The aim of this paper is to estimate in a multivariate context the factors associated with
well-being and ill-being without making the assumptions that they are opposite ends of
the same continuum, and that the factors uniformly affect both well-being and ill-being.
Using the first five waves of panel data from the Household, Income and Labour
Dynamics in Australia (HILDA) survey, we jointly model positive and negative wellbeing
in a two-equation dynamic panel data model. We found that while past ill-being
had a significant effect on current well-being there was no support for a reverse
relationship (i.e. lagged effect of well-being on current ill-being). In addition, we also
found support for asymmetry in how certain factors affect well-being and ill-being. The
implication of the findings in this paper for the happiness literature is that for future
empirical work, it would perhaps more prudent to begin with the notion that well-being
and ill-being are distinct dimensions, that the unobservables that affect well-being and
ill-being are correlated, and to specify econometric models that allow for these concepts
to be reflected.