24
Pricing Urban Water
A Marginal Cost Approach
Kala Seetharam Sridhar and Om Prakash Mathur†
Introduction
The objective of this chapter is to estimate the marginal
costs of providing water in Indian cities. As discussed
later in the chapter, marginal cost and not the average
cost, should be the basis for pricing water supply. A
city usually develops its least expensive water sources
first, but as demand grows it normally becomes increasingly expensive to produce an additional unit of
water. In such an instance, using the average leads to
an underestimation of the cost of additional water (see
Williamson 1988).
Ideally the basis for pricing water supply should be
the long run marginal cost (LRMC). However, due to
the lumpy nature of capital expenditure and difficulty in
full cost recovery of capital expenses in water supply, it
is difficult to use LRMC-based water pricing. The alternative is to use short run marginal cost (SRMC) based
water pricing, where the basis for pricing is operation
and maintenance (O&M) costs. The literature suggests
that SRMC has been used for pricing various modes
of transport (see Link 2003 for pricing of roads and
Idström and Tervonen 2004 for rail infrastructure).
This chapter applies the principles of SRMC to six
water utilities (or municipal service providers) in India
and evaluates their tariffs against the price obtained
from SRMC estimates. The next section discusses the
data on water supply and service levels in the six cities
of study. Following this, we summarize the results
from the estimation of marginal costs, the output
elasticity of costs which has implications for returns to
scale. Then, the marginal costs of supplying water and
water tariffs are compared in the cities of study, with a
view to understanding the policy implications of the
work.
Description of Data on Water Supply
Expenditure and Service Levels
The discussion in this section and the marginal cost
estimates in the following section are based on data
from six Indian cities—Bengaluru, Chandigarh, Jaipur,
Surat, Lucknow, and Pune.1 Table 24.1 presents the
data on expenditure and the level of service of water
supply for the six cities for the period 1991–2003.
†
The authors would like to thank Surender Kumar for help with econometric estimation in the paper. The authors thank the
United Nations University–World Institute for Development Economics Research (UNU–WIDER) for the time to work on this,
where part of this paper was completed when the principal author was visiting UNU–WIDER during April–May 2011. Any errors
remain that of the authors.
1
For a description of why the choice of these Indian cities makes sense, see Sridhar and Mathur (2009).
352
India Infrastructure Report 2011
Table 24.1
Capital and Operation & Maintenance (O&M) Expenditure on and
Availability of Water Supply, All Cities (No. of cities: 6)
Capital expenditure
O&M
Per capita expenditure
(at constant 1993–4 prices) (Rs )
Water supply per capita
per day* (litres)
Per capita expenditure
(at constant 1993–4 prices) (Rs)
Year
Average
Max.
Min.
Average
Average
Max.
Min.
1991
50.22
107.02
0.47
206.13
96.73
216.40
1.83
1992
26.28
73.11
0.43
195.59
99.72
249.29
1.61
1993
21.77
43.85
0.40
239.98
97.59
277.71
1.47
1994
49.76
112.77
0.39
243.06
119.69
381.78
1.30
1995
44.69
84.93
0.41
257.54
116.34
325.93
1.24
1996
65.70
173.43
0.27
250.46
149.01
395.40
1.43
1997
48.48
102.65
0.23
246.11
155.55
320.28
1.47
1998
81.10
161.23
0.25
248.09
166.11
350.74
1.54
1999
110.71
321.41
0.19
248.13
197.56
351.36
1.44
2000
130.42
397.47
0.23
246.36
215.35
425.24
1.48
2001
102.34
447.24
0.19
244.56
193.30
407.42
1.41
2002
104.53
432.31
0.23
254.96
203.31
473.52
1.35
2003
55.57
120.80
0.29
248.48
151.24
395.02
1.33
Average
68.58
240.73
150.89
Source: Sridhar and Mathur (2009).
Note: *The water supply amounts stated are net of leakages.
The volume of water supply per capita per day in
the six cities varies between 196 and 260 litres during
the time period of the study. The maximum capital
expenditure is Rs 447.24 per capita in 2001–2 while
per capita O&M expenditure is Rs 473.52 (2002–3).
