Where time is expressed now as T and g0 captures the growth effects of ignored and trended
variables on TFP. It is generally hard to include more than a few variables in the Z vector
because these growth enhancing variables are generally trended and correlated. Therefore, it
is hard to estimate accurately the individual growth effects of these variables in the Z vector.
It is well known that the Steady Sate Growth Rate, SSGR, in the Solow model equals the rate
of growth of A i.e., TFP. We have selected 4 variables for inclusion into the Z vector and
these are Dreher’s comprehensive measure of globalization (GLO), an index of institutional
reforms (INSTI), the rate of inflation (DLP) and the ratio of current government expenditure
to GDP (GRAT). Definitions of the variables and sources of data are in the appendix. DLP
and GRAT proxy good economic policies and institutional reforms have been emphasized as
a growth improving variable by aid giving agencies like the IMF and the World Bank. Other
potential variables for inclusion into the Z vector are overseas development aid, other
measures of economic stability such as the ratio of budget deficit to GDP and stock of
human capital etc. In fact there is no end to the list such potential variables that can be
included into the Z vector. In this context Durlauf, Johnson, and Temple (2005) have noted
that the number of potential growth improving variables, used in various empirical works, is
as many as 145. We have not added any more additional variables into the Z vector partly
due to the limitations of data and possible multicolinearity between the variables. However,
the intercept term viz., g0 should capture the effects of some of these ignored variables if
they have significant growth effects.
4. Empirical Results
The specifications in equations (1) and (2) can be estimated with the standard penal data
methods of fixed and random effects. However, the specifications in (3) cannot be easily
estimated with these methods because of the nonlinearity of the variables in TFP.
Generalized Method of Moment (GMM) proposed by Arellano and Bond (1991) is the
commonly employed estimation procedure to estimate the parameters in a dynamic panel
data model with nonlinearities in the variables. In this method first differenced transformed
series are used to adjust for the unobserved individual specific heterogeneity in the series.