Hi all,
My paper is looking at the most significant determinants of u.k. economic growth. I will look at macroeconomic , technological, human capital, socio-geographical and governance variables against the dependent variable of GDP growth. Anyone who has done research into growth-theories could suggest suitable statistical models which can be used on stata? I am looking at Baynesian Model Averaging currently but not sure that is the best model out there
I have been working in a somewhat related issue (efficiency of railways)
I have no simple answer given the fact that your reasearch seems to be starting.
Yet, I would have
1) be carreful of multicolinearity (if two variables are correlated the best you can do is to introduce a synthétic vairiable
2) you could test some various time lags (as in the Box-Jenkins methods)
3) a constant growth rate leads to a linear tendancy when you transform the data with log. (but in this case, I don't know exactly what is the hypothesis on error terms.
4) comparison of data from entities belonging to different currency area are difficult (especially nif there is no market to provide a value for the currencies (e;g; Chinese RMB versus dollars).
5) Even if this is more difficult, you can use panel models, as discribed in the textbook of Green and in most of the availaible software documentation, this could enable you to measure the fixed effects.
6) last but not least : in this case you have to choose the hypothesis you want to test, before the empirical work.
Good luck, hoping this can help.
Hi Adam, I can advise you to work on causality and cointegration. I tink when you work with time series, many models like arima, sarim, arma,etc used their own past value in prediction. When you decide to make a model, be carefull on stationarity of your data. When data are not stationary, you get spurious regression and coefficient have bias. Be carefull about the test to use. they ay many tests to test unit roots but all do not have same result. Check also structural break, because in case of structural breaks, dickey Fuller test are waek and fail to refuse the null hypothesis of unit root.
Good luck.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling