MODE OF TRANSPORTATION AND THE
INCOME-DISTANCE TO WORK RELATIONSHIP
IN PIRACICABA, SAO PAULO, BRAZIL
CYRUS BASSETT DAWSEY III
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
To Marshlea and Cyrus B. Dawsey, Jr.
Many people have been involved with this research project. To
name all within these limited confines would be impossible, but some
deserve special recognition. In the first place I would like to thank
the members of my supervisory committee: Dr. Raymond E. Crist,
Dr. David L. Niddrie, Dr. Carl W. Spurlock, and Dr. Robert W.
Bradbury. They have provided constant assistance with general
theoretical problems as well as mundane comma splices and split
Secondly, I am appreciative for the help received in Piracicaba.
Dr. Delmar Marquette, Professor at Escola Superior Agricola Luiz de
Queiroz, provided the airphotographs of Piracicaba. Dr. Adilson Maluf,
the city's mayor, Sr. Raul Elu, a public official, and many other govern-
mental workers cooperated with the project by acceeding to every request.
The hospitable residents of Piracicaba generously relinquished their
privacy and gave of their time to answer the inquiries of the questionnaire.
Finally, I am grateful to my family for their constant support.
My parents, Marshlea and Cyrus B. Dawsey, Jr., have always been a
source of help and inspiration. For this research my wife, Barbara, and
children, Marc and Amanda, have lovingly sacrificed time and attention
which was by right their own.
To these people and to many others, therefore, I am deeply in debt.
TABLE OF CONTENTS
ACKNOWLEDGMENTS .............. ... ... .... iii
LIST OF TABLES ........ .......... ... .... vi
LIST OF FIGURES ............................ vii
ABSTRACT .... ... . ..... . ..... ... ... x
I RESEARCH SCOPE AND OBJECTIVE . . . ... 1
Introduction .. ... .. . .. .. . .... 1
Research Objective .. . . . . . . . 5
II INCOME AND DISTANCE TO WORK: ANGLO-
AMERICA ........... ........ ... 8
The Journey to Work ... .. . . . . .... 8
Evolution of the Journey to Work . . . . 9
The Gravity Model .... .. .. ........ 12
Methods of Minimizing Distance. . . . . 15
The Bid-Price Model . . . . . 17
Development of the Bid-Price Model ........ 18
Income and the Slope of the Bid-Price Function 20
Multiple Work Locations . . .. . ...... 24
Empirical Research . . . . . . . .. 30
III INCOME AND DISTANCE TO WORK: LATIN
AMERICA .. . . . . . . . . .. .. 34
Non-Western Residential Pattern . . .. . 34
Possible Causes of the Non-Western Pattern . 35
Changes in the Non-Western Pattern . . .. 37
Transportation as an Explanation of the
Non-Western Pattern . . . . . .... 38
Time Costs and the Bid-Price Model . . . ... 41
Research Hypotheses .. . . . . . . . 46
TABLE OF CONTENTS (Continued)
IV DATA COLLECTION IN THE STUDY AREA . . .. .47
Piracicaba . . . . . . . . . . . 47
History and General Characteristics ...... 47
Piracicaba as the Study Area . . . . .. 50
Data Collection .. .. . .. . .. . . 55
Variables to be Measured . . . . .... 55
Interviews . . . . . . . . . . 57
Property Survey Data . . . . . .... 61
Definition of the Effective Statistical Sample . 63
V RESULTS ............. .. ..... 66
General Description .. ..
Residential Areas .. ...
Places of Work .. ...
Statistical Analyses .. ....
Distance to the City Center
Distance to Work .. ..
Expanded Model .. ...
Conclusions .. . ......
APPENDIX A NORMATIVE BID-PRICE MODEL;
TRANSPORTATION COST MONEY . . . ... .132
APPENDIX B NORMATIVE BID-PRICE MODEL;
TRANSPORTATION COST TIME . . . ... .140
APPENDIX C QUESTIONNAIRE . . . . .... 150
BIBLIOGRAPHY ....... ... .. . . ......... 152
BIOGRAPHICAL SKETCH ....................... 170
. . . . . . . .
. . . . ., . .
. . . . . . .
. . . . .
. . . .
LIST OF TABLES
1 Urban Population of the City of Piracicaba ...... 49
2 Correlation Matrix of Income and Surrogate
Measures of Economic Status . . . . . ... 63
3 Work Situation of Respondent Household Heads .... .64
4 Transportation Used for the Journey to Work by
the Respondents Who Had a Fixed Place of Work
and a Steady Rate of Income . . . . . .... 65
5 Passenger Cars Licensed in Piracicaba. . . . ... 94
6 Sample Means and Estimated Parameters for
Income-Distance to the City Center Relationships. . 103
7 Sample Means and Estimated Parameters for
Income-Distance to Work Relationships . . . ... 106
8 Estimated Parameters for Income-Distance to
Work Relationship for Walkers Before and After
the Inclusion of the Length of Work Variable ..... .110
9 Estimated Parameters for Income-Distance to
Work Relationship for Walkers Before and After the
Inclusion of the Length of Residence Variable. ..... .112
LIST OF FIGURES
1 Bid-Price Curves for Rich and Poor Individuals;
commuting costs are monetary . . . . .... 21
2 Bid-Price Curves of People at 5 Different Income
L evels . . . . . . . . . . . . . 23
3 Bid-Price Curves of Workers at 3 Different
Employment Sites. ........... . ... .. .25
4 Bid-Price Curves of Rich Workers at E and Poor
Workers at El, E2 and E3 . ............... 27
5 Bid-Price Curves for Rich and Poor Individuals;
commuting costs are temporal . . . . . .... 44
6 City Center . . . . . . . . ... ..... . 51
7 Southwest Panorama ................... 52
8 Water Falls ................... ... .. 53
9 Pumping Station...................... ... 54
10 General Map of Piracicaba . . . . . . . .. 56
11 Doctor's Home .......... ... ....... 67
12 Business Executive's Home. . . . . . . ... 68
13 Resident of Sao Judas Tadeu Section . . . . .. 70
14 Sao Judas Tadeu Residential Block. . . . . ... 71
15 Backyard . . . . . . . . . . . . 72
16 Peripheral Poor. . . . . . . . . . ... 73
17 Rented Rooms . . . . . . . .... ... . 74
LIST OF FIGURES (Continued)
18 Communal Water . . . . . . . . ... .. 75
19 Poor Family in a Mixed Neighborhood . . . ... 78
20 Homogeneous Architecture . . . . . . .... 79
21 Heterogeneous Architecture . . . . . .... 80
22 Beira Rio . . . . . . . . .. ... . 81
23 Swimming Area ... ..................... .. .82
24 Dopla . . . . . . . ........ 84
25' Jardim Primavera .................... 85
26 Siderurgica Dedini . . . . . . . .... . 87
27 Metalurgica Dedini . . . . . . . .... 88
28 Plant Expansion . . . . . . . . ... . 89
29 Boyes . . . . . . . . . . . . . .90
30 Engenho Central . . . . . . . . ... .. 91
31 Chair Factory. . . . . . . . . . .. 92
32 Luiz de Queiroz University . . . . . .... 93
33 Bus Terminal....................... 95
34 Taxi Point . . . . . . . . . . . 96
35 Parking Meters . . . . . . . .... ... .. 97
36 Bicycle Commuter ..... ............ 98
37 Repair Shop . . .. . . . . . .... .. 99
38 Charrete Stop . . . . . .. ...... .... 100
LIST OF FIGURES (Continued)
39 Income-Distance to City Center . . . . .... 104
40 Changes in the Income-Distance to Work Relation-
ship for Walkers After the Inclusion of Length of
Work or Length of Residence Variables . . . ... 111
41 MainStreet . . . . . . . . . . . . 119
42 Market ...... .. ... ... . .. ... 120
43 General Store. . . . . . . . .. . . 121
44 Shoppers . . . . . . . . . . . . . 122
45 Vendors. . . . . . . . . ... ....... 123
46 The Coming of the Automobile I . . . . ... 127
47 The Coming of the Automobile II . . . . .... 128
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
MODE OF TRANSPORTATION AND THE
INCOME-DISTANCE TO WORK RELATIONSHIP
IN PIRACICABA, SAO PAULO, BRAZIL
Cyrus Bassett Dawsey III
Chairman: Dr. Raymond E. Crist
Major Department: Geography
Empirical evidence suggests that a positive relationship exists
between income and distance to work within cities of Anglo-America.
The relationship appears to be negative in Latin American cities. A
normative bid-price model which applies Von-Thunen's framework to
residential land use can be transformed into a model of interaction
between the residence and place of work. In this form the model
describes the income-distance relationship characteristic of Anglo-
America when distance overcoming costs are measured in monetary
units. The model describes the relationship characteristic of Latin
America when the costs are measured in units of time.
Hypotheses of a positive relationship between income and distance
to work for automobile using commuters and of a negative relationship
between income and distance to work for commuters who use other modes
of transportation were tested in Piracicaba, Sao Paulo, Brazil. Linear
regression analyses of data collected from interviews of a 4 percent
sample of the city residents showed that the hypothesized relationship
could be inferred only for the group who walked to work. When length
(in years) of residence was added as a measure of utility maximization,
the significant relationship became not significant at the 05 level.
The research hypotheses were rejected. A negative relationship
between income and distance to the city center was significant at the 05
level for all groups differentiated according to the mode of transportation
Relative location with respect to the city center was concluded
to be more important than to the place of work. The negative income-
distance to the city center relationship for automobile users was
concluded possibly to be the result of a low marginal utility of residential
land. The results imply that the characteristic negative income-distance
to the city center relationship will not be changed by increasing dependence
on the automobile in Brazil.
