Deficit bias

Monetarists argue that both monetary policy (^{Friedman, 1968}) and fiscal policy (^{Barro, 1974}) are incapable of having a long-term impact on real economy. For this reason, they propose as optimal fiscal policy the one that keeps the tax rate constant, that is, they eliminate the use of discretionary measures, leaving the task of stabilization to monetary policy and automatic fiscal stabilizers (^{Friedman, 1948: 247-250}; ^{Barro, 1979: 941-950}; ^{Lucas and Stockey, 1982}).

However, with some exceptions, some authors do recognize the capacity of both types of policies to generate short-term effects. Therefore, just as monetary policy has an incentive to be expansive in order to promote employment at the cost of higher possible inflation (^{Nordhaus, 1975}) , fiscal policy has a bias towards deficit (^{Kydland and Prescott, 1977: 473-474}; ^{Buchanan, 1997: 120-123}; ^{Halac and Yared, 2014}).

This bias is reinforced in certain situations, such as elections. Given that voters are not aware of the intertemporal budgetary constraint (^{Corsetti and Roubini, 1993: 63-70}), the Government has an incentive to deviate from the optimal fiscal policy, with the objective of increasing the level of activity and employment, thus raising its positive image (^{Buchananan and Wagner, 1977: 114-115}; ^{Cukierman and Meltzer, 1986}; ^{Poterba, 1994: 816-818})². The incentive for diversion is not limited to the level of public spending, but also includes its composition by increasing consumption-related items and reducing investment-related items (^{Rogoff, 1990: 21}).

Even during non-electoral periods, the common pool problem shows that certain items of public expenditure can be used to favor certain groups of interest (^{Weingast et al., 1981: 650-651}; ^{Velasco, 1997: 4}).

Procyclicality bias

The common pool problem also generates a tendency towards pro-cyclicality. Indeed, in times of boom, tax revenues are plentiful, while there are opportunities for private profit, which increases the incentive to lobby. Thus, additional fiscal resources are not used to reduce deficit or debt in order to be better prepared for a reversal of the cycle (^{Lane and Tornell, 1996} and ¹⁹⁹⁸; ^{Tornell and Lane, 1998}; ^{Lane, 2002}). This voracious effect may explain why both in Europe (^{European Commission, 2001: 63-64}) and in other developed countries (^{Hercowitz and Strawczynski, 2004}; ^{Balassone and Francese, 2004}) the procyclical character is concentrated during booms, while in recessions there is usually a behavior that is rather acyclical (^{Manasse, 2006: 15-19}).

While there is evidence of procyclical fiscal policies in various regions of the world³, this bias is reinforced in developing countries (^{Talvi and Vegh, 2005}; ^{Alesina and Tabellini, 2005}), in part because credit constraints prevent them from running fiscal deficits during the recession (^{Kaminski et al, 2004}).

On the other hand, the typical volatility of these economies is reflected on the tax bases, ultimately impacting collection. In this way, any attempt to soften intertemporal public spending would mean incurring large surpluses during booms and significant deficits in recessions. However, as economies with large unmet social needs, the pressures to allocate surpluses for this purpose are very significant, so it is not so easy to generate savings in the boom (^{Talvi and Vegh, 2005: 177-180}).

Membership of a monetary union and the delay in implementation

A monetary union implies a single monetary policy and as many fiscal policies as there are members of the union. Nothing ensures that all these policies are compatible with those developed by the Central Bank (^{Buti et al., 2001}). Indeed, there is reason to argue that the deficit incentive is higher in these cases. First, the incorporation into a common monetary space generates a decrease in interest rates in many countries, which encourages indebtedness and allows high fiscal deficits to be incurred (^{Buti and Giudice, 2002: 2-4}). Second, given the interconnection between countries and the damaging spillover effects of the eventual default of a member, the possibility of being rescued creates a moral hazard that may encourage national authorities to incur larger deficits (^{Inman, 1996: 2-3}). Third, since it is not in charge of monetary policy, political short-sightedness does not allow us to see the adverse effects of a larger deficit on inflation (^{Beetsma and Uhlig, 1999: 553})⁴.