The minimum per capita capital expenditure incurred
is Rs 0.19 (1999) while the minimum per capita O&M
expenditure is Rs 1.24 (1995).
Since our sample consists of cities which differ on
several criteria, we classify them into various categories
to examine if there are common outcomes for cities in
similar situations. We consider Chandigarh and Surat
as our benchmark cities to demonstrate how much
spending can be expected from such cities to offer a
certain level of services. Over the years, we find that our
benchmark cities have supplied more water per capita
than others and that the expenditure on water supply
has also been higher for benchmark cities.
Next, we classifiedthe six cities on the basis of whether
water is supplied by a municipal corporation or nonmunicipal bodies such as a parastatal (for instance, the
Bangalore Water Supply and Sewerage Board, BWSSB
in Bengaluru) and other state level bodies. We find
that, on average, both per capita capital and O&M
expenditures are lower in cities with non-municipal
service providers than they are in cities with municipality service providers (see Tables 24.2). This could either
be a reflection of the fact that non-municipal bodies
are more efficient in the delivery of their services or
that they spend too little per capita. There appears to
be greater support for the latter since the average per
capita per day volume of water supply is also higher
in the municipality service provider cities than in the
non-municipal counterparts. However, water supply is
more volatile in the municipality provider cities than
that in the non-municipal service provider cities.
Pricing Urban Water
Table 24.2
353
Capital and O&M Expenditures on and Availability of Water Supply, Non-Municipal and
Municipal Provider Cities
Capital expenditure
Non-municipal
provider cities
(no. of cities: 3)
O&M expenditure
Municipal
provider cities
(no. of cities: 3)
Non-municipal
provider cities
(no. of cities: 3)
Average per capita
expenditure
(at constant
1993–4 prices)
(Rs)
Municipal
provider cities
(no. of cities: 3)
Year
Average per capita
expenditure
(at constant
1993–4 prices)
(Rs)
Average water
supply
per capita
per day
(litres)
Average per capita
expenditure
(at constant
1993–4 prices)
(Rs)
Average water
supply
per capita
per day
(litres)
Average per capita
expenditure
(at constant
1993–4 prices)
(Rs)
1991
62.84
149.69
12.38
262.56
70.77
135.68
1992
29.42
149.00
16.87
242.18
85.88
120.50
1993
19.99
168.00
27.13
311.97
95.45
100.82
1994
40.72
184.39
76.85
301.72
129.46
105.05
1995
31.53
184.15
64.44
330.92
111.01
124.34
1996
39.25
181.62
92.15
319.30
134.27
163.75
1997
31.05
180.03
74.62
312.19
108.99
202.12
1998
54.27
189.15
107.94
307.02
110.64
221.57
1999
69.98
183.56
151.45
312.71
118.64
276.47
2000
141.72
181.46
119.12
311.25
144.54
286.15
2001
156.22
177.80
48.45
311.33
130.19
256.40
2002
164.39
182.83
44.68
303.05
160.47
246.15
2003
18.49
183.53
80.29
291.79
5.42
248.46
Average
66.14
176.55
87.01
301.38
108.13
191.34
Source: Sridhar and Mathur (2009).
Note: *For 1991–4, the capital expenditures are just for Surat, hence the standard deviation is 0. Pune did not supply data on capital
expenditures for those years and Chandigarh became a municipal corporation only in 1994.
We distinguish the impact of Octroi-levying cities
from those that do not levy this tax. Octroi, while
being a distortionary tax, has been a major source of
revenue for cities. Hence, it is expected that cities that
have access to this revenue should be spending more
than those that do not. It may, however, be mentioned
here that most cities in India have abolished Octroi and
abolition of this tax is a major reform agenda driven by
the central government. In our sample of cities, Surat
and Pune continued to have the Octroi (during the
time period chosen for the study) whereas Bengaluru,
Lucknow, Jaipur, and Chandigarh did not.
