CRS ANALYZES EFFECTS OF CHANGES IN BUSINESS EXPENSE DEDUCTIONS.

                          Jane G. Gravelle

The President's tax proposals include a provision that disallows 50 percent of the cost of business meals and entertainment as a deduction from tax liability. Under current law, 80 percent of these expenditures can be deducted. Questions have been raised as to the effect of this provision on employment and in particular on employment in the restaurant and entertainment industries.

The analysis in this paper suggests that there will be virtually no effect on the restaurant or entertainment industries from the tax changes in the President's proposals. The principal reason is that the "price" effects, or change in relative cost, are extremely small, because higher tax rates (which encourage business meals and entertainment expenditures) offset the limit on deductions. The overall calculated effects encourage these expenditures, although the percentage decrease in cost is negligible (less than one half of one percent). In addition, only a fraction of consumption is associated with business use; this fraction is about one quarter for eating and drinking places and about three percent for amusements.

Even if the price effects were larger, however, the experience of the 1986 legislation suggests that behavioral responses to changes in tax prices are small in any case. Although increases in the cost of these expenditures were quite large, output of eating and drinking places actually grew faster than overall output following 1986. This experience suggests that there could not have been a strong behavioral response to the tax changes.

Finally, even if there were significant price effects and behavioral responses, these industries could probably absorb some contraction in demand without the loss of employment of current workers due to normal growth and attrition, and turnover. There is no reason, however, to expect that there would be an overall effect on employment in the country as long as the overall fiscal stance of the tax proposals is fixed, and thus no net loss in overall jobs available.

                              CONTENTS

The President's tax proposals include a provision that disallows 50 percent of the cost of business meals and entertainment as a deduction from tax liability. Under current law, 80 percent of these expenditures can be deducted. Questions have been raised as to the effect of this provision on employment and in particular on employment in the restaurant and entertainment industries.

The analysis presented below suggests that there will be virtually no effect on the restaurant or entertainment industries from the tax changes in the President's proposals. The principal reason is that the price effects are extremely small, because higher tax rates also contained in the proposals (which encourage business meals and entertainment expenditures) offset the limit on deductions. The overall calculated effects encourage these expenditures, although in a negligible amount. In addition, only a fraction of consumption is associated with business use; this fraction is extremely small in the case of entertainment.

Even if the price effects were larger, however, the experience of the 1986 legislation suggests that behavioral responses to changes in tax prices are small as well. Thus, there would be unlikely to be a significant effect even if tax rates were not increasing.

Finally, even if there were significant price effects and behavioral responses, the industry could probably absorb some contraction in demand without the loss of employment of current workers due to normal growth and attrition. (There may, of course, be variations across individual firms). There is no reason to expect that there would be an overall effect on employment in the country as long as the overall fiscal stance of the tax proposals is fixed.

The following analysis considers price effects of the current proposals, general behavioral responses, and overall employment consequences in turn.

A CALCULATION OF THE PRICE EFFECT

In order to calculate the overall price effect, one needs both the relative price change from the tax treatment of business meals and entertainment and the share that these business uses constitute of total industry output.

PRICE EFFECTS FOR BUSINESS EXPENDITURES

In the following two sections, the concept of tax price is first presented, and then the tax price is calculated.

The Concept of Tax Price

How do taxes affect the "price" firms face for business meals and entertainment? Assuming that businesses are maximizing profits, expenditures on meals and entertainment can be thought of largely as non-taxed consumption.

In some cases, these deductions represent abuses of the tax rules. Perhaps the best illustration of this phenomenon is the story told many years ago by Philip Stern in his book The Great Treasury Raid. ¹ Mr. Stern described a New York executive who boasted that he had not paid for his lunch in thirty-one years. He belonged to a group of business luncheon companions who took turns charging the lunch to their firms. Although this example is probably extreme, the line between personal expense and business expense can be a very fluid one for partnerships, proprietorships, and closely held businesses -- and even for corporate executives who work in larger firms.

