Statement of John G. Wilkins, Managing Director,
 Barcroft Consulting Group, on behalf of National Retail Federation

Testimony Before the House Committee on Ways and Means

Hearing on Fundamental Tax Reform

April 11, 2000

Mr. Chairman and Members of the Committee:

I am managing director of the Barcroft Consulting Group and I am here on behalf of the National Retail Federation. My statement reports on the findings of a study undertaken by PricewaterhouseCoopers ("PWC") for the National Retail Federation Foundation. I was principal author of that study, which examines the economic impact of substituting a national retail sales tax ("NRST") for the federal income tax.

Introduction

Nearly nine years ago, I testified before this committee on the issue of international competitiveness and the role of income taxation. In that statement, I noted that the United States relied less on consumption and more on income taxes to finance government than virtually all of the other industrialized nations -- even when state retail sales taxes are included in the equation. Although our income tax structure is by no means a pure income tax -- having some elements of a consumption tax mixed in with an income tax -- I nonetheless observed that a greater reliance on taxing consumption and less reliance on taxing income, would be healthy for the economy by spurring savings and helping keep income tax rates low. That was 1991 and this is 2000. And I continue to hold that view.

While I applaud the desire of the chairman to replace the current tax code with something that is simpler and fairer, I do not, however, subscribe to abandoning an income tax altogether and replacing it with a national sales tax. As our study shows, this could have very harmful short-term and mid-term economic results. In light of the remarkable economic achievements of the past decade, it would be foolish to simply get rid of a tax structure under which the economy is flourishing and replace it with an untried system with uncertain economic consequences.

The PWC study, largely completed last year when I was director of PricewaterhouseCoopers' national economic consulting group, focuses on the economic impact of replacing the income tax with a national sales tax similar to the Fair Tax proposal. The proposal examined was H.R. 1467 rather than H.R. 2525, which had not been introduced at the time of the study. One significant difference is that H.R. 1467 does not repeal federal payroll taxes while the Fair Tax, H.R. 2525, does and consequently requires a higher tax rate.

What follows is a discussion of (1) why the PWC dynamic model is particularly capable of recognizing the effects of a national sales tax on the economy in general and on retailers in particular; (2) why the proposed national sales tax rate could eventually be much higher than proponents advertise due to likely exclusions, the need for budget neutrality, and transition rules; and (3) what impact this kind of tax would have on short-term and long-term economic growth, consumption, corporate profits, employment, and trade.

The Dynamic Estimating Model

In the early 1990s, Coopers & Lybrand undertook to develop an economic model that was missing from the arsenal of tax analysis tools available to government economists and others concerned with the potential economic impact of fundamental tax reforms. The traditional microsimulation models used by the staff of the Joint Committee on Taxation ("JCT") and the Treasury Department's Office of Tax Analysis for revenue estimating purposes are based on large samples of individual and corporate taxpayers and consequently are ideal for analyzing the impacts of tax law changes on the taxpayer population. These models are not, however, capable of analyzing the impact of fundamental tax reforms on the economy. For such analysis economists turn to macroeconomic models. Unfortunately, macroeconomic models are rarely designed for tax analysis: they frequently have only two or three producing sectors; some have only a single household or a handful of households to represent the entire population of taxpayers and consumers; and most examine the economy only when it is in equilibrium, with labor and other resources artificially restricted to be at full employment.

The PWC model was designed specifically for analysis of tax reforms, such as the replacement of the income tax with a NRST. In order to retain the benefits of the microsimulation models used by the JCT and the Treasury, the PWC model has three prongs, incorporating two microsimulation models with a macroeconomic forecasting model. The two microsimulation models are an individual income tax model with 100,000 separate tax return records and a corporation income tax model with 15,000 synthetic tax return records. The third prong of the overall model is a macroeconomic forecasting model that provides year-by-year short-term forecasts of the economy as well as mid-term and longer-term forecasts. Importantly, the PWC model has the following traits that distinguish it from most other models:

· The PWC model contains an open economy, allowing changes in foreign trade, cross-border investment flows and exchange rate adjustments.
Many models are restricted to closed economies that ignore the existence of foreign trade and other international transactions.

· The PWC model permits the economy to be in disequilibrium during transition periods.
Most other models artificially force the economy to always be in equilibrium, with no unemployment of labor and other resources -- even during periods of transition from an income tax to a sales tax.

· The PWC model's database includes records for 100,000 households and 15,000 corporations.
Most models include only a handful of households and corporations. Sometimes only one household represents the entire household sector.

