Why the FairTax Will Work
Bartlett's Unfair Attack on the FairTax
In his December 24, 2007 Tax Notes article, “Why the FairTax Won’t Work,” Bruce Bartlett purports to critique the FairTax, a proposal to replace almost all federal taxes with a retail sales tax plus a rebate. In fact, Barlett’s article has little to say about the FairTax and even less to say that’s accurate. Instead, most of his article misstates research on the FairTax, criticizes unnamed proponents of the FairTax, lambasts unattributed views of the FairTax, and engages in political punditry. This paper takes a close look at Bartlett’s “analysis,” exposing his repeated use of straw men for what it is rhetoric disguised as economics. (1)
The FairTax was proposed over a decade ago. It now has hundreds of thousands, if not millions of supporters. The vast majority of these supporters are not economists. Some are radio talk show hosts, some are Congressman, some are painters, salesmen, investment bankers, homemakers -- you name it. Many FairTax supporters are wildly enthusiastic about the plan, and some have certainly made fantastic claims in its name that have no economic merit. Some of these claims, no doubt, have even found their way onto the FairTax.org website, which is manned, in the main, by volunteers, not economic PhDs.
Bartlett seems convinced that FairTax advocates and FairTax.org is out to con the public, and he makes repeated references to their dishonesty and distortions. In my own dealings with the leaders of FairTax.org, I’ve seen no evidence of any such thing. Quite the opposite. But I have learned that the way economists think and talk about tax reform is different from the way the public does and that the meaning of words like “revenue neutral” need to be spelled out in fine detail.
In any case, shooting the messenger, as Bartlett does, is no substitute for understanding the message. The FairTax represents a significant tax reform proposal that deserves to be taken seriously. Doing so requires examining the proposal itself, not some misconception or modification of what’s being proposed.
Bartlett begins his critique by accosting unnamed messengers (referenced by “FairTax advocates”) for supposedly suggesting that consumer, producer, and factor prices would be unaffected by the FairTax, with workers simply keeping the income and payroll taxes that would otherwise have been deducted from their paychecks.
Clearly, such an outcome is inconsistent with elementary economics, and no serious student of the FairTax would assert such an outcome. Nonetheless, Bartlett’s devotes, by my count, some 31 paragraphs, including a primer on the Great Depression, to demolishing this straw man. (2)
Had Bartlett wanted to find public statements in FairTax.org-sponsored research acknowledging that wages must fall if consumer prices don’t rise and that consumer prices must rise if wages don’t fall, he could easily have done so. He could simply have read the November 13, 2007 issue of Tax Notes which contains my co-authored article calculating the FairTax tax rate needed to achieve real revenue neutrality (i.e., permit federal and state governments to purchase the same real goods and services and make the same real transfer payments). (3) In the course of calculating the tax rate needed to achieve real revenue neutrality, my co-authors and I pointed out that the economics of the FairTax do not, apart from transition issues, depend on whether the Federal Reserve does or does not accommodate the FairTax by permitting the price level to rise. In discussing the possibility that the Fed would not accommodate, we pointed out that in this case, producer prices, wage rates, and rental rates would all fall.
Bartlett’s second concern lies in the calculation of the FairTax rebate. He takes issue with the proposal’s treatment of childless households, suggesting that the size of their rebates are too large. From this Bartlett surmises that Congress would raise the rebates to households with children thereby “greatly increasing the cost of the rebate.” But if the rebates to childless households are too large, the solution is not to make everyone’s rebate too large, but rather to cut rebates to childless households and, thereby, reduce required FairTax revenue.
Bartlett’s next “critique” is even less memorable. He claims that Americans won’t perceive their monthly FairTax rebate check as progressive even though the rebates will be a much higher percentage of the resources of the poor than they will be of the rich. Instead, he says, households will view the FairTax as proportional because everyone will have to pay the same FairTax rate when they spend their money, no matter the source of their money. This is no different from claiming that people judge tax fairness based on their marginal rather than their average tax rates. Were this the case, marginal tax rates under our current tax system would presumably be set to rise monotonically with income, which is certainly not the case. (4)
Bartlett’s contention here is symptomatic of a pervasive failure to stick to economics. Bartlett’s expertise does not, to my knowledge, extend to psychology or political science. So when he asks his readers to accept his assessment of perceptions or his judgment of political reactions, I, for one, start feeling queasy.
Examples in the political domain are Barlett’s statement “Nor is Congress going to send rebate checks to billionaires” (page 1243) and “… Congress would exempt some goods and services from the FairTax (page 1243).”
My guess is that billionaires won’t bother applying for the FairTax rebate and Congress wouldn’t worry about sending them rebate checks if they do. After all, Congress now sends billionaires monthly Social Security checks and pays billionaires’ healthcare bills through Medicare. As for Congress exempting particular goods and service, I see no reason this will happen. The rebate will suffice to fully insulate the poor from paying any tax on net. Exempting anyone or anything from the FairTax means raising the FairTax rate on everyone and everything. This is the real beauty of the proposal; it gives everyone a vested interest in all spending decisions.
Having said this, let me hasten to say that my political presumptions have as much professional basis and claim to acceptance as Bartlett’s, which is to say none. Rather than make what boils down to religious arguments about what will or won’t happen in the political arena we economists should, I feel, evaluate the FairTax at face value. I.e., we economists should keep to our knitting and leave others to read political entrails.
Bartlett’s first significant economic critique of the FairTax appears five pages into his article, where he states “… there would be an enormous shift in the tax burden from the wealthy to those with lower and middle incomes.” (page 1245) As proof of this proposition he reproduces a table (his table 5, p. 1245) generated by the Treasury’s Office of Tax Analysis entitled “Distribution of the Federal Tax Burden Under the FairTax.”
Notwithstanding its source, there are two major problems with the Treasury’s analysis of the FairTax’s progressivity. First, the Treasury produced this table in response to a request from President Bush’s Advisory Panel on Federal Tax Reform. The Tax Reform Panel was charged with considering reform of the personal and corporate income taxes. Its purview did not extend to reforming the payroll tax. As a consequence, although the Treasury referenced the FairTax in the table, the Treasury completely ignores one of the most progressive elements of the FairTax, namely the elimination of the highly regressive FICA tax. Bartlett mentions that the table considers replacing only the income tax. But he fails to mention that were the table to include replacing the payroll tax, the FairTax would look much more progressive.
The second problem is the static/current-income framework used to examine progressivity. Economists have long recognized that current income is not a reliable measure of remaining lifetime economic resources and that it provides highly misleading assessments of tax fairness. (5) Lifetime economic resources refers to the sum of a household’s current net worth plus the present value of its future labor earnings its so-called human wealth. It is the real measure of the household’s intertemporal spending power.
Remaining lifetime resources need bear no relationship whatsoever to current income. Take Bill Gates. His resources run into the tens of billions of dollars, but his current income could well be zero (even negative) because of poor performance of Microsoft stock or his other financial holdings. The Treasury would, nonetheless, classify Bill Gates together with the homeless as among the poorest households in America. Were one to divide Gates’ putative FairTax payments by his income of zero, the implied average tax rate would be infinite; i.e., one of the nation’s richest men, if not its richest, would be viewed, in this hypothetical, as among its poorest and judged as being taxed at an astronomical rate.
This is just one of many problems associated with using current income as a proxy for lifetime resources. A second problem is that current labor income may be temporarily low although lifetime resources is quite high. Doctors doing residencies come to mind. So do lawyers and other professionals at the early stages of their careers. The flip side of this coin is that current labor income can be temporarily high, which is the case for almost all Americans right before they retire. Since households make spending decisions based on their assets and their lifetime labor income, young doctors who spend in anticipation of the future earnings will be measured by the Treasury as paying high FairTaxes on low incomes and pre-retirees will be measured as paying low FairTaxes on high income. I.e., the analysis will discover tax regressivity where none exists.
