You cannot make a fat man skinny by tightening his belt.

–John Maynard Keynes

Introduction

Balanced budget requirements are often heralded as a panacea for public debt . The goal of such a proposed amendment (or any other form of requirement that public budgets be balanced) is laudable: preventing rampant public deficits and stemming the tide of public debt . The fear is that without such a policy solution, a financial collapse can be the only outcome. Whether this is true remains to be seen and is an important question, of which Salsman (2017) provides an excellent exploration. This paper, however, asks a different question: are balanced budget requirements likely to succeed in their stated goal? Looking through individual states in the U.S., the answer would appear to be a resounding “no” as forty-nine of the fifty states have at least some form of balanced budget rule, either at the state-constitutional level or legislatively, yet all fifty states are currently running deficits.

This is surprising for several reasons. First, the political arena is populated by individual people who, for the most part, seem quite capable of balancing their own personal budgets , as we very rarely hear of politicians who have persistent financial difficulties (Wagner 2012). Thus, we can confidently say that even though living within a budget can be challenging at times, the overwhelming majority of the people involved in determining political budgets are able to do so in their own lives. Second, the problem with political budgets cannot be due to the fact that a group is making a collective choice, as other groups throughout society are able to do so.Footnote 1 Finally, the problem cannot be one of difficulty. The amount of debt, whether done by an individual or a group, is determined by subtracting expenditures from revenues. If keeping these two numbers roughly similar were merely difficult, we would expect errors to be made on both sides of “zero,” with some years seeing revenues exceeding expenditures and others vice versa. That we persistently see error in the form of expenditures exceeding revenues points to the existence of some sort of systematic bias towards debt.

The general logic behind the support of a balanced budget requirement is that requiring revenues to equal expenditures will automatically cause this to be true without affecting revenues. But to presume this is to misapply notions that are relevant for private debt to public debt . Specifically, it conflates the process by which public debt is determined with that of private debt . While on the surface, both are determined by the difference between revenues and expenditures over a period of time, there are substantial structural differences which lie beneath the surface. These differences are conventionally ignored by traditional public finance theorists, who concern themselves with examining aggregate figures rather than the processes and institutions through which these aggregate statistics are determined. Instead, problems with balancing public budgets are imagined to be the result of a lack of e.g. political will and it is here where a written requirement that revenues equal expenditures gains its acceptance in the sense that willpower and formal requirements are viewed as substitutes. To put this another way, there is no need for a rule requiring something when there exists sufficient will to do it. Where there is a lack of willpower, formal rules are imagined to be a near-perfect substitute.

Of course, there are problems with this. Most obviously, a legislative body can always decide to grant exceptions for particular circumstances. For example, even during the golden age of “that old time fiscal religion” (Buchanan and Wagner 1977), where public debts were avoided as a matter of principle, wartime spending was understood to be an acceptable exception to the principle. In today’s world, where governments are (rightly or wrongly) expected to provide financial assistance in times of disaster or hardship, unforeseen circumstances can give rise to situations where such exceptions could be granted.

Setting this aside, there are two related problems that a balanced budget requirement faces. The first is the process by which revenues and expenditures are determined in a political arena. The second is the substitutability of public spending and regulation. Taken together, these two problems provide a myriad of possible strategies for would-be politicians to accomplish the stated goal of equalizing public revenues and public expenditures while having little to no effect on actual revenues and expenditures from the perspective of the taxpayer.

This paper contributes to the growing body of literature that considers public finance to be an explanatory, rather than a hortatory, discipline. Rather than take the approach of providing advice and counsel, this paper seeks to generate understanding. In this respect, it is most similar to Hebert and Wagner (2013) and Hebert (forthcoming, a), which explain the source of tax code complexity in democratic societies and Hebert (forthcoming, b) which explains the trifurcation within the U.S. Federal budget process among committees related to authorization, appropriation, and revenue generation.

This paper will proceed as follows. Section “Private Debt vs. Public Debt” describes some of the differences between private debt and public debt and the incentives faced when making forecasts. Section “Budgetary Chicanery” describes different means of manipulating federal budget numbers that are commonly used. Section “Spending vs. Regulation” discusses the substitutability of expenditures and regulation and in doing so introduces the ability to move implicit expenditures off-budget. Section “Conclusion” concludes.

