It’s by Donald Marron and Eric Toder. You can read the full paper here.

— “Many tax preferences are effectively spending programs. Adding these spending-like tax preferences to federal outlays and receipts makes the government appear about 4 percent of GDP larger. The 1986 tax reform cut these benefits, but they have since rebounded to a larger share of GDP than before.”

The U.S. Federal Reserve building. (Brendan Smialowski/Bloomberg)

— “Making these adjustments requires caution, however. It is tempting, for example, to simply add together all the provisions that the federal government identifies as ―tax expenditures and treat those effectively as spending. But that goes too far. Not all tax expenditures are the functional equivalent of spending.”

— “A number of authors have suggested distinguishing between tax expenditures that represent disguised spending and those that represent structural departures from a comprehensive income base but do not replace any clearly identifiable direct spending program (Fiekowsky 1980; Kleinbard 2010; Marron 2011; Shaviro 2004; Toder 2005). Although these authors use different formulations and labels, they all focus on a subset of tax expenditures that replace subsidies or transfer payments that could otherwise be delivered as outlays. In this view, which we share, it is only those ―spending-like tax preferences that should be included in a ―spending total designed to measure the size of government.”

— “Clear spending substitutes are those tax expenditures that encourage selected activities or aid specific groups of taxpayers and could be replaced by similar programs delivered as direct outlays. Examples are renewable energy credits, the home mortgage interest deduction, the exclusion from tax of employer-provided health insurance and health benefits, and tuition tax credits. All these provisions subsidize identifiable activities (renewable energy, housing investment, health insurance, and college tuition), try to promote definable social goals (reduced greenhouse gas emissions, increased homeownership, broader health insurance coverage, and increased college attendance), and could be designed as outlays.”

— “Other provisions represent broad choices in tax policy design but are not associated with any clear spending objective. For example, many economists favor consumption instead of income as a tax base, and our current income tax can be thought of as a hybrid between consumption and income taxation. The treatment of saving in qualified retirement saving plans, which allows most workers to defer tax on contributions until the proceeds of their contributions and investment earnings are withdrawn from the account, is an example of a provision that taxes the return to saving based on consumption instead of income tax principles.”

— “The ten largest tax expenditures in terms of 2012–16 budgetary costs (revenue losses plus outlays for refundable credits) identified by OMB (2011) will cost $4 trillion between 2012 and 2016—about 65 percent of the cost of all tax expenditures over that period (table 1). We classify six of them as spending substitutes: the exclusion of employer contributions for medical insurance and medical care, deductibility of mortgage interest on owner-occupied homes, exclusion of net imputed rental income on owner-occupied homes, deductibility of nonbusiness state and local taxes other than on owner-occupied homes, the earned income tax credit, and deductibility of charitable contributions other than education and health.

— “We classify the other four among the ten largest provision—step-up in basis for capital gains at death, 401(k) plans, accelerated depreciation of machinery and equipment, and the special rate on capital gains (excluding other provisions that tax income in selected industries as capital gains)—as general tax policy choices instead of spending substitutes.”

— “Between 1985 and 1988, tax expenditures dropped from 8.7 percent of GDP to 6.0 percent. Many large tax expenditures were eliminated, including the investment tax credit, accelerated depreciation on rental housing, and preferential rates for capital gains, and others were substantially scaled back. In addition, lower marginal tax rates reduced the value of remaining individual and corporate income tax expenditures structured as exclusions, exemptions, deductions, and deferrals. Since then, however, tax expenditures have grown as policymakers enacted new provisions, as marginal tax rates increased, and as certain sectors, e.g., health insurance, grew faster than the economy.”

— “We use these estimates of spending-like tax preferences to construct a revised series of total federal spending and revenue in 1985 and from 1988 through 2016...Accounting for those resources, the federal government has been around 4 percent of GDP larger in recent years than budget figures indicate. In 2010, for example, both federal spending and revenues were about $600 billion more than the official budget measures.

— “Thus far, we have emphasized that some tax provisions are really spending in disguise. It is also important to consider whether some spending provisions may be revenues in disguise. The budget records the cash inflows from some policies as negative spending rather than as revenues. These inflows are known as offsetting collections and receipts because they are viewed as reducing government spending rather than adding to revenues. The premiums that beneficiaries pay for Medicare coverage, for example, are recorded as negative spending because they reduce the net subsidy that the government provides through the program.”

— “When we add in about $230 billion in user charges, finally, effective federal spending rises to 25.4 percent of GDP. By this metric, federal spending was more than one-quarter of economic activity in 2007, almost 30 percent more than officially reported.”