· The PWC model tracks 85 different industries, producing 85 different products for intermediate and final sales.
Most models represent the entire producing economy with only two or three industries frequently producing only two or three different classes of goods.

The most important distinguishing characteristic of the PWC model is that it focuses on the near and intermediate term. The key feature is the model's ability to forecast detailed information on the performance of the economy on a year-by-year basis as the economy makes the transition from an income tax structure to a sales tax structure. The PWC model shows that this transition is not nearly as smooth and simple as some sales tax proponents would like. It is this ability of the model to provide short-term transition results on an annual basis that provided the main impetus for its construction.

The PWC model has been used successfully to evaluate a wide range of tax proposals, from the recommendations of the Kemp Commission to the national sales tax. The model was also used to produce dynamic revenue estimates for the January 1997 symposium on dynamic revenue estimating sponsored by the staff of the Joint Committee on Taxation.

Required Tax Rates

Much of the discussion surrounding national sales taxes centers on the required tax rate. This rate depends obviously upon the tax base -- all consumption or some portion of consumption after certain exclusions. It further depends upon the amount of taxes to be replaced on a neutral basis. It also depends upon the degree of compliance -- 100 percent or some lesser fraction as is the case with the income tax. Lastly, it depends upon mitigating provisions such as increased social security benefits for the elderly designed to prevent too much double taxation of their lifetime earnings, family allowances designed to prevent some of the associated redistribution of tax burden and other transition rules designed to lessen short-term economic disruptions.

Tax Exclusive and Tax Inclusive Tax Rates. The discussion of tax rates is confused by the practice of proponents of the NRST to couch the rate in so-called tax inclusive terms. Sales tax rates are usually considered on a tax exclusive basis. Under this normal tax exclusive concept, for example, a $30 tax on a $100 item is considered to represent a 30 percent tax rate. The consumer would pay the retailer $130, $30 of which would be forwarded to the tax authorities.

Proponents of national sales taxes like to measure the tax rate in this example by dividing the $30 tax by the tax inclusive price, which is $130. The tax rate calculated this way would be only 23 percent (30/130). Under the tax inclusive rate concept, confusion is likely to arise when a customer is quoted a 23 percent tax inclusive rate on a $100 purchase and finds the sales clerk asking for $130.00 ($100 plus $30 tax) rather than the expected $123.00 ($100 plus $23 tax). The confusion will be reinforced by the fact that all state sales taxes are always quoted on the normal tax exclusive basis.

Revenue Neutral vs. Budget Neutral Tax Rates. A second concern in determining the appropriate rate for a national sales tax that would replace other existing federal taxes involves the concept of neutrality. Most believe that any replacement tax ought to be neutral: the government should be left as well off under the replacement tax as it is under the current tax structure. Independent of its political appeal, neutrality focuses the spotlight on the economic pluses and minuses that could result from restructuring the tax system as opposed to the consequences of making the government bigger or smaller, which can be done without restructuring the tax system.

There are two ways to identify neutrality (revenue and budget) and two ways of measuring neutrality (static and dynamic). Revenue neutrality means that the replacement tax raises the same revenue as the current tax. Budget neutrality means that the replacement tax leaves the overall budget surplus (or deficit) unchanged. Budget neutrality is the better measure because it recognizes that, by influencing the price of government purchases, interest rates, or transfer payments, for example, tax changes can affect the spending side of the budget as well as the revenue side.

Static measurement of budget neutrality fails to take account of how a replacement tax influences the budget by accelerating or retarding economic growth. Dynamic measurement of budget neutrality corrects this shortcoming by taking into account macroeconomic effects, such as a short-term change in the level of employment, that can affect government spending and revenue. The dynamic measure of budget neutrality is the most meaningful concept for a replacement tax.

These concepts can be illustrated as follows. According to the PWC model, a very broad-base national sales tax, such as the Tauzin-Traficant proposal to replace the current federal personal and corporate income taxes, the federal estate and gift tax, and most federal excise taxes would require a tax exclusive rate of 18.8 percent under the somewhat unrealistic assumption of 100 percent compliance and revenue neutrality. This rate would be a 15.8 percent rate on a tax inclusive basis, reasonably close to the 15 percent claimed by the sponsors. In order to maintain budget neutrality -- even on a static basis -- the rate would have to be raised from 18.8 percent to 24.5 percent. This is because the federal government would want to fully maintain the purchasing power of all transfers payments (social security, welfare, unemployment, etc.) in order to protect the elderly, the poor, and other transferees. In order to do so, these transfer payments must be increased by the amount of the tax on the goods and services they would purchase. Holding recipients of federal government transfer payments harmless in this manner while maintaining real federal spending on goods and services requires a budget-neutral NRST rate of 24.5 percent (or 19.7 percent on a so-called tax-inclusive basis) -- still figured at an unrealistic 100 rate of compliance.