Private Debt vs. Public Debt

In all cases of debt, the surface-level calculation is identical: debt is determined by the difference between an entity’s revenues and expenditures over a given period of time. This entity can be an individual person, a group of people, a business, or a government—it makes no difference, at least at this surface level of analysis. For example, if a household has annual revenues equal to $100,000 and annual expenditures equal to $75,000, we can surmise that this household has an extra $25,000 left at the end of the year. This money can be carried forward into the next year, which allows the household to spend more money in this second year. Alternatively, if the household has $100,000 in annual revenue and expenditures equal to $125,000 per year, this household must borrow $25,000 in order to finance their operations. The same type of analysis can be performed with respect to public budgets , albeit typically with much larger numbers.

This analysis, however, skips over two important realities. The first is the category mistake that is made when comparing private and public debt . The second, which derives from the first, is to examine how revenues and expenditures are determined in the public setting and how this differs from how they are determined in the private setting.

As Wagner (2017, p. 118) points out, democratic governments around the world do in fact incur debts in their name. However, it would be erroneous to describe a government as indebted because there is no explicit relationship created between a debtor and a creditor created. At first blush, this should sound strange. But if we consider government as a financial intermediary which connects people who wish to save currently by purchasing public debt and everyone else in society, the notion of indebtedness breaks down. This breakdown is not caused by some ambiguous relationship between bondholders and the parliamentary assembly , as we could accurately describe saving money with a private bank in much the same way. Instead, the difference between the two stems from the idea that the bank, its employees, own the value consequences of their actions, whether they be positive or negative, and bear ultimate responsibility. In parliamentary assemblies, this is simply not the case as individual members of a parliamentary body do not gain from creating value through debts nor do they lose from destroying value from debts. In fact, in line with Buchanan and Tullock (1962) and Tullock (2005), it may even be the case that individual members stand to reap private rewards from apparent fiscal irresponsibility. Such irresponsibility is unlikely to garner sufficient support if explicitly stated and so individual members will seek to justify any proposed change in policy that affects expenditures or revenues. To understand this, it is necessary to examine the process by which these figures are determined in the public sector. This is not to say that individual members are intentionally trying to hide nefarious deeds. Often, the goals are laudable and include such language as helping to reduce, e.g. poverty.

The determination of revenues and expenditures in the private and public spheres is markedly different in several ways. Suppose for simplicity that the time period under consideration is one calendar year, beginning in January first and ending December thirty-first. Both private and public institutions engage (at least implicitly) in some form of forecasting about the upcoming year’s revenues and expenditures. For simplicity, suppose that the private household’s income is earned by one person and that this person’s annual salary is $120,000, payable on the first of every month. This household’s expected budget every month would therefore be equal to $10,000 multiplied by the probability that the person maintains employment. Perhaps, as Clark Griswold does in the 1989 classic, National Lampoon’s Christmas Vacation, this worker has also receives an annual bonus check at the end of the year. Thus, this person’s annual revenues for the upcoming year may exceed their salary and they may choose, as Clark Griswold did, to purchase an expensive purchase at the end of the year with the expected bonus check. If that bonus check were to not show up, or if it was in an amount less than expected, debt would be incurred and the responsibility for reconciling this debt would fall squarely on the shoulders of the household, which has identifiable people in it.

Government revenues at the beginning of a fiscal year are determined in a similar way. Congress determines some aggregate dollar figure of taxable activity (the tax base) and multiplies it by some aggregate figure of a tax rate and thus arrives at a figure for total revenue. Like the private household example, this involves a significant amount of forecasting. Unlike the private realm, however, should actual revenues fall short of the actual expenditures, it is not at all clear who is to blame for the resulting deficit. To be sure, blame will be cast. Congressional Democrats, for example, will likely point to a proposed tax increase that was blocked or a loophole that was created by Congressional Republicans as the culprit while Congressional Republicans will likely point to a program that Congressional Democrats supported which cost more than anticipated. In either case, determining which specific members of Congress are at fault is an impossible task just as it’s impossible to determine which specific members of Congress are at fault for a nigh-incomprehensible tax code (Hebert and Wagner 2013; Hebert, forthcoming, b).

The above analysis, like Wagner (2012), points to issues surrounding errors made in forecasting. While individuals in both the private and public spheres may have some general sense of what the future will hold, it is impossible to predict the future with certainty and thus, error is a feature of the world, not a flaw. However, the above analysis could easily describe particular instantiations of normally distributed error, which would offset itself over a sufficiently long time horizon. In thinking about this, it is necessary to describe the incentives that individual people face in different institutional settings when making such forecasts.

In the private realm characterized by individuals or individual households, prudence would seem to be likely. A household may reasonably be expected to underestimate revenues and overestimate expenditures over a given year in order to avoid finding themselves in unanticipated debt. In other word, of the myriad methods by which revenues and expenditures may be forecasted, a private household can be reasonably expected to select the forecasting method that produces a set of estimates biased downward with respect to revenues and upward with respect to expenditures.

