Separation of Powers with Party Polarization. A Lab Experiment.
Elections allow citizens to monitor and select good politicians. Canonical agency theories imply that voters make use of politicians' re-election desire to either improve leader selection or to increase incumbents' discipline. Theories of voting emphasize valence issues (e.g., honesty) and often consider policy proposals offered by different parties Besley 2005; Forteza and Pereyra 2021). Considering both dimensions together implies a trade-off for voters between positional and valence issues. Polarization may tilt this trade-off and posit an important challenge for sustaining accountability on valence issues (Ashworth and
Bueno de Mesquita 2009; Besley 2005; Eggers 2014; Kayser and Wlezien 2011). When politicians and citizens radicalize their policy positions, the stakes of elections increase and therefore electoral advantage undermines electoral accountability because partisan voters trade lower accountability for their desired policies. Under extreme polarization voters may fail to hold politicians accountable for corrupt behavior. We analyze the extent to which the separation of powers mediates the effect of policy polarization on the degree of electoral accountability. Based on a theoretical model built on Forteza and Pereyra (2021) and an original laboratory experiment, we investigate whether (i) political polarization weakens electoral accountability, and (ii) separation of powers strengthens electoral accountability by allowing voters to change the incumbent without changing the positional policies in highly polarized environments.
This pre-registered lab experiment (at the EGAP-OSF) has been implemented together with the Centre for Experimental Social Science (CESS) at Oxford in September 2021. The theoretically motivated design explores the effects of polarization and separation of powers on representatives' decisions about rents extraction, as well as on voters' choices in response to these decisions. We focus on the case of partisan electoral advantage (i.e., noncompetitive elections on the policy dimension), as this is the case where policy polarization is known to be more problematic accountability (Besley 2005; Forteza and Pereyra 2021).
The results reveal several important findings. With respect to incumbent politicians, we first observe a continuum of rent extraction rather than a theoretically expected binary distribution of choices (either zero or full extraction). We hypothesize that this unexpected result responds to the existence of representatives with intermediate ideal levels of rent extraction. When the opportunity cap is larger than the representative's bliss point, incumbents extract their most preferred level of rents, leaving the rest to the voter. Second, polarization increases the level of rent extraction conditional on the opportunities to extract. Third, the introduction of checks and balances insignificantly restores the discipline loss induced by the polarization treatment. In the voters' case, we observe that they positively react to incumbents' discipline. More importantly, polarization strongly increases the reelection rates and this effect is offset by the incorporation of check-and-balances. However, these effects only weakly go through our suggested mechanism. Relatedly, we do not find evidence that separation of powers affects voters' reaction to rents. Overall, our findings suggest that polarization is highly relevant for both accountability and discipline. The channels through which polarization operates are heterogeneous, which brings important implications for future experimental research in the field.
Bueno de Mesquita 2009; Besley 2005; Eggers 2014; Kayser and Wlezien 2011). When politicians and citizens radicalize their policy positions, the stakes of elections increase and therefore electoral advantage undermines electoral accountability because partisan voters trade lower accountability for their desired policies. Under extreme polarization voters may fail to hold politicians accountable for corrupt behavior. We analyze the extent to which the separation of powers mediates the effect of policy polarization on the degree of electoral accountability. Based on a theoretical model built on Forteza and Pereyra (2021) and an original laboratory experiment, we investigate whether (i) political polarization weakens electoral accountability, and (ii) separation of powers strengthens electoral accountability by allowing voters to change the incumbent without changing the positional policies in highly polarized environments.
This pre-registered lab experiment (at the EGAP-OSF) has been implemented together with the Centre for Experimental Social Science (CESS) at Oxford in September 2021. The theoretically motivated design explores the effects of polarization and separation of powers on representatives' decisions about rents extraction, as well as on voters' choices in response to these decisions. We focus on the case of partisan electoral advantage (i.e., noncompetitive elections on the policy dimension), as this is the case where policy polarization is known to be more problematic accountability (Besley 2005; Forteza and Pereyra 2021).
The results reveal several important findings. With respect to incumbent politicians, we first observe a continuum of rent extraction rather than a theoretically expected binary distribution of choices (either zero or full extraction). We hypothesize that this unexpected result responds to the existence of representatives with intermediate ideal levels of rent extraction. When the opportunity cap is larger than the representative's bliss point, incumbents extract their most preferred level of rents, leaving the rest to the voter. Second, polarization increases the level of rent extraction conditional on the opportunities to extract. Third, the introduction of checks and balances insignificantly restores the discipline loss induced by the polarization treatment. In the voters' case, we observe that they positively react to incumbents' discipline. More importantly, polarization strongly increases the reelection rates and this effect is offset by the incorporation of check-and-balances. However, these effects only weakly go through our suggested mechanism. Relatedly, we do not find evidence that separation of powers affects voters' reaction to rents. Overall, our findings suggest that polarization is highly relevant for both accountability and discipline. The channels through which polarization operates are heterogeneous, which brings important implications for future experimental research in the field.