Contrary to common perception, many fixed-income investors have not suffered unusually low real interest rates in and after the Great Recession of 2008. This is because taxable investors must first pay taxes on nominal interest returns, before inflation further reduces their earned real interest rates. To obtain the same real after-tax yield, investors need more than one-to-one compensation for inflation. As a result, long-term Treasury bonds have been no less attractive for taxable investors in 2016 (with a 1.0% post-tax real yield) than they were in 2006 (0.5%), 1976 (-1.7%), 1966 (0.9%), and 1956 (0.8%), although they have been less attractive than they were in 1996 (2.4%) and 1986 (2.9%). Short-term Treasury bond yields have been on the low side but have also not been particularly unusual.

When choices are made from ordered lists, individuals can exhibit biases toward selecting certain options as a result of the ordering. We examine this phenomenon in the context of consumer response to the ordering of economics papers in an e-mail announcement issued by the NBER. We show that despite the effectively random list placement, papers listed first each week are about 30% more likely to be viewed, downloaded, and subsequently cited. We suggest that a model of skimming behavior, where individuals focus on the first few papers in the list due to time constraints, would be most consistent with our findings.

This paper analyzes a new way of reducing the major individual tax expenditures: capping the total amount that tax expenditures as a whole can reduce each individual's tax burden. More specifically, we examine the effect of limiting the total value of the tax reduction resulting from tax expenditures to two percent of the individual's adjusted gross income. Each individual can benefit from the full range of tax expenditures but can receive tax reduction only up to 2 percent of his AGI.Simulations using the NBER TAXSIM model project that a 2 percent cap would raise $278 billion in 2011. The paper analyzes the revenue increases by AGI class. The 2 percent cap would also cause substantial simplification by inducing more than 35 million taxpayers to shift from itemizing their deductions to using the standard deduction. For any taxpayer for whom the 2 percent cap is binding, a cap would reduce the volume of wasteful spending and the associated deadweight loss. Even for those taxpayers for whom the cap is not binding but who are induced by the cap to shift from itemizing to using the standard deduction, the deadweight loss associated with deductible expenditures would be completely eliminated

This paper examines how the Alternative Minimum Tax (AMT) affects the weighted average marginal tax rates that apply to various components of taxable income and the subsidy rates on various income tax deductions. It also considers how several AMT reform proposals would affect the number of AMT taxpayers, total AMT liability, and weighted average marginal tax rates. On average, the AMT has only a modest impact on the weighted-average marginal tax rates for most sources of income although some tax-payers face substantially higher tax rates, and others substantially lower rates, as a result of the AMT Our projections show that modest increases in the AMT exclusion level have substantial effects on the number of AMT taxpayers, and that indexing the AMT parameters would reduce the number of AMT payers in 2010 by more than 60 percent.

Using the TAXSIM model for the period 1962-95, we consider the federal tax system's impact as an automatic stabilizer. Despite the many changes in the tax system, there has been relatively little change in its role as an automatic stabilizer. We estimate that individual federal taxes offset perhaps as much as 8 percent of initial shocks to GDP. We also suggest that the progressive income tax may help to stabilize output via its effect on the supply of labor, an additional effect that may even be of similar magnitude to the more traditional path of stabilization through aggregate demand.

Medicare currently transfers roughly $ 200 billion annually to the elderly. The difference between medicare expenditures flowing into the state and annual Medicare taxes leaving the state—both on an annual and a lifetime basis—varies widely across states. Lifetime Medicare transfers, defined as the present value of benefits less accumulated (and predicted) Medicare taxes, are, on average, $17,386 per household. However, these benefits vary widely across states, so that residents of some states benefit by more than $40,000 per household, while residents of other states actually lose, in the sense of paying more in taxes than they will receive in benefits. Expected Medicare benefits per dollar of taxes paid varies from $.84 for residents of Oregon to $1.93 for residents of Louisiana. The differences are present even after controlling for a variety of factors such as the migration pattern of retirees, the price of health care, and the underlying health status of people in each state. These variations raise questions about the fairness and efficiency of the Medicare program.

