CHAPTER 4
The economic conception of water
W. M. Hanemann
University of California. Berkeley, USA
ABSTRACT: This chapter explains the economic conception of water -how economists think about
water. It consists of two main sections. First, it reviews the economic concept of value, explains how it is
measured, and discusses how this has been applied to water in various ways. Then it considers the debate
regarding whether or not water can, or should, be treatetl as an economic commodity, and discusses the
ways in which water is the same as, or different than, other commodities from an economic point of view.
While there are some distinctive emotive and symbolic features of water, there are also some distinctive
economic features that make the demand and supply of water different and more complex than that of
most other goods.
Keywords: Economics, value ofwate!; water demand, water supply, water cost, pricing, allocation
INTRODUCTION
There is a widespread perception among water professionals today of a crisis in water resources
management. Water resources are poorly managed in many parts of the world, and many
people -especially the poor, especially those living in rural areas and in developing countries-
lack access to adequate water supply and sanitation. Moreover, this is not a new problem -
it has been recognized for a long time, yet the efforts to solve it over the past three or four
decades have been disappointing, accomplishing far less than had been expected. In addition,
in some circles there is a feeling that economics may be part of the problem. There is a sense
that economic concepts are inadequate to the task at hand, a feeling that water has value in
ways that economics fails to account for, and a concern that this could impede the formulation
of effective approaches for solving the water crisis.
My own personal assessment is that the situation is somewhat more complex than critics
suggest. On the one hand, as environmental and resource economics has evolved over the past
forty years, it has developed a conceptual toolkit that I think is well suited for dealing with many
of the issues of water supply and water resource management. On the other hand, economists
sometimes slip into older ways of thinking and characterize economic value in terms that are
inadequate or misleading. Moreover, even among economists there is an inadequate appreci-
ation of the complexities of water as an economic commodity; these render it distinctive from
other commodities and they contribute to the explanation of the current crisis in water.
This chapter examines the economic conception of water -how economists think about
water -at least partly in light of these concemsl. It consists of two main sections. Section 2
I I am well aware that there are other conceptions of water coming from other disciplines. I see those as
complements, rather than substitutes, for the economic conception of water. For a fascination account
of alternative conceptions of water in the 19th century, see Hamlin (2000).
62 WM. Hanemann
reviews the economic concept of value, explains how it is measured, and discusses how this
has been applied to water in various ways. Section 3 takes on the debate regarding whether or
not water can or should be treated as an economic commodity, and discusses the ways in which
water is the same or different as other commodities. The chapter ends with a few concluding
observations in section 4.
2 WHAT IS ECONOMIC VALUE? HOW IS IT MEASURED?
Is economic value measured by market price? If an item has a price of US$ X, is this also
the amount of its. economic value? Most people assume the answer is yes, and economists
sometimes also make this statement. For example, the following passage equates the economic
value of water with its market price:
"In a market system, economic values of water, defined by its price, serve as a guide to allocate
water among alternative uses, potentially directing water and its complementary resources into
uses in which they yield the greatest total economic return" (Ward & Michelsen, 2002).
If it were true that economic value is measured by market price, this would imply that only
marketed commodities can have an economic value. Items that are not sold in a market -
including the natural environment, and public goods generally -would have no economic
value. If this were so, economic value would indeed be a narrow concept and at variance with
many people's intuitive sense of what is valuable.
In fact, however, economic value is different than price. Price does not in general measure
economic value, and items with no market price can still have a positive economic value.
This was first pointed out by Dupuit (1844) and Marshall (1879). But, as explained below, it
took until the 1970s for this to become well accepted within modern economics. It was around
this time that operational procedures became available to measure economic value separately
from price; and it was around this time that non-market valuation emerged as a field in eco-
nomics. It so happens that water as a commodity played a role in these developments, both clar-
ifying the economic concept of value and developing operational procedures for measuring it.
2.1 The meaning of economic value
The distinction between market price and economic value was famously noted by Adam Smith
in a passage in the Wealth of Nations describing the paradox of water and diamonds:
"The word value, it is to be observed, has two different meanings, and sometimes expresses
the utility of some particular object, and sometimes the power of purchasing other goods
which the possession of that object conveys. The one may be called value in use; the other,
value in exchange. The things which have the greate$t value in use have frequently little or no
value in exchange; and, on the contrary, those which have the greatest value in exchange have
frequently little or no value in use. Nothing is more useful than water; but it \vill purchase
scarce anything; scarce anything can be had in exchange for it. A diamond, on the contrary,
has scarce any value in use; but a very great quantity of other goods may frequently be had in
exchange for it" (book I, chapter IV).
Smith was using the comparison between water and diamonds to illustrate a distinction
between two different meanings of value. In fact, neither the distinction between the definitions
The economic conception of water 63
of value nor the use of water to illustrate it wa~ original with Smith2. Two thousand years before
Smith, Plato had observed that: "only what is rare is valuable, and water, which is the best of
all things. ..is also the cheapest,,3. In fact, Plato and Smith were both expressing a thought
that had occurred to many other people over the ages, namely that the market price of an item
need not reflect its true value. Market price reflects the fluctuating circumstance of daily life,
whether the vagaries of supply (sudden scarcity, monopoly, etc.) or demand (temporary needs,
changes in taste, fads and fashions), while the true value is something more basic, enduring,'
and stable.
Just what this true value is has been seen differently at different times. For Plato, the true
value was intrinsic to the ideal form underlying the item. For Aristotle, it was intrinsic to the
natural end that the item served. Aristotle also originated the distinction between this value -
in effect, value in use -and value in exchange: "of everything which we possess there are
two uses. ..one is the proper, and the other the improper or secondary use of it. For example,
a shoe is used for wear, and is used for exchan&e; both are uses of the shoe"4. For Saint
Thomas Aquinas, the true value of an item was determined by its inner goodness, an intrinsic
quality of the item stemming from its relation to the divine purpose. In the 14th century, some
Scholastics propounded a view closer to Aristotle's that the intrinsic value of an item arises
from its inherent usefulness and ability to please man according to rules of reason. However,
starting with Davanzati in 1588, Italian humanists stressed subjective human preference rather
than objective human need as the basis of true value. Men seek happiness, Davanzati wrote,
by satisfying all their wants and desires, and they value items as these contribute to this end.
While value reflects human preference -not only wants of the body but also what one later
writer called "wants of the mind, most of them proceeding from imagination" -price reflects
not only demand but also supply, and that is influenced by scarcity. As Barbon wrote in 1690,
"things may have great virtues, but be of small value or no price if they are plentiful".
These quotations from the 16th and 17th centuries demonstrate an awareness of three key
principles. First, demand is separate from supply. Demand indicates what things are worth to
people; supply indicates what things cost. Second, market price reflects the interaction of both
demand and supply and, in principle, is separate from each of them. Third, the value that people
place on an item (their demand for the item) inevitably reflects their subjective preferences5.
Returning to Adam Smith, given his distinction between value in use and price (value in
exchange), which is fundamentally the more useful measure of value? Here I part company
from Smith because, following Hume and Locke before him, he associated the true value of
an item largely with its cost of production. This English School held that, while the market
price of an item at any particular point in time is determined by demand and supply, in the long
run this will tend towards what Hume called afundamental price, and Smith a natural price,
2 Moreover, as I shall argue, Smith's analysis of this example is incomplete, and the conclusion he drew
from it is largely incorrect.
3 Plato Euthydemus, as cited in Bowley (1973). The discussion that follows draws on Bowley, Schumpeter
(1954), Gordon (1975), Pribram (1983) and Niehans (1990).
4 Aristotle, Politics, Book I, 9.
5 This applies to water, too, even though it is obviously an essential want of the body. Without wishing
to demean the importance of water, I will present some evidence below that, compared to other items
they could buy, people sometimes place a lower value on improving their access to water than what the
public health professionals would recommend.
WM. Hanemann
which is determined by the underlying cost of production. Implicitly, they were assuming a
horizontal long-run supply curve, so that consumer demand has no influence on price in the
long run. This is now seen as a special case, and the modern economic concept of value focuses
essentially on value in use.
The modern concept was first formulated by Dupuit (1844) and Marshall (1879, 1890).
Dupuit stated that the "maximum sacrifice expressed in money which each consumer w{)uld
be willing to make in order to acquire an object" provides "the measure of the object's utility".
Marshall used a very similar formulation; he defined the "economic measure" of a satisfaction
as "that which a person would be just willing to pay for any satisfaction rather than go without
it". These definitions highlight the distinction between demand and supply: the measure of
value is what the item is worth to the individual, not what it costs. Thus, an item can be cheap
to produce, in the sense that its total cost is low, but highly valuable to the owner, in that its
total value to him is large, or conversely.
Generalizing from this, the modern econ<1mic concept of value is defined in terms of a
trade-off. When an economist states that, for some individual, X has a value of 50 in terms
of Y, this means no more, and no less, than that the individual would be willing to exchange
X for 50 units of Y. Y is said to be the numeraire in terms of which value is measured. This
numeraire can be money but it need not be; it could, for example, be some specific commodity.
The trade-off is in no way limited to market goods; it can be between any two items that the
individual values, regardless of whether these have a market price.
Before any further discussion of the relation between value and price, it is necessary to
introduce a distinction which was lacking in Smith's analysis but was understood by Dupuit
and Marshall, namely the distinction between marginal, on the one hand, and average or total,
on the other. Thus, marginal value needs to be distinguished from average or total value, and
marginal cost from average or total cost. The marginal quantity measures the change in total
value, or total cost, associated with a unit change in quantity, while the average measures total
value, or total cost, averaged over the total quantity. Admittedly, there is one case where they
are the same: if the marginal value (or marginal cost) is constant as quantity changes, then
marginal cost (or value) coincides with average cost (or value). But, in general, marginal value
and marginal cost are not likely to be constant. In particular, the general presumption is that
marginal value (and marginal utility or marginal benefit) decline with quantity.
The notion of declining marginal utility was the cornerstone of Dupuit's analysis. Dupuit
recognized that if the consumer is free to vary the quantity of an item purchased, she will
choose this quantity so as to equate her marginal value (utility) for the item to its price. In that
case, the market price provides an accurate measure of the marginal value associated with the
last unit of consumption. But, Dupuit stressed, the total payment does not accurately reflect
the total value of all units consumed. This is because of diminishing marginal utility: if the
marginal value of the last unit just equals the market price, it follows that the marginal utility
associated with the infra-marginal units will be higher than this market price. In effect, the
consumer earns a profit on the infra-marginal units because they are worth more to her than
the price she pays, which in fact is why she consumes a larger quantity thereby rendering these
units infra-marginal. Marshall, who independently formulated a similar argument thirty years
later, called this profit the consumers 'rent in 1879, and the consumers surplus in 1890.
In summary, if there is a market price for the item in question and if the consumer is free to
vary the quantity of this item that she purchases, its marginal value to her is reflected in, and
can be measured by, the market price; otherwise, not. Even when price reflects marginal value,
The economic conception of water
total expenditure does not reflect total value; instead, total expenditure understates total value
because of the presumption of diminishing marginal utility. The distinction between marginal
and total is the key to the full resolution of the diamond and water paradox: water may have a
smaller value than diamonds at the margin, but it undoubtedly has a larger total value.
Although Dupuit and Marshall correctly enunciated the economic concept of value in its
modern formulation, it actually dropped out of favor with economists around the turn of the
last century. Marshall himself came to be troubled that his use of the demand curve to measure
consumer's surplus was inexact and relied on the assumption of a constant marginal utility of
income. And, as the ordinal utility revolution took hold in economics, Marshall's analysis based
on cardinal utility appeared hopelessly out-dated and irrelevant. It took until the 1970s before
these issues were .fully resolved and the Dupuit-Marshall concept was recognized as being
both fully consistent with modern, ordinal utility theory and susceptible of rigorous empirical
measurement. This came about as a result of several impertant conceptual advances.
First, Hicks rehabilitated the Marshallian conCept of consumer's surplus in a series of papers
starting with Hicks (1939), which demonstrated that this concept is in fact consistent with
ordinal utility theory and that it could be measured exactly if one were given an indifference
map. However, this was a pyrrhic victory because the general view was that, while it is a useful
theoretical construct, the indifference map is not itself directly observable. Hence, Marshall's
measure as re-interpreted by Hicks was not measurable in practice. This view finally changed
around 1970 as the result of the development of what is known as duality theory, including the
demonstration by Hurwicz & Uzawa (1971) of a theoretically rigorous yet practical numerical
procedure for identifying the specific utility function underlying any given system of demand
equations that satisfies the formal requirements of modern ordinal utility theory. This now
made it possible to start with an econometric estimate of a suitably specified demand equation
for a marketed commodity, or a system of demand equations for a set of commodities, and
derive a theoretically consistent and rigorous estimate of the Dupuit-Marshall measure of the
economic value of these commodities.
Second, building on Hicks (1939), Henderson (1941) discovered an alternative way of
characterizing the trade-off that underlies the economic concept of value. When one says that
a person is willing to exchange X for 50 units of Y, this could mean either: I) the person would
be willing to give up (pay) 50 units of Y to obtain X; or 2) the person would accept 50 units of
Y to forego X. The first uses maximum willingness to pay (WTP) as the measure of value, and
is the measure mentioned by Dupuit and Marshall and analyzed by Hicks (1939, 1941). The
second is the new measure that was suggested by Henderson; it uses minimum willingness to
accept (WTA) as the measure of value. Together, these exhaust the logically possible ways of
expressing a trade-off. Hicks (1942, 1943, 1946) analyzed the relationship between them in
the case of a price change and showed that they differ by an income effect6.
The third development was the extension of the economic concept of value to a broader
class of items than market commodities. In fact, nothing limits X in the definition of economic
value given above to being a market good; it could actually be anything from which people
derive satisfaction. This suggests that the same definition of economic value can be applied to
6 An important paper by Willig (1976) showed how one could use Hurwicz and Ozawa's result to develop
a tight numerical bound on the possible difference between WTP and WTA in the case of a price change
for a marketed commodity. The Hurwicz-Uzawa result implies that both WTP and WTA can be derived
from an econometric estimate of suitably specified demand equations.
66 WM. Hanemann
non-market items. For example, one could say that a person values some aspect of his health
at 50 units of Y if he would be willing to exchange 50 units of Y to preserve that aspect of his
health; that he values a beautiful sunset at 50 units of Y if he would be willing to exchange
50 units of Y to experience it; or that he values an endangered species of animal at 50 units
of Y if he would be willing to exchange 50 units of Y to ensure its preservation. In each
case, it should be evident that there are two possible ways to formulate the exchange: a WTP
formulation and a WTA formulation. This was demonstrated formally by Maler (1971, 1974)
who showed that, when Y is money, the Hicksian analysis and its modem formulation in terms
of duality theory carry over from the valuation of market goods to non-market items 7. Maler's
analysis thus provides a formal justification for the field of non-market valuation, including
the monetary evaluation of the natural environment.
2.2 Non-market valuation and water
.
,
Economic valuation deals with the valuation in monetary terms of items that people might
care for. Non-market valuation applies the same notion to items that are not sold in a market.
It is important to emphasize that the Dupuit-Marshall concept of economic value carries over
to such items. This is because, even for something that is not sold in a market, it is still
meaningful to conceptualize the economic measure of the satisfaction from the item as the
monetary amount which the person would be just willing to exchange for the item if it were
possible to make such an exchange. In effect, this generates a monetary measure of the change
in the person's welfare by using the change in the person's monetary income toot she would
consider equivalent to the item in question in terms the overall impact on her satisfaction8.9.10.
The history of non-market valuation in the USA is closely intertwined with water projects,
since these were an important motivation for the development of cost-benefit analysis. The
idea of cost-benefit analysis originated in the USA, in Hammond's (1960) phrase, as "an
administrative device owing nothing to economic theory" in the context of managing the
activities of the US Army Corps of Engineers around the beginning of the last century. The
1902 River and Harbor Act had created a Board of Engineers to review navigation projects;
7 However, there is an important difference. With valuation of non-market items, the difference between
the WTP and WTA measures involves not only an income effect but also a substitution effect (Hanemann,
1991).
8 The equivalence can be conceptualized in two possible ways -the maximum amount that the person
would be willing to pay to gain the item, or her minimum WTA to forego it.
9 It should be noted that this definition provides a unified approach to welfare measurement for both
firms and households. In the case of households, whose objective function is defined in terms of utility,
the monetary measure is the change in income that is considered equivalent to the change in utility. In
the case of firms, whose objective function is defined in terms of profit, the monetary measure is the
change in profits itself.
10 Although the modern economic concept of value is defined in terms of a trade-off, it is possible that
some people find themselves unable to make a trade-off because for them the two items being compared
are incommensurable. A type of preference that gives rise to such trade-off aversion is where there is a
lexical ordering over commodities: certain goods in any quantity or quality always take precedence over
all quantities or qualities of other goods, so that no amount of increase in the latter can ever compensate
for any reduction in the former. This is known as lexicographic preferences. A modified version of
lexicographic preferences is where the lexical ordering applies only below a threshold level of the good
(Lockwood, 1996); among other things, this generates a situation where the individual might have an
infinite WTA for a reduction in an item.
The economic conception o/water 67
in conducting a review, the Board was required to consider the commercial benefits from
such projects in relation to their costs. The River and Harbor Act of 1920 further required the
separate reporting of special, or local, benefits as opposed to general, or national, benefits
for the purpose of ensuring proper local cost-sharing. In 1934, the National Resources Board
appointed a Water Resources Committee to consider "the development of an equitable system
of distributing the cost of water resource projects, which should include not only private but
also social accounting". Finally, the Flood Control Act of 1936 permitted the Army Corps of
Engineers to involve itself in flood control provided that, in a famous phrase, "the benefits to
whosoever they may accrue are in excess of the estimated costs". In 1946, a Subcommittee
on Benefits and Costs of the federal Inter-Agency River Basin Committee was appointed to
investigate the practices of the various federal agencies that were engaged in the evaluation
of federal water resource projects and to formulate some "mutually acceptable principles
and procedures". This led ultimately to the publication. in 1950 of what became known as
the Green Book which attempted to codify the "Principles of cost-benefit analysis for use by
federal agencies. The following decade saw the publication of many academic journal articles
and six major academic books dealing with the economic analysis of water projects.
The 1950s were when the field of non-market valuation began to come into existence.
The approach that emerged first is what became known as the travel cost method or, more
generally, the revealed preference method. It arose initially out of an effort by the National
Park Service (NPS) to measure the economic value associated with the national parks. At the
time there were no entrance fees at national parks, so the NPS could not use park revenues
as a measure of their value. The issue was assigned to a staff economist who wrote to ten
distinguished economic experts for their advice. All but one replied that it was impossible
to measure recreational values in monetary terms, but the tenth, Harold Hotelling disagreed.
He saw that, even though there was no entrance fee for a national park, it still cost visitors
something to use the parks because of expenses for travel, lodging and equipment. These
expenditures were not captured by the NPS but, they still set a price on the park. Moreover,
this price would vary among people coming from different points of origin. By measuring the
price and graphing it against visitation rates one could construct a demand schedule for visits
to the site, and then determine consumer's surplus in the usual manner as the area under this
demand curve.
The NPS report followed the majority view and asserted that it was not possible to set
a monetary value on outdoor recreation. However, in 1956 the State of California hired an
economic consulting company to estimate recreational benefits associated with the planned
State Water Project. This company learned of Hot elling's idea and decided to apply it. A survey
of visitors was conducted at several lakes in the Sierras and data were collected on how far
they had traveled and how much they had spent. Using these data, a rough demand curve was
traced out, and an estimate of consumer's surplus was constructed. This analysis appeared in
Trice & Wood (1958), the first published application of the travel cost method. At the same
time, Marion Clawson (1959) at Resources for the Future had begun collecting data on visits
to Yosemite and other major national parks in order to apply Hotelling's method to them, which
was the second published application. By 1964, there were at least five more applications in
various parts of the USA, and the travel cost method was an established procedure.
The insight behind the travel cost method, and revealed preference generally, is that, while
people cannot buy non-market goods such as clean water or an unspoiled environment directly,
there sometimes exist market goods that serve as a partial surrogate for the non-market good
68 WM. Hanemann
because the enjoyment of these goods is enhanced by, or depends on, the non-market good.
In that case, the demand for the market goods is used as a surrogate for the demand for the
non-market good.
The limitation of this approach is that there may not exist a market good that can serve as
surrogate for the non-market good of interest. Moreover, even if such a good exists, it may not
capture all of people's preferences for the complementary non-market good. The conceptual
identification of what might be omitted by the revealed preference approach came about as a
result of papers by Weisbrod (1964) and Krutilla(1967). Both authors started from the premise
that some of people's motives for valuing the natural environment may differ from those
for valuing a market good. People may value the natural environment out of considerations
unrelated to their own immediate and direct use of it. Weisbrod focused on uncertainty and
what became known as option value: some people who do not now visit a national park,
say, may still be willing to pay money to protect it from destruction or irreversible damage
because they want to preserve their option of visiting it in the future. Krutilla focused on
what became known as bequest value and existence value I 1. With bequest value, the notion
is that some people would be willing to pay because they want to preserve the park for future
generations. With non-use value, the notion is that some people would be willing to pay even
if they knew that neither they nor their children would ever visit it; in Krutilla's example,
people may "obtain satisfaction from mere knowledge that part of the wilderness in North
America remains". These are legitimate sources of value, Krutilla and Weisbrod felt, but they
would not be respected by private managers of the environmental resource. Nor would they
be adequately measured by a conventional revealed preference analysis such as the travel cost
method. Consequently, some other method of measurement is needed.
The alternative approach, suggested by Ciriacy- Wantrup (1947), is to interview people and
elicit their monetary value; this became known in economics as the contingent valuation (CY)
methodl2. Ciriacy-Wantrup was discussing soil conservation and he noted that several of the
benefits were non-market goods, such as reduced siltation of rivers or reduced impairment of
scenic resources. He characterized the problem as being how to obtain a demand curve for such
goods, and suggested the following solution: "[Individuals] may be asked how much money
they are willing to pay for successive additional quantities of a collective extra-market good.
The choices offered relate to quantities consumed by all members of a social group. ..If every
individual of the whole social group is interrogated, all individual values (not quantities) are
aggregated". The results correspond to a market-demand schedule. While noting the possible
objection that "expectations of the incidence of costs in the form of taxes will bias the responses
to interrogation", he felt that "through proper education and proper design of questionnaires
or interviews it would seem possible to keep this potential bias small".
Having identified a solution conceptually, Ciriacy- Wantrup never pursued it further. The
first significant application was by Davis (1963) which dealt with the economic value of
outdoor recreation in the Maine woods; to measure this Davis interviewed a sample of hunters
II The latter is now also called non-use value and passive use value.
12 The same idea was earlier suggested by Bowen (1943) who conceived of surveys as a surrogate for
using voting to determine the public's demand for what he called social goods. Recently, the term stated
preference has been used to cover CV and related approaches. They are also known as direct valuation
whereas revealed preference approaches are referred to indirect valuation because they do not measure
preferences directly but instead infer them from externally observed behavior.
The economic conception of water 69
and recreationists and asked how much more they would be willing to pay to visit the area 13.
The next application was by Ridker (1967); to measure the damages from air pollution, Ridker
included some questions in a survey about people's WTP to avoid soiling from air pollution.
In 1969, a steady stream of CV studies began to appear in the economics literature. Official
recognition was given to CV in 1979, when the US Water Resources Council included it along
with travel cost as recommended methods of non-market valuation.
The first application of non-market valuation in the USA, the 1957 valuation of recreation
benefits from the California State Water Project, was a harbinger of things to come: since
then many non-market valuation studies have been conducted in the USA in connection with
water resources management issues, and the environmental consequences of water projects
have come to playa-significant role in the design and approval of water projects. These
trends emerged slowly in the 1970s and 1980s, driven by developments in the implementation
of the 1969 National Environmental Policy Act (NEPA)14. .NEPA required federal agencies
to prepare a detailed statement of environmental impacts for proposed major actions which
significantly affect the quality of the human environment, including the identification of the
environmental impacts of the proposed action, alternatives to the proposed action, and any
adverse environmental impacts which cannot be avoided should the proposal be implemented.
In consequence, since the mid-1980s it has not been acceptable in the USA to perform an
economic assessment of a major water project without including some non-market valuation
of the project's environmental impacts. For example, non-market valuations of environmental
impacts were included in the Department of Interior's re-assessment of the operation of Glen
Canyon Dam on the Colorado River in 1984-1992, and in the Bureau of Reclamation's assess-
ment of the Central Valley Project Improvement Act in 1993-1996. In California, they were
included in the State Water Resources Control Board's review of the diversions of water from
the San Francisco Bay/Delta to the Central Valley and Southern California, conducted in 1987-
1994, and in the Board's 1993 Mono Lake Decision requiring Los Angeles to reduce its diver-
sion of water from streams feeding Mono Lake on the eastern side of the Sierra Nevadal5. In
the case of Mono Lake, the Board decided that it was in the public interest to reduce Los Ange-
les' diversion from Mono Lake by about two thirds, despite the resulting loss of hydropower
and water supply (which amounted to over 8% of Los Angeles' total water supply) primarily
in order to protect habitat for birds and other wildlife; non-use values associated with habitat
protection constituted the main component of environmental benefits (Wegge et al., 1996).
It should be emphasized that the use of non-market valuation applies to positive as well
as negative environmental impacts of water projects. The experience in the USA has been
that these can generate significant economic benefits associated with water-based recreation,
eco-tourism, and the non-use value of ecosystem protection. These environmental benefits
sometimes greatly outweigh the benefits from agricultural or even urban water use. In short,
in the USA we have now moved from the traditional situation where there was essentially a
single objective for large water projects, namely the provision of water for off-stream uses,
13 Probably the first CV study was actually conducted in 1958 for the NPS, which hired a market research
company to survey residents of the Delaware River basin about their WTP entrance fees for national
parks.
14 Subsequently, a number of states passed laws imposing similar reporting requirements on agencies of
the state government; for example, the California Environmental Quality Act was enacted in 1970.
15 I served as the Board's economic staff for 1ts investigation of both these issues.
70 WM. Hanemann
to a situation where any new water project must have environmental restoration as an explicit
objective along with the provision of any off-stream uses.
IS WATER DIFFERENT?
Now that the economic concept of value has been explained, the question arises whether
it is appropriate to apply this concept to water. Is water an economic commodity, and can
it be analyzed using the conceptual framework of economics in the same way as any other
d.?
commo Ity.
The answer is contested ground between economists and their critics. One of the four Dublin
Principles, adopted at the 1992 International Conference on Water and the Environment in
Dublin, holds that "water has an economic value in all its competing uses and should be recog-
nized as an economic good". Similarly, Baumann~ Boland (1998) write: "water is no different
from any other economic good. It is no more a necessity than food, clothing, or housing, all of
which obey the normal laws of economics". Per contra, Barlow & Clarke (2002) proclaim it
as a "universal and indivisible" truth that "the Earth's freshwater belongs to the Earth and all
species, and therefore must not be treated as a private commodity to be bought, sold, and traded
for profit. ..the global freshwater supply is a shared legacy, a public trust, and a fundamental
human right, and therefore, a collective responsibility". Vandana Shiva (2002) writes in a
similar vein about a clash between two cultures: "a culture that sees water as sacred and treats
its provision as a duty for the preservation of life and another that sees water as a commodity,
and its ownership and trade as fundamental corporate rights. The culture of commodification
is at war with diverse cultures of sharing, or receiving, and giving water as a free gift".
My own view lies somewhere between these two positions. Baumann & Bolandareundoubt-
edly correct when they point out that food, clothing and shelter, like water, are necessities of
life, and they are typically provided through the market without any complaint. Why, they
ask, should water be different? I believe there are two reasons why this is so. First, water is
clearly viewed by many people as being different. The fact that water, unlike other household
commodities, arouses such passion speaks for itself: for better or worse, water is perceived as
having a special significance that most other commodities do not possess 16. This itself has eco-
nomic consequences. Second, I believe that water has some other economic features that make
it distinctive. These features make water different from, say, bread or land, as an economic
commodity yet they are often overlooked by economists. They matter greatly because they
affect the demand for water, its value, and the social and institutional arrangements by which
it is supplied. To explain them, I need to introduce several more items from the economists'
conceptual toolkit.
3.1 Water as a private good, water as a public good
Since Samuelson (1954), economists have drawn a distinction between conventional market
goods -also known as private goods -and what are known as public goods, "which all enjoy
16 This is true in rich as well as poor countries -in the USA, for example, it is notoriously difficult
for publicly owned urban water utilities to obtain political approval for even trivial rate increases while
other household utilities such as cable television raise their rates with impunity; Glennon (2004) makes
a similar observation.
71
The economic conception of water
in common". The two key properties of a public good are non-rivalry in consumption and
non-excludability. With conventional goods, one person's consumption necessarily competes
with that of another, in that more consumption by one person renders a smaller quantity of
that good available for consumption by anybody else. With public goods, by contrast, more
consumption by one person in no way reduces the amount available for others. Conventional
consumption goods are excludable in that, if this is so desired, it is physically possible. to
exclude any person from consuming the commodity. With public goods, by contrast, if the
good is available for consumption by anybody, it is available for consumption by all. Examples
of a public good suggested by Samuelson were "an outdoor circus or national defense which is
provided for each person to enjoy or not, according to his tastes". The abatement of pollution in
a lake is another example of a public good, as are other types of environmental improvement:
my enjoyment of the clean water in the lake in no way reduces the amount of clean water
available for your enjoyment (non-rivalry) and, if the water in the lake is clean for me to enjoy,
it is clean for everyone's enjoyment (non-excludability) I?
In this fmmework, water is both a private good and a public good. When water is being
used in the home, in a factory or on a farm, it is a private good. When water is left in situ,
whether for navigation, for people to enjoy for the view or for recreation, or as aquatic habitat,
it is functioning as a public goodl8. Moreover, while the water in a reservoir is a private good,
the storage capacity of the reservoir per se may be a public good. By contrast, most of the
other commodities associated with food, clothing or shelter are purely private goods and have
no public goods aspect; this is one of the respects in which water is different than these other
commodities in economic terms.
Samuelson identified two important consequences of the public good properties. First,
while public goods are likely to be supplied collectively, for example through a voting process,
rather than through a decentralized market, it is likely that they will be undersupplied because
people have a selfish incentive to free ride on the collective decision process by understating
their true interest in the public good. Second, the valuation of public goods is fundamentally
different than that for private goods because a public good can be enjoyed simultaneously by
many while a private good can be consumed by only one party at a time. Thus, the value placed
on a given unit of a private good is that of a single user -in an efficient market, this will be the
user with the highest and best use for the item. By contrast, the value placed on a public good
is that of many people, namely all those who care for the iteml9. This is why the non-market
17 In addition to private and public goods, there is an intermediate case where there is rivalry in con-
sumption but not excludability. These are known as common pool resources. Examples include fisheries,
forests, grazing grounds, and oil fields. The other intermediate case, sometimes called club goods or
quasi-private goods, is where there is non-rivalry combined with the possibility of exclusion. Examples
include television frequencies, public libraries, and bridges, for each of which it is possible to exclude
access. Furthermore, there may be non-rivalry at low levels of aggregate consumption of a club good, but
rivalry at a high level of consumption once the item becomes congested -this can happen, for example,
with parks and bridges.
IS To the extent that water-based outdoor recreation is excludable, this would be a quasi-private good. To
the extent that groundwater or water flowing through the distribution system of an irrigation district is
non-excludable, these are common pool resources.
19 This follows from Samuelson's demonstration that the aggregate demand curve for a public good is
constructed in a radically different manner than the aggregate demand curve for a private good. With
a private good, the aggregate demand curve is the horizontal sum of every individual's demand curve
for the good; with a public good, the aggregate demand is the vertical sum; this observation had in
WM. Hanemann
benefits of environmental preservation can sometimes outweigh the use benefits associated
with the diversion of water for off-stream agricultural or urban use.
The public good nature of water in situ, historically associated with navigation, has had a
decisive influence on the legal status of water. In Roman Law and, subsequently, in English
and American common law, and to an extent in Civil Law systems, flowing waters are treated
as common to everyone (res communis omnium), and are not capable of being owned. These
waters can only be the object of rights of use (usufructuary rights), but not of rights of
ownership2o. Thus, even though water and law are often complementary inputs, there is a
crucial distinction in that land can be owned, while water cannot.
3.2 The mobility o/wate?'
A distinctive physical feature of water is its mobility. Water tends to move around. It flows, it
seeps, it evaporates. When water is applied to,plants in the field (or to an urban landscape), a
substantial portion either seeps into the ground or runs off the ground as tailwater. In addition,
in residential indoor uses and most industrial uses there is usually an outflow of wastewater
after the use is completed. The consequence is that there can be several sequential uses of the
same molecule of water since water is rarely consumed fully by a given user and what is left
is physically available, in principle, for use by others22.
The mobility of water and the opportunity for sequential use and re-use make water relatively
distinctive as a commodity -especially compared to land, for which such multiple, sequential
uses are impossible (except in nomadic societies). These properties of water have important
economic, legal and social implications. Keeping track of water flows is costly and sometimes
difficult. Consequently, it is often hard or impractical to enforce excludability or to establish
property rights to return flows. In this respect, water is very different as an asset than land,
which is relatively easy to divide and fence. The common solution is to resort to some form
of collective right of access; in effect, this internalizes the externality associated with the
mobility of return flows. A classic example of this is the riparian water right in English and
American common law. This permits any landholder whose property is adjacent to a stream
or body of water to divert a reasonable amount of water, provided this does not cause harm to
other riparian landholders or interfere with their co-equal right to divert a reasonable amount
of water. The riparian right to the use of water is not a right to a fixed quantity, and it is a
co-relative right shared with all other riparians along the same stream23.
fact already been made by both Bowen (1943) and Ciriacy- Wantrup (1947). In terms of the distinction
between use and non-use values, if there is a non-use value for an item this is a public good.
20 By contrast, groundwater beneath private land and springs or rainwater found on private land is
typically treated by the law as being privately owned by the landowner(s) on whose land the spring
occurs or under whose land the groundwater lies.
21 The analysis in this and the following two sections was influenced by reading Young & Haveman
( 1985).
22 In the process, however, there can often be some reduction in the quality of the water relative to that in
the first use. Because of the solvent properties of water, the return flows are apt to dissolve and absorb
chemicals in the media through which they pass.
