REGIONAL DEVELOPMENT AND WATER RESOURCE MANAGEMENT:
IMPLICATIONS OF A CHANGING AGRICULTURAL SECTOR
ON PUERTO RICO'S SOUTH COAST
GREGORY LEE MORRIS
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The assistance of the members of my committee is gratefully
acknowledged. In particular, I wish to acknowledge the patient and
encouraging assistance I have received from my chairman, Dr. Kylstra,
and my long-standing inspirational debt to Dr. Odum. I also appre-
ciate the understanding demonstrated by my committee in accepting my
five years of commuting between Gainesville and Puerto Rico as I at-
tempted to balance job responsibilities and income against research
requirements and expenses.
I owe a great personal debt to Douglas Pool. He was
responsible for introducing me to Puerto Rico, and has served as a
reliable and helpful friend. He also single-handedly ran a
difficult agricultural venture during my unexpectedly long absence
from the island while bringing this research to its conclusion.
There are many individuals on the island, too numerous to
name, who have offered readily of themselves and provided personal
insights into the social, cultural and political aspects of Puerto
Rico. These include many friends at the Department of Natural Re-
sources in San Juan, the Tirado family, Janet Weiss, Hugh Thorne, and
my dedicated friend and companion, Miriam Pab6n.
From my family in Sarasota I have received financial aid
and friendly advice and support at the most timely moments. In
Gainesville, my sister has provided three years of accommodations,
airport limousine service, and good humor.
Finally, I am fortunate to have the very professional typing
services of Sofia Kohli ("Superfish"). Her experience and ability
facilitated the preparation of the final document immeasurably.
TABLE OF CONTENTS
ACKNOWLEDGMENTS . .. .... .. .. ii
ABSTRACT . .... .. . .. vii
CHAPTER I: INTRODUCTION AND OVERVIEW . .. 1
Problem Statement . .. 1
Methods of Regional Analysis . 3
Input-Output . 3
Neoclassical Growth Models . 5
Simulation Models . 8
Previous Resource Studies in Puerto Rico 0
Study Methodology . 12
Regional Development Model .. 12
Resource Management Model .. 13
Report Organization . .. 14
CHAPTER II: STUDY AREA DESCRIPTION . .. 16
Location . . 16
Physical Setting . 19
Geology . . 19
Vegetation and Land Use .. 19
Climate and Hydrology . .. 21
Socioeconomic Setting . .. 33
History . . 33.
Socioeconomic Objectives .. 37
Socioeconomic Conditions .. 38
Institutional Framework . .. 43
CHAPTER III: REGIONAL MODEL .. . 48
Population Sub-Model . .. 52
General Equations . .. 52
Fertility . . 55
Mortality . . 61
Migration . . 63
Economic Sub-Model . .. 74
General Equations . .. 74
Productive Capital . .. 75
TABLE OF CONTENTS--Continued
Production Relationships ... 95
Economic Dependence . 112
Population and Economic Linkage ...... 116
Employment . . 117
Labor Force .............. 117
Unemployment and Migration ... 119
Model Calibration . ... 120
Regional Model Simulation Results .. 122
CHAPTER IV: WATER MANAGEMENT MODEL ... .. ...... .131
Supply Curve for Water . .... 132
General Considerations ... 132
Supply Curve Construction ... 133
Demand Curve for Water . .... .140
General Considerations ... 140
Municipal Water . .... .142
Industrial Water . ... 145
Recreation and Natural Systems .. 150
Agricultural Water Demand ... 151
Existing Conditions .......... 151
Water Demand in the Sugar Industry 153
Water Demand by Non-Sugar Crops .... 166
Aggregate Irrigation Demand ...... 169
Changing Pattern of Irrigation Demand 175
Water Management Model . .... .178
Safe Yield Criteria .......... 179
Water and Salt Budget . .. 179
Model Assumptions and Equations .. 183
Water Resource Scenarios ... 192
Model Results . 194
CHAPTER V: DISCUSSION . .... 197
Development Objectives ........... 197
Alternative Development Patterns ... 201
Feasibility of Development Alternatives 201
Alternative Development Patterns in
Agriculture . .... .204
Resource Implications of Agricultural
Change. ... . 207
Management Changes . .. 207
Institutional Needs and Constraints 209
A Proposal for Resolving the
Institutional Problem ... 212
TABLE OF CONTENTS--Continued
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS
Summary and Conclusions .
Overview . .
Regional Growth .
Resource Management .
Recommendations . .
Agriculture . .
Water Management .
CAPITAL STOCK ESTIMATE FOR PUERTO RICO
REGIONAL MODEL EQUATIONS .
SKETCH . . .
Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
REGIONAL DEVELOPMENT AND WATER RESOURCE MANAGEMENT:
IMPLICATIONS OF A CHANGING AGRICULTURAL SECTOR
ON PUERTO RICO'S SOUTH COAST
Gregory Lee Morris
Chairman: Chester D. Kylstra
Major Department: Environmental Engineering Sciences
As provided by the Water Resources Council's "Principles and
Standards," resource management in the United States is conducted
with the objective of providing for increases in income and maintenance
of environmental quality. However, these criteria inadequately express
development goals in Puerto Rico, where a development policy aimed
principally at providing increases in income and industrialization
through capital importation and increases in transfer payments have
served to greatly increase the island's level of economic dependency.
Over 25 percent of all income on the island was derived from unilateral
federal transfers by 1977. The savings rate in the island's private
sector has decreased and is now negative, with income increases being
used to support higher levels of debt rather than increases in produc-
A simulation model is constructed based on theoretical rela-
tionships to examine these development patterns. It is shown that
continuation of existing development patterns may provide long-run
increases in income and employment, but that this may be accompanied by
even higher levels of dependency. Under these conditions public
sector resource decisions based on income and employment will only
perpetuate these undesirable patterns.
These patterns can be broken only through the promotion of
local investment and entrepreneurial activity. A prime sector for
the promotion of entrepreneurship is the island's backwards agricultural
sector, which stands in stark contrast to the modern industrial sector.
Most of Puerto Rico's mechanizable lands, and nearly all irrigated
lands on the semi-arid south coast, are planted in sugarcane on govern-
ment operated farms. The government sugar industry suffers losses of
approximately $500 per acre per year. In response to these losses,
and in the interest of promoting agricultural modernization, government
policy has recently been implemented to lease government sugar lands
to private growers for the production of non-sugar crops.
Hydrologic and irrigation conditions on the south coast are
described, showing that a significant portion of the economic problems
in the sugar industry may be attributed to archaic irrigation practices.
Private farms are expected to be characterized by increasing irrigation
efficiency, which could increase the rate of consumptive use and
decrease recharge from irrigation percolation, creating an unfavorable
water balance in the already heavily drafted south coastal aquifer.
The conversion from sugarcane to diverse seasonal crops will also
create the potential for economically efficient transfers among private
Existing water management institutions in Puerto Rico are
poorly adapted to accommodate these changes in irrigation management.
The primary problem lies in the lack of accountability which the
resource manager has toward the resource user. An institutional
framework is synthesized in this research which is aimed at improving
this accountability through creation of a local water management insti-
tution which will be capable of providing for long-term as well as
seasonal transfers among users in response to changing crop patterns
and economic demand.
INTRODUCTION AND OVERVIEW
Regional growth creates changes in patterns of natural
resource utilization. By understanding these changes and their
implications, management policies may be implemented to increase the
value of resources to society. This value to society is determined
by the contribution which a particular resource use makes to patterns
of regional development which are themselves perceived to be desirable.
Thus, regional development and resource management policies are
intimately linked to one another. Changing development policies and
patterns of growth can foster changes in resource allocation. This
in turn, can create the need for revised resource management policies
and institutions. Conversely, the development policy itself must be
formulated within the constraints imposed by the region's resource
The relationship between changing regional development
strategy and water resource management in Puerto Rico is examined.
Following three decades of intensive development effort, Puerto Rico
in the 1970s has high levels of unemployment and has become increas-
ingly dependent on growth in transfer payments from the federal
government, which now equal 26 percent of the island GNP. A majority
of business earnings are now remitted to non-residents, reinforcing
the island's dependence on capital imports. Some of the changes in *
development policies needed to deal with the conditions will also
require changes in patterns of resource use and management policies.
A regional development model for Puerto Rico is constructed
and simulated to examine growth patterns under current and alternative
development strategies over the period 1947-2020. It is asked
whether the strategy of industrialization, emphasizing heavy reliance
on capital imports and transfer payments, can generate patterns of
development which are consistent with the island's long-run goals,
i.e., to reduce unemployment, increase income levels, and increase the
island's economic self-reliance.
Changes in regional development patterns may modify patterns
of resource use. This relationship between a changed regional
development strategy and water resource management is explored for
Puerto Rico's semi-arid south coast. Models of water resource use are
constructed and used to suggest some of the water management implica-
tions of new development policy, focusing on the potential effects of
changing irrigation management accompanying the shift from unprofit-
able sugarcane production on government-managed plantations to the
profitable production of non-sugar crops on numerous smaller private
The hydrologic changes are assessed from supply and demand
conditions and a water management model. The water management model
is run over a period of 25 years to define those changes in ground
water withdrawal, consumptive use in agriculture, and artificial
recharge which could create an unfavorable water or salt balance.
These results are used to suggest institutional changes which are
needed to accommodate these changing water management conditions.
Methods of Regional Analysis
A variety of quantitative analytical models may be utilized
to examine regional growth patterns and potential, including export
base, cumulative causation, industry structure analysis, shift share
analysis, etc. However, many of these models have serious theoretical
or operational weaknesses and cannot be used to directly address the
specific relationship of resource infrastructure to the regional
growth process. Three types of models which have been used to
evaluate the role of resources in regional growth are discussed:
input-output, neoclassical growth models, and simulation modeling.
The input-output (I-0) framework is a matrix of the linkages
among the sectors of the existing economic system. Utilizing an I-0
model an analyst can examine the response of all sectors of the
economy to any initial changes which it is desired to test. In its
static form the technologic structure of the system is held constant.
The conversion of a static I-0 model to a dynamic framework requires
not only the conversion of the model's internal structure, but also
requires that the exogenous sectors change. Richardson (1973) has
suggested that a separate model might be constructed to generate the
changes in exogenous variables over time which would be required for
the dynamic operation of an I-0 model. When used as a dynamic tool
the analyst makes judgmental estimates of the changes in the
coefficients which may occur over time, plus the effect of changing
technology and substitutions on both the input and the output sides.
Changes in industrial mix and other "add-ons" are required to make
this static model perform in a dynamic framework.
