panel cointegration

[sohail]

hi, I am new to eviews. I am using eviews6. I am doing panel data analysis.

Dependent variable: stock price volatility
independents: dividend yield , dividend payout, earning per share, liquidity, leverage and size
Data: 2005-2012
Companies: all FTSE-100 companies

Please advise:
1) how to do unit root test on panel data
2) how to do cointegration test

thanks

sohail

[EViews Glenn]

1. Structure your data as a panel, open the series in question, and perform a unit root test.

2. Panel cointegration testing requires EViews 7 or later. Panel cointegration estimation requires EViews 8.

[sohail]

please refere to my first post and please help me on the following:

opening panel data
File- new- workfile - dated regular frequency - start date 2005 - end date 2012 - ok then object - new object - pool - under cross section identifiers i pasted 100 abbrivated company names e-g ABS_ then I click proc - import pool data - then paste all variables in format like this SPV_? DY_? DP_? EPS_? SZ_? LIQ_? LEVERAGE_?
under ordinary and pool series column then i click ok. my question is that is this the right way to open panel data??

unit root test
In pool window, i click view and then unit root test. by this way, i am doing the individual variable unit root testing by typing variables in e-g spv_? format and by selecting level and then individual intercept. By doing so one variable named dy (dividend yield) is not statonary at level instead it get stationary at 1st difference and rest of the variables are stationary at level.
am i doing unit root testing in the right way?

unit root test by making group in pool window

this is how i am doing panel unit root test: in panel window by clicking proc - make a group - paste SPV_? DY_? DP_? EPS_? SZ_? LIQ_? LEVERAGE_? in series list window and then ok. a new window opens then i click view - unit root test- slect summary and then level and individual intercept the following result comes up

Group unit root test: Summary
Series: SPV_AAP_
Date: 12/28/13 Time: 13:26
Sample: 2005 2012
Exogenous variables: Individual effects
Automatic selection of maximum lags
Automatic selection of lags based on SIC: 0 to 1
Newey-West bandwidth selection using Bartlett kernel

Cross-
Method Statistic Prob.** sections Obs
Null: Unit root (assumes common unit root process)
Levin, Lin & Chu t* -145.575 0.0000 640 4274

Null: Unit root (assumes individual unit root process)
Im, Pesaran and Shin W-stat -16.0046 0.0000 639 4271
ADF - Fisher Chi-square 1915.57 0.0000 640 4274
PP - Fisher Chi-square 2136.84 0.0000 640 4462

** Probabilities for Fisher tests are computed using an asymptotic Chi
-square distribution. All other tests assume asymptotic normality.

according to these results all variables are stationary at level and i can use panel data regression. however when i use panal data regression and by doing hausman test i use random effect model then r squared value is very small like 0.014 and adjusted r square is 0.007. what should i do? I am stuck here please help