If the relationship between ηi and Zit is not specified and fixed effects are treated as parameters to be estimated, the model suffers from an incidental parameter problem. The incidental parameter problem arises, because the number of fixed effects parameters, ηi, is increasing with the number of individuals (cross-sectional units). Hence, the estimation ofηi is inconsistent, especially if time series observations are very small and the number of cross-sectional units is large. The inconsistency is not limited to the individual effects, but extends to all coefficients.