100% found this document useful (1 vote)
215 views545 pagesSyd Sae Ter Hammond Mathematics For Economic Analysis
Mathematics for Economic Analysis
Uploaded by
kalpesh_dCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
100% found this document useful (1 vote)
215 views545 pagesSyd Sae Ter Hammond Mathematics For Economic Analysis
Mathematics for Economic Analysis
Uploaded by
kalpesh_dCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 545
Mathematics
for Economic
Analysis
Knut Sydseter
Peter J. Hammond
Stanford University
PRENTICE HALL, Upper Saddle River, New Jersey 07458 Library of Congress Cataloging-in-Publication Data
Sydseter, Knut.
Mathematics for economic analysis / Knut Sydsater, Peter J
Hammond.
Poem.
Includes bibliographical references and index.
ISBN 0-13-583600-X
1. Economics, Mathematical. 2. Economics~Mathematical models.
1. Hammond, Peter J. U1. Title.
HBI35.S888 1995
330°.01'51-de20 94-4225
ce
Production Editor: Lisa Kinne
Acquisitions Editor: J. Stephen Dietrich
Copy Editor: Peter Zurita
Cover Designer: Maureen Eide
‘Manufacturing Buyers: Patrice Fraccio and Marie McNamara
Editorial Assistant: Elizabeth Becker
©1995 by Prentice-Hall. Inc.
A Pearson Education Company
Upper Saddle River, NJ 07458
All rights reserved. No part of this book may be
reproduced, in any form ot by any means,
without permission in writing from the publisher.
Printed in the United States of America
1029.8) 7-625) 4103)
ISBN O-13-S83b00-X
Prentice-Hall International (UK) Limited, London
Prentice-Hall of Australia Pty. Limited, Sydney
Prentice-Hall Canada Inc., Toronto
Prentice-Hall Hispanoamericana, S.A., Mexico
Prentice-Hall of India Private Limited, New Delhi
Prentice-Hall of Japan, Inc., Tokyo
Pearson Education Asia Pte. Ltd., Singapore
Editoria Prentice-Hall do Brasil, Ltda., Rio De Janeiro —]
Introduction
The economic world is a misty region.
The first explorers used unaided vision.
Mathematics is the lantern by which what
was before dimly visible now looms up in
firm, bold outlines. The old phantasmagoria
disappear. We see better. We also see further.
—Ining Fisher (1892)
1.1 Why Economists Use Mathematics
Economic activity has been part of human life for thousands of years. The word
“economics” itself originates from a classical Greek word meaning “household
management.” Even before the Greeks there were merchants and traders who
exhibited an understanding of some economic phenomena; they knew, for instance,
that a poor harvest would increase the price of com, but that a shortage of gold
might result in a decrease in the price of com. For many centuries, the most
basic economic concepts were expressed in simple terms requiring only the use of
rudimentary mathematics. Concepts like integers and fractions, together with the
operations of addition, subtraction, multiplication, and division, were sufficient to
allow traders, merchants, farmers and other economic agents to discuss and debate
the economic activities and events that affected their daily lives. These tools were
enough to enable merchants to keep accounts and to work out what prices to charge.
'“Phantasmagoria” is 2 term invented in 1802 to describe an exhibition of optical illusions
produced by means of @ magic lantern. 2 Chapter 1 / introduction
Even calculations of interest on loans were not very complicated. Arithmetic could
perform the tasks that merchants required of it even without the concept of 2er0
and the decimal system of notation. Where a calculating device was required, the
abacus was powerful enough.
The science of economics reached a turning point in the eighteenth cen-
tury with the publication of works such as David Hume's Political Discourses
(1752), Frangois Quesnay’s Tableau Economique (1758-1759), and Adam Smith's
The Wealth of Nations (1776). Economic arguments began to be formalized and
developed into theories. This created the need to express increasingly complex
ideas and interrelationships in a straightforward manner. By the mid-1800s, some
writers were beginning to use mathematics to communicate their theories. Some
of the first to do this were economists such as. Antoine Cournot (the first writer to
define and draw an explicit demand curve, and to use calculus in solving a maxi-
mization problem in economics) and Léon Walras (who distinguished himself by
writing down and solving the first multiequation model of general equilibrium of
supply and demand in ali markets simultaneously). They found that many of their
ideas could be formulated most effectively by means of mathematical language,
including algebraic symbols, simple diagrams, and graphs. Indeed, much more
sophisticated economic concepts and increasingly complex economic theories have
become possible as mathematical language has been used to express them.
Today, a firm understanding of mathematics is essential for any serious stu-
emt of economics. Although simple economic arguments relying on only two or
three variables can sometimes be made in a clear and convincing fashion without
mathematics, if we want to consider many variables and the way they interact, it
becomes essential to resort to a mathematical model.
As an example, suppose that some government agency is planning to allow
a large amount of new housing to be constructed on some land it controls. What
consequences will this have for employment? Initially, new jobs will be created
in the construction sector as laborers are hired for the project. Moreover, the
construction of new houses requires bricks, cement, reinforcing steel, timber, glass,
and other building materials. Employment must also grow in firms that manufacture,
these materials. These producers in tum require materials from other producers,
and so on. In addition to all these production effects, increased employment leads
to increased incomes. If these income gains are not entirely neutralized by taxes,
then a greater demand for consumer goods results. This in tam leads to an increased
need for employment among producers of consumer goods, and again the flow of
input requirements expands. At the same time, there are feedbacks in the system;
for example. increased incomes also generate more demand for housing. In this
manner, both positive and negative changes in one sector are transmitted to other
sectors of the economy.
The point of this example is that the economic system is so complex that the
final effects are difficult to determine without resorting to more formal mathematical
devices such as a “circular-fiow model” of the entire economy. An example will
be the input-output model presented in Section 12.1 Sec. 1.2 / Scientific Method in the Empirical Sciences 3
Mathematical Analysis
‘The principal topic of this book is an important branch of mathematics called
mathematical analysis. This includes differential and integral calculus and their
extensions. Calculus was developed at the end of the seventeenth century by New-
ton and Leibniz. Their discoveries completely transformed mathematics, physics,
: and the engineering sciences, giving them all new life. In similar fashion, the
introduction of calculus into economics has radically changed the way in which
economists analyze the world around them. Calculus is now employed in many
different areas of economics: for example, it is used to study the effects of relative
price changes on demand, the effects of a change in the price or availability of an
essential input such as oi] on the production process, the consequences of popula-
tion growth for the economy, and the extent to which a tax on energy use might
reduce carbon dioxide emissions.
The following episode illustrates how economists can use mathematical anal-
ysis to solve practical problems. In February 1953, the Netherlands was struck by a
catastrophic flood far more extensive than any previously recorded. The dikes pro-
tecting the country were washed away and over 1800 people died. Total damages
were estimated at about 7% of national income for that year. A commission was
established to determine how to prevent similar disasters in the future. Rebuilding
the dikes to ensure 100% security would have cost an astronomical amount, even
if it were possible at all. The real problem therefore involved a trade-off between
cost and security: higher dikes would obviously cost more, but would reduce both
the probability and likely severity of future flooding. So the commission had to
try to select the optimal height for the dikes. Some economists on the commis-
sion applied cost-benefit analysis, a branch of economics that involves the use of
mathematical analysis, in order to weigh the relative costs and benefits of different
alternatives for rebuilding the dikes. This problem is discussed in more detail in
Problem 7 in Section 8.4.
Such trade-offs are central to economics. They lead to optimization problems
of a type that is naturally handled by mathematical analysis.
1.2 Scientific Method in the Empirical Sciences
Economics is now generally considered to be one of the empirical sciences. These
sciences share a common methodology that includes the following as its most
important elements:
1. Qualitative and quantitative observations of phenomena, either directly or by
! carefully designed experiment.
2, Numerical and statistical processing of the observed data.
3, Constructing theoretical models that describe the observed phenomena and
explain the relationships between them. 4 Chapter 1 / introduction
4. Using these theoretical models in order to derive predictions.
5. Correcting and improving models so that they predict better.
Empirical sciences thus tely on processes of observation, modeling, and verifica-
tion. If an activity is to qualify fully as an empirical science, each of the foregoing
points is important. Observations without theory can only give purely descriptive
pictures of reality that lack explanatory power. But theory without observation
risks losing contact with the reality that it is trying to explain.
Many episodes in the history of science show the danger of error when
“pure theory” lacks any foundation in reality. For example, around 350 3.c..
Aristotle developed a theory that concluded that a freely falling object travels at
a constant speed, and that a heavier object falls more quickly than a lighter one.
This was convincingly refuted by Galileo Galilei in the sixteenth century when
he demonstrated (panly by dropping objects from the Leaning Tower of Pisa)
that, excluding the effects of air friction, the speed at which any object falls is
proportional to the time it has fallen, and that the constant of proportionality is,
the same for all objects, regardless of their weight. Thus, Aristotle's theory was
eventually disproved by empirical observation.
A second example comes from the science of astronomy. In the yeat 1800,
Hegel advanced a philosophical argument to show that there could only be seven
planets in the solar system. Hegel notwithstanding, an eighth planetary body,
the asteroid Ceres, was discovered in January 1801. The eighth principal planet,
Neptune, was discovered in 1846, and by 1930 the existence of Pluto was
known?
With hindsight, the falseness of these assertions by Aristotle and Hegel ap-
pears elementary. In all sciences. however, false assertions are being put forth
Fepeatedly, only to be refuted later. Correcting inaccurate theories is an important
part of scientific activity, and the previous examples demonstrate the need to ensure
that theoretical models are supported by empirical evidence.
In economics, hypotheses are usually less precise than in the physical sci-
ences, and so less obviously wrong than Aristotle's and Hegel’s assertions just
discussed. But there are a few old theories that have since become so discredited
tha few economists now take them seriously. One example is the “Phillips curve,”
that purported to show how an economy could trade off unemployment against
inflation, The idea was that employment might be created through tax cuts and/or
increased public expenditure, but at the cost of increased inflation. Conversely.
inflation could be reduced by tax increases or expenditure cuts, but at the cost of
higher unemployment.
2The process of discovery clied on looking at how the motion of other known planets deviated
from the otbits predicted by Newton's theory of gravitation. These deviations even suggested where to
look for an additional planet that could. according to Newton's theory. account for them. Unuil recently
scientists were still using Newton's theory to search for a tenth planetary body whose existence they
suspected. However, more accurate estimates of the masses of the outer planets now suggest that there
‘ae no further planets to find after all Sec. 1.2 / Scientific Method in the Empirical Sciences 5
Unlike Hegel, who could never hope to count all the planets, or Aristotle. who
presumably never watched with any care the fall of an object that was dropped from
rest, the Phillips curve was in fact based on rather careful empirical observation. in
an article published in 1958, A. W. Phillips examined the average yearly rates of
wage increases and unemployment for the economy of the United Kingdom over
the long period from 1861 to 1957. The plot of those observations formed the
Phillips curve, and the infiation-unemployment trade-off was part of conventional
economic thinking until the 1970s. Then, however, the decade of simultaneous high
inflation and high unemployment (stagnation and inflation, generally abbreviated
stagflation”) that many Westem economies experienced during the period 1973
1982 produced observations that obviously lay well above the usual Phillips curve.
The alleged trade-off became hard to discern.
Just as Aristotle’s and Hegel’s assertions were revised in the light of suitable evidence,
this stagflationary episode caused the theory behind the Phillips curve to be modified. It was
suggested that as people learn to live with inflation, they adjust wage and loan contracts to
reflect expected rates of inflation. Then the trade-off between unemployinent and inflation
‘hat seemed to be described by the Phillips curve becomes replaced by 2 new trade-off
between unemployment and the deviation in inflation from its expected rate. Moreover, this
expected rate increases as the current rate of inflation rises. So lowering unemployment was,
thought to lead not simply 10 increased inflation, but to accelerating inflation that increased
each period by more than was expected previously. On the other hand, when high inflation
came to be expected, combating it with policies leading to painfully high unemployment
would lead only to gradual decreases in inflation, as people’s expectations of inflation fall
rather slowly. Thus, the original Phillips curve theory has been significantly revised and
extended in the light of more recent evidence.
Models and Reality
In the eighteenth century, the philosopher Immanuel Kant considered Euclidean
geometry to be an absolutely true description of the physical space we observe
through our senses. This conception seemed self-evident and was shared by all
those who had reflected upon it. The reason for this agreement was undoubtedly
that all the results of this geometry could be derived by way of irrefutable logic
from only a few axioms, and that these axioms were regarded as self-evident truths
about physical space. The first person to question this point of view was the
German mathematician Gauss at the beginning of the 1800s. He insisted that the
relationship between physical space and Euclid’s model could only be made clear
by empirical methods, During the 1820s, the first non-Euclidean geometry was
developed—tbat is, a geometry built upon axioms other than Euclid’s. Since that
time it has been accepted that only observations can decide which geometric model
gives the best description of physical space.
This shows how there can be an important difference between a mathematical
model and its possible interpretations in reality. Moreover, it may happen that
more than one model is capable of describing a certain phenomenon, such as the 6 Chapter 1 / introduction
relationship between money supply and inflation in the United States or Germany.
Indeed, this often seems to be the case in economics. As long as all the models
to be considered are internally consistent, the best way to select among competing
explanations is usually to see which one gives the best description of reality. But
this is often surprisingly difficult, especially in economics.
In addition, we must recognize that a model intended to explain a phe-
nomenon like inflation can never be considered as absolutely true; it is at best only
an approximate representation of reality. We can never consider all the factors
that influence such a complex phenomenon. If we tried to do so, we would ob-
tain @ hopelessly complicated theory. This is true not only for models of physical
phenomena, but for all models within the empirical sciences.
These comments are particularly relevant for economic research. Consider
once again the effects of allowing new housing to be built. In ofder to understand
the full implications of this, an economist would require an incredible amount of
data on millions of consumers, businesses, goods and services, etc. Even if it were
available in this kind of detail, the amount of data would swamp the capacities of
even the most modern computers. In their attempts to understand the underlying
relationships in the economy, economists are therefore forced to use various Kinds
of aggregate data, among other simplifications. Thus, we must always remember
that a model is only able to give an approximate description of reality; the goal of
empirical researchers should be to make their models reflect reality as closely and
accurately as possible.
