Sage Journals: Discover world-class research

Abstract

The Swiss debt brake is widely appreciated as one of the most rationally designed fiscal rules in the world and was thus also discussed as a blueprint in the debates about fiscal rules in Germany, the European Union member states and Israel. However, evidence that this rule really contributes to sound federal finances, in the sense of reducing the cyclically adjusted deficits, does not exist yet. We investigate the effectiveness of the Swiss debt brake by employing the Synthetic Control Method. We find that the introduction of this fiscal rule improved the budget balance by about 3.7 percentage points on average in a post-intervention period covering five years.

Keywords

Swiss debt brake cyclically adjusted budget balance government debt synthetic control method H6

Introduction

On December 2, 2001, Swiss citizens voted in favor of the introduction of a new fiscal rule at the federal level. After the debt to GDP ratio of the Swiss federal government had almost doubled in only six years—from 13 percent in 1991 to 25 percent in 1997—government, parliament, and the general public wanted to break this trend and start with a consolidation of the federal budget.¹ A possible deficit bias of political decision-makers, emphasized in the surveys by Alesina and Perotti (1995), Alesina and Passalacqua (2016), and Yared (2019), in particular, during the consolidation of public finances, has been braced in Switzerland by the institutional setting of the budgetary process at the federal level. Tax rate increases require a change of the Swiss constitution whereas an increase in government expenditure only needs a simple majority vote of representatives (Geier, 2011). As a result, spending increases were rather financed by budget deficits than by tax increases because of political reasons and institutional restraints. Switzerland introduced the federal debt rule to counter this kind of institutionalized deficit bias and arrive at sound federal finances.

Several fiscal rules that came into effect later on can be traced back to the Swiss debt brake.² In 2009, Germany introduced a debt rule which targets the structural budget deficit of the federal level and the states (Laender) and, just as the Swiss debt brake, does not privilege public investment spending anymore. Moreover, the new rule includes a compensation account which is debited if actually accruing financial balances deviate from planned budget outcomes such that revenues fall below expenditures (Baumann, Doennebrink, and Kastrop, 2008; BMF, 2008; Kastrop and Snelting, 2008).³

In 2011, the European Union member states agreed on the European Fiscal Compact in response to the European debt crisis. Under this treaty, the respective states consented to implement an (almost) balanced budget rule at the general government level. Although it is the member states’ responsibility to decide on the details of their fiscal rule the compact requires some features following the Swiss balanced budget rule and the German debt brake. So it calls for a balanced budget and limits the structural deficit to 0.5 percent of GDP. The national rules must be accompanied by an automatic correction mechanism that becomes effective in the case of non-compliance with the rule. Of the twenty-five signatory states twenty-two states are bound by the contract (19 euro area Member States plus Bulgaria, Denmark, and Romania). Although the signatory states have meanwhile introduced national fiscal rules accordingly, it goes without saying that compliance with the Fiscal Compact considerably varies among them (Burret and Schnellenbach, 2013; European Commission, 2017).

In Israel, the experience with fiscal rules shows that they were hardly obeyed in the past. For Israel, it was important to obtain a fiscal rule that directly impinges on the spending trends and that directly punishes systematic deviations from the rule. Debrun, Epstein, and Symansky (2008) value the compensation account to be helpful in a way that systematic deviations from the fiscal target do not fall into oblivion. Finally, they recommend an expenditure rule for Israel including a feedback mechanism that adjusts the expenditure growth ceiling according to the long-run debt target.

Despite this development, it is contested whether fiscal rules actually contribute to fiscal sustainability. Many empirical studies find a debt reducing effect of fiscal rules (see Eyraud et al. (2018); Yared (2019); and, for a meta-analysis, Heinemann, Moessinger, and Yeter (2018)). The differentiated assessment of fiscal rules may have started with Bohn and Inman (1996), who consider the budget surplus of the general fund of Forty-seven US states covering the 1970–1991 time span and find a clear positive and statistically significant effect for strong rules, especially the no-carry-over deficit rule. In focusing on constitutional limits on guaranteed bond indebtedness in the US states for the 1961–1990 period, Kiewiet and Szakaly (1996) find that guaranteed debt is lower when the issuance of bonds is prohibited at all. A qualified majority in parliament or revenue-based limitations are, in contrast, not helpful in reducing the debt level. In also considering the US states, von Hagen (1991) does not find a significant effect of a debt rule or a strong balanced budget rule on debt per capita or the debt ratio based on a test for equal means. However, he finds that the median debt per capita is lower in the country group with a debt rule or a strong balanced budget rule. Analyzing Colorado’s Taxpayer Bill of Rights of 1992 using a synthetic control methodology, Eliason and Lutz (2018) do also not find any robust effect of fiscal rules on budget outcomes. However, in a difference-in-difference analysis for a large set of countries using historical data, Asatryan, Castellón, and Stratmann (2018) estimate that the introduction of a balanced budget rule reduces the debt to GDP ratio by eleven percentage points and reduces the probability of experiencing a sovereign debt crisis. For a sample of seventy-seven advanced and developing countries from 1984 to 2015, Gootjes and De Haan (2022) reported an evidence that fiscal rules constrain political budget cycles.

For Switzerland, there is evidence for the 1980–1998 time period, the 1986–1997 time span, and the 1980–2011 period that strong statutory fiscal constraints significantly reduce deficits and debt at the cantonal level (Burret and Feld, 2018a,b; Feld and Kirchgaessner, 2001, 2008; Schaltegger, 2002). Regarding the local level, fiscal rules reduce deficits. Beyond also providing evidence for the effectiveness of fiscal rules, Luechinger and Schaltegger (2013) show that fiscal rules improve the accuracy of budget projections. The authors interpret this result as a decrease in strategic behavior during the budget process. Using financial market data from 1981 to 2007, Feld et al. (2017) report that cantonal fiscal rules reduce spreads, although the re-establishment of a no-bailout regime regarding local jurisdictions has a quantitatively larger negative effect on cantonal spreads.

Regarding the effects of the Swiss federal debt brake on federal finances, two papers have recently been published.⁴ Building on a much earlier version of our paper, Asatryan, Castellón, and Stratmann (2018) identify a negative effect of the Swiss federal debt brake on the public debt to GDP ratio and the expenditure to GDP ratio using the synthetic control method, but only report these results in the passing without further explanations. Salvi, Schaltegger, and Schmid (2020) consider the effect of the federal debt brake on federal debt more thoroughly also using the synthetic control method and provide several robustness tests particularly on investment spending.

This article analyses the effects of the Swiss federal debt rule in more depth by focusing on the cyclically adjusted budget balances (CAB), the fiscal variable that is actually targeted by the Swiss debt brake. Our analysis thus adds to the work by Salvi, Schaltegger, and Schmid (2020), who concentrate on the stock of public debt, and provides a much more extensive analysis than Asatryan, Castellón, and Stratmann (2018), who also report results on the debt ratio and the expenditure ratio. As the Swiss debt brake is legally effective since eighteen years and practically effective since sixteen years, standard time series methods may have their identification problems. While the 2003 budget was the first under the debt brake, the government had to apply a transitional period that lasted until 2005. Therefore, and like Salvi, Schaltegger, and Schmid (2020) or Asatryan, Castellón, and Stratmann (2018), we employ the Synthetic Control Method to analyze the effect of the Swiss debt rule’s introduction on the CAB, also providing additional evidence on the federal government debt ratio as well as the general government debt ratio. Aside other results, we find an improvement of the budget balance by about 3.7 percentage points in a post-intervention period covering five years.

The article proceeds as follows: “The Design of the Swiss Balanced Budget Rule” section explains the mechanics of the Swiss fiscal rule, surveys the literature that discusses the features of this rule and deals with the start-up difficulties of the years 2003–2005. The “Synthetic Control Estimator” section then outlines the working of the empirical method we use. While the “Data” section reports the data, the “Empirical Analysis” section provides the empirical results and several robustness tests. The “Conclusions” section concludes the article.

The Design of the Swiss Balanced Budget Rule

The Swiss fiscal rule basically consists of a ceiling of total federal government expenditure, that is, expenditure in the next fiscal year must follow the predicted revenue for that fiscal year. Additionally, predicted revenue is multiplied by a factor that corrects for the position within the business cycle, thus actually targeting the CAB. The basic mode of an operation can be illustrated by equation (1)

E_{t} = k \cdot R_{t},

(1)

where the expenditure ceiling

E

for financial year

t

equals the revenue forecast

R

for the same year

t

multiplied by the adjustment factor

k

. This factor is given by

k = \frac{Y_{t}^{T}}{Y_{t}}

(2)

with

Y_{t}^{T}

being the potential real GDP (trend) and

Y_{t}

being the predicted real GDP of the year

t

. In case of an under-utilization of capacities, the ratio of potential real GDP over predicted real GDP is larger than one and expenditures can exceed predicted revenues. In times of booms, in contrast, the factor

k

forces the budget to generate a (yearly) surplus which results in a balanced budget over the cycle. Eventually, the calculation of the factor

k

relies on an adjusted Hodrick-Prescott filter instead of using a production function approach.