23 In American law, there is a further requirement that riparians put the water to a reasonable use. With
surface water, the major alternative in American law to the riparian right is the appropriative right, which
was developed in the arid West. This permits the diversion of water, regardless of whether the diverter
owns the riparian land, in a fixed quantity, subject to the principle of first in time is.flrst in right. The
The economic conception o/water
3.3 The variability of water
In addition to the mobility of water in streams, another crucial feature is the variability of
supply in terms of space, time, and often quality. Spatially, water is distributed very unevenly
across much of the globe; just six countries -Brazil, Russia, Canada, Indonesia, China and
Colombia -account for half of the world's total renewable supply of freshwater (Postel &
Vickers, 2004). Even within countries and regions, there is unequal spatial distribution. In
California, for example, two thirds of the state's population live in Southern California, but this
region receives less than 10% of the state's total precipitation. For any given region, there is
substantial variation in precipitation both within the year and between years. In California, for
example, approximately 80% of the annual precipitation falls between October and March,
while three quarters. of the water use occurs between April and September24. Beyond this,
cycles of wet and dry years occur in California as a function of wider climatic phenomena
such as the interannual El Nino-Southern Oscilla,tion and" the Pacific Decadal Oscillation.
While the annual runoff in California has averaged about 87,500 Mm3 over the past 90 years,
it has been as low as 18,500 Mm3 (in 1977) and as high as 166,500 Mm3 (in 1983).
Because of this variability, the major challenge for most large water systems is the spatial
and temporal matching of supply with demand. Storage is typically the key to controlling
the temporal variability in supply, while inter-basin transfers are used to overcome the spa-
tial mismatch between supply and demand. But, the variability of supply has affected not
just the engineering of water resource systems but also the legal, and institutional arrange-
ments for the use of water. The variability of supply is yet one more point of divergence
between water and land, and it explains why the property rights regimes are different: it
would surely be difficult to apply the ownership rights in land to so variable a resource as
water.
Besides the variability in supply, the demand for water may be intermittent, especially in
agricultural uses of water where crops need to be irrigated only at periodic intervals rather
than every hour of every day. Until the advent of affordable storage, which has mainly been
a phenomenon of the 20th century and large-scale diversion of water, the intermittent nature
of traditional agricultural demand was an important factor promoting the sharing of access.
If there is water in the stream and one member of the group is not currently diverting water
theory is that, if the streamflow is inadequate to meet all the diversion requirements, those with a more
recent (junior) date of initial diversion cede to those with an older (more senior) date. On the ground
in at least some Western states, the practice seems to be rather different. The precedence of seniority
is not self-enforcing without resort to litigation, which is slow and costly. Much of the time, therefore,
what actually happens with appropriative rights may be closer to a version of the riparian system. In
California, there is not a functioning system to record the actual diversions of water, nor to check these
against the quantity associated with the water right. Consequently, much of the surface water use by
agricultural occurs outside the formal structure of California appropriative water rights law. This can
become an impediment to long-run water transfers as the inadequate documentation casts a shadow of
doubt on a seller's specific property right.
24 The seasonal variability of precipitation in California is exacerbated by the fact that it is an arid region
with a Mediterranean climate. But precipitation is distributed unevenly throughout the year in almost all
parts of the world, albeit not as severely as in California. In Europe, the major part (46%) of the runoff
occurs during April to July, and similarly in South America; in Asia, 54% occurs in June to September;
in Africa, 44% occurs in September to December; in Australia and Oceania, 40% occurs in January and
April (UNESCO, 2000).
74 WM. Hanemann
from the stream, other members of the group were allowed to divert the water rather than let
it flow to waste in the ocean25. This is another key difference with land: while the demand for
water is intermittent, the demand for land to grow crops or to locate a building is continuous,
and there can be no such sharing of the same resource among multiple users. The intermitte.nt
nature of the agricultural demand for water is conducive to the collective sharing of a right of
access as opposed to individual ownership of a property right.
3.4 The cost ofwater
Compared to other commodities, and other utility services, the cost of water has several
distinctive features which complicate its supply.
Water is bulky, and expensive to transport relative to its value per unit of weight. Con-
sequently, the transportation infrastructure for water is far less extensive than that for more
valuable liquids such as petroleum. Also, cot,npared to electricity, water is relatively expen-
sive to transport, but relatively cheap to store. Therefore, the strategy for averting shortage
takes a different form with water than electricity. If there is a sudden shortfall in supply, with
electricity this can be made up almost instantaneously by importing power over the grid from
a source that could be 1500 km away or more. With water, there is no comparably intercon-
nected transportation grid and, even if there were, it takes longer to move a comparably large
quantity of water. Thus, to deal with unexpected outages, one has to either resort to rationing
or stockpile sufficient stored water prior to the period of peak use.
Another distinctive economic feature is that water supply is exceptionally capital-intensive
compared not only to manufacturing industry generally but also to other public utilities. In the
USA, for example, the ratio of capital investment to revenues in the water industry is double
that in natural gas, and 70% higher than in electricity or telecommunications. Moreover,
the capital assets used in water supply cannot be moved to another location and are generally
unusable for any other purpose; they represent an extreme type of fixed, non-malleable capital.
Furthermore the physical capital in the water industry is very long-lived. The infrastructure
associated with surface water storage and conveyance and the pipe network in the streets can
have an economic life of 50-100 or more years, far longer than that of capital employed in
most manufacturing industry or in other public utility sectors26.
In addition, there are significant economies of scale in many components of water supply
and sanitation. These are especially pronounced for surface water storage: given a specific
25 Storage changes this, because streamflow can be stored when it is not currently being used. The
discussion here focuses on agricultural rather than urban use of water. Urban use is different because it
is more continuous in nature, and when there is a piped water supply this is typically pressurized, unlike
with agriculture which relies mainly on gravity flow. Gravity distribution fosters sharing of intermittent
access, while pressurized distribution fosters simultaneous individual access.
26 The Roman aqueduct that still stands in Segovia, Spain, is ample testimony to the physical longevity of
certain types of conveyance structure. The effective life of a dam is governed by the rate of siltation but can
be well over a century. For piping, the American Waterworks Association recommends a replacement
cycle of 67 years. A conventional drinking water treatment plant may have a useful life of 40 years,
although high technology processes such as reverse osmosis facilities have a shorter life. Compared
to surface water, the capital infrastructure associated with groundwater typically has a shorter life; a
groundwater pump might typically have a life of about 25 years in the case of an electric pump, or about
15 years for a diesel pump.
75
The economic conception o/water
dam site, within some range, by increasing the capacity of the dam one can significantly
reduce the unit cost of stored water. With a groundwater source, by contrast, the economies of
scale in production are much less pronounced. There are also important economies of scale
in the treatment and conveyance of drinking water and wastewater27.
The capital intensity, longevity, and economies of scale mean that water supply and sani-
tation costs are heavily dominated by fixed costs. In a simple surface-water supply system
with minimal treatment of drinking water, minimal treatment of sewage prior to discharge, and
a heavy reliance on gravity flow, the short-run marginal cost of water supply and sanitation
may be almost zero except for small costs associated with pumping to move water through
the system28. Even in a modem system with full treatment of drinking water and sewage
discharges, the short-run marginal costs are extremely low. There is thus an unusually large
difference between short- and long-run marginal cost in water supplr9.
The capital intensity and economies of sc&le associated with surface water supply have
profound economic and social implications. For one thing, because these are classic pre-
conditions for a natural monopoly, they make it more likely that there will be a single provider
in any given area. More generally, they foster public provision of a surface water supply rather
than individual, self-provision, whether the public provision is by a collective of the users or
a monopoly seller3°. Furthermore, the construction and operation of large-scale surface water
storage and distribution systems require a high degree of co-ordination and social control. This
was the central thesis of Witt fogel's (1957) study of ancient hydraulic societies -civilizations
that were dependent on large-scale surface water diversion and distribution systems, such
as Mesopotamia, Egypt and China. Wittfogel argued that, in such societies, the effective
provision of water required the centralization of power and an oriental despotism mode of
governance in which a state bureaucracy, headed by an absolute ruler, ruled on the basis of
its control of the hydraulic system. Wittfogel's work was subsequently criticized by other
27 With the conveyance of drinking water from the point of production to the point of use, and of treated
wastewater to the point of discharge, there are economies of scale with respect to volumetric capacity
but not length.
28 In the USA, the ratio of operating costs to total costs for efficient water firms is about 10%; by contrast,
it is 32% for gas utilities and over 57% for electric utilities (Spiller & Savedoff, 1999). In the UK water
industry, Armstrong et at. (1994) report that operating costs represent less than 20%, and fixed costs
more than 80%, of total costs.
29 When piped water supply was introduced into cities in the 19th century, water agencies chose not to
meter individual homes or small non-residential users partly because of an ethos in favor of promoting
universal service but also because water was so cheap at the margin that they felt it was not worth the cost
of metering it. By contrast, electricity and gas were metered in residential connections. Water service
was financed by charges based on the type or value of the property being served. In the USA, metering of
residential users did not become common until well into the 20th century, and there were some notable
handouts (Denver and New York City did not meter until about 15 years ago). In the UK, nearly all
residential water users were unmetered prior to privatization in 1990.
30 For urban water, the main alternatives to a public supply of surface water are water vendors or household
self-provision through pumped wells or rainwater catchments. Where there are water vendors, the unit
cost of vended water is always much higher than that of water from piped supply, typically by a factor of
10 or more. The primary reason for the cost differential is economies of scale: it is far more expensive to
deliver water in relatively small quantities through multiple trucks rather than in large quantities through
a single pipe network. However, the up-front capital investment required for vended water is far lower
than for a piped water supply, and this can make vended water a viable alternative.
76 WM. Hanemann
scholars31. Nevertheless, his notion still resonates; it was applied to the American West by
Worster (1986), who characterized this as a hydraulic society based on the development and
control of water infrastructure by a political elite. This characterization is contested by Kupel
(2003), who argues that the history of water projects in Arizona is one of response by civic
leaders to requests for service by urban and suburban residents, closely resembling other
aspects of the history of modern urban infrastructure. This is not necessarily a contradietion:
the commonality in urban infrastructure is capital intensity and economies of scale, with the
consequent need for public sector leadership and social co-ordination and, also, the consequent
prospect of a handsome increase in land value in the area being served32.
Another problematic consequence of the capital intensity, longevity of capital, and
economies of scale in surface water infrastructure is the propensity to what might be called
lumpiness or, less politely, gigantism in these systems. Because of the economics, there is a
strong incentive to make a substantial expansion of capacity at a single point in time rather
than to plan for a series of incremental chang!?:s pread out over time33. The drawback is that it
may take many years, or decades, before the demand materializes to utilize this capacity (and
the willingness to pay -WTP -to finance it). When fully utilized, the project provides water
at a low cost; but there is uncertainty whether and when it will be fully utilized, and meanwhile
it ties up scarce capital. Large surface water projects are risky, and difficult, inter-temporal
balancing acts34.
3.5 The price ofwater
It is important to emphasize that the prices which most users pay for water reflect, at best,
its physical supply cost and not its scarcity value. Users pay for the capital and operating
costs of the water supply infrastructure but, in the USA and many other countries, there is
no charge for the water per se. Water is owned by the state, and the right to use it is given
31 It was pointed out that large-scale irrigation works in Mesopotamia were developed after the rise of a
centralized state, so that hydraulic society could be the result rather than the cause of state formation. The
Maya civilization, where irrigation was of marginal importance, was cited as evidence that centralized
states might not always be associated with hydraulic systems. It was also noted that there are several
modem communities in Mesopotamia where small-scale cooperative irrigation works without centralized
external control.
32 As noted above, the increase in land value made it possible to finance urban water infrastructure with
property taxes rather than through user charges. This may have been economically rational not only
because of the very low marginal cost of urban water, but also because of the public good benefits of
urban water supply associated with improved fire protection and also what, in the 19th century, were
believed to be the public health benefits of washing down streets (Anderson, 1980).
