Victor Champonnois*, João V. Ferreira**

phd seminar

Victor Champonnois*, João V. Ferreira**

The influence of institutions and tax systems on protest rates in stated preferences surveys*
Choice with Time**
Co-écrit avec
Nicolas Gravel**

VC Salle A

Centre de la Vieille-Charité - Salle A

Centre de la Vieille Charité
2 rue de la Charité
13002 Marseille

Mardi 10 janvier 2017| 12:30 - 14:00

Edward Levavasseur : edward.levavasseur[at]
Lara Vivian : lara.vivian[at]


*Some respondents to valuation surveys refuse to state their true preferences because they protest against some characteristics of the survey. Among others, the studies' payment vehicle and the country where the survey was conducted are two significant determinants for this behavior. While differences in protest rates across payment vehicles are well-studied, almost no study has tried to explain differences across countries. I explore an hypothesis by which trust in the institutions and tax pressure are determinants of the protest rate for studies that use a tax payment vehicle. For instance, respondents should be less likely to accept a tax payment vehicle if there is a high level of government corruption, or if they feel too much tax pressure.  I exploit country and time variation in meta-data on stated preferences studies merged with various source of country panel data. I find that for countries with a low level of trust in the institution, an increase in this level leads to a decrease of the probability to protest against a tax vehicle, but this effect becomes positive beyond a certain threshold. The opposite relation is observed for tax pressure.

**We propose a framework for the analysis of choice behavior when the later explicitly depends upon time. We relate this framework to the traditional timeless choice-theoretic models. We illustrate the usefulness of our framework by proposing three possible models of choice behavior in such a setting: (i) changing tastes, (ii) preference formation by trial-and-error and (iii) choices with endogenous status-quo bias. We provide a full characterization of each of these three choice models by means of revealed preference-like axioms.