Why the Treasury persists in conducting highly misleading and professionally irresponsible incidence analyses is beyond me. But Bartlett, who is, notwithstanding my concerns about his FairTax views, an outstanding economist, should treat such studies for what they are: content free.
As Bartlett knows, consumption taxation (ignoring any rebate and assuming a single, fixed rate) is neither progressive nor regressive when properly measured relative to remaining lifetime resources. Instead, it’s proportional. It’s proportional for the simple reason that remaining lifetime consumption, C, equals remaining lifetime income, R, so taxing consumption at a fixed rate is the same as taxing lifetime resources at that rate.
To see this algebraically, note that C = R, which is the household’s budget constraint. Thus, taxing consumption at rate θ generates the same lifetime taxes (θC), measured in present value, as one would collect in present value by taxing resources directly at rate θ; i.e., θC = θR. Thus, under a consumption tax, whether you tax a household’s spending or tax what the household uses to finance its spending (its wealth plus its current and future labor earnings) doesn’t matter to what the household ultimately pays. (6)
The household’s average lifetime tax rate, call it ζ, is calculated by dividing lifetime taxes, θC, by lifetime resources, R. Thus ζ = θC/R, but since C = R, we have that ζ = θ regardless of the size of R, which is another way of saying that consumption taxation is neither regressive nor progressive, but simply proportional.
Now consider the FairTax, which includes a rebate whose present value, I’ll assume, equals B regardless of the size of R. In this case, the average lifetime net tax rate R is given by the formula ζ = (θC - B)/R or ζ = θC/R - B/R or ζ = θ - B/R, which rises with R. Hence, by including the rebate, the FairTax transforms what would otherwise be a proportional consumption tax into a progressive consumption tax.
How does this square with Bartlett’s statement that “… there would be an enormous shift in the tax burden from the wealthy to those with lower and middle incomes.”? It doesn’t. Since the FairTax is obviously progressive when properly measured relative to lifetime resources, Bartlett’s statement requires the current tax system to be even more progressive when measured relative to lifetime resources. Showing this to be the case requires actually comparing remaining lifetime net tax rates for households with different levels of lifetime resources. And it requires doing this analysis on a cohort-by-cohort basis. The reason is that a given amount of lifetime resources means less to a younger household than it does to an older one because the younger one has to spread those same resources over many more years of life.
Unfortunately, Bartlett doesn’t present his own cohort-specific remaining lifetime tax progressivity comparisons. Nor does he cite the other studies that either perform this analysis directly or provide indirect evidence on the results of such an analysis. (7)
The table above presents results from Kotlikoff and Rapson (2006), which directly compares the FairTax’s progressivity relative to remaining lifetime resources with that under our current federal tax system. The analysis is performed for a set of stylized households that differ with respect to earnings and other economic characteristics, like house value, mortgage payments, college tuition payments, that are scaled to their labor earnings.
As indicated, the FairTax lowers average remaining lifetime net tax rates for all the stylized households, but the percentage reductions are larger at the lower end of the earnings distribution; i.e., when it comes to working households, the FairTax makes our tax system more progressive.
The reduction in average net tax rates is quite striking, prompting the question, "Is there a free lunch?" How can the FairTax generate lower net tax rates for everyone and still pay for the same real government expenditures? The answer is that not everyone’s average net tax rate falls. The table focuses on workers, including some very high-earning workers. It doesn’t consider households whose major economic resource is their wealth. For such households the FairTax will deliver a net tax hike compared to the current system.
Consider, for example, your typical billionaire, of which America now has more than 400. (8) These folks are invested, no doubt, primarily in equities on which they pay taxes at a 15 percent rate, whether their income comes in the form of capital gains or dividends. In addition to having the income from their wealth taxed at a low rate, the principal of their wealth is completely untaxed either directly or indirectly. Assuming they and their heirs spend only the income earned on the wealth each year, the tax rate (the ratio of the present value of income taxes to the current value of resources, which in this case is just their wealth) is 15 percent. In contrast, under the FairTax, the effective tax rate is 23 percent. Hence, the wealthy will see their average tax rate almost double when the FairTax is introduced.