1.3 The Use of Symbols in Mathematics
Before beginning to study any subject, it is important that everyone agrees on
common “language” in which to discuss it. Similarly, in the study of mathematics,
which is in a sense a “language” of its own, it is important to ensure that we all
understand exactly the same thing when we see a given symbol. Some symbols in
smathematics nearly always signify the same definite mathematical object. Examples
are 3, V2, =, and {0. 1], which respectively signify three special numbers and
a closed interval. Symbols of this type are called logical constants. We also
frequently need symbols that can represent variables. The objects that a variable
is meant to represent are said to make up the domain of variation. For example,
we use the letter x as a symbol for an arbitrary number when we write
x 16 = (x +4)(x —4)
In words the expression reads as follows:
The difference between the square of the number (hereby called x)
and 16 is always equal to the product of the two numbers obtained by
adding 4 to the number and subtracting 4 from the number x. ‘Sec. 1.3 / The Use of Symbols in Mathematics 7
The equality x? — 16 = (x + 4)(x — 4) is called an identity because it is valid
identically for all x. In such cases. we sometimes write x7 ~ 16 = (x +4)(x — 4),
where = is the symbol for an identity.
‘The equality sign (=) is also used in other ways. For example, we write
A= ms? as the formula for the area A of a circle with radius r. In addition, the
equality sign is used in equations such as
e+x-12=0
where x stands as a symbol for the unknown number. If we substitute various
numbers for x, we discover that the equality sign is often invalid, in fact, the
equation is only true for x = 3 and for x = —4, and these numbers are therefore
called its solutions.
Example 1.1
A farmer has 1000 meters of fence wire with which to enclose a rectangle.
If one side of the rectangle is x (measured in meters), find the area enclosed
when x is chosen to be 150, 250, 350, and for general x. Which value of x
do you believe gives the greatest possible area?
Solution If the other side of the rectangle is y, then 2x + 2y = 1000.
Hence, x + y = 500, so that y = 500 — x. (See Fig. 1.1.) The area A (in
m*) of this rectangle is, therefore,
A =x(500 — x) = 500x — x?
Because both sides must be positive, x must be positive and 500 — x
rust be positive, This means that x must be between 0 and 500 m. The areas
when x = 150, 250, and 350 are 150. 350 = 52,500, 250 - 250 = 62,500,
and 350 - 150 = 52,500, respectively. Of these, x = 250 gives the greatest
value. In Problem 7 of Section 3.1 you will be asked to show that x = 250
really does give the greatest possible area.
When studying problems requiring several (but not 100 many) variables, we
usually denote these with different letters such as a, b, c, x, y. z, A, B, and so on.
Often, we supplement the letters of the Latin alphabet with lowercase and capital
FIGURE 1.1 8 Chapter 1 / Introduction
Greek letters such as a, B, 7, T°, and Q. If the aumber of variables becomes large,
we can use subscripts or superscripts to distinguish variables from each other. For
example, suppose that we are studying employment in a country that is divided into
100 regions, numbered from 1 to 100. We can then denote employment in region 1
by Mi, employment in region 2 by No, and so on. In general, we can define
N; = total employment in region i,
+2,.--,100
The suffixes i = 1, 2, ..., 100 suggest that the index i can be an arbitrary number
in the range from 1 to 100. If Nsp = 2690, this means that 2690 people are em-
ployed in region 59. If we want to go further and divide the employed into men
and women, we could denote the number of women (men) employed in region i
by Wi (wi), Then, we would have Ni") + NM"? = N;, fori =1,2,..., 100.
Note that this notation is actually much clearer than if we were to use 100 different
letters to represent the variables N;—even if we could find 100 different letters
from some combination of the Latin, Greek, Cyrillic, and Sanskrit alphabets!
Many students who are used to dealing with algebraic expressions involving
only one variable (usually x) have difficulties at first in handling expressions in-
volving several variables. For economists, however, the previous example shows
how important it is to be able to handle algebraic expressions and equations with
many different variables. Here is another example.
Example 12
Consider the simple macroeconomic model
Y=C+I, C=at+by a
where Y is the net national product, C is consumption, and J is the total
investment, which is treated as fixed.’ The three letters, 7, a, and b, denote
positive numerical constants—for example, | = 100, a = 500, and b = 0.8
are possible values of these constants. Rather than thinking of the two mod-
els with T = 100, C = 500+0.8Y and with J = 150, C = 600 + 0.9¥ as
entirely different, however, it is often more sensible to regard them as two
particular instances of the general model (1], where 7, a, and 6 are un-
known, and can vary; they are often called parameters. But they should be
distinguished from the variables C and ¥ of the model.
‘After this discussion of constants as parameters of the model. solve (1)
for ¥.
Solution Substituting C = a + bY from the second equation of (1] for
C into the first equation gives
Y=a+byst
*In economics. we often use a bar over a symbol to indicate that itis fixed. Sec. 1.3 / The Use of Symbols in Mathematics 9
Now rearrange this equation so that all the terms containing Y are on the
left-hané side. This can be done by adding ~bY to both sides, thus canceling
the bY term on the right-hand side to give
Y¥-by=at+i
Notice that the left-hand side is equal to (1 — 6)¥. so (1—b)¥ =a +i.
Dividing both sides by 1 — b, so that the coefficient of Y becomes 1, then
gives the answer, which is
This solution is a formula expressing Y im terms of the three parameters /,
a, and b. The formula can be applied to particular values of the constants,
such as 7 = 100, a = 500. b = 0.8, to give the right answer in every case.
Note the power of this approach: The model is solved only ance, and then
‘numerical answers are found simply by substituting appropriate numerical
values for the parameters of the model.
Problems
1. a, A person buys x1, x2. and x3 units of three goods whose prices per unit
are, respectively, pj, p>, and ps. What is the total expenditure?
b. A rental car costs F dollars per day in fixed charges and 6 dollars per kilo-
meter. How much must a customer pay to drive x kilometers in 1 day?
. A company has fixed costs of F dollars per year and variable costs of ¢
dollars per unit produced. Find an expression for the total cost per unit
(otal average cost) incurred by the company if it produces x units in one
year.
. A person has an annual salary of $Z and then receives a raise of p%
followed by a further increase of 9%. What is the person’s new yearly
salary?
e. A square tin plate 18 cm wide is to be made into an open box by cutting
‘out equally sized squares of width x in each corner and then folding over
the edges. Find the volume of the resulting box. (Draw a figure.)
2. a. Prove that
can be written as 10 Chapter 1 / introduction
b. An item initially costs $2000 and then its price is increased by 5%.
Afterwards the price is lowered by 5%. What is the final price?
¢. An item initially costs a dollars and then its price is increased by p%.
Afterwards the (new) price is lowered by p%. What is the final price
of the item? (After considering this problem, look at the expression in
Part (a).)
4. What is the result if one first lowers a price by p% and then increases it
by p%?
