Hi,

First of all, thanks to everybody, I really found very useful your suggestions.

Regarding the suggestion by Pat, I agree it suits better to use the 
word hypothesis rather than principle. Though there is one point I 
would like to make. As I understand it (and it is also in the book of 
Hywell Jones, I think), the accelerator hypothesis says that the 
change in investment is a function (in lack of another word) of the 
EXPECTED change of output, because it is the best predictor of 
expected profitability (I add the latter). The fact that it has 
usually been modeled as a function of the past change in output, is 
because it was the easiest way to treat the formation of expectations.

Thanks again to all.

Pablo Bortz