INTRODUCTION TO FINANCE

REVIEW OF STATISTICS













2
A Review of Statistics

In a world with uncertainty, we typically describe possible
future values or rates of return of a security and the relative
likelihoods of these values by means of a probability
distribution over future values or rates of returns.


For example, a continuous distribution is given by:
Probability


% Return





3
A Review of Statistics
Discrete distribution - states of the world model

Security Return in %
State Prob. 1 2 3 4
Expansion 0.10 5.00 3.50 6.00 2.50
Normal 0.40 7.00 4.50 5.00 -0.50
Recession 0.30 6.00 4.00 7.00 1.00
Depression 0.20 -2.00 0.00 4.00 13.00


4

A Review of Statistics
Expected Return:

=
= =
S
1 s
s , i
R
~
s
p R )
i
R
~
( E
where E R R mean ected return
i
(
~
) (exp ) = =
s state in i urity sec of return unknown
s , i
R
~
=
s state of y probabilit
s
p =
S total number of states occurring = .
Examples:
E R (
~
)
1
= 0.10 (5.0) + 0.40 (7.0) + 0.30 (6.0) + 0.20 (-2.0)
= 4.70%
E R (
~
)
2
= 0.10 (3.5) + 0.40 (4.5) + 0.30 (4.0) + 0.20 (0)
= 3.35%
E R (
~
)
3
= 0.10 (6.0) + 0.40 (5.0) + 0.30 (7.0) + 0.20 (4.0)
= 5.50%
E R (
~
)
4
= 0.10 (2.5) + 0.40 (-0.5) + 0.30 (1.0) + 0.20 (13.0)
= 2.95%.


5
A Review of Statistics
Spread of distribution of outcomes:
.
i
2
1
S
1 s
2
) R
s , i
R
~
(
s
p )
i
R
~
( Std o =

=
=

Note: { } .
2
i
2
)
i
R
~
( Std )
i
R
~
( Var o = =


Examples:
o
1
2
= 0.10 (5.0 - 4.7)
2
+ 0.40 (7.0 - 4.7)
2

+ 0.30 (6.0 - 4.7)
2
+ 0.20 (-2.0 - 4.7)
2

= 11.61
o
1
= 3.41%.
o
2
2
= 0.10 (3.5 - 3.35)
2
+ 0.40 (4.5 - 3.35)
2

+ 0.30 (4.0 - 3.35)
2
+ 0.20 (0.0 - 3.35)
2

= 2.90.

o
2
=1.70%.

o
3
= Calculate
o
4
= Calculate


6
A Review of Statistics
Covariance:
.
j , i
S
1 s
)
j
R
s , j
R
~
)(
i
R
s , i
R
~
(
s
p )
j
R
~
,
i
R
~
( Cov o =

=
=

When the returns of securities i and j tend to deviate from
their means in the same direction they are called positively
correlated or they covary positively [that is, 0 )
j
R
~
,
i
R
~
( Cov > ].

When they tend to deviate from their means in opposite
directions, they are negatively correlated or they covary
negatively [that is, 0 )
j
R
~
,
i
R
~
( Cov < ].


When the returns of securities i and j are unrelated with each
other, then 0 )
j
R
~
,
i
R
~
( Cov = and they are uncorrelated.




7
A Review of Statistics
Examples:
o
1 2 ,
= 0.10 (5.0 - 4.7) (3.5 - 3.35)
+ 0.40 (7.0 - 4.7) (4.5 - 3.35)
+ 0.30 (6.0 - 4.7) (4.0 - 3.35)
+ 0.20 (-2.0 - 4.7) (0.0 - 3.35)
= 5.81.
o
1 3 ,
= 0.10 (5.0 - 4.7) (6.0 - 5.5)
+ 0.40 (7.0 - 4.7) (5.0 - 5.5)
+ 0.30 (6.0 - 4.7) (7.0 - 5.5)
+ 0.20 (-2.0 - 4.7) (4.0 -5.5)
= 2.15.

o
1 4 ,
= Calculate
o
2 3 ,
= Calculate
o
2 4 ,
= Calculate
o
3 4 ,
= Calculate







8
A Review of Statistics

Correlation--A standardized Covariance:

. 1
j , i
1 where
,
j i
j , i
)
j
R
~
( Std )
i
R
~
( Std
)
j
R
~
,
i
R
~
( Cov
j , i
+ < p <
o o
o
= = p



- Since )
i
R
~
( Std and )
j
R
~
( Std are always positive, the sign of
j , i
p depends on the sign of .
j , i
o

- What is the relation between
j , i
p and
i , j
p ?






9
A Review of Statistics
Regression Analysis:
- The general idea of a regression is that it involves fitting a
line through a set of points. In particular, consider the
following regression model:


~ ~ ~
y x = o + | + c
where


~
y = the value of the dependent variable

~
x = the value of the independent variable

~
c = a random error with E(
~
c ) = 0
o = the Y axis intercept
| = the slope of the line.



10
A Review of Statistics

The picture that we are trying to capture is as follows:


{
Y
Y
i
o
|
c
i
X
X
1




11
A Review of Statistics

- In general the idea is that there is a linear relationship
between y and x which is not quite exact.

Since we do not know the relationship we want to estimate
o and | that minimize the sum of the deviations squared
from the fitted line.

The best fit is given by the slope:


( )
.
2
x
yx
2
x
x
~
, y
~
Cov
x , y
o
o
=
o
= |


- How would you write the relation between security 1 and
2?
- What is the relation between
1 , 2
| and
2 , 1
| ?

- What is the relation between
1 , 2
| and
2 , 1
p ?



12
A Review of Statistics
Examples of Correlation and Regression:
Perfect Positive Correlation

( )( )
. 00 . 1
1.70 3.41
5.81
2 1
1,2
1,2
= =
o o
o
= p
Note:
~ ~
. R
2
1.00 0.50R
1
= +

Less Than Perfect Correlation:

( )( )
p
o
o o
1,3
1,3
1 3
2.15
3.41 1.12
= = = 0563 . .
Note:
~ ~
~
. R
3
4.63 0.185R
1
e
3
= + +

Perfect Negative Correlation

( )( )
p
o
o o
1 4
1 4
1 4
17 42
341 511
100
,
,
.
. .
. . = =

=
Note:
~ ~
R
4
10.00 1.50R
1
. =