This is not the case in the public realm, for here the coin of the realm is the ability to sell voters and constituents on the efficacy of a proposed piece of legislation. Thus, proponents of a new tax scheme are likely to select the methodology that produces the most lavish of estimates for projected revenues. Likewise, proponents of a new spending plan will likely select the methodology that produces the lowest total cost. In other words, where the private realm could be characterized by “hope for the best but plan for the worst,” the public realm could be characterized by “assert the best.”

Budgetary Chicanery

As Block (2008) notes, “one pesky reality of [government] budgeting is that it requires the use of numbers.” In an ideal world, these numbers would bear some semblance on reality. The problem with this is that in order to accurately forecast the future, one must know, precisely, what the effect of a bill would be on the citizenry. Increased sophistication of modeling techniques has done little to resolve the difficulties of predicting the future accurately but have given us what Graetz (1995) refers to as the “illusion of precision.” However, if getting the numbers required for balancing the budget were merely difficult, then all we would be concerned with would be error.

As it turns out, the error in forecasting revenues and expenditures are almost entirely one-sided. In the case of U.S. Federal revenues, each and every year for the last ten years has fallen short of projections. Likewise, actual expenditures exceed projected expenditures over this time period. The challenge, then, becomes one of explaining these systematic biases.

In order to assess the financial impact of any proposed change in legislation, two numbers must be established. The first number that must be established is a sort of base for purposes of comparison. This number, appropriately called the “baseline,” is a projection of what will happen under current legislation if nothing changes in current law. The second number is an estimation of revenue that would be generated or costs that would be imposed if the proposed legislation were to pass. This number is referred to as the “score.” Taken together, these numbers provide Congress with an estimation of where they are currently heading financially and how the proposed legislation will affect that trajectory. While both are highly important numbers, the baseline is perhaps the more important of the two as a high baseline would reduce the apparent cost of any proposed legislation.

These numbers became particularly important after the 1990 Budget Enforcement Act. This act imposed two restrictions on Congress: First, Congress was not to adopt spending legislation that would cause annual appropriations on discretionary spending to exceed caps established in the budget resolution. Second, all new tax legislation or changes to entitlement programs must be revenue neutral, meaning that decreases in revenue or increases in spending had to be offset by increases in revenue or decreases in spending elsewhere in the budget.

In an ideal world, the economic assumptions and methodologies used to construct these numbers would not matter. Unfortunately, this is not the case. As Block (2008) describes, “in the imprecise world of budgetary mathematics, even seemingly small changes in estimation methodologies and economic or behavioral assumptions can lead to significantly different scores.” There will always be legitimate differences of opinion over which assumptions to make or which estimation methodologies to use, however it would be foolish not to also acknowledge that particular assumptions and methodologies that best suit a particular ideology will be used in order to advance a particular agenda.

To add a further complication to this, the task of classifying an item as taxation or spending is not as straightforward as one would intuitively believe. Both Reagan and Clinton, for example, have argued that increases in Social Security benefit taxes should be scored as spending cuts rather than tax increases since the tax increases would effectively reduce the amount of benefit that the recipients actually received.

Beyond this, Congress is able to manipulate each year’s budget through selectively timing when revenues or expenditures occur. For example, this can be accomplish through inconsistent uses of a budget window. Budgets are, after all, tied to a specific amount of time. Every household has its own daily budgets, but also weekly, monthly, yearly, and perhaps even lifetime budgets. In setting these budgets, households set aside a certain amount of money that can be spent over the course of a specified amount of time. The same applies to any committee or office that makes budget projections and recommendations—they must put forth a plan that sets aside a certain amount of money that is available to be spent over a certain amount of time.

In the past, Congress used a short, one-year basis for budgeting purposes. While this provided Congress with the flexibility to amend the budget resolution each year, it also meant that individual Congresses were not taking into account the long-run costs of any proposed legislation, instead only taking into account the current-year costs.

In 1990, Congress passed the Budget Enforcement Act as a means of resolving this by moving to a statutory five-year minimum window for purposes of budget resolutions. In 1997, Congress began requesting ten year budget information from the Congressional Budget Office (CBO) and the President’s Office of Management and Budget (OMB) followed suit, similarly supplying 10 year budget information. The important thing here is that both the legislative and the executive branches both used the same budget window, the same scoring practices, etc. in their evaluations. Doing so allows for a more fruitful comparison between the two branches’ budget proposals.