33 By contrast, systems supplied from groundwater are considerably less lumpy and more scaleable.
34 This is well illustrated by the experience of the Central Arizona Project, the most recent large water
project in the USA, which actually went bankrupt (Hanemann, 2002). The two key parameters for the
economic viability of a water project are til,e discount rated use to assess the present value of net benefits,
and the rate of growth in the public's ability and, more importantly, WTP for the water. Public agencies
can generally borrow at a lower interest cost than private firms, and also are more apt to take a long-term
perspective in evaluating investment. The public sector's ability and willingness to apply a low discount
rate is a major reason for its predominant role in the provision of water supply infrastructure. For all the
rhetoric on privatization of water, the private sector seems more interested in taking over the operation
of existing infrastructure rather than financing new infrastructure, except for water treatment facilities
which, as noted above, have a shorter life than other water supply infrastructure.
The economic conception of water 77
away for free. Water is thus treated differently than oil, coal, or other minerals for which the
USA government requires payment of a royalty to extract the resource. While some European
countries, including England, France, Germany and Holland, do levy an abstraction charge
for water, these charges tend to be in the nature of administrative fees and are not generally
based on an assessment of the economic value of the water being withdrawn. Thus, in places
where water is cheap, this is almost always because the infrastructure is inexpensive, or the.
water is being subsidized, rather than because the water per se is especially abundant.
In the USA, it has long been noted that the prices charged to farmers are far lower than
those charged to urban residents, often by a factor of 20 or more. It is often assumed that this
is because the irrigation water has been subsidized by the federal government, but this is not
in fact the main reason.
It certainly is true that the federal government has subsidized irrigation in the West by
waiving interest charges and other means. Between 1902 and 1994, the federal government
spent US$ 21,800 million to construct 133 watersu~plyprojects in the West. Although most of
the water from these projects is used for irrigation, the cost allocated for repayment by irrigation
users was set at US$ 7100 million (33%). Of this, US$ 3700 million was subsequently waived.
Of the remaining US$ 3400 million payable by irrigators, only US$ 950 million had actually
been repaid as of 1995 (General Accounting Office, 1996). The remaining balance will not be
paid off until well into this century, if at all. The combined effect is that recipients of irrigation
water from federal projects will have repaid, on average, about US$ 0.10 on each dollar of
construction cost. However, these projects account for only about 19% of total irrigation supply
in the West. The remainder comes from groundwater or non-federal surface supply projects,
none of which is subsidized to a significant degree35. While the non-federal irrigation supply
is more expensive than the federal supply, it still is much cheaper than urban water supply.
The reason is the sharp difference in the real cost of agricultural versus urban water supply.
Unlike urban water, irrigation water is not treated, and it is generally not available on demand
via a pressurized distribution system. Moreover, the physical capital used for irrigation supply
is often old and long-lived, and it may have been paid off long ago.
There is an additional tendency to under price water in the USA -urban as well as agri-
cultural- because most water agencies set price to cover the historic (past) cost of the system
rather than the future replacement cost36. There is typically a large gap between these two costs
because of the extreme lumpiness and longevity of surface water supply infrastructure. The
capital intensity of the infrastructure exacerbates the problem because, after a major surface
water project is completed, since supply capacity so far exceeds current demand, there is a
strong economic incentive to set price to cover just the short-run marginal cost (essentially, the
operating cost), which is typically minuscule37. As demand eventually grows and the capacity
becomes more fully utilized, it is economically optimal to switch to pricing based on long-
run (i.e. replacement) marginal cost, but by then water agencies are often politically locked
35 If there is a subsidy, it is likely to be mainly for the electricity used in pumping water.
36 This is due partly to the conservatism of the conventional engineering emphasis on cost recovery,
and partly to the fact that, since most water supply agencies in the USA are publicly owned rather than
investor-owned, there is a strong ethos to avoid making a profit on the sale of water.
37 Erie & Joassart-Marcelli (2000) argue that this type of water pricing encouraged urban growth, and
urban sprawl, in Southern California, but they fail to recognize the economic logic that drives it by virtue
of the lumpiness and capital intensity of water supply infrastructure.
78
WM. Hanemann
into a regime of low water prices focused narrowly on the recovery of the historical cost of
construction.
3.6 The essentialness of water
Water is essential for all life -human, animal, or plant. In economics, there is a concept,
also called essentialness, that formalizes this notion. The concept can be applied eitlier to
something that is an input to production or to something that is directly enjoyed by people as a
consumption commodity. In the case of an input, if an item has the property that no production
is possible when this input is lacking, the item is said to be an essential input. In the case of
a final good, if it has the property that no amount of any other final good can compensate
for having a zero level of consumption of this commodity, then it is said to be an essential
commodity. Water obviously fits the definition of an essential final good: human life is not
possible without access to 5 or 10 Lid of water per perSon. Water also fits the definition of an
essential input in agriculture and in several manufacturing industries (e.g. food and beverages,
petroleum refining, lumber and wood products, paper, chemicals, and electronic equipment)
that cannot function without some input ofwater38.
However, essentialness conveys no information about the productivity or value of water
beyond the vicinity of the threshold. It implies nothing about the marginal value associated with,
say, applying 76 versus 89 cm of water to irrigate cotton in the Central Valley of California.
It says nothing about the marginal value of residential water use at the levels currently expe-
rienced in Western Europe or the USA -the latter averages about 455-530 Lid per person,
more than two orders of magnitude larger than the minimum quantity that is needed for human
survival39.
The latter statement is not meant to belittle the uses of water by households in Western
Europe or the USA. My point is that, in addition to being essential for human life, water
contributes in important ways to the enjoyment of the satisfactions of life. Consequently,
there are many other residential end uses of water besides its use for drinking. Indeed, if one
examines the history of residential water use in the USA from the early 19th century to the
present, it is striking how water consumption has grown over time through the steady accretion
of end-uses, each representing the discovery of a new way to employ water for people's use
and enjoyment. When a piped water supply first became available in the 19th century, the
initial household uses were the same ones that had existed when family members had to fetch
water from an external source -drinking, cooking, hand washing, and limited bathing40.As
time passed, many other uses were found -tubs for bathing, water borne sanitary waste
disposal, outdoor landscape and garden watering, automatic clothes washers, swimming pools,
automatic dish washers, car washing, garbage disposal, indoor evaporative cooling, hot-tubs,
lawn sprinklers, etc. The result has been a constantly rising trajectory of per capita household
water use.
38 These are the largest water-using industries in the USA in terms of freshwater intake.
39 Total urban water use in the USA averages about 680-830 Ud per capita, depending on the location.
40 For example, Blake (1956) notes that out of 15,000 houses with running water in Philadelphia in 1849,
only about 3500 were equipped with private baths. In 1871, by contrast, 112,457 Philadelphia houses
had running water, and 80,000 of these had bathtubs and fixtures for hot and cold water (Anderson,
1980). The use of water closets to remove human wastes did not become widespread until almost two
decades after the introduction of piped water into homes (Tarr, 1979).
The economic conception of water 79
Two conclusions can be drawn from this historical experience. First, in developed countries,
the fact that water is essential for human life is almost certainly irrelevant when assessing
the value of residential water supply because the ways in which water is used are nowhere
near the threshold level at which essentialness applies. Second, there is a possibility that
some of the developing countries may also move along a rising trajectory of residential water
consumption because the things that have made abundant water use an element of a comfortable
modern life style in developed countries could also become attractive to people in these
countries as their income rises.
For reasons that are entirely understandable, there is a tremendous emphasis in the water
literature on the need to secure at least a minimal water supply for the nearly 1100 million
people around the globe who currently live without access to an improved water source. In
developing this estimate, the United Nations and the WHO used a figure of20 L/d per person
as the minimum human requirement for water for drinkirrg and basic sanitation. Gleick (1999)
has argued that this is too low, and has advocated the adoption of a basic human right to 50
L/d per person as the minimum required for bathing and cooking as well as the other basic
needs. As noted above, these low levels of consumption are not very different from the initial
levels of water use in developed countries when piped water was first introduced in the 19th
century. It is important to recognize that, if the efforts to provide improved water supply and
sanitation in the developing countries are successful, the future levels of water use that might
ultimately emerge in these countries could diverge from these minimum levels, as happened
in the developed countries.
In short, while it is obviously appropriate to think in terms of human needs for water, it is
also appropriate to recognize that people also have demands for water as a commodity that
generates pleasure by utilizing it in various ways. As they become more affluent, poor people
are likely to choose to allocate more of their resources to satisfying not just their needs of the
body but also their "wants of the mind, most of them proceeding from imagination", perhaps
including some domestic end uses of water that might seem outlandish today.
If and when this broadening of domestic water use beyond the basic minimal level occurs,
an important implication is that planners will need to adopt a behavioral approach to the anal-
ysis and projection of urban demand, as opposed to the engineering/public health approach
that dominates the literature on water and poverty today. The behavioral approach focuses
not on how much water people need but rather on how much water they are willing to
pay for.
The difference between the engineering and behavioral perspectives is well illustrated by
the experience of the World Bank over the past fifteen years. Water planners in the Bank
originally thought that water and sanitation projects in developing countries were not viable
if they required households to pay more than 3-5% of their income for the project services,
because this would be more than they would be willing or could afford to pay, rendering
the projects infeasible (World Bank, 1975). It became evident from detailed household-level
studies sponsored by the Bank in the 1980s and 1990s that, in many developing countries,
some households spend considerably more than this on access to traditional, unimproved water
and sanitation. Some households purchase water from vendors at prices which can be much
higher than the cost of piped water. Where there is piped water supply, many households incur
expenses on installing storage capacity in the home to ensure that they have water when the
pipes run dry; others undertake a wide variety of practices to treat contaminated water in
their home to make it safe to drink. Moreover, some carefully designed contingent valuation
80 WM. Hanemann
surveys by Dale Whittington and his colleagues showed that some households have a WTP
for improved water supply that can exceed 3-5% of household income41.
The issue is not how much a household values access to water versus no access to water
at all! but, rather, how much it values a piped, public water supply relative to the existing
alternatives. In this context, it is interesting to review what is known about how the adoption
of water relative to other utility services varies with household income in developing countries.
Komives et al. (2003) have analyzed data on the percentage of households at different income
levels with four utility services: piped water, sewer, electricity, and telephone42. As monthly
household income increases from very low levels to US$ 300 per month, coverage of all of
these infrastructure services increases rapidly; above US$ 300 coverage increases at a slower
rate. However, what is most striking is that, for households in this sample, at all income levels,
more people have electricitY than piped water or sewer. Very few of the poorest households
have piped water or sewer, but almost a third of these households have electricity service. In
Kathmandu, Nepal, for example, all of the hous~holds surveyed had the option to connect to
electricity, water, and sewer; the majority of the very poor households chose electricity but
not water or sewer. As income increases, the percentage of households choosing water and
sewer increases, but the percentage with electricity is always higher. Almost no one, at any
income level, has only a piped water service; but many households have electricity and not
water. Thus, although most households would certainly like improved water and sanitation
services, it is not their most important development priority; given their limited resources,
many of them want electricity before an in-house piped water or sewer connection43.
In short, the fact that water itself is a necessity does not necessarily mean that people
prefer piped water over electricity service. Indeed, because water is a necessity, households
must already have some access to water supply. The question is thus how much they value an
improved supply. This will depend on how bad the existing water service is, and how much
better the improved service is expected to be44.
3.7 The heterogeneity of water
It is common to talk of the value of water as though it were a single, homogenous commodity.
This is obviously false: water has many dimensions besides just quantity. These include:
(a) location; (b) timing; (c) quality; and (d) variability/uncertainty. To a user, one liter of water
is not necessarily the same as another liter of water if it is available at a different location, at a
different point in time, with a different quality, or with a different probability of occurrence.
41 The subsequent experience when the Bank went ahead and implemented the projects has borne out
the predictions of the contingent valuation surveys quite well (Griffin et al., 1995). It would be wrong
to infer too much from the data on vended water purchases because the vast majority of households in
developing countries do not buy from vendors. This implies that their WTP for vended water is less than
the cost.
42 The data come from the World Bank's Living Standards Measurement Surveys covering over 55,000
households in 15 developing countries.
43 This does not mean that water and electricity are not complements -they often are. It simply suggests
that people with limited budgets who cannot afford both generally prefer to have electricity first and
piped water later.
44 It may also depend on the quantity of water involved -because of diminishing marginal utility, what
a household would be willing to pay per unit for the first 20 Lid per capita is not necessarily the same
as it would pay to go from 40 to 60 Lid per capita.
81
The economic conception of water
There are two ways to incorporate the multi-faceted nature of water in a formal economic
analysis. The first approach is simply to define different types of water as different commod-
ities. For example, the consumption of water in January is represented by Xl, that in February
is represented by X2, that in March by X3, etc. The consumer is then assumed to have a
utility function defined over monthly consumption throughout the year and also over other
commodities whose consumption is denoted by z (which can be a vector or a scalar), leading.
to the formulation:
U = U(X\, X2, ..., X\2, z)
(1)
The significance of this formulation is that it leads to separate demand functions for con-
sumption in each month. The demand for water in the ith month will be a function of the price
of water in that month, the prices of water in the other months (which mayor may not be
different), and the price ofz, as well as the consumer's income, Xi = hi (Pl,P2, ...,PN,Pz, y).
The differences between one month's demand function and that of another will reflect the dif-
ferent ways in which the two monthly consumptions enter the underlying utility function (1).
Given this approach, the annual demand for water is the aggregate of the 12 separate demand
functions for the individual months. There is no demand function for annual consumption
of water per se, except in the special case where the underlying utility function takes the
particular form:
u = U(XI, Xl, ..., XIZ,Z) = U (Lx;, z)
(2)
This is the only formulation that generates a well-defined demand function for aggregate
annual consumption, X == Lx;. Note, however, that the formulation in (2) implies that water
consumption in any month is a perfect substitute for consumption in any other month45. More
generally, if one discounts the difference between facets of water use and treats water as a
single, homogeneous commodity, this is equivalent to assuming that the different types are
all perfect substitutes for one another. It is an empirical question whether this is a plausible
assumption.
The alternative framework for analyzing differentiated commodities was provided in a
somewhat simple form by Lancaster (1966) and then broadened by Maler (1974), and is
known as the characteristics approach to consumer demand46. The Lancaster-Maler model
extends the utility model (1) by offering an explicit account of why the x's are viewed as
separate commodities, based on their specific characteristics. The notion is that there is a set
of characteristics or attributes associated with each commodity. Suppose there are K relevant
characteristics (attributes), and let qik denote the amount or level of the kth characteristic asso-
ciated with one unit of consumption of commodity i. The characteristics of each commodity
45 Two commodities are said to be a perfect substitutes in consumption if the consumer is willing to trade-
off one for the other at the same, fixed rate of exchange regardless of how much or how little is consumed;
in his eyes they can always be used in exactly the same way, with exactly the same outcome. When two
commodities are perfect substitutes, they have essentially the same value. The polar opposite of perfect
substitute is perfect complement. Two commodities are perfect complements if they are valued in fixed
proportions to one another; consequently, they will always be purchased together in fixed proportions.
In this case, no value is placed on increasing one of the items unless there is a corresponding increase in
the other; an old-fashioned example (in England) was tea and milk.
46 The application of Maler's work to the modeling of differentiated commodities was exposited in
Hanemann (1982).
82 WM. Hanemann
are taken as given by the consumer who is free to vary only the quantity of the commodity, Xi.
Thus, if the consumer wishes for more of the kth characteristic, she accomplishes this not by
changing the characteristics of any good, since these are fixed to her, but rather by switching her
consumption towards commodities with a high level of this characteristic (i.e. a high value of
qik); quality variation is accomplished through quantity variation. If there are N separate differ-
entiated commodities together with undifferentiated consumption, z, the utility function takes
the form:
, qN, Z)
(3)
U = U(x\, X2,
, XN, ql, Q2,
where qi = (qit, ..., qiK). The demand functions for commodities now depend on the attributes
as well as the prices, and take the form xi=hi (Pt, P2, ...,PN, qt, q2, ...,qN, pz, Y),
i = 1, ..., N. Thus, this formulation provides a framework for analyzing the effect of
characteristics/attributes on demand -it provides a .model of the demand for attributes
(i.e. for q).
';
I have focused so far on the multi-faceted nature of water with respect to consumer choice,
but this obviously applies to producer choice also. Both of the approaches described above
can be incorporated in a production function just as in a utility function. For example, the
production analog of (3) is a production function of the form:
qN, Z)
(4)
y =f(Xl, X2,
XN, q\, Q2,
where y is output, the x's are forms of water input, and z is a vector of non-water inputs.
Suppose, for example, that y is crop production, N = 2, Xl is the quantity of groundwater
pumped by a farmer, and X2 is the quantity of surface water delivered to the farmer by the
irrigation district in which he is located. Even if the price were the same, the experience in
California has been that most farmers do not consider groundwater and surface water to be
equally attractive. They find groundwater more convenient because they totally control its
supply and can obtain it at the flick of a switch while, with surface water, they have to wait
until the irrigation district is able to route the water to them through the canal system. On the
other hand, in some parts of California there are differences in water quality, with groundwater
being more saline than surface water. Thus, immediacy of access and salinity are two of the
attributes that enter the crop production function (4) in this case.
In addition to providing a framework for conceptualizing the demand for q, whatever this
may be, the Lancaster-Maler model also provides a framework for the economic valuation of q.
It can thus be used to measure water users' WTP for better availability of water, or less saline
water, or a more reliable water supply, or more generally water of one type versus another
(e.g. the premium on groundwater versus surface water, or on water at one location versus
another).
It should be emphasized that the attributes in q that differentiate one type of water from
another are by no means limited to the type of physical characteristics mentioned so far, such
as location, timing, quality, and reliability. Other aspects, such as how the water is provided,
can be the object of people's concern and the focus of their preference. With water, users often
care greatly about fairness in allocation or payment -what is known as procedural justice -
and this may differentiate one source of water from another in their eyes. The Lancaster-Maler
formulation permits one to incorporate in q such psychological or sociological attitudes within
an economic model of the demand for water, so that one can analyze how these attitudes might
The economic conception o/water 83
generate a different value for water when provided in a particular way or obtained from a
particular source47.
3.8 The fallacy of using average value
In most policy-related applications of economic valuation involving water, the relevant quan-
tity that needs to be known is the marginal value of water rather than the average or total value.
Precisely because water is a necessity of life, most people have some access to some amount of
water, and most policy interventions therefore involve changing the quantity and/or quality of
access rather than transforming the situation from no access to some access48. Ceteris paribus,
there is likely to 1?e some degree of diminishing marginal utility for consumers, and dimin-
ishing returns for producers; for this reason there can be a substantial difference between the
marginal value of an increase in water supply and its average value. This needs to be empha-
sized because researchers often use an estimate, of the a~erage value of water to measure the
benefits of a policy intervention; the resulting estimate is likely to be inaccurate.
An example comes from the recent Spanish National Hydrological Plan, which proposed
a major water transfer from the Ebro River to the Mediterranean coast, from Barcelona
in the North to Murcia in the South (MIMAM, 2000). Of the 1050 Mm3 to be transferred,
560 Mm3 was targeted for delivery to agricultural areas along the coast that have been relying
on depleting supplies of groundwater. The correct way to measure the benefit from an incre-
ment in water supply for farming in the receiving areas is to estimate the marginal value of
water (marginal net profit) in the agricultural uses that would go out of production without
the importation of project water. Instead, the economic assessment performed by the Spanish
government (MIMAM, 2002) valued the imported water using an estimate of the average value
of water in current uses, calculated as the simple ratio of aggregate farming profit in the area
divided by aggregate water use. There are two flaws in this approach: I) it interprets all profit
from farming in the area as exclusively a return to water; and 2) it treats the return to water as
constant regardless of the amount of water used.
Rather than just being a return to water, the profit from agriculture is likely to be a return
to the farmer's investment in land and other fixed assets, and also a return to the farmer's
own labor and his family's labor. And, rather than the average value of water being constant
in the receiving areas, there are several reasons to believe that the average value declines as
more water is supplied, causing the marginal value of water to be less the average value. With
varying land quality and the opportunity to grow different crops, farmers in the region are
likely to respond to any reduction in water supply by idling their least productive land and
discontinuing their least profitable crops. Furthermore, some users of groundwater also have
access to some surface water supply -in varying ways and to varying degrees, the supplies
of water for farmers in the receiving area are interconnected so that, within the area, water is
a somewhat fungible commodity. Since the imported water is one among several sources of
water for irrigation, the relevant demand function is the farmers' demand for project water,
not their demand for water overall. Because of the availability of substitutes, the agricultural
demand for project water is likely to be more elastic than the demand function for water overall.
47 What Frey & Stutzer (2005) refer to as procedural utility can be represented by this model.
48 By this, I am not implying that interventions are unimportant or trivial in their consequences.
WM. Hanemann
Therefore, the marginal value of the imported water in the region is likely to be substantially
less than the existing average value49.
The key difference between the two concepts is that the marginal value involves the deriva-
tive of a relationship, and to estimate this one needs a (formal or informal) model of how
water generates value. The average value, by contrast, can be estimated crudely by dividing
two quantities without any understanding of how they are related in reality, and without any
assurance that this ratio will remain constant. Consequently, the use of an estimate of average
value, and the assumption of its constancy, although common, are almost certainly a mistake.
3.9 The benefits afwater
There are numerous ways in which an increment in access to water might produce benefits,
whether to those who use the water directly or to others, Examples include: the use of water
for agricultural or industrial production, its use. for hydropower or for navigation, residential
use, flood control, water based recreation, or aquatic habitat. A key tool used by economists
in formalizing many of these benefits is the concept of a production function. A production
function is conceived as an empirical, causal relationship between the levels of inputs required
to produce an output, or an outcome, and the level of output or the outcome that results. One
example is the production function for an industrial or agricultural output as a function of water
and other inputs to the production process. Another example would be a health production
[unction relating inputs (including behavior patterns and levels of resource availability) to the
production (attainment) of health status outcomes. A third example would be an ecological
production function relating inputs and resource endowments to the production (attainment)
of ecosystem outcomes.
However, while the notion of a production function is undoubtedly useful as a conceptual
tool for organizing one's thought about these matters, it may work less well as a dependable
empirical construct. It may work better on a micro-scale (i.e. at a factory level) rather than
at the level of an entire regional economy, and it implies a notion of causation that may be
oversimplified. In practice, it can often turn out to be surprisingly difficult to measure a
production function on a regional scale or, more generally, to measure the specific increment
in benefits associated with an increment in water availability; these difficulties are clearly
evident in the literature on water and economic development50.
The notion that water supply contributes to economic growth and development seems
intuitively obvious. After all, it is known that many of the world's major cities owe their ori-
gin to their location along coasts or rivers where water-borne transportation was facilitated.
But, the relevant question is whether an increment in water availability now would generate
an increment in economic activity now, and how much. In the USA, federal water projects
49 If one were considering the total elimination of farming across a large portion of the irrigated acreage in
the receiving areas, the difference between the average and marginal value could be of little significance,
but in this particular case the change involves about 10% of the irrigated farmland and 9% of the water
supply in the receiving area. An important piece of evidence suggesting that the NHP was substantially
overstating the value of the project water is the fact that the cost of irrigation water in the receiving areas,
including the cost of groundwater pumping, is mainly in the range of 0.06-{}.27 E/m3, while the NHP
valued the imported water at 0.75 E/m3.