RESEARCH SCOPE AND OBJECTIVE
A dominant characteristic of man has been his tendency to seek
out his own kind. Since the beginning of their presence on earth, people
have protected and enriched themselves by clustering in groups, and
ultimately towns evolved. From the small early villages in China, Egypt,
Babylonia and India, settlements have developed into the giant metro-
politan areas of today.
For primitive man, gathering into groups was probably natural.
He, just as the primates, compensated for physical weakness by hunting
and fighting along side others of his own species, and through a process
of natural selection, this efficient behavior survived. Those who opted
for societal life more often lived on to transmit their culture to offspring.
Although the exact definition of a city varies (Mayer, 1971), the
growth of early settlements involved many complex factors. Improve-
ments in agricultural techniques created food surpluses, and groups of
people not directly involved in subsistence production could be supported.
At physically favorable locations, people gathered to exchange goods,
so that a professional merchant elite eventually emerged. The in-
creasing knowledge was dealt with by an educational elite; rites and
cultural functions were conducted by a religious elite; and the whole
agglomeration was ruled by a governmental elite. The growth of
cities, therefore, was related to technological innovation and increased
productivity as well as to professional diversification and specialization.
History since the industrial revolution has been marked by an
ever increasing shift of people from rural to urban areas. In 1800,
only 3 percent of the world's population lived in places of 5, 000 people
or more, and 1. 7 percent lived in places of 100, 000 or more. By 1950,
over 30 percent lived in places of 5,000 or more while 13 percent resided
in cities containing 100, 000 or more inhabitants (Hauser, 1967, p. 7).
Although most of the people of the world still live in rural areas, the
trend toward urbanization is evident.
The growing importance of the urban area as a home for man
has led to increased interest in cities. Curiosity has been aroused by
apparent regularities and differences encountered among and within cities.
What, for example, causes the sprawl of Los Angeles or the high density
of Calcutta? Why is Paris the fashion capital, or why does Detroit
produce automobiles? What causes financial executives to seek downtown
offices while supermarket operators locate in suburbs ?
Curiosity has not been the only reason for focusing attention on
cities. While a growing number of people are choosing an urban area
for a home, cities are less than ideal environments. Much investigation,
therefore, has been aimed at the solution of particular problems. To try
to improve the quality of urban life, researchers look for answers
to questions like: why is social pathology associated with certain racial
and economic groups; what causes squalid squatter settlements around
large Latin American cities; or how can visual and aural blight as well
as water and air pollution be controlled?
The study of cities can be divided into two categories: inter-urban
and intra-urban. The first concerns the external relationships of a city.
These would be the city's association with other cities, ties to a rural
hinterland, or how the town fits into a larger regional or national pattern.
Intra-urban study, on the other hand, involves internal characteristics
of an urban place. Interest is concentrated on the component parts of a
city and how they interact with each other.
Although inter-urban studies are an important field of investigation
(witness the attention devoted by geographers to Central Place Theory),
the scope of the research presented in this paper is intra-urban. It
concerns events and characteristics found within cities.
Intra-urban studies are conducted by researchers from various
fields. History and geography are different from most social sciences
in that they focus on a dimension rather than a particular subject or
activity. While sociologists study group behavior and economists study
economic behavior, historians and geographers look at any or all activity
through time and over space. This spatial aspect is one of the
characterizing features of the field of geography.
The orientation of this research is geographic. Although some
of the concepts and models are developed using economic theory, the
basic dimension of interest is spatial. How certain factors vary over
physical space and distance is of primary concern.
One of the factors which varies over space is residential land use,
and much attention has been devoted to the study of where people live
within cities. Some researchers are content to map or describe
residential patterns (Jones, 1931; Applebaum, 1952; Niedercorn and
Hearle, 1964; Loewenstein, 1963), but others are concerned with dis-
covering regularities from one city to another. These regularities can
be stated in terms of descriptive models. The concentric zone (Burgess,
1925), sector (Hoyt, 1939), and multiple nuclei (Harris and Ulman, 1945)
models are the ones most often cited, and recent studies show how the
patterns described by each of these is related to different factorial
dimensions found within cities (Anderson and Egeland, 1961;
Berry, 1965; Simmons, 1965; Murdie, 1969).
Regularities in residential location, therefore, appear to exist
in urban areas. What causes these patterns? Why do people live
where they do? The residential location decision is a very complex
psychological process that involves a variety of known and unknown
factors. Basically, however, it concerns the evaluation of information
about the perceived environment. Each individual considers desirable
and undesirable residential characteristics, compares them to the
perceived locational possibilities, and makes the decision of where to
This decision need not be a logical one. What is or is not
desirable to a given person depends on what his physical and psychological
needs are. These needs may be different from those of other people.
Furthermore, the decision is based on a perception of reality rather than
the real world itself. This perception also varies from one individual
to another (Wolpert, 1966; Pred, 1967 and 1969; Wood, 1970).
Many features of the site itself might weigh heavily in the process.
Characteristics of the land and the house are obviously important factors.
Another group of considerations involves relative location. The proximity
to a polluting factory or a busy street, for example, might be negative
features of a particular location, while nearness to a shopping center,
the children's school, or a parent's job could be positive aspects. Proximity
to a large number of urban features, therefore, is an important component
of the locational decision.
The aggregate location of residences within cities, is the sum of
a large number of individual decisions, each of which is based on a
variety of considerations. Site considerations are important, but so too
are those pertaining to the relative location with respect to other features
found within the city. Proximity to certain items is avoided while
accessibility to others is sought.
The objective of this study is the investigation of the relationship
between residential sites and one of the items to which accessibility
is desired: the place of work. Empirical evidence indicates that
accessibility to the place of work may be more desirable for some
income groups than for others. A relationship, therefore, appears to
exist between income and the distance people are willing to travel to
The income-distance to work relationship, however, is not the
same for people living in industrialized countries and developing areas.
Research conducted in Anglo-American cities shows that distance usually
increases with income; poor workers live nearer their jobs than rich
individuals do. Descriptive accounts of Latin American cities, however,
indicate that the pattern in these towns may be the opposite with poor
people travelling farther to work than wealthier individuals. The
relationship between income and distance to work, therefore, appears
to be positive in industrialized societies and negative in developing ones.
The study presented here considers a possible explanation for the
difference between the income-distance to work relationship characteristic
of cities in Anglo-America and Latin America: the type of transportation
used for the journey to work. The following specific question is asked:
is the type of transportation used for the trip to work important in
determining the nature of the relationship between income and distance
to work? If transportation mode is indicated to be an important variable,
then it must be included with income in an explanation of the distance
people choose to live from their jobs.
Partial causality is implied. Income and transportation type
influence distance to work which is in turn related to where people
live. If a change in the income-distance relationship occurs when a
change takes place in the type of transportation used, then the inclusion
of the new variable, mode of transportation, adds some information to
a general understanding of why people live where they do. Transportation
mode can thus take part in the explanation of residential location.
INCOME AND DISTANCE TO WORK: ANGLO-AMERICA
The Journey to Work
As stated in Chapter I, relative location is a factor in the
locational decision. While proximity to various features within a
city might be desirable, accessibility to the place of work is one of
the most important considerations.
The ebb and flow of the working force is the single most important
component of movement within cities. In Chicago the number of trips
originating from home going to work is almost twice as large as the
next highest destination category (Chicago, 1959), and in Toronto
roughly half of all trips involve job commuting (Toronto, 1966). Other
studies place the figure for various cities at 40 percent and higher
Many aspects of the journey to work are being widely investigated
(Wheeler, 1969a) since commuter flow is related to many urban problems.
The concentration, for example, of travel at peak hours leads to in-
efficiencies in the use of transportation networks. Social (Wolforth,
1965) and marginally increasing economic (Walters, 1961) costs
associated with traffic congestion have led to the advocacy of variable
pricing schemes. These would discourage the use of highways at peak
hours (Vickrey, 1963). On the other hand, the financial viability of
various public transit alternatives to the automobile have been compared,
and the present bus system is seen as being the most efficient in the short
run for most cities (Meyer, Kain, and Wohl, 1965).
Evolution of the Journey to Work
The journey to work is the result of the separation between residence
and job, and it is a recent phenomenon. In earlier times production
was small scale, and people worked at home or next door in small shops.
In-many instances gainful employment was not distinguishable from house-
hold activity (Sjoberg, 1960).
The industrial revolution brought increased capitalization and plant
size as producers took advantage of economies of scale. As industry
became larger and more complex, a greater and more diverse labor
force was needed at each plant (Liepman, 1944). Domestic functions
gradually became different from income producing work, and, with
improved means of transportation, the activities became physically
Small, pre-industrial domestic or shop industry gave way to
larger factories which were ringed by worker's houses. These mills
in turn were gradually replaced by complex plants which were spatially
removed from the residences of the labor force. The process was
shown in a study of Chorley, England (Warnes, 1970) for the period
1780-1850, where distance to work is compared to two characteristics
of industry: size (positive relationship) and age (negative relationship).
Workers tended to cluster around older and smaller mills but not
newer, larger ones.
Improvement in transportation technology was the main factor
involved in the separation of workers and their jobs. Streetcars, rail-
roads, buses, and the automobile made it possible for people to commute
to work. They could cover large distances in a short period of time,
and there was, therefore, no longer the necessity to live next door to the
factory. Technological change brought about a reduction in the costs
associated with overcoming distance, and spatial interaction over a much
larger area was made possible (Janelle, 1971).
Not only did workers move away, but industry as well decentralized.
The use of trucks and electricity allowed plants to break away from sites
favored by railroads or rivers. Physical proximity to such features was
no longer necessary.