Aggregating the O&M and capital expenditure
across all the years, we find that the Octroi levying cities
(that is, Surat and Pune) indeed spent higher amounts
per capita on water supply than their non-Octroi counterparts (see Tables 24.3 and 24.4).2 The per capita per
day supply of water in the Octroi cities was also, on
2
As of June 2011, Octroi continues to be levied by Pune. For Octroi rates for 2010–11 see http://www.punecorporation.
org/pmcwebn/index.aspx, last accessed on 9 June 2011. The continued existence of Octroi has also been confirmed by talking to
an official.
354
India Infrastructure Report 2011
Table 24.3
Year
Capital and O&M Expenditures on and Availability of Water Supply, Cities with and without Octroi
Cities with Octroi (no. of cities: 2)
Average capital per capita expenditure
(at constant 1993–4 prices) (Rs)
Cities without Octroi (no. of cities: 4)
Average capital per capita expenditure
(at constant 1993–4 prices) (Rs )
1995
64.44
31.53
1996
121.41
37.84
1997
102.65
34.94
1998
137.97
52.67
1999
195.39
68.37
2000
158.43
116.41
2001
57.52
124.74
2002
36.81
138.39
2003
60.04
52.59
103.85
62.34
Average
Source: Sridhar and Mathur (2009).
Table 24.4
O&M Expenditures on Water Supply, Cities with and without Octroi
Cities with Octroi (no. of cities: 2)
Cities without Octroi (no. of cities: 4)
Year
Average O&M
per capita expenditure
(at constant 1993–4 prices) (Rs)
Average water supply
per capita per day
(litres)
Average O&M
per capita expenditure
(at constant 1993–4 prices) (Rs)
Average water supply
per capita per day
(litres)
1991
135.68
205.75
70.77
206.31
1992
120.50
183.62
85.88
201.57
1993
100.82
296.37
95.45
211.79
1994
105.05
283.87
129.46
222.65
1995
124.34
275.98
111.01
248.31
1996
148.13
268.45
149.45
241.47
1997
151.65
260.48
157.51
238.93
1998
156.99
262.06
170.67
241.10
1999
245.95
279.51
173.36
232.44
2000
237.20
285.84
204.42
226.62
2001
180.89
281.44
199.50
226.12
2002
179.31
274.67
215.31
241.82
2003
175.18
265.87
135.29
236.90
Average
158.59
263.38
146.01
228.92
Source: Sridhar and Mathur (2009).
Pricing Urban Water
average, higher than that in the non-Octroi cities, a
finding that again reinforces the relationship between
spending and level of service in the case of water supply.
While spending may or may not translate into higher
levels of service, it is possible that where cities are efficient (for instance, those that ensure minimal leakages)
in their provision of the service, higher spending does
result in higher volume of the service.
Further, we made a distinction between cities whose
populations grew rapidly in the 1990s and those that
Table 24.5
355
grew more slowly during this period (Table 24.5).3
Surprisingly, the slow-growth cities spent more (capital
as well as O&M) per capita on water and were able to
supply higher volume of water per capita. We noted
that Bengaluru, which is the highest spender on water
in absolute terms, was a slow-growing city during the
1990s. So it is possible that the findings in Table 24.5,
of the slow-growing cities spending more per capita
than the fast-growing ones, are influenced by the
figures for Bengaluru. Bengaluru’s growth in the 1990s
Capital and O&M Expenditures on and Availability of Water Supply, Cities by Population Growth
Capital expenditure
Fast-growing cities
(no. of cities: 3)
O&M expenditure
Slow-growing cities
(no. of cities: 3)
Fast-growing cities Slow-growing cities
(no. of cities: 3)
(no. of cities: 3)
Year
Average per capita
expenditure
(at constant
1993–4 prices)
(Rs)
Average water
supply
per capita
per day
(litres)
Average per capita
expenditure
(at constant
1993–4 prices)
(Rs)
Average water
supply
per capita
per day
(litres)
Average per capita
expenditure
(at constant
1993–4 prices)
(Rs)
Average per capita
expenditure
(at constant
1993–4 prices)
(Rs)
1991
6.43
199.80
94.02
212.45
91.07
208.00
1992
8.65
187.59
43.91
203.58
80.87
214.02
1993
13.76
242.15
29.78
237.82
67.70
209.70
1994
38.62
234.40
60.89
251.71
70.47
268.12
1995
43.10
230.84
47.08
284.23
83.31
264.94
1996
81.03
226.18
50.36
274.75
99.23
278.75
1997
51.44
221.56
46.51
270.66
101.59
291.11
1998
92.07
220.89
70.14
275.28
105.17
308.18
1999
130.32
229.73
91.10
266.53
164.44
326.81
2000
105.70
234.12
155.14
258.59
158.63
363.69
2001
38.41
229.29
166.26
259.83
121.06
339.39
2002
24.62
226.12
184.45
298.22
119.99
363.78
2003
40.12
220.30
78.74
290.76
117.23
306.38
Average
51.87
223.31
86.03
260.34
106.21
287.91
Source: Sridhar and Mathur (2009).
