Econometrics: why co-integration is needed - page 28
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
...it all makes sense, but next I still don't understand how to get the coefficient. For example:
Thus, we arrive at the following error correction model (ECM model):
dY1 = -a1*S + lagged(dY1, dY2)
dY2 = -a2*S + lagged(dY1, dY2)
Where the two variables a1 and a2 come from I still don't understand. And whatlagged(dY1, dY2) means too. ))
For two time series x(t), y(t) the vector of cointegration is very simple. You estimate a pairwise linear regression y(t)=a+b*x(t)+N(0, sigma) using any method. Then the cointegration vector is [-b; 1], its scalar multiplication by the vector [x(t), y(t)] gives a stationary process of the form N(a, sigma), in your ECM model it is denoted as S.
There are two solutions - to use orthogonal OLS or to choose the regression with larger residuals variance (for trading purposes as it is easier to trade such a process).
ECM-model is another way of writing the cointegration relation. Coefficients a1, a2 reflect influence of a deviation S from its average on further increments of the processes x(t), y(t). I won't describe it all, because it is too long. It seems to me that Eliseeva's econometrics textbook contains a detailed explanation.
There are no traders in this thread, only theoretical mathematicians engaged in blatant k-eye))
Please stop flooding, since the level of intelligence does not allow you to make money with the help of the methods discussed. And you're wrong about the traders.
Thank you. But I would like to do cointegration tests in MT5. I don't need anything else yet. I don't have any special hopes. It's just a research tool for me.
Your point of view is widespread on this (and not only on this) forum. Voluntarily or involuntarily people draw an analogy between TA and econometrics: in TA you can take several indicators and build a TS. The analogy in econometrics: take several methods and build a TS. Unfortunately, econometrics is a set of a large number of interrelated tools + understanding of applicability of these tools + experience in applying these tools.
Using your example. If you calculate the cointegration vector, as you have been prompted to do, it is obligatory to answer the question: will the residual from the regression be stationary? Besides, it is important to look not only at the value of the coefficients in the vector, but also to look at the information that will accompany the applied ISC.......
The algorithm you have proposed is the tip of the iceberg and it is important to have a complete set of tools, ready-made, to apply them as needed. In this sense, Matlab is not very suitable, as it is not a statistics package, and one has to understand very well the meaning of the results at each step and decide if further steps are necessary.
...
Thank you. I'll see what I can figure out.
faa1947:
...That's why I wrote "...bye I don't need anything else". But I will of course expand my toolkit as needed. Thank you.