This paper makes three principal contributions. First, we propose a new estimator for the unobservable variable models with endogenous causes. We show that under factor analysis type of assumptions, Robinson and Ferrara (1977)'s procedure is not fully efficient, and a more efficient procedure can be obtained by merging Robinson and Ferrara' procedure with that in Joreskog and Goldberger (1975) procedure. The asymptotic properties of the new estimator are compared with these two estimators and results indicate that, with our mixture approach, significant efficiency gains can be achieved. Second, we model the Quality of life (QOL) as an unobservable or latent link variable between observable causes and observable effects, which mitigates problem of bias, inconsistency, and arbitrary weightings of explanatory factors. Third, we estimate and compare QOL indices for 43 countries for the year 1990s, noting differences between countries and over time.