1: \begin{abstract}

2: \noindent

3:    This paper studies linear panel regression models in which the

4:    unobserved error term is an unknown smooth function of  two-way unobserved fixed effects. 

5:    In standard additive or interactive fixed effect models the individual specific and 

6:    time specific effects are assumed to enter with a known functional form (additive

7:    or multiplicative), while we allow for this functional form to be more general and unknown.

8:    We discuss two different estimation approaches that allow consistent estimation of 

9:    the regression parameters in this setting as the number of individuals and the number of time periods

10:    grow to infinity.

11:     The first approach uses the interactive fixed effect estimator in \cite{Bai2009}, which is still applicable here,

12:    as long as the number of factors in the estimation grows asymptotically. The second 

13:    approach first discretizes the two-way unobserved heterogeneity 

14:    (similar to what \citealt{bonhomme2017discretizing} are doing for one-way heterogeneity)

15:   and then estimates a simple linear fixed effect model with additive two-way grouped fixed effects. For both estimation methods we obtain asymptotic convergence results, perform Monte Carlo simulations,

16:   and employ the estimators in an empirical application to UK house price data.

17:    

18: \end{abstract}

19: