Dropping a cross section at a time

[sazad]

Hi there,

I am trying to check for random cross section effects by dropping a cross section at a time as done in Engle and Rangel (2008): Table 9.

Do I use the proc mixed to make the period fixed and country random or just use the proc panel with ranone? Also, I don't know how Engle and Rangel (2008) found autocorrelation coefficient in Table 9 for panel specification.

Could anyone help, please?

proc mixed data = mydata;

class country year;

model my model year/ solution;

random country;

run;

or

proc panel data = mydata;

id country year;

model my model/HAC ranone;

run;

Thanks,

Sohel

Reference:

Engle, R.F. and Rangel, J.G., 2008. The spline-GARCH model for low-frequency volatility and its global macroeconomic causes. The Review of Financial Studies, 21(3), pp.1187-1222.

[mkeintz]

I'm not familiar with proc panel or proc mixed.  BUT ...

... I don't see why YEAR is a class variable, which makes it no different than, say color.  Shouldn't you allow year to be a continuous variable?  Which I believe is the only way to get autocorrelation (assuming you - and Engle and Rangel - mean serial autocorrelation).