A CLEAR UNDERSTANDING

OF THE INDUSTRY
IS CFA INSTITUTE INVESTMENT FOUNDATIONS THE BIG PICTURE
RIGHT FOR YOU?
Investment Foundations is a comprehensive global education
Investment Foundations is a certificate program designed to give certificate program that provides a clear understanding of investment
you a clear understanding of the investment management industry. industry essentials. The certificate program is designed for all
Whether youre just starting a career in financial services or want to professional disciplines outside of investment roles, including IT,
learn essential industry concepts, Investment Foundations offers an operations, accounting, administration, and marketing. There is no
accessible solution for breaking through the complexities of the global education or experience requirement and the exam can be taken at
investment industry and raising your professional profile. your convenience at available test centers around the world.

WHAT YOU LEARN

HOW WILL YOU BENEFIT [CFA Institute Investment
The certificate program covers Foundations] is very relevant to
the essentials of the investment Clarity the current market and I can study
management industry: Benefit from having a common on the move. The program cleared
understanding of industry
up a lot of concepts for me and
Module 1: structure and terminology,
now I am much more comfortable
speaking with clients about what
Industry regardless of your job function
is happening in the market.
Overview or geographic location.
MITALI BHANDARE
Module 2: MORNINGSTAR,
Ethics and Collaboration INVESTMENT FOUNDATIONS
Regulation Work more effectively CERTIFICATE HOLDER
with global colleagues by
Module 3: understanding industry
Inputs and functions, building stronger
Tools relationships and raising your
professional competence. The main benefit of [CFA Institute
Module 4: Investment Foundations] was an
Investment ability to see a bigger picture of
Instruments Confidence the finance industry and the role
Gain the knowledge to identify
of our business within it.
Module 5: issues and the confidence to ALEXANDER TARASOV
Industry speak up. Get a better sense of CITCO FUND SERVICES,
Structure your role and how you connect INVESTMENT FOUNDATIONS
with the complex global industry CERTIFICATE HOLDER
Module 6: at large.
Serving Client
Needs

Module 7: HOW TO FIND OUT MORE

Industry Go to: cfa.is/InvFound
Controls

www.cfainstitute.org

2016 CFA Institute. All rights reserved.

CHAPTER 8
QUANTITATIVE CONCEPTS
by Michael J. Buckle,PhD, James Seaton,PhD, and Stephen Thomas,PhD
LEARNING OUTCOMES

After completing this chapter, you should be able to do the following:

a Define the concept of interest;

b Compare simple and compound interest;

c Define present value, future value, and discount rate;

d Describe how time and discount rate affect present and future values;

e Explain the relevance of net present value in valuing financial

investments;

f Describe applications of time value of money;

g Explain uses of mean, median, and mode, which are measures of fre-
quency or central tendency;

h Explain uses of range, percentile, standard deviation, and variance, which

are measures of dispersion;

i Describe and interpret the characteristics of a normal distribution;

j Describe and interpret correlation.

Time Value of Money 177

INTRODUCTION 1
Knowledge of quantitative (mathematically based) concepts is extremely important to
understanding the world of finance and investing. Quantitative concepts play a role
in financial decisions, such as saving and borrowing, and also form the foundation
for valuing investment opportunities and assessing their risks. The time value of
money and descriptive statistics are two important quantitative concepts. They are
not directly related to each other, but we combine them in this chapter because they
are key quantitative concepts used in finance and investment.

The time value of money is useful in many walks of life: it helps savers to know how
long it will take them to afford a certain item and how much they will have to put
aside each week or month, it helps investors to assess whether an investment should
provide a satisfactory return, and it helps companies to determine whether the profit
from investing will exceed the cost.

Statistics are also used in a wide range of business and personal contexts. As you
attempt to assess the large amount of personal and work-related data that are part of
our everyday lives, you will probably realise that an efficient summary and description
of data is helpful to make sense of it. Most people, for instance, look at summaries of
weather information to make decisions about how to dress and whether to carry an
umbrella or bring rain gear. Summary statistics help you understand and use informa-
tion in making decisions, including financial decisions. For example, summary infor-
mation about a companys or markets performance can help in investment decisions.

In short, quantitative concepts are fundamental to the investment industry. For any-
one working in the industry, familiarity with the concepts described in this chapter is
critical. As always, you are not responsible for calculations, but the presentation
of formulae and illustrative calculations may enhance your understanding.

TIME VALUE OF MONEY 2

Valuing cash flows, which occur over different periods, is an important issue in finance.
You may be concerned with how much money you will have in the future (the future
value) as a result of saving or investing over time. You may want to know how much
you should save in a certain amount of time to accumulate a specified amount in the
future. You may want to know what your expected return is on an investment with
specified cash flows at different points in time. These types of problems occur every
day in investments (e.g., in buying a bond), personal finance (e.g., in arranging an
automobile loan or a mortgage), and corporate finance (e.g., in evaluating whether
to build a factory). These problems are known as time value of money problems
because their solutions reflect the principle that the timing of a cash flow affects the
cash flows value.

2014 CFA Institute. All rights reserved.

178 Chapter 8 Quantitative Concepts

2.1 Interest
Borrowing and lending are transactions with cash flow consequences. Someone who
needs money borrows it from someone who does not need it in the present (a saver)
and is willing to lend it. In the present, the borrower has money and the lender has
given up money. In the future, the borrower will give up money to pay back the lender;
the lender will receive money as repayment from the borrower in the form of interest,
as shown below. The lender will also receive back the money lent to the borrower. The
money originally borrowed, which interest is calculated on, is called the principal.
Interest can be defined as payment for the use of borrowed money.

Lends Money

Pays Interest

Lender Borrower

Interest is all about timing: someone needs money now while someone else is willing
and able to give up money now, but at a price. The borrower pays a price for not being
able to wait to have money and to compensate the lender for giving up potential current
consumption or other investment opportunities; that price is interest. Interest is paid
by a borrower and earned by the lender to compensate the lender for opportunity cost
and risk. Opportunity cost, in general, is the value of alternative opportunities that
have been given up by the lender, including lending to others, investing elsewhere,
or simply spending the money. Opportunity cost can also be seen as compensation
for deferring consumption. Lending delays consumption by the term of the loan (the
time over which the loan is repaid). The longer the consumption is deferred, the more
compensation (higher interest) the lender will demand.

The lender also bears risks, such as the risk of not getting the money back if the
borrower defaults (fails to make a promised payment). The riskier the borrower or
the less certain the borrowers ability to repay the loan, the higher the level of inter-
est demanded by the lender. Another risk is that as a result of inflation (an increase
in prices of goods and services), the money received may not be worth as much as
expected. In other words, a lenders purchasing power may decline even if the money is
repaid as promised. The greater the expected inflation, the higher the level of interest
demanded by the lender.

From the borrowers perspective, interest is the cost of having access to money that
they would not otherwise have. An interest rate is determined by two factors: oppor-
tunity cost and risk. Even if a loan is viewed as riskless (zero likelihood of default),
there still has to be compensation for the lenders opportunity cost and for expected
inflation. Exhibit1 shows examples of borrowers and lenders.
Time Value of Money 179

Exhibit1 Examples of Borrowers and Lenders

Borrowers and lenders can be people, companies, financial

institutions, and so on. Here are some examples of borrowers and
lenders that you may be familiar with.

Borrow Money If you get money from the bank

and pay interest, you are the
borrower and the bank is the
Pay Interest lender.

If you put money into a savings

Save Money account and earn interest, you
are a saver who lends your
savings to the bank and the bank
Receive Interest
is the borrower.

If people invest in a
Invest Money company and earn interest
by buying bonds, they are
the lenders and the
Receive Interest
company is the borrower.

2.1.1 Simple Interest

A simple interest rate is the cost to the borrower or the rate of return to the lender,
per period, on the original principal (the amount borrowed). Conventionally, interest
rates are stated as annual rates, so the period is assumed to be one year unless stated
otherwise. The cost or return is stated as a percentage rate of the original principal so
the rates can then be compared, regardless of the amount of principal they apply to.
For example, a loan with a 5% interest rate is more expensive to the borrower than a
loan with a 3% interest rate. Similarly, a loan with a 5% interest rate provides a higher
promised return to the lender than a loan with a 3% interest rate.

The actual amount of interest earned or paid depends on the simple interest rate, the
amount of principal lent or borrowed, and the number of periods over which it is lent
or borrowed. We can show this mathematically as follows:
Simple interest = Simple interest rate Principal Number of periods

If you put money in a bank account and the bank offers a simple interest rate of 10%
per annum (or annually), then for every 100 you put in, you (as a lender to the bank)
will receive 10 in the course of the year (assume at year end to simplify calculations):
Interest = 0.10 100 1= 10

If your money is left in the bank for two years, the interest paid will be 20:
Interest = 0.10 100 2= 20
180 Chapter 8 Quantitative Concepts

Simple interest is not reinvested and is applied only to the original principal, as shown
in Exhibit2.

Exhibit2 Simple Interest of 10% on 100 Original Principal

160
Interest per Year
150
10
140
10
130
Pounds ()

10
120
10 End of Previous
110 Year Balance
10 (Principal + Interest)
100

...
Original Principal
0
Original Year 1 Year 2 Year 3 Year 4 Year 5
Principal

If the interest earned is added to the original principal, the relationship between the
original principal and its future value with simple interest can be described as follows:

Future value = Original principal [1+ (Simple interest rate

Numberofperiods)]

To extend our deposit example: 100 [1+ (0.10 2)] = 100 (1.20) = 120. The
value at the end of two years is 120.

2.1.2 Compound Interest

Interest compounds when it is added to the original principal. Compound interest
is often referred to as interest on interest. As opposed to simple interest, interest
is assumed to be reinvested so future interest is earned on principal and reinvested
interest, not just on the original principal.

If a deposit of 100 is made and earns 10% and the money is reinvested (remains
on deposit), then additional interest is earned in the course of the second year on
the 10 of interest earned in the first year. The interest is being compounded. Total
interest after two years will now be 21; 10 (= 100 0.10) for the first year, plus
11 (= 110 0.10) for the second year. The second years interest is calculated on the
original 100 principal plus the first years interest of 10. As shown in Exhibit3, the
total interest after two years is 21 rather than 20 as in the case of simple interest
shown in Exhibit2.
Time Value of Money 181

Exhibit3 Compound Interest of 10% on 100 Original Principal

160
Interest per Year 14.64
150

140 13.31
130
Pounds ()

12.10
120
11.00 End of Previous
110 Year Balance
10.00 (Principal + Interest)
100

... Original Principal

0
Original Year 1 Year 2 Year 3 Year 4 Year 5
Principal

The relationship between the original principal and its future value when interest is
compounded can be described as follows:
Future value = Original principal (1+ Simple interest rate)Number of periods

In the deposit example, 100 (1+ 0.10)2 = 100 (1.10)2 = 121. With compounding,
the value at the end of two years is 121.

2.1.3 Comparing Simple Interest and Compound Interest

Compound interest is extremely powerful for savers; reinvesting the interest earned
on investments is a way of growing savings. Somebody who invests 100 at 10% for
two years will end up with 1 more by reinvesting the interest (121) than with simple
interest (120). This amount may not look very impressive, but over a longer time
period, say 20 years, 100 invested at 10% for 20 years becomes 300 with simple
interest {100 [1+ (0.10 20)] = 100 3= 300} but 673 with compound interest
[100 (1+ 0.10)20 = 100 (1.10)20 = 673]. This concept is illustrated in Exhibit4.
182 Chapter 8 Quantitative Concepts

Exhibit4 Effects on Savings of Simple and Compound Interest

700

600

500

Balance ()
400

300

200

100

0
0 5 10 15 20
Years
Simple Interest Compound Interest

2.1.4 Annual Percentage Rate and Effective Annual Rate

Unless stated otherwise, interest rates are stated as annual rates. The rate quoted is
often the annual percentage rate (APR), which is a simple interest rate that does not
involve compounding. Another widely used rate is the effective annual rate (EAR).
This rate involves annualising, through compounding, a rate that is paid more than
once a yearusually monthly, quarterly, or semi-annually. The following equation
shows how to determine the EAR given the APR.

Number of periods per year

APR
EAR = 1 + 1
Number of periods per year

Example 1 shows a few types of financial products and their simple interest rates
(APRs) and their compound rates (EARs).

EXAMPLE1. SIMPLE AND COMPOUND INTEREST RATES

A credit card charges interest at an APR of 15.24%, compounded daily. A bank

pays 0.2% monthly on the average amount on deposit over the month. A loan
is made with a 6.0% annual rate, compounded quarterly. The following table
shows what the expected annual rate is for each of these situations. The rate is
higher than the APR because of compounding.
Time Value of Money 183

Simple Interest Rate Compound Interest Rate

or APR or EAR

365
0.1524
Credit card 15.24% 16.46% = 1 + 1
365

12
0.024
Bank deposit 2.4% (= 0.2% 12) 2.43% = 1 + 1
12

4
0.06
Loan 6.0% 6.14% = 1 + 1
4

As can be seen in Example1, in general, whenever an interest rate compounds more

often than annually, the EAR is greater than the APR. In other words, more frequent
compounding leads to a higher EAR.

2.2 Present Value and Future Value

Two basic time value of money problems are finding the value of a set of cash flows
now (present value) and the value as of a point of time in the future (future value).

2.2.1 Present Value and Future Value

If you are offered 1 today or 1 in a years time, which would you choose? Most
people say 1 today because it gives them the choice of whether to spend or invest
the money today and avoid the risk of never getting it at all. The 1 to be received in
the future is worth less than 1 received today. The 1 to be received in the future is
today worth 1 minus the opportunity cost and the risk of being without it for one
year. The present value is obtained by discounting the future cash flow by the interest
rate. The rate of interest in this context can be called the discount rate.

Present Interest Rate (r%) Future

Value Value

Present Future
Time Time

Discount Rate (r%)

Time affects the value of money because delay creates opportunity costs and risk. If
you earn a return of r% for waiting one year, 1 (1+ r%) is the future value after
one year of 1 invested today. Put another way, 1 is the present value of 1 (1+
r%) received in a years time.
184 Chapter 8 Quantitative Concepts

A saver may want to know how much money is needed today to produce a certain
sum in the future given the rate of interest, r. In the example in Exhibit3, todays value
is 100 and the interest rate is 10%, so the future value after two years is 100 (1+
0.10)2 = 121. The present valuethe equivalent value todayof 121 in two years,
given that the annual interest rate is 10%, is 100.

100 121
Present Interest Rate (10%) Future
Value Value

Today In 2 Years

100 121
Present Discount Rate (10%) Future
Value Value

Before you can calculate present or future values, you must know the appropriate
interest or discount rates to use. The rate will usually depend on the overall level of
interest rates in the economy, the opportunity cost, and the riskiness of the invest-
ments under consideration. The following equations generalise the calculation of
future and present values:

Future value = Present value (1+ Interest rate)Number of periods

Future value
Present value =
(1 + Discount rate) Number of periods
Note that the interest and discount rates are the same percentage rates, but the termi-
nology varies based on context. Calculating present values allows investors and analysts
to translate cash flows of different amounts and at different points in the future into
sums in the present that can be compared with each other. Likewise, the cash flows can
be translated into the values they would be equivalent to at a common future point.

Example2 compares two investments with the same initial outflow (investment) but
with different future cash inflows at different points in time.

EXAMPLE2. COMPARING INVESTMENTS

1 You are choosing between two investments of equal risk. You believe
that given the risk, the appropriate discount rate to use is 9%. Your initial
investment (outflow) for each is 500. One investment is expected to pay
out 1,000 three years from now; the other investment is expected to pay
out 1,350 five years from now. To choose between the two investments,
you must compare the value of each investment at the same point in time.

Present value of 1,000 in three years discounted at 9%

1, 000 1, 000
= = = 772.18
(1.09)3 1.295
Time Value of Money 185

Present value of 1,350 in five years discounted at 9%

1,350 1,350
= = = 877.41
(1.09)5 1.5386
As you can see, the investment with a payout of 1,350 five years from
now is worth more in present value terms, so it is the better investment.

2 You are choosing between the same two investments but you have reas-
sessed their risks. You now consider the five-year investment to be more
risky than the first and estimate that a 15% return is required to justify
making this investment.

Present value of 1,350 in five years discounted at 15%

1,350 1,350
= = = 671.19
(1.15)5 2.0114
The investment paying 1,000 in three years (discounted at 9%) is, in this
case, preferable to the investment paying 1,350 in five years (discounted
at 15%) in present value terms. Its present value of 771.28 is higher than
the present value of 671.19 on the five-year investment.

Example2 shows three elements that must be considered when comparing investments:

the cash flows each investment will generate in the future,

the timing of these cash flows, and

the risk associated with each investment, which is reflected in the discount rate.

Present value considers the joint effect of these three elements and provides an effec-
tive way of comparing investments with different risks that have different future cash
flows at different points in time.

2.2.2 Net Present Value

Present value is appropriate for comparing investments when the initial outflow for
each investment is the same, as in Example2. But investments may not have the same
initial cash outflow, and outflows may occur at times other than time zero (the time of
the initial outflow). The net present value (NPV) of an investment is the present value
of future cash flows or returns minus the present value of the cost of the investment
(which often, but not always, occurs only in the initial period). Using NPV rather than
present value to evaluate investments is especially important when the investments
have different initial costs. Example3 below illustrates this.

EXAMPLE3. COMPARING INVESTMENTS USING NET PRESENT VALUE

The NPV of the investment in Example 2 that is paying 1,350 in five years
(discounted at 15%) if it initially cost 500 is:
186 Chapter 8 Quantitative Concepts

671.19 500.00= 171.19

The NPV of the investment paying 1,000 in three years discounted at 9%
if it initially cost 700 is:
772.18 700= 72.18.
This amount is less than 171.19, making the investment paying 1,350 in
five years discounted at 15% worth more in present value terms. This conclusion
differs from that reached when present value only was used.

If costs were to occur at times different from time zero, then they would also be dis-
counted back to time zero for the purposes of comparison and calculation of the NPV.
If the NPV is zero or greater, the investment is earning at least the discount rate. An
NPV of less than zero indicates that the investment should not be made.

Calculating the NPV allows an investor to compare different investments using their
projected cash flows and costs. The concepts of present value and net present value
have widespread applications in the valuation of financial assets and products. For
example, equities may pay dividends and/or be sold in the future, bonds may pay
interest and principal in the future, and insurance may lead to future payouts.

Estimating values by using cash flows is also important to companies considering a

range of investment opportunities. For example, should the sales team be supplied
with tablets or laptops, or should the company open a new office in Asia or carry on
visiting from the companys European headquarters? In order to choose, decision
makers estimate the expected future cash flows of the alternatives available. The
decision makers then discount the estimated cash flows by an appropriate discount
rate that reflects the riskiness of these cash flows. They work out the discounted cash
flows for each opportunity to estimate the value of the cash flows at the current time
(the present value) and to arrive at the net present value. They then compare the net
present values of all the opportunities and choose the opportunity or combination of
opportunities with the largest positive net present value.

2.2.3 Application of the Time Value of Money

The time value of money concept can help to solve many common financial problems.
If you save in a deposit account, it can tell you by how much your money will grow
over a given number of years. Time value of money problems can involve both posi-
tive cash flows (inflows or savings) and negative cash flows (outflows or withdrawals).
Example4 illustrates, with two different sets of facts, how cash inflows and outflows
affect future value.

EXAMPLE4. FUTURE VALUE

1 You place 1,000 on deposit at an annual interest rate of 10% and make
regular contributions of 250 at the end of each of the next two years.
How much do you have in your account at the end of two years?
Time Value of Money 187

The initial 1,000 becomes 1,000 (1+ 0.10)2 = 1,210

The first annual 250 payment becomes 250 (1+ 0.10) = 275
The second annual payment is received at the end and earns
no interest = 250
The total future value = 1,735

2 You place 1,000 on deposit and withdraw 250 at the end of the first
year. The balance on deposit at the beginning of the year earns an annual
interest rate of 10%. How much do you have in your account at the end of
two years?

At the end of the first year, you have 1,000 (1+ 0.10) = 1,100
You withdraw 250 and begin the second year with an amount = 850
At the end of the second year, you have 850 (1+ 0.10) = 935

Time value of money can also help determine the value of a financial instrument. It
can help you work out the value of an annuity or how long it will take to pay off the
mortgage on your home.

2.2.3.1 Present Value and the Valuation of Financial Instruments People invest
in financial products and instruments because they expect to get future benefits in
the form of future cash flows. These cash flows can be in the form of income, such
as dividends and interest, from the repayment of an amount lent, or from selling the
financial product or instrument to someone else. An investor is exchanging a sum of
money today for future cash flows, and some of these cash flows are more uncertain
than others. The value (amount exchanged) today of a financial product should equal
the value of its expected future cash flows. This concept is shown in Example5.

EXAMPLE5. VALUE OF A LOAN

Consider the example of a simple loan that was made three years ago. Two years
from today, the loan will mature and the borrower should repay the principal
value of the loan, which is 100. The investor who buys (or owns) this loan should
also receive from the borrower two annual interest payments at the originally
promised interest rate of 8%. The interest payments will be 8 (= 8% 100),
with the first interest payment received a year from now and the second two
years from now.

How much would an investor pay today to secure these two years of cash flow
if the appropriate discount rate is 10% (i.e. r = 0.10)? Note that the rate used for
discounting the future cash flows should reflect the risk of the investment and
interest rates in the market. In practice, it is unlikely that the discount rate will
be equal to the loans originally promised interest rate because the risk of the
investment and interest rates in the market may change over time.
8
The first years interest payment is worth = 7.27.
1.101
188 Chapter 8 Quantitative Concepts

8
The second years interest payment is worth = 6.61.
1.102
The repayment of the loans principal value in two years is
100
worth = 82.64.
1.102
So today, the cash flows returned by the loan are worth 7.27+ 6.61+ 82.64=
96.52. So this loan is worth 96.52 to the investor. In other words, if the original
lender wanted to sell this loan, an investor would pay 96.52.

Through the understanding of present value and knowing how to calculate it, investors
can assess whether the price of a financial instrument trading in the marketplace is
priced cheaply, priced fairly, or overpriced.

2.2.3.2 Time Value of Money and Regular Payments Many kinds of financial arrange-
ments involve regular payments over time. For example, most consumer loans, including
mortgages, involve regular periodic payments to pay off the loan. Each period, some of
the payment covers the interest on the loan and the rest of the payment pays off some
of the principal (the loaned amount). A pension savings scheme or pension plan may
also involve regular contributions.

Most consumer loans result in a final balance of money equal to zero. That is, the
loan is paid off. Two time value of money applications that require the final balance
of money to be zero are annuities and mortgages.

An annuity involves the initial payment of an amount, usually to an insurance company,

in exchange for a fixed number of future payments of a certain amount. Each period,
the insurance company makes payments to the annuity holder; these payments are
equivalent to the annuity holder making withdrawals. These withdrawals can be viewed
as negative cash flows because they reduce the annuity balance. The initial payment
to the insurer is called the value of the annuity and the final value is equal to zero.

A repayment or amortising mortgage involves a loan and a series of fixed payments.

The initial amount of the loan is referred to as the principal. Although the payment
amounts are fixed, the portion of each payment that is interest is based on the remain-
ing principal at the beginning of each period. As some of the principal is repaid each
period, the amount of interest decreases over time, and thus the amount of principal
repaid increases with each successive payment until the value of the principal is
reduced to zero. At this point, the loan is said to mature.

Example6 illustrates the reduction of an annuity to zero over time and the reduction
of a mortgage to zero over time. To simplify the examples, the assumption is that the
annuity and the mortgage each mature in five years and entail a single withdrawal or
payment each of the five years.
Time Value of Money 189

EXAMPLE6. ANNUITY AND MORTGAGE

1 A retired French man pays an insurance company 10,000 in exchange for

a promise by the insurance company to pay him 2,375 at the end of each
of the next four years and 2,370 at the end of the fifth year. The insurance
company is in effect paying him 6.0% interest on the annuity balance.

Withdrawal
Annuity Balance (Payment by
at Beginning Balance at End of Year Insurance
Year of Year before Withdrawal Company)

1 10,000 10,600 (= 10,000 1.06) 2,375

2 8,225 8,719 (= 8,225 1.06) 2,375
3 6,344 6,725 (= 6,344 1.06) 2,375
4 4,350 4,611 (= 4,350 1.06) 2,375
5 2,236 2,370 (= 2,236 1.06) 2,370
6 0

2 You borrow 60,000 to buy a small cottage in the country. The interest
rate on the mortgage is 4.60%. Your payment at the end of each year will
be 13,706.

Mortgage
Outstanding Total
at Beginning Mortgage Principal
Year of Year Payment Interest Paid Reduced

1 60,000 13,706 2,760 (= 60,000 0.046) 10,946

2 49,054 13,706 2,257 (= 49,054 0.046) 11,449
3 37,605 13,706 1,730 (= 37,605 0.046) 11,976
4 25,630 13,706 1,179 (= 25,630 0.046) 12,527
5 13,103 13,706 603 (= 13,103 0.046) 13,103
6 0

As you can see in Example6, both the annuity and mortgage balances decline to zero
over time.
190 Chapter 8 Quantitative Concepts

3 DESCRIPTIVE STATISTICS

As the name suggests, descriptive statistics are used to describe data. Often, you are
confronted by data that you need to organise in order to understand it. For exam-
ple, you get the feeling that the drive home from work is getting slower and you are
thinking of changing your route. How could you assess whether the journey really is
getting slower? Suppose you calculated and compared the average daily commute time
each month over a year. The first question you need to address is, what is meant by
average? There are a number of different ways to calculate averages that are described
in Section 3.1, each of which has advantages and disadvantages.

In general, descriptive statistics are numbers that summarise essential features of a data
set. A data set relates to a particular variablethe time it takes to drive home from
work in our example. The data set includes several observationsthat is, observed
values for the variable. For example, if you keep track of your daily commute time
for a year, you will end up with approximately 250 observations. The distribution of
a variable is the values a variable can take and the number of observations associated
with each of these values.

We will discuss two types of descriptive statistics: those that describe the central ten-
dency of a data set (e.g., the average or mean) and those that describe the dispersion
or spread of the data (e.g., the standard deviation). In addition to knowing whether
the drive to work is getting slower (by comparing monthly averages), you might also
want to find a way to measure how much variation there is between journey times
from one day to another (by using standard deviation).

Similar needs to summarise data arise in business. For example, when comparing the
time taken to process two types of trades, a sample of the times required to process
each trade would need to be collected. The average time it takes to process each type of
trade could be calculated and the average times could then be compared. Descriptive
statistics efficiently summarise the information from large quantities of data for the
purpose of making comparisons. Descriptive statistics may also help in predicting
future values and understanding risk. For example, if there was little variation in the
times taken to process a trade, then presumably you would be confident that you had
a good idea of the average time it takes to process a trade and comfortable with that
as an estimate of how long it will take to process future trades. But if the time taken
to process trades was highly variable, you would have less confidence in how long it
would take on average to process future trades.

3.1 Measures of Frequency and Average

The purpose of measuring the frequency of outcomes or central tendency is to
describe a group of individual data scores with a single measurement. The value used
to describe the group will be the single value considered to be most representative
of all the individual scores.

Measures of central tendency are useful for making comparisons between groups
of individuals or between sets of figures. Such measures reduce a large number of
measurements to a single figure. For instance, the mean or average temperature in
Descriptive Statistics 191

country X in July from 1961 to 2011 is calculated to be 16.1C. Over the same period
in September, the average temperature is 13.6C. Because it is a long time series, you
can reasonably conclude that it is usually warmer in July than September in country X.

Common measures of central tendency are

arithmetic mean,

geometric mean,

median, and

mode.

The appropriate measure for a given data set depends on the features of the data and
the purpose of your calculation. These measures are examined in the following sections.

3.1.1 Arithmetic Mean

The arithmetic mean is the most commonly used measure of central tendency and is
familiar to most people. It is usually shortened to just mean or average. To calculate
the mean, you add all the numbers in the data set together and divide by the number of
observations (items in the data set). The arithmetic mean assumes that each observation
is equally probable (likely to occur). If each observation is not equally probable, you
can get a weighted mean by multiplying the value of each observation by its proba-
bility and then summing these values. The sum of the probabilities always equals 1.

Exhibit5 shows the annual returns earned on an investment over a 10-year period. The
information contained in Exhibit5 will be used in examples throughout this section.

Exhibit5 Ten Years of Annual Returns

25
26.4%

20
Annual Returns (%)

15

10
8.0% 7.2%
4.2% 5.2%
5 3.7% 3.7%
2.4%
1.3% 0.8%

1 2 3 4 5 6 7 8 9 10
Year

Example7 shows the calculation of the arithmetic mean.

192 Chapter 8 Quantitative Concepts

EXAMPLE7. ARITHMETIC MEAN

An investment earns the returns shown in Exhibit5 over a 10-year period.