On a dynamic basis, this 24.5 percent budget neutral tax rate could be lowered slightly to 24.1 percent by the tenth year of the tax, thanks to a somewhat improved economy; however, there are still other concerns involving base erosion and compliance that need to be factored in before an ultimate budget neutral tax rate can be determined.

Compliance. Most experts concede that it is difficult to estimate the rate of compliance under a NRST. What is not in dispute, however, is that the compliance will be lower the higher the NRST rate. Moreover, while state sales tax rates appear to have relatively high rates of compliance, these compliance rates are not comparable to the NRST. State sales tax compliance is high because the sales tax rates are relatively low -- typically 4 to 6 percent -- and their tax bases are relatively narrow. Unlike the value added tax, there is an incentive on the part of both consumers and sellers to avoid the tax. Individuals could easily avoid the NRST in a number of ways, such as disguising personal consumption expenditures as business costs that would not be subject to tax.

If tax compliance is no better than the 83 percent overall compliance rate under the federal income tax, the budget neutral tax rate would have to be raised from 24.1 percent to 29.0 percent. If compliance matches compliance rate for proprietors under the federal income tax, then the budget neutral rate would be raised another 5.9 points to 34.9 percent. Keep in mind that the compliance rates under the income tax are strengthened by forced compliance of withheld tax on wages and by numerous checks and balances payer-provided information returns and audits.

Exemptions and Allowances. Many have observed that enactment of a NRST would encourage many special exemptions from the base. Three of the most frequently mentioned are:

· removing consumption of financial sector services entirely from the base, since the taxation of such services is, at best, extremely complicated and would be difficult to administer;

· removing rental housing services and the resale of existing homes from the tax base so as to continue a current-law tax preference and to mitigate problems arising from unequal treatments of owner-occupied and rental housing; and

· removing employer-provided fringe benefits from the tax base so as to vastly simplify the tasks of businesses, which would otherwise be untaxed.

The effect of removing these items from the NRST base would cause the budget neutral tax rate to rise to 37.5 percent with compliance equal to overall income tax compliance and to 45.1 percent with compliance equal to income tax compliance of proprietors.

Effect of State and Local Tax Piggybacking. Once states lose their ability to piggyback their income taxes off the federal income tax, it is anticipated that many would elect to instead piggyback their revenue needs by adding a state sales tax rate to the federal rate. For consumers, this would further boost the overall rate on consumption to 53.6 percent assuming overall income tax compliance and to 64 percent assuming proprietor's income tax compliance. These figures assume the above exemptions and allowances would be established.

The PWC model is not alone in estimating tax rates for the NRST that are considerably higher than proponents frequently cite. A National Bureau of Economic Research study place the tax rate at 27.3 percent if payroll taxes are not included in those taxes to be replaced and 45.4 percent if payroll taxes are also replaced. A Joint Economic Committee ("JEC") study concluded that the NRST rate would have to be at least 32 percent unless imputed items of consumption, like "rent" that the national income accounts assume homeowners pay themselves were also included in the base. Furthermore, if food, medicine, and physician's services were excluded (as is commonplace among many state sales taxes) the rate would have to rise from 32 percent to 49.3 percent. Alternatively, they found that if all services were excluded from the base but food and medicine continued to be taxable, the rate would have to rise to 64.6 percent.

Impact of the NRST on the Economy

In the long run, the economy will perform somewhat better under a NRST. The results of the PWC model show:

· By 2010 real GDP will be higher by $178 billion (1.8 percent) and will remain above baseline throughout the forecast period.

· By 2010 real personal consumption expenditures will be higher by $16.5 billion (0.3 percent) and will remain above baseline throughout the forecast period.

· Throughout the forecast period national private savings is higher than under the baseline thanks to higher personal savings as consumers delay consumption and higher corporate savings as businesses reinvest a large portion of undistributed corporate profits.

These salutary accomplishments are similar in direction to findings of other studies. Only the magnitudes may differ. It is in the short run that the PWC study finds harmful results.

