Theoretical Foundations of the Latent Regression Approach:

The latent regression described above may be  interpreted in  terms of utilities. Let  U' be  the individual's  utility  if  she decides  to  work/seek  work and U⁰ be  the individual's utility if she decides to abstain from working. We may formulate

Here  x  represents a  common  set of factors,  such  as age,  marital  status, gender, education, and work history, β₀ and β₁ are vectors of unknown parameters and ε₀ and ε₁ are unobservable, alternative specific ,  taste components. The  individual will  decide  to  worklnot work based on  calculations  of marginal benefit-marginal cost of labor force participation, which  in  turn  are based  on  the utilities derived  from the two alternatives. We observe the choice made but not the unobservable utilities. The observed indicator Y = 1 whenever U¹ ≥ U⁰ and  Y = 0 whenever U¹ - U⁰ ≤ 0. For the i^th individual in our sample,