Microeconomics with Calculus: Tutorial #1

Calculus and Economics

Daniel S. Christiansen
ALBION COLLEGE
[email protected]
August 14, 2002

• Instructions for Viewing

• View Full Screen
• Table of Contents
• Begin tutorial

Copyright
c 1999–2002 Daniel S. Christiansen
 Table of Contents
1 Introduction 4

2 Using Calculus to Learn Economics 7

3 The Notion of a Derivative 10

3.1 The Mathematics of a Derivative . . . . . . . . . . . . 10
3.2 The Geometry of a Derivative . . . . . . . . . . . . . . 13
3.3 The Economics of a Derivative . . . . . . . . . . . . . 18
3.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . 22

4 Derivatives of a Few Simple Functions 24

4.1 Power Functions . . . . . . . . . . . . . . . . . . . . . 24
4.2 Exponential and Logarithmic Functions . . . . . . . . 27
4.3 Section Quiz . . . . . . . . . . . . . . . . . . . . . . . 29

5 Rules of Differentiation 31
5.1 Sums, Differences, and Constants . . . . . . . . . . . . 32
5.2 Product Rule . . . . . . . . . . . . . . . . . . . . . . . 33
Table of Contents (cont.) 3

5.3 Quotient Rule . . . . . . . . . . . . . . . . . . . . . . . 35

5.4 Chain Rule . . . . . . . . . . . . . . . . . . . . . . . . 36
5.5 Inverses . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.6 Section Quiz . . . . . . . . . . . . . . . . . . . . . . . 42

6 Optimization 44
6.1 Maximization and Minimization . . . . . . . . . . . . . 45
6.2 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . 51

7 A Few Words about Integration 52

8 Chapter Quiz 56

Answers to Quizzes 60

Solutions to Exercises 61

References 68
Section 1: Introduction 4

1 Introduction
The idea of this project is to present an intermediate-level course
in microeconomic theory with the help of some simple notions from
calculus. There are two distinguishing features. The first is that each
chapter comes in two versions — one designed for the printed page
(the paper version) and one intended to be read on the computer
screen (the web version). The paper version comes in the form of a
traditional textbook. The web version presents the same material but
has interactive features, for example quizzes that can be corrected on
the spot. Furthermore, all of the material is hyperlinked, so reading
does not have to be done in a linear manner from beginning to end.
This is useful in leaving and then returning to figures, equations,
solutions to exercises, and other points of reference. It is also useful
in linking to other documents or web sites.
The other distinguishing feature is the way calculus is used to de-
velop microeconomic theory. We presume a minimal background in
calculus — just one course. You will surely benefit from having taken
more than one calculus course, and we eventually use some concepts
Section 1: Introduction 5

and techniques from more advanced courses. But when we do so, we

develop what we need with the assumption that you have not seen it
before. What is truly unique about this approach is the order in which
the economics is presented. Since you already know about calculus
with one variable, we first present that portion of microeconomic the-
ory that can be handled with this set of mathematical tools. It turns
out that we can get a lot of mileage out of one-variable calculus and
that we can cover almost all of the basic topics from principles of mi-
croeconomics. At a later point we introduce partial derivatives and
constrained optimization and take up the topic of consumer theory.
It is with consumer theory that almost every other intermediate mi-
croeconomics textbook begins. But this is not a good place to start
using calculus with the background we have assumed.
Thus the prerequisites for following this text are one course in
principles of microeconomics and one course in calculus.
If you should require further background in microeconomics at the
principles level, almost any principles of economics textbook should
provide what you need. The book by Mankiw [4], for example, is
a popular current textbook that provides a good grounding in basic
Section 1: Introduction 6

economics.1
The best background in mathematics is the most rigorous calculus
course you can find — the kind a serious mathematician would like to
offer you. If you are looking to make up deficiencies in your training
or simply looking for review, this advice is not likely to be timely. In
this case any calculus textbook should be able to serve as reference
material for the mathematics used here. One possibility is the book
by Hughes-Hallett [2]. Another good source of help is Donald P.
Story’s e-calculus web site [6], an interactive tutorial for a first course
in calculus.2 Still another possibility is the old but classic Calculus
Made Simple [7] written by Silvanus P. Thompson in 1910.3
1 In particular, see Mankiw’s Supply and Demand I: How Markets Work, Supply

and Demand II: Markets and Welfare, and Firm Behavior and the Organization
of Industry.
2 The e-calculus web site inspired this project, and Story’s exerquiz and web

packages, together with the mathematical typesetting system LATEX, made it pos-
sible.
3 Thompson’s book has the long and suggestive subtitle, Being a Very Simplest

Introduction to those Beautiful Methods of Reckoning Which are Generally Called

by the Terrifying Names of the Differential Calculus and the Integral Calculus.
Section 2: Using Calculus to Learn Economics 7

There are many good textbooks on microeconomic theory at the

intermediate level, among them those by Landsburg [3] and Varian [9].
Neither of these employs calculus except in the appendices. Two
books that use calculus are Binger and Hoffman [1] and Nicholson [5].
You need to be aware that in each of these textbooks the material is
presented in an order different from that of this book.
In this chapter we review the concepts from calculus that are
needed to read the chapters that make up the first part of this project,
Decisions and Markets: Economics with One Variable.

2 Using Calculus to Learn Economics

It is a premise of this book that a calculus approach to the study
of microeconomics pays large dividends. There are several related
reasons why this is so.
First, the use of mathematics helps develop problem-solving skills.
It imposes a rigor that mandates keeping track of costs and benefits,
and it provides a framework for determining which variables and pa-
Section 2: Using Calculus to Learn Economics 8

rameters are important. We apply calculus and other mathematics to

a wide variety of problems. Our view is that you can’t learn economics
without solving problems.
Second, the calculus approach helps in learning to think clearly.
You will be forced to translate weak verbal arguments into precise,
consistent statements. Making a statement formal clarifies what is
being said. When you are called upon to verify what you think is
obvious, you may discover that it is not obvious or perhaps not even
true.
Third, the use of calculus unifies the material by focusing on the
common economic structure of problems. When we strip the specific
details away, many problems look surprisingly alike and have com-
mon solutions. More generally, this is an argument for the power of
abstraction. Calculus makes it easier, not harder, to learn economics.
There are, in fact, only a small number of basic ideas in intermediate
microeconomics; much of what we do is working out the details for
particular situations.
Fourth, calculus helps in becoming literate in the language of mod-
ern economics. For better or worse, mathematics is the language of
Section 2: Using Calculus to Learn Economics 9

economics. If you want to read the economic literature or get an

advanced degree in economics, you will have to deal with the mathe-
matics. The best business schools demand mathematical maturity of
their students; and if you want to undertake a serious study of finance,
for example, you will find that calculus is only the beginning.
Fifth, this approach provides the opportunity to apply mathemat-
ics to social science problems. Whereas many of the examples in a
first course in calculus come from physics, there are also examples
from economics that illustrate most of the same points. This shows
the power of calculus and aids in its understanding.
Sixth, this approach allows us to take the subject to a more ad-
vanced point and thus gives a better appreciation of the power of eco-
nomic theory. Economics is ultimately a policy science, and theory
plays an essential role. Hal Varian [8] has some enlightening answers
to the question, “What use is economic theory?” Among other things,
he suggests that theory is useful as a substitute for data that we need
but do not have; and he explains how theory can generate useful in-
sights in explaining economic phenomena, even when the theory is
only approximately true. We will see examples of these insights as we
Section 3: The Notion of a Derivative 10

proceed.

3 The Notion of a Derivative

The concept of a derivative plays a central role in this book. We
review this notion from the perspective of mathematics, geometry,
and economics.

3.1 The Mathematics of a Derivative

Consider an independent variable x and a dependent variable y, so
that y is a function of x. We choose to write this relationship as
y = y(x). We have to be a bit careful because we are using the
letter y in two different ways; it stands both for the value that y takes
and for the function itself.
We discuss what is meant by the derivative of the function y at a
particular point x1 . For another point x2 close to x1 , let ∆x = x2 −x1
be the change in x and let ∆y = y(x2 ) − y(x1 ) be the corresponding
change in y. Then the derivative of y with respect to x at the
Section 3: The Notion of a Derivative 11

point x1 is defined to be the limit of the ratio of the change in y to

the change in x, i.e.,
y(x1 + ∆x) − y(x1 )
lim .
∆x→0 ∆x
If y is a linear function, y = a + bx, then the ratio of the change
in y to the corresponding change in x is always the same and is equal
to b. This is true regardless of the size of the change in x or the
beginning point x1 . Thus the derivative of the function y = a + bx is
the constant b.
In principle, the limiting operation above could be used to calcu-
late the derivative of an arbitrary function.4 For our purposes this is
not necessary; instead we presume that this has been done to gener-
ate the derivatives of a few simple functions as given in Section 4 and
some rules of differentiation as given in Section 5. It is important to
observe that the derivative is typically not a constant (as is the case
4 There are some technical issues related to whether or not the limit, and hence

the derivative, exists. We put these aside.

Section 3: The Notion of a Derivative 12

for a linear function) but depends on the value of x at which we are

calculating it.
We need notation for the derivative of y with respect to x. The
two most common ways of writing it are either as dy/dx or as y 0 . We
often want to indicate where the derivative is being evaluated. The
derivative of the function y at the point x1 may be denoted either as

dy 
or y 0 (x1 ).
dx x=x1
We use both the dy/dx and y 0 (x) notation, freely switching from one
to the other.
Since the derivative of y with respect to x is a function of x, we can
also work with its derivative, the second derivative of y with respect
to x. This is written as d2 y/dx2 or as y 00 . As the second derivative is
also a function of x, we can write its value at the point x1 in either of
the following ways:
d2 y 

or y 00 (x1 ).
dx2 x=x1
Section 3: The Notion of a Derivative 13

3.2 The Geometry of a Derivative

We have seen that, for a linear function y = a + bx, the derivative is
equal to the slope b. If both a and b are positive, then Figure 1 shows
the function y on the left and the derivative y 0 or dy/dx on the right.
The first thing to note is that to do this right, two separate graphs
are required. This is because the units of the function and the units
of the derivative are not the same. This is evidenced by the most
common example used in calculus courses: distance y, measured in
feet, depends on time x, measured in seconds. The derivative, velocity,
is measured in feet per second.5 In general the units of dy/dx are those
of y/x (and not those of y). Since the derivative is everywhere equal
to b, a positive constant, its graph appears on the right as a horizontal
line. The second derivative (not shown) is constant at zero.
5 One could, of course, plot distance and time on the same axis, placing 30 feet

at the same point as 30 feet per second. But this leads to nonsensical comparisons
between distance and velocity: is 20 feet per second more or less than 30 feet?
To complicate the matter, the answer changes when we consider speed in feet per
minute.
Section 3: The Notion of a Derivative 14
0
y y
y

b y0
a

x x
Figure 1: (Left) A linear function y = a + bx. (Right) The derivative dy/dx
or y 0 of the same function, which is the constant b.

The geometric interpretation is based on identifying the derivative

of a function with its slope. If the function is linear, it is clear what we
mean by its slope. But what is meant by the slope of any arbitrary
function or curve? Given a point on its graph and another point
nearby, we can take the ratio of the change in y to the change in x.
If we let the two points become closer and closer together, then in
the limit we obtain the slope at the point. Geometrically, this is the
Section 3: The Notion of a Derivative 15
y 0
y
y y0

x x
Figure 2: (Left) A function y(x) with positive and increasing derivative.
(Right) The derivative y 0 (x) of the same function.

tangent to the function. This is the same limiting operation we used

in finding the derivative mathematically.
For various shapes of the function y = y(x), we plot both the
function and its derivative. This has already been done for a linear
function in Figure 1.
Students often confuse the words “positive” and “increasing.” The
linear function y on the left of Figure 1 is positive (i.e., it lies above
Section 3: The Notion of a Derivative 16

the x axis) and increasing (i.e., it slopes up). Its slope, the derivative,
is plotted on the right. The slope is positive, but it is not increasing.
(The slope as a function of x is constant!)
Consider next the function y(x) shown on the left of Figure 2. As x
increases the slope gets larger. Thus y 0 , plotted on the right, must
increase as x increases. Here we can say that the function y is positive
and increasing. The derivative is also positive and increasing. If the
derivative is linear, as we have drawn it, then the second derivative
(not shown) is a positive constant.
Now examine the function y(x) shown on the left of Figure 3.
As x increases the slope gets smaller. The slope, plotted on the right,
must thus decrease as x increases. Again the function y is positive
and increasing. But its derivative is positive and decreasing. If the
derivative is linear, as we have drawn it, then the second derivative
(not shown) is a negative constant.
Finally look at the function y(x) shown on the left of Figure 4.
For the values of x shown, the function is always positive. But it
first increases and then decreases. Thus the derivative, shown on
the right, is positive at first but after a point becomes negative. It is
Section 3: The Notion of a Derivative 17
y 0
y
y

y0
x x
Figure 3: (Left) A function y(x) with positive but decreasing derivative.
(Right) The derivative y 0 (x) of the same function.

always decreasing. If it is a linear function, then the second derivative

must be a negative constant.
The graph on the left of Figure 2 is an example of a convex function
(one for which d2 y/dx2 is positive or zero) and those on the left of
Figures 3 and 4 are concave functions (ones for which d2 y/dx2 is
negative or zero).
Section 3: The Notion of a Derivative 18
y y 0

0
x

y y0
x
Figure 4: (Left) A function y(x) with derivative first positive and then negative.
(Right) The derivative y 0 (x) of the same function.

3.3 The Economics of a Derivative

In economics a derivative typically represents a marginal concept. If
benefits (B) and costs (C) depend on the level of an activity (x), as
in Chapter 2, then the derivative of B with respect to x represents
marginal benefit and the derivative of C with respect to x represents
marginal cost. If revenue depends on the quantity of a good sold, then
the derivative of revenue with respect to quantity represents marginal
Section 3: The Notion of a Derivative 19

revenue. If production depends on labor input, then the derivative

of production with respect to labor input represents the marginal
product of labor. If resource cost depends on labor input, then the
derivative of resource cost with respect to labor input represents the
marginal resource cost of labor. We will see many such examples
throughout the book.
We examine one of these concepts, marginal cost (MC), more
carefully. In a first course in economics, such an expression is usually
defined as MC = ∆C/∆x and represents the cost of producing one
more (or the last) unit. In the formula it is understood that x takes on
only the discrete values 1, 2, 3, etc. This is precisely the point of view
that must be taken, at least in principle, when a good is indivisible.
Supertankers and dams, for example, come only in whole units.
Seeing the formula for marginal cost in the previous paragraph,
students sometime wonder why they must divide by the change in x.
The reason is to obtain the extra cost per unit of x. If $10 of extra
cost is incurred in producing two more bushels of output, then the
marginal cost is $5 per bushel.
We will assume that the variable x is continuous, i.e., that it can
Section 3: The Notion of a Derivative 20

take on any real value (or at least any nonnegative value) and not just
1, 2, 3, etc. This is appropriate when we want to measure, in pounds,
the amount of cheese produced in a day. Although commercial scales
may not be capable of measuring cheese to the one-millionth of a
pound, we are not misled by presuming that we could do so. Even
when we deal with goods such as automobiles that are less homoge-
neous than cheese, we can still make sense of requiring a production
of 61.5 automobiles per day if we interpret that to mean 123 automo-
biles over a two day production run. The advantage of working with
a continuous variable is that it greatly simplifies the mathematics.
For a continuous variable x the appropriate definition of marginal
cost is as the derivative of the cost function: MC = dC/dx or MC =
C 0 (x). An advantage of using the C 0 (x) notation is that it emphasizes
that marginal cost is itself a function of x.
In a continuous setting we often speak loosely of marginal cost
representing the extra cost incurred by increasing x by one unit or
the cost saved by decreasing it by one unit. We should remember
that, strictly speaking, the kinds of increments we are dealing with
are infinitesimal ones. But the units are still those of rates: dollars
Section 3: The Notion of a Derivative 21

per bushel, dollars per pound, dollars per automobile, etc.

We reexamine the graphs in Figures 1-4 from a total cost and
marginal cost perspective. We consider whether the figure on the left
might represent a cost function. For concreteness, let us suppose that
we are examining the cost of producing cereal. Then x is measured
in boxes of cereal per month and y is measured in dollars per month.
Marginal cost, pictured on the right, is then measured in dollars per
box of cereal.
The graph on the left of Figure 1 represents a linear cost function
where the cost goes up by the same amount for each extra box of
cereal produced. This is another way of saying that marginal cost
is constant. In these examples there is some fixed cost as well. The
graph on the left of Figure 2 represents a cost function where the
marginal cost gets larger as more boxes are produced. The graph on
the left of Figure 3 shows a situation where marginal cost falls as more
boxes are produced.
The graph on the left of Figure 4 is not a good candidate for a
cost function. The problem is that, after a point, cost goes down as
more is produced. While it might be appropriate for the marginal
Section 3: The Notion of a Derivative 22

cost of producing cereal to fall in some situations, it is not reasonable

to think that total cost would go down as more boxes of cereal are
produced. This would mean that marginal cost, the cost of producing
an extra box of cereal, is negative.

3.4 Exercises

Exercise 3.1. If the amount of taxes paid (T ) depends on income (x),

how would you use calculus notation to describe the marginal tax rate?
If taxes and income are both measured in dollars per year, what are
the units of the marginal tax rate?

Exercise 3.2. Consider a function y(x) as shown in Figure 5. Draw

a graph to show the derivative y 0 (x). Which of the words “increas-
ing,” “decreasing,” “positive,” and “negative” apply to the function y?
Which of these terms apply to the derivative y 0 ? Which of these terms
apply to the second derivative y 00 ?
Section 3: The Notion of a Derivative 23
y
y

x
Figure 5: See Exercises 3.2 and 3.3.

Exercise 3.3. If the function plotted in Figure 5 is a cost function,

what can you say about the shape of marginal cost?
Exercise 3.4. If the derivative y 0 of a function y is a straight line,
can you be sure that the function y is also a straight line? Explain.
Exercise 3.5. Geometrically, the first derivative of a function at a
point measures its slope. What does the second derivative measure?
Section 4: Derivatives of a Few Simple Functions 24

4 Derivatives of a Few Simple Functions

We can accomplish most of our goals by working with derivatives of
only a few functions. The process of taking a derivative is called
differentiation. There are three kinds of functions that we need
to be able to differentiate: power functions such as x2 , exponential
functions such as ex , and logarithmic functions such as ln x.

4.1 Power Functions

A power function is a variable raised to a constant power: y = xa
where a is a constant. From the definition of a derivative, it can be
shown that the derivative of a power function is given by
dxa
= axa−1 . (1)
dx

The formula given in equation (1) covers more functions than you
might initially think. First, the formula works when a is a positive
Section 4: Derivatives of a Few Simple Functions 25

integer. Thus
dx2 dx3 dx10
= 2x, = 3x2 , and = 10x9 .
dx dx dx
For the special case a = 1, the derivative of x is the constant 1
(since x0 = 1).
The formula is also valid when a is a negative integer. You should
make sure you understand what is meant by a variable x raised to
a negative power: it is the same as one divided by x raised to the
corresponding positive power. Thus
1 1 1
x−2 = 2 x−3 = 3 and x−10 = 10 .
x x x
If we apply the formula from equation (1) to the three functions above
we find that
dx−2 dx−3 dx−10
= −2x−3 , = −3x−4 , and = −10x−11 .
dx dx dx
Given the meaning of negative exponents, we can write these as




d 1/x2 2 d 1/x3 3 d 1/x10 10
= − 3, = − 4, and = − 11 .
dx x dx x dx x
Section 4: Derivatives of a Few Simple Functions 26

The case a = −1 is worth special attention:

dx−1
= −x−2 .
dx
This is the same as writing
d (1/x) 1
= − 2.
dx x
The same formula works, in fact, for any real number a. For
example, it works for a = 21 , a = 13 , and a = − 12 . The meaning of
xm/n , when m √and n are integers, is the√ nth root of xm . Thus x1/2
is √
the same as x, x 1/3
is the same as 3 x, and x−1/2 is the same as
1/ x.
Once again we can use equation (1) to calculate derivatives:
√
dx1/2 1 d x 1
= x−1/2 or = √
dx 2 dx 2 x
and √
dx1/3 1 d3x 1
= x−2/3 or = √
3
.
dx 3 dx 3 x2
Section 4: Derivatives of a Few Simple Functions 27

Similarly,
√
dx−1/2 1 d (1/ x) 1
= − x−3/2 or =− √
2
.
dx 2 dx 2 x3

4.2 Exponential and Logarithmic Functions

In a first course in calculus, you discussed the constant e = 2.718 . . ..
Two functions are built from from this fundamental mathematical
constant. The first is the exponential function y = ex . Generalizations
of this simple function are important in economics because they allow
for constant proportional rates of growth. Thus they appear naturally
when dealing with interest rates and other growth rates over time.
The derivative of ex is especially easy to remember as it is a func-
tion whose derivative is the function you start with:
dex
= ex . (2)
dx
Section 4: Derivatives of a Few Simple Functions 28
y
y = ex

y = ln x

x
Figure 6: The exponential function y = ex and the logarithmic function y =
ln x. The two functions are inverses of one another. (As such each is the
reflection of the other around the 45◦ line, the dotted line in the graph.)

The second function is the logarithm to the base e of the variable x,

written as y = ln x. The derivative of the logarithmic function y = ln x
is
d ln x 1
= . (3)
dx x
Section 4: Derivatives of a Few Simple Functions 29

The two functions ex and ln x, plotted in Figure 6, are closely

related as they are inverses of one another. This means, for example,
that if y = ln x, then x = ey . Geometrically it means that one function
is transformed into the other by rotating about the 45 degree line.
Note that ex is equal to one when x = 0 and is well-defined for
negative values of x, where it is less than one but greater than zero.
The function ln x is zero when x = 1 and is negative when x is between
zero and one. For zero and negative real numbers, however, ln x is
not defined. Note also that the exponential function is convex while
the logarithmic function is concave.

4.3 Section Quiz

Click on the “Begin Quiz” button to begin, choose your answers, click
on the “End Quiz” button to see how many questions you answered
correctly, and click on the “Correct” button to see the correct answers.

1. If y = x4 then dy/dx is equal to

Section 4: Derivatives of a Few Simple Functions 30

4x4 4x3 3x4 3x3

2. If y = x−4 then dy/dx is equal to
−4x−3 −4x−4 −4x−5 −4x
3/2
3. If y = x then dy/dx is equal to
3 1/2 3 −1/2 3 2/3 3 −2/3
2x 2x 2x 2x
x
4. If y = e then dy/dx is equal to
x e ex ln x
5. If y = ln x then dy/dx is equal to
√
ex x 1/x x
−6
6. If y = x then dy/dx is equal to
−7
−6x −6x−5 −x−7 −x−5
√
7. If y = x then dy/dx is equal to
√ √ √ √
x x/2 2 x 1/ (2 x)
8. If y = x.8 then dy/dx is equal to
.8x.2 .8x−.2 .8x−.8 .2x.8
Section 5: Rules of Differentiation 31

9. If y = x−1.6 then dy/dx is equal to

1.6x−2.6 1.6x−.6 −1.6x−2.6 −1.6x−.6
10. If y = 1/x then dy/dx is equal to
1/x 1/x2 −1/x −1/x2

5 Rules of Differentiation
New functions can be made out of the simple functions introduced in
Section 4 by combining them in obvious ways. For example, we might
add or multiply two functions together. We need to discuss how to
differentiate these new functions.
Section 5: Rules of Differentiation 32

5.1 Sums, Differences, and Constants

Suppose we have two functions, y1 (x) and y2 (x). We can construct
new functions by adding the two together and by subtracting one from
the other. The derivative of the sum of the two functions is the sum of
the derivatives. Thus the derivative of y1 (x) + y2 (x) is y10 (x) + y20 (x).
For example, the derivative of x2 + x3 is 2x + 3x2 . Likewise, the
derivative of the difference of the two functions is the difference of the
derivatives. Thus the derivative of y1 (x) − y2 (x) is y10 (x) − y20 (x). For
example, the derivative of x2 − x3 is 2x − 3x2 .
Constant functions are particularly easy to work with. The deriva-
tive of a constant is zero. But the derivative of a constant times a
function is the constant times the derivative of the function. Thus
if we have a function y = y(x) and a constant c, the function c has
derivative zero but the function cy(x) has derivative cy 0 (x). From the
rule for sums given above, the derivative of y(x)+c is just y 0 (x). Thus
the derivative of the function 5 is zero, the derivative of 5x2 is 10x,
and the derivative of 5x2 + 5 is 10x.
Example 5.1. We calculate the first and second derivatives of the
Section 5: Rules of Differentiation 33

function y(x) = x3 + 3x2 − 6x + 2. The first derivative is y 0 (x) =

3x2 + 6x − 6. The second derivative is the derivative of the first
derivative. Thus y 00 (x) = 6x + 6.
Example 5.2. We calculate the first and second derivatives of the
function y(x) = x−1 +ln x. The first derivative is y 0 (x) = −x−2 +x−1 ,
and the second derivative is y 00 (x) = 2x−3 − x−2 .
As economists we are interested in functions such as those we have
constructed here. For example, we care about sums because we add
the actions of various participants together to get the market result.
We work with differences when we construct profit as revenue minus
cost. We face constant functions when we deal with fixed or overhead
cost. And we see a constant times a variable when we construct
revenue as a price times quantity in cases where price is given.