Economic Growth. Gross domestic product ("GDP"), the value of all goods and services produced in the country, would increase in real terms in anticipation of the enactment of the NRST, as consumers speed up purchases they would otherwise make at a later date.

The aftermath of the speedup is a sharp economic downturn in the year the tax becomes effective. Instead of achieving an expected 2.0 percent rate of real growth in 2001, the assumed first year of the tax, the economy would shrink by 1.1 percent. Although the economy would begin to grow again in subsequent years, it would take until the fourth year of the tax for GDP to reach its pre-NRST level. Before the economy fully recovered, the cumulative loss in real GDP (measured in 1992 dollars) would reach $180 billion and annual employment would dip by 1.5 million jobs. By 2010, real GDP under the NRST would be 1.8 percent above the level that the current income tax would have achieved. Given a growth forecast of about 2.3 percent per year, the 1.8 percent of additional GDP represents only a 9-month speedup in economic growth over a ten-year period.

Consumption. Personal consumption expenditures in real terms would be below the current-law baseline until the ninth year after the NRST was introduced. During the 2001-2008 period, the cumulative decline in consumer spending (measured in 1992 dollars) would exceed $500 billion. Over the first five years of the tax, consumption would be 1.5 percent lower on average than it would be under current-law tax; and over the second five years, consumption would be down 0.2 percent from expected levels.

The overall drop in consumption and the subsequent pickup as the economy recovers masks many important details. Consumption changes will vary greatly according to income levels and according to items of consumption. Changes in consumer purchases reflect the fact that the NRST generally shifts the tax burden away from higher income families and toward lower income families. Although the poorest of the poor may be roughly compensated for their loss of the refundable earned income credit by repeal of the payroll taxes, the moderately poor and many in the vast middle class must have higher overall tax burdens in order to balance those with the highest incomes whose consumption taxes would be far smaller than their income taxes. Households with incomes in the bottom fifth of the income scale would have to reduce their purchases of durable goods by 13 percent on average for the second five years after introduction of the NRST and purchases of nondurable goods by 6 percent. In contrast, households with incomes in the top fifth of the income scale would increase their purchase of durables by an average of 2.4 percent and their purchases of nondurables by 0.3 percent for that same period.

Saving and Investment. In nominal terms, the net private saving of U.S. residents and businesses under the NRST would be about 18 percent higher than under current law for the ten-year period, 2001-2010. Government saving is assumed to be virtually unaffected (that is, the NRST is assumed to be a budget-neutral replacement for current-law taxes that would be repealed). For that same period, personal saving would be higher by about 15 percent. The nominal dollar increase in private saving would come about equally from personal saving and corporate saving. Corporations would be expected to finance about half of all induced new investment through their own saving, by retaining approximately 15 percent of the repealed corporate income tax. The remaining 85 percent of the corporate income tax would be distributed to shareholders in the form of increased dividends.

Over the ten-year period, 2001-2010, induced gross private domestic investment would add another 7 percent to the amount of nominal investment under current law. This increase in investment is nearly twice as large as the increase in gross national saving. Consequently, roughly half of all induced investment is foreign-owned investment flowing into the United States in response to lower financing costs and the elimination of the Federal corporate tax on equity income. Although most of the growth in real investment occurs in the business sector, most of the increase in nominal investment can be attributed to the rise in the price of residential investment due to the tax on new construction under the NRST.

Corporate Profits. On average, corporate profits are about 2 percent lower over the ten-year period, 2001-2010; however, there are notable exceptions for certain industries and certain years. In the aggregate, profits return to the level expected under current law by the year 2010, and are expected to improve thereafter.

Employment. Due to the near term decline in consumer spending, private sector jobs and civilian employment are expected to be lower than they would be under current law through the fourth year of the NRST. The near term estimate would indicate a drop of 1.5 million jobs. Thereafter, jobs and employment will pick up. The labor force is expected to expand by about 1 percent as potential second earners and others are lured into the workplace by vastly lower taxes on wages and salaries and entrepreneurial income.

Anticipatory Consumption Speedup. Introduction of a NRST is expected to create a speedup in purchases of goods and services between the time the tax is announced and the time it becomes effective. If the NRST were imposed as of January 1, 2001, a surge in personal consumption of both domestically produced and imported goods and services would occur in 2000. In addition, equipment investment would accelerate to take advantage of depreciation deductions in the year 2000 before the income tax is repealed. Together, these factors produce a temporary drop in personal saving and a temporary rise in the real rate of economic growth in 2000. Real GDP in 2000 is estimated to be 2.8 percent above the level that would have been obtained in the absence of a proposed NRST. Thereafter, real GDP is depressed by $180 billion from 2001 through 2003. By 2004, real GDP has recovered to pre-tax change levels, and remains above the pre-tax change baseline throughout the remainder of the forecast period.