If total expenditure exceeds the ceiling, the (additional) deficit is booked in a compensation account. Deficits in this account must be redeemed in the subsequent fiscal years (without specifying the terms of the amortization, however). If the deficit of the compensation account becomes too large, that is, it exceeds 6 percent of effective expenditure of the last fiscal year, the government must reduce it below 6 percent of expenditure within three years.

Beyond that, the calculation of the expenditure ceiling comprises investment spending but disregards windfall revenue (hence not endangering the rules’ stringency) and the social security system. The law also considers the need for extraordinary expenditure. Yet, this is possible only for a finite number of cases which are conclusively stated in the law. Every single case needs the confirmation by the majority of parliament. Since 2010 the rule also covers the extraordinary budget (BV, 2002; FHG, 2006, 2010; Geier, 2011).

As mentioned, the Swiss debt rule was given credit from scientists and practitioners alike. So it is esteemed that the rule targets the deficit instead of debt as the former is under the direct control of policy-makers. Furthermore, a balanced budget target is welcomed because a surplus target would certainly raise the desire to spend these surplus amounts in line with short-run political demands instead of using them to unburden future generations. In addition, a balanced budget target is better understood by the public than a somewhat arbitrary threshold for the debt ratio (Danninger, 2002; Debrun, Epstein, and Symansky, 2008).

The rule’s way of considering the business cycle is also acknowledged. Instead of demanding the budget to be balanced ‘‘over the cycle’’ the rule is based on revenue and GDP forecasts for the next financial year only. Prediction errors, which often go along with medium-term forecasts of macroeconomic variables, are thus minimized (Danninger, 2002).

The centerpiece of the rule, the compensation account, is explicitly appreciated by Debrun, Epstein, and Symansky (2008). Thanks to this error correction mechanism the violation of the fiscal rule does not jeopardize the sustainability of public finances as the ultimate aim of a fiscal rule. Finally, the rule is commended for being enshrined in the constitution and for the fact that the escape clauses are listed conclusively in the law (Danninger, 2002).

However, there is also some criticism. Danninger (2002) points out that the Hodrick-Prescott filter comes up with an endpoint bias that is not convincingly solved in the Swiss debt brake. Regarding long-lasting recessions he, moreover, states that redeeming deficits from the compensation account might become problematic under such circumstances as it takes place in a situation of under-utilization of capacities. Fiscal policy acts in a procyclical way then. Similarly, fiscal policy becomes procyclical in a long-lasting recession because the HP filter adjusts the GDP trend downwards if the effective GDP decreases over a longer time period. This reduces the output gap and forces fiscal policy to be more restrictive although the recession is still running (Danninger, 2002).

Finally, it is argued that investment spending will become too low under the rule. Since it is contended that investment spending positively impacts on economic growth it might be helpful or even necessary to foster investment spending. Moreover, politicians tend to cut investment spending when forced to cut overall spending. Yet, the difficulty in terms of investment spending is its precise definition. There is a danger that consumption spending is declared as investment spending which would undermine the functioning of the rule. The Swiss Federal Finance Administration (SFFA) argues that a special fund might be put in place in the case of a need of investment spending. However, this again leads to a decrease in transparency (Colombier, 2004, 2006).

Before turning to the empirical method, we briefly look at the introduction phase of the rule (2003–2005). The budget of the year 2003 and the medium-term financial plan 2004–2006 were both the first of its kind under the new rule. To calculate the expenditure ceiling, the SFFA used the classical Hodrick-Prescott (HP) filter at that time. Because of the endpoint bias, the expenditure ceiling was inappropriately low and the budget ran into a deficit.⁵ Moreover, the adjustment factor $k$ does not react very precisely to corrections of the GDP forecast. So at the time of the budget preparation in year $t - 1$ , only a GDP forecast is available, that is, the calculation of the expenditure ceiling is based on this GDP forecast and the respective trend GDP value. If then later that year the GDP forecast is being revised, the GDP trend changes due to the relatively large impact of the last trend GDP value on the overall trend. However, the ratio of $Y_{t}^{T}$ over $Y_{t}$ is affected only weakly because both values $Y_{t}^{T}$ and $Y_{t}$ have changed. This was precisely the case in 2003 when the adjustment factor $k$ did not appropriately react to the revision of the GDP forecast (Colombier, 2004). Consequently, the federal government introduced a transition period, and the debt ratio increased in the years 2003–2005 in spite of the introduction of the debt brake.

