiple businesses to increase the
the lowest rates, with the average at approximately      viability of the enterprise. Revenues from  these other
$0.1 3/im3. Additionally, significant differences in tariffs  activities, such as pipeline construction, or supplying water
exist between the North and the South, most notably
between the Red River Delta and the Mekong Delta. On     Figure 6  Tariff levels (US$/m3)
average, water tariffs in the Mekong Delta are twice as
high as those charged in the Red River Delta.               0.20
Profitability. Factors that influence the "profitability of
water supply systems", whereby profit margin is defined
as the difference between revenues and operating costs
and depreciation, are:
o Consumption: The higher the water consumption per
capita, the higher the "profits" assuming the tariff is   0.10_    *-
adequate. It is likely that the size'' of the system, in
terms of how     much water is consumed        and
subsequently paid for, explains a large part of the
difference between a "profitable" and "non-proftable"
water provider.
o Tariffs. The higher the water rates, the more likely it is
that a positive return is generated.                                 -l    -L
o Investment cost per capita: The higher the investment
per capita, the more unlikely that the profit margin is                       "'s °       0<0 \
positive.                                                           o     o
10 A flat rate is defined as a single rate per cubic meter regardless of how much water is being used Block or progressive rates are using different rates for
different consumption levels Even though, many small towns and/or townlets use flat rates, they may have different rotes for different types of consumers
(for example, residentiol and commercial customers).
I Although the average town population is higher for the group of "profitable' companies, correlations show that the link between population size and
system size is tenuous
16



service materials, may end up as tariff revenues in the  no maintenance and/or treatment costs are registered.
books of the water company. Hence, in many cases actual  This is likely to have a positive impact on profitability in
water supply revenues may be significantly over-  the short-run, but is in the long-run an unsustainable
estimated.                                        strategy. In general, operation and maintenance costs
are underestimated.
In the remaining systems, billing efficiency seems to be
significantly lower than the survey data suggests. In Que,  Assuming a depreciation rate of 5 percent per annum
where a flat rate is in place, tariff revenues comprise less  (using a 20 year lifetime of assets), almost all water
than 60 percent of the tariff. In the case of systems that  systems register significant losses. The notable exceptions
use a block rate, for instance the PWSC of Tan Hoi Dong  are the towns of Ben Luc and Thu Thua (under SOE
and Tan Hiep, tariff revenues are also far below the first  management), Can Giouc, Cai Be, Tan Hoi Dong and
block of the water rate, suggesting that billing efficiency  Tan Hiep (under PWSC management) and Nam Giang
is far from perfect.                              (cooperative). If financing cost, operation and
maintenance and treatment cost, taxes and
A problem   also occurs with the operation and    overestimation of revenues are taken into account , it is
maintenance costs. The reliability of the data is hard to  possible that several of the "profitable" systems would
check as underlying information is often missing (such  turn unprofitable. Management models have only a
as electricity use, or staff numbers). In several systems,  marginal impact on the profit margin.
Figure 7 Profit Margin by Management Model
STPC/CPC                       -                      _
Cooperatives
Communities
Private
SOEs
PwScs
7          ~~~~7 
-20%           -15%           -10%           -5%            0%             5%
This Figure excludes the two major outliers
Tu Ha, a PWSC managed system and Quang An, a privately managed water system
17



Figure 8  Funding Pattern by Management Model    Funding Patterns. There is quite a variety between
management models and regions in terms of funding,
00%                                         with the exception of the STPC/CPC management model.
On average, the state puts in about 40 percent of total
investment costs. Customers pay most of the remainder,
80% _                          1             either through direct up-front contributions or through
tariffs that provide the companies with the cash flow to
pay for the investments. The remaining funds are provided
60%                                         through loans.
The funding pattern per region is substantially more
40%                      diverse than that of the different management models,
40%      _.                - 
with large variations in government contributions and
hence in the contributions of customers/communities.
Most government subsidies are directed to the North East
20%      _        *                          and North Central Regions. In the Red River Delta, the
government on average contributes 50 percent of the
investment costs, compared to only 12 percent in the
0%                     __-L _      -_        Mekong Delta. From a poverty alleviation perspective,
c\CC  &0 1  f    4        t   2         this distribution makes sense as the average per capita
(P o-                                      income in the North is much lower than that of the
Mekong Delta.
|   Siate [7 Cusiommer/community  Other
The funding patterns of the PWSCs and other SOEs are
intriguing. Some PWSCs and SOEs obtain investment
PWSCs show the best performance if the major outliers  funds from the parent company (Khe Tre, Can Giuoc
are excluded. However, if the outliers are included, the  and Thu Thua), others through preferential loans (Tan
results are dramatic - the PWSC management model with  Hoi Dong, Thu Thua) and some others are funded
a small profit margin turns into the management model  through Provincial Water Supply Development Funds
with the largest loss margin.                    (Cai Be). The questions then arise:
With one exception, all "profitable" companies are  a What is the precise relationship between these water
located in the Mekong Delta. The superior performance  supply stations and the parent company?
of the water supply systems in the Mekong Delta is due  a How important are cross-subsidies in PWSCs and SOEs
to the higher than average per capita consumption, lower  - are small towns being cross-subsidized by larger
than average per capita investment costs and relatively  towns or by other than water suplly business activities
high tariffs.                                     the companies engage in?
18