3
We used the average growth rate of population during 1991–2001 for the six cities to distinguish between fast-growth and
slow-growth cities. Based on this, cities that grew relatively rapidly during the 1990s were Surat, Jaipur, and Pune while Bengaluru,
Chandigarh, and Lucknow were classified as being the slow-growth cities.
356
India Infrastructure Report 2011
of course does not include the eight urban areas, which
were merged into the Bruhat Bengaluru Mahanagara
Palike (BBMP) only in 2006.
Results from Estimation of
Marginal Costs
As McNeill and Tate (1991), Link (2003), and Idström
and Tervonen (2004) point out, the marginal cost is
equal to the marginal operating cost, which includes
variable costs. Turvey (1976), however, points out that
the capital costs required to meet incremental demand
for water tend to be lumpy, and cannot be determined
statistically. Others (for example, Warford 1997) also
generally accept that for capital expenditures, a statistically determined function would be rarely appropriate. Hence we estimate the SRMC, based on O&M
expenditures.
The first step is to develop a cost (expenditure)
function based on the city’s/utility’s budgets for O&M
expenditures. The cost function shows the relationship
between the water supplied and the costs incurred.4
Other factors such as topography, input prices, and
expenditure responsibilities of the local government also
determine the expenditure/cost levels (for a complete
discussion of the methodological challenges involved
in separating costs from expenditures, see Sridhar and
Mathur 2009).
To estimate the cost function, a random effects panel
data model is estimated. ‘Random effects’ is the most
suitable procedure here as it accounts for unobserved
heterogeneity.5 Table 24.6 summarizes the random
effects estimates of the marginal cost of supplying one
extra kilolitre (kl) of water. The results indicate that
marginal cost estimate for Bengaluru’s is about Rs 2.43
per kl of water provision (followed by Chandigarh at
Rs 1.83 per kl). The estimates also show that ownership
(city versus parastatal or utility) has the biggest effect on
the cost of providing water supply—the non-municipal
bodies incur a marginal cost of Rs 141 per kl of water
supply when compared with that incurred by municipal
bodies.6
Based on the estimates in Table 24.6, Table 24.7
presents the output elasticity of cost. An inference for
increasing or decreasing returns to scale can be made
on the basis of these estimates. In cities where municipal bodies offer water supply (Chandigarh, Surat,
and Pune), the output elasticity is >1 (with decreasing returns to scale), whereas in the case of all cities
where non-municipal bodies offer water supply (Jaipur,
Lucknow, and Bengaluru), the output elasticity is <1
(relatively inelastic) with the result that they experience
increasing returns to scale, quite in line with what one
would expect with utilities or parastatal bodies.
Based on the estimates in Table 24.6 (for all cities),
we arrive at the predicted expenditures, predicted costs
(based on various factors included in the estimation)
and compare these with actual average expenditures incurred by these cities on water supply (see Table 24.7).
4
What determine this volume of water supply (or of any other service considered here) is of course subject to debate—migration,
increasing population, or simply demand. We do not have data to determine the demand schedule for water for which micro,
household-level data on water tariffs paid and quantity of water consumed would be required. Education is a normative characteristic,
which could affect the preference for water. But it may not necessarily affect the actual expenditure/cost incurred, at least not in the
context of India. If education affects the demand for water, then it must be the case that the highest water spending municipalities
should also be the ones with educated population since that indicates water demand. We did attempt to get data on the proportion
of population with bachelors and masters’ degrees in the six cities of our study over the entire time period. This was available only for
a few years, which substantially reduced the size of our already small sample.
In alternative specifications, we were exploring the possibility of using average household income in the city as an exogenous
determinant of the level of expenditure on water supply, which is a different variant of the education characteristic. But that may not
be necessary or desirable. That would involve mixing positive and normative issues. Further, income data at the city level in India are
rarely collected; for a single year we could use data published by the National Council of Applied Economic Research (NCAER). But
such data are not available in a time-series fashion, required for the study. So, effectively, we were unable to adequately control for
local preferences for public services in determining expenditures.
5
We performed a Hausman test on the suitability of a fixed versus random effects; and found that random effects is most suited
for estimating this model.
6
Since the dependent variable is the log of the deflated O&M expenditure on water supply, we converted the coefficient estimates
into numbers for interpretation by taking the exponent of the logs.
Pricing Urban Water
Table 24.6 Random Effects Estimation of Expenditure
on (Net) Water Supply, Dependent Variable: Log of O&M
Expenditure, All Cities (deflated in 1993–4 prices)
Variable
Coeff.
Z
Log of net water supply (net of leakages)
0.89
3.57
Leakages
0.01
2.16
Log of city’s land area
0.40
2.07
Ownership[municipal body (0)
versus a parastatal (1)]
4.95
1.84
Log of duration of watersupply (in hours)
0.05
0.20
Net watersupply dummy for Chandigarh
0.60
1.90
Net watersupply dummy for Jaipur
–0.27
–17.90
Net watersupply dummy for Surat
0.56
1.81
Net watersupply dummy for Pune
0.53
1.72
Net watersupply dummy for Lucknow
–0.25
–11.10
Constant
–4.87
–3.35
Source: Public Health Engineering Department, Government of
Rajasthan.
Note: Number of observations is 53.
Table 24.7
City
Output Elasticity of Cost and
Returns to Scale
Output
elasticity
of cost
Economies
of
scale
Increasing
or decreasing
returns to scale
Chandigarh
1.49
0.67
DRS
Jaipur
0.62
1.63
IRS
Surat
1.45
0.69
DRS
Pune
1.42
0.70
DRS
Lucknow
0.64
1.56
IRS
Bengaluru
0.89
1.13
IRS
Source: Government of Rajasthan.
Notes: DRS–Decreasing returns to scale; IRS: Increasing returns
to scale.
Table 24.8 shows that all cities, notably Jaipur and
Lucknow, spend very little on water supply (with their
actual expenditure as a proportion of expenditure predicted on the basis of various factors, being less than
2 per cent), when compared with what we predict on
357
the basis of various characteristics. Bengaluru’s expenditures on water supply are in line with our projections
from the model.
Comparison of Marginal Costs and
Water Tariffs
Part of the rationale for estimating marginal costs is
that many cities might find it economically efficient
to price their services appropriately (as reflected in the
tariffs) and offer a better level of public services rather
than close their doors to in-migration. Table 24.9 summarizes the water tariffs for the six cities (as of 2006).
The rationale for setting water tariffs by cities is
based on some notion of affordability. Such a notion
is delinked from the O&M and capital expenditures
(which they perceive to be the same as costs, although
they are not the same. (See Sridhar and Mathur 2009
for an explanation of differences between expenditures
and costs incurred by the cities). Cities such as Jaipur
have always kept the price of water low and affordable
for major sections of the population. Political considerations have played a major role, and typically no cost
or expenditure considerations are taken into account
while determining tariffs.
In Lucknow, the water tax is set at 12.5 per cent
of the annual rental value of the property, so it is primarily related to the consumption (of water) which is
assumed to depend on the size and other characteristics
of property. No considerations of coverage of capital or
O&M expenditures or costs are taken into account by
the Lucknow Jal Sansthan. Similarly, in Pune, the water
Table 24.8
Predicted and Actual Expenditures
City
Predicted
expenditure
Actual
expenditure
Actual/
Predicted
(per cent)
Chandigarh
515,146,668
333,694,057
64.78
Jaipur
447,074,592
2,840,345
0.64
Surat
650,225,825
241,288,782
37.11
Pune
1,143,062,189
581,299,405
50.85
776,737,233
12,715,307
1.64
1,159,286,604
1,262,246,076
108.88
Lucknow
Bengaluru
Source: Sridhar and Mathur (2009).
358
India Infrastructure Report 2011
tax is set at a certain proportion of property taxes which
are based on the annual rental value of property.7 This
is based on the assumption that the consumption of
water is related to carpet area of the household. Thus,
while the cities relate water tariff to consumption of
the good, most are unable to recover their actual costs
or expenditures of supplying water, due to concerns of
affordability or political considerations.
On the other hand, in Bengaluru, the increase in
water tariffs is based on proportionate increases in the
electricity expenditures which account for nearly half
of the total expenditures, thus confirming the role that
topography plays in increasing expenditures though
affordability concerns also play a role. Surat switched
to a system of metered connections in March 2008. In
Table 24.9
City
Chandigarh
this system, consideration is paid to both expenditures
and the level of consumption of water. Water tariffs are
based on the O&M expenditures of supplying water,
the carpet area of the household for which the connection is given. The cost (expenditure) of salaries of
the employees and water treatment are covered by the
water tariff. Currently Surat is able to recover about 70
per cent of the O&M cost (expenditure) through the
tariffs. By 2011, as required by the Jawaharlal Nehru
National Urban Renewal Mission, the city will be covering 100 per cent of O&M expenditures through its
water tariffs. However, the city is not covering depreciation charges in its water tariff. Chandigarh Municipal
Corporation has attempted to cover nearly 80 per cent
of its O&M costs through the tariff.8
Current Water Tariff Structure for Metered Water Connections
Rate of Water Tariffs (rate per kl)*
Duration
From 31 March 2002 till now
Surat
Domestic
Non-domestic
1–15 kl @ Rs 1.75 per kl
15–30 kl @ Rs 3.50 per kl
30–60 kl @ Rs 5.00 per kl
above 60 kl @ Rs 6.00 per kl
Weighted average: Rs 5.01 per kl
Institutional: Rs 9
For government andsemi-government
offices: Rs 12.
For industrial, semi-industrial,
commercial establishments: Rs 11
All unmetered monthly Rs 240)
(not consumption-based)
13.0**
Pune
January 2000 to 31March 2005
from January 2005 till now
Rs .3.00 per kl
Rs 3.00 per kl
Rs 16.00
Rs 21.00
Bengaluru
Current
Rs 19.44 per kl
Rs 6 to Rs 60.00
Jaipur
From 1 June 1998 till now
Upto 15 kl @ Rs 1.56 per kl
15–40 kl @ Rs 3.00 per kl
Above 40 kl @ Rs 00.
Weighted average: Rs 3.39 per kl
Limit
Up to 15 kl
15–40 kl
Above 40 kl
Lucknow
Current
Rs 2.45 per kl
Non-domestic: Rs 12.25
Commercial: Rs 7.35
Government: Rs 90
Non-domestic
Rs 68
Rs 8.25
Rs 11.00
Industrial
Rs 11.00
Rs 13.75
Rs 16.50
Sources: Individual cities, service providers, and authors’ computations.
Notes: *These tariffs are current as of 2006, when this work was originally completed.
**For non-domestic uses, depending on the purpose, various tariff rates apply, the highest being applicable for industrial uses (Rs 24
per kl), and the minimum (of Rs 4 per kl) for use in educational institutions. What is reported here is the average of the non-domestic
rate for various purposes. The full schedule of rates for non-domestic uses is summarized in Table 24.9.
