If one is to use Transparency International rating of a country, for example, for a number of years (let say 5 years is 136, 132, 135 128 130), can these number be used directly and interpreted in terms of higher value for low transparency of that country (since higher number indicate lower rating in Transparency International Rating) or it has to be transformed. If it is to be transformed, into what level(s) will it be transformed.
there are two reasons for log transformation of the variable one is theoretical reason and another one is statistical reason.
if the variable really has compounding effect, you should make the log transformation. it is the theoretical reason for the transformation.
statisticla reason is if the variable does not satisfy required satistical properties(non-stationarity,asymmetric distribution etc.,) you should do the log transformation to get more consistence results .
Before going into the depth of the problems of econometric techniques, three suggestions on the data:
1. Try to use the raw "Corruption Perceptions Index", instead of the ranking. You will get something like 30, 29, 31, 30, 31. In this series, higher values (e.g. 31) mean more transparency and lower values (e.g. 29) mean more corruption.
2. Try to introduce a time lag. It depends on the type of dependent variables you wish to test. If it is related to investment, it is advisable to introduce at least two years of lag.
3. Be careful with the interpretation of the CPI, and for at least two reasons. One is that it reflects perceptions, which are a bit tricky, as compared with hard data. The second is that it is a compound index, as referred to by Dr Potharla. He is right that something has to be done about it. In any case, you have to be more moderate with the interpretation of the effects of CPI than with the interpretation of the effects of hard number variables.
In any case, it is essential to be confident about the good quality of variables if you want meaningful results.
All the best for the exercise.
Thank you Kalman. Your explanation is quite clear to me and the caution in interpretation is noted. However, my country' s CPI from Transparency international for a number of years now is as high as the figure I provided in my question.
Regards.
Dear Friend
I use the CPI to calculate the impact on human development in Argentina, and I could use it as a correlation of the development variables, or as an independent varaible against the dependent variable (a development example, poverty)
Thus showing correlation of corruption and development variables, correlation and intensity, and see the impact of corruption on these variables, greetings
It's in my book "Economia de la Corrupcion" at Amazon