Political support for reforms of the pension system - two - PowerPoint PPT Presentation

Information Priming Design Conclusions Political support for reforms of the pension system - two experiments Ana Fontoura Gouveia Banco de Portugal and NOVA SBE SPARC seminar - ICS Lisboa - 12 July 2018 Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Motivation Old-age dependency ratio in the EU: 29 in 2015, 50 in 2050 Even worse for PT: 31 in 2015, 65 in 2050 Public opinion awareness: 61% of the Europeans not confident about the future of their pension (Eurobarometer71) Portuguese are: among the least confident in the EU... ... and the less keen to reform less aware of the demographic challenges Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Research question 1 Research question 1 What role for information on support for reforms? Self-interest Boeri et al (2002), Heinemann et al (2009), Chong et al (2001) Ideology Blinder and Krueger (2004), Lynch and Myrskyl (2009) Information Boeri et al (2002), Walstad et al (2002), Boeri and Tabellini (2012) Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Randomized control trial Trial 1 - Portuguese population over 18 years old 428 replies (June/July 2013) Average response time: 25/30 minutes Treatment: Informative text upfront randomly assigned to half of the sample Preferences Support CGA reform Knowledge (6/15) Behavior Support SS reform Demographics Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Impact of treatment on information Table: Correct responses, questions for which answer provided in text - by treatment intensity (average) Not treated Least treated Somewhat treated Most treated N 229 58 100 41 Info text 2,72 2,83 3,03* 4,12** Info PAYG 0,91 0,84 0,82** 1** Info ratio 0,17 0,28* 0,32** 0,54** Info ratio30 0,15 0,28** 0,29** 0,61** Info sust. factor 0,57 0,50 0,52 0,78** Info formula 0,62 0,71 0,71 0,80** Info fund 0,30 0,22 0,37 0,39 No differences for info not in text (Table 2 v. Table 3 in the paper): most treated not ex-ante more informed Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Impact of treatment on support Table: Share of respondents in favor of reforms, by treatment intensity Not treated Least treated Somewhat treated Most treated N 229 58 100 41 Favor at least 0,70 0,69 0,72 0,85** one reform Retirement age 0,33 0,35 0,38 0,49** Pensions active 0,19 0,19 0,25 0,34** Pensions all 0,25 0,31 0,29 0,37* Contributions active 0,34 0,22 0,29 0,29 Contributions retired 0,22 0,29 0,23 0,27 Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Impact of treatment on support - probit Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions What drives effective treatment? Table: Determinants of treatment intensity Dependent variable: 1 most treated; 0 other treated Info out text 0,03 Unemployed -0,07 Union -0,14 Student -0,02 Left 0,01 Married -0,12 * Right 0,01 Male -0,05 Young 35 -0,07 Media -0,08 Old 55 0,40 ** Attention 0,06 High income -0,08 Covered SS 0,01 Low income 0,15 ** Pseudo R-squared 0,11 N 198 Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Main results - information experiment Information problems exist: 35% know the contribution rate; 17% know the young to old ratio More informed (info out text) more willing to reform Treatment increases information (treated not ex-ante more informed) Intensity of treatment varies (+old, +low income, -married) Only the most treated significantly more willing to support reforms Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Research question 2 Research question 2 What role for priming? Knowledge stored in long-term memory activated by exposure to a stimulus (Hopkins, 2011) Consistency Political opinions v. Economic opinions (Wilcox and Wlezien, 1993; Palmer and Duch, 2001) Media covergae Impact on elections (Althaus and Young, 2006) Self-interest Prompting self-interest induces self-interested responses (Chong et al., 2001) Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Randomized control trial Trial 2 - University students (UNL, UL, UTL) 522 replies (May 2013) Average response time: 20/25 minutes Treatment: Randomized blocks [Preferences & Information] and [Support for reforms] Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Main results - priming experiment (overview) In relation to voters on the center of the political spectrum: Those on the left consistently less keen to support reforms, irrespective of priming; Those on the right more supportive only after priming. . (Hopkins, 2011 finds behavior replicated in actual voting situations) Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Additional results: partial privatization (overview) In both surveys: willingness to partially privatize the system, potentially at a cost (based on Schokkaert et al., 2000). Reduction in contributions by half, paid directly in the salary. Upon retirement, the pension would be lower, as if the worker had worked for 50% to 30% of her initial salary. Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Additional results: partial privatization (overview) 20% to 25% of repondents do not know whether they favor such reform (5% to 10% for more standard reforms) Support cut in half when less favourable terms apply 25% to 30% of those in favor, refuse if investment fund 40% to 50% of those against, accept if investment fund Thus, design and information crucial Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments

Information Priming Design Conclusions Conclusions More informed voters more willing to reform pension system How to increase information levels? Public policy More research needed on determinants of treatment effectiveness Priming effects relevant - role of public debate Design matters (even non-costly details matter: Saez, 2009) Ana Fontoura Gouveia Political support for reforms of the pension system - two experiments