The separation of the home from the place of work, occasioned by
reductions in the costs of transportation, has led some authors to contend
that the relationship between the residence and the job is becoming
weaker (Wolforth, 1965; Boyce, 1969; Forrest and Tan, 1970). The
argument can be made, however, that what has changed is the "form"
of the journey to work, and not the "process. "
Theoretically, form refers to the static characteristics or
morphology of events, while process refers to changes in these events
through time. Ultimately, everything can be defined as form--where
the change in time is 0--and process--where the change in time is
infinite. The static or form feature of residential location, for example,
is concerned with where people live and it can be mapped or described
by using the concentric zone, sector, or multiple nuclei models. On the
other hand, residential location process refers to where people move
(i.e., change residence) (Rossi, 1955; Moses, 1962; Boyce, 1969).
The journey to work is by and large a process. It occurs twice
daily for most individuals, and the aggregate volume of flow varies
greatly from hour to hour. Spatially, it is the process of interaction
which occurs between the two locations; home and work.
The form of the journey to work, or the physical distance between
home and employment, has indeed grown longer. At the aggregate level,
commuting occurs over a much broader area than it did in the past. The
increase in distance does not mean, however, that the travel to work
process has changed. Perhaps some modification has taken place
because of a shorter work week or owing to increased lunch box use but
not as a result of changes in the cost of transportation.
If a century ago a causal link existed between residential location
and accessibility to work, the influence of one upon the other is equally
present today. Patterns are simply blown up to a larger dimension
involving greater physical distances. A decrease in the cost of trans-
portation does not mean that the basic shape or the strength of these
patterns has changed, and proximity to work should still be important
in determining residential location. Accessibility is why people live
in cities, and it follows, therefore, that nearness to work is a major
factor. Most urban travel is, after all, made up of commuters.
The Gravity Model
A series of models is often used to describe the relationship between
the residence and the place of work. These are various forms of the
"gravity" or "potential" model originally conceived by Zipf (1947). The
main function of the models is the prediction of the volume of interaction
which takes place between two or more locations. The models are
allegorical in that they apply physical science relationships between
mass and distance to human events. The general form of the gravity model
is as follows:
I.. = +M
I = interaction between locations i and j.
M. = measure of mass (population, for example) at i.
Mj = measure of mass at j.
d.. = distance between i and j.
b = empirically derived exponent.
With basically the same format, the gravity model is expanded
to include the actual or possible interaction between an area i and all
other areas j. In this form it is referred to as a "potential" model.
The use of such a framework in a number of studies indicates
that as distance increases the volume of interaction does decrease. Even
though the model has been applied to interaction between home and other
possible destinations (Huff, 1963; Wheeler and Stutz, 1971; Hale, 1973),
most interest has centered on the link between the residence and the
place of work.
Empirical research shows that the number of employees at a
given place of work decreases as distance from that place increases.
The number, for example, of people working in factories in Massachusetts
declines significantly with distance from the plant after the first two or
three miles (Carroll, 1949). Much later, Getis (1969), using a
probabilistic approach, wrote that within a certain radius of the place
of work a frictionlesss" area exists within which workers are indifferent
about their place of residence. Outside this area the decay of distance
occurs as predicted by the gravity model.
Distance is also minimized by commuters in Montreal who tend
to select the shortest possible route to work (Scarlett, 1970). Although
people of Athens travel farther to work than they do for other reasons,
the number of trips decreases significantly with distance (Pappas, 1970).
Hecht (1973) discovered that individual stress increases when the length
of the journey to work increases.
Distance to work is frequently used as a variable in the prediction
of future growth. A potential model can describe present accessibility
of urban sectors with respect to the working force, and future growth
in each sector can be evaluated when a parameter of total regional growth
is included. Development would depend on the number of employees who
commute from a sector and how far they must travel to their jobs. The
farther they go, the less likely it is that their home sector will grow
Gravity models are often used for planning purposes. Predictions
of future growth are based on a series of input factors including present
accessibility between employment and employees. Lowry (1968) has
summarized the characteristics of the most important of these models
and discussed drawbacks and promises involved in their use.
In order to increase its predictive value, the gravity model is
continually being modified. The inclusion of items such as residential
attractiveness (Wilson, 1969), and intervening opportunity (Schneider,
1959), as well as the mating of the model to probability and game theory
(Malm, Olsson and Warneryd, 1966) are all attempts to better relate
the framework to reality.
Problems of the gravity model have been widely recognized
(Lowry, 1964, p. 22; Colenutt, 1970, p. 116; Berry and Horton, 1970,
p. 493). The major criticism has been that the model is basically
descriptive. It is used as a tool for making predictions of the future
based on present empirically determined relationships. This use is
limited because predictions are dependent on current or past conditions
and not on general explanatory concepts.
The theoretical underpinning of the model consists of laws
borrowed from physics and of an informal knowledge concerning the
costs of overcoming distance. This knowledge leads to the conclusion
that the statement of the model--that interaction decreases as
distance increases--is obviously true. The nature, however, of the
various costs is not dealt with, and the model says nothing about how
distance is related to human behavior in general.
Methods of Minimizing Distance
Studies of interaction using gravity type models do prove that a
relationship exists between the residence and the place of work. While
the commuting distance has increased over the years, the journey to
work is still a strong link which shapes the character of cities. The
model indicates that the distance between home and job tends to be
minimized but it does not state how this is accomplished. The possible
means of shortening the distance are: a change in the residential site,
a change in the employment location, and a change in the journey itself.
Barring changes in transportation technology, the journey itself
should already be optimal. The worker already uses the shortest
route or the quickest and cheapest means of getting to work (Scarlett,
Changes in the location of employment are very possible. Harris
(1954), and Vance (1966) have argued that the labor component is an
important factor which is often overlooked in industrial location theory.
The historical spreading of job sites, discussed above, is a contemporary
process. Taaffe, Garner, and Yeates (1963) showed that in Chicago many
firms have moved to locations in peripheral residential areas. Although
many workers still go downtown, a large proportion of them commute to
these non-central jobs. The trend has been confirmed in Sydney (Logan,
1968) and London (Daniels, 1973).
A change in the place of work, therefore, is recognized as a
possible distance minimizing response. Workers can change jobs
and employers can change the site of their plant or office. If employers
locate according to where people live, the distance to work does not
explain residential patterns. It explains the location of jobs. Investigating
changes in employment opportunities and labor migration in Los Angeles,
Burns (1964) found a strong correlation between the two but no time lag.
This would have possibly indicated which causes the other by showing
which occurs first.
Most researchers seem to agree that the residential location is
more flexible than that of the place of work, and they have used the
commuting distance to explain residential land use patterns (Carroll,
1952; Duncan, 1956; Kain, 1962; Taaffe, Garner and Yeates, 1963;
Goldstein and Mayer, 1964; Loewenstein, 1965; Wheeler, 1967, 1968b;
Halvorson, 1973). Lowry (1968), in formulating his planning model,
allowed the residential site to be a free variable; one that responds to
the location of industry and commerce.
The study presented here considers the mode of transportation
variable as being a factor which modifies the relationship between income
and the distance to work. Other research indicates that the commuting
distance is most often, although not always, modified by changing the
place of residence. If this is true, and if transportation mode does
influence the income-distance relationship, the results of this investigation
should add more to an understanding of the location of residences than
The Bid-Price Model
The major function of the gravity model, as applied to the
journey to work, is the description of empirically observed relationships
which exist between the volume of movement and distance. The model,
however, states nothing of how the various costs associated with distance
are related to human behavior. Why, for example, do people tend to
minimize the amount of time devoted to commuting?
A group of regional economists and urban geographers are
increasingly investigating the journey to work from a different
perspective. They use the normative approach. Rather than seek
aggregate patterns or relationships in nature, these researchers are
concerned with developing a deductive model of what man does under
The advantage of the normative methodology is that general theory
is built into the model, and "why" or "how" questions are answerable.
The problem is that if the general theory is incorrect, the model
must likewise be incorrect. The test of validity is whether or not
the explicit or implicit assumptions of the model apply to real situations.
Development of the Bid-Price Model
One normative residential land use model is an urban application
of Von Thinen's ideas (Hall, 1966) about the rural landscape. The value
of property in the city, as in the country, is said to be tied to benefits
which accrue to the owner in terms of savings in expenses of transportation
to a central location. Land users with large transportation costs value
accessibility more highly and will, consequently, pay more for property
than other potential owners. The land goes to the highest bidder.
The basic idea was presented in an early work by Hurd (1903) who
related accessibility to the amount of money individuals or firms are
willing to bid for property. Land value was said to depend on proximity,
which is equal to convenience.
Later, Haig (1926) added the concept of transportation costs,
stating that they are the reflection of the "friction of distance. The
value of a given piece of land depends on the savings in transportation
expenses which the location provides, as well as on characteristics
of the site itself. Haig was the first to consider residential land. Space
for a home was said to be evaluated like any other personal purchase
with the costs and benefits weighing in the consideration. In 1949
Ratcliff (1949) added the idea that maximum aggregate utility is reached
when total transportation costs in a city are minimized. This point
was later attacked (Alonso, 1964) and defended (Goldberg, 1970).
Despite some criticism (Wendt, 1957), the general model was
continued and expanded in the early 1960's. Kain (1962) developed a
utility function which includes as a negative component the cost of
transportation in the journey to work. Wingo (1961b) treated trans-
portation in depth and added the opportunity costs of, or the loss of
income due to, time spent commuting.
During the past decade, two landmarks have emerged in the
development of the normative model. William Alonso's Location and
Land Use, published in 1964, served two important purposes. It
incorporated the variability of the size of the residential site, and the
book formally developed the normative ideas in precise mathematical
Finally, Richard Muth (1969) broadened the model and included
the quantity of housing as a variable. He stated that the site which
is bid on possesses two characteristics: land and building; the
demand for each is variable with income. Muth also included more
than one place of work but did not completely develop the potential
Use of the model has continued with work by geographers and
economists. The individual demand for land and transportation has
been variously related to population densities (Cassetti, 1969; Cassetti
and Papageorgiou, 1971), the aggregate cost of distance (Goldberg,
1970), the third spatial dimension of high rise apartments (Wright,
1971), and open versus closed urban systems (Wheaton, 1974). Goldstein
and Moses (1973) have summarized the more recent developments in
The normative residential land use model is based on rational
behavior by economically oriented individuals. All extraneous variables
are controlled with general assumptions while the few items of interest
are manipulated mathematically. A simple version of the model is
presented and discussed in Appendix A.
Income and the Slope of the Bid-Price Function
The normative land-use model shows that transportation costs
cause an individual's bid-price, or the amount he is willing to pay per
unit of residential land to achieve a given level of satisfaction, to
decline as distance from the place of work increases. The bid-price
function over distance, therefore, is negative. Moreover, when trans-
portation costs are monetary and directly variable with distance, the
slope, or rate of change, of the bid-price curve varies with income.
The bid-price functions of two individuals are presented in
Figure 1. Both workers commute to a central location (t0) and they
are identical in every way except for income. When distance is plotted
horizontally, the high income or "rich" person has a gentle sloped bid-
price function, and the low income or "poor" worker has a steeper one.
to tl t2 di
Figure 1. Bid-Price Curves for Rich and Poor Individuals;
commuting costs are monetary.
Because property goes to the highest bidder, land in the t0-ti interval
is purchased by the poor commuter, and the distance between tl and t2
goes to the rich bidder.
Many different income (y) levels could be added to the model.
Each person in the city would be identical to the others except for the
amount of money each earned. The effect is shown in Figure 2 where
the more gentle is the slope of the bid-price curve. Each higher level
of income acquires land farther away from the place of work. The
result of the addition of many people earning different amounts of
money is to "round" or "smooth" the linear individual functions into
an exponential aggregate price of land curve.
Generalizing for many bidders, therefore, changes in the slope
component of the bid-price function with changes in income lead to
a direct relationship between individuals' income and the distance they
live from the central work location. This occurs under the condition
that the people are identical in every way except in the amount of money
The presence of low income groups on high-priced land at the
center of large American cities is thus explained by the bid-price
model. The cost of transportation (k) is not variable with income, and
the expense constitutes an ever greater proportion of the budget as income
gets smaller. The quantity of land can be modified, though. Poor people
can live on ever smaller pieces of land (or more crowded in apartment
Figure 2. Bid-Price Curves of People at 5 Different Income Levels.
buildings) as their earnings decrease. The value of accessibility
becomes a more important commodity in comparison to the quantity of
Multiple Work Locations
The normative urban land use model, as developed by Alonso,
Wingo, and others, is generally tied to one central place of work. Rationale
for using accessibility to the place of work and not to just any trip
destination is that income itself is made dependent on paying the commuting
costs. The worker does not have the option to go or not to go. For
economic man total utility = 0 if y = 0.
The model need not be restricted to one place of work (Muth,
1969, p. 87). At any residential site, the individual who submits the
highest bid, regardless of his place of employment acquires the use
of the land. The incorporation of many job locations can be shown more
clearly with schematic diagrams than with mathematical equations.
Assume a linear city with several employment sites. Workers
in the city are homogeneous in every way except the location of their
work place. Their bid-price functions are likewise identical except that
k is measured from each individual's place of employment rather than
from one central location. The bid-prices "peak, or reach their
highest value (where k = 0), at the work site of the person who is bidding.
A city with three places of work (Fl, E2 and F3) and the bid-price curves
of their workers are shown in Figure 3. Employees of F1 acquire the
tl F1 t2 F2 t3 F3 t4 distance
Figure 3. Bid-Price Curves of Workers at 3 Different Employment Sites.
the land between t1 and t2, those of E2 live between t2 and t3, and
the workers at E3 occupy the t3-t4 interval.
Assume that another group of workers is added. These are
identical to all others in the city except that they work at F1 and receive
a higher income. Let the whole population, therefore, be divided into
poor (all previous workers) and rich (new, higher income group). The
bid-price function of the rich is not as steep as that of the poor (see
Figure 1). The effect of the addition of this new rich group is shown
in Figure 4.
Land in this linear city is acquired in the following manner:
Interval Purchasing Group
t -t1 rich--E1
t1 -t2 poor--E1
t2 -t3 rich--EF
Land is purchased by the rich--E1 group of workers only in those
intervals where their more gently sloped bid-price curve is higher
than the one of the other groups. Land close to intermediate job sites
is not rented because of high competition from the employees of those
places of work.
E1 E2 F3
Figure 4. Bid-Price Curves of Rich Workers at E1 and Poor Workers at
E1, E2 and EF3
Many different income levels could be added for each employment
site. The effect would be to "round" the curves as in Figure 2, but
with the intermediate peaks of Figure 4. Generalizing further, many
places of work with differing numbers of employees who earn different
incomes could be considered in a city of two spatial dimensions.
Graphically, the result of these modifications would be an
aggregate land value cone which is centered on the general area
where the most people are employed. At different locations the
generalized cone surface would be altered with subsidiary peaks
rising around intermediate places of work. The model concerns
only residential land use, but the shape of this conic surface closely
resembles the often reproduced (Chorley and Haggett, 1967, p. 337;
Yeates and Garner, 1971, p. 252) allegorical "circus tent" diagram
which is used by Berry (1963) to describe urban land values. Empirical
land value studies (Knos, 1962; Yeates, 1965; Mills, 1969) confirm the
For the argument to be true, an important requirement is that
there be competition among the different groups of land users. Various
income levels must be represented at each of the employment sites.
Otherwise, there would be no distance differentiation; the rich would
live near their place of work on one side of town, and the poor would
live close to their job site in another part of the city.
In the hypothetical, two dimensional city, on which the model is
based, residential patterns of different income groups could be very
complex. There would be accessibility oriented lower income people
clustered around the various employment sites with the intervening
space going to higher income bidders of more distant places of work.
Differentiating residential land use according to income could be difficult,
if not impossible, and a general pattern might not be evident on a residential
land use map.
If competition is present, the basic relationships between income,
the value of land, and accessibility would still hold true, however. The
bid-price functions continue to be shaped by the same variables of the
linear city. Distance to work is still the item which determines how much
money an individual is willing to pay for land to achieve a certain level
As discussed, the bid-price model becomes one of interaction rather
than land use. Different income groups may be mixed or segregated,
concentrated or dispersed because jobs are unevenly distributed across
the city. Where people live can be normatively explained, however, when
the distance to work variable is included.
The statement of the model is unambiguous. Despite the residential
land use picture, the positive relationship between distance to work
and the level of income holds true. The rich live farther from their
jobs than the poor do.
As stated previously, the test of a normative model lies in the
validity of the assumptions which are made. Obviously cities are not
homogeneous plains, human behavior is not rational, not all land is
exchangeable, information is imperfect, nothing remains fixed through
time, and commuting costs are not measured only in money. Further-
more, there is no guarantee that residential location is the dependent
A wide gap exists, therefore, between the hypothetical world of
the model and the real world of "everyday. While the deductive logic
of the normative framework may be perfect, many of the factors which
the model leaves out bear strongly, if not decisively, on the final
location decision. It would be hopeless to expect real city patterns to
conform completely to those which are predicted by the normative model.
On the other hand, the normative approach is a valid one (Chisholm,
1971). The assumptions which are made need not be flawlessly true. If
there are elements of truth in them, then there is some validity to the
conclusions of the model. If statements such as "man is basically
rational, or "information is by and large available, can be made, the
normative model does provide some understanding.
The important question concerns those variables which are made
constant by the assumptions. For example, is distance to work and the
cost of commuting really the main consideration of the decision maker,
or are items such as the availability of housing, social segregation, and
historical preference more important.
The validity of the bid-price model must be determined through
empirical investigation. When commuting costs are money expenditures,
the model predicts a positive relationship between income and the
distance to work, and this relationship is generally confirmed in Anglo-
America. The early concentric zone model described an increase in
the social status of neighborhoods as distance from the city center grew
larger. Duncan (1956) confirmed a direct relationship between distance
to work and occupational status in Chicago, and other investigators have
obtained similar results in different American cities (Kain, 1962; Lowry,
1963; Goldstein and Mayer, 1964; Loewenstein, 1965; Wheeler, 1969b;
Catanese, 1970; Bederman and Adams, 1974).
Variations, however, are present in the general pattern.
Wheeler (1968a) discovered that the relationship holds true for
white workers in Pittsburgh, but not for blacks. Furthermore, black
people tend to work in certain areas of the city, regardless of distance
from home or access to public transit (Whiting, 1952; Wheeler, 1966;
Soot (1974) found that a negative relationship existed between
distance to work and income in Milwaukee for the 1949 and 1959
periods but that this relationship is disappearing at the present. Also
examining general trends, Wheeler (1974, p. 46) stated that for the
richest and poorest people the distance to work should be greater than
it is for middle income groups. In Charleston, West Virginia,
Halvorson (1973) discovered exactly the opposite. The middle income
groups live the fartherest from the place of work, and the highest and
lowest income groups live the closest.
Differences among types of commuting are also evident. Although
the positive relationship between income and distance is present for
those who travel to work in the central business district, no such
pattern is discernable for those who go to suburban jobs (Taaffe, Garner, and
Yeates, 1963; Logan, 1968; Wheeler, 1970a). This tends to discredit the
normative model and to suggest that access to the city center is more
important than proximity to work.
The verdict is not yet in on the normative residential location
model. The empirical findings, to some extent, support the predicted
relationship between income and distance to work, but many questions
remain. The wide difference between reality and the idealized world
of the model have led many to reject it (Wolforth, 1965; Boyce, 1969;
To explain the distribution of rich and poor, those who reject the
normative model often rely on a general theory about the nature of housing.
As people earn more money, they seek better quality residences,
and since these are built by developers on the periphery, the higher income
group moves farther away from downtown. In the center of the city the
old homes are left behind to "filter" down to lower and lower income
groups (Vernon, 1964; Johnston, 1971, Edel, 1972). Racial and status
discrimination encourage the process. Poor, predominantly black
people, therefore, are said to be trapped downtown because of racial
segregation and characteristics of the housing market.
The filtering theory also presents some problems, however. It
depends on urban expansion, and, as Alonso (1964, p. 109) pointed out,
cities which are not growing, present the same positive relationship
between income and distance to work. Furthermore, the theory assumes
that housing is not controlled by demand but by developers. Are people
really forced to build new homes on virgin peripheral land instead of using
present or buying new property in their current neighborhood? Growing
cities in non-industrial countries do not present the direct relationship
between income and distance, even though houses are continuously being
In conclusion, neither the normative model nor the filtering theory
are completely adequate in explaining the positive relationship between
income and distance to work. A cross-cultural comparison of this
relationship might shed light on what causes it.
INCOME AND DISTANCE TO WORK: LATIN AMERICA
Non-Western Residential Pattern
Little research has been done concerning the interaction
between residence and place of work in non-industrial countries.
The distance relationship between home and job in these areas,
therefore, must be largely inferred from descriptive studies of
residential land use. The majority of these studies indicate that the
distribution of families by income is different from that found in
industrialized societies. The people of greater wealth tend to choose
homes at the city center, and poorer groups live on the urban periphery.
The income distribution pattern, commonly termed "non-western, "
is most prevalent in Latin America. Seven empirical land use studies
stand out (Schnore, 1967; Amato, 1968) as landmarks in the description
of urban structure. These investigations show that in Merida (Hansen,
1934), Oaxaca (Hayner, 1944), Mexico City (Hayner, 1945), and
Guadalajara (Dotson and Dotson, 1954), Mexico; La Paz (Leonard, 1948),
and Sucre (Hawthorn and Hawthorn, 1948), Bolivia; and Guatemala City
(Caplow, 1949) the residential areas of the upper classes have traditionally
been nearer the city center than those of the middle and lower status
groups. The findings are confirmed in various other cities of Spanish
America (DeLaubenfels, 1957; Whiteford, 1960; Chaves, 1965;
Amato, 1970) and Brazil (Araujo, 1951; Bernardes, 1954; Prandini,
1954; Pyle, 1970, Nakagawara, 1973).
The location of residential areas within Latin American cities is
indicative of the relationship between workers' home and job locations.
Land use patterns indicate the distance relationships between residences
and work sites because most employment tends to be concentrated at the
city center. Professionals and bureaucrats as well as vendors, clerks,
and laborers are most often employed in the downtown area. In the
larger cities, factories are, or have been until recently, centrally
located (Hoyt, 1963; Harris, 1971); and industrial workers often, if
not always, commute to the city center.
Possible Causes of the Non-Western Pattern
Why is the spatial distribution of rich and poor classes opposite
that found in the developed world? Several arguments are advanced.
One is that the cultural history of Latin America is basically different
from that of other areas. The rich are said to prefer to live downtown
because the center of the city has traditionally been the most attractive
location. The importance of the central square as a focus for pro-
fessional and social activity, the characteristic house types, and the
checkerboard street patterns are all deeply rooted in the Spanish and
Portuguese heritage. Gabenheimer (n. d.) and Reis (1968) have presented
interesting accounts of the importance of culture in shaping the morphology
of early Latin American cities.
Although culture is undeniably important, the argument that
it is the primary cause of the non-western residential pattern can be
questioned. A variety of non-Iberian societies, including pre-
industrial Anglo-America, exhibit, or have exhibited, the same
distribution of rich and poor. Furthermore, recent evidence (see
below) indicates that the pattern in Latin America may be changing.
This means either that the cultural attachment of the rich to the center
is weakening or that other events are shaping urban morphology.
Rural-urban migration is also often considered to be a factor
in the distribution of the rich and poor within cities. The migrants
are largely low-income and unskilled people who settle at the urban
fringe in squatter communities. A popular myth is that they are
generally unemployed and live, therefore, marginal to, or outside of
the urban economic system. Although generalization is impossible,
90 percent of the residents of the Lima barriadas do have jobs (Andrews
and Phillips, 1971). In a survey of the rural-urban migrants of Rio de
Janeiro, Sao Paulo, and five other major Brazilian cities, Hutchinson
(1963) found that 85 percent of the household heads secured a job during
the first month after arrival.
Another misconception is that the inhabitants of the poor sections
always come from agricultural backgrounds. Portes (1971) summarized
recent studies dealing with slums in seven Spanish American cities
and found that the majority of the dwellers were born at their present
residence or came to the city at an early age. Of the remainder, most
had had urban experience elsewhere. Morse (1971) stated that two-
thirds of those who are rural migrants are urbanized in their education
These findings show that the peripheral poor of Latin America
are an integral part of the urban population. They have lived in
cities most of their lives and are interacting with the economic
system. Although the pay is meager, they go to work daily as does the
rest of the population. Slum sections, therefore, are integrated
components of the urban system. Recent studies confirm that the low
income sections are not squalid, parasitic growths as once was thought
but are communities with high levels of social organization (Eyre,
1972; Souza and Porter, 1974).
Changes in the Non-Western Pattern
Various studies show that the spatial distribution of income
groups in Latin America is changing. Cities may be evolving toward
a "western" pattern where the rich live farther from the center than
do the poor. The process is altering the traditional spatial arrange-
ment of Merida (Hansen, 1934), Mexico City (Hayner, 1945),
Guadalajara (Dotson and Dotson, 1954), Bogota (Amato, 1968; McCallum,
1974), Rio de Janeiro (Bernardes, 1967), Caracas, and Lima (Harris,
1971, p. 249) where the elite groups are leaving their central location.
Other cities, however, are not following the trend (Hayner, 1944;
Leonard, 1948; Hawthorn and Hawthorn, 1948; Araujo, 1951; Bernardes,
1954; DeLaubenfels, 1957). Throughout the continent the non-western
residential pattern characterizes small and medium size towns. In
these communities, the wealthy families remain firmly attached to
the downtown area.
Some researchers have incorporated the current changes into a
general evolutionary explanation of the non-western residential
location pattern. Schnore (1967) and Hoyt (1963) have believed that
all cities may pass through different evolutionary stages as economic
development occurs. The recent abandonment of the city center by
the rich could evidence the beginning of the process of filtering (Johnston,
1972) which is said to be characteristic of industrial society.
Others (McGee, 1971; Herrera, 1971; Alonso, 1973), however,
have maintained that the evolution of cities in industrial areas cannot
be used to explain present conditions in the developing countries.
The urbanization process, population of the cities, and socio-economic
infrastructure of these areas are basically different from that of the
United States and Europe during the last century.
Transportation as an Explanation of the Non-Western Pattern
Transportation has also been cited in order to explain residential
distributions found in Latin American cities. Schnore (1967) and
Berry, Simmons, and Tennant (1963) referred to transportation
technology as being the important factor which differentiates the western
and non-western patterns. Amato (1968) stated that improved transpor-
tation made it possible for the rich to move away from the center of
Bogota, and in a study of Buenos Aires, Sargent (1972) developed a
dynamic model which has speculation and settlement occurring within
a transportation framework.
Transportation modifies residential location by affecting
intra-urban spatial interaction. A change in transportation technology
does not change residential location directly, but alters the mobility
of people. Transportation, therefore, has a direct impact on the
journey to work and a secondary influence on residential location.
If transportation is the important modifying variable, the non-
western pattern is really a result of interaction between residences
and work sites. What appears to be an urban spatial income distribution
is in fact a relationship between income and distance to work.
Wheeler and Thomas (1973), in a pioneer study, investigated
journey to work interaction in Tegucigalpa. The results showed that
the population of the city is highly immobile, with most of the people
working in the same section of town in which they live. Those who
use motorized transportation generally travel farther than the workers
who walk. Factor analyses of an origin-destination matrix, after a
technique used previously by Wheeler (1970b), however, showed
little spatial patterning.
The changes which are occurring in the residential pattern
are often explained by stating that the increased availability of
motorized vehicles makes it possible for the wealthy to leave the
center of town. They can afford cars or train rides and are thus
able to live where they want. The lower income groups, meanwhile,
are restricted to an area which is within walking distance from the
city center. The rich have a greater mobility because they use a
different type of transportation.
Although such a framework explains current trends, questions
concerning the causes of the western and non-western residential
patterns remain unanswered. If Latin America is evolving into a
motorized society where few people walk, the mobility of the rich
and poor will again be the same. What happens then to the spatial
distribution of the different income groups? Clearly, transportation
used as an explanatory variable in this form cannot show how the non-
western residential distribution is evolving into the pattern of
the motorized societies of today. A comparison of the distance to
work of the immobile poor and the mobile rich explains neither
the western nor the non-western patterns. In primitive societies
everyone walks while in developed cultures the vast majority uses
In present-day Latin America primitive means of travel are
rapidly being substituted by buses, automobiles, or motorbikes, and
the current transitional period is unique in that two distinct systems
are present side by side. Are residential distributions evolving from
a non-western to a western pattern? If so, is changing transportation
technology truly an important variable influencing the process? To
understand the effect of changes in the transportation system, care must
be taken not to compare the distance to work of two groups with different
mobility levels. Regardless of income, someone who walks should
always live closer to his place of work than someone who uses an
automobile. Income and transportation, therefore, must be considered
separately if the distance to work is to be explained.
Time Costs and the Bid-Price Model
The normative bid-price model, as discussed in Chapter II, can be
used to explain the relationship between income and distance to work
in industrialized areas. It is also applicable to residential patterns
in Latin America. As presented by Alonso, the model is, in many ways,
unrealistic, and the assumption made concerning the nature of transportation
costs is one of the least valid. The cost of commuting, even in a motorized
society, involves more than monetary expenses. The price of gasoline,
for example, might be of minor concern when compared to the nervous
stress produced by travelling on congested urban streets.
An important non-monetary cost is that of time. The hours
spent commuting to work are hours which are not available for other
uses. Eating, sleeping, playing, and money making are limited by
the amount of time used for the journey to work.
For many years social scientists have recognized time as a
component of the aggregate cost of commuting (Haig, 1926, Liepman,
1944; Lillibridge, 1952). The change in transportation which made
the separation between home and place of work possible was a reduction,
not in money, but in the amount of time which was necessary to travel
a given distance (Vance, 1966).
Developers of the bid-price model have also been aware of time
costs, and leisure has generally been included in the personal utility
function. The estimation of the value of leisure time is not new in
economics, but it was first applied to commuting by Wingo (1961b).
People were said to attach a certain utility to money and to
time spent off the job. The point at which time and money are optimally
combined can be obtained when the time utility function and wage rate
are known. Wingo (p. 55-62) showed that when commuting is introduced,
employers are forced to pay a higher wage for a given period of work
because the amount of leisure available to the worker is reduced.
This extra wage is the monetary price of commuting.
As developed by Wingo, the cost of time can be included in
generalized mathematical formulations of the normative model.
Operationalizing and testing such a model, however, is difficult
because money and time utility values are unknown. Utility is a
conceptual tool and not an observable and measurable quantity. A
model that shows a change in the slope of the bid-price curve with a
change in income (as developed in Chapter II) and that also includes time
as a commuting distance cost requires that the utility relationship between
time and money be estimated. When this is done, the rate of change of
the bid-price curve is unpredictable. As shown in Appendix B, the
component which involves the monetary cost of distance tends to lessen
the slope with an increase in income, while the component related to
the time cost of distance increases the slope.
Some of the problems encountered in the attempts to test the
normative model in industrial countries may be the result of the
counteracting effects of time and money costs on the slope of the bid-
price function. Halvorson (1973) for example, obtained inconclusive
results in his investigation of the relationship between income and
distance to work in Charleston, West Virginia (see above p. 31). When
time is used as a surrogate for distance, however, he found that the
higher income groups opt for shorter work trips and outbid the poorer
people for homesites which are accessible to the place of work.
In a motorized society, time and money costs are interwoven
in a complex manner. The automobile commuter, for example, uses
amounts of both money and time for the trip to work, and items such as
traffic congestion and parking influence the ratio of the two costs.
Isolating the time factor in such a society is difficult.
When monetary expenses are excluded, the statement of the model
is clear. As shown in Appendix B, the value of time spent commuting
increases directly with income, and the bid-price function of a rich
individual, therefore, is steeper than that of a poor person (see Figure 5).
The generalizations of many different income levels (see Figure 2),
multiple work locations (see Figure 4), and two spatial dimensions (See
discussion on p. 28 ) can be made with the only difference being the
reversal of the bid-price curves of the rich and the poor. In all
instances where the groups compete for land there is a negative
relationship between income and distance to work; the rich live
Figure 5. Bid-Price Curves for Rich and Poor Individuals; commuting
costs are temporal.
nearer their jobs than do the poor.
The normative bid-price model, therefore, predicts two different
relationships between income and distance to work. Where motorized
transportation is used, a positive relationship between income and
distance to work prevails, but where non-motorized transportation
is used, the function becomes negative.
These conclusions were also graphically derived by Lave (1970)
and Harvey (1972). The bid-price curve of the high income walking
group was said to be steeper because time is "more valuable" to the
rich. The authors, however, did not formally deal with the relationship
between opportunity costs and money income.
A society in which time commuting costs are predominant is
one where transportation is primitive; one where bicycle riding or
horse cart use, for example, are important means of travel. Walking
is the ultimate commuting time cost. Money outlay, except for shoe
expense, is non-existent, but time expenditure is greater than for
any other transportation mode.
At the present, motorized and non-motorized societies coexist
in Latin America. The normative bid-price model, if confirmed, can
not only explain the non-western residential pattern of the pre-
automobile era, but it can also predict what the spatial structure will
become as fewer and fewer people walk to work.
The bid-price model predicts two different spatial relationships
between income and distance to work. The variable which differentiates
the relationships is the means of transportation which is used for
commuting. The purpose of this research is that of searching for these
predicted relationships in a Latin American city. The hypotheses to
be tested are:
1. For workers who use a means of transportation for which the
cost of overcoming distance involves primarily money expenditures, a
positive relationship exists between income and the distance between
residence and place of work.
2. For workers who use a means of transportation for which the
cost of overcoming distance involves primarily time expenditures, a
negative relationship exists between income and the distance between
residence and place of work.
DATA COLLECTION IN THE STUDY AREA
The research hypotheses presented in the previous chapter describe
certain relationships which exist between variables of the bid-price
model. To test the hypotheses, these variables must be translated
into items of information which can be collected and measured in a real
environment. For this study, Piracicaba in the state of Sao Paulo, Brazil
was selected as the study area. One hundred miles northwest of Sao
Paulo city, Piracicaba is a regional center with a population of 123, 806
History and General Characteristics
Although officially founded in 1776, the settlement dates back
to the 17th century. During this period, the interior of the state of
Sa'o Paulo was penetrated by groups of people (bandeiras) who searched
for gold and Indians to be enslaved (Morse, 1965). Common practice
was for the group to settle down periodically and grow food crops to
replenish supplies. The favorable soil, navigational barrier, and fish
supply made the site near the rapids on the Piracicaba river a favorite
for these temporary stops. In time, the settlement became permanent.
Little changed, however, until the early 18th century when
work on a projected road link between Sao Paulo and the gold mines
of Cuiaba brought a boom to the area. This growth culminated in the
incorporation of the city in 1776.
The 19th and 20th centuries have been marked by continued
economic progress. Coffee was important until the 1920's, but
recent development has been a result of the shift of sugar cane pro-
duction from northeast Brazil to the interior of the state of Sao Paulo.
This move was caused by the decline of Brazilian export markets, the
increase in consumption in the populous southeast, and the modern
cultivation and refinery techniques which are used in Sao Paulo.
In the southeast, the physical setting has been most favorable
around Piracicaba. Paleozoic sedimentary bedrock of the area has
led to the formation of a clay-rich podzol soil series (Queiroz, 1968,
p. 500) which, coupled with a favorable subtropical climate, makes
conditions ideal for sugar cane cultivation. Piracicaba, therefore,
became the focus of the crop's production in southeastern Brazil. Today,
practically all agricultural activity near the city involves sugar cane.
For the 1972-73 growing season, 79 percent of the county 's cultivated
acreage was devoted to the crop.
Piracicaba has developed into a commercial distribution and
supply center for all goods related to the cultivation, refinery, and
distillation of cane. The industrial sector of the city's economy has
greatly expanded since the first plant was built during the 1800's.
Substantial activity within the city is related to sugar refining and
distillation, but most production involves the fabrication of heavy
machinery components for sugar mills. The "Grupo Dedini" consortium
supplies 80 percent of all refinery and distillery equipment used in
Latin America and accounts for over half of Piracicaba's total productive
output. Many other goods are also produced with furniture and textiles
representing important segments of the industrial economy.
The growth of the urban population has been most dramatic in
recent years (See Table 1). The racial and ethnic composition is
not significantly different from that of other Sao Paulo cities. Portuguese,
Spanish, and Italian backgrounds are well represented, but Piracicaba
does not contain a large Japanese colony as do some of the other interior
Urban Population of the City of Piracicaba
The sugar boom, industrial development, and population increase
are related to the overall prosperity which the state of Sao Paulo
has experienced in recent years. Piracicaba is forecast to continue
to grow if the international price and demand for sugar products remain
high. At the present, an incentive program aimed at reducing the
concentration of factories around greater Sao Paulo is leading to a
diversification of production in Piracicaba. A large industrial park
is being developed in which the Caterpillar Tractor Company, for example,
is building a plant second only to its home installation in Peoria, Illinois.
Piracicaba as the Study Area
A number of factors contribute to make Piracicaba a good city
in which to test the research hypotheses of this study. In the first place
the size of the town is such that almost all of the urban area is accessible
to people who walk. On the other hand, distances are great enough,
for those who walk, for there to be a true "friction" due to the distance.
Secondly, motorized and non-motorized transport systems do
exist side by side. The number of inhabitants per motor vehicle (8. 7)
is less in Piracicaba than for any other city in Brazil, but most of the
people still walk to work.
Thirdly, secondary and tertiary activities are equally well
represented. Of the males over ten years old, 15,584 are industrial
workers, and 15, 870 have commercial, administrative, or other service
Finally, the city approaches, in many ways, the closed system
assumed by the bid-price model. Piracicaba is a regional center
(Muller, 1966) and although goods flow in and out of the city, most of
the people live and work within the urban area. Few people either
Figure 6. City Center. View of the central area with the
traditional square and Catholic church.