This paper presents new information on the fraction of adjusted gross income, and of wages and salaries, that is reported by taxpayers in the top one half of one percent of the income distribution.This corresponds to roughly five hundred thousand households in the late 1990s.This paper relies on data from the Treasury's Individual Income Tax Model for the period 1960-1995.The definition of adjusted gross income is standardized, so that changes in the tax law do not affect the measured concentration of AGI.The results suggest that the share of AGI reported by the highest income households increased significantly between the early 1980s and the mid-1990s, with most of the increase taking place in the years immediately following the Tax Reform Act of 1986.While we find some evidence of transitory changes in the concentration of income around major tax changes, which may be the result of income retiming by high income taxpayers, re-timing does not seem to explain most of the changes since 1986.

This paper describes a new household-level data file based on merged information from the IRS Individual Tax File, the Current Population Survey, the National Medical Expenditure Survey, and the Consumer Expenditure Survey. This new file includes descriptive data on household income as well as consumption. The data file can be linked to the NBER TAXSIM program and used to evaluate the distributional effects of changing the federal income tax code, as well as the distributional effects of replacing the individual income tax with a consumption tax. We use this data file to analyze the long-run distributional effects of adopting a national retail sales tax that raises enough revenue to replace the current federal individual income tax and corporation income tax, as well as federal estate and gift taxes. Our results highlight the sensitivity of the change in distributional burdens to provisions such as lump sum transfers, sometimes called "demogrants," in the retail sales tax plan, and to the choice between income and consumption as a basis for categorizing households in distribution tables.

During the early years of its existence, the National Bureau of Economic Research (NBER) assembled an extensive data set on all aspects of the pre-WWII macroeconomy. Until 1978, this data set existed only on the handwritten sheets to which the early NBER researchers copied the data from original sources. In 1978, the Inter-University Consortium for Political and Social Research (ICPSR) transferred the data to magnetic tape. A number of researchers have used the ICPSR tape, but two key problems discourage many from taking advantage of this unique data set. The first is that modern econometric software does not have the ability to read the obsolete ICPSR format. The second is that the process of transferring the data from the NBER's handwritten sheets to the ICPSR tape introduced a number of mistakes. We have eliminated these two impediments to the use of the NBER data set by converting the ICPSR tape to a portable format and by verifying the accuracy of the data using the NBER's original handwritten sheets. The data set is now available on the Internet and can be accessed using standard gopher or web-browser software.

The 1993 tax legislation raised marginal tax rates to 36 percent from 31 percent on taxable incomes between $140,000 and $250,000 and to 39.6 percent on incomes above $250,000. This paper uses recently published IRS data on taxable incomes by adjusted gross income class to analyze how the 1993 tax rate increases affected taxable income, tax revenue, and economic efficiency. Our estimates are based on a difference-in-difference procedure comparing growth of taxable incomes among taxpayers with AGIs over $200,000 to the growth of incomes of lower income taxpayers. We use the NBER TAXSIM model to adjust for interyear differences in the composition of the two taxpayer groups. The results show that high income taxpayers would have reported 7.8 percent more taxable income in 1993 than they did if their tax rates had not increased. Because of the high threshold for the increase in tax rates, this decline in taxable income caused the Treasury to lose more than half of the extra revenue that would have been collected if taxpayers had not changed their behavior. The deadweight loss caused by the higher marginal tax rates (including the effects on labor supply and on consumption of goods and services favored by deductions and exclusions) is approximately twice as large as the $8 billion in revenue raised by the 1993 tax rate. Several possible statistical biases could cause the estimated effect of the tax changes to either underestimate or overestimate the true long-run effect. The paper concludes with a discussion of these problems and of plans for future analysis.