50 Similar difficulties are to be found in the literature on water and health.
The economic conception of water 85
have long been advocated for their claimed contribution to regional economic development.
However, the actual empirical evidence is less obvious and more negative. As Howe (1968)
noted, in industrial processes, water costs are a relatively small fraction of total production
costs even in water-intensive industries, and there are many examples of firms in such indus-
tries choosing to locate plants in water deficient areas because of market or non-water input
considerations.
In the USA in the late 1960s, there was a flurry of efforts to conduct formal, ex post statis-
tical analyses of the impact of water availability on economic development; the findings were
generally quite negative5!. The studies were motivated in part by Bower's (1964) hypothesis
that the availability of water at the intake end and/or the effluent end is not a major factor
in macro-location. decisions of industry relating to location in major geographical areas or
regions, such as river basins, but it can be a major determinant of micro-location decisions
relating to location within the region or basin. Ben-David (1966) found that employment
in the major water-intensive industries was not-significantly related to a measure of water
availability in a cross-section regression of USA states, but there was a significant positive
impact at the county level in a regression of counties within Pennsylvania. Howe (1968)
extended this analysis to include all the counties in the USA, and found the evidence for such
an effect to be extremely limited. Cox et al. (1971) examined counties in the Northeastern
USA in which large water projects had been constructed between 1948 and 1958, and found
no relationship between project size and economic growth over the period 1950-1960. By
contrast, Garrison & Paulson (1972) examined counties in the Tennessee Valley region and
found a significant micro-location relation between water-oriented manufacturing employ-
ment and a measure of water availability. At the macro-location level, Carson et al. (1973)
sampled counties in geographic sub-regions from all parts of the country, both rural and urban,
and found no significant relationship between federal water resource projects and population
growth. Cicchetti et al. (1975) extended this study using economic sub-regions as the unit of
analysis and found that variables representing federal investment in irrigation'facilities had
no significant impact on regional income and growth, and only a small and not convinc-
ingly significant impact on the value of farm output. There was some relationship between
economic growth and federal investments in flood control, hydropower and recreation, but
the coefficients were often unstable52. A similar study by Fullerton et al. (1975) of coun-
ties in seven western states found no relationship between water investment and economic
growth.
It seems clear that an investment in water supply does not automatically guarantee economic
growth. But, what conclusion can be drawn? Is there never an economic case for investing in
water supply?
I want to suggest that part of the problem arises from the inadequate concept of causation
that is being utilized by economists in conceptualizing the notion of a production function.
51 The focus of these studies was long-run employment and economic growth a:fter the completion of
project construction: it was recognized that there would be a short-term increase in employment during
the period of construction.
52 Some of these results and those of Cox et ai. (1971) raise the question of the direction of causation: do
federal investments in flood control, say, cause economic growth, or do growing areas use their political
clout to attract federal water investments? Walker & Williams (1982) suggest that the latter is the causal
connection.
WM. Hanernann
The philosophy literature makes a distinction between necessity and sufficiency: X could be
a necessary but not sufficient condition for Y to occur, or it could be a sufficient but not a
necessary condition. However, this distinction is generally ignored in the economic literature;
including both the theoretical and empirical analyses of the relationship between water and
growth. The production function as conventionally formulated as a relationship along the
lines ofY=! (X, Z, ...), implies that X, Z and the other factors on the right-haad side of the
equation are each a sufficient condition for producing Y: changing any individual element X
or Z is sufficient to induce a change in the value of Y. Similarly, the conventional forms of
regression equation used in the statistical literature imply that the regressors are each sufficient
conditions for a change in the dependent variable.
However, the true relationship may be different, and perhaps more complicated. For one
thihg, it seems plausible that having an adequate supply of water might be a necessary but not a
sufficient condition for economic growth. While water does not automatically generate growth,
it may be the case that areas which persist in lacking an adequate water supply (regardless
of whether or not they started out with adequate water) will not flourish economically. For
example, one can expect that people will eventually leave those areas and migrate to other
areas that do have an adequate water supply. Thus, lack of water could be a sufficient condition
for economic decline or, to put it another way, water may be a necessary but not sufficient for
economic growth. But, this is not a relationship that is captured in the existing formulations
of production functions and regression equations53.
In fact, the relationship between water and growth might be even more complicated. It
may be that there are multiple possible causal pathways, such that while there is some causal
linkage between water and growth, the linkage is sufficiently imprecise and variable that water
is neither a necessary nor a sufficient condition for growth. In effect, there is sometimes a
causal linkage, but not always. If this is so, it would require a new formalism to express this
type of relationship.
The statistical literature suggests at least two possible methods for estimating a relationship
that is heterogeneous in the manner just suggested. Both allow for multiple possible relation-
ships for the determination of Y, with a mechanism that is at least partly stochastic determining
the specific relationship that applies at any particular instance; within this framework, one
of the possible relationships can involve Z as a causal determinant of Y (e.g. as a sufficient
condition), while in others Z is not a determinant of Y. One statistical approach is the finite
mixture model (McLachlan & Peel, 2000), in which it is assumed that different observations
53 One way to operationalize the concept of necessity is through what is known as a switching regression.
The general structure of a switching regression, framed around a variable Z, is:
/i(Z,X,;{JI)+El
ifZ50
ifZ>O
h(Z,X2; .82) + f:;
Here Y is the dependent variable, Ji( ) represents a possibly non-linear functional relation, XI and
X2 are vectors of explanatory variables (which mayor may not be different), and .81 and .82 are the
corresponding coefficient vectors which are to be estimated. In the present context, 2 measures the level
of water supply and to represent it as essential one would have to impose the condition that fi (2, XI;
.81) + ~I = 0 whenever 2.5 o.
The economic conception of water 87
within the data belong to different groups to each of which a separate statistical relationship
applies. Here one estimates both the underlying relationships and also the group membership
probabilities for each observation in the data. The other approach is model averaging, of which
Bayesian model averaging (Raftery et al., 1997; Hoeting et al., 1999) is receiving considerable
attention. This approach assumes that there are many possible models that could apply to the
given data set taken as a whole. One proceeds by estimating each model under consideratio.n
using the entire data, and then averaging these models. In the Bayesian version, the weights
used for averaging are the posterior probabilities that each of the models is the correct one.
This approach produces confidence intervals for coefficient estimates that formally account
for model uncertainty in addition to estimation uncertainty.
The twin issues of model uncertainty and the need to reflect other causal relations besides
sufficiency deserve to receive greater attention in the empirical evaluation of the benefits
of water not only for economic development but also for health, fish habitat, and other
outcomes 54.
'\
THE
PROBLEM OF WATER FROM AN ECONOMIC PERSPECTIVE
4
It is commonly said that the problem of water is not one of economics but politics, not one of
physical shortage but governance. This is partly correct, but not entirely. The generic problem
of water is one of matching demand with supply, of ensuring that there is water of a suitable
quality at the right location and the right time, and at a cost that people can afford and are
willing to pay.
The difficulty in accomplishing this is partly institutional and certainly includes prob-
lems of governance. However, some of the problems of governance themselves have an
economic explanation. The omnipresence of fixed costs in surface water supply creates a
classic economic problem of cost allocation which has no satisfactory technical solution55.
The extraordinary capital intensity and longevity of surface water supply infrastructure, and
the predominance of economies of scale, create a need for collective action in the provision
and financing of water supply that simply does not arise with most other commodities. It has
been recognized since Olson (1965) that the provision of goods through collective action may
be flawed because of a failure of incentives.
Olson set out to challenge the optimistic notion that individuals with common interests
can necessarily be counted on to act voluntarily to further those common interests. The prob-
lem arises from harmful coincidences of rivalness/non-rivalness in benefits combined with
excludability/non-excludability in costs. Examples are free riding by members of the group
who withhold their individual contribution but can still benefit from the results of their col-
leagues' efforts, and rent seeking by individuals who seek to capture for themselves the benefits
54 A non-water-related health application of Bayesian model averaging is Koop & Tole (2004). A recent
book by political scientists has been devoted entirely to the empirical modeling of necessary conditions
(Goertz & Starr, 2003).
55 The most convincing solutions are rooted in bargaining theory and identify a cost allocation based
on relative bargaining strength; this is more a political than an economic approach (Young, 1986). An
early application of bargaining theory to water is Rogers (1969); for a recent survey see Carraro et af.
(2005).
88 WM. Hanemann
of collective action while throwing the cost on others56. Consequently, as Olson concluded:
"unless the number of individuals is quite small, or unless there is coercion or some other spe-
cial device to make individuals act in their common interest, rational self-interested individuals
will not act to achieve their common or group interest".
The challenge, thus, is to find a suitable non-coercive mechanism that motivates collective
action. This has become the subject of vast literature in economics, political science, sociology
and game theory. Success can be achieved, in principle, in several different ways. Cultural and
social norms shape preferences and may tilt the balance of individual choice in favor of col-
lective action. Homogeneity can help in some circumstances57. The nature of the institutional
arrangements is crucial58: if the rules are simple, transparent and devised locally, if moni-
toring and enforcement are relatively cheap, with graduated sanctions for non-compliance, if
low-cost and fair adjudication is available, then, ceteris paribus, successful collective action is
more likely. HQwever, the extent to which these conditions are met depends partly on people's
outlook and disposition, and partly on the phySical reality. For the reasons enumerated above,
water supply and sanitation are not always well situated in this regard.
In short, while there clearly are some distinctive emotive and symbolic features of water that
make the demand for water different, there are also some distinctive physical and economic
features which make the supply of water different and more complex than that of most other
goods. This fact has often been overlooked by economists and non-economists alike.
REFERENCES
Agrawal, A. (2002). Common Resources and Institutional Responsibility. In: E. Ostrom; T. Dietz;
N. Dolsak; P.C. Stern; S. Stonich & E.U. Weber (eds.), The Drama of the Commons. National Research
Council.
Anderson, L.D. (1980). The diffusion of technology in the 19th century American City: municipal water
supply investments. Ph. D. Dissertation. Northwestern University, Chicago, USA.
Armstrong, M.; Cowan, S. & Vickers, J. (1994). Economic analysis and British experience. MIT Press.
Baland, J.M. & Platteau, J.P. (1999). The ambiguous impact of inequality on local resource management.
World Development, 27(5): 773-788.
.;).56 Spiller & Savedoff (1999) emphasize the possibility of rent seeking by the government. They
argue the combination of economies of scale and massive sunk costs of investment encourage gov-
ernments in developing countries to act opportunistically against the private (and often foreign)
companies that operate the water supply system by pressuring them to lower prices, disallowing their
costs, requiring them to undertake special investments, controlling their purchasing or employment
patterns, or trying to restrict the movement or composition of capital. "All these are attempts by
politicians. ..to capture the rents associated with the company's sunk costs by administrative mea-
sures". Spiller & Savedoff argue that this leads to a low-level equilibrium of low prices and bad
service.
57 The ambiguous implications of inequality for the success of collective action are noted by Baland &
Platteau (1999) and by Bardhan & Dayton-Johnson (2002). A disequalizing shift in the access to common
resources has two effects which work in opposite directions. On the one hand, the people who benefit
from the change have a larger stake in the resource and, therefore, a greater incentive to act towards the
collective good. But, there is a corresponding disincentive to the other individuals whose share of the
outcome has been reduced.
58 This has been emphasized most influentially by Elinor Ostrom; see, for example, Ostrom (1990, 2003).
I draw here on Agrawal (2002) who provides a useful summary of the literature.
The economic conception o/water
Bardhan, P. & Dayton-Johnson, J. (2002). Unequal irrigators: heterogeneity and cornrnons management
in large-scale multivariate research. In: E. Ostrom; T. Dietz; N. Dolsak; P.C. Stern; S. Stonich &
E.U. Weber (eds.), The Drama of the Commons. National Research Council.
