Eviews example4

Example #4: Stock Market Modeling using unit roots, cointegration, VAR

Attached files, see folder content at the bottom

  • data and EViews programs
  • explanatory notes and slides

The EViews exemples below allow you to reproduce example results I used in some of my courses (Financial Economic Analysis and Empirical Methods, Financial Methods and Techniques). These examples are based on the data for the UK financial markets supplied by T.C. Mills home page and used in his book The Econometric Modelling of Financial Time Series. Download the files provided in the links below (right click, and save as (keeping the same file name)). Read the document with the short guide and comments on the step-by-step operations in EViews. Start EViews, open the workfile (.wf1) with the necessary data and follow the directions in the guide. Compare your results with the results in your sheets/handouts.

EViews workfile with data (fta.wf1)

The EViews programs below shows how the use of corresponding EViews commands and shows how you can calculate variables and statistics in EViews using your own formulas. The program will open the fta.wf1 workfile and expects to find this file on a floppy in your a:\ drive. Change the path name in the program when necessary. Check the program for compatibility.

Unit roots

    • Unit root tests POWERPOINT SLIDES

NOTE: Trend in empirical research seems to be towards using the Ng-Perron unit root test (given limitations of ADF/PP tests) and a careful examination of possible structural breaks in deterministic trends.

EViews exercise: short guide (Ex1_unit root.doc)

EViews program: fta_ur.prg

Cointegration and error correction

    • Cointegration POWERPOINT SLIDES

NOTE: Trend in empirical research is towards using the Johansen VECM approach (given biases and limitations of OLS based estimators) and, increasingly, the ARDL or bounds testing approach (given weakenesses of unit root tests).

Eviews exercise: short guide (Ex2_cointegration.doc).

EViews program: fta_coint.prg

Suggested literature:

* Brooks (2002) Introductory Econometrics for Finance, Chapter 7

* Harris and Sollis (2003) Applied Time Series Modelling and Forecasting, Chapters 2-6

Example applications of unit root tests and cointegration in the empirical literature

* Mehra (1998) The bond rate and actual future inflation, FRB Richmond Economic Quarterly, vol. 84 (2) Spring 1998

Unit roots and random walks

    • Random walk tests POWERPOINT SLIDES

CAUTIONARY NOTE: Random walk tests find particular application in some empirical tests of the Efficient Market Hypothesis (EMH) in financial economics. Careful consideration of economic theory shows that these tests are absolutely useless, because random walks are neither necessary nor sufficient conditions for efficient markets. Predictability of prices or returns is not incompatible with efficient markets (although the degree of predictability is subject to arbitrage constraints). The random walk model that is perpetuated in many textbooks is an oversimplified version of economic theory that relies on unrealistic assumptions. The only valid test of efficient markets is the test on abnormal returns (returns after correction for risk-free interest rate, risk premium, and trading costs).

Eviews exercise: short guide (ExA_random walk.doc), short EViews program to calculate variance ratio statistics as described in CLM (1997) (vratiotestclm.prg)

EViews program: fta_rw.prg

Suggested literature:

* Campbell, Lo, MacKinlay (1997) The Econometrics of Financial Markets, Chapter 2

Example applications in the empirical literature

* Liu and He (1991) A variance-ratio test of random walks in foreign exchange rates, Journal of Finance, vol.46 (2)

* Lo and MacKinlay (1988) Stock market prices do not follow random walks: Evidence from a simple specification test, Review of Financial Studies, vol. 1 (1) 1988

Vector Autoregression models

    • VAR POWERPOINT SLIDES

CAUTIONARY NOTE: VAR models are a useful tool to build small economic models with limited number of variables and relying on the persitence of economic dynamics. These models are particularly useful in a forecasting environment. Empirical results suggest that extrapolation of the recent past is, by default, the best approach to reasonable forecasts (the major trick is to capture long-run equilibrium relationships, i.e. cointegration). The weakness of VAR models becomes important when researchers start to attempt to give structural interpretations to shocks and move on to impulse response functions etc.

Eviews exercise: short guide (ExB_var.doc).

EViews program: fta_var.prg

Suggested literature:

* Stock and Watson (2001) Vector autoregressions, Journal of Economic Literature, vol.15 (4)

* Keating (1992) Structural approaches to vector autoregressions, FRB StLouis Review

* Brooks (2002) Introductory Econometrics for Finance, Chapter 6

Example applications of VARs in the empirical literature

* Keating (1992) Structural approaches to vector autoregressions, FRB St.Louis Review, vol. ( ) September/October 1992

* Balke and Emery (1994) The federal funds rate as an indicator of monetary policy: Evidence from the 1980s, FRB Dallas, Economic Review, 1st Quarter 1994

* Boivin and Giannoni (2002) Assessing changes in the monetary transmission mechanism: A VAR approach, FRB New York Economic Policy Review, vol. ( ) May 2002