THE PERSONAL INCOME TAX AS AN AUTOMATIC

STABILIZER

E. Cary Brown, professor of economics, Massachusetts Institute of Technology The purpose of this note is to bring to the committee's attention certain recent empirical findings regarding the automatic sensitivity of the personal income tax and to consider some of their implications for tax policy.

1. SENSITIVITY OF EXISTING PERSONAL INCOME TAX

A study made by Pechman,1 and brought up to date by Cohen,2 examined into the response of the personal income-tax base and of effective tax rates to changes in economic activity. It was found that changes in taxable income were sensitively related to changes in total income (around 60 percent of potential taxable income showed up in the tax base), but that the average tax rate applicable to this base changed relatively little. In combination these two types of behavior indicated a change in tax yield under, say, 1953 definitions of income and tax rates of around 14 percent of personal income in the period 1949-53.

These are important findings, although they are based on only a limited number of observations in the postwar period. They are also limited to a particular set of personal exemptions and do not indicate how the income sensitivity of the personal income tax would be affected by modifications in the exemption structure. In an attempt to get more general results, Dr. Kruizenga and I took a longer period of time-1929 to 1953-to derive an explanation of the income-tax base with as few variables as possible, one of which was the average level of exemptions. A close statistical relationship was found between the ratio of taxable to personal income, the dependent variable, and the two independent variables of the average per capita level of exemptions and per capita personal income. There was a good deal of experimenting with the precise character of relationship. While

1 Joseph A. Pechman, "Yield of the Individual Income Tax During a Recession," National Tax Journal, VII (March 1954), later reproduced with minor corrections in Conference of the Universities-National Bureau Committee for Economic Research, "Policies To Combat Depression" (Princeton: Princeton University Press, 1956).

2 Leo Cohen, "An Empirical Measurement of the Built-In Flexibility of the Individual Income Tax," American Economic Review, Papers and Proceedings, XLIX (May 1959). E. Cary Brown and Richard J. Kruizenga, "Income Sensitivity of a Simple Personal Income Tax," Review of Economics and Statistics, XLI (August 1959).

2357

there was no clearcut choice on statistical grounds, analytic considerations favored the following: *

4

where Z taxable income, Y=personal income, P=population, and E=aggregate amount of exemption for the entire population. As one would expect, the ratio varies directly with the level of per capita income and inversely with per capita exemptions. Income operates more strongly on this ratio than do exemptions.

Theoretically, the definition of income and the deductions allowed could modify these results. But changes in the definition of income over as long a period as from 1929 to the present seemed to have had relatively little effect. Our attempt to use another variable, like deductions, for example, failed to improve the statistical relationship. The reason seems to lie in the fact that potential deductions, such as medical expenses, contributions, taxes, interest, and the like, correlate closely with personal income, and, therefore, they offer no additional explanation of the income sensitivity of the tax base. Most of the tax deductions, then, seem not to change the yield response of the personal income tax if rate adjustment is made to compensate for the smaller tax base. In other words, they act largely as a qualification on the average tax rate.

5

Given the income sensitivity of the tax base, the sensitivity of the tax yield will depend on the rate structure. But the progressive structure in this country places major reliance on the starting or first bracket rate, especially in view of income splitting in the family unit. Most findings emphasize the high degree of stability in the average tax rate in the postwar period under a given rate structure. They must be regarded as tentative, however, since they are supported by as yet fairly scanty data. Until the large statistical gap-allocating for many years taxable income to various taxable-income brackets-is filled, we cannot be very sure of this conclusion. So far such information is limited to tabulations for recent years and to estimates for the postwar period.

I was hopeful that the amount of taxable income in each bracket could be explained by a technique similar to the one we had used to explain aggregate taxable income. But when each bracket was handled in the same way as an exemption, the results were unsatisfactory for the few available postwar observations and estimates. Odd things are happening to income in some of these brackets. In some years income in some brackets has decreased despite the fact that aggregate taxable income has increased. This may be the result of a long-run tendency and not a cyclical factor. But until we know more, at the moment we can only say that economic fluctuations of postwar size and type did not modify the distribution of taxable income sufficiently to result in significant changes in the average tax rate. This means

The linear arithmetic relationship, of equally good quality statistically, was: E .1669+.0002 -.0016. But because this showed no limit to the fraction of income

Y

taxable, the text equation was favored.

