BETTINA DALESSANDRO
Sdertrns hgskola
Advanced Econometrics
Introduction
In this paper it will be analyzed the annual data for the United States
of America on unemployment and inflation, by Ramanathan, which data set
in available on data examples on the Gretl software. The source is 1996
Economic Report of the President. The analysis will use the following
variables year: 1959-95, unemp: civilian unemployment rate, cpi:
consumer price index, infl: percent change in CPI (inflation rate), wggr:
percent change in average weekly earning, in current dollars.
Regarding the nature of the data, the type used will be Time Series. It
is a set of observations that a variable takes at different times, such as
daily (e.g. stock prices, weather reports), weekly (e.g. money supply),
monthly (e.g. the unemployment rate, the consumer price index CPI.
Succesive observations may be correlated, leading to problems for the
regression, one of them being autocorrelation, which shall be analyzed and
discussed later in this paper. To estimate the linear regression model, the
Ordinary Least Squares shall be used.
Multicollinearity:
or another common
endogeneity case is when the regressor does not only explain the
dependent variable but is also explained by the dependent variable, this is
called simultaniaty.
The analysis starts by using OLS in two stages, when I will select
l_unemp as to be tested to see if it is endogenous. The result tell us that
although not very significant, there is endogeneity in the model.
Using the Jarque-Bera test for normality of residuals we get that for
the test of null hypothesis of normal distribution:
const
alpha(0)
alpha(1)
beta(1)
0,0391259
0,346292
0,318482
1,434955
1,4611
1,8549
0,2531
0,14400
0,06361
0,80016
***
*
0,440177
Log da verossimilhana
Critrio de Schwarz
20,01082
58,07623
Critrio de Akaike
Critrio Hannan-Quinn
Bibliography
50,02165
52,86127