The Synthetic Control Estimator

In principle, the method builds on the trajectory of an outcome variable for one single unit of interest (treated unit) in a time period ahead of an intervention (pre-intervention period) as well as the period following this intervention (post-intervention period). The method then aims at approximating the path of the outcome variable of the treated unit in the pre-intervention period as best as possible with that path of a (‘‘synthetic’’) control group. Eventually, the trajectory of the outcome variable of this synthetic control group is depicted for both periods. So the path of the control group in the post-intervention period shows how the outcome variable of the treated unit would have evolved without the intervention. The difference between both trajectories in the post-intervention period thus provides the synthetic control estimator (SCE).⁶

A stylized representation of the method is given in Figure 1 where the solid line represents the outcome variable (e.g., the budget balance) of the treated unit and the dashed line represents the outcome variable of the control group, respectively. The effect of the intervention on the outcome variable is then represented by the gray triangle.

Figure 1.

Stylized representation of the synthetic control method.

More formally the method is based on a sample of $i + j$ units. While $i$ depicts the treated unit, $j = 1, \dots, J$ is a positive number of potential comparison units. All these units together constitute the donor pool. Time runs from $t = 1, \dots, T$ , whereas $T_{p r e}$ is a positive number of pre-intervention time periods. $T_{p o s t}$ , in contrast, entails a positive number of post-intervention time periods.⁷ Thus it holds that $T = T_{p r e} + T_{p o s t}$ .

The approximation of the outcome variable of the treated unit is not obtained by simply using the weighted average of the outcome variables of the synthetic control group but it is based on some $k = 1, \dots, K$ (economic) characteristics that are predictor variables of the outcome variable. $X_{i}$ then is a ( $K$ $\times$ 1) vector of these predictor variables for the treated unit in the pre-intervention period. The predictors are averaged over this period. Likewise, the ( $K$ $\times$ J) matrix $X_{J}$ covers the same $K$ (averaged) predictor variables for the same time period but for all $J$ potential comparison units in the donor pool. Since the predictor variables differ in their ability to predict the outcome variable a weight $v_{k}$ is assigned to every predictor variable $k$ based on a regression.

In a next step, the synthetic control group is obtained by compiling the ( $J$ $\times$ 1) vector of weights $W = (w_{1}, \dots, w_{J})^{'}$ . The control group thus is the weighted average of the comparison units. These weights are required to be non-negative and sum up to one so that $0 \leq w_{j} \leq 1$ and $w_{1} + \dots + w_{J} = 1$ holds. Usually some units in the donor pool are not chosen to be part of the control group which means that the respective comparison units get a zero $w$ -weight. The $W$ -vector is then chosen to as best as possible approximate the matrix $X_{J}$ with the vector $X_{i}$ . That is, it minimizes the difference $v ‖ X_{i} - X_{J} W ‖$ . The ultimate aim of the optimization is to minimize the root-mean-square prediction error (RMSPE)⁸ of the outcome variable in the pre-intervention period. The vector $W^{*}$ is then used to depict the outcome variable of the synthetic control group.⁹

The estimation of the intervention effect is given by the difference between the $T_{p o s t}$ -values of the outcome variable of the synthetic control group and that of the treated unit, respectively. So for any point in time $t > T_{p r e}$ , the SCE can be written as

S C E_{t} = Y_{i, t} - \sum_{j = 1}^{J} w_{j}^{*} Y_{j, t} .

(3)

In the subsequent analysis, we calculate the treatment effect of the overall post-intervention period

T_{p o s t}

as the average treatment effect.¹⁰

In order to explore whether the treatment effect occurred by chance, we conduct cross-unit placebo tests. The intervention then is assigned to and the analysis is conducted for every single country in the donor pool. “Placebo” then means that the introduction of a debt brake is imposed to a country although it did not happen.¹¹ According to the case of Switzerland, we choose 2002 to be the year of intervention. Subsequently, we can calculate the post-intervention RMSPE as well as the ratio of the post-intervention RMSPE over the pre-intervention RMSPE for every country (post–pre-ratio). The intervention effect can be labeled significant if the treated unit comes up with a post–pre-ratio that is large relative to the ratios of the donor pool countries.

To quantify the statistical significance we refer to the concept of the p-value and calculate the probability to find a country in the donor pool with a post–pre-ratio of the same size or even larger than the post–pre ratio of Switzerland. This probability is then compared to a certain significance level (e.g., $α =$ 5 percent) to evaluate the statistical significance of the result obtained in the analysis.¹²

Basically, we restrict our sample to OECD member states so that countries with substantial economic or political structural breaks different from Switzerland and the OECD states are excluded. Similarly, we restrict the post-intervention period to five years in order to protect the trajectory of the synthetic control group against influences unrelated to the intervention.

One particular characteristic that hardly can be measured but is particularly discussed when it comes to Switzerland is fiscal conservatism. The reasoning here is that the introduction of the fiscal rule does not cause an improvement of the budget balance by itself because the Swiss citizens and with them the Swiss politicians inherently prefer low spending and slight borrowing. To test for fiscal conservatism Dafflon and Pujol (2001) create an index based on the referenda dealing with fiscal affairs regarding the central level which were submitted to voters in the Swiss cantons. Voters are classified to be conservative if they show a high acceptance rate in favor of tax rate increases, spending cuts or measures that aim at reducing the deficit. They find that the demand for public borrowing is lower in a canton the more conservative the residents are. Consequently, Krogstrup and Waelti (2008) are exclusively testing for voter preferences as omitted variable. In their analysis covering the Swiss cantons the impact of a fiscal rule on the budget balance remains significant even when controlling for the fiscal preferences of the voters. Similarly, Funk and Gathmann (2013) augment their estimation of the effect of direct democracy on government spending by the inclusion of some measures that capture voter preferences. They find that this relationship gets considerably weaker in terms of size and statistical significance once voter preferences are incorporated.

Hence it can be questioned what the separate impact of the Swiss fiscal rule is. Or put differently, we risk to get a treatment effect that is too large if we do not capture fiscal conservatism. More importantly, we would not be able to quantify the real size of the treatment effect as we cannot separate the treatment effect of the fiscal rule from the impact of fiscal conservatism. Again, we follow Abadie, Diamond, and Hainmueller (2010) in these arguments. If citizens and politicians in Switzerland behave fiscally conservative this should affect the outcome variable in the pre-intervention period. Since this method is based on matching it would assign a positive $w$ -weight to countries with citizens and politicians who also behave in a fiscally conservative way or countries that are at least able to approximate the conservative spending behavior in Switzerland. This means that the fiscal conservatism of Switzerland is captured by the approximation of the pre-intervention period and the treatment effect can be attributed to the introduction of the fiscal rule.¹³ Beyond that, Funk and Gathmann (2013) clearly show that fiscal conservatism is rather stable over time. In this respect, its inclusion would not help to considerably improve the pre-intervention RMSPE as the SCE emulates the variation of the outcome variable over time. In a parametric regression model, the oblivion of a relevant covariate leads to the mis-specification of the econometric model (Angrist and Krueger, 1999). Under the matching estimator, in contrast, the covariates mainly contribute to the fulfilllment of the unit homogeneity assumption (Gangl and DiPrete, 2004). Thus, we are convinced that our results are unbiased even if fiscal conservatism is not explicitly controlled for.

Eventually, we shall consider reverse causality and the existence of spillover effects. Reverse causality is present if a certain shape of the treatment effect caused the introduction of the Swiss debt brake. As the treatment effect occurs undoubtedly after the policy intervention, we feel confident that reverse causality is not a serious obstacle in our context. Regarding the existence of spillover effects, there must not be an impact of the intervention on the outcome variable of the treated unit or the donor pool units during the pre-intervention period. The same holds for the outcome variable of any donor pool unit in the post-intervention period. An effect of the intervention on Switzerland’s outcome in the pre-intervention period could come up as an anticipation. However, anticipation effects are captured by the approximation similar to fiscal conservatism. We do not know a case in which the Swiss balanced budget rule influenced the spending behavior of another country in the pool during the time period considered in this article.

Data

First and foremost, we are interested in the effect of the Swiss debt brake on the budget balance. Thus, we look at the total federal government budget balance which is total federal government revenue net of total federal government expenditure, either in national currency or in current prices (OECD, 2007). The budget balance is expressed as percentage of GDP and is adjusted by the output gap to obtain the CAB (see the “Baseline Results” section).

Additionally, we are interested in the effect on the federal government debt as percentage of GDP.¹⁴ Moreover, we look at general government debt as percentage of GDP.¹⁵

In order to find out whether compliance with the rule is achieved by an increase in government revenue or a decline in government expenditure, we also tested the effect of the debt brake on total revenue (as percentage of GDP) as well as total expenditure (as percentage of GDP). However, these variables vary considerably over time and thus an approximation in the pre-intervention period is not possible to a satisfying extent. We also subtract welfare spending from government expenditure which did not prove useful.

In choosing variables that are good predictors of debt development, we first consider the natural logarithm of real GDP per capita as Wagner’s law claims that government spending is higher the more developed the economy is. Secondly, we consider the annual growth rate of GDP, the growth of unemployment as well as the rate of unemployment as debt development depends on the business cycle. Moreover, the growth of GDP reflects a country’s capability to repay debt. Thirdly, the natural logarithm of total population is included as it captures the idea that some government spending categories are affected by economies of scale. Finally, we include the share of population aged sixty five and above (as percentage of total population) as well as the sum of imports and exports which serves as an indicator for the openness of the economy. Whereas the former impacts on health spending, the latter affects social spending as governments of open economies are incentivized to compensate the losers of globalization (Rodrik, 1998). These predictor variables are always supported by the lagged outcome variable of certain years which often is the first and the last year of the pre-intervention period. Among these predictors we search for a combination that best supports the approximation of the Swiss data.

As stated above, we restrict our sample to OECD member states. However, data is not consistently available for all of these nations and all variables of interest. Thus, we restrict our sample to countries for which data is available for the budget balance, the central government debt ratio and the general government debt ratio. Eventually, our sample contains Australia, Austria, Belgium, Canada, the Czech Republic, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Luxembourg, the Netherlands, New Zealand, Norway, Portugal, the Slovak Republic, Spain, Sweden, the United Kingdom, and the United States of America. Hence our donor pool basically consists of twenty-two countries.¹⁶

For the countries Israel, South Korea, and Poland, we have data for two of the dependent variables. We choose Israel to be part of the analysis of the budget balance and the general government debt ratio as data is available for these two series. Likewise, we use data of South Korea and Poland to examine the effect of the debt brake on the central government debt ratio as well as the general government debt ratio.

Empirical Analysis

Baseline Results

At first glance the variable “government expenditure” is the rule’s target figure. However, expenditure is related to predicted revenue of the same fiscal year rendering the Swiss fiscal rule a balanced budget rule. As the Swiss debt brake takes the business cycle into account, we focus on the CAB and calculate the component of the business cycle by applying the HP filter to available GDP data. Unlike the SFFA, we can apply the classical HP filter (with a smoothing factor $λ = 100$ ) as we do not face an endpoint bias of the trend-GDP time series (see “The Design of the Swiss Balanced Budget Rule” section). Our calculation of the CAB is in line with the standard methodology used, for example, in the European Union fiscal framework (Mourre et al., 2013) and can be expressed by

C A B = \frac{B}{Y} - \frac{Y - Y^{T}}{Y^{T}}

(4)

with

\frac{B}{Y}

as the nominal budget balance to GDP ratio and

\frac{Y - Y^{T}}{Y^{T}}

as the percentage of the output gap on the trend-GDP. Regarding the minuend of the equation, we use GDP data in current prices since revenue and expenditure are also expressed in current prices. Consequently, we also use (trend) GDP data in current prices in the subtrahend of the equation. Concerning the minuend, price changes should not affect the public finance variables seriously as they are expressed as percentage of GDP. Similarly, price movements should largely be canceled out in the subtrahend as the nominal business cycle element

Y - Y^{T}

is also expressed as percentage of nominal GDP. Of course, this assumes that revenue and expenditure are affected by price changes just as the output variable. The output gap usually is multiplied by the factor

ϵ

that captures the cyclical sensitivity of the budget balance. However, since the SFFA in their calculation of the budget balance implicitly assumes the revenue elasticity to be one (Colombier, 2004), we refrain from using an overall elasticity different from one.

