The development of water rights in Colorado: an empirical analysis.
Penn, David A. ; Zietz, Joachim
I. Introduction
An important insight offered by the property rights economics
literature is that institutions, or the rules of the game, affect how
well markets function (Pejovich 1990). Properly functioning markets
require well-defined and enforceable property rights. When property
rights are poorly defined or weakly enforced, market incentives that
encourage entrepreneurial activity, innovation and invention, creative
activity, and hard work will diminish in effectiveness (Demsetz 1967).
Property rights do not spring up from the ground well defined and
enforceable. Rather. they change over time due, in part, to changing
economic circumstances (Demsetz 1967; Anderson 1982; Pejovich 1990).
People not only pursue their self-interest within the rules, they also
allocate resources to changing the rules of the game to their own
benefit (Anderson 1982, p. 761). In fact, establishing and protecting
property rights can be considered a productive activity toward which
resources will be devoted.
The manner in which certain property rights emerge and change over
time is the focus of this study. The origins and evolution of Western
water law offer an important example of how property rights change in
response to changing economic incentives. The paper focuses on the
Colorado experience largely due to the fact that Colorado was one of the
first states to establish a system of water rights based exclusively on
the system of prior appropriation. Many of the developments in water
rights in the rest of the Western United States derive in one way or
another from the Colorado system.
Our study offers the first quantitative evidence that links
economic incentives with water rights defining and enforcing activity.
Our model suggests that water claimants will more carefully define their
rights to water when either the demand for water increases or the supply
of water decreases. The results also indicate that not all empirical
facts easily match a theory of property rights evolution, as developed,
for example, in Anderson (1982). To understand historical developments
it appears useful to also incorporate the predictions of the rent
seeking literature (e.g., Krueger 1974), especially as it applies to
agriculture (Honma and Hayami 1986; Anderson and Hayami 1987; Gardner
1987).
The manner in which water rights evolve and adapt to increasing
demand for water and frequent periods of scarcity is at least as
relevant today as it was in the late 19th century. Rapidly rising demand
coupled with periodic severe droughts exert great pressure on
today's water allocation institutions, in the U.S. as in other
countries. Competition for water has multiplied for a number of reasons
including rapid urban population growth, protection of instream water
rights, competition for water among states, and the recognition of
Native American water rights.
The remainder of the paper is organized as follows. The next
section provides a brief account of water rights development in
Colorado, offering qualitative evidence of how the rules of the game
evolve with changing economic circumstances. This is followed by an
outline of a simple model for water rights development. The data are
presented next, followed with a presentation of empirical model
estimates. The paper ends with a brief summary and some conclusions.
II. Historical and Legal Background
Agricultural development, irrigation, and the evolution of water
rights law were inseparably linked during the late 19th and early 20th
centuries in Colorado. The next section discusses the historical
development of irrigation in Colorado, followed with an analysis of the
formation of legal precedents.
1. Farming development and irrigation
Farmers migrating to Colorado found a large percentage of sunny
days and low humidity, both very favorable for crop production given
adequate moisture. But with annual precipitation of just 12 to 15 inches
in the plains east of the mountains, irrigation was employed where water
was available (Census 1920). Consequently, farming development spread
out mostly adjacent to the three major watercourses located in the north
(Platte River), the south (Rio Grande River), and the southeast
(Arkansas River). Farming by irrigation produced much more output per
acre compared with farming in the more humid areas of the country. But
irrigated farming also entailed much higher capital expenditures and
required more attention by the farmer to matters related to securing and
protecting rights to water (Taylor 1907).
Early irrigation in Colorado produced vegetables for the miners;
these ditches were small, narrow structures that irrigated just a few
acres. Lands nearest the streams were cultivated first; as these lands
became fully claimed, additional irrigation development required larger
and more elaborate irrigation works that could carry water to lands
farther from the streambed (Dunbar 1983). The diversion of water for use
on lands not adjacent to the watercourse arose from the necessities
presented by the dry and arid environment.
Following the Civil War, westward expansion of the railroad network
transformed local agriculture from small-scale production for local
markets to large-scale, more specialized, enterprises producing largely
for markets in the East and overseas. By providing connections to large
market centers in Chicago and New York, rail expansion created strong
incentives for farmers to produce more than could be consumed by local
markets. Access to New York for farmers in the West also provided access
to large European markets via transatlantic shipments.
Falling rail and shipping costs quickened the pace of market
integration. From the 1880s to 1910, for example, the cost of shipping a
bushel of wheat from Chicago to New York fell from 8.6 cents to 5.4
cents, and the cost of transatlantic shipment dropped by nearly
two-thirds. Declining transportation costs caused grain prices in
Chicago, New York, and Liverpool to move much closer together. By World
War I, the price of wheat was nearly the same at all three locations.
Market integration greatly increased access to large markets for
Colorado farmers, but also made them more dependent on economic and
weather conditions far from home (Atack, Bateman, and Parker 2000).
