Democracy in Deficit: The Political Legacy of Lord Keynes

These early Keynesian views took form in the precepts of “functional finance,” the unadulterated Keynesian substitute for the principle of the annually balanced budget.
*59 A government that followed the precepts of functional finance should let the state of its budget be determined wholly by the needs of national macroeconomic management. Budget surpluses would be incurred to curb inflationary pressures, and budget deficits would be created when unemployment appeared. There was no acknowledgment in this early Keynesian discussion that inflation might, in fact, show up before satisfactorily “full” employment is attained, a situation for which the principles of functional finance offered no obvious policy guidance. This aside, however, the focal point of governmental activity under functional finance was the stabilization of prices and employment.

In contrast, the focal point under the balanced-budget principle was the provision of various goods and services through government. The balanced-budget principle, however, was never operationally replaced by an accepted regime of functional finance. Very few economists (and even fewer politicians) were willing to go to the extreme limits suggested by functional finance. Hence, various intermediate principles emerged that attempted to achieve a truce between the implied Keynesian norms of functional finance and the old-fashioned precepts of the balanced budget. In this chapter, we shall describe these various replacements for the balanced-budget norm. Accumulating experience indicates that the constraining impact of these alternative principles has been frail indeed.

Budget Balance over the Cycle

The principle that the government’s budget should be balanced over the course of the business cycle represented an effort to bridge the gulf that appeared to separate the Keynesian precept of functional finance from the classical precept of an annually balanced budget. Budget balance over the cycle appeared to retain an ultimate balancing of revenues and outlays, of costs and benefits, which is the essential feature of the strict budget-balance concept. The old-fashioned rule was to be modified only in the accounting period over which the balancing criterion was to be applied; the period was lengthened from the arbitrary accounting year to that which described the full sequence of the business cycle. It appeared that this would allow for the discretionary use of the budget for purposes of countercyclical macroeconomic management. The modified old-fashioned rule and the new Keynesian use of the budget looked to be fully harmonious.

This apparent reconciliation of these two sets of budgetary principles would possibly have been successful if, in fact, business cycles were somehow known to be regular in their amplitude and time sequence and, in addition, were known to exist exogenous to economic policy. In this situation, and only in this situation, business activity would rise and fall with a regular and predictable rhythm. Budget deficits and surpluses could be applied symmetrically, and the amplitude of the fluctuations could be diminished. The deficits and surpluses would cancel out over the entire cycle, and yet macroeconomic management would smooth out both the peaks and the troughs of economic activity. When business cycles occur in a predictable pattern of regular oscillations and when political constraints on budgetary policy are ignored, the principle of a balanced budget over the cycle seems to bridge the gulf that otherwise would seem to separate the classical and the Keynesian prescriptions.

Even if the underlying oscillations in economic activity were known to be regular and exogenous, this budgetary rule would probably be applied asymmetrically in a democratic setting, for reasons we have already developed in Chapter 7. Once the irregularity and subsequent nonpredictability of cycles are acknowledged, this political bias would become even stronger. However, in a setting in which cyclical swings are
known to be regular, the presence of a cumulative budget deficit over the whole cycle would offer a clear and unambiguous indicator that the balancing rule had been violated. Knowledge that such a criterion might be present would itself act as a constraint on irresponsible budgetary behavior. When cycles are irregular, however, the rule for a cyclically balanced budget and that for functional finance are inconsistent, quite independent of political bias in application. One has to give way to the other. Either the cyclically balanced budget must be pursued at the expense of functional finance, or functional finance must take precedence over the norm of a cyclically balanced budget. The irregularity of cycles undermines the bridge between the two norms and, therefore, negates the possibility that the rule of budget balance over the cycle could serve as an effective substitute for the constitutional constraint of budget balance within the accounting period.

