Kam Kup <[email protected]>:
Is it censored at 0 and 100 (i.e. a negative "true" value shows up in
the data as zero)? Or constrained to lie in that interval? If the
latter, see Papke and Wooldridge (1996) and use a command like -glm-.
I guess you would like to include fixed effects (making some strong
add'l assumptions, then using -xtgee- and tricks described by
Wooldridge at http://www.stata.com/meeting/snasug08/; download
http://www.stata.com/meeting/snasug08/wooldridge.zip for materials)
but you should at minimum cluster by country.
Leslie E. Papke and Jeffrey M. Wooldridge. 1996. "Econometric Methods
for Fractional Response Variables with an Application to 401(k) Plan
Participation Rates." Journal of Applied Econometrics, 11(6): 619-632.
[see also http://www.nber.org/papers/t0147.pdf]
On Mon, Sep 22, 2008 at 1:36 PM, Kam Kup <[email protected]> wrote:
> I'd like to sovereign credit spreads (a number between 0 and 100) over three years for about 20 countries. Some of the explanatory variables are updated daily (such as the market spread), others are updated monthly (such as the country's overall debt level), and some are unchanging (a dummy for membership in OPEC) and a country-specific dummy
>
> I have 3 questions:
>
> 1. Should I model each day's credit spread, using the (at the time) latest released debt levels as a dummy variable?
>
> 2. Which command should I use in Stata to estimate this? Would -xtreg, fe- work, even though the dependent variable is censored below (at 0) and above (at 100)?
>
> 3. For anyone who has worked with this kind of data, should I make any adjustments (e.g. for heteroskedasticity, autocorrelation, etc)?
>
> Thank you very much,
> Kam
*
* 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/