The proposition that the FairTax lowers taxes on workers and raises taxes on the rich is not surprising. It’s precisely what the mathematics of consumption taxation implies. Indeed, moving from an income and payroll tax to a consumption tax is, mathematically speaking, equivalent to moving to a tax system that taxes labor income at a lower rate (than under the income and payroll taxes) and also taxes wealth. (9) Bartlett’s suggestion that from workers’ perspective “It’s all a wash.” is simply wrong.
How can a proposal that lowers taxes on workers and imposes a tax on wealth be branded as regressive? Good question. But it would behoove those who care about social justice to think again about this proposal. It’s not every day that wealthy Americans, many of them Republicans, propose taxing their wealth, albeit indirectly.
Well rich members of today’s older generations may be a concern of Barlett. They aren’t a concern of mine. Our country has spent the past five decades transferring ever greater sums from young workers to contemporaneous older generations, including extremely wealthy members of older generations. The most recent example his the introduction of Medicare Part D’s prescription drug benefit. This transfer to current and near-term elderly has a present value cost of some $10 trillion.
This process of passing the generational buck constitutes fiscal child abuse, pure and simple, and has left the country with a fiscal gap (difference between the present value of future federal spending and taxes), which, by my estimate, now totals $70 trillion. There is simply no way we can deal with this huge insolvency without asking the wealthy, including wealthy older Americans, to pay substantially more taxes.
What about the poor elderly living off of Social Security? How do they fare under the FairTax? They are made better off. The CPI-indexation of Social Security benefits will maintain the real purchasing power of their benefits, but they’ll also receive the rebate. Indeed, for elderly living at the poverty level, the living standard increase will be roughly 23 percent!
Quoting the FairTax Rate
Bartlett claims that “FairTax supporters” are being dishonest when they reference the 23 percent rate because the mark up at the store will be 30 percent. Here’s his statement: “I think it is reasonable to conclude that the FairTax’s emphasis on the 23 percent tax-inclusive rate rather than the 30 percent tax-exclusive rate is a ruse designed solely to increase support for the proposal above what would be the case if it were generally known that the appropriate rate assumption is 30 percent.”
The way we quote the FairTax in public discussions shouldn’t be governed by how Bruce Bartlett feels about FairTax supporters. It should be quoted to make an apples-to-apples comparison with the tax rate under our existing tax system, which is, after all, the one whose replacement is being contemplated. Consequently, the only appropriate way to quote the FairTax is on a tax-inclusive basis, i.e., to reference the proposed 23 percent.
If Bartlett insists on discussing the FairTax in terms of the tax-exclusive 30 percent rate, that’s his prerogative. But it then behooves him to quote the income plus FICA tax on a tax-exclusive basis. For example, typical middle-class workers paying a 15.3 percent marginal payroll tax (including the employer’s portion) and a 25 percent income tax, would be quoted as having a 66 percent effective tax rate! One wonders why Bartlett failed to provide such a quote.
The Size of the FairTax Rate and the Treatment of Government
Here, again, Bartlett is off the mark. The FairTax is designed to be spending neutral, not simply neutral with respect to nominal revenues. And our analysis, which shows that a 23.82 percent rate is needed for real revenue neutrality, does, indeed, ensure that governments have enough revenue to purchase the same real goods and services and make the same real transfer payments. We point this out in the third paragraph of our paper, and it’s obvious from our equations.