3. Solve the following equations for the variables specified:
ax=2y—3)+y for y bo ax-b=cx+d for x
© AKVL=Y for L d. pxt+qy=m for y
1
—-a
e te =e forr f& ¥=a(¥-1¥ —k)+b+1,+G forY
mart
4. The relationship between a temperature measured in degrees Celsius (or
Centigrade) (C) and in Fahrenheit (F) is given by C = $(F — 32).
a. Find C when F is 32; find F when C = 100.
b. Find a general expression for F in terms of C.
c. One day the temperature in Oslo was 40°F, while in Los Angeles it was
80°F. How would you respond to the assertion that it was twice as warm
in Los Angeles as in Oslo? (Hint: Find the two temperatures in degrees
Celsius.)
. If a rope could be wrapped around the earth’s surface at the equator, it
would be approximately circular and about 40 million meters long. Suppose
we wanted to extend the rope to make it 1 meter above the equator at every
point. How many more meters of rope would be needed? (Guess first,
and then find the answer by precise calculation. For the formula for the
circumference of the circle, see Appendix D.)
w
Harder Problems
6. Solve the following pair of simultaneous equations for x and y:
pre(l-g)y=R and gxt(l—p)y=S
7. Consider an equilateral triangle. and let P be an arbitrary point within the
triangle. Let fj, hz, and hs be the shortest distances from P to each of the
three sides. Show that the sum fy + hy + hs is independent of where point
P is placed in the iangle. (Hini: Compute the area of the triangle as the
sum of three triangles.) Sec. 1.4 / The Rea! Number System 11
1.4 The Real Number System
God created the integers;
everything else is the work of man.
—L. Kronecker
Real numbers were originally developed in order to measure physical characteristics
such as length, temperature, and time, Economists also use real numbers to measure
prices, quantities, incomes, tax rafes, interest rates, and average costs, among other
things. We assume that you have some knowledge of the real number system, but
because of its fundamental role, we shall restate its basic properties.
Natural Numbers, Integers, and Rational
Numbers
The everyday numbers we use for counting are 1, 2, 3, .... These are called
natural numbers. Though familiar, such numbers are in reality rather abstract and
advanced concepts. Civilization crossed a significant threshold when it grasped the
idea that a flock of four sheep and a collection of four stones have something in
common, namely “fourness.” This idea came to be represented by symbols such
as the primitive :: (still used on dominoes or playing cards), the modem 4, and
the Roman numeral IV. This notion is grasped again and again as young children
develop their mathematical skills.
During the early stages of many cultures, day-to-day problems motivated
the four basic arithmetic operations of addition, subtraction, multiplication, and
division. If we add or multiply two natural numbers, we always obtain another
nafural number. Moreover, the operations of subtraction and division suggest the
desirability of having a number zero (4~4 = 0), negative numbers (3-5 = -2),
and fractions (3 +5 = 3/5). The numbers 0, +1, +2, +3,... are called the inte-
gers. They can be represented on a number line like the one shown in Fig. 1.2.
The rational numbers are those like 3/5, that can be written in the form a/b,
where a and 6 are both integers, An integer n is also a rational number, because
n= n/\. Examples of rational numbers are
1 wu 425 10 0 126
Lr S.-i, -126=-22
op Tt T : 1265 —s05
The rational numbers can also be represented on the number line. Imagine that
we first mark 1/2 and all the multiples of 1/2. Then we mark 1/3 and all the
FIGURE 1.2. The number fine. 12 Chapter 1 / introduction
=1 0 |
FIGURE 1.3
multiples of 1/3, and so forth. You can be excused for thinking that “finally” there
will be no more places left for putting spore points on the line. But in fact this is
quite wrong. The ancient Greeks already understood that “holes” would remain in
the number line even after all the rational numbers had been marked off. This is
demonstrated in ‘the construction in Fig. 1.3.
Pythagoras’ theorem tells us that s* = 1? + 1? = 2, sos
shown, however, that there are no integers p and q such that /2
¥2is not a rational number. (Euclid proved this fact in about 300 B.c. See Problem
3 in Section 1.6.)
The rational numbers are therefore insufficient for measuring al! possible
lengths, let alone areas and volumes. This deficiency can be remedied by ex-
tending the concept of numbers to allow for the so-called irrational numbers.
This extension can be carried out rather naturally by using decimal notation for
numbers.
The Decimal System
The way most people write numbers today is called the decimal system, or the
base 10 system. It is a positional system with 10 as the base number. Every
natural number can be written using only the symbols, 0, 1, 2, ..., 9, that are
called digits. You will note that “digit” also means “finger,” or “thumb,” and that
most humans have 10 digits. The positional system defines each combination of
digits as a sum of exponents of 10. For example,
1984 =1-10°+9-10°+8-10' +4-10°
Each natural number can be uniquely expressed in this manner. With the use of
the signs + and ~, all integers, positive or negative, can be written in the same
way. Decimal points also enable us to express rational numbers other than natural
numbers. For example,
3.1415 = 3+ 1/10! +4/10 + 1/10? + 5/10"
Rational numbers that can be written exactly using only a finite number of decimal
places are called finite decimal fractions.
Each finite decimal fraction is a rational number, but not every rational num-
ber can be written as a finite decimal fraction, We also need to allow for infinite Sec. 14 / The Real Number System 13.
decimal fractions such as
100/3 = 33.333
where the three dots indicate that the decimal 3 recurs indefinitely.
If the decimal fraction is a rational number, then it will always be periodic—
thatis, after a certain place in the decimal expansion, it either stops or continues to
repeat a finite sequence of digits. For example, 11/70 = 0.1 5714285714285...
2T1S28 371428.
Real Numbers
The definition of a real number follows from the previous discussion. We define a
real number as an arbitrary infinite decimal fraction. Hence, a real number is of the
form x = +:m.oa203..., where m is an integer, and a, (1 1,2...) is an infinite
series of digits, each in the range 0 t0 9. We have already identified the periodic
decimal fractions with the rational numbers. In addition, there are infinitely many
new numbers given by the nonperiodic decimal fractions. These are called irra-
tional numbers, Examples include /2, V5, x, 2¥?, and 0.12112111211112....
It tums out that, in general, it is very difficult to decide whether a given
smamber is rational or irrational. It has been known since the year 1776 that 2 is
irrational and since 1927 that wis irrational. However, we still do not know as
of 1993 whether 2%? + 33 js irrational or not. One might gain the impression that
there ace relatively few irrational numbers. In fact, there are (in a certain precise
sense) infinitely more irrational numbers than there are rational numbers.
‘We mentioned earlier that each rational number can be represented by a
point on the number line. But not all points on the number line represent rational
numbers. It is as if the irrational numbers “close up” the remaining holes on
the number line after all the rational numbers have been positioned. Hence, an 14 = Chapter 1 / introduction
unbroken and endless straight line with an origin and a positive unit of length is
a satisfactory model for the real numbers. We frequently state that there is a one-
to-one correspondence between the teal numbers and the points on a number line.
The rational and irrational numbers are said 10 be “dense” on the number
line. This means that between any two different real numbers. irrespective of how
close they are to each other, we can always find both a rational and an irrational
number—in fact, we can always find infinitely many of each.
