Let \( X_1, X_2, X_3 \) be a random sample from a distribution with the probability density function
\[
f(x|\theta) = \begin{cases}
\frac{1}{\theta} e^{-x/\theta}, & \text{if } x > 0, \, \theta > 0 \\
0, & \text{otherwise}
\end{cases}
\]
If \( \hat{X}(X_1, X_2, X_3) \) is an unbiased estimator of \( \theta \), which of the following CANNOT be attained as a value of the variance of \( \hat{X} \) at \( \theta = 1 \)?