VI. Social Performance
Social performance was measured in terms of affordability   Figure 9   Affordability of Tariff (% Household Income)
of water tariffs and connection fees, and customer
satisfaction with the quality of services.                      3.0%
Affordability of water supply services. Affordability or        2.5%
ability to pay is often measured as the sum households
spend on water as a percentage of their household income.       2.0%
Because no separate surveys were being undertaken, the
average per capita income was estimated for each small
town/townlet investigated12 multiplied by household size        1.5%
and used as a proxy for household income. Using that
proxy, water tariff costs are rather affordable at an average   10
2 percent of household income.                                  0.5%
High connection charges are a major obstacle for poor
households'3. Payment ability and willingness to pay for        0.0%_/
connections are different in each studied area and                       G                         o�@    ,cs
dependent on the economic conditions of each family. In             0'9   c      °
areas where people can easily access alternative good
quality water sources, demand for piped water tends to be
12 The average per capita income was determined by the district authority, which calculates this data annually using the methods as described in government
regulations
13 Poor households usually do not own fields Their only livelihood is farming, or they do not hove a stable job Average monthly per capita income is low -
normally below VND 60,000-100,000 They live in simple houses that lack basic facilities and convenience They usually have more children than the
average household, and their children are not properly educated because of the lack of meons
19



lower Vice versa, in areas where there are not many  o Installment payments are allowed over a period of 1 to 2
alternative water sources and/or water quality is      years (such as in Que, Ben Luc, Thu Thua, Tan Hoi Dong,
inadequate, demand for piped water tends to be higher.  Nhi Quy).
In most of the studied areas, however, many households  o Subsidy policies for poor households by offering lower
face difficulties in paying for the connection fee. The  connection fees and construction contributions (for poor
connection fee is one important and decisive factor for  and so-called priority households).
poor households in getting access to piped water.
Satisfaction with Water Pressure and Water Quality. In
Figure 10 Connection Costs as o Percentage of       some towns, complaints focus on problems such as water
Household Income                           clarity or odor. Customer satisfaction varies substantially
40%                                             among management models and with different
technologies. Although many of these complaints result from
35%                                             differences in the water sources, operation and maintenance
may also play a role.
30%
25%  -     ~      -                  [          Figure 11 shows that, in general, customer satisfaction with
the pressure in the system is quite high. SOEs show much
20%                      ^[ lower performance in delivering well pressurized systems
15%               _[l than the other management models.
15%
Figure 11   Csm     Sst 0i         Pe
0%90
C C,  44x4° <oNS4  oNe  \C,9680% _
70%  -        _ ___
The average total water connection cost in some small town  60%  -       -                   _
water supply stations is relatively high, up to VND 3 million  50%  -
(US$197) against a minimum wage of about $1 per day.     40%  -
The reasons for the high connection fees are manifold:
high construction costs; incomplete secondary network;   30%  -           *_
choice of materials in project design (for instance, the use  20% -           -I
of galvanized steel pipes instead of PVC pipes,) and the          l._         _
use of indirect expenses and additional charges.         10%                ,       -     L
In order to provide incentives to ensure affordability, some                    &
water supply systems have applied the following                 c°   c
supportive policies:
20



