QUANTITATIVE ANALYSIS

Lecture Notes, Past Quiz, Exam, & Home Work Questions

Prof. Dr. Serhan FTOLU Gazimausa 2005

Foreword
These Lecture Notes have been prepared to serve as a study guide for the students of Quantitative Analysis course (MGMT 322), that I offer for business students. They are designed to outline the critical topics that are covered by the course but also attempt to give an example about the application of each quantitative and statistical tool covered by the course. At the end of these Lecture Notes I also added examples of past quizzes, exams, and homework questions, to help interested students both to improve their analytical capacity regarding the application of the quantitative tools, that they have learn, and at the same time better prepare for exams. I would like to express my gratitude for the excellent work that our teaching assistants and former students of MGMT 322, Erhan ERDOAN and Sleyman EFE, and Gift, have done in organizing and typing my in-class lectures.

Assoc. Prof. Dr. Serhan FTOLU Department of Business Administration Eastern Mediterranean University

Quantitative Analysis MGMT 322

INDEX CHAPTER 2: A REVIEW OF PROBABILITY THEORY ..................................................- 4 A. PROBABILITY ......................................................................................................................... - 4 B. EVENT ..................................................................................................................................... - 4 EXAMPLE ..................................................................................................................................... - 4 C. EXPERIMENT .......................................................................................................................... - 4 D. SAMPLE SPACE ...................................................................................................................... - 4 EXAMPLE ..................................................................................................................................... - 4 E. MUTUALLY EXCLUSIVE EVENTS .......................................................................................... - 5 EXAMPLE ..................................................................................................................................... - 5 ALTERNATIVE APPROACHES IN PROBABILITY THEORY.......................................- 5 1. CLASSICAL APPROACH .......................................................................................................... - 5 EXAMPLE ..................................................................................................................................... - 6 2. RELATIVE FREQUENCY APPROACH ...................................................................................... - 6 EXAMPLE ..................................................................................................................................... - 6 3. SUBJECTIVE APPROACH ........................................................................................................ - 6 EXAMPLE ..................................................................................................................................... - 6 PROBABILITY RULES...........................................................................................................- 7 A. MARGINAL (UNCONDITIONAL) PROBABILITY: ................................................................... - 7 B. THE ADDITION RULE: ........................................................................................................... - 7 1. THE ADDITION RULE FOR MUTUALLY EXCLUSIVE EVENTS ................................................... - 7 EXAMPLE ..................................................................................................................................... - 7 2. THE ADDITION RULE FOR NON-MUTUALLY EXCLUSIVE EVENTS .......................................... - 7 EXAMPLE ..................................................................................................................................... - 8 EXAMPLE ..................................................................................................................................... - 8 TYPES OF EVENTS.................................................................................................................- 9 1. STATISTICALLY INDEPENDENT EVENTS ............................................................................... - 9 2. STATISTICALLY DEPENDENT EVENTS ................................................................................ - 10 TYPES OF PROBABILITIES UNDER INDEPENDENCE ................................................................ - 10 1. MARGINAL (UNCONDITIONAL) PROBABILITY ....................................................................... - 10 2. JOINT PROBABILITY ............................................................................................................... - 10 3. CONDITIONAL PROBABILITY ................................................................................................. - 10 TYPES OF PROBABILITIES UNDER DEPENDENCE ................................................................... - 11 1. MARGINAL (UNCONDITIONAL) PROBABILITY ....................................................................... - 11 2. JOINT PROBABILITY ............................................................................................................... - 11 3. CONDITIONAL PROBABILITY ................................................................................................. - 11 EXAMPLE ................................................................................................................................... - 11 -

-1-

Quantitative Analysis MGMT 322

BAYES THEOREM ..............................................................................................................- 12 EXAMPLE ................................................................................................................................... - 12 EXAMPLE ................................................................................................................................... - 13 PROBABILITY DISTRIBUTION.........................................................................................- 15 DISCRETE VS CONTINUOUS RANDOM VARIABLES .................................................................... 15 DISCRETE PROBABILITY DISTRIBUTION ........................................................................................ 15 CONTINUOUS PROBABILITY DISTRIBUTION .................................................................................. 15 BINOMIAL PROBABILITY DISTRIBUTION ................................................................................ - 18 CHARACTERISTICS OF A BERNOULLI TYPE PROCESS ................................................................ - 18 BINOMIAL FORMULA:................................................................................................................ - 18 EXAMPLE ................................................................................................................................... - 19 NORMAL PROBABILITY DISTRIBUTION .................................................................................. - 23 CHARACTERISTICS ..................................................................................................................... - 24 TWO PARAMETERS: MEAN () & STANDARD DEVIATION ()................................................... - 24 EXAMPLE: .................................................................................................................................. - 20 3 MATHEMATICAL FACTS ABOUT NORMAL DISTRIBUTION ................................................. - 24 EXAMPLE: .................................................................................................................................. - 25 CHAPTER 3: FORECASTING...............................................................................................-29THE METHODOLOGY OF FORECASTING................................................................................. - 29 TECHNICAL ANALYSIS

FUNDAMENTAL ANALYSIS BASIC STEPS IN FORECASTING FORECASTING MODELS ........................................................................................................... - 31 FOR NON-SEASONAL PRODUCTS WITH LINEAR TREND TYPE OF PATTERN........................ - 31 NAVE MODEL ........................................................................................................................... - 33 MOVING AVERAGE MODEL. .-39SIMPLE EXPONENTIAL MODEL .................................................................................................. - 41 IMPORTANT STEPS ..................................................................................................................... - 42 TIME SERIES REGRESSION MODEL ............................................................................................ - 42 IMPORTANT POINTS ................................................................................................................... - 44 SMOOTHING LINEAR TREND MODEL......................................................................................... - 46 IMPORTANT STEPS ..................................................................................................................... - 46 CHAPTER 4: ALTERNATIVE DECISION MAKING ENVIRONMENTS ....................- 48 BASIC STEPS IN DECISION MAKING ALTERNATIVE DECISION MAKING ENVIRONMENTS ALTERNATIVE CRITERIA FOR DECISION MAKING UNDER UNCERTAINTY .......................... - 48 THE MAXI-MAX CRITERION ...................................................................................................... - 49 THE MAXI-MIN CRITERION ....................................................................................................... - 49 THE CRITERION OF REALISM ..................................................................................................... - 50 THE MINI-MAX REGRET CRITERION ......................................................................................... - 50 ALTERNATIVE CRITERIA FOR DECISION MAKING UNDER RISK .......................................... - 51 -

-2-

Quantitative Analysis MGMT 322

EXPECTED VALUE CRITERION ................................................................................................... - 51 EXAMPLE ................................................................................................................................... - 51 CRITERION OF RATIONALITY ..................................................................................................... - 52 THE CRITERION OF MAXIMUM LIKELIHOOD ............................................................................. - 52 SUPPLYING THE NUMBERS: HOW TO OBTAIN MEAN AND STANDARD DEVIATION WHEN DATA IS MISSING OR INCOMPLETE........................................- 57 -

CHAPTER 5: COST VOLUME PROFIT ANALYSIS .................................................- 61 GENERAL DEFINITION AND USE OF COST- VOLUME- PROFIT ANALYSIS ITS APPLICATION COMBINING UNIT MONETARY VALUES AND PROBABILITY DISTRIBUTION.............................. - 64 3 STEP PROCEDURE TO OBTAIN EXPECTED PROFITS .............................................................. - 67 3 STEP PROCEDURE TO OBTAIN EXPECTED LOSS ................................................................... - 68 -

Home Works-70-

-3-

Quantitative Analysis MGMT 322

Chapter 2: A Review of Probability Theory

A. Probability
Probability is the chance or likelihood of something occurring (happening). We express probabilities as either fractions or decimals E.g: P(H) = 9/10 (as a fraction) or P(H) = 0.9 (as a decimal) Note: The probabilities should neither be < 0, nor >1.

B. Event
An event is defined as either one or more than one of possible outcomes of doing something or an activity.

Example
ACTIVITY: Inflationary process in Turkey in the year 2005: Events: A: If annual inflation for 2005 is exactly 10% B: If annual inflation for 2005 is less than 8% C: If annual inflation for 2005 is either 5% or 7%

C. Experiment
Is the activity that generates events. For example, tossing a coin twice.

D. Sample Space
The sample space of an experiment is the list of all possible outcomes of that experiment.

Example
Experiment: Tossing a coin. Sample Space: S = {HH, TT, HT, TH} Events: A: Exactly one outcome, that is, getting 2heads = HH B: More than one outcome, that is, getting at least one tail = (TT, HT, TH) C: Exactly three outcomes, that is, getting 2tails or getting one head and one tail = (TT, HT, TH) -4-

Quantitative Analysis MGMT 322

E. Mutually Exclusive Events

Mutually exclusive events are those events which can take place only one at a time, and not together at the same time.

F. Non Mutually Excusive Events

Are those events which can happen together at the same time.

Example 1
Experiment: Tossing a coin twice Events: X: Getting two Heads: P(HH) Y: Getting two Heads or two Tails: P(HH or TT) Z: Getting Head-Tail or Tail-Head: P(HT or TH) X and Y are non - mutually exclusive but X and Z are mutually exclusive.

Example 2
Experiment: Randomly selecting a student in this class. Events: A: Selected student is 21 years old. B: Selected student is male C: Selected student is female D: Selected student is 22 years old. Therefore: Events AB are non-mutually exclusive because they can happen together at the same time. Events AD are mutually exclusive because they can not happen together at the same time. Events AC are non-mutually exclusive because they can happen together at the same time. Events BC are mutually exclusive because they can not happen together at the same time.

Alternative Approaches in Probability Theory

1. Classical Approach
P(A) = # of outcomes favorable to the occurrence of event A Total # of possible outcomes

-5-

Quantitative Analysis MGMT 322

Example 1
Experiment: Randomly selecting a student in a class of 40 students. Events: A: Selecting a female student, given total number of females equal to 10. B: Selecting a male student, given total number of males equal to 30. The total number of students is 40 P(A) = 10/40 P(B) = 30/4

2. Relative Frequency Approach

The probability of the event found by using the given past data.

Example
Experiment: Selecting a student randomly. Events: A: Selecting a student who gets A from MGMT 322. Past Data: 20% of all students who have taken MGMT 322 in the past get A. P(A) = 0.20 B: Galatasaray wins the championship in 2005, using the past data obtain how frequently the similar event occurred in the past. Past Data: 40% of the last 9 years, Galatasaray was a champion. P(B) = 0.40

3. Subjective Approach
The probability of the event found by using the personal feeling or experience.

Example
Event: A: Galatasaray wins. P(A) = 0.90 As a supporter or a football watcher you may have the feeling that Galatasaray will win 90%.

-6-

Quantitative Analysis MGMT 322

Probability Rules
A. Marginal (Unconditional) Probability:
Only one event can take place. P(A)

B. The Addition Rule:

1. The Addition Rule for Mutually Exclusive Events
If Events A, B, and C are mutually exclusive (meaning they can not happen together at the same time.) Then: P(A or B) = P(A) + P(B) P( A or B or C) = P(A) + P(B) + P(C) Questions that can be asked and answered by using Addition Rule: Q1: What is the probability of A or B or C taking place? Q2: What is the probability of A or B taking place? Q3: What is the probability of B or C taking place?

