One can, of course, not find out (decompose) different effects one a variable (oil price) on the basis of this variable alone. The program needs information on the influencing variables, too. And here you have the problem, that statistics show only show the effected transactions and it is difficult to differentiate between demand and supply. In principle, this is not only true for the oil transactions, but also for aggregate demand. But here you can assume, that this is mainly demand-driven and you can take real GDP as demand variable. The question is, which countries' GDP to take.

But the main problem is, that oil markets are far from normal, and therefore one cannot expect the supply/demand-price effects, economics assumes for "normal" (perfect?) markets.

It needs a lot of effert to develop a stucture of the study, before you can go into statistical work.

Ok, thank you very much. Your answers always help me a lot. I am really interested in those models but i obviously need to take more care about the requirements as young researcher. Anton Rainer

In R, you can create a Cholesky decomposition of a matrix using the "chol " function from the base package. Here is an example of how you could create a Cholesky decomposition of a matrix representing oil price shocks:

library(base) if it is installed otherwise you should install it

# Define the matrix representing oil price shocks

oil_shocks

I do not understand, why one should use a Cholesky decomposition in this case. This decomposiion is (in my understanding) used for solving linear equation systems. This means, you have to specify this equation (system) before. Obviously, the equation for which the form and paramters to find-out is: oil price=f(oil supply,aggregate demand, specific deman, other factors). Under the assumption, that the relations are linear(in some transformations), one can apply the usual estimation methods, if one has suffient data. One need not care how the method solves the system, e.g how it calculates the parameters.

The problem I wanted to show in my first answer that it is not easy to fifferentiate between supply and demand. E.g. if you have the GDP for a certain period, is it supply or demand? What is a shock? I know this term is frequently used by New Classical and New Keynesian economists, but the meaning is unclear.

Thank you all for your support. I really appreciate it and it makes life much easier.