This leads to an important econometric point. Cross-sectional data ignore unobserved hetero-geneity between individuals, such as personality traits and hormonal routines. In general this leads to inconsistent estimates.20 In order to exploit the advantages of panel data, the model can be further generalized. In particular, the error term uit can be decomposed into three elements: individual effects, ηi, time effects, θt, and an idiosyncratic error term, εit, which looks as follows: