Pass the Bucks: Investment Incentives as Political Credit-Claiming Devices Evidence from a Survey Experiment Nathan M Jensen Edmund Malesky Mariana Medina Ugur Ozdemir

Public Policy Puzzle Tax incentives have little impact on investment decisions Tax incentives have little impact on investment decisions (Wells et al 2001; Bobonis and Shatz 2007) Incentives are too large relative to benefits Incentives are too large relative to benefits (Morris and Pirnia 1999; Blomstrom and Kokko 2003; Buettner and Ruf 2007) 17 projects in the U.S. since 2002 over $100 million each 17 projects in the U.S. since 2002 over $100 million each VW in Tennessee: $450 million ($225,000 per job) VW in Tennessee: $450 million ($225,000 per job) AMD and IBM in New York ($1.1 billion and $0.5 billion) AMD and IBM in New York ($1.1 billion and $0.5 billion) Most countries offer tax incentives Most countries offer tax incentives Ford in Romania: $193 million ($83,000 per job) Ford in Romania: $193 million ($83,000 per job)

Motivation: Globalization and Credit Claiming Economic performance affects perceptions of government and voting Economic performance affects perceptions of government and voting Voters blame politicians economic impact of international markets, natural disasters and shark attacks (Hellwig et al 2008) Voters blame politicians economic impact of international markets, natural disasters and shark attacks (Hellwig et al 2008) “Globalization” as a justification or mechanism for preferred policies “Globalization” as a justification or mechanism for preferred policies Increases rejection costs (Vreeland 2003) Increases rejection costs (Vreeland 2003) Inefficient redistribution (Acemoglu and Robinson 2001) Inefficient redistribution (Acemoglu and Robinson 2001) Fiscal Illusion/Obfuscation (Wagner 1976; Kono 2006) Fiscal Illusion/Obfuscation (Wagner 1976; Kono 2006)

Theory Model: Pandering Model: Pandering High vs low quality politicians, voters, and imperfect information High vs low quality politicians, voters, and imperfect information Implication: Politicians can offer tax incentives, even when they are the incorrect policy Implication: Politicians can offer tax incentives, even when they are the incorrect policy Extensions Extensions More direct focus on voters, specifically partisan and independent voters. More direct focus on voters, specifically partisan and independent voters.

Incentives, Economic Performance, and Voter’s Decisions

Three Hypotheses H1: Voters are more likely to reelect governors who attract investment to their state. H2: Voters are more likely to reward governors who attract investment and offer fiscal incentives. H3: Ideologically independent voters are more likely than partisan voters to reward the use of incentive policies.

Research Design Simple survey experiment in the U.S. Simple survey experiment in the U.S. Treatments on investment and incentives Treatments on investment and incentives Affect of treatment on voting for the governor Affect of treatment on voting for the governor Data: 2009 Cooperative Congressional Election Survey Data: 2009 Cooperative Congressional Election Survey Representative U.S. sample (1,974) Representative U.S. sample (1,974)

Experimental Question Your state competed with a number of other states over a new manufacturing plant that will create 1,000 jobs. With the support of the governor, your state offered a tax incentive (break/reduction) package that was [greater/equal or less] than that of the other states. If your state [receives/does not receive] this investment, how would this effect your evaluation of your governor’s performance in office? 1) I would be much more likely to vote for the governor in the next election. 2) I would be slightly more likely to vote for the governor in the next election. 3) My vote choice would not be altered. 4) I would be slightly less likely to vote for the governor in the next election. 4) I would be slightly less likely to vote for the governor in the next election. 5) I would be much less likely to vote for the governor in the next election.

Graph of Interaction

Results from Ordered Probit Model Attracting Investment affects approval: Among the general population, offering an incentive offers an additional 1.4% votes in favor of the governor. Attracting Investment affects approval: Among the general population, offering an incentive offers an additional 1.4% votes in favor of the governor. Partisan voters are less susceptible to changing their voting intentions for the governor. Partisan voters are less susceptible to changing their voting intentions for the governor. The vote bonus for independents from governors winning projects with the use of incentives is 5.6%. The vote bonus for independents from governors winning projects with the use of incentives is 5.6%. The vote bonus for incentives is higher for losing governors (about 3.5% from all respondents, 11.2% from independents). The vote bonus for incentives is higher for losing governors (about 3.5% from all respondents, 11.2% from independents). Dominant strategy is to offer an incentive. Dominant strategy is to offer an incentive.

Additional Tests Survey Questions Survey Questions Knowledge of incentives Knowledge of incentives Experimental Questions Experimental Questions Attribution of economic performance Attribution of economic performance [recovery, decline] Perceptions of investment by origin Perceptions of investment by origin [foreign, Chinese, Japanese]

Conclusions and Future Research Conclusion Conclusion Offering tax incentives, while economically inefficient, is a dominant strategy. Offering tax incentives, while economically inefficient, is a dominant strategy. Mediated by Approval: Popular governors have less need to gamble for resurrection by pursuing economically damaging but politically advantageous policies. Mediated by Approval: Popular governors have less need to gamble for resurrection by pursuing economically damaging but politically advantageous policies. Future Research Future Research Theory: Focus on individual behavior Theory: Focus on individual behavior Empirics: Empirics: 2010 CCES 2010 CCES Testing outside of the United States Testing outside of the United States