|
Following Mankiw, Romer, and Weil (1992), Temple (1998) estimates a linear conditional convergence regression, or growth regression using data for 78 countries and covering the period 1960-1985, that is of the form:
These variables reflect differences in factor
accumulation across countries and are expected to control for growth
differences in equilibrium. The fifth variable () is the logarithm of
output per working-age person at the beginning of the period, and is
expected to capture the Neoclassical convergence effect due to diminishing
returns to reproducible factors, that tends to favor poorer countries. The
last four exogenous variables (
) are dummies for
respectively sub-Saharan Africa, Latin America and the Caribbean, East Asia,
and the industrialized countries of the OECD plus Israel. These variables
allow us to control for differences in efficiency, variation of which has
been found to be essentially intercontinental.
To estimate such a multiple linear regression, we first read the data Temple (1998) analyzed and that are stored in temple.dat and define both the independent and the dependent variables.
z=read("temple.dat") x=z[,2:10] y=z[,1]
Second, we load the
stats
quantlib and use the following
XploRe
code that computes the linear regression of y on x, and stores
the values of the estimated parameters as well as their respective standard
error, -statistic, and
-value.
library("stats") {b,bse,bstan,bpval}=linreg (x,y)
This quantlet also provides as an output the following ANOVA (ANalysis Of VAriance) table that allows us to infer:
A N O V A SS df MSS F-test P-value _________________________________________________________________ Regression 10.957 9 1.217 33.397 0.0000 Residuals 2.479 68 0.036 Total Variation 13.436 77 0.174 Multiple R = 0.90305 R^2 = 0.81550 Adjusted R^2 = 0.79108 Standard Error = 0.19093 PARAMETERS Beta SE StandB t-test P-value _________________________________________________________________ b[ 0,]= 4.2059 0.7425 0.0000 5.664 0.0000 b[ 1,]= 0.2522 0.0354 0.5934 7.122 0.0000 b[ 2,]= 0.3448 0.0635 0.3966 5.426 0.0000 b[ 3,]= 0.0674 0.0533 0.1364 1.263 0.2108 b[ 4,]= -0.4411 0.2480 -0.1476 -1.778 0.0798 b[ 5,]= -0.3981 0.0543 -0.8488 -7.330 0.0000 b[ 6,]= -0.2038 0.0828 -0.2178 -2.461 0.0164 b[ 7,]= 0.0642 0.0810 0.0676 0.793 0.4303 b[ 8,]= 0.3910 0.1175 0.2078 3.328 0.0014 b[ 9,]= 0.1611 0.1177 0.1747 1.368 0.1757