These equations require the federal government to pay the FairTax, but also compensate it with extra revenue to cover these payments. Our formula for the tax rate, ignoring the rebate and other non-essentials, is ti (C+G) = G , where ti is the effective (tax-inclusive) rate, C is real household consumption expenditure, and G is real government expenditures. Now ti = te/(1+te), where te is the nominal tax rate (the 30 percent mentioned above). Plugging this formula for ti into the budget equation, we have the following: te (C+G) = G(1+te); i.e., the nominal revenue (the equation’s left-hand side) collected by levying taxes on both household and government consumption provides enough nominal revenue to cover the government’s nominal costs of buying the initial amount of G, where these costs include paying the FairTax.
Had Bartlett read our paper more carefully he would also have learned that state governments don’t need to be compensated by the federal government when the FairTax is implemented for the following reasons. First, if, as we assume, the states levy their sales taxes on top of the FairTax, their nominal revenues from sale taxes will rise by the nominal FairTax rate, thereby ensuring that the real purchasing power of the sales tax revenue remains fixed.
As for state income taxes, in raising consumer prices relative to wages (upon which state income taxes are levied), the FairTax reduces the real value of state sales taxes. To see this, suppose that Joe, a Massachusetts taxpayer, pays $3000 in state income taxes and assume, for simplicity, that the FairTax leads to zero decline in his wages and a 30 percent increase in consumer prices. If Massachusetts doesn’t change its state income tax rate, Joe’s real state income tax payment will fall. The reason is that the $3,000 will now only purchase $3,000/(1.3) in goods and services. The FairTax, in effect, gave Joe a state income tax cut, measured in real terms in terms of the goods and services his tax payment can purchase. The loser here is the Massachusetts state government, which sees its real revenue decline. But if Massachusetts simply raises its income tax rate by 30 percent, it will restore its real revenue collection.
This point is rather subtle. (12) One way to think about it is that the state income taxes become, in effect, deductible from the FairTax insofar as state taxpayers hand over nominal, not real revenue to the state governments when making their state income-tax payments.
But, getting back to compliance, it’s clear that the incentive to cheat will be significant. But it’s also clear that the incentive to evade income taxes is greater than would be the incentive to evade the FairTax. Moreover, much of the evasion under the FairTax would come in the same form and done by the same actors as evasion under the income tax; for example, a waitress who fails to collect FairTax on tips is also, presumably, not paying income taxes on those tips under the current system. Hence, under the current system, she’d not only failing to disclose part of national income. She’s also failing to disclose part of national consumption.
NIPA (National Income and Produce Account) imputations for underreported income are, therefore, very closely related to NIPA imputations for unreported consumption. These consumption imputations appear to represent about 2 percent of the NIPA household consumption total that my co-authors and I used to measure the requisite FairTax rate. Incorporating a 2 percent lower base assumption would certainly raise the requisite FairTax rate. (14) But incorporating reasonable growth assumptions would lower it. My sense is that the two corrections would roughly offset each other.
The reasons I’m not overly worried about evasions number five. First, the vast majority of retail sales occur in major retail outlets, like WalMart. Second, we’ll have vastly fewer taxpaying entities (14 million rather than more than 100 million) on which to focus our enforcement efforts. Third, we’ll have hundreds of thousands of otherwise unemployed IRS agents, accountants, and tax attorneys to enlist to enforce a single tax. Fourth, we can always compel firms to report, via 1099-type forms, their sales to other firms. This will provide the same records that arise under a VAT. Fifth, the government can stipulate that all retail sellers provide buyers with a written receipt, regardless of whether the transaction is or is not in cash, specifying that the FairTax has been paid. Thus if sellers don’t mail in the taxes, there will be a paper trail of the evasion. Sellers who try to bribe buyers to forgo receipts and split the FairTax among themselves run the risk of having the buyer turn them in.
The FairTax is, I should add, equivalent to a VAT except for excluding taxation of business-to-business sales, which yields, under the VAT, no net revenue. Virtually every developed country with the exception of the U.S. has a VAT. Since the FairTax is a VAT, except for its treatment of businesses transactions and its inclusion of a rebate, no one should view the FairTax with alarm. Collecting revenues by taxing retail sales is a tried and true practice.