The principle of a balanced budget over the cycle, which acknowledged the necessity for budgetary manipulation to achieve common stabilization objectives, had a relatively short existence in the arsenal of fiscal weapons. A line of argument soon surfaced to suggest that prevailing budgetary-fiscal institutions operated automatically to stabilize the aggregate level of economic activity. This idea of built-in flexibility shifted the focus from the need for discretionary fiscal manipulation to the prospects for institutional adjustments that would ensure automatic fiscal reactions in the direction of desired objectives.

Built-in Flexibility

As our narrative in earlier chapters has shown, one of the first steps on the way to full-fledged Keynesian budgetary management was the recognition that attempts to maintain strict budget balance during periods of economic distress might accentuate depression and, in turn, worsen rather than improve the prospects of budget balance. It was noted that the deficits of the early depression years of the 1930s were due to the simultaneous fall in tax revenues and the increase in relief outlays, and that these deficits themselves were forces leading toward recovery. Before the 1930s, when strictly classical principles were dominant, such “built-in flexibility” was considered to be an undesirable property of a fiscal system. Instability in revenues and outlays was something to be avoided, since this made the task of balancing the budget more difficult. With the deficits of the Great Depression, however, there came to be increasing awareness that such instability was desirable in itself. Macroeconomic considerations began to emerge, and built-in flexibility was adjudged to moderate fluctuations in aggregate economic activity.
*60

Once these side effects of budgetary unbalance were fully recognized to be desirable, a natural extension, and especially to those economists who had accepted the then new Keynesian paradigm, was to propose structural changes in the budget that would increase the macroeconomic adjustments, that would increase the built-in flexibility. That is to say, Keynesian-inspired norms for both the taxing and the spending structure appeared alongside the traditional norms which were derivative from considerations internal to the fiscal system rather than external. Proposals for tax reform came to be evaluated against macroeconomic criteria, for their effects on real income and employment, over and above those familiar criteria of “justice,” “equity,” “convenience,” “certainty,” and others. In some cases, the macroeconomic criteria achieved dominance in the economists’ lexicon. Similar changes occurred in the analysis of spending. Macroeconomic arguments, based on Keynesian economics, came to be advanced for the acceleration of welfare and relief types of budgetary outlay.

In their early phases, however, the policy implications of these arguments for built-in flexibility were confined to structural features on both the revenue and spending sides of the fiscal account. There was nothing here that offered a norm or criterion for the state of budget balance or imbalance. In one sense, the discovery of built-in flexibility, both in its positive and in its normative aspect, was supplemental to the accompanying discussion of alternatives to the balanced-budget rule.

Budget Balance at Full Employment

The incorporation of the structural norms for built-in flexibility provided almost a natural lead-in to an alternative to that of the annually balanced budget. This alternative is the principle of budget balance at full employment, or, as it has come to be known in the 1960s and 1970s, budget balance at high employment. Once it came to be recognized that the budget deficits that emerged passively during periods of economic recession exerted desired pressures toward recovery, it was but a small step to the recognition that the recovery itself exerted a simultaneous effect on the size of the deficits. From this, it followed that the expansiveness or the restrictiveness of fiscal policy might be measured, not by looking at the generation of surpluses or deficits in any current-period setting, but by predictions of the effects that would emerge in a hypothetically postulated setting of full employment.
*61 In the Keynesian economic environment, if the economy is operating at full employment, there is no Keynesian-inspired reason for departure from budget balance. The norm should become, therefore, that of setting revenues and expenditures such that the two sides of the budget will come into balance if the full-employment level of income should be achieved. Under the operation of this norm, and because of the built-in flexibility, surpluses would automatically be created in all settings in which demand pressures are excessive. This alternative rule seems to incorporate both the norms for Keynesian budgetary management and the classical principle of balancing revenues against spending. In particular, the fiscal choice process, in the Wicksellian emphasis, seems to be balanced here in that new programs of spending proposed would be weighed against the tax costs of these programs, at full-employment income.