Vil. Institutional Performance
Ownership. One of the major issues still to be  the facility owners - especially when projects were
addressed in Vietnam's small town management of  partially financed by the State. The PC decides on the
water supply systems is ownership. Generally,    management and operation model forthewater supply
ownership of the water supply systems in small towns  system. For townlet water supply systems in which the
is fairly clear. The DPC or PWSC, state-owned    investments were funded by private investors, or the
enterprises, are assigned by PPCs to be investors in  community, the private sector and the community are
water supply prolects and later to bethe facility owners.  the owners of the facility.
In townlets, PCERWASS played the role of investor in
almost all of the water supply systems investigated.  Ownership issues arise, however, in cases where
Upon prolect completion, PCERWASS hands the facility  multiple sources, such as private investors and
over to local authorities - DPC, STPC, CPC or   community have contributed to the investment capital,
agricultural cooperatives. Local authorities formally are  as illustrated in Box 1.
Box 1 Ownership Who is in Charge2
In Nam Giang, 1,250 households have connections for which they contributed 375 million VND. At the some
time, all shareholders of the cooperative have drawn parts of their shares from the Cooperative, so the contribution
of the users and the CPC (representing the state fund aid) have become disproportionately larger but these
stakeholders are not considered as shareholders. Tien Giang Province has recognized this problem, and has
promulgated a regulation that does not allow state-owned and private water supply enterprises to mobilize
funds from customers/users for the construction of water supply facilities. In case funds have been mobilized,
they must be refunded. The legality and the regulation on the contribution by water users to the expansion
investments must be considered and guided more adequately, ensuring the customers' participation but also the
legality in terms of the rights of the water users and the owners of the facilities.
21



Lack of clarity in ownership rights affects the process of  It is obvious that very new systems have lower production
mobilizing people to make contributions. Some water  to capacity ratios than more mature systems, as systems
supply systems receive large contributions from their  are usually built with a long term time horizon (at least
customers for construction and/or rehabilitation (such as  1 0 years). The lowest production by capacity ratio is for
Thu Thua, Ben Luc, Nam Giang). This may lead to   STPC/CPC run systems, where only 21 percent of the
confusion over ownership rights, causing disputes  capacity is used, although these systems have been
between different stakeholders. Clarifying property rights  operating for at least 9 years.
at the outset will help avoid later conflicts in management
and operation.                                    In small towns, two thirds of the systems do not comply
with technical design standards. The PWSC and the STPC/
For the purposes of this evaluation, since it was difficult  CPC-run companies used technical designs far above
to find a good indicator for ownership, this aspect is  the standard (up to 120 lcd). The deviation for townlets
reflected in community and government contributions.  is much less stark; with most providers using design
standards closer to those set by the Government.
Adherence to Standards. The variable that was included  However, these are average deviations and they may
as a proxy for institutional performance is the adherence  reveal very different patterns within the same category. If
to technical design standards as set by the Government.  the two major outliers are taken out of the comparison,
The Ministry of Construction has designated the standard  PWSC's average deviation reduces to 19 lcd above the
per capita water consumption per day as follows:  design standard, while the private companies are under-
designing their systems.
Table 2  Design Standards for urban and rural
domestic water consumption                Figure 12 Average deviation of design capacity from
Categories              Consumptfon                technical standards
standards
(liter/capita/day)      100%
Towns category I and 11           200-250
Towns category III and IV         150-200             80%
Towns category V (Small Town)      80-120
60%
Rural areas (including townlets)    25-50
Allowed Tolerance of 10 - 20% according to concrete situotion  40%
Even though these standards may have been set, they     20%
are hardly adhered too. There is a gap between actual
standards and what is applied in the different systems,  0*
with many being over-designed. For all management
models on average, only 36 percent of capacity is used
for production. For very recently constructed systems such  -20L
as the systems run by cooperatives, communities and              GI.'  e.,&e  efr
private operators, the actual use of capacity is much lower.           +    c     c'
22



Box 2   Technical Design Options and their Impact on Financial Viability
The capacity and service level of townlet water supply systems are usually determined on the basis of the funds
which can be mobilized from the State, donors, community contributions or other sources. Because these funds
are usually in short supply, investment in water supply systems is usually done gradually, closely following
demand. Initially only a small well is drilled, a small pump is used. Once demand increases, additional investments
are made (drilling more wells, purchasing more pumps, expanding the pipeline network, etc.). Good examples
are: Tan Hoi Dong, Nhi Quy and Ben Luc which at present operate at around 70 to 80 percent of their design
capacity. This investment model allows making full use of plant capacity, avoiding the construction of excessively
large facilities, and reducing operation cost and depreciation, and thus helping to keep the cost of water down.
However, in the design of the pipe network the need for expansion in the future should betaken into consideration:
the main pipeline should be of sufficient diameter, the treatment station should be designed in modules which
can be easily integrated when expansion is needed.
On the other hand, some water supply systems have invested large sums compared to the ability of its inhabitants
to pay, so that the actual amount of water pumped is much lower than the capacity invested in. This case is
common where the principal investor is not also the manager and operator of the system, such as in Tu Ha,
where the DPC was the principal investor, constructing a water station with a capacity of 4,000 m3/day and a
cost of 18 billion VND to provide water to the two small towns of Sia and Tu Ha. Tu Ha operates on only less than
5 percent of its design capacity, just like Quang An. Other small towns and townlets like Vu Thu, Kien Khe, Que,
Trung Lao and Le Loi 1 also have systems which are seriously over-designed.
23



VilI. Conclusions and Recommendations
Access to water supply services, not to mention sanitation  Figure 13 Overall performance by Management Model
services, is still rudimentary in small towns and even more
so in townlets and communes. In the study sample, about
38 percent of the population in these settlements owns house  140
connections, while a slightly higher number representing
45 percent of the population has access to water supply  120
services. However, these data are rather rosy as they reflect
the situation in settlements that have a basic piped water
network, which is only the case in a minority of these                * 0
settlements nationwide.
80
Is there a "best practice"
management model?                                         60
The ten selected indicators reflecting these dimensions were
ranked - the best performance within the sample getting   40
the highest number of points. The total values show that
sustainability is rather marginal for all systems with an  20
average value hovering around 1 10 out of a maximum score
of 202. On the basis of this indicator, which sums all these
dimensions of sustainability together, the performance by
management model is shown in Figure 13.                       "C     o     ', 9       &'   ('0S
There are no serious outliers in how management models
perform. Cooperatives, communities, and PWSCs are doing
24



better than average, while STPC/CPC-managed systems are  o Better technical and financial performance of a system's
performing below average. The SOEs and privately operated  management does not necessarily result in overall more
systems are performing more or less average.           successfully performing systems - customer satisfaction
may still be low if tariffs or connection costs are high and
The implications are interesting, showing that:        the needs of the poor are not met
* The use of more focused customer-oriented approaches
results in better overall performance, which supports  Cooperatives, PWSCs, and SOEs have, on average, higher
evidence elsewhere in the world that demand-responsive  financial scores than the other management models.
approaches have a positive impact on systems'     Technical performance is best by PWSCs, SOEs and
sustainability. Cooperatives, communities and private  cooperatives. Part of the better performance of PWSCs and
operators show above-average social performance, while  to a lesser extent SOEs may be due to the higher age of the
SOEs and STPCs/CPCs that are also engaged in outside  systems which influences the coverage and consumption
business activities, may be less effective than companies  levels in systems that are more mature than the systems that
that deliver water supply services exclusively.    only recently came on stream  Age of a system is strongly
correlated with the technical performance of the system,
and hence adjusting for age is likely to make the differences
Figure 14 Impact of different indicators on performance  in technical performance much less evident
140                                              The largest differences between the various management
models are shown in the social dimension of performance
120 ____________________________                  Systems managed by communities, cooperatives and private
operators have much higher scores than systems managed
100             ,-, f      ..under other management models SOEs score significantly
below average in this dimension of performance. STPC/
CPCs score below average on institutional performance.
80
The data presented here are average data, as such they do
60                                               not reflect differences within the same management models
The PWSCs show very mixed results with companies in the
40   _-.                                         Mekong Delta (in the province of Tien Giang) showing the
water supply systems among the top performers in the
201       _          - *             , _  _      sample, while PWSCs in other regions are scoring below
20          *l    *     *    *average in overall performance. The question arises what
makes the PWSC of Tien Giang so much more successful2
0                         -                     The reasons for the success of Tien Giang province are
9GC3  59~0              X <  o  \            manifold, but the most likelyto have contributed are (i) clear
s     c°~'               institutional arrangements between the different agencies
involved in the water supply sector, with Tien Giang having
* Financial * Technical * Social  Institutional  three different types of arrangements (PWSC, rural WSC,
and private operator); (ii) entrepreneurial set-up, with PWSC
25