Strong demand, favorable local climate, and good soil enhanced by
irrigation created ideal conditions for crop production and livestock.
Nearly seven in ten dollars from irrigated cultivation in Colorado
during the late nineteenth and early twentieth centuries can be
attributed to five crops: alfalfa, potatoes, sugar beets, fruit (mostly
apples), and wheat. By 1909, Colorado produced threefifths of the entire
western irrigated crop of sugar beets and thirty-five percent of the
irrigated potatoes (Census 1910). The production of irrigated alfalfa
was important as a source of livestock feed during the long and bitter
winter seasons when forage is scarce.
The need to spread fixed costs over more acres of production
combined with increasing market demand for crops and livestock created
strong incentives for the development of larger farms. Consequently,
farmers took on more debt to purchase additional land and new or
expanded irrigation works (Boyd 1897).
Early irrigation works in the 1860s were often simple, consisting
of a crude stream diversion and a short ditch. Soon, however, farmers
discovered that they must combine to build and maintain expensive
irrigation works; these organized efforts were termed 'irrigation
enterprises' by the Census Bureau. Between 1860 and 1880, nearly
1,600 new irrigation enterprises came into existence. The number of
irrigation enterprises more than doubled in the 1880s so that by 1890,
58.9 percent of all farms and 19.4 percent all land in farms were
irrigated. By 1900, 71.3 percent of all farms in the state were
irrigated (Census 1910).
Capital expenditures for new irrigation enterprises in Colorado
averaged $11 per acre in the 1870s and 1880s, climbing to $19 per acre
in the 1890s, $49 per acre between 1900 and 1910 and $82 per acre from
1910 to 1914, all in 1900 dollars. Once the easily reclaimed lands were
developed, providing irrigation to additional lands required moving
water a greater distance from the watercourse to higher elevations,
requiring much greater capital expenditure per acre. In addition,
capital spending in the 1890s and later had much to do with building
storage reservoirs, improving existing ditches, and building new ditches
within existing irrigation enterprises.
After 1900 the number of irrigated farms continued to rise, but
more slowly than the number of farms with no irrigation, reducing the
proportion of farms under irrigation to 56 percent in 1910 and 48
percent in 1920. The rapid growth of dryland farming was partly a
consequence of the escalating cost of irrigation and partly due to the
development of techniques for moisture-saving cultivation,
drought-resistant varieties of crops, and the introduction of the
gasoline tractor in the early twentieth century; these influences
combined to greatly improve the ability of farmers to grow crops without
expensive irrigation (Hundley 1988).
Farmers enjoyed favorably wet weather in the early 1880s; from that
time until World War I, the weather alternated, with one or two wet
years followed with one or two dry years. Below average precipitation
and low stream flow caused crop output to fall from the previous year in
1890, 1893, 1896, 1899, 1902, and so on. Due to the depression beginning
in 1890, world markets for grain were in surplus due to high production
in Canada, South America, and Australia. Falling grain prices in
conjunction with very dry weather caused widespread economic hardship;
real crop prices had dropped 21 percent in Colorado.
A lesson learned from the dry years was that more irrigation is
needed to smooth out the persistent spells of low precipitation and more
reservoirs are needed to smooth out high and low periods of stream flow.
Consequently, the number of storage dams doubled in the 1880s and
continued to be built at a robust pace until 1915. Between 1889 and 1899
the number of acres under irrigation increased 81 percent, mostly in
wheat and other cereals and alfalfa hay and Colorado surpassed
California in the number of irrigated acres and was second largest to
California in the value of irrigated crops (Census 1910 and 1920).
Between 1910 and 1920, production on irrigated lands rose partly
due to increased demand domestically and partly because of higher demand
from Europe to feed the hungry population of warring nations. Corn,
wheat, alfalfa, and sugar beets all experienced significantly greater
output, although the greatest impact on farmers was the much higher
prices for crops during this period.
2. Development of legal precedents
Three overlapping but distinct periods of water rights litigation
can be identified. In the earliest period, from about 1872 to the early
1890s, the majority of court cases dealt with upholding and refining the
doctrine of prior appropriation. During this period the riparian water
rights system in place was set aside in favor of doctrine prior
appropriation, a system more suited to the dry climate. In the second
period, from about 1892 through about 1904, cases involving irrigation
enterprises were prominent. These cases dealt with defining the rights
of irrigation cooperatives and partnerships relative to the irrigators
and other irrigation enterprises. Finally, cases regarding storage
reservoirs, water rights transfers, and changes in the point of
diversion were prominent in the third period beginning in about 1905.
Figure 1 shows a spurt of litigation from 1890 to 1892, mostly
involving irrigation enterprises, with the volume of court cases
dropping until 1905 when a second surge of litigation occurs. One might
expect in the absence of well-defined rules during the earliest period
of settlement, the courts provided the most attractive means of dispute
resolution (Khan 2000). As the rules of water usage became more clearly
established over time, the number of disputed issues will decline.
Legal precedents established in a body of case law provide courts
an important method of using an earlier decision to decide a similar
case in the same way (Landes and Posner 1976). Precedents can be viewed
as a stock of legal knowledge that, over time, improves the
predictability of the law (Miceli 2004). An example can be found in a
recent study of litigation on the Australian frontier. Khan (2000) shows
an initial surge in litigation was followed with a long-term declining
trend in the late 1800s. She argues that the as the stock of legal
capital grows, standards became more routine and predictable, reducing
the probability of future litigation on similar issues.
[FIGURE 1 OMITTED]
The earliest important case was decided by the Colorado Supreme
Court in 1872. In Yunker v. Nichols, (1) the court first recognized an
absolute need for diverting water from the watercourse. Reversing the
decision of the lower court, the court held that the dry Colorado
climate, the right to divert water from the watercourse and the right to
convey water through land owned by others are necessities. In the
opinion of Chief Justice Hallett,
In a dry and thirsty land it is necessary to
divert the water of streams from their natural
channels, in order to obtain the fruits of the
soil. The value and usefulness of agricultural
lands, in this territory, depend on the supply
of water for irrigation, and this can only be
obtained by constructing artificial channels
through which it may flow over adjacent
lands (p. 552).
Priority of water rights was specifically recognized later in the
Colorado constitution in 1876, following the customs and rules laid down
by the miners nearly two decades earlier. In 1878 the Colorado Supreme
Court again re-affirmed priority of rights to water in Schilling v.
Rominger. (2)
Legal challenges from riparian water users continued to arise,
however. In 1882 the Colorado Supreme Court issued a very clear
statement of the right of diversion and right of priority. In Coffin v.
Left Hand Ditch, (3) the court found that without the rights to divert
and transport water, hallmarks of the system of prior appropriation,
farmers would have little incentive to grow crops:
... vast expenditures of time and money
have been made in reclaiming and fertilizing
by irrigation portions of our unproductive
territory. Houses have been built, and permanent
improvements made; the soil has been
cultivated, and thousands of acres have been
rendered immensely valuable, with the understanding
that appropriations of water
would be protected. Deny the doctrine of priority
of appropriation, and a great part of the
value of all this property is at once destroyed
(pp. 448-449).
The court ruled that the common law doctrine of riparian rights is
inapplicable in Colorado due to the arid climate and the necessity of
diverting water from the watercourse. The court also held that the first
person to appropriate, or divert, the water and put to a beneficial use
has the prior right.
Although the Coffin decision definitively repudiated riparian
rights, legal challenges to the doctrine of prior appropriation still
occurred but with less frequency. From 1882 to 1909 the court ruled
against riparian rights in five separate cases. In the last case,
Sternberger v. Seaton, (4) an exasperated court argued that to recognize
a riparian right "... would require the reversal of decisions of
the court, tearing up the statute laws, and nullification of provisions
of the constitution." Eventually the rights of diversion and prior
appropriation were accepted as the undisputed law, but only after a
period of challenge after challenge by those who felt threatened.
The next period of litigation involved irrigation enterprises and
corporate canal companies. As lands nearest the rivers and streams were
settled, the development of additional cropland required transporting
water to lands much farther from the watercourse. Much more capital
investment was needed to build these new, more expensive irrigation
works, creating strong incentives for the development of irrigation
enterprises that could develop and maintain diversion works and ditches.
The most prevalent form of these enterprises consisted of cooperative
ventures, partnerships between farmers, and corporate-owned enterprises.
As can be imagined, the ambiguities of the rights of irrigation
enterprises vis-a-vis the rights of irrigators generated a great deal of
litigation* During the 1890s, the Supreme Court and appellate courts in
Colorado decided 21 cases dealing with irrigation enterprises, compared
with just three cases in the previous decade. In the Wheeler (5) case,
for example, the Colorado Supreme Court ruled that the rights to water
belong to the irrigator, and the irrigation enterprise could legally
impose fees related to canal upkeep and transportation costs, but could
not charge a perpetual fee for the use of the water* The number of court
cases that dealt with the irrigation enterprises peaked during the 1890s
and held steady in the 1900s and 1910s.
Most of the legal issues relating to prior appropriations and
irrigation enterprises had been settled by the early 1900s. Farming
growth, however, continued unabated, causing incentives for irrigators
to search for new ways to increase the amount of water available for
growing crops. As the density of diversion facilities increased over
time, irrigators found that they could more efficiently use their water
by changing the point of diversion from one place to another. Other
irrigators found profit in selling water rights to rapidly growing
municipalities, thus raising legal issues regarding the transferability
of water rights.
In addition, irrigating enterprises built expensive reservoirs that
allowed storage of waters during the spring when stream flow typically
is swollen by melting snow cover in the mountains* Stored water could be
used to irrigate crops in the late summer when stream flow is much
smaller* The first storage reservoirs began to appear in the mid-1880s
in the Cache la Poudre Valley (Dunbar 1983).
The water-supply enhancing innovations of water storage, water
rights transfers, and changes in the point of diversion raised a new set
of issues for the courts to sort out. The volume of litigation on these
issues increased from just four cases in the 1890s to sixteen cases
1900-1909 and seventeen cases 1910-1919.
The total volume of water rights litigation peaked in about 1906,
then gradually declined. Several factors may have contributed to the
decline in water rights defining and enforcing activity. First, as
mentioned earlier, the emergence of case law precedents reduced the need
for the courts to address the same issues over and over. Second, dryland
farming increased greatly during the 1900s in response to the rapidly
escalating capital costs of irrigation. Dryland farms add to crop
production without increasing the demand for water. Third, increasing
storage of water during the wet season for use during the growing season
helped mitigate periods of water shortage. Court cases recognizing the
right to build reservoirs on the stream bed and financing generated by
the federal Reclamation Act, created incentives for new reservoirs that
reduced the effects of droughts on the availability of water for
irrigation (Fox 1918). As noted by Tarlock (2001), dams reduced the need
to enforce water rights in the courts, thus lowering the number of
lawsuits undertaken.
The history of Colorado water rights development demonstrates how
the water rights defining and enforcing activity changed over time as
economic circumstances changed. Early irrigators defended the rule of
diversion and priority of right in order to protect their investments in
the farm enterprise. Increased demand for food fostered large capital
investments in ditches and canals, resulting in the emergence of
irrigation enterprises and canal corporations, raising new issues for
property rights defining and enforcing activity. Increases in water
storage, dryland farming, and changes in the point of diversion
characterize the final period of water rights litigation.
III. Theoretical Background
Anderson (1982) offers testable propositions regarding property
rights defining and enforcing activity. First, higher market values or
greater scarcity will cause individuals to strengthen their claims to
resources. Second, an increase in the probability of losing an asset
will increase property rights enforcing activity. In this section, we
outline how these propositions can be examined empirically in the
context of Colorado water rights litigation.
In the present model, water rights defining and enforcing activity
depends on the value of water: the higher its value, the greater the
benefits from additional defining and enforcing activity. The value of
water depends at a minimum on five factors, (a) prices of irrigated farm
crops, (b) production costs, (c) farm production per acre or
productivity, (d) the number of acres under irrigation and (e) the
quantity of water available in the stream.
Following Anderson (1982) the value of water will rise if the
output price of irrigated crops goes up. This assumes that the price
elasticity of supply is non-negligible. Conversely, a rise in production
cost can be expected to reduce production and, hence, the value of
water. The value of water is anticipated to increase as farm
productivity goes up, as measured by higher output per acre. The demand
for water and, hence, its value will rise if more land is irrigated. As
more water is available in rivers, the value of water and, hence, water
rights enforcement activity should decrease.
An important implication of this model inspired by Anderson (1982)
is that irrigators will increase their efforts to protect and define
their rights to water when their demand for water increases or when the
supply of water decreases.
IV. Data
The number of court cases (cases) dealing with water rights is
chosen as a measure of water rights defining and enforcing activity
(Figure 1). This is the variable to be explained by the model. Court
cases dealing with water rights at the Supreme Court and appellate
levels are collected from several editions of the American Digest, a
publication that lists headnotes of court cases by legal category. Cases
are selected from the category "Water and Water Courses" from
four editions of the Digest. Case headnotes are scanned to ensure that
the case is related, in general, to the protection or definition of the
right to water. Some cases and categories of cases are excluded. Cases
involving damages to land due to water seepage from canals, for example,
do not involve the value of water and so are excluded. Subcategories
such as "Bed and Banks" are also excluded. The types of cases
included range from issues such as rights of way to damages due to loss
of flow, and from appurtenance to the validity of the system of prior
rights. (6)
The regressors to explain the number of water rights court cases
per year are identified in Table 1. The list of variables includes more
variables than are used for the preferred models presented in the
results section.
The demand for water and its value is directly tied to the level of
irrigated farming. Two alternative variables are employed to account for
this relationship: the number of irrigated farms (ifarms) in Colorado
and the acreage under cultivation by irrigated farms (acresifarms). The
variable ifarms is estimated using decennial census figures. (7) Annual
figures for ifarms and acresifarms are assumed to grow at the same rate
within a given decade, determined by the annualized census-to-census
growth rate.
The output price of irrigated crops is represented by the real
production-weighted output price of five key crops (pout) in the early
history of Colorado. An estimated consumer price index from the
Historical Statistics of the United States is used to deflate crop
values per bushel. As seen in Figure 2, real prices declined from the
1880s to the 1890s, probably due to the depression and deflationary
pressures during the mid 1890s. During the period 1904-11, however,
output prices soared, probably reflecting rising world demand for crops.
