Random walk test
Random Walk Test
For KOREAN MARKET
CONTENTS
Part 1
Introduction
Part 2
Data
Part 3
Empirical Analysis
Part 4
Extension Study
Part 5
Summary
INTRODUCTION
ABOUT THE TOPIC
Market efficiency and random walk theory
This topic is fundamental and closely related to other topics.
The validity of the random walk hypothesis is prerequisite for a lot of financial models
Allow to apply knowledge learn from class
INTRODUCTION
WHY KOREA MARKET
We are interested in emerging market
There are already a lot of researches focus on regions like
Hong Kong and China, Japan
Not sufficient data for countries like India
Korean market has nearly 40 years data
INTRODUCTION
DATA
KOSPI (Korean Composites Stock Price Index)
From 1/1/1975 to 12/31/2012, 38 years in total
We divide the data into small groups according to the upward and downward trend
INTRODUCTION
DATA
Time series plot of the index
INTRODUCTION
DATA
Time series plot of the index (group by complete cycle)
INTRODUCTION
DATA
IMPIRICAL ANALYSIS
UNIT ROOT TESTS
We use the ADF test to test weather unit root exists in the stock price series by estimating the following equation through OLS regression.
null hypothesis
H0: ρ0=0 (There is unit root)
H1: ρ0≠0 (There is no unit root)
If we reject the H0, it implies that the there is no random walk. TEST RESULT 1
Random walk appears in most times, as the null hypothesis of a unit-root is not rejected for any period. But we should also be careful about this result according to the paper.
TEST RESULT 2
Random walk exists in each period, as the null hypothesis of a unit-root is not rejected for any period. RUNS TEST
TEST WHETHER SUCCESSIVE PRICE CHANGES ARE
INDEPENDENT OF EACH OTHER OR RANDOM WALK.
TWO APPROACHES TO DEFINE THE SIGN OF RETURNS:
1: RETURN>0, THEN POSITIVE
RETURN<0, THEN NEGATIVE
2: RETURN>MEAN RETURN, THEN POSITIVE
RETURN<MEAN RETURN, THEN NEGATIVE
AN ADVANTAGE OF SECOND APPROACH: ALLOW FOR AND
CORRECT THE EFFECT OF AN EVENTUAL
For KOREAN MARKET
CONTENTS
Part 1
Introduction
Part 2
Data
Part 3
Empirical Analysis
Part 4
Extension Study
Part 5
Summary
INTRODUCTION
ABOUT THE TOPIC
Market efficiency and random walk theory
This topic is fundamental and closely related to other topics.
The validity of the random walk hypothesis is prerequisite for a lot of financial models
Allow to apply knowledge learn from class
INTRODUCTION
WHY KOREA MARKET
We are interested in emerging market
There are already a lot of researches focus on regions like
Hong Kong and China, Japan
Not sufficient data for countries like India
Korean market has nearly 40 years data
INTRODUCTION
DATA
KOSPI (Korean Composites Stock Price Index)
From 1/1/1975 to 12/31/2012, 38 years in total
We divide the data into small groups according to the upward and downward trend
INTRODUCTION
DATA
Time series plot of the index
INTRODUCTION
DATA
Time series plot of the index (group by complete cycle)
INTRODUCTION
DATA
IMPIRICAL ANALYSIS
UNIT ROOT TESTS
We use the ADF test to test weather unit root exists in the stock price series by estimating the following equation through OLS regression.
null hypothesis
H0: ρ0=0 (There is unit root)
H1: ρ0≠0 (There is no unit root)
If we reject the H0, it implies that the there is no random walk. TEST RESULT 1
Random walk appears in most times, as the null hypothesis of a unit-root is not rejected for any period. But we should also be careful about this result according to the paper.
TEST RESULT 2
Random walk exists in each period, as the null hypothesis of a unit-root is not rejected for any period. RUNS TEST
TEST WHETHER SUCCESSIVE PRICE CHANGES ARE
INDEPENDENT OF EACH OTHER OR RANDOM WALK.
TWO APPROACHES TO DEFINE THE SIGN OF RETURNS:
1: RETURN>0, THEN POSITIVE
RETURN<0, THEN NEGATIVE
2: RETURN>MEAN RETURN, THEN POSITIVE
RETURN<MEAN RETURN, THEN NEGATIVE
AN ADVANTAGE OF SECOND APPROACH: ALLOW FOR AND
CORRECT THE EFFECT OF AN EVENTUAL