Professional Documents
Culture Documents
DO FinECT AppendixA 2020 PDF
DO FinECT AppendixA 2020 PDF
Daniela Osterrieder
Topics:
The Summation Operator and Descriptive Statistics
Properties of Linear Functions
Proportions and Percentages
Some Special Functions and Their Properties
Appendix A
APPENDIX A:
Basic Mathematical Tools
(3) Let yi be another series with n elements and let a and b be any
constant numbers, then
n n n
∑ (axi + byi ) = a ∑ xi + b ∑ yi
i=1 i=1 i=1
n
xi ∑ni=1 xi
∑ ≠ n
i=1 yi ∑i=1 yi
Descriptive Statistics
1 n
x̄ ≡ ∑ xi
n i=1
Descriptive Statistics
Sample Variance
n
̂) ≡ 1 ∑ (xi − x̄ )2
Var(x
n i=1
describes the variation of the data relative to the sample mean (squared
distance)
→ Use properties of summation operator to find
1 n 2 1 n 2 2
∑ (xi − x̄ ) = ∑ x − (x̄ )
n i=1 n i=1 i
Descriptive Statistics
Sample Standard Deviation
¿
Á1 n √
À ∑ (xi − x̄ )2 1 n
≡Á (≠ ∑ (xi − x̄ )2 !)
n i=1 n i=1
describes the variation of the data relative to the sample mean (absolute
distance)
→
¿ ¿
Á1 n Á1 n 2
À ∑ (xi − x̄ )2 = Á
Á À ∑ x − (x̄ )2
n i=1 n i=1 i
Descriptive Statistics
Sample Covariance
̂ 1 n
Cov(x , y) ≡ ∑ (xi − x̄ ) (yi − ȳ)
n i=1
describes the common variation of two data sets (x and y) relative to the
respective sample mean
→ Use properties of summation operator to find
1 n 1 n
∑ (xi − x̄ ) (yi − ȳ) = ∑ xi yi − ȳ x̄
n i=1 n i=1
Descriptive Statistics
Sample Correlation
̂, y) ≡ √
1
n∑ni=1 (xi − x̄ ) (yi − ȳ)
Corr(x √
1 n 2 1 n 2
n (x
∑i=1 i − x̄ ) n ∑i=1 (yi − ȳ)
describes the common variation of two data sets (x and y) relative to the
respective sample mean, standardized by the two respective sample
standard deviations. Note that
̂, y) ≤ 1
−1 ≤ Corr(x
Descriptive Statistics
Is the number in the set {xi } such that 50% of the values in {xi } are
larger than it and 50% are smaller
Example: Let {xi } = {5, 20, 0.5, −3, −10, 2, 0, 9, −1.5} and
{yi } = {5, 20, 0.5, −3, −10, 2, 0, 9, −1.5, 1}
Sort the data in increasing order, i.e. {−10, −3, −1.5, 0, 0.5, 2, 5, 9, 20}
and {−10, −3, −1.5, 0, 0.5, 1, 2, 5, 9, 20}
If n is odd (as in set {xi }), then the median is simply the number in
the middle
→ The median of {xi } is 0.5
Descriptive Statistics
If n is even (as in set {yi }), then the median is the average of the
two numbers in the middle
→ The median of {yi } is 21 (0.5 + 1) = 0.75
Linear Functions
If x and y are two variables that are related by the following function
y = β0 + β1 x
Example
Assume that the relationship between a country’s inflation and
unemployment level is given by
Example
If x1 , x2 and y are three variables that are related by the following function
y = β 0 + β 1 x1 + β 2 x2
Proportion
42
⇒ The proportion of adults with a high-school degree is 50 = 0.84
Percentage
42
⇒ The percentage of adults with a high-school degree is 50 × 100 = 84%
→ Proportion × 100 = Percentage
Percentage Change
⇒ The percentage change in income between 1994 and 1995 is
x1 −x0 ∆x
x0 × 100 = x0 × 100
→ (Proportionate/relative change) × 100 = Percentage change
Example:
The (one-day holding period) return on the S&P 500 between
December 31, 2014 and January 2, 2015 was 0.034%
Returns are often reported as proportionate price changes, i.e. the net
return on the S&P 500 between December 31, 2014 and January 2,
2015 was 0.00034
The gross return on the S&P 500 between December 31, 2014 and
January 2, 2015 was 1.00034
Special Functions
A nonlinear function that plays a crucial role in financial econometrics, is
the natural logarithm. Often referred to as the log function
y = log(x ) = ln(x ) = loge (x ), x >0
x1 x1 − x0 ∆x ∆x
∆ log(x ) = log(x1 ) − log(x0 ) = log( ) = log(1 + ) = log(1 + )≈
x0 x0 x0 x0
∆x
if the proportionate/relative change is small (i.e. if x0 ≈ 0)
It follows that
∆x
100 × ∆ log(x ) ≈ 100 ×
x0
that is approximately equal to the percentage change
E.g.
100 × (log(41) − log(40)) = 2.46926
41 − 40 ∆x
100 × = 2.5 (Note: = 0.025 ≈ 0)
40 x0
D. Osterrieder (RBS) Financial Econometrics Spring Semester 2020 24 / 32
Some Special Functions and Their Properties
∆ log(y)
Recall that β 1 is the marginal effect of log(x ) on log(y), i.e. β 1 = ∆ log(x )
Percentage change in y
=
Percentage change in x
Example:
Let q be the demanded quantity of e.g. oil and let p be the price. Assume
log(q) = 4.7 − 1.25 × log(p)
→ If the price of oil increases by 1%, the demanded quantity decreases by
1.25%. The price elasticity of demand is -1.25.
log(y) = β 0 + β 1 x
then
∆y
100 × ∆ log(y) 100 × y0
100 × β 1 = ≈
∆x ∆x
Percentage change in y
=
Unit change in x
Example:
Let wage be hourly wage in US$ and educ is years of education. Assume
log(wage) = 2.78 + 0.094 × educ
y = β 0 + β 1 log(x )
then
β1 ∆y ∆y
= ≈
100 100 × ∆ log(x ) 100 × ∆x
x0
Unit change in y
=
Percentage change in x
β1
If x changes by 1%, then y changes by 100 units.
Example:
Let hours be hours worked per week. Assume hours = 33 + 45.1 × log(wage)
→ If hourly wage increases by 1%, the hours worked per week increases by
0.451 (i.e. approximately half an hour).
Finally, notice that the inverse function of the log function is the
exponential function
y = log(x ) ⇔ x = exp{y} = e y
Can be re-written as