Example
Suppose that the number of students in each age group in the class is as follows: Age 20 21 22 23 24 Total Male 0 1 6 3 4 14 Female 2 3 3 1 0 9 Total 2 4 9 4 4 23

Events: A: Selecting a random student who is 20 years old. B: Selecting a random student who is 21 years old. C: Selecting a random student who is 22 years old. D: Selecting a random student who is 23 years old. P(A or B) = P(A) + P(B) = 2/23 + 4/23 = 6/23 P(A or C) = P(A) + P(C) = 2/23 + 9/23 = 11/23 P(A or B or C or D) = P(A) + P(B) + P(C) + P(D) = 2/23 + 4/23 + 9/23 + 4/23 = 19/23

2. The Addition Rule for Non-Mutually Exclusive Events

If A and B are non-mutually exclusive events, Then: P(A or B) = P(A) + P(B) P(A and B)

-7-

Quantitative Analysis MGMT 322

Example 1
Suppose that the sex and the eye color of students in a class of 40 people is distributed as follows: Black 13 8 21 Green 9 10 19 Total 22 18 40

Male Female Total

Experiment: Selecting a random student. Events: A: Selecting a male student. B: Selecting a female student. C: Selecting a black eye student. D: Selecting a green eye student. P(A or C) = P(A) + P(C) P(A and C) = (22/40 + 21/40) 13/40 = 30/40 P(B or C) = P(B) + P(C) P(B and C) = (18/40 + 21/40) 8/40 = 31/40

Example 2
Suppose that the Ford Motor Corporation produces 3 different models for Turkish consumers. Annual volume of production and sales of all 3 model combined is 700 units. However, some of the autos sold turned to be defective. Based on the past data the production manager estimated the following distribution of defective and non-defective autos for each model manufactured. A 10 140 150 Models B 20 180 200 C 70 280 350 Total 100 600 700

Defective Non-Defective Total

a. What is the probability that a randomly selected consumer will buy either model A or C? b. What is the probability that the consumer will buy either model B or a Defective auto? c. What is the probability that the consumer will buy either model A or a Nondefective auto? a. P(A or C) = P(A) + P(C) = 150/700 + 350/700 = 500/700 b. P(B or D) = P(B) + P(D) P(B and D) = (200/700 + 100/700) 20/700 = 280/700 c. P(A or N) = P(A)+P(N) P(A and N) = (150/700 + 600/700) 140/700= 610/700 -8-

Quantitative Analysis MGMT 322

Example 3
There are 500 companies whose stocks are traded on Istanbul stock exchange. These companies operate in these basic sectors of the economy; Industry, Service, and Mining. Each one of these companies can report their positive profits and losses for each year. Based on the past data, financial analysts estimated the following distribution of the number of companies in each sector that are likely to report positive profits and losses in the year 2005: Sectors Industry Service Mining Total 240 30 420 Positive Profits 150 50 10 20 80 Losses 200 250 50 500 Total a. What is the probability of randomly selected company (out of these companies listed on Istanbul Stock Exchange) to report positive profits in year 2005? b. What is the probability of a randomly selected company to report either losses in year 2005 or to be from Mining sector? c. What is the probability of a randomly selected company to be either from Industry or from Service sector? a. P( Positive profits) = 420/500 = 0.84 b. P( Losses or Mining) = P( Losses) + P( Mining) P( Losses and Mining) = 80/50 + 50/500 20/500 = 110/500 = 0.22 c. P( Industry or Service) = P(Industry) + P(Service) = 200/500 + 250/500 = 450/500 = 0.9

Types of Events
1. Statistically Independent Events
These are those events whereby the probability of occurrence of an event is not affected by ( or dependent upon) the occurrence of the other event.

Example
Events: A: Izlem goes to Istanbul next Tuesday B: Istanbul stock market decreases by more than 10% next Friday. Therefore: Event A and B are Independent events because the occurrence of event A does not have an impact on the probability of the occurrence of event B.

-9-

Quantitative Analysis MGMT 322

2. Statistically Dependent Events

These are those events whereby the probability of the occurrence of an event is affected by ( or dependent upon) the occurrence of the other event.

Example 1
Events: A: Stock Market Index rises tomorrow B: Interest rates go down today Therefore: Event A and B are Dependent because the occurrence of event B is expected to increase the probability of the event A

Example 2
Events: A: The quantity of automobiles purchased in Turkey rises B: The quantity of petroleum consumed in Turkey rises Therefore: Event A and B are Dependent because the probability of the occurrence of event B is affected by or dependent upon the occurrence of event A, as the more the automobiles Turkish people are purchasing, the more the petroleum they will buy to run their automobiles.

Types of Probabilities under Independence

1. Marginal (Unconditional) Probability
P(A): This is interested in the probability of a single event.

2. Joint Probability
P(AB): This is where you are interested in either the probability of event A and event B happening together at the same time or in succession. P(AB) = P(A) * P(B) P(BA) = P(B) * P(A) A and B are independent events.

3. Conditional Probability
P(A/B): Probability of A, given B! As long as A and B are independent events P(A) is not affected by P(B) so; P(A/B) = P(A) P(B/A) = P(B)

- 10 -

Quantitative Analysis MGMT 322

Example
Events: A: It rains in Cyprus today B: Exchange rate of dollar/yen increases today Therefore: P(A/B) = P(A) PB/A) = P(B)

Types of Probabilities under Dependence

1. Marginal (Unconditional) Probability
P(A) : Is interested in just the probability of a single event.

2. Joint Probability
P(AB) = P(A/B) * P(B) P(BA) = P(B/A) * P(A) P(AB) = P(BA)

3. Conditional Probability
P(A/B) = P(AB) / P(B) P(B/A) = P(BA) / P(A)

Example
Suppose that the eye color of students in a class of 40 people as follows; Male Female Total Black 20 10 30 Green 8 2 10 Total 28 12 40

Events: A: Selecting a Male student. B: Selecting a student with Black Eyes. C: Selecting a Female student. D: Selecting a student with Green Eyes. Q1. If the selected student is a Male student, what is the probability that he has got Green Eyes? P(D/A) = P(DA) / P(A) = (8/40) / (28/40) = 0.074

- 11 -

Quantitative Analysis MGMT 322

Q2. If the selected student is a Female student, what is the probability that she has got Green Eyes. P(D/C) = P(DC) / P(C) = (2/40) / (12/40) = 0.167

Bayes Theorem
Is used for revising prior (before) estimates of probabilities using limited new information to obtain posterior ( new or after) probabilities. Event A gives us P (A); after sometime passes event B occurs which is dependent on event A. We can use Bayes Theorem to obtain posterior probability of A: P (A/B) = P (AB)/ P(B)

Example 1
Suppose we have 100 dices in a basket, and we have two types; half of it is Type1 and the other half is Type2. Given that the probability of getting Ace (1) when you roll a Type1 die is 0.3 and that of rolling Ace (1) of Type2 is 0.6. You are also given the initial probability of Type1 as 0.5 and that of Type2 as 0.5. If you randomly select a die and roll it and you find out that Ace (1) has come up, what is the probability of this die to be a Type1 die? NB: These two events (Ace and Type1 or 2) are dependent because the probability of getting Ace is affected by the occurrence of Type1 or 2. P( Type1/Ace) = P(Type1 Ace) / P(Ace) Note: When you are not given the values for the formula like in the example above, use the following five steps below in order to get the Answer: Step 1: Prepare the following table:
Marginal Probability of Elementary Events Probability of Secondary Event Conditional Probability of Secondary Event, Given Each of the Elementary Event

P(Type 1) = 0.5 P(Ace) P(Type 2) = 0.5

P(Ace / Type 1) = 0.3 P(Ace / Type 2) = 0.6

Step 2: Computation of probability of secondary event: P(secondary event) = summation of its joint probabilities of each one of the elementary events.

- 12 -

Quantitative Analysis MGMT 322

P(Ace) = P(Type1, Ace) + P( Type2, Ace) Step 3: Computation of joint probability of secondary events with each of the elementary events: P(Type 1, Ace) = P(Ace/Type1)* P(Type 1) = 0.3 * 0.5 = 0.15 P(Type 2, Ace) = P(Ace/Type2) * P(Type2) = 0.6 * 0.5 = 0.30 P(Ace) = 0.15 + 0.30 = 0.45 Step 4: Use the Bayes Theorem to get the final answer P(Type1/ Ace) = P(Type 1, Ace) / P(Ace) = 0.15 / 0.45 = 1/3 This is posterior probability

Example 2
Economists believe that the annual inflation rate in Turkey is affected by the changes in the petroleum prices. The probability of inflation rate increasing in 2001 is estimated to be 0.60. The probability of petroleum prices rising is 0.40. The probability of both inflation rate and the petroleum prices rising is 0.35. On the other hand, the probability of inflation rate rising and at the same time, the petroleum prices not rising is estimated to be 0.20. If petroleum prices do not rise in 2001, what is the probability of inflation rate rising? Events: I: Inflation rate increasing. P: Petroleum prices rising. N: Petroleum prices not rising

Given: P(I) = 0.6 P(P) = 0.4 P(IP) = 0.35 P(IN) = 0.2 Solution: P(N) = 1 P(P) = 1- 0.4 = 0.6 P(I / N) = P(IN) / P(N) = 0.2 / 0.6 = 0.333 0.3

- 13 -

Quantitative Analysis MGMT 322

Example2:
Annual profits of construction sector depend on wage rate paid to labor employed by the sector. The past data about the behavior of annual profits suggested that the probability of profits rising in a given year is 0.80 and the probability of profits falling is 0.20. The past data also suggest that in 60% of all the years during which the profits have increased, wage rate has declined, and it has increased in the remaining 40%. However, in 80% of all years during which profits have fallen, wage rate has increased, and the only remaining 20% wage rate has decreased. Due to the economic crisis in Turkey, wage rate has been declining since the beginning of 2005. Given this downward trend in wages, what is our revised estimate for the probability of annual profits of construction sector to rise in the year? Events: A: Profits rising B: Profits falling C: Wage rate declining D: Wage rate increasing Given: P(A) = 0.80 P(B) = 0.20 P(C/A) = 0.60 P(D/A) = 0.40 P(D/B) = 0.80 P(C/B) = 0.20 Solution: Without using the table P(A/C) = P(AC) / P(C) or P(CA) / P(A) 0.60 = P(CA) / 0.80 P(CA) or P(AC) = 0.60 * 0.80 = 0.48 P(C) = P(CA) + P(CB) = [P(C/A) * P(A)] + [P(C/B) * P(B)] = (0.60 * 0.80) + (0.20 * 0.20) = 0.48 + 0.04 = 0.52 P(A/C) = 0.48 / 0.52 = 0.92

Example 3
A child chosen at random in a community school system comes from low-income family 15% of the time. Children from low-income families in the community graduate from college only 20% of the time. Children not from low-income families have a 40% chance of graduating from college. As an employer of people from this community, you are reviewing applicants and note that the first one had a college degree. What is the probability that the person comes from a low-income family?

- 14 -

Quantitative Analysis MGMT 322

Events: L: Child coming from a low-income family G: Graduate from college H: Child coming from not a low-income family Given: P(L) = 0.15 P(H) = 1- P(H) = 1- 0.15 = 0.85 P(G/L) = 0.20 P(G/H) = 0.40 P(L/G) = ? Solution: P(L/G) = P(LG) / P(G) P(G/L) * P(L) / P(G) P(G) = [P(G/L) * P(L)] + [P(G/H) * P(H)] = (0.20 * 0.15) + (0.40 * 0.85) = 0.03 + 0.34 = 0.37 P(L/G) = 0.03 / 0.37 = 0.08 = 8%

Probability Distribution
Random Variables
A Random Variable is a variable which takes on values unpredictably as a result of a random process. EG: Events: X: Age of a randomly selected student Y: The value of stock market Index in Istanbul by the end of todays session. Z: Interest rates on a new Bond issued by the Turkish Government that will take place next week. DISCRETE VERSUS CONTINUOS RANDOM VARIABLES

1. Discrete Random Variable

This is allowed to take on only limited number of possible values. E.g: X = Age of a randomly selected student. a. Binomial Distribution b. Poisson Probability Distribution

2. Continuous Random Variable

This is allowed to take any value over a certain range. They are infinite numbers of possibilities. a. Normal Probability Distribution - 15 -

Quantitative Analysis MGMT 322

b. Exponential Probability Distribution

E.g: Possible values of X: X1 X2 X3 X4 X5 X6 20 21 22 23 24 26

Probability Distribution of a Random Variable

Is the listing of probabilities that correspond to each possible value of that random variable. Two Steps to Obtain the Probability Distribution of X: Step 1 Obtain all the possible values that X can take on. Step2 Obtain the probability of the occurrence of each possible value of x. Then, present the results in a form of a Table or a Graph. E.g: Age of a randomly selected student Age 20 21 22 23 24 26 Total Number of Students 1 5 7 4 3 1 21

Step 1: Possible X values:

- 16 -

Quantitative Analysis MGMT 322

X1 X2 X3 X3 X5 X6

20 21 22 23 24 26

Step 2: Obtain the probability of each possible X values x 20 21 22 23 24 26 P(x) 1/21 5/21 7/21 4/21 3/21 1/21

Expected Value of a random variable

Its donated as; E(X) It is the weighted value of all the possible values of x.

E(X) for a Discrete Random Variable

E(X) = P(Xi)Xi P(Xi) = Probability of occurrence of the ith value E.g: E(X) = (20)(1/21) + (21)(5/21) + (22)(7/21) + (23)(4/21) + (24)(3/21) + (26)(1/21) = 0.95 + 5 + 7.33 + 4.38 + 3.43 + 1.24 22 years

- 17 -

Quantitative Analysis MGMT 322

E(X) for Continuous Random Variable

E(X) = P(X) dx

Example
Suppose that the average growth rates of GDP of countries in the world are given below: Average Growth Rate 8% 6% 4% 2% Total Number of Countries 20 30 50 1oo 200

What is the Expected (average) growth rate of a randomly selected country in the world? Solution: E(Growth Rate) = (0.08)(20/200) + (0.06)(30/200) + (0.04)(50/200) + (0.02)(100/200) = 0.008 + 0.009 + 0.01 + 0.01 = 0.037 = 3.7%

Binomial Probability Distribution

It is the probability distribution of a binomial random variable which takes on values as a result of Bernoulli Process. It is a type of Discrete Probability Distribution.

Characteristics of a Bernoulli Type Process

i. ii. iii. iv. Its made up fixed number of trials of the same activity. Each trial has two out comes only. (Success or Failure) The probability of success on each trial remains fixed over all trials. Each trial is statistically independent.

Binomial Formula:
P(r) = N! p r q nr r!(n r )!

- 18 -

Quantitative Analysis MGMT 322

n: Number of trials. r: Number of success outcomes. p: Probability of success outcomes. q = (1 p) : probability of failure. P(r) = probability of getting exactly r (successes) out of n trials. Kind of Questions that can be asked and answered using Binomial Formula: Q1: Obtain the probability distribution of a Binomial random Variable X Q2: What is the probability of getting at least r* successes out on n trials; P(r r*) Q3: What is the probability of getting at most r* successes out of n trials; P(r r*) Q4: What is the probability of getting exactly r* successes; P(r*) General Answers: Q1:Using the Binomial Formula, find the probability distribution of X X X1 X2 X3 Xn P(X) P(X1) P(X2) P(X3) P(Xn)

Q2: P(r r*) = P(r*) + P(r* + 1) + P(n) Q3: P(r r*) = P(0) + P(1) + .P(r*) Q4: Use the formula

Example 1
Suppose that we toss a coin 4 times. Q1. What is the probability of getting exactly 2 Heads? Q2. What is the probability of getting at least 3 Heads? Q3. What is the probability of getting at most 3 Heads? Q4. Using the binomial formula obtain the binomial distribution of the possible number of tails have may come up. Solution: Q1. n = 4 p = 1/2 (Head) q = 1/2 (Tail) 4! P(2) = 2!(4 - 2)! (1/2)2(1/2)4-2 = 0.37

- 19 -

Quantitative Analysis MGMT 322

Q2. P(r3) = P(3) + P(4) 4! P(3) = 3!(4 - 3)! (1/2)3(1/2)4-3 = 0.25

4! P(4) = 4!(4 - 4)! P(r3) = 0.313 Q3. P(r 3) = P(0) + P(1) + P(2) + P(3) 4! P(0) = 0!(4 - 0)! (1/2)0(1/2)4-0 = 0.063 (1/2)4(1/2)4-4 = 0.063

4! P(1) = 1!(4 - 1)! P(r 3) = 0.938 Q4. p: probability of getting Tail q: probability of getting Head. P(0) = 1/16 P(1) = 4/16 P(2) = 6/16 P(3) = 4/16 P(4) = 1/16 (1/2)1(1/2)4-1 = 0.25

Example 2:
Suppose there are 2 outcomes for inflation in Turkey for each year: High or Low. If inflation rate is greater than 100%, inflation is high. If it is less than 100%, inflation is low for that year. Sabanci Corp. makes $50 million of positive profits for each year of inflation and loses $20 million for each year of low inflation. Economists estimated that the probability of high inflation is 80% for each of the next 4 years.

- 20 -

Quantitative Analysis MGMT 322

a) What is the probability of getting exactly 2 years of high inflation in the next 4 years? n= 4 r= 2 p=0.8 q=0.2 P (2) = 4! / 2! (4-2)! *[(0.8)2 * (0.2)4-2] = 0.1536 b) What is the prob. of getting at least 3 years of HI over the next 4 years? P (r 3) = P (3) + P (4) = 0.4096 + 0.4096 = 0.8192 P (3) = 4! / 3! (4-3)! * [(0.8)3 * (0.2)1] = 0.4096 P (3) = 4! / 4! (4-4)! * [(0.8)4 * (0.2)0] = 0.4096 c) What is the probability of getting at most 3 years of LI? p = 0.2 q = 0.8 P (r 3) = P (0) + P (1) + P (2) +P (3) = 0.0016 + 0.0256 + 0.1536 + 0.4096 = 0.5904 P (0) = 0.0016 P (1) = 0.0256 P (2) = 0.1536 P (3) = 0.4096 d) What is the expected net profit for Sabanci over the next 4 years? X1 X2 X3 X4 X5 X All 4 years of HI 4*50 = 200 3 years of HI and 1 year of LI [(3X50) (1*20)] = 130 2 years of HI and 2 years of LI (2*50) (2*20) = 60 1 year of HI and 3 years LI (1*50) (3* 20) = -10 0 year of HI and 4 years of LI (0*50) (4*20) = -80 P (X) P (200) = P (4) P (130) = P (3) P (60) = P (2) P (-10) = P (1) P (-80) = P (0)

- 21 -

Quantitative Analysis MGMT 322

Example 3
Prof. TT has put $2 on each of the 3 horses running three different races in Istanbul Horse Racing center. He feels that each of the bets he has made has a 0.2 probability of winning. A winning ticket on any of the three horses will earn Prof. TT $40. Assuming a Bernoulli process, answer the following questions: a. What is the probability of at least 2 horses ( on which Prof. TT ha s bet) winning? b. What is the Expected Earnings from all the three horses? c. What is the probability of at most 1 out of the three horses (on which Prof. TT has bet) winning? d. What is the probability distribution of his possible earnings from all the three races? Solution: n= 3 p = 0.2 q = 1- 0.2 = 0.8 x = number of horses P(0) = 3! (0.2)0 (0.8)30 = 0.512 0!(3 0 )! 3! (0.2)1 (0.8)31 = 0.384 1!(3 1)!

P(1) =

P(2) =

3! (0.2)2 (0.8)32 = 0.096 2!(3 2)! 3! (0.2)3 (0.8)33 = 0.008 3!(3 3)!
Profit (0*40) 6 = -6 (1*40) 6 = 34 (2*40) 6 = 74 (3*40) 6 = 114 P(X) 0.512 0.384 0.096 0.008

P(3) =
X 0 1 2 3

a. P(r 2) = P(2) + P(3) = 0.096 + 0.008 = 0.104 b. E(X) = -6(0.512) + 34(0.384) + 74(0.096) + 114(0.008) = -3.072 + 13.056 + 7.104 + 0.912 = $18 c. P(r 1) = P(0) + P(1) = 0.512 + 0.384 = 0.896

- 22 -

Quantitative Analysis MGMT 322

d.

X Profit 0 -6 1 34 2 74 3 114 216 Total

P(X) -6/216 34/216 74/216 114/216

Example 4
Suppose that the inflationary process in Turkey is a Bernoulli process over the next 5 years. The probability of inflation rate to be high (meaning that it exceeds 100%) for each year over the next 5 years is estimated to be . Note that the low inflation corresponds to the case when the annual inflation rate is less than or equal to 100%. If the company XYZ earns approximately $10 million in a typical year of high inflation, and loses $5 million in a typical year of low inflation, what is the expected profit of XYZ for the next 5 years?
Solution: N = 5 years High Inflation Years Profits 0 (0*10) (5*5) = -25 1 (1*10) (4*5) = -10 2 (2*10) (3*5) = 5 3 (3*10) (2*5) = 20 4 (4*10) (1*5) = 35 5 (5*10) (0*5) = 50

Prob(x) 0.0001 0.0147 0.0879 0.2637 0.3955 0.2373

E(profits) = (-25) * (0.0001) + (-10) * (0.0147) + (5) * (0.0879)+(20)*(0.2637)+(35)*(0.3955)+(50)*(0.2373) = $31.250 Million

Normal Probability Distribution

This is a type of continuous probability distribution and its range extends from negative infinity to positive infinity.

- 23 -

Quantitative Analysis MGMT 322

Characteristics
1. Bell shaped probability curve. 2. It has a single peak 3. Mean () lies at the center and the distribution is symmetrical at the vertical line erected at . 4. Two tails extend indefinitely.

Two Parameters: Mean () & Standard Deviation ()

: Located in the center of the population and it is the average of the data. : It gives information about relative spread of data around the mean.

Application of Normal Distribution 3 Mathematical Facts about Normal Distribution

- 24 -

Quantitative Analysis MGMT 322

1) Approximately 68% of the population lies in the interval ranging 1 below the to 1 . 2) 95% of the population lies in an interval ranging 2 below the to 2 above the 2 . 3) 99% of the population lies in an interval ranging 3 below the to 3 above the 2 .

X~N (50, 5)

= x/ N = $ 50, = 5
Z ~ N (0,1)

Example 1:
Suppose that the stock price of Corp. X (s) is distributed normally with a of 100 TL and of 10 TL?

- 25 -

Quantitative Analysis MGMT 322

a) What is the probability of s to be more than 120 TL in any day? b) What is the probability of s to be less than 90 TL in any day? a)

z = 120 100/ 10 = 2 P (s > 120) = P (z > 2) = 1 P (z < 2) = 1 0.97725 = 0.02275

b)

z = 90- 100 / 10 = - 1 P (s < 90) = P (z <-1) = 1- P (z > 1) = 1 0.84134 = 0.15866

- 26 -

Quantitative Analysis MGMT 322

Example 2
Prof. TT believes that the annual profit of Turkish bank is a normal random variable with a mean of $600,000 and a standard deviation of $100,000. Prof. TT is currently analyzing those banks whose annual profit volume lies between $500,000 and $650,000. a. If total number of bank is 270, what is the approximate number of bank that Prof.TT will analyze? b. If Prof. TT randomly selects a bank to analyze, what is the probability that this bank will have an annual profit volume of more than $400,000?
Solution: = $600,000 = $100,000

X
500,000 600,000

Y
650,000 $

a. P(500,000 $ 650,000) = (z1 z z2) z1 = 500,000 600,000 / 100,000 = -1

z2 = 650,000 600,000 / 100,000 = 0.5 Area X = P(z +1) 0.5 = 0.84134 0.5 = 0.34134 Area Y = P(z 0.5) 0.5 = 0.69146 0.5 = 0.19146 Area X + Area Y = 0.34134 + 0.19146 = 0.5328 The Approximate number of Banks = 270*0.5328 = 143.856 144 b. P($ > 400,000)

- 27 -

Quantitative Analysis MGMT 322

400,000

600,000

Z* = 400,000 600,000 / 100,000 = -200,000 / 100,000 = -2 P(z +2) = 0.97725

Example 3
A financial analyst computed the return on stockholders equity for all the companies listed on the New York Stock Exchange. She found that the mean of this distribution was 10%, with a standard deviation of 5%. She is interested in examining further those companies whose return on stockholders equity is between 16% and 22%. Of the approximately 1,300 companies listed on the exchange, how many are of interest to her?
Solution: = 10% = 5% n = 1,300 (Total Population)

10%

16% 22%

Z1 = 16% - 10% / 5% = 1.2 Z2 = 22% - 10% / 5% = 2.4 P(1.2 Z 2.4) For: Z2 = P(Z 2.4) = 0.99180 Z1 = P( Z 1.2) = 0.88493 Number of Companies of interest to her = (0.99180 0.88493) * 1,300 = 0.10687 * 1300 = 139

- 28 -

Quantitative Analysis MGMT 322

CHAPTER 3: FORECASTING
Definition of Forecasting
It is the prediction of the future behavior of a variable. E.g. stock prices, inflation rates, interest rates, exchange rates, profits, gold prices, land prices, etc. The most critical variable that companies are interested to forecast is Demand (sales).Usually the companies are interested to forecast monthly sales (next month), or quarterly sales (next quarter).