7
For instance, for annual rental value ranging from Rs 0–3000, the water tax is Rs 1000 a year (see Sridhar and Bandopadhyay
2007).
8
For instance, on average, about Rs 65 crores is incurred annually on O&M costs, out of which nearly Rs 50 crores is recovered
through the tariff.
Pricing Urban Water
The SRMC estimates obtained here represent only
the O&M expenditures. Due to this, they appear to
be lower than the international evidence regarding
marginal costs of providing water. A World Bank (1994)
study finds that in Lima, the LRMC9 of providing
water supply was $0.45 per cubic metre (that is, per
kl) whereas the actual tariff was only around $0.28 per
cubic metre.
In the case of Chandigarh and Jaipur, we computed
weighted average tariffs based on the quantities and
rates for various categories. This weighted average tariff
turns out to be Rs 5.05 in Chandigarh and Rs 3.39
359
per kl in Jaipur. Based on the estimates in Table 24.6,
Chandigarh, Surat, and Pune incur positive marginal
costs in supplying water to their residents. Lucknow
and Jaipur spend very little on water supply and for this
reason an additional kl of water does not impose much
burden for these cities, as Table 24.6 confirms.
The results indicate that there is a potential for
increasing tariffs as the current tariffs are lower than
the SRMC. For better management of urban water services, efficient pricing would have to be complemented
with reduction of leakages, thefts, and unaccounted for
water, in the distribution system.
References
Idström, Tiina and Juha Tervonen (2004),‘Marginal Rail
Infrastructure Costs in Finland 1997–2002’, Finnish Rail Administration, Traffic System Department.
Helsinki 2004. Publications of Finnish Rail Administration A 6/2004, (ISBN 952-445-102-8, ISSN
1455-2604), available at http://rhk-fi-bin.directo.fi@
Bin0f0fb8897ed6c16aebf2c2d7a11bda79/13142643
30/application/pdf/32220/Rhk-a604.pdf last accessed
on 25 August 2011.
Link, Heike (2003) ‘Estimates of marginal infrastructure
costs for different modes of transport’, Paper submitted
to the 43rd Congress of the European Regional Science
Association, 27–31 August, Jyvaskyla, Finland.
McNeill, Roger and Donald Tate (1991), Guidelines for
Municipal Water Pricing, Social Science Series No. 25,
Inland Waters Directorate, Water Planning and Management Branch, Ottawa, Canada.
Sridhar, Kala Seetharam and Om Prakash Mathur (2009),
Costs and Challenges of Local Urban Services: Evidence
from India’s Cities, Oxford University Press, New Delhi.
Sridhar, Kala Seetharam and Simanti Bandopadhyay (2007),
Improving the Fiscal Health of Indian Cities: A Pilot Study
of Pune, World Bank, December, National Institute of
Public Finance and Policy, New Delhi.
Turvey, R. (1976) ‘Analyzing the marginal cost of water
supply’, Land Economics, Vol. 52, No. 2 (May), pp.
158–68.
Warford, Jeremy (1997) ‘Marginal opportunity cost pricing
for municipal water supply’, Special paper, International
Development Research Centre, Ottawa, Canada.
Williamson, J.G. (1988), ‘Migration and Urbanization’, in
Hollis Chenery and T.N. Srinivasan (eds) Handbook
of Development Economics, Vol. I, North Holland,
Amsterdam.
World Bank (1994), Lima Water Rehabilitation and Management Project, Report No. 13206-PE, Environment and
Urban Development Division, Latin America and the
Caribbean Regional Office.
9
For purposes of computing LRMC, data on expenditures by projects, disaggregated by civil works, and plant and equipment, is
required. However, we did not get such information from the cities. If we had had access to the disaggregated data, we could have
attempted computation of LRMC, using the approach suggested by Turvey (1976). This hinges upon the use of discount rates and
arriving at different capital recovery factors for plant and equipment vis-à-vis civil works.