When applied to the real numbers, the four basic arithmetic operations always
result in a real number. The only exception is that we cannot divide by 0.
i is not defined for any real number a
This is very important and should not be confused with 0/a = 0, forall a #0. No-
tice especially that 0/0 is not defined as any real number. For example, if a car re-
quires 60 liters of fuel to go 600 kilometers, then its fuel consumption is 60/600 = 10
liters per 100 kilometers. However, if told that a car uses 0 liters of fuel to go 0 kilo-
meters, we know nothing about its fuel consumption: 0/0 is completely undefined.
Inequalities
In mathematics and especially in economics, inequalities are encountered almost
as often as equalities. it is important, therefore. to know and understand the rules
for carring out calculations involving inequalities. These are presented in Section
A.7 in Appendix A. The following example is of interest in statistics.
Example 13
Show that if a > 0 and b > 0, then
vab
aa
Solution (You shouid first test this inequality by choosing some specific
numbers. using a calculator if you wish.) To show the given inequality. it is
enough to verify that ab < (a +5)°/4 because then the square root of the
Jeft-hand side cannot exceed the square root of the right-hand side—that is.
Vab < 3(a +b). To verify this, it is enough to check that the right-hand
side minus the left-hand side is nonnegative. But indeed
(a-by
ae
0
In fact. essentially the same proof can be used to show that Jab < $(a +b)
unless a = b. Sec. 14./ The Real Number System 15
‘The number 4(a +5) is called the arithmetic mean of a and b, and Jab
is called the geometric mean. What does the inequality in [1.1] state about the
different means?
Intervals
If a and b are two numbers on the number line, then we call the set of all num-
bers that lie between a and b an interval. In many situations, it is important 10
distinguish between the intervals that include their endpoints and the intervals that
do not. When a < &, there are four different intervals that all have a and b as
endpoints, as shown in Table 1.1. Note that the names in the table do not distin-
guish (a, 6) from (a, b}. To do so, one could speak of “closed on the left,” “open
‘on the right,” and so on. Note, too, that an open interval includes neither of its
endpoints, but a closed interval includes both of its endpoints. All four intervals,
however, have the same length, b — a.
‘We usually illustrate intervals on the number line as in Fig. 1.4, with in-
cluded endpoints represented by dots, and excluded endpoints at the tips of arrows.
The intervals mentioned so far are all bounded intervals. We also use the word
“interval” to signify certain unbounded sets of numbers. For example, we have
[a, 00) = all numbers x, with x > @
(00, b) = all numbers x, with x 0, thatis, x >
=x ifx—2 <0. thatis. x < Sec. 1.4 / The Real Number System 17
Hence.
cya fer2 fre
me \eeet ifx<
(Check this answer by trying the values of x tested in part (a).)
Let x; and x2 be two arbitrary numbers. The distance between x; and x2 on
the nomber line is equal to x1 — x2 if x1 > x2, and equal to —(x) ~ x9) if x1 < xy
‘Therefore, we have
lx; — x2| = distance between x; and x2 on the number line 13)
In Fig. 1.5, we have indicated geometrically that the distance between 7 and 2 is
5, whereas the distance between —3 and —S is equal to 2, because |—3 — (—5)| =
|-3 +5) = 2) =2.
‘Suppose |x| = 5. What values can x have? There are only two possibilities:
either x = 5 of x = —5, because no other numbers have absolute values equal to 5,
Generally, if a is greater than or equal to 0, then |x| =a means that x = a or x =
—a. Because |x| > 0 for all x, the equation |x| = a has ne solution when a < 0.
If a is a positive number and |x| < a, then the distance from x to 0 is less,
than @, and so
4
xl y—(r-y) +h
ce pr+qy0)
ye
. Consider Problem 1(c) in Section 1.3, Set up an inequality that determines
‘how many units x the company must produce before the average cost falls
below Sq. Solve the inequality for x. Put F = 100.000. ¢ = 120, q = 160,
and solve the problem for this case.
6 Calculate [2x ~ 3), for x= 0, 1/2, and 7/2.
7. a, Calculate [5 - 3x), for x =—1. 2. and 4.
b. Solve the equation |5 — 3x| = 0.
¢. Rewrite |5 — 3x| by using [1.2].
8. Determine x such that
a. |3 — 2x] b. ix) <2 e-2s1
d. 3-8x/ <5 e |x] > v2 f. jx? -2) <1
9. A S-meter iron bar is to be produced. It is necessary that the length does not
deviate more than 1 mm from its stated size. Write a specification for the Sec. 1.5 / A Few Aspects of Logic | 19.
rod’s length x in meters: (a) by using a double inequality and (b) with the
aid of an absolute-value sign.
1.5 A Few Aspects of Lagic
An astronomer, a physicist, and a mathematician were travelling on @ train in
Scotland. Through the window they saw a flock of sheep grazing in a meadow.
The astronomer remarked, “In Scotland all sheep are black.”
The physicist protested, “Some Scottish sheep are black.”
The mathematician declared, “In Scatland there exists a flock of sheep all of
which are black on at least one side.”
So far we have emphasized the role of mathematical models in the empirical
sciences, especially in economics. The more complicated the phenomena to be
described, the more important it is to be exact. Errors in models applied to practical
situations can have catastrophic consequences. For example. in the early stages of
the U.S. space program, a rocket costing millions of dollars to develop and build
had to be destroyed only seconds after launch because a semicolon had been left
out of the computer program intended to control the guidance system.
Although the consequences may be less dramatic, errors in mathematical
reasoning also occur rather easily. In what follows, we offer a typical example
of how a student (or professor) might use faulty logic and thus end up with an
incorrect answer to a problem.
Example 1.6
Find a possible solution for the equation x +2 = Va=3.
“Solution” Squaring each side of the equation gives (x + 2)? =
(4x), and thus x7 + 4x +4 = 4~ x. Rearranging this last equation
gives x? +5x =0. Canceling x results in x +5 = 0, and therefore x = —S.
According to this reasoning, the answer should be x = ~5. Let us
check this, For x = ~5, we have x +2 = —3. Yet V@—x = V9 = 3,
so this answer is incorrect. In Example 1.9, we explain how the error arose.
(Note the wisdom of checking your answer whenever you think you have
solved an equation.)
This example highlights the dangers of routine calculation without adequate
thought, it may be easier to avoid similar mistakes after studying more closely the
structure of logical reasoning.
Propositions
Assertions that are either tue or false are called statements, or propositions. Most
of the propositions in this book are mathematical ones, but others may arise in daily
life. “All individuals who breathe are alive” is an example of a true proposition, 20 © Chapter 1 / introduction
whereas the assertion “all individuals who breathe are healthy” is an example of
a false proposition. It should be noted that if the words used to express such
assertions lack a precise meaning, it will often be difficult to distinguish between
a true and a false proposition.
Suppose an assertion such as “x? — 1 = 0” includes one or more variables.
By substituting various real numbers for the variable x, we can generate many
different propositions, some ue and some false. For this reason we say that the
assertion is an open proposition. In fact, the proposition x> — 1 = 0 happens
to be tue if x = 1 or —1, but not otherwise. Thus, an open proposition is
not simply true or false. It is neither true nor false until we choose a particular
value for the variable. In practice we are somewhat careless about this distinction
between propositions and open propositions; instead, we simply call both types
propositions.