Figure 7. Southwest Panorama. Residences and privately owned
vacant property are in the fore and middle ground. Cuestas
in the distance are evidence of differential erosion between
Jurassic basalt and more ancient shale and sandstone.
Figure 8. Water Falls. Partial view of Piracicaba river rapids
and the public sightseeing park.
Figure 9. Pumping Station. Old building houses water intake
controls of the Boyes textile mill.
commute from residences or travel to jobs which are outside the
Variables to be Measured
Before data could be collected in Piracicaba, a decision was
necessary as to what measurable information most closely corresponded
to the variables of the bid-price model. For this study, distance was
assumed to be airplane, or straight line distance. Because of the natural
barrier of the Piracicaba river (See Figure 10), distance for those who
commuted across the river was measured from the work place to the
most convenient bridge, and from there to the residence.
Walking and bicycle riding were assumed to be the obvious
transportation modes for which overcoming distance involves time
costs and automobile use the mode which involves money expenses.
The price of a bus ticket is not variable with distance in Piracicaba.
Although a ticket costs money, the expense which is directly variable
with distance for bus users is time. Bus travel, therefore, was also
assumed to be a mode for which overcoming distance involves time
costs. Other means of transportation are not important in Piracicaba.
Income was assumed to be the per-month earnings derived by the
household head from his principal job. While other members of the
household might hold jobs, the locational decision, and hence the distance
to work, was thought to be primarily based on the family head's income
Figure 10. General Map of Piracicaba.
from his principal job. Rationale was as follows: Each household
probably has a variety of income sources, including, for example,
money from the father's work, the mother's washing and ironing,
the son's shoeshine job, pensions, retirements, and rent paid to the
family. Any number of non-monetary items such as vegetables from a
garden or gifts from a friend might also be received.
Determining all sources of income for each household would be
impossible, and the principal job income was adopted for this reason
as a standardized surrogate for total income. Furthermore, the
valuation of time and the opportunity cost of commuting are developed
(See Appendix B) by considering income to be the amount of money earned
on the job (during the time period J). Other surrogates for total
income might be various items related to wealth or affluence, and so
data were collected on certain of these variables which were available.
The major sources of data were personal interviews conducted
in Piracicaba during the June, 1973-May, 1974 period. A questionnaire
was prepared in which the household head, or the family member earning
the most money, was asked his place of work, the type of transportation
which he used to go to work, and his principal job income. Information
concerning the length (number of years) of residence and work at the
present location, as well as the length (number of years) of transportation
use was also requested.
Of the total households in Piracicaba, interviews were conducted
in a sample of 4 percent (1 of 25). An adequate frame from which to
draw the sample was unavailable, and steps had to be taken, therefore,
to ensure an adequate representation of the city's entire population.
An areal sampling technique was used, and, for this, a household
density map of the city had to be constructed. Household was defined
as being the family group permanently inhabiting one housing unit,
and census criteria (direct access to outside, independent kitchen
facilities, and separate living quarters in general) were used to
define housing units. Hotels, boarding houses, and other such
establishments were not considered to be permanent places of residence.
In order to draw an areal sample based on a density map, an
assumption had necessarily to be made. The population of interest in
Piracicaba was that of workers with a fixed residence and job location.
This group, therefore, was assumed to be randomly distributed in
the general population with no particular spatial pattern distinguishing
the residences of the commuting workers from those of the rest of
the city inhabitants. Evidence subsequently gathered during the inter-
views indicated that the assumption was valid.
Two sources were used in the construction of the density map.
The first was a series of air photographs (taken in 1969 at a scale of
1: 12,000) which were available at the local university (University of
So Paulo agricultural extension "Luiz de Queiroz"). Based on informa-
tion derived from these photographs, the city was spatially divided into
different sectors with the sole criteria for differentiating the sectors
being the number of residential units per square area. For each
sector a housing density measure was recorded.
The second source used in the construction of the density map
was information gathered from ground observation. Each of the
sectors was visited, and densities of randomly chosen blocks were
checked against the density figure derived from the photographs.
Particular attention was given to areas where closely grouped buildings
made difficult the distinction of separate residential units from the
photographs. Care was also taken in the commercial sectors where
stores and residences were mixed, and in parts of town which had
developed rapidly between 1969 and 1973.
The number of sectors and the density values of each were
modified when the airphoto information did not agree with observed
conditions. The final map contained sixty-three sectors with densities
ranging from 1 to 65 housing units per block (1 hectare or 10,000 square
meters). An added 64th sector included Piracicaba's seventeen multi-
storied apartment buildings.
The final map provided a density figure for each sector. The
total area of each sector was also determined, and the combination
of both measurements yielded the number of housing units per sector.
The total number of residences in Piracicaba estimated from the density
map (25,252) was slightly greater than that arrived at during the 1970
census (22,125). The difference between these figures can be
attributed to growth from 1970 to 1973, difference in the criteria
used, and human error. Where the borders of the sectors of the
density map approximated those of historical sections of town for
which census information was available, the census and density map
values were also not substantially different.
The collection of data was sectorially stratified, and a randomized
cluster sampling technique was used within each sector. The
procedure was to select a location in a sector by using a grid, map, and
table of random numbers. The residences on the street block nearest
to this randomly chosen point constituted the sampling unit, and
a maximum of eight interviews were then conducted at the location.
If there were less than eight houses with people at home on the block,
the number of interviews equalled this lesser number.
The procedure was then repeated with the selection of another
random point and in turn repeated until the full 4 percent of all
residences in the sector had been visited. All sectors were treated
identically until the required 1,010 interviews were made in the city.
Blank questionnaires soliciting the desired information were
distributed during the morning or early afternoon hours, and the
completed forms were gathered the following day. Accompanying each
questionnaire was a standard instructional paragraph and a copy
of a letter of introduction from the city mayor.
Every effort was made to leave a copy at each house in the
sample area (until the maximum of eight had been distributed). Skeptics
were assured that the interviews were of scientific value and that
complete confidentiality would be maintained. If no one was present,
the house was necessarily skipped and not included in the sample.
If for some reason the questionnaire had not been completed by the
following day, additional calls were made on subsequent days.
Some problems were posed by the fact that the head of the house-
hold was usually at work during the time of the interview. Special
care had to be taken to ensure that the instructions were well grasped
by the person who was spoken to (usually the wife) and that she could
communicate them to the head of the family.
Another problem was that an unavoidable bias against unmarried
workers was built into the study. Their homes were generally empty
at the time of the interviewing. In special instances when the members
of the visited family could neither read nor write, the instructions and
questions were read orally by the interviewer who also recorded the
Property Survey Data
Data were also gathered from confidential files at city government
offices. Most of the results of a recent city-wide property survey
were made available, and from this source the surrogates for economic
status were collected. The availability of this second information source
made it possible to compare land value, house quality and house size
where interviews had been made with the value, quality, and size of
the property and residences where the residents had not responded
(Property value was obtained from a confidential map which is revised
yearly by city officials; house size referred to general floor space;
house quality was a tabulated point score compiled from individual
ratings on fourteen different criteria. ).
Thus, in a general manner, differences in economic status
could be detected between those who had responded to the questionnaire
and those living on the same block who had rejected, not received,
or failed to return the form. In this manner, the built-in bias
against groups such as unmarried workers and unresponsive people
could be, albeit crudely, tested for.
Student's "t" tests were performed to check this bias. The
property value, house size, and house quality of the individuals
who responded to the questionnaire were compared to the same measures
pertaining to people who were not at home at the time of the interview.
Similar comparisons were made between respondents and persons who
received but failed to return the questionnaire as well as between
respondents and the individuals who had refused to accept the blank
For all tests, null hypotheses of no difference between respondent
and non-respondent statistical populations were accepted at the 05
level of significance. The economic status of the people who answered
the questionnaire was not appreciably different from that of their
neighbors who did not.
A comparison was also made between the principal job income
data obtained from the interviews and the surrogate measures of economic
status (property value, house size, and house quality) obtained from
the municipal government. Student's "t" tests indicated a greater
than zero value for coefficients in a correlation matrix (See Table 2)
of the four variables. Null hypotheses of no correlation between income
and each of the economic status variables were rejected at the 05 level
of significance. Principal job income and the surrogates, therefore, were
inferred to be different measures of the same general quality; total
wealth. Income information, as obtained from the interviews, is a valid
measure of total wealth.
Correlation Matrix of Income and
Surrogate Measures of Economic Status
Income House size House quality Land value
Income 1.00 0.69 0.73 0.40
House size 1.00 0.78 0.48
House quality 1.00 0.42
Land value 1.00
Definition of the Effective Statistical Sample
Of the 1,010 questionnaires distributed, 759 were completed.
Of these positive responses, some could not be included in the sample
because the situation of the family head did not meet certain necessary
conditions. The first requirement was that the household head have
a permanent place of work within the city to which he commuted on a
daily basis. As shown by the breakdown of the positive responses with
respect to the place of work in Table 3, 63. 1 percent met this condition.
Work Situation of Respondent Household Heads
Number of Total
Fixed place of work 479 63. 1
Work location not fixed 59 7. 8
Out of town work location 81 10. 7
Retired 80 10.5
Unemployed 31 4, 1
Work at home 27 3.6
Student 2 0.2
Total positive responses 759 100. 0
A second requirement concerned income. Some individuals
received highly variable incomes from month to month, and others
refused completely to answer questions concerning earnings. Responses
of this nature were not included in the sample. Of the 479 responses
with a fixed place of work, 464 contained valid income data.
Finally, information concerning the means of transportation
was necessary. All of the 464 responses met this requirement.