also undertaking other business activities, including the  population, households with access to water supply services
design and construction of small towns and commune water  or the different set of standards that apply for small towns
schemes; (iii) additional financing arrangements, in which  and townlets. However, "economies of scale" do exist when
a water development fund has been set up for the    consumption is taken into account.
rehabilitation and expansion of investments, funded by
existing customers through a surcharge of VND200 per cubic  Government regulation, incentive structures and procedures
meter                                               (including technical standards, government subsidy policies,
for instance) seem to have an overall adverse impact on the
Within these three systems in the same province, the oldest  system's performance. Lack of government regulation in
PWSC systems perform significantly better than the more  townlets has resulted in a sharp proliferation of different
recently established rural water supply company or the  types of management models which are tailored to local
private operator, suggesting that age should not be  needs. Technical standards in townlets are more in line with
overlooked as a factor explaining performance.      actual consumption patterns, resulting in a lower probability
of over-designed systems. Also, the average townlet depends
Yet, the large variety in performance between models also  more on other sources of funding than that provided by the
shows that the more important aspect of management is  government. In general, townlets generate larger community
not so much the organizational model around which it is  contributionsthan small town-based systems, a feature linked
organized, but what "rules of the game" are being used  to above average performance.
and applied. Some "rules of the games" like autonomy in
managing the water supply business, proper tariff levels,  As the benefits of "economies of scale" are not
affordability of services are better indicators for success than  straightforward, the issue regarding the value of aggregation
are the management models per se.                   or regionalization of systems needs to be addressed.
Aggregation or regionalization has the advantage that
systems within a province can pool scarce human and
What about "economies of scale"?                    financial resources more efficiently Aggregation also offers
the possibility to use different forms of cross-subsidies, such
Correlation shows there is a weak, but inverse relationship  as those between different locations and between different
between the size of settlement and performance: the higher  types of business activities.
the model's performance rating, the more likely it is located
in a townlet. This is an intriguing result - and it seems to  Yet, this does not necessarily result in a better outcome for
contradict the intuition that small towns can benefit more  society than establishing a more competitive water market,
from "economies of scale" than townlets.            characterized by an enabling environment for a variety of
management systems. It is important that the cross-subsidy
The difference in population size between small towns and  system is understood so as to arrive to a proper tariff design.
townlets is smaller than one may expect. The average small
town has 10,805 inhabitants compared to about 8,500  In the sample, PWSCs as regional or provincial entities, show
inhabitants for the average townlet. Correlations between  a reasonable, above-average performance, but this
coverage, population size and performance show      performance is mainly due to the fact that the one provincial
inconclusive results suggesting that other factors besides size  PWSC (represented with three systems) in the Mekong Delta
are more important in defining overall performance- either  is performing very well; other PWSCs in other provinces
26



Box 3  One poor family in Vu Thu town talked about its ability to pay.
"We have two children but we have only 4 "sao" of field. If we want to have piped water, we should pay at least
400,000 VND, equivalent to some tens of kilograms of rice. How we can find so much rice? At the beginning of
the school year, we had to borrow money with a high interest rate to pay the school fees for our children. We
have to wait for the crops or to sell chickens and ducks to pay for the debt. Hence, it will be difficult to get piped
water to our home."
Tran Thi My - prioritized family in Cai Rong town, Quang Ninh
"Currently we are using well water, but we want to change to piped water because the water from the well is
becoming increasingly polluted. In order to have piped water, we should have a service line from the other side
of the road to our house-the line would be nearly 1 00m long, and the estimated cost ranges from 1.5 million
VND to several million VND. But we have no savings and therefore we can't pay for the pipe in order to have a
water connection."
perform below average Moreover, factors such as age of  they are not able to purchase tanks for storing rainwater or
the system, hydrological and socio-cultural factors, rather  to dig wells, they have to buy water from neighbors with a
than the benefits of aggregation only, may influence  higher price In a group discussion, people in Can Giuoc
performance significantly.                            town said "It is possible to live without eating for 3 days,
but how can we live without water for 2 days?", or people
Serving the Poor                                      from Nam Giang "People sometimes need water more than
electncity, short of water for one week is really a nightmare"
Water supply services are still not reaching the poor14 in
small towns and townlets For the moment, the beneficiaries  Cross subsidies need to be studied further to fully understand
of pro-poor pricing schemes and policies are generally more  the role they may play in reaching the poor, especially with
well off residential customers. As elsewhere in the region,  regard to the system of cross-subsidies that is at work in
the poortend to be excluded from accessto piped networks,  PWSCs and SOEs and possibly STPC/CPCs managed
because of the prohibitive cost of connecting to the system  systems The fact that many PWSCs have profitable
(especially so in systems run by PWSCs, SOEs and STPC/  businesses flourishing alongsidetheir water supply business
CPCs) and the location of the poor, often on the periphery  (including construction-related businesses - pipeline
of settlements The above-mentioned government subsidy  installation contracts with districts, and communes, bottled
policy could also start addressing how the poor can gain  water, and the like), the question arises, how sustainable
access to services The question is how to target the poor,  the actual water supply business in these companies is?
specifically, what is the best manner to ensure their access  Although cross-subsidies are a means of funding water
to services.                                          supply investments, more insight into how the different
systems are subsidized could help rationalize the
A majority of households prefer to have connections and  government's subsidy policy and the water company's tariff
use piped water. If there is no substitute water source and  policies.
14 Poor households usually do not own fields Their only livelihood is farming, or they do not have a stable iob Average monthly per capita income is low -
normally below VND 60 ,000-100,000 They live in simple houses that lack basic facilities and convenience They usually hove more children thon the
average household, and their children are not properly educated because of the lack of means
27



