The first and second moments of the monopulse azimuth estimates are used to characterize the estimator performance in a background of mainbeam or sidelobe multipath or ATCRBS interference. An exact expression for the bias error is obtained that not only accounts for the bias due to targets separated in azimuth but also demonstrates the effect of fading or signal cancellation. For a target signal at a 20 dB signal-to-noise ratio, it is shown that the fading effect produces negligible bias unless the signal-to-interference ratio is between ±2.5 dB. A first order expression for the variance is obtained that demonstrates that an increase in the random error will occur due to the target separation in azimuth and due to signal fades. The exact nature of the bias and variance depend strongly on the relative phase between the target and interference signals with the poorest performance being obtained at the in- and out-of-phase conditions. It is knowledge of this behavior that is essential in evaluating the idea of azimuth estimation data editing.