Title: The semiparametric Bernstein-Von Mises theorem

Abstract: In a smooth semiparametric estimation problem, the marginal posterior for the parameter of interest is expected to be asymptotically normal and satisfy frequentist criteria of optimality if the model is endowed with a suitable prior. It is shown that under certain straightforward and interpretable conditions, the assertion of Le Cam's acclaimed but strictly parametric Bernstein-Von~Mises theorem (Le Cam, 1953) holds in the semiparametric situation as well. As a consequence, Bayesian point-estimators achieve efficiency, for example in the sense of Hajek's convolution theorem (Hajek, 1970). The model is required to satisfy differentiability and metric entropy conditions, while the nuisance prior may not have pointmasses and must assign non-zero mass to certain Kullback-Leibler neighbourhoods, analogous to (Ghosal, 2000). In addition, the marginal posterior is required to converge at parametric rate, which appears to be the most stringent condition in examples. As such, the results constitute a relatively straightforward and immediate way to assess whether a model-prior pair will achieve Bernstein-Von~Mises optimality. We also formulate an existence theorem for a nuisance prior such that the corresponding posterior displays the desired limiting behaviour. The results are applied to estimation of the linear coefficient in partial linear regression, with a Gaussian prior for the nuisance.