5 See Pechman and Cohen, notes 1 and 2 above.

C. Harry Kahn has undertaken the estimates for the National Bureau of Economic Research.

that the yield response of the progressive personal income-tax structure was essentially no greater than that of a proportional tax structure of equal initial yield.

This yield response of the personal income tax can, therefore, be raised or lowered primarily by changes in the basic rate or in the level of exemptions. Tinkering with the progressive rate structure will be of relatively little consequence for its yield sensitivity. Our study found that yield sensitivity was about 12 percent of personal income under 1953 rates and exemptions, and around 13 percent under the present structure. This may seem like a rather small amount. Yet, if there were no offsets between gross national product and personal income, a multiplier of 4 could be cut by such an offset to a multiplier of 3-an impressive performance. Naturally, as more and more offsets are interposed between GNP and personal income, the absolute stabilizing significance of the personal income tax is reduced. This should not lead us to a denigration of the relative contribution it can make, however.

In automatically coping with inflation, however, heavier demands are made upon stabilizing instruments. In order to provide a drag on inflation, tax yields must rise in real terms, or, put in another way, money yields must increase more than prices rise. In the 1946-48 period, when prices were rising rapidly and the Federal personal income-tax structure remained essentially unchanged, money tax yields apparently just kept pace with the Consumer Price Index (table 1). If our fiscal instruments must be prepared to cope effectively with inflation as a major problem of the future, this behavior is a cause of much concern. Serious consideration should be given, therefore, to techniques that would improve this performance. TABLE 1.-Federal personal tax receipts, 1946-48

Source: (1) Current prices: U.S. Department of Commerce, "U.S. Income and Output" (Washington: GPO, 1959) table III-3. (2) Constant prices: Current-price series divided by Consumer Price Index based on 1935-39 prices.

II. IMPROVING THE SENSITIVITY OF THE PERSONAL INCOME TAX

1. Changing the rate and exemption structure

Without altering the basic structure of the personal income tax, yield sensitivity may be enhanced by raising exemptions and compensating for the revenue loss by increasing the initial tax rate. The

following approximate consequence on the yield sensitivity of the tax at 1953 income levels of such a modification would be as follows:

These changes in yield sensitivity are not as marked as one would expect. An implication of these data are that, given the existing broadly based tax and bracket structure, very little can be done in a moderate way to enhance yield sensitivity. Or, put in another way, it is the average yield of taxes as heavy as ours that shapes the yield response, not the marginal adjustments that are feasible. 2. Changing the bracket structure

One of the major impediments to a more sensitive response of the income tax is attributable to the present bracket structure. The average tax rate seems largely unaffected by changes in income, because the bracket distribution of taxable income seems relatively stable. Information is lacking on whether a different bracket structure would make the distribution of taxable income react more sensitively to income changes. But it seems likely that finer bracket graduations at the bottom of the taxable-income scale would make effective tax rates somewhat more responsive to income changes. Perhaps the first two brackets should be sliced into $500 sizes, as has been proposed from time to time. It seems worth a try, at modest complication. 3. Automatic rate modifications

It may even be desirable, if more automatic stability is seriously desired, to take a more radical step. There would be much merit in combining a proposal for automatic rate adjustment made by a group of university professors some years back with the interesting proposals of last year for temporary interim reduction (or increase) of withholding rates, with a proper adjustment of liabilities in the subsequent period. These adjustments would revolve around a permanent rate structure which would be altered only on the basis of long-range considerations.

What little we know about the complexities of a dynamic economic system suggests that early action on a small scale to correct maladjustments leads to smaller gyrations of the economy than late, massive action. While discretionary action can, in theory, accomplish this task, postwar fiscal history suggests that appropriate action will generally not be taken promptly enough for such a light touch. Almost invariably discretionary policy gets caught up in a long debatewhether to adjust the personal income tax or some other tax, whether to change rates or exemptions, and whether the time is ripe for such

A statement by the National Planning Association Conference of University Economists, 1949, reprinted in American Economic Association, "Readings in Fiscal Policy" (Homewood: Richard D. Irwin, Inc., 1955).