Public Choice 78: 205-218, 1994. © 1994 Kluwer Academic Publishers. Printed in the Netherlands. Equilibrium in fiscal choices: Evidence from a budget game* PIERVINCENZO BONDONIO CARLA MARCHESE Department of Economics, University of Turin, via S. Ottavio 20, 1-10124 Turin Accepted 22 april 1992 Abstract. This is an applied study about the stability of collectivedecision-making in fiscal matters and the features of individual preferenceswhichare sufficientto reach a socialordering. It is based on data about citizens' preferences collectedthrough a budget game played by a sample of voters in Turin, a large Italian city. By simulating a series of indMdual choices, a social ordering of the municipal budget items is arrived at. The rote playedby restricted (i.e., single-peaked, single-caved etc.) preferences to avoid cyclesin choice simulations is then assessed. 1. Introduction Building on the pathbreaking contributions of Black (1948) and Arrow (1963), investigation of the voting paradox and the stability of collective decision- making continues in the current literature: a good overview is provided by Sen (1970). Within this stream of thought, this paper focuses on the features of individu- al preferences that suffice to achieve a social ordering (Sen and Pattanaik, 1969). The paper makes use of data resulting from a survey of fiscal preferences made on a sample of voters in a large Italian city, Turin, to test whether the paradox of voting may actually occur. It checks also whether restrictions upon individual preferences, sufficient to achieve a social ordering, are present or not. To this end, the most significant of these restrictions are reviewed: single- peakedness, separability, single-cavedness, limited agreement and ext~cemal restriction (See Kramer, 1973; and Sen, 1986). 1 Collective choices are made in democracy through some sort of voting process, usually majority voting. This, however, does not guarantee that all the Arrow conditions can be met simultaneously (see Sen, 1986: 1077-1084). Literature has tried to find an exit-path from Arrow's dilemma - which can be described summarily as implying "either irrationality of collective choices of dictatorship" - working along many streams. Two of these research streams seem to be more interesting than others. One contrasts the "ideal