5.2 Product Rule

The product rule applies to calculating the derivative of a function
obtained by multiplying two functions together when neither one is
Section 5: Rules of Differentiation 34

necessarily a constant. The result is that the derivative of the product

of two functions is the first times the derivative of the second plus the
second times the derivative of the first. Thus if we have two functions
y1 (x) and y2 (x), we calculate the derivative of the product function,
y(x) = y1 (x) · y2 (x), as follows:
y 0 (x) = y1 (x) · y20 (x) + y2 (x) · y10 (x)
or, using the dy/dx notation,
dy dy2 dy1
= y1 · + y2 · .
dx dx dx
A legendary calculus teacher teaches this as the hi-ho rule because
“the derivative of hi-ho equals hi-dee-ho plus ho-dee-hi.” (Here
hi is the function y1 (x), ho is the function y2 (x), and dee refers to
taking the derivative.)
Example 5.3. We calculate the derivative of the function y(x) = xex .
The product rule is appropriate because the two functions x and ex are
multiplied together. Since the derivative of ex is ex and the derivative
of x is 1, we have that y 0 (x) = xex + ex .
Section 5: Rules of Differentiation 35

Example 5.4. As another example we calculate the derivative of

y(x) = x2 · x3 . Using the product rule, y 0 (x) = x2 · 3x2 + x3 · 2x = 5x4 .
(Note that y(x) = x5 and, from our results in Section 4.1, its derivative
must be 5x4 .)

Product functions will be of great interest to us because expen-

diture is the product of price and quantity; and, in some situations,
price is not constant.

5.3 Quotient Rule

The quotient rule applies when one function is divided by another.
Suppose we have two functions y1 (x) and y2 (x). We consider the
“quotient” function y(x) = y1 (x)/y2 (x). (We presume that we are
not dividing by zero.) The derivative of the quotient is given by
y2 (x) · y10 (x) − y1 (x) · y20 (x)
y 0 (x) = 2 .
[y2 (x)]
Section 5: Rules of Differentiation 36

Example 5.5. Let y1 (x) = ln x and y2 (x) = x. We consider the

quotient function y(x) = (ln x)/x. Then
x · (1/x) − ln x 1 − ln x
y 0 (x) = = .
x2 x2
Example 5.6. We calculate the derivative of the function y(x) =
x3 /x2 (which is equal to x and should have derivative 1):
x2 · 3x2 − x3 · 2x
y 0 (x) = = 1.
x4
Average cost, which is total cost divided by quantity, is a quotient.
We will want to differentiate this function early in Chapter 3.

5.4 Chain Rule

The chain rule applies when we have two functions with a connection:
the independent variable of the first function is the dependent variable
of the second.
Suppose we have two functions y = y(w) and w = w(x). Since y
depends on w and w depends on x, ultimately y depends on x. We
Section 5: Rules of Differentiation 37

write this composite function as

y(x) = y(w(x)). (4)
We start with x, first apply the function w, and then apply the func-
tion y to the result. We have to be a bit careful here because we
are using the same letter y to denote the first function of w and the
composite function of x. It should be clear that when we differentiate
with respect to w we are referring to the former function and when
we differentiate with respect to x we are referring to the latter. Do
not confuse the composite function with a product function.
The chain rule says that the derivative of the function in equa-
tion (4) is equal to the product of the derivative of y with respect to w
and the derivative of w with respect to x. We can write this using our
dy/dx notation as
dy dy dw
= · (5)
dx dw dx
0
and using our y (x) notation as
y 0 (x) = y 0 (w(x)) · w0 (x). (6)
Section 5: Rules of Differentiation 38

Equation (5) has the advantage of being easy to remember. It appears

as if the dw terms cancel out. Equation (6) makes it very clear where
each function is being evaluated.

Example 5.7. Consider the function y(x) = (1 + x2 )10 . This can be

written as y = w10 with w = 1 + x2 . Thus
dy dw10 d(1 + x2 )
= · = 10w9 · 2x = 20x(1 + x2 )9 .
dx dw dx
It is much easier to solve the problem this way than to expand the
original function and then differentiate. Notice that we must substi-
tute for w so that the answer is in terms of x alone.
2
Example 5.8. Consider the function ex . This can be written as
y = ew with w = x2 . Thus
dy dew dx2 2
= · = ew · 2x = 2xex .
dx dw dx
Section 5: Rules of Differentiation 39

Example 5.9. Consider the function y = 1/ ln x. Here y = 1/w with

w = ln x. Then
dy d(1/w) d ln x 1 1 1
= · =− 2 · =− .
dx dw dx w x x ln2 x
Applied to Example 5.7 the chain rule says that the derivative of
something to the 10th power is 10 times the something to the 9th
power times the derivative of the something. Applied to Example 5.8
it says that the derivative of e raised to the something is e raised
to the something times the derivative of the something. Applied to
Example 5.9 it says that the derivative of something to the minus
one is minus something to the minus two times the derivative of the
something. The most common mistake students make with the chain
rule is forgetting to multiply by “the derivative of the something.”
We are interested in composite functions in economics so that we
can attribute revenues and costs to their ultimate sources. If, for
example, revenue depends on output and output depends on labor
input, then revenue ultimately depends on labor input. We will need
to be able to differentiate such functions.
Section 5: Rules of Differentiation 40

5.5 Inverses
The function y = y(x) is a rule that leads from any value of inde-
pendent variable x to a unique value for the dependent variable y.
The inverse function x = x(y) is a rule that goes the other direc-
tion: for a given value of y it provides the unique value of x such that
y = y(x). We will restrict ourselves to working with cases where the
inverse function is well-defined.6
The inverse function is quite different from the reciprocal of the
function y(x), which is 1/y(x).7
As long as we avoid dividing by zero, the derivative of the inverse
function is one divided by the derivative of the original function:
dx 1 1
= or x0 (y) = 0 . (7)
dy dy/dx y (x(y))
6 We have to be careful because there may be more than one value of x associ-
√
ated with each value of y. If y = x2 , for example, then x = ± y; i.e., there are
two values of x associated with a given y. Thus x = x(y) is not well defined over
all real numbers. If we look only at positive values of x, then there is no problem.
7 Note that x(y(x)) = x; this involves composition of two function. Alterna-

tively, y(x) times 1/y(x) equals one; this involves the product of two functions.
Section 5: Rules of Differentiation 41

Again the first set of notation is the most suggestive: it appears that
the dx and dy symbols can be manipulated as algebraic quantities.
The second set of notation is much more precise about where deriva-
tives are evaluated.

