The Wald statistic remains a perfect approximation to the Likelihood Ratio Test.

The Wald statistic tends to produce coverage errors because it assumes a symmetric quadratic curvature.

The Wald statistic converges to a standard normal distribution regardless of the skewness.

The Wald statistic is always more conservative than the Score test.