Importance of Sales Forecasting

There are many decisions that critically depend on sales forecasts and these include: 1) Input ordering decisions: Depends on how much working capital and other inputs you will need. 2) Production decisions: Depends on how much you will produce in a certain time period, that is, it determines the output. 3) The pricing decisions: This depends on whether the market is monopolistic or oligopolistic. 4) Advertising decisions: Depends on how much you will spend on advertising. 5) Inventory decisions: This is how much you should maintain or change depending on demand (sales) forecast.

The Methodology of Forecasting

1) Our Methodology (Technical analysis) This methodology uses past sales data only to predict (forecast) future sales. The limitation of this method is that it only depends on past data to predict the future sales.

Formula: Ft+1

= f (Xt, X t-1, X t-n )

- 29 -

Quantitative Analysis MGMT 322

2) The Econometric Methodology (Fundamental Analysis) This methodology is made up of statistical tools, like Economic theory and Regression Analysis. It also uses past data about all the variables below to forecast future demand. Formula: QDx = f (Px, Ps, Pc, I, Expectations, Population) 1) BASIC STEPS IN FORECASTING( Technical Analysis) Step 1: Determine the type of the product based on its sales pattern. Step 2: Determine the forecasting models that can be used for this type of product. Step 3: Apply each model on the actual past data and obtain MSE of the model. Best model is the one which yields smallest MSE (Mean Square Error) Step 4: Use the BEST MODEL to obtain a forecast for the future.

MSE test is the test of forecasting accuracy MSE =

et2/ n

- 30 -

Quantitative Analysis MGMT 322

TYPES OF PRODUCTS BASED ON THEIR SALES PATTERN How To Determine The Type Of Product: First collect past actual sales data (monthly/ quarterly) and plot it against time and see the pattern.

The following two are the possibilities: NON- SEASONAL a) Constant level b) Linear Trend c) Exponential Smoothing d) Damped Trend SEASONAL a) Constant Level b) Linear Trend c) Exponential Trend d) Damped Trend

Forecasting Models
FOR NON-SEASONAL PRODUCTS WITH CONSTANT LEVEL TYPE OF PATTERN
1) Nave Model 2) Moving Averages (M.A) 3) Simple Exponential Smoothing

FOR NON-SEASONAL PRODUCTS WITH LINEAR TREND TYPE OF PATTERN

1) Nave Model 2) Time Series Regression 3) Smoothing Linear Trend (A Linear Exponential Smoothing)

a) Constant- level (Non-seasonal): It has a fixed amount of sales.

- 31 -

Quantitative Analysis MGMT 322

b) Constant-level (Seasonal): It has fixed amount of sales but with regular ups and downs.

- 32 -

Quantitative Analysis MGMT 322

c) Linear Trend (Non-Seasonal): General increase of sales over time.

d) Linear Trend (Seasonal): Sales increase at a constant rate.

- 33 -

Quantitative Analysis MGMT 322

e) Exponential Trend (Non-Seasonal): Sales increase over time at an increasing rate.

f) Exponential Trend (Seasonal): Sales increase over time with regular ups and downs.

- 34 -

Quantitative Analysis MGMT 322

g) Damped Trend (Non-Seasonal): Sales increase over time at a decreasing rate.

h) Damped Trend (Seasonal):

- 35 -

Quantitative Analysis MGMT 322

Has regular ups and downs.

Ft+1 = Xt F April = X March et =X t - Ft F2nd Quarter 2002 = X1st Quarter 2002

Steps in Computing The Mean Squared Error (MSE) Steps 1: Obtain as many data as possible Steps 2: If you have enough data, meaning that if you have at least 12 units of data in case of monthly data or at least 8 units of data in case of quarterly data divide the data set into 2 halves, the 1st half is called the Warm-Up Sample and 2nd half is called the Forecasting Sample. Steps 3: Apply the model from the very beginning and obtain forecasts and errors for each time period. a) If you have enough data, apply each model on all the data and obtain Forecasting Errors, but use only the forecasting errors coming from the Forecasting Sample ( the 2nd half of the data) to obtain the MSE. b) If you do not have enough, use all the available or given data as forecasting sample and use all the forecasting errors coming from this sample to obtain MSE. Steps 5: Best model is the one which yields smallest MSE. (Use the BEST MODEL to obtain a forecast for the future.)

- 36 -

Quantitative Analysis MGMT 322

Application of Nave Model

Formula: Ft + 1 = Xt Ft = Forecast for next period Xt = Actual sales in the current period. Et = Xt Ft Et = Forecasting error for period t Xt = Actual sales in period t Ft = Forecasts in period t

E.g. If FJan, 2005 = 5,500 XJan, 2005 = 4,000 EJan, 2005 = 4,000 5,000 = -1,500

Example 1
Suppose that we are given the following past sales data of Televisions in North Cyprus, obtain the forecast of sales for the month of January, 2003 using the Nave Model and find the MSE of this model:. WARM-UP SAMPLE Month 2002 Jan. Feb. March April May June July August Sept. Oct. Nov. Dec. 2003 Jan. T 1 2 3 4 5 6 7 8 9 10 11 12 13 Xt 28 27 33 25 34 33 35 30 33 35 27 29 Ft 28 27 33 25 34 33 35 30 33 35 27 29 et -1 +6 -8 +9 -1 +2 -5 +3 +2 -8 +2

FORECASTING SAMPLE

- 37 -

Quantitative Analysis MGMT 322

Solution: Nave Models Formula = Ft + 1 = Xt NB: There is no forecast and no error for the first month (F1 and e1), but we obtain the forecasts and the errors for the next and the following months (periods) according to the Nave Model. Thus the forecast for the next period (month) is the forecast for the previous period.

F2 = 28, e2 = X2 F2 = 27 28 = -1 F3 = 27, e3 = X3 F3 = 33 27 = +6 F4 = 33, e4 = X4 F4 = 25 33 = -8 Follow the same procedure up to F12. a) Forecasts for January, 2003 (F13) = 29 b) MSE: Since we have enough data, we are going to take only the forecasting errors from the forecasting sample. Therefore:
MSE =

e
t =7

12

2 t

( 2) 2 + ( 5) 2 + (3) 2 + ( 2) 2 + ( 8) 2 + ( 2) 2 6

= 18.3

Example 2
Suppose that you are the manager of HSBC Bank. You believe that the pattern of quarterly amount of savings deposits is non-seasonal and exhibits approximately a constant-level type of behavior. You need a forecast for the savings deposits at your bank for the 1st quarter of 2005 in order to make better decision about the portfolio investment and interest rate policies of your bank. Using the Nave Model, obtain a forecast for the amount of savings deposits for the 1st quarter of 2005 and obtain the MSE of the model, given the following past quarterly data about the amount of savings deposits ( in million of dollars) of your bank:
Year 2003 2003 2003 2004 2004 2004 2004 2005 T 1 2 3 4 5 6 7 8 Quarter 2nd 3rd 4th 1st 2nd 3rd 4th 1st Deposits 120 100 110 130 90 110 100 Ft et

120 100 110 130 90 110 100

-20 +10 +20 -40 +20 -10

- 38 -

Quantitative Analysis MGMT 322

Solution: F2 = 120, e2 = 100 120 = -20 F3 = 100, e3 = 110 100 = +10 Follow the same procedure up to F7.

a) Forecasts for 1st quarter, 2005 (F8) = 100 b) MSE: Since we do not have enough data, use all the forecasting errors to obtain the MSE. (20) 2 + (10) 2 + (20) 2 + (40) 2 + (20) 2 + (10) 2 MSE = 6 = 500

Application of Moving Averages Model

Unweighted M.A Model 2-period (Un.W) M.A 3-period (Un.W) M.A Weighted M.A Model 2-period (W) M.A 3-period (W) M.A

2- Period (Unw) M.A Model: Ft+1 = Xt+ Xt-1/ 2 3- Period (Unw) M.A Model: Ft+1 = Xt + Xt-1 + Xt-2 / 3 2- Period (W) M.A Model: Ft+1 = wt Xt + wt-1 Xt-1 3- Period (W) Ft+1 = wt Xt + wt-1 Xt-1 + wt-2 Xt-2 wt + wt-1 + wt-2 = 1 wt >wt-1 > wt-2

- 39 -

Quantitative Analysis MGMT 322

Example 1
From the data given below, find the forecast for F13 using 3-period (Unw) MA Model and obtain the MSE of this model.
WARM-UPSAMPLE

t 1 2 3 4 5 6 7 8 9 10 11 12 13

Xt 28 27 33 25 34 33 35 30 33 35 27 29

Ft

et

29.3 28.3 30.7 30.7 34 32.7 32.7 32.7 31.7 30.3

-4.3 +5.7 +2.3 +4.3 -4 +0.3 +2.3 -5.7 -2.7

FORECASTING SAMPLE Solution: 3-Period (Unw) MA Models Formula: Ft + 1 = Xt + Xt-1 + Xt 2 / 3

F4 = 33 + 27 + 28 / 3 = 29.3, F5 = 25 + 33 + 27 / 3 = 28.3,
F13 = 29 +27 +35 / 3 = 30.3

e4 = 25 29.3 = -4.3 e5 = 34 28.3 = +5.7

MSE =

e
t =7

12

2 t

( 4.3) 2 + ( 4) 2 + (0.3) 2 + ( 2.3) 2 + ( 5.7 ) 2 + ( 2.7 ) 2 = 6 = 13.3

- 40 -

Quantitative Analysis MGMT 322

Example 2
Given the Dow-Jones Index of stock market prices as below, use a 4-period (W) MA model to find the MSE of the model using the following weights:0.5, 0.3, 0.1 and 0.1 respectively.
Month t Dow-Jones Index (Xt) 820 840 857 823 784 793 817 820 826 824 830 842 840 839 Ft et

May June July August September October November December January February March April May June
14

1 2 3 4 5 6 7 8 9 10 11 12 13 14

835 809 800 807 813 820 824 827 835 838

-51 -16 17 13 13 4 6 15 5 1

MSE:

et
t =8

(13) 2 + (13) 2 + (4) 2 + (6) 2 + (15) 2 + (5) 2 + (1) 2 = 91.6 7

Application of Simple Exponential Model

Formula: Ft+1 = Ft + et

0 < <1

= smoothing parameter

To apply this model, we need; 1) A value for F1 2) 2) The best value of Note = Possible values of = 0.1,0.2, 0.3,., 0.9

- 41 -

Quantitative Analysis MGMT 322

Important Steps
Step 1: To obtain F1, we take the mean (average) of the warm-up sample. If we do not have enough data, we take the MEAN of all data as F1. Step 2: To find the Best Value of , we apply the model with each possible value of on the warm-up sample (or on all of the data in case we do not have enough data) and obtain MSE for each case. Best value of is the one which yields smallest MSE. Step 3: Then apply the model with the best value of on the forecasting sample to get MSE of the model and obtain Forecast for the future.

If we did not have enough data, smallest data MSE obtained in step 2 becomes the MSE of the model.