Barlow, M. & Clarke, T. (2002). Blue Gold: the fight to stop corporate theft of the Worlds Water. The
New Press.
Baumann, D.D. & Boland, J.J. (1998). The case for managing urban water. In: D.D. Bauman; J.J. Boland
& M. Hanemann (eds.), Urban water demand management and planning. McGraw Hill: 1-30.
Ben-David, S. (1966). Impact of water resource development on economic growth. In: Quarterly Progress
Report to the Corps of Engineers, July. Raleigh, North Carolina, USA.
Blake, N.M. (1956). Water for the cities. Syracuse University'Press.
Bowen, H.R. (1943). The interpretation of voting in the allocation of economic resources. Quarterly
Journal of Economics, 58(4), November: 27-48.
Bower, B. T. (1964). The location decision of industry and its relationship to water. Paper presented at
the Conference of the Committee on the Economics of Water Resources Development. December, San
Francisco, California, USA.
Bowley, M. (1973). Studies in the History of Economic Theory before 1870. Macmillan.
Carraro, C.; Marchiori, C. & Sgobbi, A. (2005). Application of negotiation theory to water issues.
Fondazione Eni Enrico Mattei. Working Paper 65, May.
Carson, J.M.; Rivkin, G.W. & Rivkin, D.M. (1973). Community Growth and Water Resources Policy.
Praeger.
Cicchetti, C.J.; Smith, \-:K. & Carson, J. (1975). An Economic Analysis of Water Resource Investments
and Regional Economic Growth. Water Resources Research, 11(1), February: 1- 6.
Ciriacy- Wantrup, S. \-: (1947). Capital returns from soil-conservation practices. Journal of Farm
Economics, 29: 1181-1196.
Clawson, M. (1959). Methods of measuring the demand for and value of outdoor recreation. Reprint
no. 10. Resources for the Future, Inc., Washington, USA.
Cox, P.T.; Grover, W.W. & Siskin, B. (1971). Effect of water resource investment on economic growth.
Water Resources Research, February: 32-38.
Davis, R.K. (1963). Recreation planning as an economic problem. Natural Resources Journal, 3(2):
239-249.
Dupuit, J. (1844). De la Mesure de l'Utilite des Travaux Publiques.Annales des Ponts et Chaussees, 2nd
Series, 8. Reprinted in translation (1952) as: "On the Measurement of the Utility of Public Works",
in: International Economic Papers, 2: 83-110.
Erie, S.P. & Joassart-Marcelli, P. (2000). Unraveling Southern California's Water/Growth Nexus:
Metropolitan Water District Policies and Subsidies for Suburban Development, 1928-1996. California
Western Law Review, 36, Spring: 267-288.
Frey, B.S. & Stutzer, A. (2005). Beyond outcomes: measuring procedural utility. Oxford Economic
Papers, 57(1), January: 90-111.
Fullerton, H.; Lewis, W.C.; Anderson, J.C.; Keith, J.E. & Willis, R. (1975). Regional development: an
econometric study of the role of\\'ater development in ~ffectuating population and income changes.
Report PRRBEO89-1. Utah Water Research Laboratory, Utah State University, Logan, USA.
Garrison, C.B. & Paulson, A.S. (1972). Effect of water availability on manufacturing employment in the
Tennessee Valley region. Water Resources Research, April: 301-307.
General Accounting Office (1996). Information on allocation and repayment of costs of constructing
water projects. Bureau of Reclamation (GAO/RCED-96-109).
Gleick, P.H. (1999). The Human Right to Water. Water Policy, 1(5): 487-503.
Glennon, R. (2004). The Price of Water. Journal of Land. Resources & Environmental Law, 24(2):
337-342.
Goertz, G. & Starr, H. (2003). Necessary Conditions. Rowman & Littlefield.
Gordon, B. (1975). Economic anao'sis before Adam Smith. Macmillan Press.
Griffin, C.; Briscoe, J.; Singh, B.; Ramsasubban, R. & Bhatia, R. (1995). Contingent Valuation and
actual behavior: predicting connections to new water systems in the State of Kerala, India. The World
Bank Economic Review, 9: 373-395.
Hamlin, C. (2000). Waters or Water? -Master Narratives in Water History and their implications for
Contemporary Water Policy. Watet: Policy, 2: 313-325.
90 WM. Hanemann
Hammond,
R.J. (1960).
Benefit-Cost
analysis and water pollution
control.
Food Research -Institute,
Stanford
University
Press.
Hanemann, W.M. (1982). Quality and Demand Analysis. In: G.C. Rausser (ed), New Directions in
Econometric Modeling and Forecasting in US. Agriculture. ElsevierINorth Holland, Inc., New York,
USA: 55-98.
Hanemann, WM. (1991). Willingness to Pay and Willingness to Accept: how much can they differ?
American Economic Review, 81(3): 635-{j47.
Hanemann, WM. (2002). The Central Arizona Project. Department of Agricultural
and Resource
Economics. Working Paper 937, October. University o~California, Berkeley, USA.
Henderson, A. (1941). Consumer's surplus and the compensating variation. Review of Economic Studies,
..
8:117-121.
Hicks, J.R. (1939). Value and Capital. Oxford University Press.
Hicks, J.R. (1941). The rehabilitation of consumer's surplus. Review of Economic Studies, 8: 108-116.
Hicks, J.R. (1942). Consumer's surplus and index numbers. Review of Economic Studies, 9: 126-137.
Hicks, J.R. (1943). The four consumer's surpluses. Review of Economic Studies, II: 131-141.
Hicks, J.R. (1946). The generalized consumer's surpluses. Review of Economic Studies, 15: 27-33.
Hoeting, J.A.; Madigan, D.; Raftery, A.E. & Volinsky, C. T. (1999). Bayesian model averaging: a tutorial.
Statistical Science, 14(2): 382-417.
Howe, C.W (1968). Water resources and regional economic growth in the United States, 1950-1960.
Southern Economic Journal, 34(4), April: 477-489.
Hurwicz, L. & Uzawa, H. (1971). On the integrability of demand functions. In: J.S. Chipman;
L. Hurwicz; M.K. Richter & H.F. Sonnenschein (eds.), Preferences, utility and demand. Harcourt
Brace Jovanovich, New York, USA: 114-148.
Komives, K.; Whittington, D. & Wu, X. (2003). Infrastructure coverage and the poor: a global perspective.
In: P.J. Brook & T.C. Irwin (eds.), Infrastructure for poor people: public policy for private provision.
The World Bank: 77-123.
Koop, G. & Tole, L. (2004). Measuring the health effects of air pollution: to what extent can we really
say that people are dying from bad air? Journal of Environmental Economics and Management, 47:
30-54.
Krutilla, J. v: (1967). Conservation reconsidered. American Economic Review, 57: 787-796.
Kupel, D.E. (2003). Fuel for growth: water and Arizona s urban environment. University of Arizona
Press, USA.
Lancaster,
K.
(1966).
A
new
approach
to
consumer
theory.
Journal
of Political
Economy,
74:
132-157.
Lockwood, M. (1996). Non-compensatory preference structures in non-market valuation of natural area
policy. Australian Journal of Agricultural Economics, 40(2): 85-101.
Maler, K.G. (1971). A method of estimating social benefits from pollution control. Swedish Journal of
Economics, 73: 121-133.
Maler,
K.G.
(1974).
Environmental
Economics:
a theoretical
enquiry.
Johns
Hopkins
University
Press
for Resources
for the Future.
Marshall, A. (1879). The pure theory of (domestic) values. London School of Economics, UK.
Marshall, A. (1890). Principles of Economics. 1st edition. Macmillan, London, UK.
McLachlan, G. & Peel, D. (2000). Finite .inL-rture models. John Wiley.
MIMAM
(Ministerio
de Medio Ambiente
[Ministry
of Environment
of Spain]) (2000). Plan Hidrologico
Nacional:
analisis economicos.
September.
Madrid,
Spain.
MIMAM
(Ministerio
de Medio
Ambiente
[Ministry
of Environment
of Spain])
(2002).
Strategic
assessment
of the National
Hydrological
Plan. Summary
Document.
January. Madrid,
Spain.
Niehans, J. (1990).
A History
of Economic
Theory:
classic
contributions,
1720-1980.
Johns Hopkins
University
Press.
Olson,
M.
(1965).
11Ie Logic
of Collective
Action:
Public
Goods
and
the Theory
of Groups.
Harvard
University
Press.
Ostrom, E. (1990). Governing the Commons. Cambridge University Press.
Ostrom, E. (2003). How types of goods and property rights jointly affect collective action. Journal of
11Ieoretical Politics, 15(3): 239-270.
91
The
economic conception of watel
Postel, S. & Vickers, A. (2004). Boosting Water Productivity. In: Worldwatch Institute, State of the World
2004. W.W. Norton.
Pribram, K.H. (1983).A History of Economic Reasoning. Johns Hopkins University Press.
Raftery, A.E.; Madigan, D. & Hoeting, J.A. (1997). Bayesian model averaging for linear regression
models. Journal of the A,!,erican Statistical Association, 92, March: 179-19 I.
Ridker, R.G. (1967). Economic costs of air pollution. Praeger.
Rogers, P. (1969). A Game Theory Approach to the Problems of International River Basins. Water
Resources Research, 5(4), August.
Samuelson, P.A. (1954). The pure theory of public expenditure. Review of Economics and Statistics,
36(4), November: 387-389.
Schumpeter, J.A. (1954). History of Economic Analysis. Oxford University Press.
Shiva, v: (2002). Water wars: privatization, pollution, and profit. South End Press.
Spiller, P. T. & Savedoff, W.D. (1999). Government opportunism and the provision of water. In: P. T. Spiller
& W.D. Savedoff (eds.), Spilled water: institutional water. Institutional Commitment in the Provision
of Water Services. Inter-American Development Bank. Washington, USA.
Tarr, J.A. (1979). The separate vs. combined sewer problem: a case study in urban technology design
choice. Journal of Urban History, 5, May.
Trice, A.H. & Wood, S.E. (1958). Measurement of recreation benefits. Land Economics, XXXIV: August.
UNESCO (2000). World Day for Water 2000. See: http://www.unesco.org/science/waterday2000/
Variability.htm
Walker, R.A. & Williams, M.J. (1982). Water from power: water supply and regional growth in the Santa
Clara Valley. Economic Geography, April: 95-119.
Ward, F.A. & Michelsen, A. (2002). The economic value of water in Agriculture: concepts and policy
applications. Water Policy, 4: 423-446.
Wegge, T.C.; Hanemann, W.M. & Loomis, J. (1996). Comparing benefits and costs of water resource
allocation policies for California's Mono Basin. Economic institutions and increasing water scarcity.
In: D.C. Hall (ed.), Advances in the economics of environmental resources. Volume 1: Marginal cost
rate design and wholesale water markets. JAI Press, Greenwich, Connecticut, USA: 11-30.
Weisbrod, B.A. (1964). Collective consumption services of individual-consumption goods. Quarterly
Journal of Economics, 78(3): 471-477.
Willig, R.D. (1976). Consumer's surplus without apology. American Economic Review, 66: 589-597.
Wittfogel, K.A. (1957). Oriental despotism: a comparative study of total power. Random House Inc.
World Bank (1975). Issues in Village Water Supply. Washington, D.C., USA.
Worster, D. (1986). Rivers of Empire: water; aridity, and the growth of the American West. Random
House.
Young, H.P. (ed.) (1986). Cost allocation: methods. principles. applications. Elsevier Science Publishers.
Young, R.A. & Haveman, R.H. (1985). Economics of water resources: a survey. In: A.V: Kneese &
J.L. Sweeney (eds.), Handbook of Natural Resource Energy Economics, Vol. II. Elsevier Science
Publishers: 465-529.