Annex 'b Description oi small towns and towingefs studied
Nam Dinh        NamGiang (T)   15785       1000       03 / 1998      Water Supply Cooperative                 Surface
(North)/        Trung Lao (T)  6,811       500        04/ 1998       Agricultural Cooperat,ve                  Surface
Red River Delta  Giao Tien (T)  17,877     900        01 / 1999      Commune People's Committee               Surface
Than Binh       Vu Thu (ST)    6,585        1000      2/1999          District Water Supply Company            Surface
(North)/        Lu (T)      . 1316          100        1999           Community self-managed.                 Surface
Red River Delta  Village 4 (T)  2746       120         10/2000       Community self-managed                   Surface
Dong Sam (T)   3949        200         6 / 2000       Private investment and operation         Surface
Ha Nam          Kien Khe (ST)  8,977       600        2 / 1999       Community self-managed.                   Surface
(North)/        Que (ST)       5,024       1000       8 / 1997       Town People's Committee direct management  Surface
Red River Delta
Long An         Ben Luc (ST)   17450       1000        1996          District Water Supply Company and        Ground
(South)/                                                              Multi-Service Enterprise
Mekong River    Can Giuoc (ST)  10,670     1000       4 / 1998        Provincial WSC                           Ground
Delta           Thu Thua (ST)  14,497      1000        1996 Almost    District WSC                             Ground
Rehabilitated
1940
Tien Giang      Tan Hoi Dong (T) 7,367     400         4 / 1996       Provincial Rural Water Supply Exploitation  Ground
(South)/                                                             and Management Company
Mekong River    Nhi Quy (T)    11 153      800         3 / 1996       Private investment and operation         Ground
Delta           Cai Be (ST)    14990       3000        1994 partially  PWSC                                    Ground and
rehobilitated                                           surface
1940
Tan Hiep (ST)  5,747        640        1994 fully     PWSC                                     Ground
Rehabilitated
1970
Quang Ninh      Quang Yen (ST) 8,176       2,000       1,976          District Division of Transport and Public Works  Surface
North-East      Mao Khe (ST)   31,559      2,000       1,976          PWSC                                     Surface and
ground
Cai Rong (ST)  6,989        1 ,000     1,993          DPC                                      Springs
Mong Cai (ST)  30,000       4,500      1,992          Environmental Services Company           Surface
Thua Thien Hue  Tu Ha (ST)     7,574       4,000       1,970          PWSC                                     Surface
North-Central   Sia (ST)       9,632                   1 999          From Tu Ha and Hue systems               Surface
Quang An (T)   10,190       1,000      1,999          Private                                  Surface
Khe Tre (ST)   3,163        500        1 ,995         PWSC                                     Surface
28






o FlO E PM@    tg g M  o        mn
e ~ )o - E3  ° lfluq Q° l °  Oq
U"Rpw roco AAAQg oqa-*, pe&Q l oo °  og °   ( 
¾   '- - -   N~~~~~~- -- - - - -  -  - <---
#ig  W Dt3  w  f5I@)  bD  Xu[Fe(ii EoQSDD@DGD <, °B3)  A
QD >3 G328QS3            IQ°Q 
O4Q~~~~~~~~~~~~~
D   -         \   -      v    / 9'  -
_; -     -        '~~~~- ,/) 
lN _N . 
_ _ _ -         - 
CX   >  J '  ~~~~ /   y