Experiments and the Land
Total Page:16
File Type:pdf, Size:1020Kb
An Experimental Economics Investigation of the Land Value Tax: Efficiency, Acceptability, and Positional Goods Joshua M. Dukea,*, TianHang Gaoa September 18, 2017 aDepartment of Applied Economics and Statistics, University of Delaware, 531 S College Ave, Newark DE 19716 *Corresponding author. E-mail addresses: [email protected], [email protected] Acknowledgements The authors are grateful for funding from the Lincoln Institute of Land Policy and support from its fellows and staff members, including Joan Youngman, Semida Munteanu, and Mark Skidmore. The authors appreciate the helpful feedback from the participants at the 2016 David C. Lincoln Fellowship Symposium. The authors thank Emerson Paradee, Xingguo Wang, and Greg Vitz for assistance with administering the experiments. Abstract This research offers the first economic experiment investigating the land tax, where landowners invest under different property tax regimes. A voting treatment assesses the relative acceptability of land value taxation. Results show a land tax produced greater overall welfare in only 37.5% of the experiment sessions. Systematic over-investment arises from the positional- good characteristic of residential land investment, but this effect vanishes when the positional- good indicator is removed. The experiments show that the participants unexpectedly voted in favor of the land tax, suggesting that the efficiency and acceptability of the land tax may be more complex than in non-behavioral economics modeling. JEL Classification: C91; H21; H31; H71 C91, Design of Experiments--Laboratory, Individual Behavior; H21, Taxation, Subsidies, and Revenue-- Efficiency • Optimal Taxation; H31, Fiscal Policies and Behavior of Economic Agents-- household; H71; State and Local Taxation, Subsidies, and Revenue Keywords: Land Use, Behavioral Economics, Taxation 1. Introduction Economists famously led by Henry George have argued in favor of a pure land value tax (LVT) or its close cousin, a split-rate tax (SRT), in part because they encourage land investment and raise revenue for public goods without distortions. As a fixed resource, land will capitalize locational benefits collectively supplied. Targeting these rents with a land tax will not distort the incentive to invest in parcel improvements, as does the ubiquitous uniform property tax (UPT) that taxes improvements and land equally. Most of the recent work on George focuses on theoretical and simulation models of optimal city size (c.f., Behrens et al. 2015). This paper seeks to complement those studies with the first economic lab experiment on how property taxation affects household investment. The data explore three dimensions of the relative performance of the land tax, including the investment incentives, wealth created, and political acceptability. The experiment also explores a new possibility in land value tax research; land value taxation can exacerbate the positional goods characteristics of housing and produces lower aggregate welfare through overinvestment. The efficiency of land taxation has been assessed using simulated general equilibrium models of open and closed cities and a handful of empirical studies. Some of the results show that LVT generates more intensive capital investments in land (Pollack and Shoup 1977; Plassmann and Tideman 2000; Banzhaf and Lavery 2010). Other land tax benefits, including attenuated tax distortions and increased investment, have also been explored under a system of inaccurate assessments with positive results found for LVT (Chapman, Johnston, and Tyrrell 2009). Several studies examine whether the ostensibly progressive land tax may be regressive through a tax regime change in real-world settings (England and Zhao 2005; Bowman and Bell 2008; Choi and Sjoquist 2015). Prior results show that LVT can provide sufficient revenue 1 (DiMasi 1987) and can increase density in cities (Brueckner and Kim 2003; Banzhaf and Lavery 2010; Choi and Sjoquist 2015). In reaction to this wealth of theory and evidence, Fischel (2015, p. 15) asked: “If economists like (the land tax) so much, why does it seem so rare in practice?” The land tax must be politically unacceptable or practically un-implementable. Research on Pennsylvania’s efforts suggests although SRT led to greater efficiency, few municipalities have adopted the tax and, surprisingly, some LVT adoptees are returning to UPT (Fischel 2015, pp. 16-17; Banzhaf and Lavery 2010). Youngman (2016, p. 18) writes that land taxes have suffered from both “administrative failures” and a “lack of political support.” Bourassa (2009, p. 195-6) identified five objections to the land tax; three reasons involve planners’ professional challenges such as difficulty to set rates, adjusting zoning to prevent