Fifty meter-class external occulters have been proposed to detect earth-like planets. The THEIA concept1, a forty-meter diameter occulter with twenty ten-meter petals has the necessary nominal performance to achieve this goal. This paper examines whether this design is robust against expected manufacturing and deployment errors. The development of a numerical algorithm that represents the mask defects as a collection of rectangular apertures mitigates the problems associated with modeling diffraction phenomena produced by an occulter with characteristic physical dimensions that span five orders of magnitude. The field from each of these rectangles, which is proportional to a two-dimensional sinc function at the telescope, is added to the diffracted field from the nominal occulter. Results for a set of representative defects are presented. A single-petal, single-defect error budget, based on a minimum contrast of 10-12 at 75 or 118 milli-arcseconds from the host star from 0.3 μ to 0.9 μ, is quoted. A Monte Carlo-type simulation that predicts the performance of the occulter in the presence of random combinations of all of the error demonstrates that the system contrast can maintained to better than 10-11 from 0.3 μ to 0.9 μ if the values in the error budget can be achieved.