Example 1
Given the following sales data, find the forecast for F13 using Simple exponential Smoothing Model and obtain the MSE of the model.
WARM-UPSAMPLE

t 1 2 3 4 5 6 7 8 9 10 11 12 13
Forecasting As Sample

Xt 28 27 33 25 34 33 35 30 33 35 27 29

Ft 30 29.8 29.5 29.9 29.4 29.9 30.2 30.7 30.6 30.8 31.2 30.8 30.6

et -2 -2.8 3.5 -4.9 4.6 3.1 4.8 -0.7 2.4 4.2 -4.2 -1.8

Assume = 0.1
F1 = 28 + 27 + 33 + 25 + 34 + 33 / 6 = 30 Ft + 1 = Ft + et F2 = F1 + et1 = 30 + 0.1(-2) = 29.8

- 42 -

Quantitative Analysis MGMT 322

F3 = F2 + et2 = 29.8 + 0.1(-2.8) = 29.5

F13 = F12 + et12 = 30.8 + 0.1(-1.8) = 30.6

MSE =

e
t =7

12

2 t

( 4.9) 2 + ( 0.7 ) 2 + ( 2.4) 2 + ( 4.2) 2 + ( 4.2) 2 + ( 1.8) 2 = = 11 .3 6

Example 2
You are the production manager of Arcelik Corporation in Turkey which specialized in production of TV sets among other kinds of durables. You need a forecast for next quarter in order the plan production process and determine working capital needs. Apply Simple Exponential Smoothing Model to obtain the forecast for the 2nd Qtr of 2005. What is the MSE of the model? Assume that the best value of = 0.5 The past actual sales data about TV sales are given below:
Year 2004 2004 2004 2004 2005 2005 Qtr 1 2 3 4 1 2 Data 5000 4800 5200 5100 4900 t 1 2 3 4 5 6 Ft F1 = 5000 F2 = 5000 F3 = 4900 F4 = 5050 F5 = 5075 F6 = 4987.5 Et 0 -200 300 50 -175

Solution: F1 = 5000 + 4800 + 5200 + 5100 + 4900 / 5 = 5000 F2 = 5000 + 0.5(0) = 5000 F3 = 5000 + 0.5(-200) = 4900 F4 = 4900 + 0.5(300) = 5050 F5 = 5050 + 0.5(50) = 5075 F6 = 5075 + 0.5(-175) = 4987.5

MSE =

et
t =1

(0) 2 + (200) 2 + (300) 2 (50) 2 (175) 2 = 32625 5

- 43 -

Quantitative Analysis MGMT 322

Application of Time Series Regression Model

Salest = a + b (time).Ft = a +b Sales is dependent variable while Time is an independent variable.
Least-Square Formulas
b=

tx nt x t n(t )
2

a = x bt

b= slope of the linear trend line a= intercept of the linear trend line X= values of the dependent variable (sales) t = values of the independent variable (time)
Note: You can start the values of t from any number as long as they are consecutive.

X = mean of the X values t = mean of the t values

Important Points
1) If we have enough data, we use the warm-up sample data for X and t in leastsquares formulas to get a and b. If we do not have enough data, we use all the data to get a and b. 2) First we obtain b and then a. 3) After we get Ft = a + bt, we apply this equation on the Forecasting Sample to get MSE of the model. Obtain forecasts for the future periods.

Example
Given the following sales data, find forecasts for January and February, 2002 (F13 and F14) using Time-Series Regression Model and obtain the MSE of this model.

- 44 -

Quantitative Analysis MGMT 322

WARM-UP SAMPLE

Months 2001 Jan. Feb. March April May June July August Sept. Oct. Nov. Dec. 2002 Jan.
FORECASTING SAMPLE

T 1 2 3 4 5 6 7 8 9 10 11 12 13

Xt 60 55 64 51 69 66 83 90 76 95 72 88

Ft

et

66.8 68.5 70.2 71.9 73.6 75.3 77

16.2 21.5 5.8 23.1 -1.6 12.7

Formula for Time Series Regression Model: Ft = a + bt

b=

tx nt x t nt
2 2

tx = (1*60) + (2*55) +( 3*64) + (4*51) +( 5*69) + (6*66) = 1307 t = 1 2 3 4 5 +6 = 91

2 2+ 2+ 2+ 2+ 2 2

t- = t/ n = 1+2+3+4+5+6 / 6 = 3.5 X- = 60+55+64+51+69+66 / 6 = 60.8 b = 1.7 a = X- - bt- = 60.8 (1.7)(3.5) = 54.9

Therefore: F1 = 54.9 + 1.7(1) = 56.6 F7 = 54.9 + 1.7 (7) = 66.8 F13 = 54.9 + 1.7(13) = 77 F14 = 54.9 + 1.7(14) = 78.7

- 45 -

Quantitative Analysis MGMT 322

MSE =

e
t=7

12

2 t

(16.2) 2 + (215 . ) 2 + (58 . ) 2 + (231 . ) 2 + ( 16 . ) 2 + (12.7) 2 = 6 = 242.6

Application of Smoothing Linear Trend Model (Linear Exponential Smoothing)

1) Ft+1 = St + Tt 2) St=Ft + 1et 3) Tt = Tt-1 + 2 et

Note = 0<1<1, 0< 2<0.1 1, 2 = smoothing parameters 1 = 0.1, 0.2, 0.3,, 0.9 2 = 0.01, 0.02, 0.03, 0.09 St = smoothed level at the end of period Tt = smoothed Trend at the end of period t+1

Important Steps
Step 1: To get F1, we use least-squares formulas on warm-up sample. If we have enough data and get a and b which have used as estimates for S0 and T0 to get

F1 = S0 + T0 = a+b If we do not have enough data, we obtain a and b using all of the data.
Step 3: Then we apply model with best (1, 2) values on the forecasting sample to get MSE of the model and obtain forecast for the next period. Note that if we did not have enough data smallest MSE obtained in Step 2 becomes the MSE of the model.

Example
Given the following sales data, find F13 and obtain MSE of this model.

- 46 -

Quantitative Analysis MGMT 322

Months 2001 Jan. Feb. March April May June July August Sept. Oct. Nov. Dec. 2002 Jan.

T 1 2 3 4 5 6 7 8 9 10 11 12 13

Xt 60 55 64 51 69 66 83 90 76 95 72 88

Ft 56.6 58.6 59.9 62 62.5 64.9 66.5 70.1 74.2 76.5 80.7 84 86.8

et 3.4 -3.6 4.1 -11 6.5 -1.1 16.5 19.9 1.8 18.5 8.7 4

Assume 1 = 0.1 and 2 = 0.01

F1 = S0 + T0 = a +b = 54.9 + 1.7 = 56.6 F2 = S1 + T1 = 56.9 + 1.7 = 58.6 S1 = F1 + 1 (e1) = 56.6 + 0.1 (3.4) = 56.94=56.9 T1 = T0 + 2 (e1) = 1.7 + 0.01(3.4) = 1.734=1.7 F13 = S12 + T12 S12 = F12 + 1 (e12) = 84 + 0.1(4) = 84.4 T12 = T11 + 2 (e12) = 2.4 + 0.01(4) = 2.4
F13 = 84.4 + 2.4 = 86.8 MSE = et 2 =
t =7 12

(16.5) 2 + (19.9) 2 + (1.8) 2 + (18.5) 2 + (8.7) 2 + (4) 2 6 = 184.2

- 47 -

Quantitative Analysis MGMT 322

CHAPTER 4: DECISION MAKING UNDER UNCERTAINITY AND RISK BASIC STEPS IN DECISION MAKING
Step 1: List the feasible Decision Alternatives Step 2: List the states of nature. States of nature are the most important (critical factors) future events that are not under the control of decision-maker but which arte likely to affect the payoffs (benefits) that can be obtained from the decision alternatives. Step 3: Calculate the payoffs of the matrix. Pay-off table shows the amount of pay-offs that can be obtained for each possible combination of state of nature and decision alternatives.

Alternative Decision Making Environments

Certainty: A manager (decision-maker) knows for sure the state of nature which will take place. Uncertainty: This is where the decision-maker knows only the possible states of nature. Risk: This is where the decision-maker knows also the probability of each state of nature.

Alternative Criteria for Decision Making under Uncertainty

Example
Problem: We have a Tape and CD manufacturing company. It is facing increased demand beyond its capacity, what is the best (optimal) way of accommodating this increasing demand or what is the optimal way of increasing supply to the market? Solution: Step 1: List the feasible decision alternatives: A Expand the Plant B Build a new plant C Subcontract out extra production to the other companies

- 48 -

Quantitative Analysis MGMT 322

Step 2: List the states of nature to this problem: 1. High Demand e.g 2million units of sales volume each year. 2. Moderate Demand e.g 1million units of sales volume each year. 3. Low Demand e.g million units of sales volume each year. 4. Failure e.g almost zero sales. Step 3: Prepare the Pay-off Table or pay-off Matrix: Payoff Matrix States of Nature High Demand Moderate Demand Low Demand Failure

Expand $ 500,000 $ 250,000 - $ 250,000 - $ 450,000

Decision Alternatives Build $ 700,000 $ 300,000 - $ 400,000 - $ 800,000

Subcontract $ 300,000 $ 150,000 - $ 10,000 - $ 100,000

N.B: This company measures pay-offs as simple profits over the next five years.

Pay-off Table and Decision Making Criteria

1) The Maxi-Max Criterion
This criterion is preferred by optimistic managers.
Step1: Identify the maximum payoff that can be obtained for each decision alternative.

Max-pay off

Expand $ 500,000

Build $ 700,000

Subcontract $ 300,000

Step2: The optimum choice is the alternative which yields the maximum payoff in this list obtained in step 1. The optimum choice is to build a new plant. Because; $ 700,000>$ 500,000>$ 300,000

2) The Maxi-Min Criterion

This criterion is preferred by pessimistic managers.
Step1: Identify the minimum payoff that can be obtained for each decision alternative.

Min-pay off

Expand - $ 450,000

Build - $ 800,000

Subcontract - $ 100,000

- 49 -

Quantitative Analysis MGMT 322

Step2: The optimum choice is the alternative which yields the maximum payoff in this list obtained in step 1. The optimum choice is to Subcontract. Because; -$ 100,000>-$ 450,000>-$ 800,000

3) The Criterion of Realism

Step1: Choose a value of such that 0<<1, is an index of your optimism about future. Relatively large value of represents higher degree of optimism about the most optimistic state of nature. Step2: Calculate the measure of realism value for each decision alternative. Measure of Realism = (*max payoff) + ((1-)*min payoff) Step3: The optimal choice is the decision alternative which yields maximum measure of realism value. Example: Step1: Take = 0.7 Step2: MoR Expand = (0.7*500,000) + (0.3*(-450,000)) = 215,000 MoR Build = (0.7*700,000) + (0.3*(-800,000)) = 250,000 MoR Subcontract = (0.7*300,000) + (0.3*(-100,000)) = 180,000 Step3: The optimal choice is to Build a New Plant. Because; 250,000>215,000>180,000

4) The Mini-Max Regret Criterion

Step1: Obtain the regret table: Regret Value = Max Pay off of the row The payoff Step2: Identify the max regret value for each decision alternative. Step3: The optimal choice is the one which yields minimum in the list, which has been obtained in step2. Example: Step 1: Re-write the Decisions Alternative Table to come up with the Regret Table:

In the 1st Row: Regret Value for $500,000 = 700,000 500,000 = $200,000 Regret Value for $700,000 = 700,000 700,000 = 0 Regret Value for $300,000 = 700,000 300,000 = $400,000 In the 2nd Row: Regret Value for $250,000 = 300,000 250,000 = $50,000

- 50 -

Quantitative Analysis MGMT 322

Regret Value for $300,000 = 300,000 300,000 = 0 Regret Value for $150,000 = 300,000 150,000 = $150,000
N.B: Follow the same procedure to obtain the regret values for the 3rd and 4th rows. Regret Table States of Nature High Demand Moderate Demand Low Demand Failure

Expand $ 200,000 $ 050,000 $ 240,000 $ 350,000

Decision Alternatives Build 0 0 $ 390,000 $ 700,000

Subcontract $ 400,000 $ 150,000 0 0

Step 2: Write the maximum decision value for each alternative:

Max-Regret Table

Expand $ 350,000

Build $ 700,000

Subcontract $ 400,000

The optimum choice is to Expand. Because; $ 700,000>$ 400,000>$ 350,000

Application of Alternative Criteria for Decision Making under Risk

1) Expected Value Criterion
Step1: List the feasible decision alternatives. When state of nature is given as a discrete random variable; feasible decision alternatives are exactly the same as the states of nature. Step2: Obtain the probability of each state of nature using the past data. Step3: Using the cost and price data, obtain the conditional profit table. Step4: Using the probabilities obtain the expected value of profits for each decision alternative. Step5: Optimal choice is the decision alternative which yields maximum expected profit.

- 51 -

Quantitative Analysis MGMT 322

2) Criterion of Rationality
Sometimes we may not trust the past data that are used to obtain probabilities for states of nature, under those circumstances we can use this criterion which assumes that all states of nature are equally likely. We use this criterion when we have limited or unreliable data to calculate probabilities. The rest of the steps are exactly the same as those of the Expected Value Criterion.

3) The Criterion of Maximum Likelihood

We can use this criterion particularly when one of the states of nature has significantly higher probability then others. In this case: Step 1: We assure that this state of nature with the highest probability will actually occur.
Step 2: Given this assumption we chose the decision alternative which yields maximum conditional profits.

Example 1
The Mid Town Food store stocks mangoes during early summer season. These are flown in from Meritt Island, Florida, each Monday and just be sold within the week following. In the past, the store has been experiencing the following sales of mangoes:
Quantities Buyers Bought 20 25 40 60 # of weeks this occurred 10 30 50 10 Total = 100

Food Store buys mangoes for $2 and sells them for $4. Given this, apply each of the following criteria to determine the optimal quantity of mangoes that must be stocked per week: a. Expected Value Criterion b. Criterion of Maximum Likelihood c. Criterion of Rationality

- 52 -

Quantitative Analysis MGMT 322

a) Using the Expected Value Criterion: Step 1: Obtain the probabilities of states of nature

Probabilities of weekly demand States of Nature 10/100 = 0.1 30/100 = 0.3 50/100 = 0.5 10/100 = 0.1

Quantities Buyers Bought 20 25 40 60

Step 2: Obtain the Feasible Stocking Actions (Decision Alternative) N.B: Obtain exactly the number of quantity demanded from the supplier.

20

25

40

60

Step 3: Obtain the Conditional Profit table Obtain the weekly profit values for each possible combination of Stocks and Demand: Profit = TR TC TR = P * Qsold TC = C * Qstocked P = 4 and C = 2

a) Weekly profits when Stocks = 20units/week and Demand = 20 Profit = TR TC = (20*4) (20*2) = 40 b) Weekly profits when Stocks = 25units/week and Demand = 20 Profit = TR TC = (20*4) (25*2) = 30 c) Weekly profits when Stocks = 40units/week and Demand = 20 Profit = TR TC = (20*4) (40*2) = 0 d) Weekly profits when Stocks = 60units/week and Demand = 20 Profit = TR TC = (20*4) (60*2) = -40
N.B: Follow the same procedure to obtain the conditional profits for other Stockings and Demands.

States of nature

Prob. 0.1 0.3 0.5 0.1

Demand 20 25 40 60

Feasible Stocking Action 20 25 $ 40 $ 30 $ 40 $ 50 $ 40 $ 50 $ 40 $ 50

40 0 $ 20 $ 80 $ 80

60 -$ 40 -$ 20 $ 40 $ 120

- 53 -

Quantitative Analysis MGMT 322

Loss = 4 2 = $2 / unit
Step 4: Calculate the Expected Values:

E()20 = (0.1*40) + (0.3*40) + (0.5*40) + (0.1*40) = $40 E()25 = (0.1*30) + (0.3*50) + (0.5*50) + (0.1*50) = $48 E()40 = (0.1*0) + (0.3*20) + (0.5*80) + (0.1*80) = $54 E()60 = (0.1*-40) + (0.3*-20) + (0.5*40) + (0.1*120) = $22
Therefore: The optimum choice is to stock 40 units because it yields max. expected profit of $54. b) Using the Criterion of Rationality

Taking = 0.25 as our probability

Prob. 0.25 0.25 0.25 0.25 Demand 20 25 40 60 Feasible Stocking Actions 20 25 $ 40 $ 30 $ 40 $ 50 $ 40 $ 50 $ 40 $ 50 40 0 $ 20 $ 80 $ 80 60 -$ 40 -$ 20 $ 40 $ 120

States of Nature

Expected value of profits: E()20 = (0.25*40) + (0.25*40) + (0.25*40) + (0.25*40) = $40 E()25 = (0.25*30) + (0.25*50) + (0.25*50) + (0.25*50) = $45 E()40 = (0.25*0) + (0.25*20) + (0.25*80) + (0.25*80) = $45 E()60 = (0.25*-40) + (0.25*-20) + (0.25*40) + (0.25*120) = $25
Therefore: The optimal choice is to either stock 25 or 40 units because they yield the maximum expected profit of $45. c) Using the Maximum Likelihood Criterion We take the highest probability of 0.5 from the conditional probability table Feasible Stocking Actions Prob. 0.1 0.3 0.5 0.1 20 25 40 60 20 $ 40 $ 40 $ 40 $ 40 25 $ 30 $ 50 $ 50 $ 50 40 0 $ 20 $ 80 $ 80 60 -$ 40 -$ 20 $ 40 $ 120

- 54 -

Quantitative Analysis MGMT 322

The optimum choice is to stock 40 units because it yields the maximum expected profit of $80.

Example 2
Suppose that you are the sales manager of Migros supermarket chain in Istanbul. You are trying to make a decision about the optimal number of bottles of Efes beer that should be stocked for each month. The cost of each bottle to Migros is $6o and the beer is sold for a revenue of $95 per beer. For each bottle that is unsold by the end of the week and returned back to Efes, Migros receives $20.The sales data for the last 50 weeks is given as follows: Quantities Buyers Bought 20 21 22 23 Number of Months this occured 20 15 10 5

a. Using the Expected Value Criterion, obtain the optimal number of Efes bottles the sales manager of Migros should stock per month. b. Using the maximum Likelihood Criterion, obtain the optimal number of Efes bottles the sales manager of Migros should stock per month.
a) Using The Expected Value Criterion

Bottles of Beer Demanded 20 21 22 23 Total

Number of weeks 20 15 10 5 50

Probability 20/50 = 0.40 15/50 = 0.30 10/50 = 0.20 5/50 = 0.10 1.00

Conditional Profit Table: Profit = TR TC TR = P * Qsold TC = C * Qstocked P = $95 and C = $60 Profit Received from the Unsold Bottles = $20

a. Weekly Profit when Stock = 20 bottles/week and Demand = 20 Profit = TR TC = (20*95) (20*60) = 700 b. Weekly Profit when Stock = 21 bottles/week and Demand = 20 - 55 -

Quantitative Analysis MGMT 322

Profit = (TR TC) + (# of unsold bottles*Profit Received) = [(20*95) (21*60)] + (1*20) = 660 c. Weekly Profit when Stock = 22 bottles/week and Demand = 20 Profit = (TR TC) + (# of unsold bottles*Profit Received) = [(20*95) (22*60)] + (2*20) = 620 d. Weekly Profits when Stock = 23 bottles/week and Demand = 20 Profit = (TR TC) + (# of unsold bottles*Profit Received) = [(20*95) (23*60)] + (3*20) = 580
Feasible Stock Action 20 21 700 660 700 735 700 735 700 735

States of Nature

Probabilities 0.40 0.30 0.20 0.10

Demand 20 21 22 23

22 620 695 770 770

23 580 655 730 805

E()20 = (0.40*700) + (0.30*700) + (0.20*700) + (0.10*700) = $700 E()21 = (0.40*660) + (0.30*735) + (0.20*735) + (0.10*735) = $705 E()22 = (0.40*620) + (0.30* 695) + (0.20*770) + (0.10*770) = $687.50 E()23 = (0.40*580) + (0.30*655) + (0.20*730) + (0.10*805) = $655
Therefore: The optimum choice is to stock 21 bottles because it yields max. expected profit of $705. b) Using The Maximum Likelihood Criterion We take the highest probability of 0.40 from the conditional probability table Probabilities 0.40 0.30 0.20 0.10 Demand 20 21 22 23 Feasible Stock Action 20 21 700 660 700 735 700 735 700 735 22 620 695 770 770 23 580 655 730 805

States of Nature

Therefore: The optimum choice is to stock 20 bottles because it yields max. expected profit of $700.

- 56 -

Quantitative Analysis MGMT 322

Supplying the Numbers: How to Obtain Mean and Standard Deviation when Data are Missing or Incomplete
This is a technique used to obtain estimates for mean and standard deviation (of particularly demand for a product) when we dont have past data to use in formula for mean and standard deviation. X~N (x ,x)

This technique involves asking questions to some experts in the company and then proceeding their answers in a scientific manner to get estimates for mean and standard deviation.
For Example: Q1: In your opinion, what is expected annual volume of sales? Or on average how much of this product can be sold per year? Answer: 1500 units/year, your estimate for x

Annual Demand

Before you ask the 2nd question, form an interval around x (1,500) by taking any percentage of estimates for x such as 20%: 20% * 1500 = 300 Then add and subtract 300 from 1500 to get an interval around 1500 given below:

Q2: In your opinion, what are the odds (chances) for the annual sales volume to lie between 1200 and 1800?

- 57 -

Quantitative Analysis MGMT 322

Answer: Odds are 3 to 1 (3:1)

P(1200< X <1800) = (3+3)/(1+3+3+1) = 6/8 = 0.75

Step 1: Obtain the probability of X being less than or equal to the right hand value of the interval. 1+ 3 + 3 P(X < 1800) = 7/8 = 0.875= 1+ 3 + 3 +1

Step 2: Obtain the Z value that corresponds to the probability value obtained in step 1 (or closest probability value in the normal table).

Probability as close as possible to 0.875 = 0.87493 (from the table) Z* = 1.15

Step 3:

Z* =

X*-

x
- 58 -

Quantitative Analysis MGMT 322

1.15=

1800-1500

x= 261

Therefore estimates for and = X~N (1500, 261) Example

Suppose that the production manager of Phillips Electronics company in Netherlands believes that the expected annual volume of sales of the new product models that they plan to introduce to the market in 2003 is 200,000.Furthermore he believes that the odds (chances) are 4 to 3 that the annual sales volume will lie between 100,000 and 300,000. What is your best estimate for the standard deviation of annual sales volume of this product model of Phillips?
Solution: = 200,000 Odds = 4:3 Range = 100,000:300,000

=?

3 100,000

4 200,000

3 300,000 X

- 59 -

Quantitative Analysis MGMT 322

P(X 300,000) = 3+4+4 / 3+4+4+3 = 11/14 = 0.7857 Z* = 0.79 Z* = X- / 0.79 = 300,000 200,000 / 0.79 = 100,000 / = 126582
Therefore: = 126582

- 60 -

Quantitative Analysis MGMT 322

Chapter 5: Cost Volume Profit Analysis

We use Cost-Volume Analysis also called Break-Even Analysis to determine the profitability and riskness of an investment project. This technique involves asking probability questions and assessing the risk faced by the company in investing in this project, based on the answers to those questions. It also involves computing the expected value of net profit from this project which also helps the managers to make more informed investment decisions.

Application: We build our example we used earlier in estimating the mean and the
standard deviation of annual demand for a product that a company is considering to produce. Note that annual demand (X) was estimated to have a mean() of 1,500 and a standard deviation() of 261.

X~N (1500, 261)

Data About The Project

If: The cost of capital investment = $100,000 (initial capital) TFC = $50,000 P/unit = $10 (Price that the company expects to sell). AVC/unit = $5 (Average variable cost is assumed to be approximately constant for the range of output that the company expects to purchase).

Questions That Managers Can Ask Specific To This Project

Q1: What is the probability of at least breaking even on this project for each year? Q2: What is the probability of earning more than $900 profit for each year? Q3: What is the probability of losing more than $200 for each year? Q4: What is the probability of earning minimum 2% annual rate of return for each dollar invested? Q5: What is the expected value of Net Profit on this project?

