1. FEC 512 Financial Econometrics
-About the Course-
• What is Financial Econometrics?
– Science of modeling and forecasting financial time series.
• Who is this course for?
– Students in finance, practitioners in the financial services
sector.
• How is the presentation of the lectures?
– Begins with review of necessary statistics and probability
theory, continues with basics of econometrics reaches up
to the most recent theoretical results
– Uses computer applications, Eviews.
– Course materials are online at online.bilgi.edu.tr
– Students are supposed to choose a data set from the list at
the beginning of the term in order to do assignments
– Attendence is not required.
1
2. FEC 512 Financial Econometrics
-About the Course-
• Textbooks:
1. (for Statistics part only) Groebner D.F. et al.(2008)
Business Statistics.
2. Ruppert D. (2004), Statistics and Finance, Springer.
3. Brooks, C. “Introductory Econometics for Finance”
4. Stock J.H. and Watson M.W. (2003), Introduction to
Econometrics (first edition), Addison-Wesley.
• Method of Evaluation:
– Assignments (50%)
– Final Examination (50% )
2
3. Overview
Before getting into applications in financial
econometrics we will first define
• returns on assets
Then we will review
• Probability
– probability density functions, cumulative distribution functions
– expectations, variance, covariance and correlation
• Statistics
– Testing
– Estimation
• We’ll also be studying some new areas of statistics:
– Regression
• interesting connections with portfolio analysis.
– Probit, Logit Analysis
– Time series models
3
4. I. Preliminary: Asset Return
Calculations
Istanbul Bilgi University
FEC 512 Financial Econometrics-I
Asst. Prof. Dr. Orhan Erdem
FEC 512 Preliminaries and Review Lecture 1-4
5. Background
How do the prices of stocks and other
financial assets behave?
We will start by defining returns on the prices
of a stock.
Lecture 1-5
FEC 512 Preliminaries and Review
6. Prices and Returns
Main data of financial econometrics are asset
prices and returns.
Almost all empirical research analyzes
returns to investors rather than prices. Why?
Investors are interested in revenues that are
high r.t. size of the initial invstmnt. Returns
measure this: Changes in prices expressed
as a fraction of the initial price.
Lecture 1-6
FEC 512 Preliminaries and Review
7. Asset Return Calculations
Pt is the price of a stock at time t. Stock pays no
dividends.
Simple return
( Pt − Pt −1 ) = P
Rt = −1
t
P−1 P−1
t t
Simple gross return
Pt
Rt + 1 =
Pt −1
Lecture 1-7
FEC 512 Preliminaries and Review
8. Multi-period returns e.g.
Pt Pt Pt −1
R t (2) = −1 = −1
Pt − 2 Pt −1 Pt − 2
= (1 + R t )(1 + R t −1 ) − 1
In general, k-month gross return is defined as
1 + R t ( k ) = (1 + R t )(1 + R t −1 )....(1 + R t − k + 1 )
Note: For small values of Rt
1 + Rt (k ) ≅ 1 + Rt + ... + Rt − k +1 or
k −1
Rt (k ) ≅ ∑ Rt −i
i =0
Lecture 1-8
FEC 512 Preliminaries and Review
9. Example 1
Suppose that the price of Arçelik stock on January
is 100YTL, and on February is 105YTL, and that you
sell the stock now(on March) at Pt=110YTL. Assume
no dividends,then
Rt=(110-105)/105=0.0476
Rt-1=(105-100)/100=0.05
Rt(2)=(110-100)/100=0.10
Check also that 1+Rt(2)=(1+ Rt)(1+ Rt-1)
1.0476*1.05=1.1
Lecture 1-9
FEC 512 Preliminaries and Review
10. Annualizing Returns
If investment horizon is one year
1+RA =1+R(12) =(1+R1) (1+R2)... (1+R12)
One month inv. with return Rt, (assume Rt=R)
1+RA=(1+R)12
Two month inv. with return Rt(2), (assume
Rt(2)=R(2))
1+RA=(1+R(2))6
Lecture 1-10
FEC 512 Preliminaries and Review
11. Cont. to Example 1
In the first example the one month return was
4.76%. If we assume that we can get this
return for 12 months then the annualized
return is
RA=(1.0476)12-1=1.7472-1=0.7472 or 74.72%
Lecture 1-11
FEC 512 Preliminaries and Review
12. Log-Returns
The log-return is
rt = log( Pt ) − log( Pt −1 ) == log( Pt / Pt −1 ) = ln(1 + Rt )
The log return in the previous example is
rt=ln(0.0476)=0.0465 or 4.65%
The above return measures are very similar
numbers since daily returns are very rarely outside
the range of -10% to 10%.
Lecture 1-12
FEC 512 Preliminaries and Review
13. Log returns are approximately equal to net
returns:
x small ⇒ log(1 + x) ≅ x
Therefore, rt = log(1 + Rt) ≅ Rt
Examples:
* log(1 + 0.05) = 0.0488
* log(1 -0.05) = -0.0513
Lecture 1-13
FEC 512 Preliminaries and Review
14. log(1+x) and x
when x is small
Lecture 1-14
FEC 512 Preliminaries and Review
16. Returns are
scale-free, meaning that they do not depend
on monetary units (dollars, cents, etc.)
not unit-less, unit is time; they depend on the
units of t (hour, day, etc.)