Implications
In order to keep track of each step in a chain of logical reasoning, it often helps
to use implication arrows,
Suppose P and Q are wo propositions such that whenever P is true, then
Q is necessarily true. In this case, we usually write
P=@ {l
This is read as “P implies Q.” or “if P, then Q,” or “Q is a consequence of P.”
‘The symbol => is an implication arrow, and it points in the direction of the
logical implication. Here are some examples of correct implications.
Example 1.7
fa x>2 => Pod
(b) xy =0 = x=0ory=0.
(©) x is a square => x isa rectangle.
(@) x is a healthy person => x is breathing.
Notice that the word “or” in mathematics means the “inclusive or,” signifying
that “P or Q” means “either P or Q or both.”
All the propositions in Example 1.7 are open propositions, just as are most
propositions encountered in mathematics. An implication P => Q means that
for each value of some variable for which P is true. Q is also true.
In certain cases where the implication (x] is valid. it may also be possible to
draw a logical conclusion in the other direction:
Qo=P Sec 15/4 Few Aspects of Logic 24
In such cases, we can write both implications together in 4 single logical equiva-
lence:
PO
We then say that “P is equivalent to Q,” or “P if and only if Q,” or just “P iff
Q.” Note that the statement “P only if 0” expresses the implication P => Q,
whereas “P if Q” expresses the implication Q => P.
The symbol <= is an equivalence arrow. In previous Example 1.7, we
see that the implication arrow in (b) could be replaced with the equivalence arrow,
because it is also true that x = 0 or y = 0 implies xy = 0. Note, however, that
no other implication in Example 1.7 can be replaced by the equivalence arrow.
For even if x? is larger than 4, it is not necessarily true that x is larger than
2 (for instance, x might be —3); also, a rectangle is not necessarily a square;
and, finally, just because person x is breathing does not mean that he of she is
healthy.
Necessary and Sufficient Conditions
There are other commonly used ways of expressing that proposition P implies
proposition Q, or that P is equivalent to Q. Thus, if proposition P implies
proposition Q, we state that P is a “sufficient condition” for Q. After all, for
Q to be true, it is sufficient that P is true, Accordingly, we know that if P is
satisfied, then it js certain that Q is also satisfied. In this case, we say that Q
is a “necessary condition” for P. For Q must necessarily be true if P is tue,
Hence,
P is a sufficient condition for Q means: P => Q
Q is a necessary condition for P means: P => Q
For example, if we formulate the implication in Example 1.7(c) in this way, it
would read:
A necessary condition for x to be a square is that x be a rectangle.
A sufficient condition for x to be a rectangle is that x be a square.
The corresponding verbal expression for P <=> @ is simply: P is a necessary
and sufficient condition for Q, or P if and only if Q, or P iff Q. I is evident
from this that it is very important to distinguish between the propositions “P is a
necessary condition for Q” (meaning Q => P) and “P is a sufficient condition 22 Chapter 1 / Introduction
for Q” (meaning P > Q). To emphasize the point, consider two propositions:
1, Breathing is a necessary condition for a person to be healthy.
2. Breathing is a sufficient condition for a person to be healthy.
Evidently proposition 1 is true. But proposition 2 is false, because sick (living)
people are still breathing. In the following pages, we shall repeatedly refer to
necessary and sufficient conditions. Understanding them and the difference between
them is a necessary condition for understanding much economic analysis. It is not
a sufficient condition, alas!
Solving Equations
‘We shall now give examples showing how using implication and equivalence ar-
rows can help avoid mistakes in solving equations like that in Example 1.6.
Example 1.8
Find all x such that (2x — 1)? — 3x7 = 2(} —4x).
Solution —_By expanding (2x — 1)? and also multiplying out the right-hand
side, we obtain a new equation that obviously has the same solutions as the
original one:
(Qx ~ 1-327 =2(} 4x) es 4x? -4r + 1-32? = 1-8
Adding 8x — 1 to each side of the second equality and then gathering terms
gives the equivalent expression
At mart 13x? = 18x a P44 = 0
Now x? +4x = x(x +4), and the latter expression is © if and only if x = 0
or x+4=0. That is,
Pidxs0 = x@$4)=0 exa0 or xt4=0
x0 of x=-4
Putting everything together, we have derived a chain of equivalence arrows
showing that the given equation is fulfilled for the two values x = 0 and
x = —4, and for no other values of x. That is.
Example 1.9
Find all x such that x +2 = /@—z. (Recall Example 1.6.) Sec. 1.5. / A Few Aspects of Logic 23
Solution Squaring both sides of the given equation yields
(Vax)
(+2)?
Consequently, x? + 4x +4 =
equation it follows that
— x, that is, x7 +5x = 0. From the latter
x(x +5) =
which implies x = 0 or x = —5. Thus, a necessary condition for x to solve
x42 = J4=x is that x = 0 or x = —S. Inserting these two possible values
of x into the original equation shows that only x = 0 satisfies the equation,
‘The unique solution to the equation is, therefore, x = 0.
In finding the solution to Example 1.9, why was it necessary to test whether
the values we found were actually solutions, whereas this step was unnecessary in
Example 1.8? To answer this, we must analyze the logical structure of our solution
to Example 1.9. With the aid of numbered implication and equivalence arrows, we
can express the previous solution as
0
xt2avinzr Se
® 3 ® 6
S Pesrs0S 4520S x e007r=
Implication (1) is true (because a = 6 => a? = B and (Ya) =a). Iris
important to note, however, that the implication cannot be replaced by an equiva-
lence. If a? = 6”, then either a = b or a = —b; it need not be tue that a = b.
Implications (2), (3), (4), and (5) are also all true; moreover, all could have been
written as equivalences, though this is not necessary in order to find the solution.
Therefore, a chain of implications has been obtained that leads from the equation
x+2= J4—x to the proposition “x = 0 or x = —5.” Because the implication
(1) cannot be reversed, there is no corresponding chain of implications going in the
‘opposite direction. We have verified that if the number x satisfies x +2 = /4—x,
then x must be either 0 or —5; no other value can satisfy the given equation.
However, we have not yet shown that either 0 or ~S really satisfies the equation
Until we try inserting 0 and —5 into the equation, we cannot see that only x = 0
is a solution. Note that in this case, the test we have suggested not only serves to
check our calculations, but is also a logical necessity.
Looking back at Example 1.6, we now realize that two errors were committed
Firstly, the implication x +5x = 0 = x +5 =0 is wrong, because x = 0 is also
a solution of x7 + 5x =0. Secondly, it is logically necessary to check if 0 or —5
really satisfies the equation.
The method used to solve Example 1,9 is the most common. It involves
setting up a chain of implications that starts from the given equation and ends with 24 Chapter 1 / introduction
8 Set of its possible solutions. By testing each of these trial solutions in tum, we
find which of them really do satisfy the equation. Even if the chain of implications
is also a chain of equivalences (as it was in Example 1.8), such a test is always a
useful check of both logic and calculations.
Problems
1. Implications and equivalences can be expressed in ways that differ from those
already mentioned. Use the implication or equivalence arrows to mark in
which direction you believe the logical conclusions proceed in the following
propositions:
a, The equation 2x — 4 = 2 is fulfilled only when x = 3.
b. If x =3, then 2x —4=2.
¢. The equation x? — 2x +1 = 0 is satisfied if x = 1.
@ If x? >4, then x > 2 or x < —2, and conversely.
2. Consider the following six implications and decide in each case: (i) if the
implication is true, and (i) if the converse implication is true. (x, y. and z
are real numbers.)
ax=2landy=Sax+y
ji
fLx>y—ox>0
3. Consider the proposition 2x +5 > 13.
a. Is the condition x > 0 necessary, sufficient, or both necessary and suffi-
cient for the proposition to be satisfied?
b. Answer the same question when x > 0 is replaced by x > 50.
©. Answer the same question when x > 0 is replaced by x > 4.
4. Solve the equation
G+ G@-P _i3x+) _y
XSI xe $1) OPT
5. Solve the following equations:
ax+2=Var¢13 bd. [x+2)=VS—x cc. x? —2Ix]-3 =0
6. Solve the following equations:
a ve-4=VE+5-9 v. Vr—4=9-Vxt5
. Fill in the blank rectangles with “iff” (if and only if) when this results in a
true statement, or alternatively with “if” or “only if”
arava x=2
x Sec. 1.6 / Mathematica! Proot 25
e. x(x +3) <0 x>-3
8. Consider the following attempt to solve the equation x+/x +4 = 2: “From
the given equation, it follows that /x +4 = 2—-x. Squaring both sides gives
x44=4—41+27. After rearranging the terms, itis seen that this equation
implies x? — 5x = 0. Canceling x, we obtain x — 5 = 0 and this equation is
satisfied when x = 5.”
a. Mark with arrows the implications or equivalences expressed in the text,
Which ones are correct?
b. Give @ correct solution to the equation
9. For each of the following 6 propositions, state the negation as simply as
possible.
a x>Oandy 20.
b. All x satisfy x > a
¢. Neither x nor y is less than 5.
4, For each ¢ > 0, there exists a 5 > 0 such that B is satisfied.
e. No one can avoid liking cats,
£. Everyone loves someone at certain times.
10. “Supreme Court refuses 10 hear challenge to lower court's decision approving
a trial judge’s refusal 10 allow a defendant to refuse to speak.” Has the
defendant the right not fo speak?
1.6 Mathematical Proof
In science, what can be proved should not be believed without proof.
—R. Dedekind (1887)
In every branch of mathematics, the most important results are called theorems.
Constructing logically valid proofs for these results often can be rather complicated.
For example, the “four-color theorem” states that any map in the plane needs at
most four colors in order that all contiguous regions should have different colors,
Proving this involved checking hundreds of thousands of different cases, a task
that was impossible without a sophisticated computer program.
Jn this book, we often omit formal proofs of theorems. Instead, the empha-
sis is on providing a good intuitive grasp of what the theorems tell us. However,
“Yleve is the German original: “Was beweisbar ist, soll in der Wissenschaft nicht ohne Beweis
geglaubt werden.” 26 © Chapter 1 / Introduction
although proofs do not form a major part of this book, itis still useful to understand
something about the different types of proof that are used in mathematics, In fact,
a proof that is actually readable is likely to some extent to rely on the reader's
intuition. Although many mathematical logicians do take care to present every step
and every argument. and this may indeed be a necessary step in enabling computers
to check a proof, the overall result is usually unreadable by most people.
Every mathematical theorem can be formulated as an implication
P=0 [#]
where P represents a proposition or a series of propositions called premises (“what
we know”), and Q represents a proposition or a series of propositions that are called
the conclusions (“what we want to know”). A statement of the form P => Q
can be regarded as two theorems,
Usually, it is most natural to prove a result of the type [+] by starting with the
premises P and successively working forward to the conclusion Q; we call this a
direct proof. Sometimes. however. it is more convenient to prove the implication
P = Q by an indirect proof. In this case, we begin by supposing that Q is not
tue, and on that basis demonstrate that neither can P be true. This is completely
legitimate, because we have the following equivalence:
P==Q ~~ isequivalentto = not Q = nor P [1.6]
It is helpful to consider how this rule of logic applies to some concreve examples:
If it is raining, the grass is getting wer
asserts precisely the same thing as
If the grass is not getting wet, then it is not raining.
If T denotes a triangle, then *
The base angles of T are equal implies that T is isosceles asserts the
same as If T is not isosceles, then its base angles are not equal.
There is a third method of proof that is also sometimes useful. It is called proof by
contradiction. The method is based upon a fundamental logical principle: that it
is impossible for a chain of valid inferences to proceed from a true proposition to a
false one. Therefore, if we have a proposition R and we can derive a contradiction
on the basis of supposing that R is false, then it follows that R must be tue. Sec. 1.6 / Mathematical Proof 27
Example 1.10
Use three different methods to prove that
ax +5x-4>0=2>0
Solution
(a) Direct proof: Suppose —x? + 5x —4 > 0. Adding x? +4 to each side
of the inequality gives Sx > x7 +4. Because x? +4 > 4, for all x, we
have 5x > 4, and so x > 4/5. In particular, x > 0
(b) Indirect proof: Suppose x < 0. Then 5x < 0 and so —x7+5x —4, as
a sum of three nonpositive terms, is < 0.
(©) Proof by contradiction: Suppose that the statement is not true. Then
there has to exist an x such that —x° + 5x —4 > 0 and x <0. But if
x 50, then -x7 45x — 4 < -x? — 4 < —4, and we have arrived at a
contradiction.
Deductive vs. Inductive Reasoning
The three methods of proof just outlined are all examples of deductive reasoning,
that is, reasoning based on consistent rules of logic. In contrast, many branches
of science use inductive reasoning. This process draws general conclusions based
only on a few (or even many) observations. For example, the statement that “the
price level has increased every year for the last n years; therefore, it will surely
increase next year too,” demonstrates inductive reasoning. Owners of houses in
California know how dangerous such reasoning can be in economics. This induc-
tive approach is nevertheless of fundamental importance in the experimental and
empirical sciences, despite the fact that conclusions based upon it never can be
absolutely certain.
In mathematics, inductive reasoning is not recognized as a form of proof.
Suppose, for instance, that the students taking a course in geometry are asked
to show that the sum of the angles of a triangle is always 180 degrees. If they
painstakingly measure as accurately as possible 1000 or even 1 million different
triangles, demonstrating that in each case the sum of the angles is 180, would this
not serve as proof for the assertion? No; although it would represent a very good
indication that the proposition is true, it is not a mathematical proof. Similarly,
in business economics, the fact that a particular company’s profits have risen for
each of the past 20 years is no guarantee that they will rise once again this year.
Nevertheless, there is a mathematical form of induction that is much used in
valid proofs. This is discussed in Section B.5 in Appendix B.
Problems
1. Consider the following (dubious) statement: “If inflation increases, then un-
employment decreases.” Which of the following statements are equivalent? 28 Chapter 1 / introduction
a, For unemployment to decrease, inflation must increase.
b. A sufficient condition for unemployment to decrease is that inflation in-
creases.
¢. Unemployment can only decrease if inflation increases.
4. If unemployment does not decrease, then inflation does not increase.
€. A necessary condition for inflation to increase is that unemployment de-
creases.
2. Analyze the following epitaph: (a) using logic and (b) from a poetic view-
point
Those who knew him, loved him.
Those who loved him not, knew him not.
3, Fill in the details of the following proof that /7 is irrational. Suppose it were
tue that /2 = p/g, where p and q are integers with no common factor.
Then p? = 2g*, which would mean that p*, and hence p, would have 2 as
a factor. Therefore, p = 2s for some integer s, and so 4s* = 2q?. Thus,
q° = 2°. It follows that g would also have 2 as a factor, a contradiction of
the hypothesis that p and q have no common factor.
1.7 Set Theory
Ifyou know set theory up to the hilt, and no other mathematics, you would
be of no use to anybody. If you knew a lot of mathematics, but no set theory,
you might achieve a great deal. But if you knew just some set theory, you
would have a far better understanding of the language of mathematics.
=I. Stewart (1975)
In daily life, we constantly group together objects of the same kind. For instance,
wwe refer to the university faculty to signify all the members of the academic staff
at the university. A garden refers to all the plants that are growing in it. We talk
about all firms with more than 1000 employees, all taxpayers in Los Angeles who
camed between $50,000 and $100,000 in 1992, and so on. In all these cases, we
have a collection of objects viewed as a whole. In mathematics, such a collection
is called a set, and the objects are called the elements of, or the members of. the
set
How is a set specified? The simplest way is to list its members, in any order,
between the two braces { and }. An example is the set
S=(a,b.c}
whose members are the first three letters in the alphabet of most languages of Eu-
ropean origin, including English. Or it might be a set consisting of three members See. 1.7 / Set Theory 29
represented by the letters a, b. and c. For example, if a@ = 0, 6 = 1. and ¢ =
then S = {0.1.2}. Also $ denotes the set of roots of the cubic equation
(x — a(x = db) — 0)
in the unknown x, where a 6, and c are any three real numbers.
Alternatively, suppose that you are to eat a meal at a restaurant that offers
a choice of several main dishes. Four choices might be feasible—fish, pasta,
omelette, and chicken. Then the feasible set F has these four members, and is
fully specified as
F = (fish, pasta, omelette, chicken}
Notice that the order in which the dishes are listed does not matter. The feasible
set remains the same even if the order of the items on the menu is changed.
Two sets A and B are considered equal if each element of A is an element
of B and each element of B is an element of A. In this case, we write A = B.
This means that the two sets consist of exactly the same elements. Consequently,
(1, 2,3} = (3.2.1), because the order in which the elements are listed has no
significance; and (1,1,2,3) = {1,2,3}, because a set is not changed if some
elements are listed more than once.
Specifying a Property
Not every set can be defined by listing all its members, however, Some sets can
be infinite, that is, they contain an infinite number of members.
‘Actually, such infinite sets are rather common in economics. Take, for in-
stance, the budger set that arises in consumer theory. Suppose there are two goods
with quantities denoted by x and y that can be bought at prices p and q, respec-
tively. A consumption bundle (x, y) is a pair of quantities of the two goods. Its
value at prices p and q is px +qy. Suppose that a consumer has an amount m to
spend on the two goods. Then the budget constraint is px +qy < m (assuming
that the consumer is free to underspend). If one also accepts that the quantity con-
sumed of each good must be nonnegative, then the budget ser, that will be denoted
by B, consists of those consumption bundles (x, y) satisfying the three inequalities
px-+qy Sm,x> 0. and y > 0. (The set B is shown in Fig. 2.41.) Standard
notation for such a set is
B={(x,y): px+ay sm, x20, y20} {71
The braces { } are still used to denote “the set consisting of.” However, instead
of listing all the members, which is impossible for the infinite set of points in the
triangular budget set B, the set is specified in two parts. To the left of the colon,
(x, y) is used to denote the form of the typical member of B, here a consumption
bundle that is specified by listing the respective quantities of the two goods. To the 30 Chapter 1 / Introduction
right of the colon, the three properties that these typical members must satisfy are all
listed, and the set thereby specified. This is an example of the general specification:
S = (typical member : defining properties}
Note that it is not just infinite sets that can be specified by properties—finite sets
can also be specified in this way. Indeed, even some finite sets almost have to be
specified in this way, such as the set of all human beings currently alive, or even.
(we hope!), the set of all readers of this book.
‘Mathematics makes frequent use of infinite sets. For example, in Section
14, we studied the set of positive integers, which is often denoted by NV, as well
as the set of rational numbers, denoted by Q, and the set of real numbers, denoted
by R. All these sets are infinite.
Set Membership
AS we stated earlier, sets contain members or elements. There is some convenient
standard notation that denotes the relation between a set and its members. First.
xeS
indicates that x is an element of S. Note the special symbol € (which is a variant
of the Greek letter €, or “epsilon”). Occasionally. one sees $5 x being used to
express exactly the same relationship as x € S. The symbol “ 5” is generally read
as “owns,” but is not used very often. To express the fact that x is not a member
of S, we write x ¢ S. For example, d ¢ (a, b, c} says that d is not an element of
the set (a,b,c).
For additional iNustrations of set membership notation, let us return to our ear-
lier examples. Given the budget set B in [1.7], let (x*, y*) denote the consumer's
actual purchases. Then it must be tue that (x*, y*) € B. Confronted with the
choice from the set of feasible main courses F = (fish, pasta, omelette, chicken},
let s denote your actual selection. Then, of course. s € F. This is what we mean
by “feasible set"—it is possible only to choose some member of that set but nothing
outside it.
In the example of choice, from four main courses, it may be argued that if
none is to the customer's Jiking, then she cannot be prevented from ordering nothing
at all from the menu. She can eat somewhere else instead. or simply go hungry. If
that is what she does, she is not really choosing outside her feasible set. Rather, our
previous description of the feasible set should be expanded to include the option of
ordering none of the four available dishes. Thus, the customer's true feasible set is
(fish, pasta, omelette, chicken, none of the previous four}
In the end, she can only avoid choosing something from this by choosing more
than one item. If this is not allowed, then Fs is her true feasible set. Sec. 1.7 / Set Theory 31
Subsets
Let A and B be any two sets. Then A is a subset of B if it is true that every mem-
ber of A is also a member of B. So A is smaller than B in some sense. even though
A and B could actually be equal. This relationship is expressed symbolically by
ACB:
ACB => [xeAsxeB]
A special case of a subset is when A is a proper subser of B, meaning that A.C B
and A #¢ BS
Set Operations
Sets can be combined in many different ways. Especially important are three
operations: union, intersection, and the difference of sets, as shown in Table 1.2.
TABLE 1.2
Notation Name The set consists of
AUB Aunion 8 ‘The elements that belong to at least
ane of the sets A and B.
ANB —— Aintersection 8 The elements that belong to both
Aand B.
A\B — Aminus 8 ‘The elements that belong to A,
but not to B.
Thus,
AUB=(x:xeAorxe B}
ANB=(x:x€Aandx € B}
A\B=(x:x€A andx¢ B}
Example LI
Let A = (1, 2,3, 4,5} and B = (3,6). Find AUB, ANB, A\B, and B\ A.
Solution AUB = (1.2,3,4,5,6), ANB = (3), A\ B= (1.2.4.5),
B\ A= {6}.
SSometimes the notation A CB is reserved for the case when A is a subset of B satisfying
As B. just as a