too much density, and educating the public about land taxation. Bourassa also identified two economic challenges from land taxation, which will be addressed in this paper: (1) it is a tax on unrealized capital gains, that is, wealth rather than cash flow; and (2) it is a policy change that creates winners and losers, and losers will object. That LVT taxes unrealized capital gains may lead some to object on ethical or economic grounds—especially as it contrasts with most taxes that are levied on transactions. The fact that most of the United States currently raises local revenue with UPT means that a switch to LVT would be a policy change, which unavoidably would create policy losers; Plummer (2009) explored these horizontal and vertical inequities in tax incidence and LVT acceptability with both qualitative arguments and numerical examples. In sum, these studies suggest that acceptability depends on more than arguments about social efficiency or tax progressivity. This paper extends these behavioral arguments about LVT acceptability by introducing another challenge to land taxation: positionality. Housing has long been recognized as a 2 positional good (Veblen 1899; Frank 1985a,b; Hirsch 1977). When a good has positionality, consumer spending will be affected by interpersonal comparisons (Frank 1985b). Recent work by Patacchini and Venanzoni (2014) finds a significant effect from social comparison and a conformity peer effect in shaping the demand for housing quality. One implication of these findings is that one’s housing improvements can cause a negative externality for neighbors, causing neighbors to overinvest to re-attain their prior level of housing utility. Housing utility is derived in the context of one’s neighborhood, and this can lead to an expenditure cascade: “a process whereby increased expenditure by some people leads others just below them on the income scale to spend more as well, in turn leading others just below the second group to spend more, and so on” (Frank, Levine, and Dijk 2014, 57). This paper argues that LVT efficiency and acceptability ought to be complicated by positionality. The equity of LVT depends on (1) whether those similarly situated in society are treated “the same” or “the same enough” to satisfy norms of fairness; and (2) whether those differently situated accept their differential burdens. Tax equity associated with cash flow, or income, is easier for the general public to observe and understand. Land capital inequities introduce a second “wealth” dimension upon which a tax policy is evaluated, and imperfectly observed. The public will likely be skeptical of LVT because it is more difficult to observe land capital differences and yet the tax rates will be “the same” and perceived to be vertically inequitable, on the face. As Fischel (2015, p. 17) explained, Pittsburgh rolled back its SRT system when people rejected the idea that owners in the same neighborhood pay the same taxes even when their houses are of different sizes. In other words, LVT does not seem vertically equitable because “people in larger houses should pay proportionately more” (Fischel 2015, p. 17), but they do not. LVT acceptability therefore conflates with the positional-goods aspect of 3 housing, and this produces an obstacle for acceptance for which behavioral economics are well positioned to explore. The contribution of this work arises from using behavioral economics to analyze the efficiency of LVT, the positional-goods characteristics of housing, and the acceptability of LVT. To our knowledge, this is the first economic experiment on LVT. Another innovation in this experiment is that it makes endogenous the process by which land values appreciate; therefore, capturing the fundamental rationale behind a Georgian land tax—to capture the rent imbued by location. The treatments test (1) the social efficiency of LVT and SRT relative to UPT, or status quo, baseline; (2) the effect of a positional-goods signal on investment behavior; and (3) the public acceptability of the LVT and SRT relative to UPT using a voting mechanism. Experiment participants will take on the roles of different types of landowners making decisions about investing in their property (termed “improvements” in this experiment). Experimental economics techniques are ideally suited for studying LVT because there are limited instances of its adoption, the few examples of adoption of LVT is often incomplete (as SRT), actual LVT policies do not vary as much as might be desired in empirical analysis, implementation is complicated by inaccurate assessments, and any one location only has one policy treatment. Lab experiments with induced values overcome all of these shortcomings, providing the researcher with control and allowing analyses that test causal hypotheses. Furthermore,