Table 4 shows the transportation used for the journey to work by the
respondents who had a fixed, non-home place of work and a steady
Transportation Used for the Journey to Work
by the Respondents Who Had a
Fixed Place of Work and a Steady Rate of Income
Number of Total
Walk 199 42.9
Private car 126 27.2
Bus 95 20.5
Bicycle 36 7.7
Other 8 1.7
Total 464 100.0
The responses which included one of the "other" (motorcycle,
horsecart, and company vehicle) transportation types were excluded
from consideration. For each of these transportation groups, the
sample size was too small for effective statistical analyses.
Of the 759 total positive responses, therefore, 456 were of such
a nature as to make them usable in the testing of the research hypotheses.
This group of responses became the effective sample, and all statistical
inferrences are based on the data contained in this sample.
The residential areas of Piracicaba are made up of many different
sectors which feature widely varying characteristics. Some sections
are inhabited exclusively by upper income groups, while within others
only the poorest people live.
The highest status area is the Jardim Europa section northeast
of the city center. In this exclusive neighborhood all of the measures
of socio-economic status are higher than for anywhere else in the city;
the mean income per month is over $1,000 whereas the city average is
$225. Houses in this area are architecturally elaborate, and they
occupy relatively large areas. Yard space is minimal, however, and
the empty land that is available is used for decorative rather than
functional purposes. Workers are most often merchants, doctors,
lawyers, corporate executives, or professors at the university, and
a commuter generally drives to work in one of the family cars.
The Sao Judas Tadeu section is also composed of high income
families, but the wealth of Jardim Europa is not matched. Although
also expensive, the residences are generally smaller and feature less
luxury than those of the richer neighborhood. Most families own cars,
Figure 11. Doctor's Home. Jardim Europa residence
featuring decorative yard and expensive Brazilian-
Figure 12. Business Executive's Home. Large and luxurious
Jardim Europa residence.
and household heads work in various professional positions.
The poorest sections of town are the peripheral areas of the
south. These form a semi-circle originating in the southeast along
the Sao Paulo highway and extending to the western areas where the
urban limit intersects the Piracicaba river. The residents of these
sections lead difficult lives. Although unemployment is not proportionally
greater than elsewhere, the majority of those who work are severely
underemployed. Many earn only the $48 minimum wage at a variety
of industrial jobs. Others receive even less as house painters, brick-
layer's assistants, maids, and street vendors. Amenities such as
running water, sewage, indoor plumbing, and paved streets are scarce.
Although most houses are made of brick, many families inhabit no
more than a 15 or 20 square meter wooden structure which offers
no protection against a climate where temperatures are comparable to
those of north Florida.
Although some migrants have come from surrounding rural areas,
many of the low income residents have spent their life in Piracicaba.
Walking and the city bus system are the most often used methods of
travel to work or to the city center.
Most of the residential sections of Piracicaba are not economically
segregated. The majority of the people in these mixed neighborhoods
earn moderate incomes (as compared with other groups and not by
United States standards), and the family heads work in a wide range
of blue and white collar jobs.
fh fii7kitiligL ntM41
Figure 13. Resident of Sao Judas Tadeu Section. Husband of
the pictured woman is a traveling salesman who earns
approximately $500. 00 U.S. per month.
Figure 14. Sao Judas Tadeu Residential Block. Typical housing
of the sector. All residents of the pictured houses
S? 9 .
c ,-, . .
Figure 15. Backyard. Typical morning activity of the family and
neighbor of a low income blue-collar worker.
~ ~s~t~p~t- -:,
Figure 16. Peripheral Poor. Wood and brick building is
inhabited by three families.
Figure 17. Rented Rooms. Property owner in a southern peripheral
area rents five small units in his backyard at about
$10. 00 U.S. apiece per month. Tennant with the maximum
total family income earns $48. 00 U. S.
Figure 18. Communal Water. The public water source is a
social gathering spot for poor women, many of whom
earn money from washing clothes.
Although generalizations for the large middle income group are
difficult, certain characteristics are evident. Houses are modest but
do feature running water, bathrooms, kitchens, and separate bedroom
areas. Sometimes the family owns a car, but more often, commuting
is performed on foot or by bus.
Some basic attitude differences seem to exist between the
middle and lower income groups. The poor people appear to be resigned
to their fate. They complain about the lot which has been dealt them,
but they feel, and often are, powerless to improve their situation.
The middle class, however, has adopted a work ethic. Many
families set economic goals and strive to achieve them. A positive
outlook toward future earning potential is common, and many goods
are purchased on credit. Consequently, this group probably spends
more time at work than any other group.
Although the middle class inhabits a large proportion of the mixed
residential areas, the poor and rich make up a substantial share of the
neighborhoods. Residences comparable to those of Jardim Europa are
scattered throughout these sections, but they are most prevalent
near the central commercial district and along the major avenues. Many
low quality houses are also evident, but they are not very often located
near the city center.
The intermixture of the different status groups does not
necessarily indicate a social integration of the various classes.
On the blocks which register a wide economic variation, there appears
to be very little personal contact among the neighbors. Non-mixed
blocks show evidence of more social interaction.
Some anomalies exist in the general residential patterns of
Piracicaba. The Beira Rio section is the oldest and most colorful
part of town. The poor residents of the left bank of the river possess
a culture and tradition that differentiates them from the rest of the
urban community. Most activity in this neighborhood is directly
tied to the river. The waterfalls attract sightseers as well as fisher-
men, and some of the Beira Rio residents do a brisk business of selling
bait and renting boats.
Others make a living by fishing, and the fish festival is a
traditional event. The practice of wading into the dangerous turbulent
water at the falls and barehandedly catching small fish is in itself
a tourist attraction, but recent deaths and stricter enforcement of city
ordinances have led to a decline in the practice. Twenty years ago
fish of 200 pounds or more were often caught by the serious fishermen,
but the Jau, Dourado, and Pintado have abandoned the now polluted
river. Piracicaba, which in the Tup" language means "The Place
Where the Fish Come, no longer lives up to its name.
Another group of neighborhoods which stand out from the general
residential pattern are the subsidized low and middle income housing
developments. Since 1965, the federal government has made a conscious
effort to improve the quality of housing in Brazil. The National Housing
Bank (B. N.H. ) was created to provide low interest mortgage loans to
Figure 19. Poor Family in a Mixed Neighborhood. Most sections
of Piracicaba contain high as well as low income
families mixed in the predominantly middle income
areas. Small yard space forces children to play
on the sidewalk,
Figure 20. Homogeneous Architecture. Housing in a middle
income neighborhood. Some units are rented but
most are owned privately by the residents.
Figure 21. Heterogeneous Architecture. Homes near the city
center, remodeled at different times, reflect taste
changes in style and in the choice of material used for
Figure 22. Beira Rio. Section of the colorful residential area
which is the oldest part of Piracicaba.
Figure 23. Swimming Area. Woman washing clothes and fisher-
man in boat are Beira Rio residents.
people of low income, and a total of 772, 000 units had been financed
in the country by 1972 (Costa, 1972). Rios (1974) and Trindade (1974)
present opposite positions concerning the success of the program.
Jardim Esplanada, Jardim Primavera, and a general section
known as Dopla are the principal areas of subsidized housing in
Piracicaba. The residents are generally of a low economic status.
Most workers are blue collar employees, and those who live on the
north side of the river are almost invariably employed at the nearby
Dedini foundary or the metal plants of Vila Rezende.
Houses are of uniform size, quality, and general aspect. The
inhabitants are highly sociable and consider themselves fortunate to
have been chosen for the low interest loans necessary to finance the
purchase of their new homes.
Places of Work
Unlike many Latin American cities, the place of work in Piracicaba
does not invariably coincide with the city center. While a substantial
number of people are employed in downtown shops and offices, an
equally large group commutes to non-central locations.
The Vila Rezende section north of the Piracicaba river is
the primary industrial area of town. Machinery plants and distilleries
combine to form a center of blue collar employment, but the largest
plant is the Siderurgica Dedini located farther north on the urban
periphery. Heavy industry is not confined to this section, however.
The Morlet and Mausa machinery plants, Boyes textile mill, Engenho
Figure 24. Dopla. New construction in the Dopla subsidized
Figure 25. Jardim Primavera. Subsidized housing finished in
1971 for low income families.
Central sugar refinary, and Alvarco autoparts factory also draw large
numbers of commuters.
White collar workers are most often employed in the downtown
area. Secondary commercial centers, although small, also exist in
the Paulista, Vila Rezende, and Cidade Alta sections. White collar
workers do not necessarily earn higher incomes than blue collar
employees. Many menial bureaucratic tasks are performed for low
wages. All utility bills and bank statements, for example, are
delivered from door to door by messengers because the postal system
is not reliable.
The standard work day is eight hours with a two-hour lunch break.
Saturday mornings are also spent on the job. The relatively long lunch
period is used by many to return home for the noon meal which is the
most important for Brazilians. Automobile users invariably eat
lunch at home, while walkers and bus riders stay at work if the distance
is too great.
Some commuters, therefore, travel to work twice as often and
incur greater distance costs during a given time period than others
do. Information about whether or not the family head eats lunch at
work was also solicited in the interview. Personal conversation and
an examination of the data showed that "lunch at home" is more likely
to be a result and not a cause of the distance to work. The factor
could not, therefore, be used as an independent variable, and it was
excluded from the statistical analyses.
Figure 26. Siderurgica Dedini. Foundary where steel is made
from scrap metal. The plant employs more workers
than any other in Piracicaba.
Figure 27. Metalurgica Dedini. Interior view of a portion of
metal plant where machinery for sugar refineries