Even in larger firms where participants are employees rather than owners or top executives, economic theory still leads to viewing these expenditures primarily as non-taxed consumption.

To explain this point, consider two firms selling the same product. Firm A spends some resources on business meals and entertainment for its employees and customers, while its competitor, firm B, forgoes such expenditures. Firm B can sell its products for a lower cost and can pay its employees higher wages because it does not incur the additional expense. Employees at firm A are willing to accept lower salaries because of the non-monetary compensation. Customers of firm A are presumably willing to pay a higher price because their employees receive some benefit from the value of entertainment and they can in turn pay lower wages to their employees. At the margin, the expenditure on meals and entertainment should just equal the value to customers and employees.

Of course, one response to this perception is that such expenditures are a crucial aspect of doing business, standard across all firms. If these expenditures are, however, so vital, it is unlikely that a firm will be influenced by tax treatment in the first place. Unless these expenditures can be seen as a substitute for other costs (wages in the two firms) there is no reason to expect that taxes have any influence. ²

This consumption aspect is important because it means that the tax rates of the individual employees are relevant to the firm's decision to make such expenditures.

With this background in mind, and thinking for the moment of a single individual tax rate, denoted as t, the price for a partner or proprietor of an expenditure on business meals or entertainment is (1-t). That is, if he spends a dollar on these purposes, he gives up a dollar of income that would be subject to tax. If there is a partial deduction, the price is (1-bt), where b is the share deducted.

The relative price for an employee of a firm or a target individual is (1-bt[sub f])(1-t)/(1-[sub tf]) where [sub tf] is the tax rate of the firm. This calculation reflects the fact that gross income is subject to an overall firm tax (hence the term in the denominator).

The Measurement of Tax Price

To determine the price effects, one must measure old and new values of the tax rates; b simply changes from 0.8 to 0.5. Individual tax rates will increase because of the rate increases for higher income individuals and because of the additional HI tax proposed. ³

Based on overall product in the economy (excluding government and household production) corporations account for 68 percent of expenditures if their expenditures are proportional to output. The calculations should, however, take account of the fact that a lot of corporate product is produced in firms that do not pay taxes. The latest Statistics of Income data (1989) indicate that only 77 percent of business receipts are for corporations that have net income. A further adjustment should be made for returns that have net income, but do not pay income tax because of statutory deductions or credits. Based on the ratio of returns with net income that are taxable, a further reduction is in order, with the result that only 57 percent of corporate expenditures are assumed to occur in firms with tax liability. When the corporate tax rate is zero, the price of consumption is simply the individual tax rate (1-t).

Measuring the individual tax rate (and the changes in the tax proposals) is difficult. Most of these individuals would be assumed to have relatively high incomes. The best distribution for measuring average marginal tax rates appears to be partnership and subchapter S income for those firms that had net income. (Partnerships with losses are more likely to be passive investment instruments). Based on this distribution the average marginal tax rate of individuals (including payroll and income taxes) rises in the President's proposal from 0.306 to 0.375.

The corporate tax rate is estimated at 0.304. This rate assumes that 20 percent of firm income is subject to the alternative minimum tax (at a rate of 20 percent). ⁴ The remaining income is taxed at 33 percent, slightly below the statutory rate of 34 percent due to graduated corporate rates. The President has proposed an increase in the top rate from 34 percent to 36 percent which would increase the weighted rate to 0.318 percent. The Ways and Means Committee has adopted an increase to 35 percent, which would increase the rate to 0.311 percent. There is some possibility that the rate will not be increased.

Calculations are performed for all three possibilities. The price changes are weighted for the shares that are noncorporate (32 percent), corporate taxable (39 percent), and corporate nontaxable.

In all three cases the price actually DECLINES, although the percentage changes in price are negligible: -0.2 percent for a 36 percent tax rate, -0.4 percent for a 35 percent tax rate, and -0.7 percent for a 34 percent tax rate. A price decline means that there will be more, not less, spending on business meals. Effectively, the individual rate increases, which lower the price of business meals and entertainment relative to other commodities, have roughly offset the restrictions in deductibility that increase the relative price.