The analytical basis of the budget-balance-at-full-employment rule is starkly simple. Assume, for purposes of argument here, that federal budget outlays are currently at an annual rate of $400 billion, and that tax collections are at the annual rate of $350 billion. In current account terms, there is a budget deficit of $50 billion on an annualized basis. Suppose, however, that the economic situation is also characterized by an unemployment rate of 7 percent, clearly adjudged to be a less-than-full-employment rate.

Now assume further that it is predicted that a fall in the unemployment rate (i.e., an increase in real income and employment) acts both to increase tax collections (because of the increase in income) and to reduce somewhat the rate of federal spending (on unemployment compensation, on food stamps, on relief payments). Let us say that it is predicted that a 2-percent shift in the unemployment rate, to 5 percent, will increase tax revenues to $390 billion and will cut spending to $390 billion. The $50 billion budget deficit, now observed, would vanish and the budget would be in balance. But what if “full employment” is defined to be 4 percent? The budget would have reached balance at the 5-percent unemployment rate. If unemployment should be reduced to 4 percent, further operation of the built-in stabilizers would generate total revenues at a predicted $410 billion rate, and outlays would be predicted to fall to $385 billion. Hence, at the defined full-employment rate of 4 percent, the federal budget would be in
surplus, not in deficit. The initially observed $50 billion shortfall in revenues behind outlays would actually be indicative of a “full-employment surplus.” Politicians, the public, and the professors might then talk themselves into thinking that a fully responsible fiscal policy would require
increasing the observed deficit.

“Full-employment surplus,” a phrase that became prominent in the Economic Report of the President’s Council of Economic Advisers after 1962, is the difference between anticipated federal revenues and federal government outlays or expenditures that are projected at some arbitrarily designated level of employment and income.
*62 As early as 1947, the Committee for Economic Development proposed that the federal budget should be arranged so that a $3 billion surplus would emerge at an unemployment rate of 4 percent.
*63 This particular projection also yielded a balanced budget at an unemployment rate of 6 percent. In this framework, any given budget is considered to be expansive if the full-employment surplus is negative, and contractionary if this full-employment surplus is positive. The larger the surplus, the more restrictive the budget structure is considered, regardless of the fact that actual budget may be running a substantial deficit. Similarly, the more negative the full-employment surplus, the more expansive the budget is considered, regardless of the actual, contemporaneous relation between taxes and expenditures.

The conception of budget balance at full employment has wide support, and, as a norm, it might appear to be a reasonable compromise between the old-time fiscal religion and the Keynesian precepts. The implied norms for budget balance at full employment seem to meet minimal classical requirements while allowing functional finance to operate. This service to two masters might be pardonable if the principle should be able to achieve its logical promise in practice. But it should be clear that this promise will not likely be met. Politically, what appears to be a vehicle to promote fiscal responsibility may become little more than an excuse for the budgetary license. Quite apart from the political implications, however, the whole conception is deceptive in its unacknowledged dependence on the existence of the presumed economic environment of the 1930s.

We shall initially examine the possible application of this principle independently of the political- or public-choice biases that emerge in exaggerated form under its influence. To do so, we may assume that democratic political pressures exert no influence on fiscal decisions, which are made exclusively by “wise” persons in Washington. This is, of course, a wholly unrealistic setting for fiscal policy choices, but it helpfully allows us to isolate the fundamental economic implications of this proposed alternative rule for budget making.

The rule is deceptive because it tends to conceal the implicit assumption about the state or condition of the economy that it contains. This is that the national economy resembles the economy of the early Keynesian models. In these, as we have previously noted, “full-employment income” is sharply and clearly defined; expansions in total spending exclusively affect real output and employment until this level of income is attained; there is no upward pressure on prices. These naive models have substantially disappeared from sophisticated economic discourse, but it is perhaps a mark of the distance between the realm of ideas and that of practice that these naive models seem to persist beneath the surface of most current discussions of the full-employment surplus.

In earlier chapters, we traced the history of the changes in economists’ attitudes as these naive models were replaced by those that embodied the Phillips-curve sort of trade-off between unemployment and inflation. A specific definition for “full-employment income” does not emerge from the trade-off models, and macroeconomic policy of any kind requires the selection of some point on a curve which is seen as continuous, with no kinks or corners. But what does this do to the rule for budget balance at full employment? The dilemma has, of course, been recognized, and attempts have been made to rescue the efficacy of the widely accepted principle for fiscal “responsibility” by specifying some combined unemployment-inflation targets. These have ranged from the more restrictive norm of “budget balance at that level of income and employment that can be achieved without inflation” to the norm that simply defines an employment target, say 4-percent unemployment, and disregards the potential inflationary consequences.

What is relevant for our purposes is that any such combination of targeted levels of unemployment and inflation reflects an explicit selection of an arbitrary point along an alleged inflation-unemployment trade-off curve. As such, there is no demonstrably unique point that dominates all others, and reasonable persons may differ concerning the relative weights to be assigned to the two conflicting objectives, increased employment and reduced inflation. Assignment of the decisions to “wise” persons will in no way remove the necessity of establishing some weights.

Economists in increasing numbers have gone beyond even the simplistic trade-off models, however, and it is now widely acknowledged that a Phillips curve, even if one exists, will shift through time as expectations are modified. What does this do for the potential applicability of the “budget balance at full employment” rule? Suppose, for example, that a combination of a 4-percent unemployment rate and a 4-percent rate of inflation is selected as the objective for policy, as the economic setting for which an attempt would be made to bring federal government revenues and outlays into balance. Suppose, further, that this combination does, in fact, describe a possible position for the economy at the time the target is selected, say, in 1977. Let us say that budgetary plans for 1978 and 1979 are then made on the basis of this target, and budgetary authorities rearrange institutions of taxing and spending to this end (such decisions are now made, we assume, by the bureaucrats appointed by the wise persons). But this whole procedure, to be at all workable, depends on the stability of the presumed Phillips curve, on the attainability of the 4-4 points. Suppose, however, that, by the time the budgetary adjustments take effect, by 1979, the Phillips curve has shifted, and the rate of inflation that will emerge at a 4-percent unemployment rate is 6 percent, rather than the 4 percent incorporated in the plans. The emptiness of the proposed norm for budgetary management becomes apparent.
*64

At the extreme other end of the spectrum from the early naive Keynesian models in which a full-employment level of income is sharply defined is the model that denies the existence of any trade-off and that embodies a “natural rate” of unemployment, which cannot be permanently modified by shifts in the rate of aggregate spending.
*65 If this model should, in fact, describe empirical reality, while, at the same time, the “budget balance at full employment” norm is pursued, either in its pristine form or in some Phillips-curve variant, the avenue is opened for a regime of continuous budget deficits which would be wholly ineffective in achieving the macroeconomic objectives for which they are created.

To this point, we have ignored what is perhaps the most important limitation on the operation of the alternative budgetary rule, which calls for balance at full-employment income. This is the extreme vulnerability of the rule in a political setting where fiscal decisions are made, not by wise persons immune from constituency pressures, but by ordinary politicians who are responsive to demands of the voting public. The directional biases toward budgetary expansion and toward deficit financing that we have discussed in Chapter 7 are exaggerated under any attempt to apply the alternative rule for budgetary management. The predictions made about the employment-increasing potential of budget deficits are likely to be unduly unrealistic on the optimistic side, while, at the same time, the inflation-producing results of these deficits are likely to be discounted. The targeted combinations of unemployment and inflation rates that emerge from the choices of politicians are quite likely to lie off of the short-run Phillips curve for the economy rather than on it. Even if a point is selected that lies on the relevant curve, the relative weight given to the employment objective is likely to be high relative to that accorded to monetary stability.

