To complete the objective of measuring the elasticity of substitution between capital and labour you will need to complete the following steps (see at the bottom for what you need to do to obtain a distinction on this assignment):
Data Construction:
download from the EU_KLEMLinks to an external site. data (for all countries):
the national accounts (the R file);
the capital accounts (the R file);
the growth accounts (the R file);
sort each dataset according to time, country and sector;
check (for yourself) that this makes sense!
how many sectors do you observe?
how many countries do you observe?
how many periods do you observe? (do you observe the complete history for all countries?)
select the following variables:
out of the national account data: ″COMP″,″EMP″,″EMPE″,″H_EMP″,″H_EMPE″;
out of the capital account data: ″I_GFCF″,″Ipyp_GFCF″,″Iq_GFCF″,″Ip_GFCF″,″K_GFCF″,″Kpyp_GFCF″,″Kq_GFCF″;
what do these variables measure? are they always present for all countries, sectors and time period?
merge the two datasets according to the variables : ″geo_code″ , ″nace_r2_code″, ″year″.
in the merged dataset, select observations:
exclude the diverse european groupings (EU##, EA##);
for each observation country, year, sector, create :
a measure of the rental price of capital as the ratio of K_GFCF/Kq_GFCF;
a measure of the price of labour as the ratio of COMP/EMPE;
why do these measures make sense?
a measure of the ratio of the spending on labour to the spending on capital as COMP/K_GFCF (as well as its log);
a measure of the ratio of the price of labour to the price of capital (as well as its log);
exclude any observation where these new variables are missing (or are non-sensical).
Data Desсrіption:
calculate the mean value of the new variables:
within sector;
within country and sector;
within period and sector;
for each new variable calculate
the difference between the variable observed at a point in time, in a country and a sector from the sum of its mean within that country and that sector, plus its mean within that period and that sector. Add its (grand) mean within that sector to the result. Let us call the final outcome the demeaned version of the variables;
why is this a ″useful″ operation?
produce some separate graphical evidence of the evolution over time of the new variables, their demeaned version, and the variables in first difference.
Model Estimation:
using OLS and for each sector, suggest a model and using this model produce the implied estimates of the elasticity of substitution between capital and labour.
produce a table that summarizes your findings. In each case report
the estimate of the elasticity of substitution;
a measurement of the uncertainty that characterizes your estimate;
a measurement of the fit of the model you have estimated;
the number of observations;
In addition to obtain a Distinction level grade for this assignment (complete these tasks and discuss your findings in your report):
for each new variable calculate, within country and sector,
the 1st difference over time, i.e. Dx(t,c,s) ＝ x(t,c,s) – x(t-1,c,s)
produce some additional graphical evidence of the evolution over time of the new variables in first difference.
based on the data in first differences, using OLS and for each sector, produce an alternative set of estimates of the elasticity of substitution between capital and labour.
using a Instrumental Variables Estimator and for each sector, produce an alternative set of estimates of the elasticity of substitution between capital and labour (using for example past observations to generate IVs for current observations).
produce a table that summarizes your additional findings.