Crop values in Colorado increased 27 percent from their level of
1882-89. Real output prices jumped again during World War I, although to
a much larger degree than before. By far, the largest proportion of the
crops represented by the output price index was grown on irrigated
farms. In value terms, according to the 1890 census, irrigated crops
accounted for 89 percent of all crops grown in Colorado. (8)
The variable output per acre (outacre) represents farm
productivity. It is calculated from data for five key crops. (9) As
discussed above, most of the crops grown during this period were
irrigated. In 1909, for example, 96 percent of the alfalfa production
was irrigated as was 65 percent of the wheat produced, 71 percent of the
potato production, and 99 percent of the sugar beet production (Census
1920).
Stream flow data (flow) are taken from U.S. Geological Survey
(1954, 1958) publications. Criteria for selection of rivers and
recording stations include the completeness of early records and the
location of the river. Based on these criteria, two rivers are selected:
the Cache la Poudre, recorded at Ft. Collins, and the Arkansas River,
recorded at Canyon City. Much of the farmland irrigated in Colorado
depends on the flow of these two rivers. An index of stream flow is used
since it reflects variations of flow around a base level without regard
to the nominal quantities. The flow index is constructed from the mean
of the nominal flows of the Cache la Poudre and Arkansas Rivers, with
the base set at the mean level of the combined flows.
Farmers must be concerned not only with the value of crops or
revenue but also with the cost of production. Unfortunately, direct cost
data do not exist for Colorado farms. However, some variables can be
identified that are likely to capture key cost components. The railway
freight cost of moving wheat from Chicago to New York is such a cost
component. If it can be taken as a proxy for transportation costs in
general, it would be relevant for farmers as an input cost. To the
extent that farmers rely on credit for their operations, the commercial
paper rate could be considered a measure of short-term credit cost.
[FIGURE 2 OMITTED]
V. Estimation Results
The time series properties of the data do not allow for
cointegation methods. In particular, as is evident from Figure 1, the
dependent variable does not follow an obvious trend. As a result,
ordinary least squares is employed to identify the determinants of water
rights litigation. The estimation results are summarized in Table 2.
Particular emphasis is placed on checking that the empirical models
capture the data generating process. (10) To document model fit,
probability values for a number of statistical specification tests are
provided. No statistical problems are evident for any of the models at
the five percent level. Simultaneous equations problems are excluded
because all regressor variables enter with a lag of at least one year.
In fact, most variables have a significantly longer lag structure
associated with it. Two models are estimated in levels and two in
log-linear format. There are few differences in the statistical adequacy
tests to prefer one specification over the other. The main reason for
including both specifications is to allow for alternative economic
interpretations of the estimated coefficients. The estimates in levels
(Models 1 and 2) provide for marginal effects, the log-linear estimates
(Models 3 and 4) for elasticities. (11)
As predicted by the theoretical considerations of Anderson (1982),
stream flow as measured by variable flow has a negative impact on the
incidence of water rights litigation, while output per acre (outacre)
and the acreage planted by irrigated farms (acresifarms) have positive
coefficients (Models 1 and 3). As an alternative to the variable
acresifarms the number of irrigated farms (ifarms) is used in Models 2
and 4. Both variables are statistically highly significant. The
elasticity of ifarms is close to unity, that of acresifarms is somewhat
less. Water rights litigation is also proportional, with an elasticity
of one, to farm productivity (outacre). These results are consistent
with the predictions provided in the theoretical background. This also
applies to the negative unitary elasticity of water rights cases with
respect to stream flow (flow).
By contrast, the estimation results collected in Table 2 do not
confirm the idea that a rise in output price (pout) or a reduction in
production cost, as proxied by freight cost (freight), lead to more
water rights litigation. In fact, the estimated coefficients point in
exactly the opposite direction and at least for variable pout the
estimated effect is highly significant statistically. To check on the
sensitivity of the results to the large increase in pout during the war
years (Figure 2), we re-estimate the equations of Table 2 for a sample
that ends in 1914 instead of in 1920. Compared to Table 2, the
coefficients of variables pout and freight increase in absolute value.
This suggests that the inclusion of the war years for the regressions
reported in Table 2 are not responsible for the negative and significant
impact of pout or the positive and significant impact of freight on the
number of court cases.
We propose the following explanation of this inconsistency with the
theory related to water rights. Falling output prices and rising
production costs do indeed lower the implied value of water, but they
also threaten the very survival of farmers. The latter threat is likely
to dominate farmers' actions. However, how farmers react to strong
price and cost pressures is widely discussed in the literature on rent
seeking and the demand for agricultural protection mentioned earlier. In
particular, lower output prices or higher production costs tend to raise
the demand for internal or external protection by farmers. Litigating
water rights can be interpreted in this context as just another way of
farmers to seek internal protection. (12) In this case, protection is
sought in the courts, not from legislators. Seeking protection in the
courts is the preferred mode in the late 1800s and early 1900s not only
because the problems are localized, but also because farming is still
too large a sector to make government protection feasible (Zietz and
Valdes, 1993). Later in time, as agriculture shrinks in relative size,
farmers' demand for protection is directly addressed to
legislators. In the case of the U.S., it has resulted in an extensive
system of government price supports and input subsidies (Gardner 1987).
