KYIV SCHOOL OF ECONOMICS Financial Econometrics (2nd part): Introduction to Financial Time Series May 2011 Instructor: Maksym Obrizan Lecture notes II.

Slides:



Advertisements
Similar presentations
Autocorrelation Functions and ARIMA Modelling
Advertisements

Autocorrelation and Heteroskedasticity
Volatility in Financial Time Series
11-1 Empirical Models Many problems in engineering and science involve exploring the relationships between two or more variables. Regression analysis.
Threshold Autoregressive. Several tests have been proposed for assessing the need for nonlinear modeling in time series analysis Some of these.
Hypothesis Testing Steps in Hypothesis Testing:
CHAPTER 21 Inferential Statistical Analysis. Understanding probability The idea of probability is central to inferential statistics. It means the chance.
Part II – TIME SERIES ANALYSIS C5 ARIMA (Box-Jenkins) Models
6-1 Introduction To Empirical Models 6-1 Introduction To Empirical Models.
Correlation and regression
© 2010 Pearson Prentice Hall. All rights reserved Least Squares Regression Models.
The Multiple Regression Model Prepared by Vera Tabakova, East Carolina University.
The Simple Linear Regression Model: Specification and Estimation
Chapter 10 Simple Regression.
Introduction to Volatility Models From Ruey. S. Tsay’s slides.
On Threshold ARCH Models with Gram- Charlier Density Xuan Zhou and W. K. Li Department of Statistics and Actuarial Science,HKU.
1 Ka-fu Wong University of Hong Kong Volatility Measurement, Modeling, and Forecasting.
PREDICTABILITY OF NON- LINEAR TRADING RULES IN THE US STOCK MARKET CHONG & LAM 2010.
Financial Econometrics
KYIV SCHOOL OF ECONOMICS Financial Econometrics (2nd part): Introduction to Financial Time Series May 2011 Instructor: Maksym Obrizan Lecture notes III.
Volatility Chapter 9 Risk Management and Financial Institutions 2e, Chapter 9, Copyright © John C. Hull
Slide Copyright © 2010 Pearson Education, Inc. Active Learning Lecture Slides For use with Classroom Response Systems Business Statistics First Edition.
Dealing with Heteroscedasticity In some cases an appropriate scaling of the data is the best way to deal with heteroscedasticity. For example, in the model.
Business Statistics - QBM117 Statistical inference for regression.
Relationships Among Variables
Lecture 5 Correlation and Regression
Time-Varying Volatility and ARCH Models
Marketing Research Aaker, Kumar, Day and Leone Tenth Edition
Chapter 13: Inference in Regression
1 MADE WHAT IF SOME OLS ASSUMPTIONS ARE NOT FULFILED?
Review of Statistical Inference Prepared by Vera Tabakova, East Carolina University ECON 4550 Econometrics Memorial University of Newfoundland.
Regression Method.
STATISTICS: BASICS Aswath Damodaran 1. 2 The role of statistics Aswath Damodaran 2  When you are given lots of data, and especially when that data is.
© 1998, Geoff Kuenning Linear Regression Models What is a (good) model? Estimating model parameters Allocating variation Confidence intervals for regressions.
Random Regressors and Moment Based Estimation Prepared by Vera Tabakova, East Carolina University.
Various topics Petter Mostad Overview Epidemiology Study types / data types Econometrics Time series data More about sampling –Estimation.
Byron Gangnes Econ 427 lecture 3 slides. Byron Gangnes A scatterplot.
Inference for Regression Simple Linear Regression IPS Chapter 10.1 © 2009 W.H. Freeman and Company.
© Copyright McGraw-Hill Correlation and Regression CHAPTER 10.
NEW FRONTIERS FOR ARCH MODELS Prepared for Conference on Volatility Modeling and Forecasting Perth, Australia, September 2001 Robert Engle UCSD and NYU.
STAT 497 LECTURE NOTE 9 DIAGNOSTIC CHECKS 1. After identifying and estimating a time series model, the goodness-of-fit of the model and validity of the.
MULTIVARIATE TIME SERIES & FORECASTING 1. 2 : autocovariance function of the individual time series.
Estimation Method of Moments (MM) Methods of Moment estimation is a general method where equations for estimating parameters are found by equating population.
Copyright © 2011 by The McGraw-Hill Companies, Inc. All rights reserved. McGraw-Hill/Irwin Simple Linear Regression Analysis Chapter 13.
GARCH Models Þættir í fjármálum Verkefni 1-f Bjartur Logi Ye Shen
Example x y We wish to check for a non zero correlation.
The Box-Jenkins (ARIMA) Methodology
13 th AFIR Colloquium 2003 The estimation of Market VaR using Garch models and a heavy tail distributions The dynamic VaR and The Static VaR The Garch.
ARCH AND GARCH V AIBHAV G UPTA MIB, D OC, DSE, DU.
Linear Regression Models Andy Wang CIS Computer Systems Performance Analysis.
Lecture 8 Stephen G. Hall ARCH and GARCH. REFS A thorough introduction ‘ARCH Models’ Bollerslev T, Engle R F and Nelson D B Handbook of Econometrics vol.
MODELING VOLATILITY BY ARCH- GARCH MODELS 1. VARIANCE A time series is said to be heteroscedastic, if its variance changes over time, otherwise it is.
Analysis of financial data Anders Lundquist Spring 2010.
Forecasting. Model with indicator variables The choice of a forecasting technique depends on the components identified in the time series. The techniques.
K. Ensor, STAT Spring 2004 Volatility Volatility – conditional variance of the process –Don’t observe this quantity directly (only one observation.
Vera Tabakova, East Carolina University
Prediction, Goodness-of-Fit, and Modeling Issues
Chapter 6: Autoregressive Integrated Moving Average (ARIMA) Models
Part Three. Data Analysis
Charles University Charles University STAKAN III
STAT 497 LECTURE NOTE 9 DIAGNOSTIC CHECKS.
Linear Regression Models
CHAPTER 29: Multiple Regression*
Undergraduated Econometrics
Chapter 3 Statistical Concepts.
Tutorial 1: Misspecification
Lecturer Dr. Veronika Alhanaqtah
Essentials of Statistics for Business and Economics (8e)
Threshold Autoregressive
BOX JENKINS (ARIMA) METHODOLOGY
Presentation transcript:

KYIV SCHOOL OF ECONOMICS Financial Econometrics (2nd part): Introduction to Financial Time Series May 2011 Instructor: Maksym Obrizan Lecture notes II # 2. This lecture: (i)Very brief summary of ARCH-GARCH and their shortcomings (ii)A few more advanced models (TAR, MSA) Sometimes series {r t } may be with no or minor serial correlation but it is still dependent… Log of Intel stock returns from January 1973 to December 1997 on the bottom left slide # 3 and of squared Intel returns on slide # 4 # 3.# 4.

# 5.# 6. Series are uncorrelated but dependent: volatility models attempt to capture such dependence Define shock or mean-corrected return # 7. Then ARCH(m) model assumes In practice, the error term is assumed to follow the standard normal of a standardized Student-t distribution # 8. Practical way of building an ARCH model (i)Build an ARMA model for the return series to remove any linear dependence in data If the residual series indicates possible ARCH effects – proceed to (ii) and (iii) (ii)Specify the ARCH order and perform estimation (iii)Check the fitted ARCH model for necessary refinements

# 9. Fitting an ARCH Model: Model Checking: Obtain standardized shocks Then use Ljung-Box statistic on to check the adequacy of the mean equation and on to check the validity of the volatility equation. Use kurtosis, skewness and QQ-plot of to check if normal distribution is applicable # 10. Shortcomings of ARCH model (i)Positive and negative shocks have the same effects on volatility  MOTIVATION FOR TAR, MSA MODELS! (ii)ARCH model is restrictive – parameters are constrained by certain intervals for finite moments etc (iii) Sometimes not parsimonious models: use GARCH # 11. GARCH model: Weaknesses of GARCH model are similar to those of ARCH: symmetric response to negative and positive shocks etc # 12. NOTES

# 13. Application: daily log returns of IBM stock from July 3, 1962 to December 31, 1999 All the estimates (except the coefficient of r t-2 ) are highly significant # 14. In addition, the Ljung-Box statistics of the standardized residuals is Q(10) = (p- value of 0.33) and of the squared standardized residuals is Q(10) = (p- value of 0.29) # 15. The unconditional mean of r t in this model is while in the sample it is only What if the model is misspecified?  Motivation for nonlinear models such as TAR and MSA # 16. Threshold Autoregressive (TAR) model

# 17. Consider a simple two-regime TAR model # 18. # 19. Observe that this model has coefficient -1.5 However, despite this fact it is stationary and geomertically ergodic if Ergodic theorem – statistical theorem showing that the sample mean of x t converges to the mean of x t # 20. Model behavior depends on x t -1 : When it is negative then When it is positive then Question: Which regime will have more observations?

# 21. In addition, TAR model has non-zero mean even though the constant terms are zero (think of an AR(m) model with zero constant for a comparison) Re-consider slide # 15 with AR(2)-GARCH(1,1) model of IBM stock: the unconditional mean of overpredicted the sample mean of Estimate AR(2)-TAR-GARCH(1,1) model and refine it (remove insignificant term in volatility equation) # 22. AR(2)-TAR-GARCH(1,1) of IBM stock # 23. Model fit All coefficients are significant at 5% The unconditional mean? The Ljung-Box statistics applied to standardized residuals does not indicate serial correlations or conditional heteroscedasticity # 24. NOTES:

# 25. Convenient to re-write TAR-GARCH(1,1):# 26. Recall the integrated GARCH model (IGARCH is a unit-root GARCH model) For example, IGARCH(1,1) is defined as The unconditional variance of a t, and thus of r t, is not defined Meaning: Occasional level shifts in volatility? IGARCH(1,1) with is used in RiskMetrics (Value at Risk calculating) # 27. Thus, under nonpositive deviation the volatility follows an IGARCH(1,1) model without a drift With positive deviation, the volatility has a persistent parameter =0.931 which is <1 giving rise to GARCH(1,1) Conclusion: # 28. NOTES:

# 29. Markov Switching Model# 30. Application to the US quarterly real GNP # 31. Cont’d# 32. Notes

# 33. Nonlinearity tests: Parametric tests The RESET Test for a linear AR(p) model Basic idea: if a linear AR(p) model is adequate then a 1 and a 2 should be zero. # 34. Apply F statistic with g and T-p-g degrees of freedom # 35. Nonlinearity tests: Nonparametric tests Q-statistic of Squared Residuals The null hypothesis of the statistic is # 36. NOTES

# 37. Application to the US quarterly civilian unemployment from 1948 to 1993 based on Montgomery, Zarnowitz, Tsay and Tiao (1998) # 38. TAR model # 39. MSA model# 40. NOTES