Quite apart from biases in the selection of targets, however, the proposed rule is politically open to abuse because of the absence of any feedback constraint on error stemming from overoptimistic projections. Both the annually balanced budget and the balanced budget over the cycle contain some benchmarks that make it possible for citizens and politicians to judge whether the stated principle is being adhered to or being violated. Error becomes visible for all to see. No such possibility exists with the modern fiscal norm—”budget balance at full employment.” There simply is no way that actual budgetary performance can be evaluated with a view to determine whether the rule is or is not being followed. A large current deficit can quite readily be passed off as actually reflecting a restrictive fiscal policy. And there is no way that this allegation can be tested, because the idealized conditions that would be required for a test will never exist, and could never exist.

The Budget Reform Act of 1974

The Congressional Budget and Impoundment Control Act of 1974 has been described by
U.S. News & World Report as “a revolutionary budget reform intended to give Congress a tighter grip on the nation’s purse strings….” The passage of this act indicates that the elements of the Keynesian modification of our fiscal constitution have not gone wholly unrecognized. Such recognition that some things are awry with our fiscal conduct is an essential prerequisite to reversing the tendencies of the past generation. The act itself, however, is unlikely to be the revolutionary reform that
U.S. News & World Report suggested.

The Budget Reform Act emerged from a recognition that previous budgetary procedures generated a bias toward spending and budget deficits. The total amount of spending emerged as the product of many individual appropriations decisions. No decision was ever made as to the total amount of public expenditure. Moreover, decisions regarding taxation were made independently of decisions regarding expenditure. A decision to increase expenditure could be made in isolation from decisions about whether that expenditure should be financed by increased taxation, by issuance of public debt, or by reduced expenditure for other services. It increasingly came to be recognized that such institutional practices created biases toward public spending and budget deficits.

This growing recognition inspired and informed the Budget Reform Act of 1974.
*66 A Budget Committee was created for each house, and these committees were given the task of setting overall targets for revenues, expenditures, and the resulting deficit or surplus. As well as setting a target for the overall level of expenditure, these committees were supposed to apportion this amount among sixteen functional categories of public expenditure. A Congressional Budget Office was created to assist in this process, as well as to make five-year projections. These projections were designed to help gauge the future impact of present decisions. Under the previous setting for budgetary choice, programs would often be created with small initial expenditure requirements, but would soon undergo an explosive growth in spending requirements.

The fiscal year was changed to start on 1 October, rather than on 1 July. Accordingly, fiscal 1977 began on 1 October 1976, instead of 1 July 1976, as it would have done previously. By 15 May preceding the start of the new fiscal year, Congress is required to have passed its first concurrent resolution. This resolution contains tentative targets for outlays and revenues, as well as for such residually determined magnitudes as the budget deficit or surplus and the amount of national debt. Furthermore, this target amount for total outlay is apportioned among the sixteen functional categories. After the passage of this first concurrent resolution, the former congressional procedures take effect. Separate committees examined the various proposals for appropriations and revenues, with these various examinations being conducted in light of the first concurrent resolution. By late September, Congress is required to have passed the second concurrent resolution, with the fiscal year then starting on 1 October. This second resolution is supposed to resolve any discrepancies that arise between the first resolution and the decisions that were made subsequently.

The Budget Reform Act is not the first attempt at instituting a comprehensive process of budget review. The Legislative Reorganization Act of 1946 also attempted to impose a congressional assessment of the entire budget. That legislation established a joint committee of the Senate and House that would determine an appropriation ceiling, the intention being that this ceiling would control the growth in expenditure. Only in 1949, however, did Congress manage to establish such a ceiling. And it promptly ignored that ceiling by approving appropriations that exceeded the ceiling by more than $6 billion. Until 1974, there were no further attempts at overall congressional control of the budget.