Input-output techniques were used by Kelso et al. (1973) to
study the effects of a declining water resource base on the Arizona
economy. These declines are being caused by withdrawals from desert
aquifers in excess of recharge rates, resulting in declining ground
water availability. In its dynamic form the structure of this I-0
model was altered in accordance with a variety of different growth
scenarios to show the possible effects of water shortage on the economy
over a 50-year period.
Serious impediments to the implementation of an I-0 model are
the lack of readily quantifiable available data and the amount of
time and expense involved in data collection. Puerto Rico has
detailed information on inter-regional trade, a serious data deficiency
in most regions, and has had several I-0 models constructed over the
past 15 years. However, in these models the agricultural sector is
aggregated and the role of irrigated agriculture is thereby masked.
There is also considerable uncertainty as to the stability of the
pattern of I-0 linkages in Puerto Rico's economy over the long run.
Other general weaknesses of the I-0 model include the assumption of
linear relationships throughout and the essentially ad-hoc procedures
which are necessary to convert it to a dynamic framework.
Neoclassical Growth Models
Lewis et al. (1973) emphasized the general inadequacy of the
regional growth models formulated todate, and underscored the need
to gain an improved understanding of regional growth processes as a
prerequisite to being able to predict the effect of water resource
investment (or any other investment) on a region's economy. They
proposed that an adequate regional model should explicitly consider
the sources of regional growth and interactions among those growth-
promoting factors, with particular emphasis on those factors which
are amenable to policy manipulation. Inasmuch as no general type of
analytical framework exists which can adequately account for the
variation among structures and growth processes in different regions,
it was stressed that each regional model would require separate syn-
Lewis et al. (1973) proposed that regional economic charac-
teristics could be described using a neoclassical production function
model of the type
Q = f(K,R,L,T)
Q = output,
K = capital,
R = resources, including land,
L = labor, and
T = technological advance.
This relationship may be solved for the marginal products of each input
3Q; aQ,; D; DQ;
3K' aR' 3L' aT'
or it may be rewritten in the growth accounting form:
q = ak + Br + c9 + t
q = -; k = K; r = ; A = AL. and t = A
Q K R L Q
AQ' = output increase from technological advance
The elasticities of production for K, R and L, respectively, are a,
a, and e. These elasticities represent the percentage change in output
corresponding to a percentage change in the respective input factor,
a MK Q
Under the assumption of perfectly competitive factor and
product markets and full utilization of resources (i.e., full employ-
ment), it can be shown that the marginal revenue product of each factor
is equal to its marginal product times the product price. As summar-
ized by Denison (1974, p. 51) in his growth accounting study:
If a small percentage increase in the number of units of
all of the factors would increase output of the sector
by x percent, then a percentage increase of the same
amount in the number of units of only one factor would
increase output by x times the share of that factor
in total earnings in the sector.
Richardson (1973), however, has opposed using the neoclassical
approach in regional growth analysis, noting that several of the
theoretical assumptions made in computing the elasticities of factor
inputs are unrealistic at the regional level. Perhaps the most
serious defect is the assumption of full employment. Departures from
full employment among regions are generally an issue of central concern
in regional analysis, but these departures are not consistent with
the neoclassical assumption that factor inputs be fully utilized.
This is particularly important in Puerto Rico where the unemployment
rate has never registered less than 10 percent. An additional problem
in the neoclassical models is the difficulty with which uncertainty
can be admitted, and the explicit consideration of uncertainty is a
central issue in Puerto Rico.
Finally, the neoclassical approach assumes a world of perfect
competition, along with its correspondingly perfect information flows,
marginal adjustment of production, and relatively easy entry to and
exit from production. These assumptions are far from realistic in
Puerto Rico, particularly in the island's irrigated agricultural
sector. To date, this sector has been operated as a government
monopoly which has been so unresponsive to prices that it continues
producing sugarcane, even though the income earned barely covers
50 percent of the production cost. The entry to production of non-
sugar crops is proving to be extremely difficult due to innumerable
institutional and physical problems such as government policy, credit
availability, land tenure, lack of local experience in non-sugar
Simulation techniques may be utilized to incorporate a variety
of relationships within a single model. These relationships are
expressed as a system of equations (often nonlinear) obtained through
the application of other analytical techniques such as regression
analysis, etc. These relationships may be combined in the simulation
model to create a system of interacting relationships and feedbacks,
the response of which may be simulated over time.
The organization of information and relationships within the
simulation model may follow definite theoretical patterns, such as
the energy analysis framework developed by Odum (1971) based on
modeling patterns of energy flow. Others have used a less rigorously
organized approach in which the modeler devises causal relationships
not necessarily following a specific theoretical pattern, such as
the models of system dynamics (Forrester, 1968). This latter class
of simulation models also includes the combination of selected
parameters for the purpose of forecasting, such as the model of the
Susquehanna River Basin by Hamilton et al. (1969). Undoubtedly the
best-known utilization of simulation models has been the formulation
of various world models (Meadows et al., 1974; Mesarovic and Pestel,
1974). Simulation studies have also been used frequently in eco-
logical studies, as summarized by Patten (1971).
Perhaps one of the more severe problems with simulation tech-
niques lies in the ease with which a relatively simple model can be
expanded into increasing size and complexity, exploiting the confusion
between complexity and accuracy. The outcome of even complex socio-
economic models generally rests on only a few critical assumptions,
and by changing these basic assumptions the behavior of the model may
be completely altered. Thus, complex socioeconomic models have the
ability to obscure basic relationships with extraneous and essentially
Once the model is constructed and operational, if the output
does not appear to be "right," then its structure may be altered
until the "correct" results are produced. However, undue emphasis can
be placed on model manipulation without necessarily creating improved
understanding of the socioeconomic system being examined, as suggested
by Lee (1973, p. 168):
There is a popular illusion that confronting a computer
with one's ideas enforces rigor and discipline, thereby
encouraging the researcher to reject or clarify fuzzy
ideas. But in a more useful sense, the effect is the
opposite; it is all too easy to become immersed in the
trivial details of working with a problem on the computer,
rather than to think it through rationally. The effort
of making the computer understand is then taken for intel-
lectual activity and creative problem solving.
This criticism becomes increasingly valid as the complexity of the
model is increased. Nonetheless, the simulation methodology is
unique in its ability to incorporate within the framework of a single
model a variety of different relationships and techniques. A properly
constructed and applied simulation model is an excellent analytic
tool, perhaps unparalleled as a means to integrate complex systems.
Previous Resource Studies in Puerto Rico
This research examines Puerto Rico's post-war development
period, and particularly those patterns which have evolved in the 1970s.
Macroeconomic studies undertaken prior to the 1973 oil embargo tend
not to reflect the realization that development conditions are changing
in Puerto Rico. Thus, earlier studies reflect entirely different
conditions of economic performance and shed less light on the nature
of the difficulties the island faces in the 1970s. Historic patterns
of socioeconomic development have been examined in works such as
those of Lewis (1963), Ross (1966), and Curet-Cuevas (1976). Some of
the development patterns and problems which emerged in the 1970s were
documented and analyzed by the Committee to Study Puerto Rico's
Finances (1975) and the ComitO Interagencial (1975).
No study has been undertaken which directly addresses the
problem of resource policy and its relationship to the overall long-
range development process. A major social benefit-cost study of a
proposed copper mining project was undertaken (Logic, 1974), but the
social benefit-cost framework utilized was limited in scope and did
not address overall development issues in depth. A major water supply
study was conducted by the U.S. Army Corps of Engineers (1977) con-
currently with this research, but was focused primarily on defining
technical supply alternatives. The island's new agricultural plan
(Vicente-Chandler, 1978) addressed some facets of agricultural water
management, but this treatment was necessarily limited in scope.
The most comprehensive overall treatment of Puerto Rico's water
resource issues is contained in the 1973 and 1975 Water Resource
Assessments for Puerto Rico (Puerto Rico Department of Natural Re-
sources, 1974; Morris et al., 1977). Many of the findings reported
in this research stemmed directly from background studies prepared
for the latter of these assessments.
A considerable record of hydrologic data exists for Puerto
Rico and is available through a combination of local and federal
agencies. The U.S. Geological Survey is the most active agency on
the island with respect to water resources data collection and analy-
sis. The water resource characteristics of most portions of the
island as well as many problems of specific interest have been
studied and results published or made available in unpublished
form. Hydrologic conditions in south coast watersheds are described
in reports by Anderson (1977), Crooks et al. (1969), Dfaz (1974),
Giusti (1968, 1971), Grossman (1969), Heisel and Gonzalez (1976),
McClymonds (1971), and McClymonds and Dfaz (1972).
Geological Survey reports are augmented by the reports of
other government agencies and their consultants, plus the University
of Puerto Rico contribution which is promulgated primarily through the
Water Resources Research Institute in Mayaguez. The total number of
reports published in the past 20 years dealing with water supply in
Puerto Rico probably numbers over 100. Yet, despite the number of
publications related to water resource issues on the island, there
has been little focus on integrating the overall problem within the
framework of long-range development.
Regional Development Model
This research begins with the construction of a regional
development model for Puerto Rico. This regional model has been
constructed in two basic sections focusing respectively on demographic
and economic parameters. Theoretical relationships for each parameter
are developed, quantified, and combined to create the composite model.
The simulation model is calibrated using data from the 1947-1977 period,
and is run until the year 2020.
In the demographic sector the population is divided into nine
age groups, with smaller age intervals being used for younger age
groups. Fertility and mortality rates are modeled as exogenous
variables, and net migration is modeled to be a function of the island
unemployment rate and the magnitude of income transfers.
The economic sector focuses on the role of capital as a
variable describing economic performance, recognizing that capital
is a quantifiable parameter which reflects a variety of other factors
such as entrepreneurship, level of economic organization, savings and
investment behavior, etc. Land, labor and resource availability are
examined as input variables describing economic performance in
Puerto Rico over the historic period.
Capital is divided into those stocks which are owned by
residents and by non-residents to examine the long-run effects of the
steadily increasing portion of productive capital which is externally
owned. Transfers from the federal government are shown to be an
increasingly important source of local income, but contribute neglig-
ibly to the formation of local productive investment. The relationship
between capital and production and the level of employment in the
modern sector of the economy are both determined by the capital-labor
ratio, an exogenous variable. Employment in the traditional agricul-
ture sector is assumed to decline steadily to a low but stable level.
The complete model, containing the demographic and economic
sections and their feedback via the net migration function, is simulated
using the DYNAMO II language (Pugh, 1973). Alternative values for the
foreign investment rate, the local investment rate, and the capital-
labor ratio are applied to generate four future scenarios. The results
of simulations to the year 2020 are examined to suggest those types
of policy which can generate more favorable patterns of regional
Resource Management Model
Important resource management implications of changing regional
development policy are examined on the island's south coast. Water
demand conditions are described to ask what types of use patterns may
occur in the future. An income and cost model for sugarcane farms
is constructed to ask if a rational framework can be constructed for
evaluating irrigation demand in the sugar industry.