25
26.4%

20

Annual Returns (%)

15

10
8.0% 7.2%
Mean
5
2.4% 4.2% 5.2%
1.3% 0.8% 3.7% 3.7%

1 2 3 4 5 6 7 8 9 10
Year

(1.3 + 2.4 + 0.8 + 3.7 + 8.0 + 3.7 + 7.2 + 26.4 + 4.2 + 5.2)
= 6.3% Mean
10

The arithmetic mean return or average annual return over the 10-year period
is 6.3%. The weighted mean return (shown in the following equation) is the same
as the arithmetic return because the probability assigned to each return is the
same: 10% or 0.1.
Weighted mean annual return
= (0.1 1.3) + (0.1 2.4) + (0.1 0.8) + (0.1 3.7) + (0.18.0)
+(0.13.7) +(0.17.2) +(0.126.4) + (0.1 4.2) + (0.1 5.2)
= 6.3%

The arithmetic mean annual return is 6.3%.

The mean has one main disadvantage: it is particularly susceptible to the influence of
outliers. These are values that are unusual compared with the rest of the data set by
being especially small or large in numerical value. The arithmetic mean is not very
representative of the whole set of observations when there are outliers. Example8
shows the effect of excluding an outlier from the calculation of the arithmetic mean.

EXAMPLE8. EFFECT OF OUTLIER ON ARITHMETIC MEAN

In the case of the annual returns in Exhibit5, there is one value26.4%that

is much larger than the others. If this value is included, the mean is 6.3%, but
excluding this value reduces the mean to 4.1%.
Descriptive Statistics 193

25
26.4%
Annual Returns (%)
20
Outlier

15

10
8.0% 7.2%
5.2%
5 Mean without Outlier 4.2%
1.3% 0.8% 3.7% 3.7%
2.4%
1 2 3 4 5 6 7 8 9 10
Year

(1.3 + 2.4 + 0.8 + 3.7 + 8.0 + 3.7 + 7.2 + 4.2 + 5.2)

= 4.1% Mean
9

The arithmetic mean excluding the outlier is 4.1%.

Including the outlier, the mean is dragged in the direction of the outlier. When there
are one or more outliers in a set of data in one direction, the data are said to be
skewed in that direction. In Example7, ordering data so larger numbers are to the
right of smaller numbers, 26.4% lies to the right of the other data. Thus, the data are
said to be right skewed (or positively skewed). Other measures of central tendency
may better accommodate outliers.

3.1.2 Geometric Mean

An alternative average to the arithmetic mean is the geometric average or geometric
mean. Applied to investment returns, the geometric mean return is the average return
assuming that returns are compounding. To illustrate how the geometric mean is
calculated, let us start with the example of a three-year investment that returns 8%
the first year, 3% the second year, and 7% the third year.

8% 3% 7%

Present Year 1 Year 2 Year 3

[(1 + 8%) (1 + 3%) (1 + 7%)] 1/3 1 6.0%

The first step to calculate the geometric mean return is to multiply 1 plus each annual
return and add them together, which gives you the amount you would have accumu-
lated at the end of the three years per currency unit of investment: [(1+ 8%) (1+
3%) (1+ 7%) 1.1903]. This value of 1.1903 reflects three years of investment, but
the geometric mean return should capture an average rate of return for each of the
194 Chapter 8 Quantitative Concepts

three years. So, the second step requires moving from three years to one by raising
the accumulation to the power of one over the number of periods held, three in this
particular case; this calculation can also be described as taking the number of peri-
ods held root of the value (1.19031/3 1.060). This value of 1.060 includes both the
original investment and the average yearly return on the investment each year (1 plus
the geometric mean return). The last step is, therefore, to subtract 1 from this value
to arrive at the return that would have to be earned on average each year to get to the
total accumulation over the three years (1.060 1 0.060 or 6.0%). The geometric
mean return is 6.0%, which in this case is the same as the arithmetic mean return.
Geometric mean is frequently the preferred measure for the investment industry.

The following formula is used to arrive at the geometric mean return:

1/ t
Geometric mean return = (1 + r1) (1 + rt ) 1

where

ri = the return in period i expressed using decimals

t = the number of periods

Example9 shows the calculation of the geometric mean return for the investment
of Exhibit5.

EXAMPLE9. GEOMETRIC MEAN RETURN

If 1 currency unit was invested, you would have 1.8 currency units at the end
of the 10 years.
Total accumulation after 10 years
= [(1+ 1.3%) (1+ 2.4%) (1+ 0.8%) (1+ 3.7%) (1+ 8.0%) (1+
3.7%) (1+ 7.2%) (1+ 26.4%) (1+ 4.2%) (1+ 5.2%)]
= [(1.013) (1.024) (1.008) (1.037) (1.08) (1.037) (1.072)
(1.264) (1.042) (1.052)]
= 1.8
Average accumulation per year = 10th root of 1.8= (1.8)1/10 = 1.061
Geometric mean annual return = 1.061 1= 0.061= 6.1%
This can also be done as one calculation:
Geometric mean annual return
= {[(1+ 1.3%) (1+ 2.4%) (1+ 0.8%) (1+ 3.7%) (1+ 8.0%) (1+
3.7%) (1+ 7.2%) (1+ 26.4%) (1+ 4.2%) (1+ 5.2%)](1/10)} 1
= 6.1%

The geometric mean annual return is 6.1%. One currency unit invested for 10
years and earning 6.1% per year would accumulate to approximately 1.8 units.
Descriptive Statistics 195

An important aspect to notice is that the geometric mean is lower than the arithmetic
mean even though the annual returns over the 10-year holding period are identical.
This result is because the returns are compounded when calculating the geometric
mean return. Recall that compounding will result in a higher value over time, so a
lower rate of return is required to reach the same amount. In fact, if the same set of
numbers is used to calculate both means, the geometric mean return is never greater
than the arithmetic mean return and is normally lower.

3.1.3 Median
If you put data in ascending order of size from the smallest to the largest, the median
is the middle value. If there is an even number of items in a data set, then you average
the two middle observations. Hence, in many cases (i.e., when the sample size is odd
or when the two middle-ranked items of an even-numbered data set are the same)
the median will be a number that actually occurs in the data set. Example10 shows
the calculation of the median for the investment of Exhibit5.

EXAMPLE10. MEDIAN

When the returns are ordered from low to high, the median value is the arith-
metic mean of the fifth and sixth ordered observations.

Annual Returns Ordered Low to High

0.8% 1.3% 2.4% 3.7% 3.7% 4.2% 5.2% 7.2% 8.0% 26.4%

(3.7 + 4.2)
4.0% Median
2
25
26.4%

20
Annual Returns (%)

15

10
8.0% 7.2%
4.2% 5.2%
5 Median
1.3% 0.8% 3.7% 3.7%
2.4%
1 2 3 4 5 6 7 8 9 10
Year

The median investment return over the 10-year period is 4.0%.

196 Chapter 8 Quantitative Concepts

An advantage of the median over the mean is that it is not sensitive to outliers. In the
case of the annual returns shown in Exhibit5, the median of close to 4.0% is more
representative of the datas central tendency. This 4.0% median return is close to the
4.1% arithmetic mean return when the outlier is excluded. The median is usually a
better measure of central tendency than the mean when the data are skewed.

3.1.4 Mode
The mode is the most frequently occurring value in a data set. Example11 shows how
the mode is determined for the investment of Exhibit5.

EXAMPLE11. MODE

Looking at Exhibit5, we see that one value occurs twice, 3.7%. This value is the
mode of the data.

