Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Panel data estimation - econometric advice needed


From   Rodrigo Refoios Camejo <refoioscamejo@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: Panel data estimation - econometric advice needed
Date   Wed, 8 Jun 2011 12:47:42 +0100

Hi there, I have a few questions which are more econometrics related
rather than STATA-specific and wondered if you guys could help.

I’m looking at modelling pharmaceutical prices. I have 30 years data
on the pricing of 25 molecules, which are divided in 5 therapeutic
classes. So, basically, I was looking to estimate how different
variables (like: #molecules available; molecule age; market share;
whether there is a generic in the subclass/molecule; clinical efficacy
proxies; price of the market leader; etc) affect molecules’ unit price
throughout their life cycle.

Initially, I was thinking of using dynamic models but as I didn’t
manage to get much out of it, so am now cosnidering to go for a
simpler model. Which type of model would you recommend, and how can I
solve the two problems I have:

·        molecules entered (and exited) the market at different times
during the study period, so the panel is unbalanced, but this
unbalance is not random

·        I expect the prices to be non-stationary, although the unit
root test in STATA not allowing for unbalanced panels – should I
deflate them or should I work on first differences?

Also, I would expect that the therapeutic class to work as clusters as
the substitutability between them is not perfect. Would including a
dummy for each subclass be sufficient to control for that, knowing
that the subclass they are part of is correlated with time, e.g. older
molecules tend to belong to a subclass and younger molecules to
another? Or would I need to build an hierarchical model? I guess some
of these issues could be seen as similar to those of modelling pricing
stock shares, i.e. where  molecule = company and therapeutic class =
industry cluster the company belongs to, so would appreciate any
advice of people who had worked on similar models.

Cheers and many thanks in advance!

R

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index