You are on page 1of 15

Linear Regression with One Regressor

1-15

FRMFinancial Risk Manager

Regression Analysis
A regression analysis has the goal of measuring how changes in one variable,
called a dependent or explained variable can be explained by changes in one or
more other variables called the independent or explanatory variables. The
regression analysis measures the relationship by estimating an equation (e.g.,
linear regression model). The parameters of the equation indicate the
relationship.
A scatter plot is a visual representation of the relationship between the
dependent variable and a given independent variable. It uses a standard twodimensional graph where the values of the dependent, or Y variable, are on the
vertical axis, and those of the independent, or X variable, are on the horizontal
axis.

2-15

FRMFinancial Risk Manager

Population Regression Function


Lockup(yrs)
5
6
7
8
9
10

10
17
16
15
21
20

Fugure 1:Hedge Fund Data


Returns(%) per year
14
14
15
12
12
15
16
10
19
19
13
13
20
19
15
16
20
16
20
18
17
21
23
19

Average Return
13
14
16
17
19
20

Figure 2:Return Over Lockup Period

3-15

FRMFinancial Risk Manager

Population Regression Function


E (return | lockup period) = B0 + B1 (lockup period)
Or more generally:

E (Yi | Xi) = B0 + B1 (Xi)


Intercept Coefficient

Slope Coefficient

Yi B0 B1 X i i
Error Term

4-15

FRMFinancial Risk Manager

Sample Regression Function and Ordinary least squares (OLS)

Yi b0 b1 X i ei
minimize e Yi - (b0 b1 Xi )

2
i

Yi b0 b1X i ei

e4

e2

e1

e3

Yi b0 b1X i
X

5-15

FRMFinancial Risk Manager

The Results of Ordinary least squares (OLS)

6-15

FRMFinancial Risk Manager

201405

7-15

FRMFinancial Risk Manager

201405
48. You are conducting an ordinary least squares regression of the returns on stocks Y
and X as Y=a + b X + based on the past three years daily adjusted closing price
data. Prior to conducting the regression, you calculated the following information
from the data:
Sample covariance
Sample Variance of Stock X
Sample Variance of Stock Y
Sample mean return of stock X

0.000181
0.000308
0.000525
-0.03%

Sample mean return of Stock Y

0.03%

What is the slope of the resulting regression line?


A. 0.35
B. 0.45
C. 0.59
D. 0.77

8-15

Notes 128

0.000181
0.5877
0.000308

FRMFinancial Risk Manager

Example

Consider two stocks, A and B. Assume their annual returns are


jointly normally distributed, the marginal distribution of each stock
has mean 2% and standard deviation 10%, and the correlation is
0.9. What is the expected annual return of stock A if the annual
return of stock B is 3%?
A.
B.
C.
D.

2%
2.9%
4.7%
1.1%

Answer: B
9-15

FRMFinancial Risk Manager

Assumptions Underlying Linear Regression


Linear regression requires a number of assumptions. Most of the major assumptions
pertain to she regression models residual term (i.e., error term). Three key assumptions
are as follows:
1. The expected value of the error term, conditional on the independent variable, is
zero ( E ( i | X i ) 0 )
2. All (X, Y) observations are independently and identically distributed (i.i.d.).
3. It is unlikely that large outliers will be observed in the data. Large outliers have the
potential to create misleading regression results.
Additional assumptions include:
4. A linear relationship exists between the dependent and independent variable.
5. The model is correctly specified in that it includes the appropriate independent
variable and does not omit variables.
6. The independent variable is uncorrelated with the error terms.
7. The variance of i is constant for all Xi : Var( i | X i ) 2
8. No serial correlation of the error terms exists

[i.e.,Corr(i ,i+j ) 0 for j=1,2,3...]

9. The error term is normally distributed

10-15

FRMFinancial Risk Manager

The Coefficient of Determination (R2)


Y

Yi b0 b1 X i

(Yi Yi ) SSR
__

(Yi Y ) TSS

(Yi Y ) ESS

__

Y
b0

Total sum of squares = explained sum of squares + sum of squared residuals

(Y
TSS

__

Y)

11-15

__

(Y Y )

ESS

ESS
SSR
R
1
TSS
TSS
2

(Y Y )
i

SSR

2 R2 R2

FRMFinancial Risk Manager

The Standard Error of the Regression


The standard error of the regression (SER) measures the degree of variability
of the actual Y-values relative to the estimated Y-values from a regression
equation. The SER gauges the "fit" of the regression line. The smaller the
standard error, the better the fit.
The SER is the standard deviation of the error terms in the regression. As such,
SER is also referred to as the standard error of the residual, or the standard error
of estimate (SEE).
In some regressions, the relationship between the independent and dependent
variables is very strong (e.g., the relationship between 10-year Treasury bond
yields and mortgage rates). In other cases, the relationship is much weaker (e.g.,
the relationship between stock returns and inflation). SER will be low (relative to
total variability) if the relationship is very strong and high if the relationship is
weak.
12-15

FRMFinancial Risk Manager

13-15

FRMFinancial Risk Manager

14-15

FRMFinancial Risk Manager

FRM

15-15

FRMFinancial Risk Manager