Annual Returns Ordered Low to High

0.8% 1.3% 2.4% 3.7% 3.7% 4.2% 5.2% 7.2% 8.0% 26.4%

3.7% Mode

The mode can be used as a measure of central tendency for data that have been sorted
into categories or groups. For example, if all the employees in a company were asked
what form of transportation they used to get to work each day, it would be possible
to group the answers into categories, such as car, bus, train, bicycle, and walking. The
category with the highest number would be the mode.

A problem with the mode is that it is often not unique, in which case there is no
mode. If there are two or more values that share the same frequency of occurrence,
there is no agreed method to choose the representative value. The mode may also
be difficult to compute if the data are continuous. Continuous data are data that can
take on an infinite number of values between whole numbersfor example, weights
of people. One person may weigh 62.435 kilos and another 62.346 kilos. By contrast,
discrete data show observations only as distinct valuesfor example, the number
of people employed at different companies. The number of people employed will be
a whole number. For continuous data, it is less likely that any observation will occur
more frequently than once, so the mode is generally not used for identifying central
tendency for continuous data.

Another problem with the mode is that the most frequently occurring observation may
be far away from the rest of the observations and does not meaningfully represent them.
Descriptive Statistics 197

3.2 Measures of Dispersion

Whereas measures of central tendency are used to estimate representative or cen-
tral values of a set of data, measures of dispersion are important for describing the
spread of the data or its variation around a central value. Two data sets may have the
same mean or median but completely different levels of variability, or vice versa. A
description of a data set should include both a measure of central tendency, such as
the mean, and a measure of dispersion. Suppose two companies both have an average
annual salary of $50,000, but in one company most salaries are clustered close to the
average, whereas in the second they are spread out with many people earning very
little and some earning a lot. It would be useful to have a measure of dispersion that
can help identify such differences between data sets.

140

120

100
Salary ($ thousands)

80

60 Average
Annual Salary

40

20

0
Company A Company B

Another reason why measures of dispersion are important in finance is that invest-
ment risk is often measured using some measure of variability. When investors are
considering investing in a security, they are interested in the likely (expected) return
on that investment as well as in the risk that the return could differ from the expected
return (its variability). A risk-averse investor considering two investments that have
similar expected returns but very different measures of variability (risk) around those
expected returns, typically prefers the security with the lower variability.

Two common measures of dispersion of a data set are the range and the standard
deviation.
198 Chapter 8 Quantitative Concepts

3.2.1 Range
The range is the difference between the highest and lowest values in a data set. It is
the easiest measure of dispersion to calculate and understand, but it is very sensitive
to outliers. Example12 explains the calculation of the range of returns for the invest-
ment of Exhibit5.

EXAMPLE12. RANGE

In Exhibit5 we see that the highest annual return is 26.4% and the lowest annual
return is 0.8%.

Annual Returns Ordered Low to High

0.8% 1.3% 2.4% 3.7% 3.7% 4.2% 5.2% 7.2% 8.0% 26.4%

26.4% 0.8% = 25.6% Range

If the extreme value at the upper end of the range is excluded, the next highest
value, 8.0%, is used to estimate the range, and the range is reduced significantly.

Annual Returns Ordered Low to High

0.8% 1.3% 2.4% 3.7% 3.7% 4.2% 5.2% 7.2% 8.0% 26.4%

8.0% 0.8% = 7.2% Range

Clearly, the range is affected by extreme values and, if there are outliers, it says little
about the distribution of the data between those extremes.

If there are a large number of observations ranked in order of size, the range can be
divided into 100 equal-sized intervals. The dividing points are termed percentiles. The
50th percentile is the median and divides the observations so that 50% are higher and
50% are lower than the median. The 20th percentile is the value below which 20% of
observations in the series fall. So, the dispersion of the observations can be described
in terms of percentiles. Observations can be divided into other equal-sized intervals.
Commonly used intervals are quartiles (the observations are divided into four equal-
sized intervals) and deciles (the observations are divided into 10 equal-sized intervals)

3.2.2 Standard Deviation

A commonly used measure of dispersion is the standard deviation. It measures the
variability or volatility of a data set around the average value (the arithmetic mean)
of that data set. Although, as mentioned before, you are not responsible for any cal-
culations, you may find it helpful to look at the formula for how standard deviation
is calculated.
Descriptive Statistics 199

2
Standard deviation =
[X1 E ( X )]2 + [X 2 E ( X )]2 + ... + [X n E ( X )]
n
where

Xi = observation i (one of n possible outcomes for X)

n = number of observations of X
E(X) = the mean (average) value of X or the expected value of X
[Xi E(X)] = difference between value of observation Xi and the mean value of
X

The differences between the observed values of X and the mean value of X capture
the variability of X. These differences are squared and summed. Note that because
the differences are squared, what matters is the size of the difference not the sign of
the difference. The sum is then divided by the number of observations. Finally, the
square root of this value is taken to get the standard deviation.

The value before the square root is taken is known as the variance, which is another
measure of dispersion. The standard deviation is the square root of the variance. The
standard deviation and the variance capture the same thinghow far away from
the mean the observations are. The advantage of the standard deviation is that it
is expressed in the same unit as the mean. For example, if the mean is expressed as
minutes of journey time, the standard deviation will also be expressed as minutes,
whereas the variance will be expressed as minutes squared, making the standard
deviation an easier measure to use and compare with the mean.

To illustrate the calculation of the standard deviation, let us return to the example of
a three-year investment that returns 8% or 0.08 the first year, 3% or 0.03 the second
year, and 7% or 0.07 the third year. The arithmetic mean return is 6% or 0.06. The
standard deviation is approximately 2.16%.

8% 3% 7%

Present Year 1 Year 2 Year 3

(0.08 0.06)2 + (0.03 0.06)2 + (0.07 0.06)2

Standard deviation =
3

(0.02)2 + (0.03)2 + (0.01)2

=
3

(0.0004) + (0.0009) + (0.0001)

=
3

(0.0014)
= = 0.0216= 2.16%
3
200 Chapter 8 Quantitative Concepts

Example 13 shows the calculation of the standard deviation for the investment in
Exhibit5.

EXAMPLE13. STANDARD DEVIATION

The arithmetic mean annual return, as calculated in Example7, is 6.3%.

Standard deviation
= square root of {[(0.013 0.063)2 + (0.024 0.063)2 + (0.008 0.063)2
+ (0.037 0.063)2 + (0.08 0.063)2 + (0.037 0.063)2 + (0.072
0.063)2 + (0.264 0.063)2 + (0.042 0.063)2 + (0.052 0.063)2] 10}
= square root of [(0.0025+ 0.0015+ 0.0030+ 0.0007+ 0.0003+
0.0007+ 0.0001+ .0404+ 0.0004+ 0.0001) 10]
= square root of 0.00497
= 0.0705, rounded to the nearest 10th percent = 7.1% (this value is used
in Example14).

The standard deviation is 7.1%.

Larger values of standard deviation relative to the mean indicate greater variation in
a data set. Also, by using standard deviation, you can determine how likely it is that
any given observation will occur based on its distance from the mean. Example14
compares the returns of the investment shown in Exhibit5 and the returns on another
investment over the same period using mean and standard deviation.

EXAMPLE14. COMPARISON OF INVESTMENTS

An investment earns the returns shown in Exhibit5 over a 10-year period:

Number of observations = 10
Mean = 6.3%
Standard deviation = 7.1%
Another investment over the same time period has the following characteristics:
Number of observations = 10
Mean = 6.5%
Standard deviation = 2.6%
Based on mean and standard deviation, the second investment is better than the
first investment. It has a higher mean return and less variability, which implies
less risk, in its returns.
Descriptive Statistics 201

3.2.3 Normal Distribution

The arithmetic mean and standard deviation are two powerful ways of describing
many distributions of data. A distribution is simply the set of values that a variable
can take, showing their observed or theoretical frequency of occurrence. For example,
consider the distribution of salaries earned by employees in two companies as shown
in Exhibit6. The observations in these distributions are grouped into different salary
ranges.

Exhibit6 Number of Employees in Various Salary Ranges

Number of Employees
Salary ($) Company X Company Y

15,00020,000 5 1
20,00125,000 8 1
25,00130,000 20 3
30,00135,000 30 8
35,00140,000 22 10
40,00145,000 12 15
45,00150,000 6 20
50,00155,000 2 9
55,00160,000 1 7

Sometimes it is helpful to look at a picture of the distribution to understand it. The

shape of the distribution has a bearing on how you interpret the summary measures
of the distribution. This data can be shown pictorially using a histograma bar
chart with bars that are proportional to the frequency of occurrence of each group
of observationsas shown in Exhibits 7A and 7B.

Exhibit7A Salaries of Employees at Company X

35
Number of Employees

30
25
20
15
10
5
0
1520 2025 2530 3035 3540 4045 4550 5055 5560
Salary Range ($ thousands)
202 Chapter 8 Quantitative Concepts

Exhibit7B Salaries of Employees at Company Y

25

Number of Employees
20

15

10

0
1520 2025 2530 3035 3540 4045 4550 5055 5560
Salary Range ($ thousands)

Note that the two distributions are not symmetrical. A symmetrical distribution
would have observations falling off fairly evenly on either side of the centre of the
range of salaries ($35,001$40,000). Instead, in each of these distributions, the bulk
of the observations are stacked towards one end of the range and tail off gradually
towards the other end. The two distributions are different in that each is stacked
towards a different end. Such distributions are considered skewed; the distribution
for Company X is positively skewed (i.e., the majority of the observations are on the
left and the skew or tail is on the right), whereas the distribution for Company Y is
negatively skewed (left skewed).

Although the range of the observations is the same in each case, the mean for each
is very different. Company Xs mean is approximately $35,000, whereas Company Ys
mean is approximately $44,000.

For a perfectly symmetrical distribution, such as a normal distribution (see Exhibit8),

the mean, median, and mode will be identical. If the distribution is skewed, these three
measures of central tendency will differ. Looking again at Company Xs salary data, for
instance, we do not have enough detailed information to identify a mode. The mean
is larger than the median because the mean is more affected by extreme values than
the median. The distribution is skewed to the right, so the mean is dragged towards
the extreme positive values. The reverse is true for distributions that are negatively
skewed, such as in Company Ys salary data. In this case, the mean is smaller than the
median because the mean is pulled left in the direction of the skew.

A normal distribution is represented in a graph by a bell curve; an example of a bell-

shaped curve is shown in Exhibit 8. The shape of the curve is symmetrical with a
single central peak at the mean of the data and the graph falling off evenly on either
side of the mean; 50% of the distribution lies to the left of the mean, and 50% lies to
the right of the mean. The shape of a normal distribution depends on the mean and
the standard deviation. The mean of the distribution determines the location of the
centre of the curve, and the standard deviation determines the height and width of
the curve. When the standard deviation is large, the curve is short and wide; when
the standard deviation is small, the curve is tall and narrow.
Descriptive Statistics 203

A normal distribution has special importance in statistics because many variables have
the approximate shape of a normal distributionfor example, height, blood pressure,
and lengths of objects produced by machines. This distribution is often useful as a
description of data when there are a large number of observations.

A normal distribution is a distribution of a continuous random variable (i.e., a variable

that can take on an infinite number of values). The vertical axis for the normal distri-
bution is the probability or likelihood of occurrence. By contrast, on the histogram
shown earlier, the vertical axis was frequency of occurrence. The mean (and median)
is the centre of the distribution and has the highest probability of occurrence. Half
of the observations lie on one side of the mean and half on the other. Approximately
two-thirds of the observations are within one standard deviation of the mean, and
95% of observations are within two standard deviations of the mean. Exhibit8 shows
a normal distribution.

Exhibit8 Standard Deviation (SD) and Normal Distribution

0.3413 0.3413

0.0228 0.0228
0.1359 0.1359

SD
3 2 1 0 1 2 3
68.26%
95.44%

The total area under the curve or bell is 100% of the distribution. The area under
the curve that is within one standard deviation of the mean is about 68% of all the
observations. In other words, given a mean of 0 and a standard deviation of 1, about
68% of the observations fall between 1 and +1, and 32% of the observations are more
than one standard deviation from the mean. The area under the curve that is within 2
standard deviations of the mean is about 95% of the observations. Given a mean of 0
and a standard deviation of 1, about 95% of the observations fall between 2 and +2,
and 5% of the observations are more than two standard deviations from the mean. The
area under the curve that is within three standard deviations of the mean represents
about 99% of the observations. Given a mean of 0 and a standard deviation of 1, about
99% of the observations fall between 3 and +3, and less than 1% of the observations
occur more than three standard deviations away from the mean.

The observations that are more than a specified number of standard deviations from
the mean can be described as lying in the tails of the distribution. Assuming that
returns on a portfolio of stocks are normally distributed, the chance of extreme losses
(a return more than three standard deviations lower than the mean return) is relatively
204 Chapter 8 Quantitative Concepts

small. The chance of the return being in the left tail more than two standard deviations
from the mean (which would be an extreme loss under typical circumstances) is just
2.5%. In other words, out of 200 days, 5 days are expected to have observations that
are more than two standard deviations from the mean. But during the financial crisis
of 2008, the losses that were incurred by some banks over several days in a row were
25 standard deviations below the mean.

To put this in perspective, if returns are normally distributed, a return that is 7.26
standard deviations below the mean would be expected to occur once every 13.7billion
years. That is approximately the age of the universe. The frequency of extreme events
during the financial crisis of 2008 was, therefore, much higher than predicted by the
normal distribution. This inconsistency is often referred to as the distribution having
fat tails, meaning that the probability of observing extreme outcomes is higher than
that predicted by a normal distribution.

Exhibit9 gives examples of different types of bell-shaped distributions. How would

you describe each curve? What does each tell you about the likelihood of extreme
outcomes?

Exhibit9 Bell-Shaped Distributions with Fat and Thin Tails

Normal Distribution Distribution with Fat Tails

Distribution with Thin Tails

In Exhibit9, the curve with the solid line represents the normal distribution. The curve
with the dotted line is an example of distribution with thinner tails than the normal
distribution, indicating a reduced probability of extreme outcomes. By contrast, the
curve with the dashed line is an example of a distribution with fatter tails than the
normal distribution, indicating increased likelihood of extreme outcomes.
Descriptive Statistics 205

3.3 Correlation
Another way of using and understanding data is identifying connections between
data sets. The strength of a relationship between two variables, such as growth in
gross domestic product (GDP) and stock market returns, can be measured by using
correlation. Essentially, two variables are correlated when a change in one variable
helps predict change in another variable.

When both variables change in the same direction, the variables are positively cor-
related. If we take the example of traders at an investment bank, salary and age are
positively correlated if salaries increase as age increases. If the variables move in the
opposite direction, then they are negatively correlated. For example, the size of a
transaction and the fees expressed as a percentage of the transaction are negatively
correlated if the larger the transaction, the smaller the associated fees. When there is no
clear tendency for one variable to move in a particular direction (up or down) relative
to changes in the other variable, then the variables are close to being uncorrelated.
In practice, it is difficult to find two variables that have absolutely no relationship,
even if just by chance.

Correlation is measured by the correlation coefficient, which has a scale of 1 to +1.

When two variables move exactly in step with each other in the same directionif
one goes up, the other goes up in the same proportionthe variables are said to be
perfectly positively correlated. In that case, the correlation coefficient is at its maximum
of +1. When the two variables move exactly in step in opposite directions, they are
perfectly negatively correlated and the correlation coefficient is 1. Variables with no
relationship to each other will have a correlation coefficient close to 0.

Correlation measures both the direction of the relationship between two variables
(negative or positive) and the strength of that relationship (the closer to +1 or 1, the
stronger the relationship). In practice, it is unusual to find variables that are perfectly
positively or perfectly negatively correlated. The stronger the relationship between two
variablesthe higher the degree of correlationthe more confidently one variable can
be predicted given the other variable. For example, there may be a high correlation
between stock market index returns and expected economic growth. In that case, if
economic growth in the future is expected to be high then returns on the stock market
index are likely to be high too.

It is important, however, to realise that correlation does not imply causation. For
example, historically in the United States, stock market returns and snowfall are both
higher in January, and from that you may assume a correlation. But obviously snowfall
does not cause an increase in stock market returns, and an increase in stock market
returns clearly does not cause snowfall. There may be situations in which a correla-
tion implies some causal relationship. For example, a high correlation has been found
between power production and job growth. It may follow that the more workers there
are, the more power is consumed, but it does not necessarily follow that an increase
in power generation will create jobs.

Correlation is important in investing because the rise or fall in value of a variable may
help predict the rise or fall in value of another variable. It is also important because
when two or more securities that are not perfectly correlated are combined together in
a portfolio, there is normally a reduction in risk (measured by the portfolios standard
deviation of returns). As long as the returns on the securities do not have a correlation
206 Chapter 8 Quantitative Concepts

of +1 (that is, they are less than perfectly correlated), then the risk of the portfolio will
be less than the weighted average of the risks of the securities in the portfolio because
it is not likely that all the securities will perform poorly at the same time.

The practice of combining securities in a portfolio to reduce risk is known as diver-

sification. An extreme example of an undiversified portfolio is one holding only one
security. This approach is risky because it is not unusual for a single security to go
down in value by a large amount in one year. It is much less common for a diversified
portfolio of 20 different securities to go down by a large amount, even if they are
selected at random. If the securities are selected from a variety of sectors, industries,
company sizes, asset classes, and markets, it is even less likely. One caveat is that the
benefits of diversification are much reduced in periods of financial crisis. In such
periods, the correlation between returns on different securities (and different asset
classes) tends to increase towards +1.

SUMMARY

The better your understanding of quantitative concepts, the easier it will be for you
to make sense of the financial world. Knowledge of quantitative concepts, such as
time value of money and descriptive statistics, is important to the understanding of
many of the key products in the financial industry. Understanding the time value of
money allows you to interpret cash flows and thus value them. Meanwhile, knowledge
of statistical concepts will help in identifying the important information in a large
amount of data, as well as in understanding what statistical measures reported by
others mean. It is easy to misinterpret or be misled by statistics, such as mean and
correlation, so an understanding of their uses and limitations is crucial.

Interest is return earned by a lender that compensates for opportunity cost and
risk. For the borrower, it is the cost of borrowing.

The simple interest rate is the cost to the borrower or the rate of return to the
lender, per period, on the original principal borrowed. A commonly quoted
simple interest rate is the annual percentage rate (APR).

Compound interest is the return to the lender or the cost to the borrower when
interest is reinvested and added to the original principal.

The effective annual rate (EAR) of interest is calculated by annualising a rate

that is compounded more than once a year. The EAR is equal to or greater than
the annual percentage rate.

The present value of a future sum of money is found by discounting the future
sum by an appropriate discount rate. (The present value of multiple cash flows
is the sum of the present value of each cash flow.)

Three elements must be considered when comparing investments: the cash

flows each investment will generate in the future, the timing of these cash flows,
and the risk associated with each investment. The discount rate reflects the
riskiness of the cash flows.
Summary 207

All else being equal (in other words, only one of the three elements differs):

the higher the cash flows, the higher the present and future values.

the earlier the cash flows, the higher the present and future values.

the lower the discount rate, the higher the present value.

the higher the interest rate, the higher the future value.

The net present value is the present value of future cash flows net of the invest-
ment required to obtain them. It is useful when comparing alternatives that
require different initial investments.

Financial instruments can be valued as the present value of their expected

future cash flows.

An annuity involves an initial payment (outflow) in exchange for a fixed number

of future receipts (inflows), each of an equal amount. Mortgages are amortising
loans; the periodic payment is fixed, and in each period some of the payment
covers the interest on the loan and the rest of the payment pays off some of
the principal. Over time, the portion of the payment that reduces the principal
increases.

The role of descriptive statistics is to summarise the information given in large

quantities of data for the purpose of making comparisons, predicting future
values, and better understanding the data.

The purpose of measures of frequency or central tendency is to describe a

group of individual data scores with a single measurement. This measure is
intended to be representative of the individual scores. Measures of central ten-
dency include arithmetic mean, geometric mean, median, and mode. Different
measures are appropriate for different types of data.

The arithmetic mean is the most commonly used measure. It represents the
sum of all the observations divided by the number of observations. It is an easy
measure to understand but may not be a good representative measure when
there are outliers.

The geometric mean return is the average compounded return for each
periodthat is, the average return for each period assuming that returns are
compounding. It is frequently the preferred measure of central tendency for
returns in the investment industry.

When observations are ranked in order of size, the median is the middle value.
It is not sensitive to outliers and may be a more representative measure than the
mean when data are skewed.

The mode is the most frequently occurring value in a data set. A data set may
have no identifiable unique mode. It may not be a meaningful representative
measure of central tendency.

Measures of dispersion are important for describing the spread of the data, or
its variation around a central value. Two common measures of dispersion are
range and standard deviation.
208 Chapter 8 Quantitative Concepts

Range is the difference between the highest and lowest values in a data set. It is
easy to measure, but it is sensitive to outliers.

Standard deviation measures the variability of a data set around the mean of the
data set. It is in the same unit of measurement as the mean.

A distribution is simply the values that a variable can take, showing its observed
or theoretical frequency of occurrence.

For a perfectly symmetrical distribution, the mean, median, and mode will be
identical.

A common symmetrical distribution is the normal distribution, a bell-shaped

curve that is represented by its mean and standard deviation. In a normal
distribution, 68% of all the observations lie within one standard deviation of the
mean and about 95% of the observations are within two standard deviations.

The strength of a relationship between two variables can be measured by using

correlation.

Correlation is measured by the correlation coefficient on a scale from 1 to

+1. When two variables move exactly in tandem with each other, the variables
are said to be perfectly positively correlated and the correlation coefficient is
+1. When two variables move exactly in opposite directions, they are perfectly
negatively correlated and the correlation coefficient is 1. Variables with no
relationship to each other will have a correlation coefficient close to 0.

It is important to realise that correlation does not imply causation.