It is, of course, always possible that the 1974 act will prove to be more successful than the 1946 act. The act itself, however, does nothing to curb deficit spending. Rather, it merely requires that the projected level of the deficit be made explicit.
*67 Moreover, any divergence of the second concurrent resolution from the first is likely to be in the direction of larger spending and larger deficits. It is unlikely indeed that the appropriations committees would generate projected outlays below the targets set in the first concurrent resolution. Any discrepancies would almost certainly be in the direction of increased appropriations. It is also unlikely that the second concurrent resolution would simply disallow all of these resulting discrepancies. While some may be disallowed, the second resolution is also likely to set higher expenditure levels, along with larger deficits. The most reasonable assessment of the Budget Reform Act seems to be, as Lenin might have put it: “Expenditure ceilings, like pie crusts, are meant to be broken.”

Short-Term Politics for Long-Term Objectives

Since the Keynesian destruction of the balanced-budget principle, we have witnessed a parade of alternative principles, all pretending to function as constraints on budgetary excesses. To date, none of these alternatives have been successful in this regard. Budgets continually become ever more bloated, and we have now become accustomed to thinking of a 2- or 3-percent inflation as an objective we might possibly attain once again in the dim future. One now hears little about price stability, and the pace of the extension of bureaucratic controls is still quickening.

The complex rules and principles that have been advanced over the past generation have given an illusion of control. They appear to mitigate the apparent disparity between common-sense notions of responsible fiscal behavior and the widely sensed irresponsibility of present budgetary outcomes. These complex rules suggest that the appearance of irresponsibility is illusory, and that such observed budgetary behavior is really necessary to fulfill the precepts of fiscal responsibility. These arguments say we must travel the deficits road to surplus (or balance). Since our political conversion to Keynesianism during the Kennedy administration, we have been told that deficits today will stimulate the economy into producing full-employment surpluses tomorrow. Only tomorrow never seems to come. Deficits have become ever more firmly ensconced as a way of life, and the imminency of surpluses has receded—once we were told that it would take only a few years before the federal budget would be in surplus, but the number of years lengthens as the size of the deficits grows. Can any honest person realistically predict balance in the federal budget?
*68

Such complex budgetary rules as “balance at full employment” serve to rationalize budgetary irresponsibility by playing upon the sense that the present is unique and involves special circumstances, and that once these circumstances have been dealt with, we can revert to the rules applicable to “normal” settings. This is like the alcoholic who has some sense that all is not well with his conduct of his life, and who resolves to get hold of himself once the particular tensions he currently finds unbearable have passed him by. Only each day, week, or month presents a fresh set of tensions, unusual circumstances, and special conditions, so “normalcy” never returns, for either the alcoholic or the Keynesian political economy.

We have been witnessing the political working out of a conflict familiar to all of us, that between short-term and long-term considerations. We all recognize that the fixation on some long-term objective or goal is necessary to provide a disciplinary base for judging short-term choices. But if that long-term objective is not fixed in mind, or if it is permitted to be swamped by momentary, short-term considerations, the result almost surely will produce a drift in directions far removed from those that would be considered desirable.

Politicians themselves have, for the most part, short time horizons. For most of them, each election presents a critical point, and the primary problem they face is getting past this hurdle. “Tis better to run away today to be around to fight again another day” might well be the motto. This is not to say that politicians never look beyond the next election in choosing courses of action, but only that such short-term considerations dominate the actions of most of them. Such features are, of course, an inherent and necessary attribute of a democracy. But when this necessary attribute is mixed with a fiscal constitution that does not restrain the ordinary spending and deficit-creating proclivities, the result portends disaster.

We do not suggest that we relinquish political and public control of our affairs, but only that politicians be placed once again in an effective constitutional framework in which budgetary manipulation for purposes of enhancing short-run political survival is more tightly restrained, thereby giving fuller scope to the working of the long-term forces that are so necessary for the smooth functioning of our economic order. Just as an alcoholic might embrace Alcoholics Anonymous, so might a nation drunk on deficits and gorged with government embrace a balanced budget and monetary stability.