Negative adjusted R square
Effiezal
I have run 2SLS + SUR (3SLS?) with three equations (three endogenous
variables). One of my equation has a -ve adjusted R square. my current
equations are:
Performance = Institutional ownership + Corporate Governance Index +
instruments(mkt risk, market to book value, managerial
ownership(piecewise method), blockholders,remuneration and director
ethnicity) + control (leverage, size, industries)
Institutional onwership = performance + coporate governance index +
instruments(mkt risk, firm risk, managerial
ownership(direct),blockholders, dividend yield, research adn
development, director enthnicity) + control (leverage, size,
indsutries)
Corporate Governance Index = institutional ownership +performance +
instruments( auditors identity, directors ethnicity,
intangibles/assets, board size) + control (leverage,size,industries)
The institutional ownership equation has a negative adjusted r square
AFTER running 2SLS+SUR.
Note:
This is in panel data (unbalanced) + FEM (year dummies). Dropping the
FEM will not improved my R square.
Need some comment. A paper by Klapper and Love (2004) argued that there
is no statistical siginficant regarding adjusted R square after running
3SLS , any comments?
Efi.