Lecture 1-16
FEC 512 Preliminaries and Review
17. Portfolio Return
N
R p = ∑ wi Ri
i =1
where wi is the weight of each asset in the
portfolio.
Example:
Lecture 1-17
FEC 512 Preliminaries and Review
18. About Returns
Returns cannot be perfectly predicted, they
are random.
This randomness implies that a return might
be smaller than its expected value and even
negative, which means that investing involves
RISK.
It took quite some time before it was realized
that risk could be described by probability
theory
Lecture 1-18
FEC 512 Preliminaries and Review
19. II. Review of Probability &
Statistics
FEC 512 Preliminaries and Review Lecture 1-19
20. Probability and Finance
Because we cannot build purely
deterministic models of the economy, we
need a mathematical representation of
uncertainty in finance (probability, fuzzy
measures etc…)
In economic and finance theory, probability
might have 2 meanings:
As a descriptive concept
1.
As a determinant of the agent decision
2.
making theory.
Lecture 1-20
FEC 512 Preliminaries and Review
21. Probability as a Descriptive Concept
The probability of an event is assumed to be
approx. equal to the rel.freq. of its occurrence
in a large # experiments.
There is one difficulty with this interpretation:
Empirical data have only one realization.
Every estimate is made on a single time-evolving
series.
If stationarity(!) is not assumed, performing
statistical estimation is impossible.
Lecture 1-21
FEC 512 Preliminaries and Review
22. Probability Concepts
Experiment – a process of obtaining
outcomes for uncertain events
Outcome – the possible results of an
observation, such as the price of a security
at t.
However, probability statements are not
made on outcomes but on events, which are
sets of possible outcomes.
The Sample Space is the collection of all
possible outcomes
Lecture 1-22
FEC 512 Preliminaries and Review
23. Examples
Event Example 1: The probability that the price of a security be in a
given range, say (10,12)YTL
Example 2:
Outcome
Sample Space=
The Set of Odd numbers is an Event
Probabilities are defined on events.
Lecture 1-23
FEC 512 Preliminaries and Review
24. Mutually Exclusive Events
If E1 occurs, then E2 cannot occur
E1 and E2 have no common elements
E2 A die cannot be
E1
Odd and Even at
Even the same time.
Odd Numbers
Numbers
Lecture 1-24
FEC 512 Preliminaries and Review
25. Independent and Dependent
Events
Independent: Occurrence of one does not
influence the probability of occurrence of
the other
Dependent: Occurrence of one affects the
probability of the other
Lecture 1-25
FEC 512 Preliminaries and Review
26. Independent vs. Dependent Events
Independent Events
E1 = heads on one flip of fair coin
E2 = heads on second flip of same coin
Result of second flip does not depend on the
result of the first flip.
Dependent Events
E1 = rain forecasted on the news
E2 = take umbrella to work
Probability of the second event is affected by the
occurrence of the first event
Lecture 1-26
FEC 512 Preliminaries and Review
27. Assigning Probability
Classical Probability Assessment
Number of ways Ei can occur
P(Ei) =
Total number of elementary events
Relative Frequency of Occurrence
Number of times Ei occurs
Relative Freq. of Ei =
N
Subjective Probability Assessment
An opinion or judgment by a decision maker about
the likelihood of an event
Lecture 1-27
FEC 512 Preliminaries and Review
28. Rules of Probability
Rules for
Possible Values
and Sum
Individual Values Sum of All Values
k
0 ≤ P(Ei) ≤ 1
∑ P(e ) = 1
i
For any event Ei Rule 1 Rule 2
i=1
where:
k = Number of individual outcomes
in the sample space
ei = ith individual outcome
Lecture 1-28
FEC 512 Preliminaries and Review
29. Addition Rule for
Elementary Events
The probability of an event Ei is equal to
the sum of the probabilities of the
individual outcomes forming Ei.
That is, if:
Ei = {e1, e2, e3}
then:
P(Ei) = P(e1) + P(e2) + P(e3) Rule 3
Lecture 1-29
FEC 512 Preliminaries and Review
30. Complement Rule
The complement of an event E is the
collection of all possible elementary events
not contained in event E. The complement of
event E is represented by E.
E
Complement Rule:
P( E ) = 1 − P(E) E
P(E) + P( E ) = 1
Or,
Lecture 1-30
FEC 512 Preliminaries and Review
31. Addition Rule for Two Events
Addition Rule:
■
P(E1 or E2) = P(E1) + P(E2) - P(E1 and E2)
+ =
E1 E2 E1 E2
P(E1 or E2) = P(E1) + P(E2) - P(E1 and E2)
Don’t count common
elements twice!