BUSINESS SHARE AND OVERALL PRICE EFFECT

The price change calculated above applies only to the portion of demand attributable to business use and the overall price effect would be multiplied by the business share.

The input/output accounts show that for eating and drinking places, intermediate use accounts for 27.7 percent of total demand. ⁵ This intermediate use should be reduced by the own industry use, and the use by government and non-profit institutions to obtain business use. This adjustment is small, and reduces the total to 25.7 percent.

Although there is substantial intermediate demand in the amusements sector of the economy, most of that intermediate use is simply use in own industry or in radio and television broadcasting. (Motion picture production and distribution accounts for a large fraction of this industry output.) Overall intermediate use after excluding these industries and the non-profit industries is only 2.9 percent. ⁶

Taking the estimate derived above for the 35 percent rate (-0.4 percent) and multiplying by the business share leads to an overall price reduction of -0.1 percent, that is, one tenth of one percent, for eating and drinking places. For entertainment, the price effect is -0.01 percent, or 1/100 of one percent.

BEHAVIORAL RESPONSES

Although the behavioral responses become moot with such small price changes unless they are enormous, it is nevertheless of some importance to touch briefly on this issue, since there is some room for error in the tax calculations presented above.

Only one econometric study of the behavioral response to these expenditures by business has been done. ⁷ This study examined the relationship between expenditures on travel, entertainment, and gifts and tax rates in a sample proprietorships. Because of the complexity of the model and data, one cannot obtain a simple elasticity from the statistical results. There are some simulations that encompass current price levels and estimated price changes that suggest elasticities in the neighborhood of 1.5 for travel and entertainment.

If these elasticities are applied to the numbers derived above, the overall estimate of the aggregate demand effect for eating and drinking places is 1.5 times 0.257 times -0.004, or slightly above 15/100 of one percent. For amusements, the estimate is 1.5 times 0.029 times -0.004, or 16/1000 of one percent.

It seems likely that this elasticity is probably too high, however, based on the experience with the Tax Reform Act of 1986. In 1986, the price of business entertainment and meals did change substantially. First, the change disallowed 20 percent of deductions, where before a full deduction had been allowed. This restriction in deductibility was magnified, rather than offset, because marginal tax rates fell in 1986. For example, consider the top marginal tax rate of 50 percent, which fell to 28 percent in 1986. In that case, the tax price of business meals rose by 52 percent. Multiplying this amount by the share of business use would lead to a weighted price change of 13 percent. If the elasticity of 1.5 is applied to this change, the overall contraction in demand would be 20 percent. Even if we assume the tax rate began at 40 percent rather than 50 percent, the price change is 28 percent, for an overall drop in demand of 11 percent.

One must be extremely cautious about drawing conclusions from any given episode, since there are many factors that can affect the fortunes of an industry at any given time. Nevertheless, evidence on the behavior of the industry after 1986 is worth examining.

Despite the large relative price effects on business meals and entertainment expenditures, business receipts in eating and drinking establishments actually grew faster than overall output in the years following the 1986 act. ⁸ While overall output grew at a 6.3 percent rate from 1986 to 1987, receipts from eating and drinking places grew by 8.7 percent. In 1988/87 the rates were respectively 7.9 percent and 10.0 percent. In 1989/88, the rates were 7.2 percent and 5.3 percent. The average annual growth rates in the three year period were 7.1 percent (overall) and 8.0 percent (eating and drinking places).

While many factors influence growth rates, it seems highly unlikely that offsetting factors could have countermanded a response as large as ten to twenty percent. This discussion suggests that the elasticity is probably much, much lower than that estimated by Clotfelter, probably considerably below one.

It is perhaps not surprising that subsequent evidence does not support the elasticity estimated in the Clotfelter study. Cross section studies in other areas (such as capital gains and charitable contributions) have produced widely varying estimates and have been the subjects of considerable controversy. Cross-section studies of effects of tax prices on expenditures are difficult to perform, in part because tax rates are correlated with income. Moreover there is no method to estimate the responses of corporations from such data, or to estimate the response of target firms. Despite the uncertainties of drawing conclusions from a single episode, the best evidence on the behavioral response may well be drawn the apparent lack of response to the 1986 tax changes.