In fact, this trend to protect U.S. agriculture from the free forces of
the market started after the output price spike triggered by World War I
(Figure 2).
VI. Summary and Conclusions
The paper examines the development of water rights in Colorado
during the latter part of the 18th and early part of the 19th century.
The model is based on theoretical work on the evolution of property
rights for natural resources and in particular water. According to the
theoretical model, greater water rights defining and enforcing activity
is expected whenever (a) the demand for water increases or (b) the
supply of water decreases.
The empirical model explains the annual number of water rights
court cases in Colorado for the years 1884 to 1920. Among the
determinants of water rights cases are the number of irrigated farms,
farm productivity, farm output price, and level of stream flow. The
empirical model confirms that proxies for the demand and supply of water
can explain a good part of the level of water rights litigation over
time: litigation increases when the potential benefits of additional
property rights defining and enforcing activity rise. This result
confirms a basic hypothesis in the economics literature on property
rights. However, it is also apparent from the estimation results that
property rights alone cannot account for all the empirical regularities.
The fact that litigation activity rises with a decrease in farm output
price and with an increase in cost requires some alternative
explanation. It is suggested that this behavior is consistent with the
predictions of the literature on agricultural protection: farmers seek
protection against the specter of default by litigating in the courts.
After Word War I, they redirect their demand for protection from the
courts to the government.
Enforcing and protecting property rights to water is clearly no
less of an issue today than it was in 19th century Colorado. In fact,
there is significant potential for serious future conflict not only in
the arid parts of the Western U.S. but also in many parts of the world
outside the U.S., including the Middle East. Many of the property rights
issues for water outside the U.S. are likely to require international
litigation and the development of a set of new international rules and
regulations.
References
American Digest 1658 to 1896, Vendor and Purchaser to Willful
Injuries, (St. Paul, MN: West Publishing, 1904).
American Digest 1897 to 1906, Vendor and Purchaser to Zinc, (St.
Paul, MN: West Publishing, 1910).
American Digest 1907 to 1916, War to Zoological Societies, (St.
Paul, MN: West Publishing, 1922). American Digest 1916 to 1926, Trover
and Conversion to Wharves, (St. Paul, MN: West Publishing, 1929).
Anderson, T.L., "Institutional Underpinnings of the Water
Crisis," Cato Journal, Vol. 2, 1982, 759-792.
Anderson, T.L., Water Crisis: Ending the Policy Drought,
(Baltimore, MD: Johns Hopkins University Press, 1983).
Anderson, K. and Y. Hayami, The Political Economy of Agricultural
Protection: East Asia in International Perspective. (Sydney: Allen and
Unwin, 1987).
Atack, J., F. Bateman, and W.N. Parker, "The Farm, the Farmer,
and the Market," Chapter 6 in The Cambridge Economic History of the
United States, Volume 1, Stanley L. Engerman and Robert E. Gallman,
editors, (New York: Cambridge University Press, 2000), 245-284.
Belsley, D.A., E. Kuh, and R.E. Welsch, "Identifying
Influential Data and Sources of Collinearity", Regression
Diagnostics, (New York: Wiley, 1980).
Boyd, D., "Irrigation Near Greeley Colorado," U.S.
Geological Survey, (Washington, D.C: U.S. Government Printing Office,
1897).
Demsetz, H., "Toward a Theory of Property Rights,"
American Economic Review, Papers and Proceedings, Vol. 57, 1967,
347-359.
Doornik, J.A. and H. Hansen, "An Omnibus Test for Univariate
and Multivariate Normality," Working paper W4&91, December,
(Oxford, U.K: Nuffield College, 1994).
Dunbar, R., Forging New Rights in Western Waters, (Lincoln and
London: University of Nebraska Press, 1983).
Durbin, J., and S.J. Koopman, Time Series Analysis by State Space
Methods, (Oxford and New York: Oxford University Press, 2001).
Eschelbach Gregson, M., "Rural Response to Increased Demand:
Crop Choice in the Midwest 1860-1880," Journal of Economic History,
Vol. 53, 1993, 332-345.
Fox, L.P., "State Regulation of the Canal Corporation in
Colorado," Michigan Law Review, 1918, Vol. 16.
Fox, L.P., "Origins and Early Development of Populism in
Colorado," Ph.D. dissertation, University of Pennsylvania, 1916.
Gardner, B.L., "Causes of U.S. Farm Commodity Programs,"
Journal of Political Economy, Vol. 95, 1987, 290-310.
Honma, M., and Y. Hayami, "The Structure of Agricultural
Protection in Industrial Countries," Journal of International
Economics, Vol. 20, 1986, 115-129.
Hundley, Jr., N., "The Great American Desert Transformed:
Aridity, Exploitation, and Imperialism in the Making of the Modern
American West," in Water and Arid Lands of the Western United
States, Chapter 2, Mohamed T. El-Ashry and Diana C. Gibbons, editors,
(New York: World Resources Institute: 1988), 21-84.
Khan, B.Z., "Commerce and Cooperation: Litigation and
Settlement of Civil Disputes on the Australian Frontier,
1860-1900", Journal of Economic History, Vol. 60, 2000, 1088-1119.
Krueger, A., "The Political Economy of the Rent-Seeking
Society," American Economic Review, Vol. 64, 1974, 291-303.
Landes, W.M. and R.A. Posner., "Legal Precedent: A Theoretical
and Empirical Analysis," Journal of Law and Economics, Vol. 19,
1976, 249-307.
Ljung, G.M., G.E.P. Box, "On a Measure of Lack of Fit in Time
Series Models," Biometrika, Vol. 65, 1978, 297-303.
Miceli, T.J., The Economic Approach to Law, (Stanford, CA: Stanford
University Press, 2004).
Newell, F.H., "1890: Report on Agriculture by Irrigation in
the Western United States," in Eleventh Census of the United
States, U.S. Department of the Interior, Census Office, 1894.
Pejovich, S., The Economics of Property Rights: Toward a Theory of
Comparative Systems, (Boston, MA: Kluwer Academic Publishers, 1990),
Online version.
Schwarz, G., "Estimating the Dimension of a Model,"
Annals of Statistics, Vol. 6, 1978, 461-464.
Tarlock, A.D., "The Future of Prior Appropriation in the New
West," Natural Resources Journal, Vol. 41, 2001, 769-793.
Taylor, H.C., "Economic Problems in Agriculture by
Irrigation," Journal of Political Economy, Vol. 15, 1907, 209-228.
U.S. Department of Agriculture, Bureau of Statistics, "Corn
Crops of the United States, 1866-1906," Bulletin No. 56.
(Washington, D.C.: Government Printing Office, 1907a).
U.S. Department of Agriculture, Bureau of Statistics, "Wheat
Crops of the United States, 1866-1906," Bulletin No. 57.
(Washington, D.C: Government Printing Office, 1907b).
U.S. Department of Agriculture, Bureau of Statistics, "Hay
Crops of the United States, 1866-1906," Bulletin No. 63.
(Washington, D.C: Government Printing Office, 1908a).
U.S. Department of Agriculture, Bureau of Statistics, "Potato
Crops of the United States, 1866-1906," Bulletin No. 58.
(Washington, D.C: Government Printing Office, 1908b).
U.S. Department of Agriculture, Division of Publications, Yearbook
of the United States Department of Agriculture, (Washington, D.C:
Government Printing Office, 1907-1920).
U.S. Department of Commerce, Bureau of the Census,
"Agriculture, 1909 and 1910," Thirteenth Census of the United
States, Volume V, (Washington, D.C.: Government Printing Office, 1913).
U.S. Department of Commerce, Bureau of the Census, Historical
Statistics of the United States, Colonial Times to 1970, Part 1.
(Washington, D.C.: Government Printing Office, 1976).
U.S. Geological Survey, Compilation of Records of Surface Waters of
the United States through September 1950, Part 9: Colorado River Basin,
Water Supply Paper No. 1313. (Washington, D.C: Government Printing
Office, 1954).
U.S. Geological Survey, Compilation of Records of Surface Waters of
the United States through September 1950, Parts 6-8: Missouri River
Basin below Sioux City, Iowa, Water Supply Paper No. 1310. (Washington,
D.C: Government Printing Office, 1958).
Williams, W.D., "Irrigation Law in Colorado, Part II,"
Rocky Mountain Law Review, Vol. 10, 1938, 87-104.
Zietz, J. and A. Valdes, "The Growth of Agricultural
Protection," Chapter 5 in Trade and Protectionism, NBER-East Asia
Seminar on Economics, Vol. 2, T. Ito and A. O. Krueger, editors,
(Chicago, IL: University of Chicago Press, 1993) 115-143.
Notes
(1.) 1 Colo. 551 (1872).
(2.) 4 Colo. 100 (1878).
(3.) 6 Colo. 443 (1882).
(4.) 45 Colo. 401 (1909).
(5.) Wheeler v. Northern Colorado Irrigation Co., 10 Colo. 582
(1886), discussed more fully in Williams (1938).
(6.) See American Digest 1658 to 1896 (1904), American Digest 1897
to 1906 (1910), American Digest 1907 to 1916 (1922), American Digest
1916 to 1926 (1929).
(7.) Census figures on the number of farms are obtained from the
"Historical Statistics of the United States," Colonial Times
to 1970, Part 1, 1976.
(8.) U.S. Department of the Interior, Census Office (1894), p. 90.
Wheat cultivation dominated early irrigation in Colorado until the early
1880s. Falling relative wheat prices encouraged diversification into
corn, oats, and alfalfa. But alfalfa required three to four times more
water per acre than did wheat, thereby exacerbating water shortages in
the late 18808 (Fox 1916, pp. 133-34).
(9.) The data are collected from U.S. Department of Agriculture,
Bureau of Statistics (1907a, 1907b, 1908a, 1908b) and the Yearbook of
Agriculture, volumes 1907 to 1920.
(10.) All models are estimated in Stamp 6.0, which is a structural
time series package discussed in Durbin and Koopman (2001). The program
allows for the inclusion of unobserved trend components. This can be of
importance for model fit and interpretation if critical regressor
variables that drive the dependent variable over time are unobserved.
The empirical results, however, do not suggest that unobserved
components play a significant role in the current case. Hence, ordinary
least squares emerges as the appropriate technique.
(11.) The variable ishort is not included in the preferred models
because it is not significant at typical levels of statistical
confidence or its coefficient is not sufficiently robust to slight
modifications of the model.
(12.) Compare on this also the interpretation of Eschelbach Gregson
(1993): "German rural economists ... were among the first to
recognize that as transport costs fell, inherent soil suitability became
more important ... As the constraints imposed by high transportation
costs are ameliorated, the constraints imposed by nature become
binding." (p. 334).
David A. Penn, Director, Business and Economic Research Center
Middle Tennessee State University, Box 102 Murfreesboro, TN 37132,
E-mail:
[email protected] Phone: (615) 904-8571
Joachim Zietz, Department of Economics and Finance Middle Tennessee
State University, Box 129 Murfreesboro, TN 37132, E-mail:
[email protected] Phone: (615) 898-5619
TABLE 1.
Definition of Variables and Basic Statistics, 1884-1920
Variable Definition Mean Minimum Maximum
cases number of Colorado water 9.38 1 19
rights cases heard by the
supreme and appellate
courts
acresifarms acres on irrigated farms 1879.92 342.70 3348.40
in Colorado, 1,000s
ifarms number of irrigated farms 18.30 4.40 28.80
in Colorado, 1,000s
pout production weighted real 117.34 73.70 265.70
price index of corn,
wheat, potatoes, hay, and
oats (adjusted by the
consumer price index),
Colorado, 1900 = 100
outacre output of corn, wheat, 1.83 1.20 2.52
potatoes, hay, and oats,
in 1900 prices, per acre,
Colorado
flow annual combined index of 104.41 49 218
stream flow for two
rivers in Colorado
freight average freight rate of 11.93 8.80 16.47
wheat, Chicago to New
York, cents per bushel
ishort U.S. Commercial Paper 4.81 3.04 7.37
Rate, New York City,
series m 13002 in NBER,
Macrohistory Database
TABLE 2.
Estimation Results, Ordinary Least Squares, 1889-1920
Levels Logs
Variables Model 1 Model 2 Model 3 Model 4
constant -2.581 -8.495 1.578 4.662
(-0.766) (0.397) (0.673) (0.131)
acresifarms_3 0.004 0.867
-0.03 (0.005)
ifarms_4 0.554 1.013
(0.019) (0.008)
pout_1 -0.067 -0.068 -1.135 -1.065
-0.001 (0.001) (0.000) (0.000)
outacre_4 5.835 5.357 1.089 1.087
-0.009 (0.017) (0.021) (0.024)
flow_4 -0.097 -0.095 -0.993 -0.988
0 (0.000) (0.000) (0.000)
freight_2 1.102 1.417 1.423 1.479
-0.055 (0.031) (0.081) (0.090)
[R.sup.2] 0.6117 0.6232 0.6889 0.6790
BIC 2.597 2.567 -1.763 -1.732
condition number 26 25 56 63
Durbin-Watson 1.8459 1.8990 1.8137 1.7704
p-values:
Ljung-Box Q (6 lags) 0.7906 0.8628 0.9918 0.9931
Heteroskedasticity 0.4212 0.4119 0.7963 0.8086
Normality-DH 0.4093 0.6569 0.2296 0.1969
Chow (50 percent) 0.3098 0.3390 0.6462 0.6893
Chow (90 percent) 0.0921 0.0737 0.0585 0.0598
Failure [chi square](5) 0.0828 0.0554 0.0544 0.0603
Cusum t(5) 0.2605 0.2227 0.0806 0.0827
Notes: cases and its natural logarithm are the dependent variables
for the level and log equations, respectively; lags are identified
by an underscore following the variable name; p-values are liven in
parenthesis. [R.sup.2] and the Schwarz's (1978) Bayesian
Information Criterion (BIC) are not directly comparable across the
level and log equations. Collinearity is checked with the condition
number suggested by Belsley et al. (1980). Ljung-Box Q is the Ljung
and Box (1978) test for autocorrelation up to lag order 6. Normality
is checked with the Doornik and Hansen (1994) test. Structural
stability is checked with two Chow tests and with two 5-period
one-step-ahead out-of-sample forecasting tests (Failure and Cusum,
respectively). All adequacy tests are passed at the five percent
level or better.