The present paper examines the efficiency and revenue effects of several alternative tax treatments of two earner families using estimates of the compensated elasticities of the labor supply of married women based on the experience with the 1986 tax rate reductions. The analysis of alternatives is based on the NBER TAXSIM model which has been modified to incorporate separate estimates of the earnings of spouses. The marginal tax rates explicitly incorporate the Social Security payroll taxes net of the present actuarial value of future retirement benefits. Three general conclusions emerge in this paper. First, the existing high marginal tax rates on married women cause big eadweight losses that can be reduced by alternative tax rules that lower marginal tax rates. Second, the behavioral responses to the lower marginal tax rates induce additional tax payments that offset large fractions of the 'static' revenue losses. Third, there are substantial differences in cost- effectiveness among these options, i.e. in the revenue cost per dollar of reduced deadweight loss. Several of the options are sufficiently cost- effective that they could probably be combined with other ways of raising revenue to produce a net reduction in the deadweight loss of the tax system as a whole. We are aware, however, that the current framework is very restrictive in three ways. It ignores the response of the primary earner to any change in tax rates on spousal income. It defines the labor supply response narrowly in terms of participation and hours, excluding other dimensions of labor supply. Taxes affect not only the labor supply of men and women but also change taxable income through changes in excluded income and deductions. These changes in taxable income are the key variable for influencing tax revenue and the deadweight loss of alternative tax rules.

We use a sample of actual tax returns to compute estimates of the marriage tax by income class under the new tax law. We predict that in 1994, 52 percent of American couples will pay an average marriage tax of about $1,244, and 38 percent will receive an average subsidy of about $1,399. These aggregate figures mask a considerable amount of dispersion in the population. Under the new law, the marriage tax for certain low-income families can exceed $3,000 annually; for certain very-high-income families it can exceed $10,000 annually.

Catastrophic medical expenses are an important economic risk facing the elderly. Little is known about the persistence of such out-of-pocket medical costs. We measure the time-series property of medical costs using information on medical deductions from a panel of tax returns. During the period of analysis, 1968-73, taxpayers could deduct medical expenses above 3 percent of income. We correct for the resulting censoring bias using multivariate Tobit estimated with a variant of the smoothed simulated maximum likelihood (SSML) method. The data suggest that the burden of out-of-pocket medical expenses is substantially larger for lower income families. Furthermore, the estimated coefficients suggest substantial time-persistence in out-of-pocket medical care costs; a $1 increase in out-of-pocket medical spending is predicted to increase future spending by an additional $2.80. These results may shed light both on the social value of catastrophic health insurance as well on aggregate saving behavior.

This paper uses tax return data for the period 1951-1990 to investigate the rising share of adjusted gross income (AGI) that is reported on very high income tax returns. We find that most of this increase is due to a rise in reported income for the one quarter of one percent of taxpayers with the highest AGIs. The share of total AGI reported by these taxpayers rose slowly in the early 1980s, and increased sharply in 1987 and 1988. This pattern suggests that at least part of the increase in the income share of high-AGI taxpayers was due to the changing tax incentives that were enacted in the 1986 Tax Reform Act. By lowering marginal tax rates on top-income households from 50% to 28%, TRA86 reduced the incentive for these households to engage in tax avoidance activities. We also find substantial differences in the growth of the income share of the highest one quarter of one percent of taxpayers, and the share of other very high income taxpayers. This casts doubt on the view that the increasing inequality of reported incomes at very high levels is driven by the same factors that have generated widening wage inequality at lower income levels.

The TAXSIM model is now 17 years old. The original tax calculator was written by Amy Taylor to estimate the effects of tax deductibility on charitable giving [Feldstein and Taylor, 1976]. The model was given its present form several years later by another Harvard student, Daniel Frisch, for a study of the incidence of a proposed integration of the corporate and personal income tax systems [Feldstein and Frisch, 1977]. Subsequent studies by Martin Feldstein, his students, and others proved the data so rich and the model so useful that TAXSIM has been updated every year since and used in scores of NBER projects. This note provides a short history and description of TAXSIM, with special emphasis on the state tax calculator used by six of the papers in this issue of the JPAM.