The econometric modelling of social preferences (Q2443951)

The `modelling of social preferences' noted in the heading is investigated on the base of normative modelling of individuals' behavior with respect to fairness. The authors follow \textit{A. W. Cappelen} et al. [``The pluralism of fairness ideals: an experimental approach'', Am. Econ. Rev. 97, No. 3, 818--827 (2007; \url{doi:10.1257/aer.97.3.818})] and use their model of income allocation between two players. The model is a game of two phases. ``In the first phase, subject \(i \in \left\{ {1,\;2} \right\}\) decides how much of her initial endowment will be `invested' \(\left( {q_{i}} \right)\).'' The pair of investments \(\left( {q_{1} ,q_{2}} \right)\) defines a total income \(X({a,q}) = a_{1} q_{1} + a_{2} q_{2}\), where the coefficients \(\left({a_{1} ,a_{2}} \right)\) are `rates of return'. In the second phase each player decides how much of the total income she would like to allocate to herself \(\left( {y} \right)\), correspondingly, \(\left( {X - y} \right)\) to her opponent. ``It is then randomly determined which of the two proposed allocations is implemented.'' Each player maximizes her utility \[ V\left({y;a,q} \right) = \gamma y - \beta \frac{{\left[ {y - m({a,q})} \right]^{2}}}{{2X({a,q})}}, \] \noindent where \(m\left( {a,q} \right)\) formulates the `fairness ideal'. In the paper under review it is assumed that there are three types of ideals \(m_{k}({a,q})\), \(k = 1,\;2,\;3,\) in mankind: `Strict Egalitarian' (SE), `Liberal Egalitarian' (LE), and `Libertarian' (L). Three approaches to econometric modelling are used: the Random Utility model, the Random Behavior model, and the Random Preference model. These models are estimated in various ways (w.r.t. SE, LE, E) to an experiment on fairness. ``Conclusions concerning fairness-related behaviour depend crucially on the choice of econometric model.''