The elimination of the housing distortion is just one of a very large and very important set of distortions that the FairTax would end or dramatically reduce. These include distortions of labor supply, saving, corporate financing decisions, corporate retention decisions, investment decisions, foreign retention decisions, bequest and gift decisions, and charitable contribution decisions. Eliminating these distortions as well as the enormous amount of time and effort involved in compiling with existing federal taxes could, itself, raise GDP by some 5 percent a year in my estimate. This is another reason to believe that the FairTax rate would be no greater than 23 percent and could be significantly lower. Unfortunately, Bartlett fails to mention this tremendous potential benefit even in passing.
Prior “Measurement” of the FairTax Rate
As for Gale’s analysis, he developed precisely the right framework to measure the FairTax rate. Indeed, it’s the one I and my co-authors used. He simply made some mistakes with respect to the numbers he entered into his formula. There is also a difference in the base year used in the calculations. But even so, Gale’s FairTax rate (at least the one that comes closest to ours in defining the base) is not much different from ours.
VAT vs. the FairTax
- Laurence J. Kotlikoff
(1) Bartlett’s views are, in addition, getting increased attention from the media. See, for example, Jonathan Weisman, “Criticism Aside, ‘FairTax’ Boosts Huckabee Campaign,” Washington Post, December 28, 2007, p. A04, available at http://www.washingtonpost.com/wp-dyn/content/article/2007/12/27/AR2007122702155.html. See also Brian Mooney, “In spotlight, ‘fair taxers’ push cause," Boston Globe, January 1, 2008, available at http://www.boston.com/business/taxes/articles/2008/01/01/in_spotlight_fair_taxers_push_cause/.
(2) Bartlett quotes Neal Boortz and John Linder on page 1245 to support his contention that FairTax advocates are assuming pre-tax wages won’t fall and consumer prices won’t rise when the FairTax is introduced. But the quotation says nothing of the kind. It simply says that workers will “… be receiving 100 percent of every paycheck…”, which is, indeed, the case regardless of the size of the paycheck (the level of pre-tax wages).
(3) See Paul Bachman, Jonathan Haughton, Laurence J. Kotlikoff, Alfonso Sanchez-Penalver and David G. Tuerck, “Taxing Sales Under the FairTax: What Rate Works?” TaxNotes, November 13, 2006, pp. 663-82.
(4) See “Does It Pay to Work and Save at the Margin?” (with David Rapson), Tax Policy and the Economy, NBER volume, 2008.
(5) See, for example, Poterba, James "Lifetime Incidence and the Distributional Burden of Excise Taxes," American Economic Review 79 (May 1989), 325-330.
(6) In the case that a household gives or bequeaths some of its resources to its children, the C includes the consumption of the children that is finance by the R. Note that the timing of consumption doesn’t matter to its present value tax liability assuming no change over time in the tax rate. While it’s true that one can try to avoid the FairTax by not spending and handing heirs a larger gift or bequest one that includes income earned on the assets the FairTax will be assessed on that income when it is spent. In present value the tax liability is the same. Namely the FairTax rate times R.
(7) See Laurence Kotlikoff and David Rapson, “Comparing Average and Marginal Tax Rates Under the FairTax and the Current System of Federal Taxation,” mimeo, Boston University, 2006. http://people.bu.edu/kotlikof/Comparing%20Average%20and%20Marginal%20Tax%20Rates%2010-17-06.pdf .
(9) See Auerbach, Alan J. and Laurence J. Kotlikoff, Dynamic Fiscal Policy, Cambridge, England: Cambridge University Press, 1987.
(11) Gentry, William M. and R. Glenn Hubbard, “Distributional Implications of Introducing a Broad-Based Consumption Tax,” NBER working paper no. 5832, November 1996.
(12) It took me and my co-authors a very long time to sort out and we only did so with the help of a brilliant tax attorney and consultant David Barton.
(14) This estimate is based on an email exchange with Brad Multon, who is in charge of the BEA office producing national income accounts.