Example 5.10. If y = 100 − 2x the inverse function is x = 50 − 21 y.

Differentiating the inverse function directly, we find that dx/dy =
− 21 . Alternatively, using equation (7) and dy/dx = −2, we find that
dx/dy = − 21 .
Example 5.11. If we consider only positive values of x and y, then
√
the function y(x) = x2 has inverse x(y) = y. Using equation (7)
dx 1 1 1
= = = √ .
dy dy/dx 2x 2 y
Notice that we substituted for x so that our answer was left in terms
of y, the variable on which dx/dy depends.
Example 5.12. Recall from Section 4.2 that the logarithmic and
exponential functions are inverses of one another so that if y = ln x
Section 5: Rules of Differentiation 42

then x = ey . Hence
dx 1 1
= = = x = ey .
dy dy/dx (1/x)
This is the same result we gave for the derivative of the exponential
function in Section 4.2.
In economics we frequently need to invert demand functions: in-
stead of the quantity that can be sold at various prices, we will often
be interested in the price at which various quantities can be sold.

5.6 Section Quiz

Click on the “Begin Quiz” button to begin, choose your answers, click
on the “End Quiz” button to see how many questions you answered
correctly, and click on the “Correct” button to see the correct answers.

1. If y = x − x2 then dy/dx is equal to

1−x 1 − 2x 2−x 2 − 2x
Section 5: Rules of Differentiation 43

2. If y = x ln x then dy/dx is equal to

1 1
1 + ln x x + ln x x+ 1+
x x
3. If y = ex /x then dy/dx is equal to
xex − ex ex − xex ex − xex xex − ex
2x
e√ e2x x2 x2
4. If y = x/x then dy/dx is equal to
− 32 x−3/2 − 12 x−1/2 − 32 x−1/2 − 12 x−3/2
5. If y = ln x2 then dy/dx is equal to
2/x2 2/x 1/x2 1/x
5
6. If y = (1 + 2x) then dy/dx is equal to
5(1 + x)4 5(1 + 2x)4 10(1 + 2x)4 10(1 + x)4
√
7. If y = 10 + x2 then dy/dx is equal to
x 2 2x x
√ √ √ √
10 + x2 10 + x2 10 + x2 2 10 + x2
−3/2
8. What is the second derivative of x ?
Section 6: Optimization 44
3 15 3 3
− 5/2 7/2 5/2
5x 4x x 4x1/2
2
9. If y = w and w = 3x + 4, the chain rule can be used to calculate
the derivative of
9x2 + 4 x2 (3x + 4) 3x2 + 4 (3x + 4)2
10. If y = 3x + 5, calculate dx/dy.
1/3 3 3x 3y

6 Optimization
Optimization is one of the basic analytical techniques employed in
microeconomics. We will get lots of mileage out of the idea that the
decisions of economic agents (households, firms, social planners, etc.)
are the result of these parties trying to do the best they can subject
to the constraints they face.
Section 6: Optimization 45

6.1 Maximization and Minimization

In this chapter we limit ourselves to a discussion of how to maximize or
minimize a function y(x) of one variable when there are no constraints,
i.e., when x may be chosen freely.8 Except for a few complications,
the idea is to find values of x for which the derivative of y with respect
to x is zero.
Suppose that we wish to find a value x∗ , if such a value exists,
that maximizes y(x). It cannot be the case that y 0 (x∗ ) > 0; if so
the function would be increasing, and a value of x higher than x∗
would give a higher value of y. Likewise, it cannot be the case that
y 0 (x∗ ) < 0; if so the function would be decreasing, and a value of x
lower than x∗ would give a higher value of y. This means that if x∗
maximizes y(x), then y 0 (x∗ ) must equal zero.
Similar reasoning suggests that if x∗ minimizes y(x), then y 0 (x∗ )
must also equal zero.
8 In many economic contexts it does not make sense for the variable x to be

negative; in those cases there is an implicit constraint.

Section 6: Optimization 46
y y

x∗ x x∗ x
∗ ∗
Figure 7: The point x gives a maximum of y(x) on the left; x gives a
minimum of y(x) on the right.

Points x∗ for which y 0 (x∗ ) = 0 are called stationary points.

A necessary condition that x∗ provide a maximum (or a minimum) is
that x∗ be a stationary point.9 We call this the first-order condition
for a maximum or a minimum. The points labeled x∗ in the two
graphs of Figure 7 are stationary points. The function on the left is
9 This might not be the case if x∗ is a boundary point.
Section 6: Optimization 47
y

x∗ x
Figure 8: The function y(x) has an inflection point at x∗ ; y(x) is neither a
maximum nor minimum at this point.

maximized at this point, and the function on the right is minimized.

From a operational point of view, stationary points provide can-
didates for points that optimize the given function.10 Unless we have
more information, we cannot be sure that a particular stationary point
is the one that solves our problem. For example, if our interest is in
10 The only other conceivable candidates are points at which the function is not

differentiable and boundary points resulting from the implicit constraints.

Section 6: Optimization 48

maximizing y(x) and we have found a stationary point x∗ , that point

could be associated with a maximum as in the graph on the left of
Figure 7. But it could also be associated with a minimum as in the
graph on the right of Figure 7 or an inflection point as in Figure 8.
It is also possible that a stationary point x∗ provides only a local
maximum (a maximization of y(x) only over points very close to x∗ )
but not a global maximum (a maximization over all values of x).
We note that there may not be a unique x that maximizes y(x), or
there may be no solution at all.11
There is a way of using calculus to help determine whether a sta-
tionary point is a local maximum, a local minimum, or neither. This
condition involves second derivatives as well as first derivatives and
is called the second-order condition. There are three possibilities:
(a) If y 0 (x∗ ) = 0 and y 00 (x∗ ) < 0, then x∗ gives a local maximum.
(b) If y 0 (x∗ ) = 0 and y 00 (x∗ ) > 0, then x∗ gives a local minimum.
(c) If y 0 (x∗ ) = 0 and y 00 (x∗ ) = 0, then x∗ may provide a local maxi-
mum, a local minimum, or neither. You should be able to tell from
11 For example, there is no value of x that maximizes the function y = x.
Section 6: Optimization 49

