Allin, 

 

thanks a lot, that was the issue. When I thus try to constrain the calculation not to go in the direction where negative lambdas are created by 

 

<hansl>

function scalar getNegativesCount(series * xSeries)

    scalar xRetVal = 0

    

    loop xIdx = 1..$nobs

        if missing( xSeries[xIdx] ) = 0

            if xSeries[xIdx] < 0.0

                xRetVal = xRetVal + 1

            endif

        endif

    endloop

 

    return xRetVal

end function

...
# ESTIMATION
mle ll = xCheck == 0 ? -ln(lambda) - sqrtPark/lambda  : NA
         series lambda = mean(sqrtPark)
         series lambda = c_ + rng * sqrtPark(-1) + rng2 * sqrtPark(-2) + err_ * lambda(-1) + err_2 * lambda(-2)
         scalar xCheck = getNegativesCount(& lambda) 
         params c_ rng rng2 err_ err_2
end mle --robust
#ESTIMATION
</hansl>
I get a "Scalar loglikelihood: can't do QML". So, how do I tell the optimizer that in case it runs to negative lambda, it should "go away" from that parameter combination? I also tried to replace the 
<hansl>
...
mle ll = xCheck == 0 ? -ln(lambda) - sqrtPark/lambda  : NA
...
</hansl>
with 
<hansl>
...
mle ll = xCheck == 0 ? -ln(lambda) - sqrtPark/lambda  : -100000
...
</hansl>
but it produced the same error.
Many thanks, 
Daniel