Having ceded monetary (and exchange rate) policy to a supranational central bank, a very important challenge for the design of fiscal rules arises: to be sufficiently rigid to avoid excessive deficits, but also sufficiently flexible to reduce volatility.

Even governments that have a strong reputation and are not expected to have the biases outlined above have incentives to implement fiscal rules. The reason for this is the time it usually takes to implement a discretionary measure in fiscal policy (^{Kopits, 2001}: 8; ^{Wyplosz, 2005: 81-82}). In fact, several quarters may elapse from the identification of the need until the measure takes effect (^{Blanchard and Perotti, 1999, 13-16}).

The basis of the rule

The basis of the rule refers to the indicator on which the numerical restriction is made (^{Ter-Minassian, 2010: 10}). This aspect is extremely important, as its choice conditions the response of the rule to the preservation of solvency and the maintenance of stability (^{Bohn and Inman, 1996: 30-34}). Consideration should also be given to how controllable it is by the Government, if it is able to provide operational guidance at all times during the cycle, and the ease with which it can be monitored (^{IMF, 2009: 4 and 20}). The most frequent candidates are the level of indebtedness, expenditure, income, and the fiscal result.

Level of indebtedness

If the objective is to ensure the solvency of the State, the first indicator referred to is public debt. By prohibiting the level of indebtedness from exceeding a certain percentage of the GDP, the development of the fiscal outcome and, consequently, also of government revenue and expenditure, is conditioned. In addition, such a rule is simple, transparent, and easy to communicate.

However, as the only instrument, it has the disadvantage that it takes some time to be influenced by changes in other fiscal variables, so it does not restrict policy in the short-term, while being affected by factors that are not within the reach of the Government, such as the interest rate (^{IMF, 2009: 21}). Similarly, it does not generate any guidance when debt is below the limit. Finally, in the face of external shocks that could significantly deteriorate the level of activity and government revenues, the deficit may generate an increase in the debt ratio, forcing adjustments to be made, in a clearly procyclical manner (^{Schaechter et al, 2012: 7}).

Expense

Total, primary or current public expenditure is a variable that is almost perfectly defined by the Government (particularly primary), while providing operational guidance at all times and being easy to monitor (^{IMF, 2009: 21}). The restriction is based on limiting the nominal or real growth rate or its share of GDP.

Conditioning to cyclical behavior is best adjusted when the growth rate is used as a reference. In effect, a limit to the annual variation of expenditure imposes a restriction on any procyclical behavior during the boom or when income is increased by the evolution of a particular variable (prices of raw materials, investment in housing, etc.). At the same time, it allows for expansionary action in the recession, albeit limited.

The main shortcoming of this type of rule is that it is not directly related to solvency, since it does not include revenue objectives (^{Schaechter et al., 2012: 8-9}). Thus, on the one hand, it is not able to influence the level of indebtedness. On the other hand, it does not seem to be a good option in the case of immature tax systems, since public revenues are expected to increase along with the modernization of the State and expenditures, in this context, may well increase at a similar rate to that of revenues without jeopardizing the solvency of the public sector (^{Machinea, 2012: 76}).

Income

A rule based on public revenues usually imposes maximum and/or minimum limits on the tax burden. In many cases they are very difficult to comply with, since fiscal revenue- unlike expenditures-does not fall under the direct control of the Government because it is highly dependent on changes in the level of activity (^{Buti et al., 2003: 16}). Even though the elasticity may be greater than one, imposing a floor or ceiling on the tax burden can limit the countercyclical nature of the tax system, being counterproductive in terms of stability (^{Schaechter et al, 2012: 9}). They are also partial rules, leaving aside expenses, and are, therefore, unable to ensure responsible behavior in terms of solvency.

Other types of income rules are those that require a portion of the proceeds to be used for a specific purpose, often to generate savings during booms in the form of sovereign funds. They are usually related to the presence of non-renewable resources, thus, in addition to stability, the objective is to improve intergenerational equity (^{Jiménez and Tromben, 2006}).

Fiscal result

After analyzing the limitations of the rules on the above variables, in view of the objectives of solvency and stability, the need arises to act on a more complete indicator such as the fiscal result (^{Sacchi and Salotti, 2015}). In fact, its calculation incorporates both expenses and income, while its behavior conditions the evolution of the debt. Thus, a balanced fiscal outcome over the whole cycle means that expenditure is placed around revenue, without generating significant changes in the level of debt, which would in principle ensure solvency (^{Tapia, 2003: 39-41}).

Similarly, a limit on the fiscal outcome in terms of GDP is relatively simple to monitor, while providing constant guidance as to the direction the policy should take. The indicator is sufficiently controllable by the Government, although there is some difficulty with cyclical components and interest payments (^{IMF, 2009: 20-21}).

Nevertheless, a rule on this basis does not provide for certain countercyclical elements, which in turn may cause solvency-related problems. If the objective is fixed, i.e., not changed according to the phase of the cycle, should the constraint be met close to the limit during the boom, it is highly likely to be missed once the economy slows down as a result of the fall in revenues. In this way, in order to comply with the rule again, the Government would be forced to consolidate the fiscal result in a clear procyclical action in the recession (^{Manasse, 2005: 8-11}).

One option to mitigate the absence of countercyclical elements is to apply the target on the structural indicator (^{Ter-Minassian, 2014}; ^{Bova et al., 2014}). In doing so, automatic stabilizers are excluded from the scope of the rule (^{Ter-Minassian, 2010: 8-11}). However, as seen since the 2008 crisis, this indicator is not satisfactory in identifying income from asset revaluation. To this end, it is very useful to develop a structural indicator that does consider these effects, such as the one developed by ^{Zack, et al. (2014) 6}.

Although, even if you want to enhance the stabilizing effect of fiscal policy, without having to resort to escape clauses or break the rule, you can consider some kind of mechanism in which the target level varies throughout the cycle. One possibility is that the objective is a fixed percentage of the Output Gap (OG). Thus, during booms, when GDP is above potential, the target SFR would be positive. Conversely, in recessions, where the OG is negative, the same rule would allow a structural deficit to be incurred.

Combination of bases

Another alternative to cover both the solvency and stability objective is to develop a rule that combines different bases. The most frequent combinations are restrictions on fiscal expenditure or result, with limits on the level of indebtedness.

As seen above, the level of indebtedness is a good indicator to encourage more responsible attitudes in terms of solvency, but in many cases, it does not generate any kind of guide for fiscal policy, and at the same time it does not imply important restrictions in cyclical terms. Expenditure, on the other hand, is useful for stability purposes, especially when its rate of change is limited, but because it does not take into account income, it is unable to act on solvency as a whole. Thus, a combination of both indicators could include the two final objectives, without generating a highly complex rule and maintaining their transparency (^{Schaechter, 2012: 9}).

The main limitation of this combination is that it still does not directly consider income. One way to include income more directly is by combining the debt target with another SFR target (^{Ter-Minassian, 2010: 11-12}). It should be noted that this combination is the one originally chosen by the European Union for its members, although recent experience, in many cases, has not produced the expected results.

There could even be a rule in which the objective of the base fluctuates with the cycle, so as to promote the stabilizing effect. On the other hand, the level of indebtedness, together with the possible existence of contingent liabilities, could define the value above which the target level would vary. In the case of a country with low debt and no contingent liabilities that wants to maintain or even increase its level of debt, this value could be zero or negative, respectively (^{Blanchard et al., 1990: 10-12}; ^{Cuddington, 1997: 6-8}). However, in a country where the debt is above the allowed maximum, or where there is a real probability of having contingent liabilities (^{Tapia, 2003: 39-41}), this value could be positive and higher the more it surpasses the limit.

Escape clauses

The basic idea of designing a rule on a cyclically fluctuating SFR objective, with flexibility and considering the level of indebtedness and contingent liabilities, is to minimize the need to temporarily suspend the rule. However, no design can be so complete as to previously consider all possible scenarios that may arise. Thus, escape clauses must always be included in the rule.

This poses a major design challenge, related to the identification of those exceptional moments. An effective method of identification is essential to allow the use of flexibility and to prevent attempts to circumvent it simply for short-term gain, without consideration for long-term solvency. Furthermore, once the decision to temporarily suspend it has been taken, the return should be specified, as well as the treatment of the accumulated deviations (^{IMF, 2009: 26}). As will be seen in the following section, the proposed generic rule and its application in the Spanish case has an escape clause designed (Equations 1 and 2). The basic idea is to condition the deactivation and subsequent reactivation to a certain value (in this case the OG) to be defined by governments on a case-by-case basis.

The case of Spain prior to 2008: an application exercise

The first step in applying this rule to the Spanish case is to have a good SFR indicator. For this purpose, the first option would be to take the estimate of the European Commission. However, while this methodology may capture some of the cyclical components of income and expenditure, it does not consider extraordinary income from asset revaluations, and is subject to large ex-post corrections, so basing a policy decision on this indicator is very risky.

In turn, the SFR will be used for Spain, estimated by Zack et al. As can be observed in Table 1, this differs significantly from the SFR estimated by the European Commission, due to the fact that it considers the extraordinary public revenues generated by the housing bubble and the construction boom. This indicator identifies that the fiscal deterioration in Spain began towards the beginning of the 2000s and already in 2004 it showed a structural deficit of the order of 4 percentage points of GDP. On the other hand, for the SFR of the European Commission, as well as for the OFR, the fiscal deterioration only began in 2007 and it was not until 2008, when the crisis began, that the first deficit was observed.

Once we have a good SFR indicator, we proceed to set the parameters of the rule. For Spain in the years prior to 2008, an appropriate percentage for the target value of SFR ("y" in Equation 1) is considered to be 50% of the OG. Thus, for example, when the GDP exceeds potential by 1%, then an SFR of 0.5% of potential GDP will be required, plus automatic stabilizers to obtain the OFR. On the other hand, in the face of a recession in which the GDP is 2% below potential, the rule itself will make it possible to incur a structural deficit of 1% of potential GDP, to which automatic stabilizers should again be added. With regard to the average target value ("x"), it is considered that Spain did not face an over-indebtedness situation nor a need to generate long-term savings⁷. A balanced budget is thus envisaged in the medium term. As for the escape clause, the thresholds for the suspension and reactivation of the rule are set at 3% and -3% OG, respectively. Thus, it is considered that Spain has not experienced any situation in the years prior to the crisis that justifies the suspension of the rule.

In short, the formula of the rule for Spain would be as follows:

Where SFR is the structural fiscal result and OG is the output gap.

Next, the calculation of the additional fiscal savings that Spain would have achieved between 2000 (when it joined the Monetary Union) and 2008 (the year in which the crisis began) if it had complied with a rule of this nature is carried out⁸. As can be observed in Table 1, these resources would have reached 48.5% of the GDP. Even if the SFR estimated by the European Commission had been considered, instead of the one calculated by ^{Zack et al. (2014)}, the additional savings would have been 25% of the GDP.

Thus, had the proposed rule been applied in Spain since 2000, the country would have started the crisis with a very interesting flow of resources. This, on the one hand, would have cooled the economy in the years leading up to the crisis, causing some effect on the housing bubble. More importantly, however, it would have been possible to implement a more powerful countercyclical policy once the crisis had begun. In effect, following the proposed escape clause, the rule would have been temporarily suspended in 2009, as the OG was -3.3% of the potential GDP. In this way, there would have been free reign of discretion to carry out measures to boost aggregate demand, beyond what is allowed by the rule.

In short, this exercise shows the importance of a fiscal rule designed ad hoc (in a context of monetary union). An effective SFR indicator, which clearly identifies extraordinary revenues, is essential for this. Otherwise it would be very difficult to withstand the pressures to expand public spending when the rule required an observed surplus.

Finally, a clarification: what has been done in this section is a partial fiscal policy exercise. Thus, it is not unknown that the proposed measures would have had effects on other economic variables, which could have modified the scenario. Nor does it argue that fiscal policy alone could have prevented the crisis. Conversely, the exercise is only intended to show that different fiscal behavior in cyclical terms would have been not only prudent, but also desirable in the interest of the real stability of the economy.