the graph on the left of Figure 7 that the second derivative is negative
at x∗ . The second-order condition thus says that the function is max-
imized there. This illustrates possibility (a). At x∗ in the graph on
the right of Figure 7 the second derivative is positive and the function
is minimized there. This illustrates possibility (b). At x∗ in Figure 8
the second derivative is zero, and we have neither a maximum nor a
minimum. This is one illustration of possibility (c). Another is the
function y = x4 which does have a minimum at x = 0 but whose first
and second derivatives are zero there.
We should be very careful to distinguish between necessary and
sufficient conditions for a maximum. If x∗ maximizes y(x), then
y 0 (x∗ ) = 0 and y 00 (x∗ ) ≤ 0. This is a necessary condition for a maxi-
mum. If y 0 (x∗ ) = 0 and y 00 (x∗ ) < 0, then x∗ gives a local maximum.
This is a sufficient condition.12
We should say a few words about how to use these conditions to
find the maximum value when a specific function is given. We need to
12 Similar conditions apply for a minimum if we replace ≤ and < with ≥ and >,
respectively.
Section 6: Optimization 50

first find the stationary points, i.e., points where the first derivative is
zero. This involves solving one equation y 0 (x) = 0 in one unknown (x).
If this is a linear or quadratic equation, then a solution can be easily
obtained. For example, in determining stationary points we may have
to solve the quadratic equation
ax2 + bx + c = 0.
One way to find such values of x is to use the quadratic formula,
given by √
−b ± b2 − 4ac
x= . (8)
2a
Note that there are two solutions, one associated with the plus sign
and one with the minus sign. We need to apply the second-order
condition to each of these values to distinguish which might be the
maximum and which the minimum.
Example 6.1. We find the value of x that maximizes the function
y(x) = 10 − x2 . Since the first derivative is given by y 0 (x) = −2x,
we find the stationary point by solving −2x = 0 or x∗ = 0. The
Section 6: Optimization 51

second derivative is y 00 (x) = −2 < 0 for all x, so x∗ = 0 maximizes

the function.
Example 6.2. We determine a local maximum and a local minimum
for the function y(x) = x3 −12x2 +36x. The first derivative is y 0 (x) =
3x2 − 24x + 36, so the stationary points are solutions to the equation
3x2 − 24x + 36 = 0 or, after simplifying, x2 − 8x + 12 = 0. This is a
quadratic equation with solution given by Equation (8):
√
8 ± 64 − 48 8±4
x= = = 4 ± 2.
2 2
Thus there are two stationary points, x = 2 and x = 6. To determine
whether each corresponds to a maximum or minimum, we need to look
at the second derivative y 00 (x) = 6x − 24. Evaluating this expression
at x = 2 and x = 6, we find that y 00 (2) = −12 < 0 and y 00 (6) = 12 > 0.
Thus x = 2 is a local maximum and x = 6 is a local minimum.

6.2 Exercises
Exercise 6.1. Find the value of x that maximizes y(x) = 2x−x2 −2.
Section 7: A Few Words about Integration 52

Exercise 6.2. Determine a local maximum and a local minimum for

the function y(x) = 13 x3 − 2x2 + 3x.

7 A Few Words about Integration

We do not presume that you have studied the integral or integration
techniques. Occasionally we find it convenient to employ integral
notation and to use one basic fact about integration.
One way we use integral notation is to represent the area under a
curve. If we have a positive function y = y(x), as in Figure 9, then
Z b
y(x) dx (9)
a
gives the area below y(x) and above the horizontal axis from x = a
to x = b. Just as the derivative is defined as a limit, the integral is a
limit of sums of the areas of rectangles as the base becomes arbitrarily
small.13
13 If you read the integral sign as an S for “sum,” then the notation suggests that

the integral is a limit of sums of terms like y(x) multiplied by small changes in x.
Section 7: A Few Words about Integration 53
y
y

Z b
y(x) dx
a

a b x
Figure 9: The integral of a positive function is just the area beneath it.

There are many techniques used to calculate the value of inte-

grals. We will use nothing more than simple geometry, restricting our
attention to calculating areas of rectangles and triangles.
Finding an integral involves the reverse process of finding a deriva-
tive. In physics we differentiate distance to get velocity but then in-
tegrate (or “sum up”) velocity to get distance traveled. In economics
These products are the areas of rectangles which, added together, approximate
the area under the curve.
Section 7: A Few Words about Integration 54
0
y

y0

y(x) − y(0)

x x
Figure 10: The integral of the derivative y 0 is the original function y(x) except
for a constant.

we differentiate total cost to get marginal cost; we need to integrate

marginal cost to get back to total cost. We could say that distance
is an antiderivative of velocity and total cost is an antiderivative of
marginal cost.
Section 7: A Few Words about Integration 55

We need a way of writing this idea that a function is the integral

(or area under) its derivative. We write
Z x
y(x) = y 0 (s) ds + y(0). (10)
0

We make three points about equation (10), which is illustrated

in Figure 10 and which represents one version of the Fundamental
Theorem of Calculus. First, we have to be careful about the notation.
In equation (9) the value of the integral is a number and not a function
of x. In equation (10), however, the integral does not depend on s
but does depend on x.
Second, we are restricting our attention to nonnegative values of
the two variables. Thus we find the area under the positive curve y 0
between zero and some prescribed positive value of x.
Third, since the derivative of a constant is zero, adding a con-
stant to the total function y produces the same marginal function y 0 .
Thus we cannot hope to find the constant in the total function from
Section 8: Chapter Quiz 56

information about the marginal function alone. The y(0) term in

equation (10) provides the appropriate constant.
Example 7.1. Consider the function y 0 (x) = b as represented in the
graph on the right of Figure 1. The area under this curve from 0 to x
is bx (the base x times the height b). Using equation (10) and the fact
that y(0) = a, we get y(x) = a + bx. This is the total function in the
graph on the left of Figure 1.
Example 7.2. Consider the function y 0 (x) = a + bx as represented in
the graph on the right of Figure 2. Consider an arbitrary value of x.
The area under this function from 0 to x is the sum of a rectangle
with base x and height a and a triangle with base x and height bx.
Thus the area is ax + (1/2)bx2 . This is the function y(x) as shown on
the right of Figure 2; when differentiated it gives y 0 (x).

8 Chapter Quiz
Click on the “Begin Quiz” button to begin, choose your answers, click
on the “End Quiz” button to see how many questions you answered
Section 8: Chapter Quiz 57

correctly, and click on the “Correct” button to see the correct answers.

1. If y 0 is positive and decreasing, then y is

decreasing with decreasing slope
decreasing with increasing slope
increasing with decreasing slope
increasing with increasing slope
2. If Revenue R depends on quantity sold Q, how might we write
marginal revenue?
dQ dx
R0 (Q) Q0 (R)
dx dQ
3. If y(x) = −mx−n , what is y 0 (x)?
mnx−n+1 mnx−n−1 −mnx−n+1 −mnx−n+1
√
4. If y(x) = 3x, what is dy/dx?
Section 8: Chapter Quiz 58
3 2 3 1
√ √ √ √
2 3x 3 3x 3x 2 3x
5. If y(x) = (2 + x3 )5 , then y 0 (x) is
5(x3 )4 5(2 + x3 )4 (2 + 3x2 )
5(2 + x3 )4 5(2 + x3 )4 (3x2 )
6. Another way to write x−.75 is as
1 1
x3/4 x4/3
x3/4 x4/3
7. What is the derivative of ln x?
1 1
ln x − ln x −
x x
8. If y(x) = u(x) · v(x), then y 0 (x) is
du dv dv du dv du dv du
· · u −v u +v
dv dx du dx dx dx dx dx
dx
9. If y = a − bx, then equals
dy
1 1
−b − a
b b
Section 8: Chapter Quiz 59

10. The function y(x) = x − x2 achieves its maximum at an x value

of
.5 1 2 5
Answers to Quizzes 60

Answers to Quizzes
Answers to Section Quiz 4.3
1. b 2. c 3. a 4. c 5. c 6. a 7. d 8. b 9. c 10. d

Answers to Section Quiz 5.6

√
1. b 2. a 3. d 4. d (The easiest way to differentiate y = x/x is
to recognize it as y = x1/2 /x1 = x−1/2 so that y 0 (x) = − 12 x−3/2 .) 5.
b 6. c 7. a 8. b 9. d 10. a

Answers to Chapter Quiz 8

1. c 2. c 3. b 4. a (chain rule) 5. d (chain rule) 6. c 7. c
8. d (product rule) 9. b (derivative of an inverse function) 10. a
(stationary point)
Solutions to Exercises 61

Solutions to Exercises
Exercise 3.1.
The marginal tax rate is written as dT /dx or T 0 (x). The units of
the marginal tax rate are the units of T (dollars per year) divided by
the units of x (dollars per year). Thus everything cancels out, and
there are no units! The marginal tax rate is a pure number, such as
30%. Return to Exercise 3.1
Solutions to Exercises 62
0
y
y0

x
Figure 11: The graph of y 0 (x) for Exercise 3.2.

Exercise 3.2.
The graph of the derivative y 0 is shown in Figure 11. The func-
tion y is always positive (it lies above the horizontal axis) and increas-
ing (it slopes up). The function y 0 is positive but it decreases at first
and then increases. The second derivative y 00 (not shown) is negative
at first and then becomes positive. It is always increasing.
Return to Exercise 3.2
Solutions to Exercises 63

Exercise 3.3.
Marginal cost first decreases and then increases. (Marginal cost
is always positive as well. If this were not true, the function y would
not be a reasonable cost function.) Return to Exercise 3.3
Solutions to Exercises 64

Exercise 3.4.
No. See Figures 2 and 3 for cases where this is clearly false. The
function y will be a straight line only for the case (as in Figure 1)
where the derivative is a horizontal line. Return to Exercise 3.4
Solutions to Exercises 65

Exercise 3.5.
The second derivative measures the slope of the slope, i.e., it mea-
sures the curvature of the function at the particular point. If the
second derivative is positive, then the function looks like a convex
function near the point (as in the diagram on the left of Figure 2). If
the second derivative is negative, then the function looks like a con-
cave function near the point (as in the diagram on the left of Figure 3).
Return to Exercise 3.5
Solutions to Exercises 66

Exercise 6.1.
First note that y 0 (x) = 2−2x, so stationary points satisfy 2−2x =
0 or x∗ = 1. The second derivative, y 0 (x) = −2 < 0 everywhere, so
the function is maximized at x∗ = 1. Return to Exercise 6.1
Solutions to Exercises 67

Exercise 6.2.
The first derivative is y 0 (x) = x2 − 4x + 3, so the stationary points
are solutions to the equation x2 − 4x + 3 = 0. This is a quadratic
equation with solution given by
√
4 ± 16 − 12 4±2
x= = = 2 ± 1.
2 2
Thus there are two stationary points, x = 1 and x = 3. To determine
whether each corresponds to a maximum or minimum, we need to look
at the second derivative y 00 (x) = 2x − 4. Evaluating this expression
at x = 1 and x = 3, we find that y 00 (1) = −2 < 0 and y 00 (3) = 2 > 0.
Thus x = 1 is a local maximum and x = 3 is a local minimum.
Return to Exercise 6.2
References 68

References
[1] Binger, Brian R. and E. Hoffman, Microeconomics with Calculus,
2nd ed., Addison-Wesley, Reading, MA, 1998. 7
[2] Hughes-Hallett, Deborah, et. al., Calculus, 2nd ed., John Wiley &
Sons, New York, 1998. 6

[3] Landsburg, Steven E., Price Theory and Applications, 5th ed.,
South-Western, Cincinnati, 2002. 7
[4] Mankiw, N. Gregory, Principles of Economics, 2nd ed., The Dry-
den Press, Fort Worth, 2001. 5
[5] Nicholson, Walter, Microeconomic Theory: Basic Principles and
Extensions, 7th ed., The Dryden Press, Fort Worth, 1998. 7
[6] Story, Donald P., e-calculus, April 3, 2000, http://www.math.
uakron.edu/~dpstory/e-calculus.html. Available: February
20, 2002. 6
References 69

[7] Thompson, Silvanus P., Calculus Made Easy, Macmillan, London,

1910. 6
[8] Varian, Hal R., What Use is Economic Theory? August,
1989, http://www.sims.berkeley.edu/~hal/Papers/theory.
pdf. Available: February 20, 2002. 9
[9] Varian, Hal R., Intermediate Microeconomics: A Modern Ap-
proach, 5th ed., W. W. Norton & Co., New York, 1999. 7