This analysis of the behavioral response of the 1986 Act reinforces our conclusion that there will be virtually no effect of the tax proposals on demand in the restaurant and entertainment industries.

AGGREGATE EFFECTS AND JOB OPPORTUNITIES

The effects on demand in the restaurant and entertainment industries described above arise from a relative price effect. Assuming that the fiscal stance of the government is otherwise unchanged (as already set by reconciliation agreements) the overall level of employment in the economy will be largely unaffected. Even if demand were to fall in the restaurant and entertainment industries, it will rise elsewhere.

Because of this overall neutral stance of assumed fiscal position, if a contraction in job availability in the restaurant and entertainment industries did occur, it would be offset by an expansion in other job opportunities. The dislocation that might be of concern, therefore, is largely associated with those already employed in the industry. If there is a substantial amount of natural growth and natural attrition, a contraction can be more easily absorbed without affecting existing workers.

Unfortunately, we were unable to locate information on attrition rates in the affected industries. Some illustrations may, however, be in order. Consider first an industry where all employees enter and remain in the industry throughout their careers. In any one year, the percentage of new positions will be the normal percentage increase in employment in the industry plus attrition. Excluding recent recession years, overall employment growth in the economy as a whole (and growth in the retail trade sector) has ranged from 2.5 percent to slightly over 3 percent. Assuming a work span of forty years, slightly less than 1.5 percent of workers should retire in any year. Thus, overall new positions that would normally become available would be in the neighborhood of 4 percent of the work force, even if no one left the industry during their working career.

Attrition would be much larger in an industry characterized by rapid turnover. On the whole, one would expect turnover rates to be relatively large in the restaurant industry, at least in the less skilled jobs. Of course, highly skilled and compensated employees or owners would be less likely to leave the industry, but they are also more able to absorb some small decrease in demand through slightly lower compensation or profits.

The general point of this discussion is that even if there were a contraction in demand for the services of the restaurant industry, a small effect would unlikely to cause disruption in the employment of existing workers. The job opportunities for new entrants into the work force or for those working in other industries would be unlikely to be affected by even a significant price change and reduction in demand.

² One such instance would be meals for traveling salesmen. Even in this case, however, the firm could choose to compensate its salesman through higher wages rather than an expense account. Of course, the salesman has to eat in any case.

³ In calculating tax rates, social security taxes are included along with income taxes, allowing for the interaction because part of these taxes are excluded from the income base. The HI tax can clearly be thought of as an additional tax since future medicare coverage is not conditioned on the dollar amount of contributions. The regular social security tax is more problematic; including it slightly increases the price effect.

⁴ Gerardi, Geraldine, Hudson Milner, and Gerald Silverstein. Temporal Aspects of the Corporate Minimum Tax: Results of Corporate Panel Data for 1987-1990. Manuscript, 1992.

⁵ The input output accounts are for the year 1987 and are published in the April 1992 issue of the Survey of Current Business.

⁶ Of course, this percentage varies according to the type of amusement. For example, industry representatives testifying in 1986 indicated that 43 percent of major league baseball tickets and 50 percent of hockey tickets are purchased by businesses. Of course, professional sports also receive revenues from television and radio: the witnesses reported during questioning that 58 percent of organized baseball revenues were from broadcasting rights. Therefore, business ticket sales would be only about 18 percent of receipts. See Comprehensive Tax Reform, Hearings, Committee on Ways and Means Part 4 of 9, June and July, 1985.

⁷ Clotfelter, Charles T. Tax Induced Distortions and the Business-Pleasure Borderline: The Case of Travel and Entertainment. American Economic Review v. 73, December 1983, pp. 1053-1065.

⁸ Business receipts are the sum of amounts from corporations, proprietorships and partnerships in the Internal Revenue Service Statistics of Income data.