After the speedup in spending in 2000, personal consumption expenditures remain $500 billion lower than they would otherwise be until the year 2009. In other words, during the first eight years of the NRST, consumption would be depressed as families and individuals respond to the tax by saving more and spending less. It is only after disposable income increases sufficiently that consumption picks up enough to pass the pre-tax change baseline level.

Real investment in equipment is down in 2002 from the pre-tax change baseline due to the tax-motivated speedup of investment into 2000. Thereafter, equipment investment is higher as businesses respond to a lowered cost of capital. Real investment in non-residential structures is down in 2003 and 2004, but picks up significantly after 2004. However, real investment in residential structures remains below pre-tax change baseline levels over the entire 2002-2010 period.

Prices. Prices for consumer goods and services quickly rise by the amount of the tax, and then some. The portion of the price increase in excess of the tax is due in part to the higher cost of imports (from the weaker dollar) coupled with the ability of some domestic producers of competing goods to hike their price to that of imports. Consumer prices similarly rise 25 percent -- roughly the nominal rate of sales tax, unadjusted for any exemptions or transition rules -- by 2002 and gradually drop from that peak to a level that remains about 18 percent above the pre-change baseline.

Examined on a year-over-year basis, these price increases generally amount to a large, one-time hike in prices as the NRST is imposed, with some moderation of this increase in the longer run. Due to a weaker dollar, merchandise import prices increase by nearly 4 percent shortly after the NRST is imposed and are 6.5 percent over baseline levels in 2010. Merchandise export prices are also above baseline levels. In 2001 and 2002 they are nearly 3 percent above the baseline. However, due to lower interest rates, which reduce business costs, export prices are only slightly greater than baseline levels for most of the remainder of the forecast period. The overall impact on prices is measured by the change in the GDP deflator, which initially rises 20 percent above the baseline price level before settling back to a 13 percent price rise relative to the baseline.

The notion espoused by some that pre-tax prices would drop some 20-30 percent under a NRST (so that after-tax prices would not rise and may even decline) is a peculiar one. This could only happen if all of the personal income tax, the corporation income tax and payroll taxes are currently embodied in retail prices. Tax incidence -- that is, who actually bears the ultimate tax burden -- is an elusive question that has been the focus of many economic papers, because the answer is not clear. However, the general consensus among economists is that perhaps a portion of the corporate income tax may be passed on to consumers in the form of higher prices, but that the majority is ultimately paid by corporate owners in the form of lower after-tax profits and by employees in the form of lower compensation. Most economists concede that personal income taxes and payroll taxes are ultimately borne by labor and are not passed on to consumers in the form of higher prices.

Nominal Output. In nominal terms, personal consumption expenditures are expected to be above their baseline level by $1,582.9 billion per year on average for the 2006-2010 period. This represents an increase of 18.3 percent over the average that would have occurred in absence of the NRST. Note, because prices would be 18.5 percent higher, on average, for this period, this nominal increase is consistent with a slight real decline in real consumption expenditures during this same period.

Trade. Merchandise exports and imports are both impacted by the NRST. Exports are made relatively cheaper to foreigners because the dollar is somewhat weaker under the NRST. Imports are subject to the NRST and are also more costly for U.S. consumers to buy due to the weaker dollar. As expected, in real terms, exports grow about 4.3 percent over the baseline during the last five years of the forecast period, 2006-2010; and imports drop an average of about 1.6 percent during that same period. In nominal dollar terms, both exports and imports are larger than under current law due to the sharp price increases for imports discussed above. Real net merchandise exports increase by $448 billion (in 1992 dollars) over a ten-year forecast period. However, over the same ten-year period, net merchandise exports in nominal dollars decline by $68 billion relative to the baseline. This nominal merchandise trade deficit helps to finance domestic investment.

Conclusion

If a NRST is enacted, the U.S. economy would lag behind for at least three years and employment would dip by more than one million jobs. Beneficial effects would not be felt for at least five years after adoption. While it is admirable to seek a fairer and simpler tax structure to replace the incredibly complex income tax code, trading an income tax in for a national sales tax is an experiment that could bring serious harm to a flourishing national economy. Uncertain long-run benefits are far insufficient to risk the short-run setbacks in virtually all sectors of the economy.