The hydrologic implications of increasing agricultural
consumptive use, increasing demand in the municipal and industrial
sectors, and different levels of artificial recharge are all examined
using an aggregate water and salt budget model for the south coastal
aquifers. This model is run over a 25-year period, with the objective
of describing the types of hydrologic changes which could accompany
significant increases in irrigation efficiency. Recommendations are
made for implementing institutional changes which can promote efficient
patterns of resource use, while avoiding deleterious effects to the
ground water system.
After the introductory comments in this chapter, the second
chapter describes the island of Puerto Rico and some of the problems
which have accompanied economic growth over the post-World War II
period. The south coast area and its hydrologic characteristics are
In the third chapter socioeconomic data from the 1947-1977
period are interpreted and the regional development model is
generated, showing the manner in which each model parameter is derived.
The results of calibration with historic data are summarized and
alternative future scenarios are described.
In Chapter IV, the water resource characteristics of the south
coast are examined, describing the level of water supply and demand
and the characteristics of current and potential future patterns of
water utilization. A simple water management model is constructed
to reflect the potential hydrologic effects of changed development
policies in the agricultural sector. The scenarios which are applied
to the water management model, as well as the results of simulating
each of these scenarios, are also described.
The management implications of these models are discussed in
Chapter V, showing the relationship between regional development
policy and water management in the agricultural sector. Existing
water management institutions are described, and recommendations are
made for changes in this institutional structure to accommodate a
changing regional development strategy.
Chapter VI briefly summarizes the conclusions and recom-
mendations stemming from this research.
STUDY AREA DESCRIPTION
The island of Puerto Rico is the easternmost of the Greater
Antilles (Fig. 1). Puerto Rico consists of one main island, two
smaller inhabited islands, and several offshore cays, all having a
total area of 3,435 square miles (Pic6, 1969). The main island is
approximately 30 miles wide and 100 miles long. Being hilly or
mountainous throughout the interior, the majority of the population
and economic activity are supported on the narrow coastal alluvial
plains. The estimated 1977 population was 3.3 million. San Juan
is the capital city, and the San Juan metropolitan area has a popu-
lation of approximately one million.
The south coast study area (Fig. 2) consists of 14 municipali-
ties having a combined 1970 population of 448,000, of which 54 percent
were classified as urban. Ponce, the largest city in the study area
and the second largest city on the island, had a 1970 population of
The study area contains the steep southern slopes of the
island's mountain range, a narrow band of foothills, and a relatively
broad alluvial plain which attains a width of four miles at its
broadest point. Because it falls in a rain shadow the south coastal
plains are semi-arid, receiving an average of 35 inches of rain
annually, concentrated in a fall wet season.
Fig. 1. Location map.
,. \ a
S\ \ x -
0 o A32
- L i
SI 0nI +H
-- ; -!
/ ^-s/ l4
?e s f .2
/ -5 \
\ -I "
* ^ (-
;uo ) 0
< ^ -/ _
Puerto Rico was formed as a result of submarine volcanism
and uplifting which began approximately 120 million years ago in the
Cretaceous Period. Erosion and reef building during submergence
combined to form the coastal metamorphic and limestone deposits which
overlay the parent volcanic rock, which is exposed throughout the
interior (Beniroth, 1969). Limestone deposits exceeding 2,000 feet
in thickness occur near the coast and have developed karst landforms
including extensive sinkhole formation, haystack hills, and the absence
of surface drainage. Limestone formations outcrop primarily on the
western two-thirds of the island, and on the north coast these for-
mations have been developed as an important aquifer.
The island's maximum elevation is 4,389 feet above sea level,
and the entire island interior is characteristically hilly or mountain-
ous. Some 38 percent of the island's total land area has slopes
in excess of 45 degrees. Generalized topographic characteristics.are
summarized in Fig. 3 showing the extent of hill and mountain terrain
and the narrowness of the alluvial plains.
Vegetation and Land Use
The island has six life zones, which have been mapped by
Ewel and Whitmore (1973) using the method of Holdridge (1967). The
diversity of life zones reflects primarily differences in elevation
and exposure, giving rise to differences in temperature and rainfall.
The south coastal plain is located in the sub-tropical dry forest life
S0 0 10
Due to intensive population pressures and an earlier dependence
on subsistence agriculture, which still exists in isolated locations
(Pool and Morris, 1976), the only areas of natural vegetation are
limited to the Luquillo National Forest and the Commonwealth forest
system, encompassing a total of 88,000 acres (4 percent of the island
area). However, considerable areas of marginal agricultural lands on
highland slopes have been abandoned and are reverting to secondary
Current land use in the south coast area is summarized in
Table 1. Agriculture accounts for the single largest category,
utilizing 51 percent of the land in the area; sugarcane is the
predominant crop. Of the 90,000 acres reportedly in sugarcane, at
least 50,000 acres receive irrigation on a regular basis.
Climate and Hydrology
Climate. Puerto Rico experiences an insular tropical trade-
wind climate. The average annual temperature in San Juan is 780 F,
and the record low is 61 F. Temperatures at higher elevations inland
are lower, and may fall below 50 F overnight in thewinter. Super-
imposed on the east-northeast trade winds is a land-sea breeze pattern
which results in easterly onshore winds during the daytime and
easterly offshore winds at night. There is a regular diurnal pattern
in wind speed on the south coast which varies from 3.5 mph at midnight
to 12 mph at noon, with wind velocities in the late summer and fall
being somewhat less than other seasons.
co CN m m Pho C) W MCnW m -*I CD cli m
Sw .- a0 CM in 0 d -- i- c0 CD C 0
CL u CM CM n 0 i- M- Mn CO M mo CM oo
4 -O CU co
S1- CM 0L
*R 1 >
4-) r- m mn CD 0. o D D
a a(aa *
L (1- r M r-i- r- Ri-r O CM CM) Io 0
CO O 4_j D 4_ C\ 70
a C'j M k\ (
o 0 NmO" M M LO 00 ( C C
o an CO-o M
o k- -o C) (n C r in co -I (0 010
aU 4O C_ r- CC -'.mmm Co (a U
<- sM C- \j (
S <0 3 -- (3 4.
C S .
r" -- LO 4-) 4-)
CM LO CM CkICoCI C r0 0 m M mCM m CD u a
Z-- I Cr
C r- Lfl J O
z 0 00
0 0 e- 0 -zr a% to 0
el LC M -- ) C)oCjm(CD a- co 1 c
Qn -j ct LO C^j *:I oo -- C r- m oj r-,
4 -J (v .
wS= -, >- C M U c s
Q- = =-M 0 MC >- M S
CD.. 0 CD Wor0a--oo ) 0
Precipitation and surface water. Average annual rainfall is
70 inches per year island-wide, with a range from 30 to 200 inches per
year between the driest and wettest stations, respectively (Fig. 4).
Inasmuch as this variation is due to orographic effects, the mean
long-term runoff from watersheds may be predicted from elevational
data alone (Black and Veatch Engineers, 1971). As a result of the
impinging trade winds the north coast is characteristically moist,
whereas the south coast falls in a rain shadow. Patterns of rainfall
and pan evaporation on the north and south coasts are compared in
Rainfall records for the longest continually operating
stations on the island are graphed in Fig. 6 using a differential
mass diagram. Periods of above average rainfall appear as upward
sloping line segments and drought periods appear as downward sloping
segments, with the slope between any two points directly related
to the severity of the.departure from average rainfall. Although
relatively long-term variation is absent in the San Juan data, cyclic
patterns become evident in the data from the south coast stations,
with the past decade being generally drier than the period of record
on the south coast. The Ensenada station (GuAnica) is the driest
station on the island, receiving a mean annual precipitation of
On the south coast, a higher variation in rainfall occurs during
the drier first half of the calendar year than in the more moist second
half. In Fig. 7 the calendar year is divided into the first and second
N 0 NORTH C 0 ASTAL
NORTHERN SLOPES -
WE-STERN INTERIOR EASTERN INTERIOR
S7TuH C-STA\L SLOPES
8 i -- i -- -- i -- i -- -- ] -- -- i --- i --- ---
4 Pan Evoporation "
J F M A M J J A S 0 N 0
Vj 1 : ---------------- -----------
i- Pan Evaporation
z SOUTH COASTAL
J F M A M J J A S O N 0
M ONT HS
Fig. 5. Average rainfall and pan evaporation for the north and
south coastal climatic regions of Puerto Rico. Climatic
regions are defined by the U.S. Weather Service.
Source: National Climatic Center. 1973. Monthly averages
of temperature and precipitation for state climatic
divisions, 1941-70: Puerto Rico and Virgin Islands.
Asheville, North Carolina.
-20SAN JUAN (City)
-20 -------- i-
Fig. 6. Differential mass diagram of annual precipitation for those
stations in Puerto Rico having the longest continuous
rainfall records. Ensenada is the driest station on the
island. Average rainfall over the periods of record are: En-
senada, 29.60 inches; Aguirre, 40.94 inches; San Juan (city),
60.17 inches. Data for Aguirre are incomplete for 1970 and
are not plotted.
u / /
u u 40
6O 60 1-
0 Jan.- June / July-Dec.
n I I I
0 40 80 120 160 200
Precipitation as Percentage of Mean
Fig. 7. Cumulative probability of rainfall in the first and second
halves of the calendar year as a function of the mean
rainfall for that same six-month period, south coast of
Puerto Rico (Villalba station).
half and the cumulative probability of rainfall is compared with
respect to the mean of record for each six-month period. The most
severe drought conditions on the south coast occur during the summer,
following low rainfall in the spring.
The island's highly dissected mountainous interior gives
rise to numerous small streams characterized by small tributary
watersheds, steep slopes, and short length prior to reaching the
ocean. There are 26 "major" stream basins on the island, but only
seven have watersheds greater than 100 square miles. The largest
watershed is the Rfo Loiza which covers 310 square miles.
Streams on the south coast are small and none flow throughout
their reach over the entire year. The upper stream reaches generally
maintain a year-round streamflow, but in most seasons diversions for
irrigation plus percolation into the aquifer deplete this streamflow
as soon as it reaches the alluvium.
Water supply development. Unregulated stream diversions
represented the primary source of irrigation water on the south coast
until the latter portion of the 19th century, at which time centrifugal
pumps were installed on some farms. The American occupation of the
island brought with it an influx of investment capital. Mainland
corporations plus wealthier local families undertook large-scale
irrigation on sugar plantations and supported the implementation of
large-scale public irrigation works. The South Coast Irrigation
District was operational by the First World War, providing both irri-
gation and hydropower supplies.
The history of reservoir construction on the island is sum-
marized in Table 2. Of the 25 reservoirs constructed in Puerto Rico
to date, 68 percent have had power generation facilities, 48 percent
supply water for irrigation, and 40 percent include a domestic supply
component in their design. Two additional reservoirs are being
constructed in Ponce for combined flood control and water supply bene-
fits of 37,000 acre feet/year, but will not enter service until the
mid-1980s. Only one south coast reservoir, Toa Vaca, has municipal
water supply as its primary function; all others have been developed
primarily for irrigation supply.
The South Coast Irrigation District consists of three inde-
pendent reservoirs and canal systems, supplying an average of 70,000
acre feet/year to approximately 28,000 acres. The District has seen a
gradual deterioration in service over the decades due to the combina-
tion of reservoir sedimentation, deterioration of canal sections,
possible land use changes, and possible changes in the water management
effectiveness of both the supplying agency as well as the users.
The Southwest Puerto Rico Project supplies an average of
35,000 acre feet/year of irrigation water to 22,000 acres in the Lajas
Valley. This system entered operation in 1954, and consists of five
interconnected reservoirs and stream diversions, two power-generating
stations, a concrete-lined irrigation canal, and a drainage system.
Ground water. Aquifer areas in Puerto Rico are outlined in
Fig. 8. The development of limestone formations predominates on the
north coast west of San Juan, whereas ground water development on the
0 m0 z
,a I' ai c- J
*w j l CC =
Cr~l. i-- i)
P^-L -. VC
CT 3 3-f
o 01 '*3 '.-
S)~ u s:
(> C0 0^ *I
'U *J 0 =0S
+* -a ec a ^
* a 2 X-=
0 i 3-
(T 3 a?'
00 0000000 C 0 LC L r 00000000
0~030N OUh~ 0003
-n a0 cc
0 r 0 C CM J fC
n-- < i- M M
\ k V- 0 it oOr-o o ^^ t1 ^ LO
.. C.. 0 .
V% 7.7 l ~J Cj >e Ln ko 0 C lM r
a- '' 42' A<<
u .) v eA '3 3I
3 :0 -
Za A 1- z .n '4 -S i Lz -I') > -J O
= a' s'3 = V 3 0 C 0-3'0i .33 V 0 C '- C a 't
a( a a a '- LO k. CO o MS N
t-s^'-)i.<-}a.3<- s~-a U-^ -J^ : j --s i
south coast has occurred in the alluvial water table aquifer. Both
formations are prolific aquifers, and individual wells with yields
of 2,000 gpm or more may be developed. Throughout the island interior,
water-yielding zones of fracture in volcanic formations have been
developed for ground water supply, but these wells are generally for
domestic use and capacities are characteristically limited to 5-20
gallons per minute.
The south coastal alluvium lying east of Ponce is the most
copious aquifer in Puerto Rico, sustaining an annual withdrawal
of approximately 200,000 acre feet per year. This alluvium consists
of a loose agglomeration of interbedded and interfingering sedimentary
materials, ranging from silts to cobbles. Alluvial formations are
continuous from Ponce eastward to Guayama, but to the west of Ponce
are confined to individual river valleys. The alluvium has a maximum
depth of approximately 250 feet, and is generally underlain by limestone
formations containing ground water of high mineral content. These
limestone formations outcrop as foothills along the entire south coast,
and westward from Ponce they extend to the ocean and form hills which
separate the alluvial valleys. The south coast soils are derived from
a mixture of volcanic and limestone parent materials and are generally
neutral to basic in pH.
Limestone formations underlying the south coast alluvium have
not been developed for water supply on a significant scale. Shallow
penetration of this aquifer in the Yauco Valley has yielded water of
suitable quality (personal communication, Lee Eaves, Manager, Carib-
bean Well and Pump Service, Ponce, Puerto Rico, 1977) and limestone
formations have been tapped for water supply in the upper reaches of
the alluvium inland from Ponce (McClymonds, 1971). However, the
water from these formations is generally of inferior quality and is
Limited areas of saline intrusion occur along the coast, but
in most locations it is not yet a serious problem. The most extensive
area of intrusion has occurred beneath the City of San Juan, and
intrusion of salts into wells in several other locations has occurred
when depression cones intercepted saline wedges lying in the lower
reaches of rivers. There is a system of wells on the south coast for
monitoring water level and intrusion, but similar information for the
north coast is lacking (Dfaz, 1974).
Puerto Rico was discovered by Columbus on his second voyage
and was settled by Ponce de Le6n during the first decade of the
sixteenth century. Under Spanish rule San Juan and the island of
Puerto Rico played a military and economic role secondary to both
Santo Domingo and Havana, more important Spanish cities in the
Caribbean. Nevertheless, considerable expense was undertaken by
Spain to fortify the entrance to San Juan Harbor in order to provide
refuge for Spanish treasure ships and to maintain control of the
Under Spanish rule Puerto Rico's political and economic
development lagged. For example, in 1898, the salary of the Spanish
Governor was greater than island-wide expenditures for education on
the island's 950,000 inhabitants (Lewis, 1963).. In the late nineteenth
century,efforts to break free of Spanish domination appeared to be
bearing fruit. However, by 1898 the island was still not politically
autonomous and the Governor was still appointed by the Spanish Crown.
When Cuba revolted against Spain precipitating the Spanish-American
War, Puerto Rico did not follow suit.
In 1898 Puerto Rico became a territory of the United States
as part of the spoils of the war with Spain. Following annexation
the Governor was appointed by the U.S. President, and it was not until
1947 that the Governor was elected locally.
Heavy investment in the sugar industry by mainland corpora-
tions rapidly followed annexation, resulting in dramatic changes on
the island. A diversified but undercapitalized agricultural system
was bought out by monied American interests and converted to the
lucrative sugar export business (Fig. 9). Due to the abundance
of labor, unemployment and underemployment was the rule and wages were
maintained at poverty levels. Unlike other countries in Latin America
where additional lands were available for the development of sub-
sistence agriculture further inland, on insular and heavily populated
Puerto Rico this alternative was highly restricted. The worker was
effectively tied to the land holdings of the sugar and coffee landlords
who needed labor only seasonally (Ross, 1966). As a result of competi-
tion for limited land resources, smaller and less competitive units
were gradually absorbed by more efficient and better capitalized farms
O^ [ *
0 .- != 0 .
0 1 .* *, *
820 840 860 80 1900 1920 1940 196 1
0 0 o 0 9.
Fig. 9. Annual production of raw sugar in Puerto Rico.
Source: Puerto Rico Dept. of Agriculture, San Juan.
Source: Puerto Rico Dept. of Agriculture, San Juan.
(Mintz, 1972a), similar to the land tenure trend documented in other
land-limited agrarian economies (Sansom, 1970).
The 1920s and 1930s had seen a steady decline in the plight
of the Puerto Rican worker. The hurricanes of 1928 and 1932 virtually
destroyed coffee plantations and the tobacco crops. As an indicator
of the degraded economic conditions, Ross (1966) noted that in the
early 1930s not more than 2,000 families on the island had sufficient
income to pay taxes, yet 39,000 were employed as domestic servants
(7.5 percent of the labor force). Home needlework was estimated to
employ over 30,000, or approximately 30 percent of those engaged in
the "manufacturing" trades. In agriculture the 500-acre rule was not
enforced, and an estimated 20 percent of all agricultural land in
Puerto Rico was held in violation of it, primarily by the sugar com-
The 1940s were a period of profound economic as well as
political change on the island. Although the results of these changes
were not visible during that decade, the fundamental political and
economic directions which evolved during that period have remained
largely unchanged to the present. In the 1940s work was begun in
earnest to change the political status of Puerto Rico, beginning with
the election of the Governor locally and eventually resulting in the
formation of the Commonwealth status (Estado Libre Asociado) in 1952.
The mid-1940s, under the term of appointed Governor Maxwell Tugwell,
1U.S. Congressional Joint Resolution 20 of 1900 established
federal policy that no individual or corporation in Puerto Rico should
own over 500 acres. (Rivera-Rfos, 1978)
also saw the organization of the executive branch of the Puerto Rican
Government as it is known today. The year 1947 also marked the
beginning of reliable economic data, which was compiled by the then
new Planning Board; and in this same period the industrial development
agency, Fomento, was also created.
With limited agricultural land and high population density it
was realized that an economic system based on agriculture alone could
not be responsible for generating the income necessary to bring the
island's population out of poverty. The late 1940s saw the beginning
to the end of this poverty as a result of three factors: (1) the
attraction of mainland manufacturing investment through an industrial
incentive program known as "Operation Bootstrap," (2) the massive
migration of Puerto Ricans to the New York area which stabilized the
size of the island's labor force, and (3) the growth of transfer pay-
ments from the federal government. Taken together, these three
factors created the fundamental changes in the Puerto Rican system
which now provides adequate food, shelter and medical services for the
entire population, plus a modest measure of luxury for many.
The initial development process reflected movement toward
the obvious and primary goal of escaping from pervasive poverty.
However, as progress toward the abolishment of extreme poverty began
to be realized, a more definitive statement of social goals was
developed, as quoted below from the "Message of the State of
the Country," delivered by Governor Muioz Marfn to the Legislature
I point out six things that I judge fundamental to the pur-
pose of our people, and on which we will invest a good
portion of the continued increases of our economic assets:
Education for all in quantity, quality and depth,
Maximum health with substantial equality for all
with respect to quality of service,
Approximate balance of quality of life between rural
and urban population, development of cities in an
Home ownership for all families,
That a growing portion of the growing Puerto Rican
economy, and consequently of the decisions of its
private sector, be in the hands of the sons and
daughters of this country as has been the case in
all the presently developed countries of the world,
Abolition of extreme poverty.
These enumerated goals are the desired final products of the
island's socioeconomic development process. The pursuit of increasing
income through industrialization and economic growth has been pursued
with intensity, yet has created patterns of growth which are not
entirely compatible with achievement of the stated goals. That dilemma
is examined in this study.
Economic characteristics. Although the island is poorer than
the lowest income state, in comparison to neighboring countries in
the Caribbean and Latin America, it has gained considerable wealth.
1Diario de Sesiones, Cuarta Sesi6n Conjunta, 11 February 1964,
p. 5. Author's translation.
During the periodfollowing World War II the island enjoyed a special
set of circumstances which combined to produce a period of rapid
economic growth, industrialization and social change. The input of
industrial capital from the mainland seeking to utilize the island's
cheap labor supply, its access to U.S. markets, and exemption from
both local and federal corporate income taxes, created a period of
rapid industrial expansion starting in the late 1940s and continuing
into the early 1970s.
