AR(1) Coefficient - What does it mean?

[redquadrant]

I am trying to find an equation to estimate GDP as a function of c, t and ar(1).

I entered the following parameters in my estimation:
log(gdp) c t ar(1)

When I calculate log(GDP) at t=2 by hand, I get 11.5566 while Eviews gets 6.6316.

I think I may be interpreting the estimation output incorrectly.

I get the following estimation output:

Variable Coefficient Std. Error t-Statistic Prob.

C 11.57080 17.42185 0.664154 0.5122
T -0.008461 0.209855 -0.040318 0.9681
AR(1) 0.971538 0.056595 17.16640 0.0000

I believe it means this:
log(GDP) = 11.571 - 0.008t + 0.972[u(t – 1)]

However, when I try to calculate it using Excel, I get different results than Eviews.

For example, for t=2, I have the data:
log(GDP) = 11.571 - 0.008t + ut
log(GDP) = 11.571 - 0.008(2) + 0.972(0.02284)
I get the ut – 1 value from Eviews' own calculation of residual at t=1 from their residuals table
log(GDP) = 11.5566

However, Eviews says the fitted value is 6.6316.

What am I doing wrong?
I suspect I'm interpreting the coefficients incorrectly, but I'm not sure. :eviews6: