Summaries of Articles

Abstract

Previous articleNext article FreeSummaries of ArticlesPDFPDF PLUSFull Text Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinkedInRedditEmailPrint SectionsMoreRetailer Remittance Matters: Evidence from Voluntary Collection AgreementsYeliz Kaçamak, Tejaswi Velayudhan, and Eleanor WilkingWhat are the effects of requiring online retailers to remit sales taxes, just as brick-and-mortar retailers do? Although goods purchased online have always been subject to similar sales/use tax as their brick-and-mortar counterparts, the responsibility to remit this tax has been on the consumer until recently. This issue has received considerable attention at the state and federal levels in light of the growing importance of online commerce and consequent efforts to stem base erosion. Yet there is little empirical evidence on how shifting the remittance obligation affects online and brick-and-mortar retail, and who bears the additional tax burden, if any, of imposing parity between the two retail channels. In this paper, we study VCAs — bilateral agreements between states and large online retailers to remit sales tax — enacted between 2009 and 2017. We explore how shifting the sales-tax remittance obligation on online sales to retailers achieves intended goals of retailer collection and parity between online and brick-and-mortar stores, as well as its consequences for sales-tax incidence.This paper has two main contributions. First, we document the impact of changing remittance obligations on the relative attractiveness of online and brick-and-mortar markets through their impact on the effective tax rate on online sales. We show that online retailers add sales taxes at checkout immediately following policy adoption, indicating that the agreements were binding, complied with, and measurable in our data. Shopping trips, where sales taxes were likely added, increase by 36 percent immediately after VCA adoption among large online retailers and show no change among brick-and-mortar retailers. This represents a significant change in the ability of consumers to use online retail as a means of evasion, as expenditure at these large retailers represents about half of their total online expenditure.Further, we find that consumers shift consumption to brick-and-mortar stores from online channels, suggesting that the differential remittance rules had conferred some advantage to online retailers. Consumers immediately reduce their online taxable expenditure due to VCAs at large online retailers by about 20 percent. Over time, brick-and-mortar household expenditure rises, even above the decline in online expenditure.Our second contribution is to document the incidence of the effective tax increase from increased compliance on online retail. We find that any effective tax increase was passed through to consumers, as the tax-exclusive price does not change after the VCA, suggesting that they bear the economic burden of the increased enforcement. However, among consumers, we do not find that the VCA substantially changed the distributional burden of US state sales taxes. Although online expenditure is a higher share of income among lower-income households, middle-income households tended to shift more consumption toward brick-and-mortar stores instead of reducing overall expenditure. Therefore, the change in overall taxable expenditure as a share of income due to the VCA did not vary significantly across income groups, suggesting that VCAs do not disproportionately impact low-income households. Our findings contribute to the ongoing policy debate on online retailer remittance of sales taxes.Bias in Tax Progressivity EstimatesJohannes KönigIn debates about income inequality and societal fairness, inevitably the issue of tax progressivity comes up. A tax system is progressive if marginal tax rates exceed average tax rates, so that the tax burden on the last dollar earned is larger than the burden on the first dollar. Hence, top earners will both pay more in absolute and in relative terms. Tax progressivity does not only bring about a more equal net income distribution, but it has important incentive effects: if the last dollar is taxed more than the first, individuals are going to be less willing to exert additional effort to earn the last dollar. Therefore, it is important to be able to quantify the progressivity of the tax system.In quantitative economic research — particularly structural economic models — a way to quantify tax progressivity has become immensely popular: the power function approximation. This approximation connects gross and net incomes by an exponential function with two parameters. The approximation enables researchers to calculate net incomes as well as average and marginal tax rates. Yet it has another extremely useful property: the exponent of the power function approximation captures the global progressivity of the tax system. In fact, it measures the residual income elasticity, that is, the percentage change of net income due to a percentage change in gross income. If the elasticity is smaller than one, the tax system is progressive, and if it is larger than one, it is regressive.The approximation is not only popular because it can quantify progressivity but also because it offers a very good fit to the data and is easy to estimate. The researcher only needs data on gross and net incomes. There are two ways to estimate the progressivity parameter: either directly using a nonlinear estimator or by taking the logarithm on both sides and using ordinary least squares.The latter estimation method is the most popular way to estimate the power function approximation, but it is generally biased. However, a nonlinear estimator that is equivalent to a Poisson regression gives consistent progressivity estimates. In this paper, I document the severity of this bias in several data sets and find that differences in estimates range from 6 to 14 percent. Generally, using ordinary least squares leads to an overestimation of progressivity. Further, I show that when biased progressivity estimates are used in subsequent model calculations, errors propagate and become quite severe. Finally, I offer guidance to the practitioner on how to estimate the power function approximation.The Effect of Tax Price on Donations: Evidence from CanadaRoss Hickey, Brad Minaker, A. Abigail Payne, Joanne Roberts, and Justin SmithThe tax price of a $1 donation is the after-tax cost of the donation for the donor. We estimate the tax price elasticity of giving using a large sample of Canadian tax filers from 2001 to 2016. The richness of the data allows us to control for a variety of important factors and estimate the responsiveness of Canadian households to changes in the tax treatment of donations. Canada supports charitable giving through nonrefundable tax credits, and we find take-up of these credits throughout the distribution of income. There is a kink in the credit rate schedule at $200 of donations, which makes it difficult to convincingly estimate the effect of tax credits on donations using conventional methods. We use a statistical technique from the literature to produce estimates with nice properties.We find strong evidence of responsiveness to the tax credits, estimating the tax price elasticity to be −1.9. This means that for every dollar remitted back to tax filers for their donations, donations increase by $1.90. We also estimate the responsiveness within intervals of the distribution of household income. We find strong evidence of responses throughout the distribution of income. Interestingly, the most responsive donors are at the bottom 20 percent and 21 percent–40 percent of the household income distribution. We explore whether this response is driven by donors changing their gift sizes in response to the change in tax price or by tax filers changing when to donate. We find evidence that the large responses at the bottom of the income distribution are driven by the decision to donate, while responses for the top 20 percent, top 10 percent, and top 1 percent are driven more by the decision of how much to donate.Taken together, our research shows that tax concessions for donations to charity are effective at increasing charitable giving. Our results show larger estimated effects than previously found in the literature, which we attribute to our sample including many more lower-income households with an incentive to donate.Income Guarantee Policy Design: Implications for Poverty, Income Distribution, and Tax RatesRobert Paul Hartley and Irwin GarfinkelIncome-guarantee policies featured prominently in public discourse around the 1960s, including the presidential campaigns by Richard Nixon and George McGovern as well as the writing of Nobel laureate economists Milton Friedman and James Tobin and of Nobel Peace Prize activist Martin Luther King Jr. After 50 years of the War on Poverty, guarantee designs are resurging in both the economics literature and policy proposals. The National Academies of Science have promoted a universal child allowance, which was temporarily realized as part of the American Rescue Plan in response to the COVID-19 pandemic. To inform the expected impacts on poverty, income distribution, and tax rates, this paper simulates a moderately sized $250 monthly guarantee per person financed by a fundamental tax reform that eliminates all income tax deductions and institutes proportionally higher marginal tax rates (MTRs). We assume that households may respond with changes in labor supply relative to both the guarantee benefit and the financing mechanisms. Furthermore, we consider a general framework of important questions for an income guarantee design and discuss how the major impacts would compare.For our benchmark guarantee policy, poverty would fall by about 40 percent with a 20 percent reduction in inequality between the 90th and 10th percentiles of income. Fundamental tax reform would finance 60 percent of the total cost of $948 billion, and the remaining costs could be financed by raising MTRs by 23 percent across all brackets. Under this reform, households with incomes below $200,000 would be net beneficiaries, and the top MTR would rise from 37 to 45.5 percent. Although universal, such an income guarantee would have greater relative benefits for disadvantaged groups. Children and larger families would experience the largest antipoverty effects, as would caregivers, who are often unpaid or lower income. A universal income guarantee would also reduce the Black-White poverty gap by 16 percent — more for a larger benefit size — because of disproportionate poverty rates among those racialized as Black.We explore various options for designing an income guarantee, such as generosity amount, income eligibility, interaction with means-tested transfers, and forms of income taxation. Along with fundamental tax reform, a higher income guarantee of $500 monthly per person would reduce poverty and inequality by about 70 and 35 percent, respectively, and could be financed with effective tax rates similar to those observed in the US tax code as recently as 1981. Phasing out benefits at higher incomes can meaningfully reduce inequality without diminishing poverty reduction. Reducing the program costs of a guarantee by offsetting means-tested transfers substantially limits poverty reduction impacts. We also compare a flat surtax on all taxable income as opposed to proportionally increasing MTRs, and the implications for poverty and inequality are negligible.A primary function of a guaranteed income program is the maintenance of a livable income to support families, and moderate-sized guarantees could greatly reduce poverty at tax rates consistent with those in the optimal income tax literature. Previous articleNext article DetailsFiguresReferencesCited by National Tax Journal Volume 76, Number 2June 2023 Published for: The National Tax Association Article DOIhttps://doi.org/10.1086/726079 © 2023 National Tax Association. All rights reserved.PDF download Crossref reports no articles citing this article.