Lecture 1-31
FEC 512 Preliminaries and Review
32. Addition Rule Example
P( Even or Asal)= P(Even) +P(Asal) - P(Even and Asal)
3/6 + 3/6 - 1/6 = 5/6
2
2,4,6 2,3,5
Lecture 1-32
FEC 512 Preliminaries and Review
33. Addition Rule for
Mutually Exclusive Events
If E1 and E2 are mutually exclusive, then
E1 E2
P(E1 and E2) = 0
So
0 utualvlye
= if m lusi
P(E1 or E2) = P(E1) + P(E2) - P(E1 and E2) c
ex
= P(E1) + P(E2)
Lecture 1-33
FEC 512 Preliminaries and Review
34. Conditional Probability
Conditional probability for any
two events E1 , E2:
P(E1 and E 2 )
P(E1 | E 2 ) =
P(E2 )
P(E2 ) > 0
where
Lecture 1-34
FEC 512 Preliminaries and Review
35. Conditional Probability Example
Of the cars on a used car lot, 70% have air
conditioning (AC) and 40% have a CD
player (CD). 20% of the cars have both.
What is the probability that a car has a CD
player, given that it has AC ?
i.e., we want to find P(CD | AC)
Lecture 1-35
FEC 512 Preliminaries and Review
36. Conditional Probability Example
(continued)
Of the cars on a used car lot, 70% have air
conditioning (AC) and 40% have a CD player (CD).
20% of the cars have both.
CD No CD Total
.2 .5 .7
AC
.2 .1
No AC .3
.4 .6 1.0
Total
P(CD and AC) .2
P(CD | AC) = = = .2857
P(AC) .7
Lecture 1-36
FEC 512 Preliminaries and Review
37. Conditional Probability Example
(continued)
Given AC, we only consider the top row (70% of the cars). Of
these, 20% have a CD player. 20% of 70% is about 28.57%.
CD No CD Total
.2 .5 .7
AC
.2 .1
No AC .3
.4 .6 1.0
Total
P(CD and AC) .2
P(CD | AC) = = = .2857
P(AC) .7
Lecture 1-37
FEC 512 Preliminaries and Review
38. For Independent Events:
Conditional probability for
independent events E1 , E2:
P(E1 | E 2 ) = P(E1 ) P(E2 ) > 0
where
P(E2 | E1 ) = P(E2 ) P(E1 ) > 0
where
Lecture 1-38
FEC 512 Preliminaries and Review
39. Multiplication Rules
Multiplication rule for two events E1 and E2:
P(E1 and E 2 ) = P(E1 ) P(E2 | E1 )
Note: If E1 and E2 are independent, then P(E2 | E1 ) = P(E2 )
and the multiplication rule simplifies to
P(E1 and E2 ) = P(E1 ) P(E2 )
Lecture 1-39
FEC 512 Preliminaries and Review
40. Bayes’ Theorem
P(Ei )P(B| Ei )
P(Ei | B) =
P(B)
P(Ei )P(B| E i )
=
P(E1 )P(B| E1 ) + P(E2 )P(B| E 2 ) + K + P(Ek )P(B| E k )
where:
Ei = ith event of interest of the k possible events
A = new event that might impact P(Ei)
Events E1 to Ek are mutually exclusive and collectively
exhaustive
Lecture 1-40
FEC 512 Preliminaries and Review
41. More Simply,
P( B | A)P(A)
P(A | B) =
P( B)
Bayes Theorem allows one to recover the
probability of the event A given B from the
probability of the individual events A,B, and
the probability of B given A.
Lecture 1-41
FEC 512 Preliminaries and Review
42. Bayes’ Theorem Example
Suppose that the probability that the price of a
stock will rise on any given day, is 0.5. Thus,
we have the prior probabilities
P(Rise)=0.5 and P(No rise)=0.5.
When it actually rises, the brokers correctly
forecasts the rise 30% of the time. When it
does not rise, they incorrectly forecast rise 6%
of the time. What is the probability that the
prices will rise if the brokers forecasted that it
will rise tomorrow?
Lecture 1-42
FEC 512 Preliminaries and Review
43. Bayes’ Theorem Example (cont.)
Let A: the event that brokers forecast that the price of the stock will
rise.
P(ARise)=30%
P(ANo Rise)= 6%
P( Rise)P(A | Rise) P( Rise)P(A | Rise)
P( Rise | A) = =
P( Rise)P(A | Rise) + P(No Rise)P(A | No Rise)
P( A)
0.30 * 0.5 0.15
= = = 0.83
0.30 * 0.5 + 0.06 * 0.5 0.15 + 0.03
As it can be seen we updated the probability of a rise (0.5) to 0.83
after we heard the brokers’s forecast of rise.
Lecture 1-43
FEC 512 Preliminaries and Review
45. Importance of Bayes Law
We update our beliefs in light of new
information.
Revising beliefs after receiving additional info
is smth that humans do poorly without the
help of mathematics.
There is a tendency to put either too little or
too much emphasis on new info
This problem can be mitigated by using
Bayes’ Law.
Lecture 1-45
FEC 512 Preliminaries and Review
46. For Further Study
For Topic I:Ruppert D. (2004), Statistics and
Finance, Springer.
For Topic II:Groebner D.F. et al.(2008)
Business Statistics.
Lecture 1-46
FEC 512 Preliminaries and Review