The initial success of the island's development program was
based less on clever design or able implementation than on the island's
relationship with the United States. Not only has the mainland infused
capital, entrepreneurship and transfer payments into the island on a
large scale, but it also absorbed the massive exodus of Puerto Ricans
from the island during the 1950s.
However, the conditions which combined to create the oppor-
tunity for rapid economic growth and stable levels of unemployment
have proven to be transient. The post-war growth began to falter and
in the early 1970s conditions began to depart recognizably from the
patterns of previous decades. Rising wage levels on the island
helped stem the migration of labor to the mainland while simultaneously
making the island less desirable for labor-intensive industries. New
industries locating on the island became steadily more capital-
intensive, obtaining the benefits of tax exemption but oriented less
toward the employment of the increasingly costly labor force. Infra-
structure requirements and the discharge of pollutants have increased
as a function of capital intensity thereby increasing both monetary
and non-monetary costs of providing each additional job.
Despite intensive industrialization, the island's unemployment
problem has not been resolved. The unemployment rate has never
dropped below 10 percent over the entire development period, and in
1977 it averaged 20 percent, the highest value recorded over the
entire post-war period. The island's increasing income has not been
used to generate a stream of savings and productive local investment,
but rather has been used to support an increasing burden of debt.
The Commonwealth Government has relied increasingly on
federal aid programs as a source of income and employment. By 1977
the value of unilateral transfer payments from the U.S. equaled
26 percent of the island GNP, and transfers to individuals accounted
for 28 percent of personal income (Puerto Rico Planning Board, 1978a,
1978c).1 The implementation of the federal food stamp program in
Puerto Rico has been the greatest single contributor to increased
transfer payments during the 1970s, and over 60 percent of the island
population now participate in this program.
The income level on the south coast is lower than the island
average, and the trend has been for this gap to increase slowly.
A variety of socioeconomic parameters for south coast municipalities
are summarized in Table 3.
1Socioeconomic accounts data, hereinafter not accompanied by
a citation, shall be understood to be contained in the economic
accounts of Puerto Rico (Puerto Rico Planning Board, 1978a, 1978c) or
of the United States (Economic Report of the President, 1978).
.- -C ....
) 4) ( ,-
3- C 00-
CL 4- -
U C 0CS
0 r- a)
i- O -- 0 c 0 C\O r-
S.'3-t.0 o C\j' )m 0mCO Z C\J CJ M
h mr- mCO mfl
*4 c e (* o*a;
C ++- I+-
i + + I + i I i
-. D M n Coi-
i-u- CJ CJ CMJ
I I I I
OM Un j M.0 m C-l m CD C r 0 UO M t V N
O U'0 MM '0 U Ln C M U, LU
- 0000 00 0 00 M J 0
O o- c aCO CO CMCOn CMo, U
U, n O) CO .0 l O C Fr-
O --- -oO --
O- Cr I- m C' -r '- r- U CO U, U,
I- o r M C ca- UUr-, O M qc UO
Ul V) o- T
0- u A C C
o USC u-u-- (0 -
S-i Vo (Z to cc -j ?rC e- Cr- 3
LC 3 3 ( (D 1-) 0 0- a
) o0 -*
S- ) UUr
c LO0 0
S c S- a)
NO O O 0O
CD y C ) o -."
O OO CO..
C- a ) 0 0 S-
C:3 (0 'a 30
C C (L -- O) EC-
= r- -0
n A4-* a) -
--- E r- '0 4-
0 C U ar-
S- 3 0 (U S- S- CA
0- = 4-) 4- (A
0,- 00 ( CI. 0 0
SCD 4-) C E C= -C
4-) Cr 00 4-U
: *r- r
S". 0) >
1E(0 E 4- J-
*r- *1 0 S. L )
0- 0.) -0 0
vC -C 0 0- O
m v Cn 0Im a.
U W) U I CD) U
S- 41 S. 0 CL S- =C L.
*a a- a 0L a- a4 )
O C4- O S O iO
*>- Q.-- < S-T- U "
i- i- cu (0 s *'>
Demographic characteristics. Puerto Rico experienced a
classic demographic transition in terms of changing fertility and
mortality rates. However, migration to the mainland converted the
demographic transition period from one of rapid population growth
to one of population stagnation. Conversely, the post-transition
period which the island has already entered has been characterized
by high rates of population growth. This demographic anomaly,
created by migration, is of primary importance in explaining the
dramatic economic changes which have occurred in Puerto Rico over
the post-war period.
The migration from the island, primarily to the New York area,
was so large that it actually resulted in two consecutive years of
decreasing population on the island in 1953 and 1954, despite high
fertility. More importantly, the major portion of these emigrants
were young and just entering the labor force (Friedlander, 1965). As
a result of migration the population of Puerto Ricans living on the
mainland U.S. increased from an estimated 100,000 in 1947 to
1.5 million in 1974, representing the exportation of 400,000 or more
labor-force members over that 27-year period.1 Over that same period
the island economy never saw an unemployment rate under 10 percent
and generated a net increase of only 203,000 jobs, despite strong
industrial promotion activities.
As economic conditions on the island began to show improve-
ment in the 1960s, the pace of net emigration slowed. This flow has
1Data provided by the Puerto Rico Planning Board, Division of
Demographic Statistics, 1976.
now reversed, and during the 1970s Puerto Rico has experienced net
immigration, despite a local unemployment in excess of 15 percent.
It is possible that a significant portion of this immigration which
occurred, despite poor employment opportunities, can be attributed
to the growth in welfare benefits on the island.
Commonwealth Government institutions characteristically have
a broad and strong legal mandate for resource planning and management
action, and public policy is the most important single factor
influencing patterns of resource utilization on the island. The
government supplies all municipal water and most irrigation water
island-wide. Through the Puerto Rico Sugar Corporation the government
operates approximately 90 percent of the island's irrigated farm land,
and the Puerto Rico Land Authority is the largest owner of irrigated
lands. Government agencies supply a significant portion of the
industrial water used on the island and regulate the withdrawals by
those other industries which are self-supplied. The pattern and
location of industrial development, and its corresponding water demand,
is determined through the zoning powers of the Puerto Rico Planning
Despite these large resource management responsibilities, a
coherent strategy for water resource development has not emerged,
aside from the protection of existing beneficial uses which is required
by law. During the past decade the island's resource management
problems and conflicts have grown in severity while satisfactory
solutions have become increasingly difficult to define and implement.
Resource development decisions have been made based on poor analysis,
or even by default (Margalin and Campen, 1970).
Two pieces of legislation are of particular importance in
establishing the legal framework for environmental resource management
on the island. The Public Policy Environmental Act1 promulgated public
environmental policy and established the Environmental Quality Board
(EQB) as a watchdog agency. Section 4 of this law provides that all
agencies of the government shall use an interdisciplinary system of
study in analyzing environmental problems, guaranteeing the integrated
use of natural and social sciences.
The Law for Conservation, Development and Use of the Water
Resources of Puerto Rico2 establishes that all fresh and coastal waters
within the jurisdiction of Puerto Rico are the wealth and property of
the People, including all water both above and below ground. All
waters fall under the administrative jurisdiction of the Commonwealth
Government as represented by the Secretary of Natural Resources, who
is provided broad planning and management powers and responsibilities.
The water law provides that planning and management actions shall be
based on socioeconomic benefit-cost analysis, with the objective of
enhancing the well-being, economic development and public safety of
the island, and that all analysis shall be based on an interpretation
1Law No. 9, July 1, 1970, Legislative Assembly of the Common-
wealth of Puerto Rico.
2Law No.136, June 3, 1976, Legislative Assembly of the Common-
wealth of Puerto Rico.
of the role of the proposed management alternatives with respect to
the overall hydrologic cycle and its effects on natural systems.
Water resource planning and management responsibilities are
fragmented among numerous Commonwealth agencies, with the result that
no single agency may be said to play a lead role. The principal
agencies involved and their primary areas of responsibility are sum-
The Puerto Rico Planning Board is the ultimate planning
authority in the Commonwealth Government and establishes overall
planning guidelines which are to be followed by the other agencies.
The primary resource management functions of the Planning Board, an
Executive Branch agency, include the continual updating and implemen-
tation of the master land-use plan, which includes decision making
over the future location of industrial sites and other intensive
resource users. In addition, the Board is responsible for the co-
ordination of the plans of other government agencies.
The Puerto Rico Aqueduct and Sewer Authority (PRASA), a public
corporation, is responsible for the planning, construction and main-
tenance of all municipal water supply and wastewater collection and
treatment facilities on the island. Of the total population of
3.1 million in 1975, an estimated 2.6 million persons were served by
the Authority (house connections plus public fountains). Of the total
1975 withdrawals of 306 mgd, metered deliveries were 199 mgd, or an
island-wide average of 77 gallons per capital per day. Approximately
30 percent of all withdrawals do not appear as metered deliveries, but
are lost due to in-plant uses by PRASA plus pipe exfiltration, fire-
fighting, and clandestine use. The Authority is responsible for the
operation and maintenance of several water supply reservoirs, includ-
ing the Toa Vaca reservoir on the south coast.
The Environmental Quality Board (EQB) is the local counterpart
of the Environmental Protection Agency and is responsible for the
promulgation and enforcement of environmental quality standards. The
EQB has established water quality standards for both fresh and ocean
waters, and has programs to improve environmental quality through the
reduction of point sources of water, air and noise pollution. It is
responsible for the conduct of the 208 water quality programs on the
island and wastewater facility planning.
The Puerto Rico Water Resource Authority (PRWRA) is a public
corporation which maintains and operates the electric power generation
facilities island-wide. In addition, the Authority owns and operates
the three government irrigation systems on the island, two of which
are in the south coast area. An average of 100,000 acre feet/year is
supplied to approximately 50,000 acres lying within the South Coast
and Lajas Valley Irrigation Districts. The third district is on the
moist north coast, where only 600 acres are currently under irrigation.
The Authority also operates reservoirs having hydropower facilities,
but less than one percent of the island's total electric supply is
derived from hydropower.
The Department of Natural Resources is the newest of the
Commonwealth agencies, having been organized in 1973. Under the
provisions of the 1976 water law, the Department is required to inven-
tory all water uses and wells island-wide, establish a system of
water use permits, regulate ground water drilling and withdrawals,
and prepare a water plan which will incorporate socioeconomic cost
benefit analysis. Of all the government entities in Puerto Rico,
the Department has the broadest water resource responsibilities, but
to date the role played by this Department has been minor. The
failure of the Department to fulfill its obligatory functions, as
provided in the water law, has been the subject of recent litigation.1
Additional Commonwealth and Federal agencies are involved in
water resource planning and management on the island, with the number
of agencies having significant roles totaling at least 16.
1Mision Industrial de Puerto Rico vs. Fred Soltero Harrington,
Secretary of Natural Resources. Superior Court of Puerto Rico, 23-
23 August 1978, San Juan. As of this writing this case had not yet
The regional model for the island of Puerto Rico is formulated
in this chapter. It has been constructed as two sub-models of
population and economic performance which are linked through two
relationships: (1) the population and its age distribution determines
the size of the labor force, and (2) economic performance and labor
force variables together determine unemployment, which influences
migration. This model is summarized in Fig. 10, using the symbols
defined in Fig. 11.
In this chapter the model structure is described, historic
relationships are synthesized, and the model is calibrated using data
for the 1947-1977 period. The assumptions for projecting modeled
relationships to the year 2020 are also outlined.
The primary purpose of this model is to examine alternative
levels of future economic performance given different assumptions
concerning the rate and source of future capital formation and the
capital intensity of production. Although the population model
remains unchanged for all scenarios, different patterns of population
growth are produced in each scenario due to the effect of different
unemployment levels on the rate and direction of migration.
Fig. 10. Regional model for Puerto Rico in the DYNAMO language.
See Fig. 11 for explanation of symbols.
iPt = Population of age group "i"
dt = Crude death rate
bt = Crude birth rate
Ut = Unemployment rate in Puerto Rico (percentage)
Pt = Population of working age in Puerto Rico
P" = Labor force in Puerto Rico
qt = Labor force participation rate
Lt = Capital-labor ratio for Puerto Rico
Yt = Gross National Product of Puerto Rico (1954 dollars)
Ht = Remission of business profits abroad (1954 dollars)
aKt = Locally owned capital stocks (1954 dollars)
bKt = Externally owned capital stocks (1954 dollars)
cKt = Stock of consumer assets
Ak = Depreciation rate of capital stocks
Ac = Depreciation rate for consumer assets
Et = Total employment in Puerto Rico
E' = Employment in the traditional agricultural sector
Ft = Rate of external capital investment (1954 dollars)
St = Rate of savings and investment in the local economy
Gt = Transfer payments per capital (1954 dollars)
SSource or sink of conserved material
D State variable defining the storage level of a
Equation controlling the flow rate of a
} Exogenously specified variable
SAuxiliary equation which is computed but has
no material flow associated with it
-- Causal linkage without material flow
Flow of conserved material
Fig. 11. Description of symbols used in the regional model in
Pressat (1972) has described the growth of a closed population
t+At = Pt + Bt,t+At t,t+At (3-1)
By adding a term for net immigration the complete population equation
may be obtained:
t tP +B- D + M (3-2)
Pt+At P tt+t + t+ t,t+At + t,t+At (3-2
P = population
B = births over interval t,t+At
D = deaths over interval t,t+At
M = net immigration over interval t,t+At
t = time, and
At = time step
The rate of births, deaths and immigration may be defined as
b Pt (3-3)
t Pt (3-4)
mt = (3-5)
b = crude birth rate
d = crude death rate
m = net immigration rate
The population equation may then be written in its simulation form:
Pt+t = P + At P (bt dt+ mt)
= Pt [1 + At (bt dt + mt)] (3-6)
The age structure of the population may be generated by using a
separate equation for each age group, adding rate equations to express
the transfer of population sequentially from younger to older age
groups as a result of aging.
The population of Puerto Rico has been divided into nine age
groups, with the population from 0-19 years being broken into 5-year
classes and the remaining population being broken into 10-year
intervals. Smaller age classes have been applied to the younger age
groups because the model is more sensitive to their behavior. Mor-
tality occurs in all age groups, but is so small that it may be
neglected for all ages except infants and the population over 40 years.
Similarly, births are neglected in all but the 15-39 year groups.
The crude rate of fertility, mortality, and migration
(bt, dt, mt) are converted to the age specific rate for each age
group (ibt, idt, imt) by multiplying by the age specific coefficients
shown in Table 4. Thus
Table 4. Age Specific Coefficients for Population Equations.
Age Specific Coefficients
Births Deaths Migration
icb icd icm
0 0.2 0.1
0 0 0.1
0 0 0
1.8 0 0.25, 0.10a
4.0 0 0.45, 0.30a
1.7 0 0.1
0 0.4 0
0 1.5 0
0 8.5 0, 0.25a
second value in 1967.
were shifted from the first to the
I I- 1` 11-~1 1 I--- ~- -- IYI--~-""-- I
ibt = iCb bt
idt icd dt
imt = iCm mt
icb = age specific coefficient for birth rate
icd = age specific coefficient for death rate, and
icm = age specific coefficient for net migration rate
The population of all nine age groups may be described using two basic
equations. The equation for the youngest age group is
1Pt+l = 1Pt + At( bt iPt- d dt 1Pt
+ Icm mt Pt 1Y'Pt) (3-7)
The population of all other age groups is given by
iPt+l = it + t(ix i-1Pt + i m mt iPt
icd dt iPt- y P t) (3-8)
In these equations the coefficients ix and iy show respectively the
gain and loss of population by each age group as a result of aging.
One of the most marked social effects of industrialization and
increasing incomes is that of initiating the demographic transition,
characterized by a rapid decrease in mortality and the retarded
decrease in fertility, as summarized in Table 5. These changes are
attributable to the changes in both the level and sources of income
which occur as a result of industrialization and urbanization.
Table 5. Characteristics of the Demographic Transition.
of Industrialization Fertility Mortality Growth Rate
Nil High High Low
Developing High Decreasing High
Advanced Low Low Low
Source: Heer, 1972.
Decreases in fertility and mortality in the pattern of the
demographic transition have accompanied industrialization in Puerto
Rico, as shown in Fig. 12. Weller (1968) studied changing fertility
patterns in Puerto Rico and concluded that the changing employment
role of the woman outside of the house was a significant factor in
reducing fertility. Friedlander (1965) correlated the educational level
of Puerto Rican women to number of children (Table 6). Myers and
Morris (1966) analyzed fertility reduction in Puerto Rico as a
function of the rural-to-urban migration and determined that the
fertility of mobile women was significantly lower than for those who
had not migrated to the San Juan metropolitan area.
Population Growth Rate
Fig. 12. Historic trends in selected demographic parameters for Puerto
Rico. (a) Estimated annual population at mid-year; (b) An-
nual percentage rate of population increase or decrease;
(c) Annual rates/1,000 inhabitants for crude fertility,
crude mortality, and net migration.
Source: Puerto Rico Health Department (n.d.).
Table 6. Educational Level Achieved vs. Com-
pleted Fertility for Puerto Rican
Mother's Education Number of Children
1-2 years 5.0
3-5 years 4.0
6-8 years 3.4
High school 2.4
Source: Friedlander, 1965.
These and numerous other studies have emphasized that the
framework within which fertility decisions are made is complex. The
child is a non-market good, making it often difficult to analyze
fertility on the basis of economic parameters alone. The cost of
children is a function of the opportunities which the parents must
forego as a result of child-bearing. Family size may change in
response to the changing tastes of the parents for leisure time,
travel, economic goods, etc. The relative cost of children also in-
creases as the family moves from a traditional agrarian setting to
the city. Factors which tend to reduce fertility in the urban-
industrialized setting include: (1) increased cost of living space,
(2) the higher cost of goods in the market system, (3) lack of un-
skilled jobs for child employment such as are available in the
traditional agrarian setting, (4) compulsory school attendance,
_ _I__ __ __ __~
reducing again the child's working ability, (5) availability of
improved pension systems in old age, and (6) changing values, tech-
niques and information on contraception and abortion (Easterlin,
1971; Schultz, 1973; Willis, 1973).
Despite the apparent complexity of fertility decisions, income
and fertility decline are strongly correlated over the post-war
development period in Puerto Rico (Fig. 13). This suggests that in-
come changes may be used as a surrogate for the variety of socioeconomic
changes which characteristically occur during development and which
create the demographic transition. Inasmuch as changing income
levels are perhaps the best overall measure of the development process
(Meier, 1970), this good correlation between increasing income and
decreasing fertility rates during the initial development period is
However, as higher income levels are achieved the relation-
ship between fertility and income becomes increasingly ambiguous,
and nonmonetary factors play a larger role in the determination of
demographic characteristics. Freedman and Coombs (1966) and Willis
(1973) have noted that there is only weak correlation between income
and family size in the United States. Sweezy (1971) has noted that
fertility rates changed imperceptibly in the United States between
the four years preceding the Depression (1925-1929) and the subsequent
four most severe years of the Depression, indicating the lack of
short-term fertility response to income changes. Long-run income and
fertility trends in the United States are summarized in Fig. 14,
showing the uncertainty of this relationship in an industrialized
'* b 44.42 0.0205 Z
40 I. s \ (,2=0.96)
| I94 ^ *- -1
"-.._ 20 -'-- '--------------'-
200 300 400 500 600 700 800 900 1000 110C
S Personal Consumption Expenditures Per Capita (1954 dollars)
Fig. 13. Fertility (b) as a function of personal consumption expendi-
tures (z, used as a proxy for income) in Puerto Rico.
Data sources: Puerto Rico Health Department (n.d.) and
Puerto Rico Planning Board (1978c).
30i Crude Birth R te
"--^ f -
SGNP per Copito-- -
1880 1900 !920 1940 1960
Fig. 14. Comparison of the crude birth rate and gross national
product per capital (in 1929 dollars) for the United
States, showing the uncertain relationship between
income and fertility.
economy. Once the demographic transition is essentially completed,
social factors only poorly related to economic performance appear to
dominate fertility decisions.
A sigmoid curve is hypothesized to be most adequate for
describing the overall long-run fertility reduction process (Coale,
1974), and has been fitted to Puerto Rican fertility data in Fig. 15.
The crude birth rate in Puerto Rico is modeled to stabilize at 15
births per 1,000 of total population, which is approximately the
birth rate required to maintain a stable population. The fitted
bt :1 + e0.123(T- 1965) + 15 (3-9)
T = year
A sigmoid curve was used by Lotka (1924) to predict the long-term
population growth of the United States. However, Lotka's predicted
population levels were in error due to the unforseeable post-war
The crude mortality rate in Puerto Rico declined rapidly
during the late 1940s and early 1950s and soon stabilized at the
level at which it remains today. A rapid decrease in mortality is
characteristically obtained in the early stages of a development
~ o -
r r / *L-
0 oo.0 0 ,r
(UO.I Lndod [P .o 40 000' L/Sfqq QA.L)
a2~J q:iL apn. 3 -
/~S (3j cEa)
t o 4- 4-)
V I- 4- CM
I 2 !=-f- .
a^By q~jig apruo '
process due to the effectiveness of public health services. However,
once these basic services are provided, further increases in income
have little effect on mortality. Although the age-specific mortality
may remain constant, the crude death rate will slowly increase as the
population distribution ages. The crude mortality rate in Puerto
Rico (Fig. 16) has been modeled to decline in a linear fashion and
then to stabilize according to the following relationship:
d947-1953 = 1402.7 (5/7)T (3-10)
d1954-2020 = 7.0 (3-11)
The age specific mortality rate has been defined as a fixed multiple
of the crude mortality rate. Because changes in the mortality rate
will have little effect on the model, the use of a more elaborate
mortality function is unwarranted.
Neoclassical theory predicts the migration of population from
areas of low income and high unemployment to areas of higher income
and lower unemployment, a trend which would tend to generate equilib-
rium levels of unemployment and income nationwide. However, the real
world represents a marked departure from this model and strong
differences in income and unemployment levels persist, including
enclaves of traditionally low income such as Appalachia. Richardson
(1973) reviewed inter-regional migration models, pointing out that
differences in levels of unemployment appeared to be the most
Fig. 16. Crude mortality rate in Puerto Rico.
significant single variable explaining migration, as predicted by
neoclassical theory. However, the characteristically poor statisti-
cal fit achieved by these models indicated that other factors played
an equally important role.
At least four parameters in addition to simple differentials
in income and unemployment need be considered to explain migration
patterns among regions (Isard, 1975; Richardson, 1973). These may
be summarized as:
1. Monetary cost of moving
2. Psychic cost of leaving family, friends and
3. Changes in non-monetary income due to differences
in culture or environmental factors such as climate
4. Uncertainty and the imperfect information on which
migration decisions are based
Some of these factors cannot be quantified, but do help to
explain some of the observed temporal and spatial patterns which have
characterized Puerto Rican migration. Although migration to the
mainland has been underway since the beginning of this century, these
factors combined to prevent large-scale migration until after the
Second World War (Mintz, 1972b). Important factors delaying migration
were the relatively high cost of the passage in relation to pre-war
earnings, and the time lag which was necessary for the initial
emigrants to establish themselves on the mainland to serve as a source
of information and assistance for potential migrants yet on the island.
By migrating from the island to the Puerto Rican neighborhoods
in the New York City area it was possible to minimize the migration
distance with respect to factors such as culture and language differ-
ences. The factor of uncertainty was reduced when family or friends
preceded the migrant and could assist in finding lodging and employment.
This attempt to minimize these migration costs or distances accounts
for the continuing clustering of Puerto Ricans in ethnic communities,
a pattern which suggests that the employment opportunities available to
migrants will be significantly different from the national average.
Migrants to the mainland in the 1950s were composed primarily
of younger persons (Friedlander, 1965), and the net return of migrants
appears to follow a similar pattern (Zell, 1973). It is estimated by
the Puerto Rico Planning Board (1978b) that 69 percent of the net
immigration to the island in 1975 was composed of return migrants, an
additional 18 percent was composed of Puerto Ricans born in the United
States (children of migrants), while the remaining 13 percent repre-
sented immigration from foreign countries. The concentration of
return migrants in the 14-29 year age groups has had a major effect
on the unemployment rate in that group, as summarized in Table 7.
It is not known to what extent this higher unemployment among return
migrants reflects the frictionall" effect of higher job mobility
among returnees or their difference in competitiveness in the local
job market. A portion of the unemployment may also reflect an employ-
ment strategy designed to maximize the receipt of unemployment and
other social welfare benefits (Zell, 1973).
Table 7. Effect of Returned Puerto Rican Migrants on the Percentage
Rate of Unemployment in Puerto Rico, 1972.
Total unemployment rate
Participation rate in
14-29 year group
Percentage of population
in San Juan area
Percentage of unemployment
in San Juan area
Percentage of island population,
excluding San Juan area
Percentage of island unemploy-
ment, excluding San Juan area
Source: Zell, 1973.
Unacceptable results are obtained when the net migration
between Puerto Rico and the mainland is related to simple differences
in the rate of unemployment. The fitted linear equation is
m = 16.1 2.98 Ud (3-12)
m = net migration rate to Puerto Rico, positive values
denoting immigration (migrants per 1,000 of total
U = island unemployment rate (percentage)
Ud = difference between mainland and island unemployment
This equation yielded a correlation coefficient of only 0.37, reflect-
ing the highly scattered nature of the data (Fig. 17).
When the unemployment rate in Puerto Rico alone is substituted
for the unemployment differential a more consistent relationship is
obtained. Net migration may be modeled as a simple linear function of
island unemployment over the 1950-1971 period, obtaining a correlation
coefficient of 0.70 (Fig. 18). The fitted equation is
m = 39.2 4.04 U (3-13)
However, the data from the years since 1972 fall far outside this
historic relationship, suggesting that the parameters influencing net
migration have changed significantly during the 1970s.
It is hypothesized that high rates of net immigration to the
island, despite the highest unemployment rates of the entire development
40 0 -
of- -10 '
m 16.-2.98 U
S11 12 13 14 15
Difference in Unemployment Rates
(Puerto Rico United States, percentage)
Fig. 17. Net immigration to Puerto Rico as a function of the
difference in the unemployment rate in Puerto Rico and
the United States in the preceding year. This relation-
ship achieved a somewhat better linear regression fit
(r2 = .50) than the comparison of unemployment and net
migration during the same year (r2 = .37).
migration during the same year (r2 = .37).
m= 39.2-4.04 U
Fig. 18. Correlation of net immigration to Puerto Rico with the
unemployment rate in Puerto Rico. Data from years 1972-
1977 are plotted separately because they do not fall
within the historic relationship.
i i i I I I I i
period, have been promoted by the growing magnitude of transfer pay-
ments. When transfer payments per capital (in constant dollars) are
added as a second independent variable to the linear migration model,
the correlation coefficient is improved to 0.83. The fitted equation
m = 24.8 0.158 G + 1.55 U (3-14)
G = transfer payments per capital (1954 dollars)
This suggests that income transfers have indeed had a measurable ef-
fect on recent migration.
However, it is not possible to forecast future migration using
the function relating net migration to unemployment as well as income
transfers (Equation 3-14). Use of this relationship would generate
continuing strong net immigration to the island unless a dramatic
decrease in transfer payments occurred, which appears unlikely. It
seems more likely that the population of Puerto Ricans (or others) who
might be influenced to migrate to the island as a result of welfare
considerations will eventually be exhausted, and immigration due to
income transfers will diminish as the supply of "welfare migrants"
becomes reduced. However, there appears to be no manner in which the
total population of potential "welfare migrants" can be estimated, so
an alternative migration model has been used.
Migration between Puerto Rico and the mainland is modeled
on the sole basis of the unemployment rate in Puerto Rico:
m = 39.2 4.04 U + I (3-15)
I = adjustment parameter for immigration rate (Fig. 19)
In order to account for the high rate of net immigration during the
1970s, despite high unemployment on the island, the function "I" has
been used to temporarily shift to the right the curve which was fitted
to migration data over the 1950-1971 period in Fig. 18 (Equation 3-13).
This function reflects that the "zero migration" level of island
unemployment was temporarily shifted due to the dramatic change in
other conditions, such as the presence of increasing levels of
transfer payments and possibly reduced benefits and job opportunities
in New York City during that city's fiscal difficulties in the early
1970s. This has been accomplished by first increasing and then
decreasing the intercept value in the migration equation using the
function values from Fig. 19.
The future value for the function in Fig. 19 is maintained
at a higher level than the pre-1971 period to reflect the effect of
the increased level of transfer payments on migration patterns. Zeli
(1973) has hypothesized that the higher unemployment rate among
return migrants (Table 7) could reflect a strategy of maximizing income
assistance payments, and the empirical data used in Equation 3-14
does suggest that increasing transfer payments have been influential
in promoting return migration to Puerto Rico. If it is true that a
significant portion of the return migrants are being induced to migrate
0 -C U
= 4-> -,
0 Q. O
0 0 4- )
4 4- U
r ., L.
a -n u
U *l- c w
0^) C I-U C S
4- re -
I o S- -c E
2 0 2
E *'- a
S2- *- 0)
S- > 0 0
or (0 fl re
---- ____ ____ ____ ____ ____ ____ -; 3
due to the increased level of income assistance available on the
island, then the rate of unemployment needed to achieve zero migra-
tion is likely to be higher in the future.
Economic relationships are constructed to model historic and
future levels of income and employment in Puerto Rico.1 Only three
state variables (level equations in the DYNAMO language) are modeled
in the economic sub-model. Two of these are for locally owned and
externally owned stocks of productive capital, respectively, and the
third is for consumer assets. The general form for the stock
Kt+1 = Kt + At (St Yt A Kt) (3-16)
S = rate of savings and investment
Y = income (1954 dollars)
A = rate of depreciation
K = stock of capital (1954 dollars)
All other equations express instantaneous values, and as such are
"auxiliary" equations in the DYNAMO language.
1A11 dollar quantities in the model and the text of this
chapter have been adjusted to 1954 value.
Performance of the economic sub-model is determined largely by
the behavior of the two stocks of productive capital and by values of
functions determined exogenously and varied from one model run to
another to generate alternative future scenarios.
In contrast to the population sub-model, in which the
parameters remain constant for all model runs, parameter values in the
economic sub-model are altered for each model run to generate alterna-
tive scenarios. The parameters which are manipulated and the character-
istics of each of these four scenarios are summarized in Table 8.
Scenario "A" is a constrained extrapolation of historic growth
trends, the rate of future local savings stabilizing at zero. Scenario
"B" is perhaps a more probable scenario, reflecting a somewhat lower
growth rate of external investment plus a perturbation in investment
flows in the later 1980s caused by, perhaps, a second petroleum
embargo. Scenarios "C" and "D," respectively, consider the changes
which might accompany low or zero levels of growth in external invest-
ment, combined with increasing levels of local savings and investment.
The specific values for each parameter in each of the scenarios are
found in Figs. 26, 30, and 37 in this section.
Sources of capital. Economic development ultimately proceeds
from human development (Schumacher, 1973). An underdeveloped area is
an area populated by underdeveloped people. The development (or
importation) of human resources makes possible the mobilization of
physical resources which results in the production of wealth. The
Table 8. Input Scenarios for the Puerto Rico Regional Model.
Parameter A Bd C D
Fta High Medium Low Zero
Sb Zero Low High High
CLRc High High Low Low
aForeign investment rate.
bpuerto Rican investment rate.
dIn scenario "B" the stream of investment is perturbed in the
late 1980s resulting in a period of decreased investment, similar to
the effect of the 1973 oil embargo and subsequent recession.
utilization of capital equipment is the key element of the development
process which permits increasing economic output (and resource in-
puts) to be achieved. Industrial capital is that physical structure
which enables human labor to be effectively coupled with the energy
available from fossil fuel and other energy sources to achieve
The development process is characterized by changes in the
structure of the socioeconomic system which creates an increase in
the rate of savings, investment of these savings in the accumulation
of productive capital, and resultant increases in production. Some
of these changes are necessary preconditions to development, whereas
others may be considered a product of that process (Rostow, 1958;
Kuznets, 1958; Gunder-Frank, n.d.).
Physical investment in capital, as evidenced by structures
and machinery, is the last step in the organization of a productive
process. Capital investment may be taken as evidence of the avail-
ability of adequately skilled labor and managerial expertise, the
presence of industrial and other requisite technology, the functioning
of financial institutions, and the anticipated ability to market the
goods produced. When the decision to make productive capital invest-
ment in a region is repeated by numerous entrepreneurs, be they
individual, corporate, or government, then the development process is
begun. This process may be sustained and regular increases in output
achieved when a sufficient rate of capital investment is achieved.
Thus, capital accumulation becomes the focal point of regional or
national development (Lewis, 1958).
The income stream generated by an economy may be divided
between consumptive expenditures and savings, with the latter being
available for investment in productive capital. This capital is
created by foregoing current consumption, thereby permitting a portion
of production to be devoted to the creation of capital stocks which
can be employed to increase future levels of production.
There are large differences in the tendency for various
sectors of society to accumulate savings. The top income group
accounts for most personal savings, and lower income groups have low
or even negative savings rates. However, total personal savings is
not large when compared to business savings, which constituted an
average of 71 percent of total savings in the United States over the
1970-1977 period. Lewis (1958) has proposed that when the rate of
capital accumulation in an economy is low it may be traced to the
small size and/or low profits in the productive sectors which provide
the capitalists, be they private or government, limited resources
and poor incentive for re-investment.
Merhav (1969) has noted that the central issue lying at the
root of economic development theory is the problem of escaping from
the low-level equilibrium trap. In this trap an economic system
cannot achieve positive real per capital growth because of the in-
adequacy of productive investment, which in turn cannot be achieved
due to the low level of per capital income and savings. This low-
level trap has been described for Puerto Rico by Friedlander (1965),
and is summarized in Fig. 20.
W \ Without C1
S-'- Population Growl
0 With Emigratiol
INCOME PER CAPITAL
Fig. 20. Low-income trap.
Source: Adopted from Friedlander (1965).
The classic low-income trap is created when the rate of
population growth in an economy grows faster than the economic growth
rate, resulting in continual impoverishment and failure of the local
economy to generate capital resources needed to finance higher rates
of economic growth. Development theory holds that the trap may be
avoided by increasing national investment by using the input of foreign
capital, or through the control of population growth rates (i.e., by
exporting migrants in this case). Both methods have been applied in
The upward sloping curve of investment represents the increas-
ing rate of investment which is assumed to occur as a result of a
positive marginal rate of savings in response to income increases.
The downwards concavity of the curve reflects a hypothesized decrease
in the marginal productivity of capital at higher levels of economic
development. The manipulation of capital and investment is the
critical variable for the "escape," and is a central parameter in
the formulation of many development models and policies.
Friedlander (1965) postulated that Puerto Rico could avoid
this trap by achieving rapidly increasing income levels through
externally financed industrialization, and achieving population
stabilization through migration. The resultant income increases were
anticipated to create an increase in the savings rate until the local
investment resources became sufficient to finance the development
process. That the rate of savings should increase as a function of
income is suggested by cross-sectional data, such as shown in Fig.21.
S15 ; ----------------;
10 1- ,
C 5 10 5
(thousands of current dollars/year)
Fig. 21. Cross-sectional savings characteristics of Puerto Rican
families as a function of family income.
Source: Committee to Study Puerto Rico's Finances, 1975.
However, the extrapolation of cross-sectional savings behavior
to income increases over time is not valid, as can be inferred from
previous experience elsewhere as well as from post-war patterns of
savings in Puerto Rico. Long-term income and savings data for the
United States and Puerto Rico, graphed in Fig. 22, show that the per-
centage of income saved has not increased regularly as a function of
income increases over time. The post-war rate of savings in Puerto
Rico has declined steadily, showing clearly that increases in income
have not increased the rate of savings in the local economy.
Estimation of capital stock. The total annual addition to
capital (gross domestic investment) consists of additions to fixed
capital plus changes in inventory. The net addition to fixed capital
may be computed by discarding inventory changes and subtracting the
annual depreciation from the gross addition to fixed capital. Housing
stocks are included in the estimate of productive capital, being
necessary to produce housing services, the value of which is reflected
in the national income accounts.
An official estimate of capital stock is not prepared by the
Puerto Rico Planning Board, but an estimated capital stock series
(excluding land) has been constructed for the island (see Appendix A)
and is summarized in Fig. 23. Total productive capital is divided
into two classes, that which is locally owned (aK) and that which is
externally owned (bK), as imputed by earnings data, assuming that
the earnings rate of locally and externally owned capital is identical.
The net inflow of external capital into Puerto Rico is modeled as a
10 o o
a -8 o o
S 11940 7
S1977 Puero Rico
Personal Disposable Income per Capita
Fig. 22. Relationship between personal disposable income per capital
(in constant 1954 dollars) and percentage of income appear-
ing as personal savings for the United States and Puerto
ing as personal savings for the United States and Puerto
show9 ng total cdCapital Stock oc w d
940' o P6 s f0
Fig. 23. Capital stock series for Puerto Rico (excluding land),
showing total capital stock and capital stock owned by
residents of Puerto Rico, as imputed from earnings.
contribution to the stock of externally owned capital, while local
savings and investment is contributed to the stock of locally owned
capital (it is assumed that savings equals investments). The
depreciation rate is the same for these two stocks of capital. The
aKt+ aKt + At(S Y aKt Ak) (3-17)
bKt+l bKt + At(F bKt Ak) (3-18)
S = rate of local savings or investment
Ak = capital depreciation rate
Y = GNP of Puerto Rico (1954 dollars)
F = rate of foreign investment (1954 dollars/year)
aK = locally owned capital (1954 dollars)
bK = externally owned capital (1954 dollars)
Earned income, minus savings, plus transfer payments (G), is
equivalent to local consumption, and is contributed to the stock of
local consumer assets (cK):
cKt+l = cKt + At[Y(1- S) + G Ac cKt) (3-19)
These consumer assets represent a mixture of consumer durables and
non-durables, not including housing (which is counted as productive
capital). A depreciation rate of 50 percent is estimated for this
stock. Inasmuch as this stock has no effect on any other model
parameter, inaccuracies in this parameter will not affect model per-
External sources of capital. The capital required to finance
investment activities can come from two sources: savings from within
the economy, or some external source of savings. A bottleneck to
economic growth can be created by the insufficiency of local invest-
ment resources and the propensity to invest the available savings in
relatively non-productive endeavors (e.g., real estate). This bottle-
neck may be broken perhaps most readily by importing financial capital
and entrepreneurship. Having experienced poor results in the govern-
ment financed and managed industrialization projects in the late 1940s,
the importation of capital and entrepreneurship became the development
policy adopted by the island's government (Ross, 1966). Relatively
little emphasis was given to the development of local entrepreneurship
The time series of Gross Domestic Investment on the island is
portrayed in Fig. 24, showing the increasing role of external invest-
ment as a source of capital compared to the role of local savings,
which have actually become negative in recent years. The computation
of the percentage contribution of each major source of savings to total
investment on the island is plotted in Fig. 25, showing even more
clearly the poor record of local savings formulation and the heavy
reliance on external capital sources.
Y E A R
Fig. 24. Sources of Gross Domestic Investment (GDI) in Puerto
Rico, showing the contribution from local and external
Data source: Puerto Rico Planning Board (1978c).
1950 1960 !970
Fig. 25. Gross Domestic Investment in Puerto Rico, showing the percen-
tage contribution of four sources of savings: (A) net ad-
justed inflow of external capital, (B) depreciation,
(C) government savings, and (D) private savings. Note that
private savings is plotted with respect to zero, and all
others are plotted with respect to private savings.
Data source: Puerto Rico Planning Board (1978c).
Local investment \t
The annual rate of foreign investment has been modeled as an
exponential growth equation which has been constrained by a decreasing
rate of growth:
Ft = $850x106 3 + 6e-.03(T-1979)(T-1979) (3-20)
Ft = annual external investment (1954 dollars)
6 = initial value for annual capital growth rate
The decreasing rate of investment growth reflects the hypothesis that
many of the conditions which contributed to Puerto Rico's rapid post-
war economic growth rate are transient in nature, and are weakening as
a result of increasing local wage levels, increased local energy cost,
and changes in industrial tax incentive programs.1 The values of Ft
used in each scenario are summarized in Fig. 26.
Although capital importation has generated rapid economic
growth, it also has created external repayment obligations against the
stream of business profits since most entrepreneurs were nonresident
individuals or corporations. Not unexpectedly, net property earnings
remitted to nonresidents have increased dramatically, from 7 percent
of total property income on the island in 1947 to 62 percent in 1977.
In 1962, externally owned industries accounted for 60 percent of the
1The future of the tax incentive program is uncertain,
particularly inasmuch as it is intimately linked to the issue of
statehood. Were Puerto Rico to become a state it would no longer be
possible to offer federal corporate income tax exemption as an
Extern!a Investment (Ft)
/ I -c
Fig. 26. Scenarios for the rate of external investment in Puerto
Rico, for use in the regional model.
_ I I I
total value added in manufacture on the island, and this figure had
risen to 70 percent by 19671 (Curet-Cuevas, 1976).
The rate of profit remission is calculated as the product of
the value of the externally owned capital and the rate of return
earned by this capital:
Ht = bKt(.12 + 0.113e--07(T-1946)) (3-21)
Ht = remission of profits (1954 dollars)
The historical and projected change in the rate of earnings from ex-
ternally owned capital in Puerto Rico is summarized in Fig. 27.
Local sources of capital. Low levels of local savings have
plagued the island economy throughout its entire history, having roots
in the colonial era when business profits and savings were remitted
to Spain (Lewis, 1963). Restricted savings rates continued to be a
problem of post-war development (Comitd Interagencial, 1975; Curet-
Cuevas, 1976; MacCoby and Fielder, 1972). There is no clear explana-
tion for the low and decreasing level of local savings and investment,
which have decreased despite rapid economic growth (Fig. 28). Suggested
reasons for low or negative rates of personal savings2 have been listed
by the Committee to Study Puerto Rico's Finances (1975, p. 27):
1More recent data are unavailable.
2Includes savings from unincorporated businesses.
y= 12+ 11.3e92
Fig. 27. Historic and projected rate of return on externally owned
capital in Puerto Rico. Regression equation fitted over the