The challenge, however, is that there are no formal rules governing the budget window that must be used other than the five-year minimum. In 2004, for example, the OMB began using a ten year window in some instances and a five-year window in others in the same budget proposal. This made meaningful comparisons and calculations difficult and resulted in charges that the President’s administration was using selective changes in the budget window solely to advance the president’s legislative agenda.

Another source of timing difficulty comes from the use of various accounting gimmicks. These can range from the relatively simple to the complex. An example of a simple accounting gimmick comes from the inconsistent use of cash versus accrual accounting . Briefly, cash accounting is a type of accounting practice that records revenues in the period in which they are received and expenses in the period in which they are paid. It does not include revenues that are to be collected in the future nor does it include expenses that are to be paid in the future in any way. In contrast, accrual accounting records incomes and expenses when the right to receive or obligation to pay them arises regardless of whether funds have been received or paid. In the private sector, the accrual method dominates due to its forward-looking nature and is even required based on generally accepted accounting principles established by the Financial Accounting Standards Board. Congress, however, does not have such a requirement. In general, Congress uses the cash accounting method (though there are exceptions) and OMB/CBO reports are calculated using this method as well. This method affords legislators with several tools of budget chicanery.

Perhaps the easiest tool is to simply delay payments or receipts by one day and into the subsequent fiscal year. Any budget must be a budget over a period of time, as one collects revenues at one point in time and expends those revenues at another. This can be over the course of a day, a month, a year, or even a lifetime. In private lives, this poses no real challenge. For example, knowing that an inheritance check is coming next month, a private person can reasonably account for the additional funds in their current budget and adjust spending habits accordingly. It would be inaccurate, however, to consider this as “current income” on one’s monthly budget.

In Congress, this is most apparent by recognizing that the fiscal year begins on October first while many spending programs are based on calendar years. Because of this, revenues and expenditures can be reported as occurring in different parts of the calendar year which correspond to different fiscal years. In the event that a spending program will put an appropriations committee over its cap, this committee can simply report the spending as occurring at a different part of the calendar year and avoid being punished for going over its limit during the current fiscal year. Delaying revenues or expenditures by one day can easily be employed to move billions of dollars from one budget year to another. Advance appropriations are similar to this in that an appropriations bill can be passed in one fiscal year while the authorization bill (and thus the actual spending) can be passed in the subsequent year, again shifting the spending from one fiscal year and into another. While these budget gimmicks are powerful, they are limited in that they can only move payments and receipts from one year to the next. In order to move budget items over longer periods, more sophisticated and complex variations need to be employed.

The easiest way to move budget items across longer periods is to phase programs in over time. In doing so, legislators commit to spending over much longer periods of time than the year in which they are legislating while simultaneously tying the hands of future Congresses (unless the future Congress takes action to repeal said program). Alternatively, rather than paying the full cost upfront to, say, construct a building, Congress can contract with a private contractor to build the building, giving the contractor the title to the building at the end of construction, subsequently paying rent to occupy the space created.Footnote 2 Again, using the cash accounting method, these allow the current Congress to commit to spending programs while only reporting a fraction of that spending in the current year. Under the accrual method, the full cost would be reported immediately regardless of when the money would actually be paid.

Another, slightly more complicated version of this is to pass legislation with a fixed expiration date or sunset even if the full expectation is that the legislation will be extended beyond this sunset or perhaps even made permanent. This “trick’’ works because of the methodology that the CBO uses when scoring any proposals. The CBO must, when scoring any proposals, consider them in light of existing law without taking into account future statutory changes to the law. As a result, if a proposal that will reduce revenue includes a sunset provision of three years, the CBO will score it only as existing for those three years and its score will be improved.Footnote 3 Block (2008) calls the temporary R&D tax credit as the most notorious offender of this sunsetting game. Despite widespread bipartisan support, a permanent tax credit for R&D has not been passed, instead the temporary credit first passed in 1981 has been extended several times such that it is now viewed as “virtually a political given.”

As a result, the cash accounting method, for all its merits of keeping an accurate accounting of levels of funds currently in the accounts of the government agencies, also provides these agencies with significant scope to engage in budget gimmicks by slightly altering the timing of the spending or revenue collection. Doing so can make it appear as if Congress as a whole is saving significant money on a year-by-year basis while simultaneously increasing long-term spending obligations. Regardless of Congress’s ability to kick the can further down the road, that road will eventually end eventually become so massive that it simply cannot be kicked any longer.

Spending vs. Regulation

At first blush, it is important to note that any parliamentary assembly is not in the business of producing any sort of economic good whatsoever. Rather, governments are merely the bridge between people who have money and people who want money. As is pointed out in Wagner (2016, pp. 136–162), we can view parliaments as a peculiar form of an investment bank. Like a private bank, parliaments connect people who have money with people who want money to perform some task. Unlike a private bank, with parliamentary assemblies there is no guarantee that the people contributing the money (i.e. taxpayers) are contributing their monies voluntarily (Podemska-Mikluch and Wagner 2013). In this sense, taxes can be viewed as akin to a forced loan by the taxpayer to the recipient of government expenditures.

Figure 9.1 illustrates this, showing the flow of dollars from citizens to the providers of public enterprises through the apparatus of government. In this scenario, the public enterprise is funded entirely by dollars received from the government which, in turn, receives its dollars from citizens in the form of taxes.

Fig. 9.1
figure 1

Money flow with government intermediation

Full size image

Because money is received by the government from citizens, it would count as revenue received and would show up on their budget. Likewise, money would flow out of the government’s coffers to finance the public enterprise and would show up as an expense. However, this is not the only way in which parliaments can direct spending and, in doing so, move money from citizens to public enterprises . Instead, governments could simply regulate the behavior of citizens and, in doing so, force payment from citizens to public enterprises directly (Fig. 9.2).

Fig. 9.2
figure 2

Money flow without government intermediation

Full size image

Here, rather than receiving and distributing funds, the parliamentary assembly merely directs citizens to purchase some good or service that is provided by the public enterprise. From the perspective of the citizen, the effect is identical: money has been moved from their accounts to the accounts of the public enterprise. However from the perspective of the parliamentary assembly’s budget , the effect is wholly different, as this would not be recorded as either a revenue or an expenditure at all.

These are also not mutually exclusive, as some enterprises are financed through both tax dollars and regulations. For example, in principle there is no reason why public transportation couldn’t be financed entirely through taxation, with riders paying zero price to get on or off the bus/subway regardless of how frequently they ride, how many transitions they make, or the total distance traveled. As a matter of practice, public transportation is typically financed through both regulatory means and through some sort of use-fee. For example, in the greater Washington, DC area, there is an extensive metrorail system that is used to bring people to, from, and around the District. During the rush-hour times, driving on the freeway is regulated in the sense that the road is reserved for vehicles carrying two or more passengers. In effect, what this does is raise the cost of driving oneself into the district in the morning and out of the district in the evening. The funds to provide the metrorail service are provided through a combination of tax dollars and the fees collected from riders.

This type of shared-cost arrangement does not guarantee that the funds collected from citizens paying directly for public enterprises will be sufficient to meet its expenditures, nor does it guarantee that this is the least-cost way to provide a metrorail system. In the event that user fees are insufficient to finance the operation of the DC Metro, supplemental tax dollars may be provided, however this would require far fewer dollars flowing from Congress to the DC Metro under this shared-cost arrangement than would be were the metrorail system funded entirely by tax dollars. Similarly, it may be more costly from the point of view of society as a whole to have this shared-cost arrangement, and analyzing to the extent to which this is true or false would an interesting exercise. However, it is certainly less costly from the perspective of government budgetary authorities to have a shared-cost arrangement aided through regulation than it is to have a metrorail system financed entirely by taxes.

The use of regulation in this capacity is tantamount to taking what would have been on-budget activities and removing them from the budget. The overall effect on citizens, however, remains the same. As a result, one concrete prediction that can be made and explored is that a balanced budget amendment would have little to no effect on the operation of parliamentary assemblies other than to shift some activities that are currently financed through taxation to providing them through regulation, effectively forcing citizens to finance their operation out of their own pockets. This was explored in Bennett and DiLorenzo (1982), who found a significant increase in the use of off-budget enterprises as a result of changes in state tax laws in the 1970s, and more recently in Bowler and Donovan (2004) and Mullins (2004).

Conclusion

This chapter has presented an overview of how federal budget numbers are determined in the United States Congress. To be sure, this is but a brief and simplistic overview that glosses over many details.Footnote 4 Nonetheless, it provides evidence of a sobering conclusion: the figures determined for public revenues and public expenditures can be manipulated with relative ease. Because of this, it is difficult to imagine a balanced budget requirement having a meaningful impact on the day-to-day operations of any type of parliamentary assembly . Today’s budget rules allow for Congress to include deficit spending in their budgetary outlook and so deficit spending is included in the budget proposals and resolutions. If it were no longer allowed, then it would not be included in the budget . Were deficit spending to become necessary over the course of the year, then one of the several forms of budgetary chicanery described above or a shift from using government expenditures to regulation would be used to reconcile any accounting differences.