- 61 -

Quantitative Analysis MGMT 322

Profit = TR = TC TFC XBE = P V Where p v = unit contribution to profits For: X > XBE, Profit >0 X = XBE, Profit = 0 X < XBE, Profit < 0 Profit = TR TC TR = PQ TC = TVC + TFC TVC = VQ

At XBE

Profit = PQ (VQ+TFC) QBE = TFC / P V

Answer 1: P(Profit 0) =P(Profit XBE)

XBE = TFC / P V = 5000 / 10 5 = 1000 P(X > XBE) = P(X > 1000) Z* = Z* = -1.91 1000 1500 261

- 62 -

Quantitative Analysis MGMT 322

P(Q > 1000) = 0.97193

The probability of reaching at least break even is 97.193 %

Answer 2: P(Earnings at least more than $900 per year) = P(Profit > 900) = P(X > X*)

X* = 1000 + 900 / 5 = 1180 P(X > X*) = P(X > 1180)

Z* = Z* = -1.23

1180 1500 261 P(X > 1180) = 0.89065

- 63 -

Quantitative Analysis MGMT 322

Answer 3: P(Losing more than $200 for each year) P(Profit < -200) = P(X < X*) X* = 1000 + (-200 / 5) = 960

Z* = Z* = -2.07

960 1500 261

P(Profit <- 200) = 0.01923

Answer 4: P(Earning minimum 2% annual rate of return for each dollar invested) P( Profit Profit*) Profit* = 2% * $100,000 = $2000 X* = XBE + Profit* / P V = 1,000 + 2000/10 5 =1,400 units P(Profit $2000) = P(X X*) = P(X 1,400) 1400 1500 Z* = 261

Z* = -0.38

1,400

1,500

Z*=-0.38

- 64 -

Quantitative Analysis MGMT 322

P(Profit $2000) = 0.64803

Answer 5: Expected Net Profits E(Net Profit) = E(Annual Profit) E(Annual Loss)

The area < 1000 is Potential Loss [ P(Profit < 0)], and the area > 1000 is Potential Gain [ P(Profit 0), while 1000 is a Break-Even Point.

Example
TT Corporation is discussing to produce a new product. It has been estimated that the average annual volume of sales will be 8,000. The manager of the sales department believes that the odds (chances) are 2 to 3 that the annual volume of sales will lie between 7,000 and 9,000. If they choose to produce the product, they expect to incur an annual fixed cost of an amount equal to 300,000. The sales price is estimated to be $100 and the variable cost per unit of output is estimated to be $50. a. What is the probability of losing money on this product for each year? b. What is the Expected Annual Profit for this project? c. What is the probability of at least breaking even for each year?
Solution: = 8,000 Odds = 2:3 TFC = $30,000 P = $100 VC/unit = $50

3 9,000 X

7,000 8,000 P(X < 9,000) = 2+2+3/3+2+2+3 = 0.70

- 65 -

Quantitative Analysis MGMT 322

Z = 0.52 Z=X-/ 0.52 = 9,000 8,000 / = 1923.

a) P(Losing Money)

6,000

8,000

Z*=?

XBE = TFC / P-V = 300,000 / 100-50 = 6000 Z* = 6000 8000 / 1923.08 = -1.04 P(X < XBE) = P(Z < -1.04) = 1 0.85083 (from the probability table) = 0.14917
b) E(Annual Profit)

Since 6,000 < 8,000 (XBE < ) E(Annual Profit) = k * * (UNLI + Z) = 50*1923.08*(o.7716 +1.04) = $174192.59
c) P(At least Breaking-Even) P(X 6,000)

- 66 -

Quantitative Analysis MGMT 322

6,000 Z* =-1.04

8,000 0

X Z

P(X 6,000) = P(Z > 1.04) = 0.85083

Combining Unit Monetary Values and Probability Distribution

Procedure:
a. 3-Step procedure to obtain in Expected Profit b. 3-Step procedure to obtain in Expected Loss c. Obtain Expected Net profit [E(Profit) E(Loss)]
Given: X~N (1500, 261) TFC = 5000 Standard Deviation = 261 Mean = 1500 P = 10 V=5 XBE = TFC / P V = 1000

a. 3 Step Procedure to Obtain Expected Profits

Step1: Determine the Z value for XBE in absolute terms |Z| Step2: Determine unit normal loss integral value for this z value using UNLI table. Step3: a) If XBE < Mean, E(Profit) = k**(UNLI +z)

b) If XBE > Mean, E(Profit) = k**UNLI Where k is the profits per unit of output above XBE.

- 67 -

Quantitative Analysis MGMT 322

Note: Sometimes K and C can be different from each other, but if you are given P and V values, then always P-V = K = C

Example
Step 1: |Z| =

1000 1500 = |-1.91| = 1.91 261

Step 2: For Z = 1.91, UNLI = 0.01077, using the UNLI table. Step 3: Since 1000 < 1500 (XBE < ) E(Profit) = k**(UNLI +z) = $5 * 261 * (0.01077 + 1.91) = $2506.6

b. 3 Step Procedure to Obtain Expected Loss

Step1: Determine the Z value for QBE in absolute terms |Z| Step2: Determine unit normal loss integral value for this z value using UNLI table. Step3: a) If XBE < Mean, E(Profit) = c**(UNLI +z)

b) If XBE > Mean, E(Profit) = c**UNLI Where c is the loss per unit of output above XBE. |(P - V) = k = c|

Example
Step 1: |Z| = |-1.91| = 1.91 Step 2: UNLI = 0.01077 for Z = 1.91 Step 3: Since 1000 < 1500 (XBE < ) E(Loss) = c**UNLI = 5 * 261 * 0.01077 = $14.05

c. Obtain Expected Net Profits

E(Net Profit) = E(Profit) E(Loss) = $2506.6 - $14.05 = $2492.55

- 68 -

Quantitative Analysis MGMT 322

Example
Girne Septic Tank Company operates a truck called the Honey Cart which is used to pump out the capacitated septic tanks. The company has estimated that usage of the truck is normally distributed with a mean of 150 hours per month and a standard deviation of 40 hours per month. In order to break-even the company must operate for 90 hours per month. a. Find the expected loss for Honey Cart if the company loses $30 for each hour of operation below the break even point. b. Find the expected Profit for Honey Cart if the company earns $28 for each hour of operation above the break-even point. c. Find the Expected Net Profit for the operation of the Honey Cart for this company
Solution: = 150 = 40 C = $30 K = $28 XBE = 90

|Z| = |XBE - / | = |90 150 / 40| = |-1.5| = 1.5 For Z = 1.5, UNLI = 0.02931
a) Expected Loss = C * * UNLI = 30 * 40 * 0.02931 = $35.17 per month b) Expected Profit = K * * (UNLI + Z) = 28 * 40 * (0.02931 + 1.5) = $1,712.83 per month c) Expected Net profit = E(Profit) E(Loss) = $1,712.83 - $35.17 = $1,677.66

- 69 -

Quantitative Analysis MGMT 322

Home works
Home Work 1
Questions from the textbook: Chapter 2: Ex 2-7, 2-8, 2-9, 2-12, 2-40, 2-44*

Home work 2
1. Suppose we toss a coin 5times, and it is an unbiased coin a. What is the probability of at least 4heads that may come up? b. What is the Expected Value of number of tails that may come up? 2. A Financial Analyst computed the return on stockholders equity for all the companies listed on the New York Stock Exchange. She found that the mean of this distribution was 10%, with a standard deviation of 5%. She is interested in examining further those companies whose return on stockholders equity is between 8% and 13%. Of the approximately 13,000 companies listed on the exchange, how many are of interest to her? 3. Questions from the textbook: Chapter 2: Ex 2-32, 2-36, 2-37, 2-38

Home Work 3
1. Following is the sales data gathered for the past 12 months of a company:
Years 1999 1999 1999 1999 1999 2000 2000 2000 2000 2000 2000 2000 Months August September October November December January February March April May June July Sales 28 27 33 25 34 33 35 30 33 35 27 29

Apply the following to the above data set to obtain MSE and develop a forecast for August 2000 for each case: a. 3-period (W) MA model, taking Wt = 0.5, Wt-1 = 0.3, Wt-2 = 0.2 b. 4-period (UW) MA model

- 70 -

Quantitative Analysis MGMT 322

c. Use Simple Exponential Smoothing model, taking = 0.5 and obtain MSE of the model and develop a forecast for August 2000. d. Which specification above said has the Highest Forecasting Accuracy? 2. Questions from the textbook: Chapter 3: Ex 3-1 (Just MSE Criteria), 3-4, 3-5, Ex 3-21(a) - Please take note that the warm-up sample should be 4 periods not 6 periods. Ex 3-25

Home Work 4
1. Use Smoothing Linear Trend model to develop a forecast for August 2000 for Question 1 in Homework-3, taking 1 = 0.1 and 2 = 0.01. Also obtain MSE of this model. 2. Questions from the textbook: Chapter 3: Ex 3-21(b) Ex 3-22, 3-23 (The warm-up Sample in both Ex 3-22 and 3-23 should be 4periods) Ex 3-26

Home work 5
1. Questions from the textbook: Chapter 4: Ex 4-1, 4-2, 4-3, 4-4, 4-5, 4-15, 4-33, 4-3 2. Beth Perry, who sells strawberries in the market environment where tomorrows demand for strawberries is a discrete random variable. Beth purchased strawberries for $3 a case and sells them for $8 a case. The rather high mark up reflects the perishability of the item and the great risk of stoking it; the product has no value after the first day it is offered for sale. Beth faces the problem of how many to order today for tomorrows business. A 90-day observation of past demand gives the information shown in the table below: Daily Demand 10 11 12 13 Number of days Demanded 18 36 27 19

What quantity should Beth Perry buy for tomorrows sale to maximize the Expected Profit?
N.B: Read the textbook; page 165-168.

- 71 -

Quantitative Analysis MGMT 322

Home Work 6
Questions from the textbook: Chapter 5: Ex 5-2, 5-3, 5-5, 5-8, 5-10, 5-12, 5-14

- 72 -

Quantitative Analysis MGMT 322

The Graph areas to be shaded:

Page 27, Example 2 Graph1: Shade X and Y area. Page 28, Example 2 Graph2: Shade from 400,000 going right, up to the end. Page 28, Example 3 Graph: Shade the area between 16% and 22% Page 64, Graph: Shade the Graph from 1400 going right, up to the end Page 66, Graph: Shade the Graph from 6000 going left up to the end Page 67, Graph: Shade From 6000 going right, up to the end

- 73 -

Quantitative Analysis MGMT 322

QUIZ QUESTIONS

- 74 -

Quantitative Analysis MGMT 322

1. Suppose we are playing a game, we draw 3 cards with replacement from an ordinary deck. We need to pay $2 for each drawing and will earn $20 dollar if the heart will be drawn. What is our expected value of net profit for 4 drawings? 2. Weekly sales of automobiles at S.Cs Auto Gallery is normally distributed with a mean of 200 and a standard deviation of 50 automobiles. a. What is the probability that sales will be less than-170 in a given week? b. What is the probability that sales will be in excess of 260? c. What is the probability that sales will be between 175 and 230? 3. The probability that CIMSAs stock price will rise in a given day is estimated to be 0.60. When CIMSAs stock price rises, there is a 0.70 probability that the exchange rate of TL/$ will fall in that day. When CIMSAs stock price falls, there is a 0.4 probability that the exchange rate of TL/$ will fall in that day. Tomorrow what is the probability that CIMSAs price will rise and also the exchange rate of TL/$ will fall?

4. Suppose we have three machines for making a particular part. The first machine produces 4% defectives, the second machine produces 6% defectives, and the third machine produces 1% defectives. Suppose also that the first machine supplies 30%, the second machine 20%, and the third machine 50%, of the parts. If a part is selected at random, what is the probability that it is defective? Given that the part is defective, what is the probability that it came from the first machine?

- 75 -

Quantitative Analysis MGMT 322

5. It has been suggested that we implement a procedure for testing for child abuse. Suppose that when a child has been abused a doctor can correctly identify the abuse with probability 0.99, and that when a child has not been abused the doctor can correctly identify the non-abuse with probability 0.9 Suppose also that the probability that a child has been abused is 0.05. If the doctor says that a child has been abused, what is the probability that the child ha actually been abused? 6. The probability that CIMSAs stock price will rise in a given day is estimated to be 0.70. When CIMSAs stock price rises, there is a 0.75 probability that the exchange rate of TL/$ will fall in that day. When CIMSAs stock price falls, there is a 0.5 probability that the exchange rate of TL/$ will fall in that day. Tomorrow what is the probability that CIMSAs price will rise and also the exchange rate of TL/$ will fall? 7. Weekly sales of automabiles at S.Cs Auto Gallery are normally distributed with a mean of 300 and a standard deviation of 40 automobiles. a. What is the probability that sales will be less than 270 in a given week? b. What is the probability that sales will be in excess of 370? c. What is the probability that sales will be between 284 and 328? 8. Our professor has put $3 on each of the three horses running in three different races at ISTANBUL HORSE RACING CENTER He feels that each of the three bets he has made, has a 0.3 probability of winning. A winning ticket on any of the three will earn our professor $15. Assuming a Bernoulli process, what is the expected value of his net profit? 9. A child chosen at random in a community school system comes from a lowincome family 15 percent of the time. Children from low-income families in the community graduate from college only 20 percent of the time. Children not from low-income families have a 40 percent chance of graduating from college. As an employer of people from this community, you are reviewing applicants and note

- 76 -

Quantitative Analysis MGMT 322

that the first one had a college degree. What is the probability that person comes from a low-income family? 10. Suppose that the past annual data about Istanbul Stock Market Index suggests that the probability of index rising in a given year is 0.60, the probability of no change is 0.05, and the probability of index falling in a year is 0.35. Furthermore analysts have found out that for all the years during which index have risen, 80% of the time interest rates have fallen and 20% of the time interest rates increased. For all the years index have not changed, 50% of the time interest rated have fallen and 50 % of the time interest rates have increased. And for all the years index have fallen, 40% of the time interests rates have fallen and 60% of the time interest rates have increased. Latest developments in Turkey show that interest rates started rising and will continue to rise in year 2001. Given this positive upward trend in interest rates in 2001, what are your revised estimates for the probability of index rising, not changing, and falling in year 2001? 11. Suppose that Ford Corporation produce 3 different models for Turkish consumers, in its manufacturing plant located in Turkey. Annual volume of sales and production of all the 3 models combined is 700 units: However, some of the automobiles manufactured and sold mm out to be defective and lead to consumer complaints. Based on the past data, production manager estimated the following distribution of defective and non-defective automobiles for each model manufactured in year 2001:
A 10 140 B 20 180 C 70 280

D N

a. What is the probability that a randomly selected consumer (who decided to buy a Ford) will buy either model A or model C? b. What is the probability that the automobile that he buys will be either model B or defective?

- 77 -

Quantitative Analysis MGMT 322

c. What is the probability that the automobile that he buys will be either model A or nondefective? d. If the automobile purchased by a randomly selected consumer is found to be defective, what is the probability that it is a Model A automobile? e. If the automobile purchased by a randomly selected consumer is found to be nondefective, what is the probability that it is Model C automobile? 12. Mr X has $150,000 available for one of three investment alternatives: stock, treasury bills, $ denominated saving deposits. The investment environment can assume any one of three states depending on the rate of inflation (which will determine the rate of depreciation of TL against $). The payoff table of Mr. Y looks like this:
Type of investment Inflation High Moderate Low Stock Treasury Bills $ Saving deposits

$300,000 $100 $50,000

$100,000 $150,000 $200,000

$ 150,000 $ 150,000 $ 150,000

a). What should we suggest Mr. X to invest by using the criterion of Maximax? b). What should we suggest Mr. X to invest by using the criterion of realism? Assume = 0.6. 13. The Province of Quebec is planning to issue hunting licenses for moose through a province-wide lottery. They have found that by harvesting some animals during the hunting season, they can increase the number of animals which survive through the winter months because of the limited food supply. Their estimates of moose population (in thousands) conditional upon the number of licenses issued and the severity of the winter are as follows:

- 78 -

Quantitative Analysis MGMT 322

Number of licenses Issued

Severity of winter Mild Moderate Severe Harsh

5,000

6,000

7,000

8,000

9,000

38 35 28 22

36 33 30 26

34 33 32 30

33 29 27 25

25 22 20 16

During the years when records have been kept, 20% of the winters have been mild while 30%, 40%, and 10% have been moderate, severe, and harsh, respectively. How many licenses should be issued in order to maximize the expected moose population? How large will the expected springtime moose population be with this decision? 14. Monthly demand for product A at company B has been as follows: Month Jan. Feb. Mar. Apr. May Jun Ju1 this model. b. Use Smoothing Linear Trend Model to develop forecast for August and obtain MSE of the model. Choose 1= 0.2, 2= 0.02 c. Compare above 2 models and decide which one is more accurate. Demand 59 53 57 48 43 37 38

a. Use Naive Model to develop forecast for August and compute MSE of

- 79 -

Quantitative Analysis MGMT 322

15.

Linda, the managing editor of a magazine, needs to develop a forecasting system for monthly sales in order to schedule press runs. Sales (in thousands of copies) for the first 8 months of 1999 (the first year of publication) were: Month Jan. Feb. Mar. Apr May Jun. Jul. Aug. Demand 50 45 60 52 69 60 47 53

Linda does not believe there is a seasonal pattern. a. Use 3-period Weighted Moving Average Model to obtain a forecast for September of 1999. Wt = 0.5, Wt-l = 0.3, Wt-2 = 0.2 b. What is the MSE of this model? 16. Monthly demand for product A at company B has been as follows:
Month Demand

Jan. Feb. Mar. Apr. May Jun. Jul. MSE.

59 53 57 48 43 37 38

Use linear regression model to develop a forecast for August and compute

- 80 -

Quantitative Analysis MGMT 322

17.

Linda, the managing editor of a magazine, needs to develop a forecasting C f10 system for monthly sales in order to schedule press runs. Sales (in thousands of copies) for the first 8 months of 1999 (the first year of publication) were
Month Demand

Jan. Feb. Mar. Apr. May Jun. Jul. Aug.

50 45 60 52 69 60 47 53

Linda does not believe there is a seasonal pattern. a. Use 2-period Un-weighted Moving Average Model to obtain a forecast for September of 1999. What is the MSE of this model? b. Use Simple Exponential Smoothing Model to forecast the sales of September and compute the MSE of this model. Take a = 0.6. c. Compare above-said two models, which one is more accurate? 18. Mr X has $100,000 available for one of three investment alternatives: stock, treasury bills, $ denominated saving deposits. The investment environment can assume any one of three states depending on the rate of inflation (which will determine the rate of depreciation of TL against $). The payoff table of Mr. Y looks like this: Type of investment Inflation Stock Treasury Bills $ saving deposits.

- 81 -

Quantitative Analysis MGMT 322

Type of Investment Inflation High Moderate Low Stock $250,000 $0 -$100,000 T. Bills $50,000 $100,000 $150,000 $ Saving Deposits $100,000 $100,000 $100,000

We have learned that Mr. Y chose treasury bills to invest. According to the criterion of realism, find a range of alpha values that characterize the level of optimism implied by this decision maker. 19. The ACME Company is contemplating a new product that would sell for $11 a unit, the per-unit variable cost for this product is $8, and the fixed cost per year allocated to this product is $ 240,000. The sales manager for ACME estimates that annual sales for this product would have a mean of 200,000 with a standard deviation of 85,000 unit& Using this information, answer the following questions: a). What is the probability that ACME would loose money on this product next year? b). If the cost of capital investment necessary to undertake this project is estimated to be $3 million, what is the probability of earning at least 5 percent rate of return annually? c). What are the expected annual profit and expected annual loss from the production of this product? (Assume that unit loss equals to unit profit)

- 82 -

Quantitative Analysis MGMT 322

EXAM QUESTIONS

- 83 -

Quantitative Analysis MGMT 322

1. Monthly Demand for TV sets at TTs Electronics has been as follows: Months Jan. Feb. March Apr. May June July Demand 138 137 143 148 157 153 159

Using Linear Regression obtain forecast for August and September and compute MSE of the model. (15 pts)

2. Los Bichos, the managing editor of ASTROLOGY magazine, needs to develop a forecast for the 2nd Quarter of 2002. (10 pts) Sales in the past quarter have been as follows: Years 2000 Qtr. 2nd Qtr. 3rd Qtr. 4th Qtr. 2001 1st Qtr. 2nd Qtr. 3rd Qtr. 4th Qtr. 2002 1 Qtr.
st

Sales 50 52 47 55 53 50 54 52

- 84 -

Quantitative Analysis MGMT 322

Los Bichos does not believe there is a seasonal pattern. Using 2-period Moving Average model obtain a forecast for the 2 Qtr. of 2002 and compute MSE of the model. (15 pts)

3. Prof. Demir is planning to invest his 1 M $ in anyone of the three investment alternatives including GOLD, REAL ESTATE and STOCK. He believes that the ANNUAL RETURNS (pay-offs) from each one of these investment alternatives will critically depend on the inflation rate. Furthermore he believes that, there are 3 possible outcomes for inflation. His corresponding pay-off table is given below:

Investment Alternatives Gold States of Nature High Inf. Medium Inf. Low Inf. $100000 $50000 $10000 Real Estate $150000 $20000 $-5000 Stocks $15000 $30000 $80000

A. Using the MINIMAX REGRET criterion obtain the optimal investment alternative. (15 pts) B. Using the MAXIMIN criterion obtain the optimal investment alternative. (5 pts)

4. Suppose that you are the sales manager of MIGROS supermarket chain in Istanbul. You are trying to make a decision about the optimal number of bottles of EFES beer that should be stocked for each month. The cost of each bottle to - 85 -

Quantitative Analysis MGMT 322

MIGROS is 10 S. whereas the price that MIGROS charges from the customers is 15$ per bottle. For each bottle that is unsold by the end of the week and returned back to the producer of EFES, MIGROS receives $5 only. The sales data for the last 200 months is given below:

QUANTITES BUYERS BOUGHT 5000 3500. 3000 2500

NUMBER OF MONTHS THIS OCCURRED 30 70 60 40

A. Using the EXPECTED VALUE criterion obtain the optimal number of EFES bottles that the sales manager of MIGROS should stock per month. (15 pts) B. Using the MAXIMUM LIKELIHOOD criterion obtain the optimal number of EFES bottles that the sales manager of MIGROS should stock per month. (5 pts)

5. Suppose that the production manager of HONDA automobiles in Japan believes that the expected annual volume of sales of the new model that they plan to introduce to the market in 2003 is 200000. Furthermore he believes that the odds (chances) are 4 to 3 that the annual sales volume will lie between 100000 and 300000. What is your best estimate for the standard deviation of annual sales volume of this model of HONDA? (10 pts) 6. The project manager of SABANCI Corporation has been analyzing the feasibility of an investment project, which involves the production of computer chips in Turkey. They

- 86 -

Quantitative Analysis MGMT 322

estimated that the cost of capital investment to undertake this roject is $ 9 M. The selling price of each computer chip is believed to be $ 500. The variable cost per unit is estimated to be approximately $ 200. Annual fixed costs are likely to be around $ 300000. The expected annual volume of sales is 1 500 and standard deviation of sales is estimated to be 250. A. What is the probability of at least breaking-even for each year? (5 pts) B. What is the probability of earning less than $75000 profits for each year? (5 pts) C. What is probability of losing at most $ 90000 for each year? (5 pts) D. What is the probability of earning more than 2 % annual rate of return for each dollar invested in this project? (5 pts)

7. Prof. TT has put $2 on each of the 3 horses running m three different races in Istanbul Horse Racing Center He feels that each of the bets he has made has a 02 probability of winning. A winning ticket on any of the three horses will earn Prof TT $40 Assuming a Bernoulli process, answer each of the following questions: a. What is the probability of at least 2 of the horses(on Prof. TT has bet) winning? b. What is the EXPECTED EARNINGS from all three races? c. What is the probability of at most 1 out of three horses (on which Prof TT has bet) winning? d. What is the probability distribution of his possible earnings from all three races?

- 87 -