Identity equation with "dependent" variable on right side

[leah]

Is there a way in eviews to solve for an unknown variable in an identity when the unknown variable is impossible to isolate on one side of the equation? This is the equation if it helps:

beta, gamma, phi, and tau are all known series. alpha is unknown -

gamma = ((1-exp(-phi-alpha))*alpha*(gamma(-1)+tau(-1)))/(beta+alpha) + (exp(-phi-alpha))*gamma

[EViews Gareth]

Not really. You could use the optimize function to get darn close.

[leah]

Okay, sortof fumbling around in the dark. Is this on the right track? When I run it I get an error saying that the "objective function evaluates to NA for one or more observations".

wfcreate a 2000 2013
fetch(d=na) gdp gdph jfnsco jf

subroutine f(series alpha, series phi, series gamma, series tau, series beta)
gamma = ((1-exp(-phi-alpha))*alpha*(gamma(-1)+tau(-1)))/(beta+alpha)+(exp(-phi-alpha))*gamma
endsub

series out
optimize f(out, gdp, gdph, jfnsco, jf)

[EViews Gareth]

You'll have to set starting values for alpha.