To test the effect of the rule’s introduction on the budget balance, we start the pre-intervention period in 1995 as revenue and expenditure data starts in 1995 for most of the countries in the sample. We rely on the consistent sample explained in the “Data” section which is enriched by revenue and expenditure data of Israel. This leaves us with twenty-three countries in the donor pool. Among the predictor variables, we employ the natural logarithm of total population and the growth rate of unemployment in the baseline model, as these variables proved most helpful to minimize the pre-intervention RMSPE. Additionally, we use three years of the lagged CAB (1995, 1999, and 2000). As the budget of the year 2003 was the first governed by the debt brake, we declare the year 2002 to be the year of intervention. The post-intervention period starts in 2003 and ends in 2007, as it shall cover no more than five years (see “the Synthetic Control Estimator” section).

The $w$ -weights are given in Table 1. All countries in the donor pool were assigned a positive $w$ -weight and are thus part of the synthetic control group. In fact, this makes Synthetic Switzerland a real “synthetic” control, where the evolution of the budget balance of Switzerland is not replicated by a small number of countries with large $w$ -weights each. Instead, many countries contribute to Synthetic Switzerland with small $w$ -weights. And although four countries (Ireland, Israel, the Netherlands, and the Slovak Republic) stand out with a relatively large $w$ -weight, none of these weights is larger than 30 percent.¹⁷

Table 1.

Budget Balance— $w$ -Weights.

State	$w$ -Weight	State	$w$ -Weight	State	$w$ -Weight
Australia	0.009	Germany	0.004	Norway	0.006
Austria	0.006	Ireland	0.286	Portugal	0.014
Belgium	0.008	Israel	0.145	Slovak Republic	0.237
Canada	0.013	Italy	0.007	Spain	0.007
Czech Republic	0.003	Japan	0.007	Sweden	0.006
Denmark	0.011	Luxembourg	0.009	United Kingdom	0.007
Finland	0.005	Netherlands	0.181	United States of America	0.009
France	0.015	New Zealand	0.005

The predictor balance is given in Table 2. The column ‘‘Treated’’ displays the value of the respective predictor variable for Switzerland averaged over the pre-intervention period (vector $X_{i}$ ). Likewise, the column ‘‘Synthetic’’ provides the mean value of the same predictor variable for Synthetic Switzerland (matrix $X_{J}$ ). The table shows that the resemblance between Switzerland and Synthetic Switzerland is very close in the pre-intervention period with respect to the predictor variables. This is reflected by the pre-intervention RMSPE which amounts to 0.43.

Table 2.

Budget Balance—Predictor Means.

Variable	Treated	Synthetic
lnPOP	15.779	15.783
Unemp growth	$- 4.33$	$- 4.328$
CAB 1995	$- 1.292$	$- 1.297$
CAB 1999	0.408	0.414
CAB 2000	$- 2.426$	$- 2.422$

Abbreviations: CAB = cyclically adjusted budget balance; POP = population.

Figure 2 exhibits the result of our baseline model. Except for small deviations of the trajectory of Switzerland from the trajectory of Synthetic Switzerland, the accuracy of the approximation is very good. For the years after the rule’s introduction, we see a clear uplift of the budget balance of Switzerland compared to the budget balance of Synthetic Switzerland that represents the absence of the debt brake. That is, the budget balance would have performed worse without the fiscal rule. Moreover, we find a clear level of difference between Switzerland and Synthetic Switzerland. The synthetic control group copies the increase as well as the following decrease in the budget balance of Switzerland in the post-intervention period. The average treatment effect amounts to 3.68 percentage points.

Figure 2.

Budget balance.

Cross-Unit Placebo-Studies

To evaluate the significance of our main result, we conduct cross-unit placebo studies. Figure 3a shows the results of the placebo tests. The gray lines represent the yearly deviation of the budget balance of each country in the donor pool from the budget balance of its synthetic counterpart. The black line shows the respective gap calculated for Switzerland. Thus, from 2003 onwards, the black line shows the treatment effect of the Swiss fiscal rule. However, Figure 3a also shows that the evolution of the budget balance cannot be well reproduced for some donor pool countries in the pre-intervention period. For example, Synthetic Ireland is dominated by Norway. The RMSPE of Ireland amounts to 4.9 percentage points. In general, it is difficult to find a combination of donor pool countries that can reproduce the budget balance of a country under scrutiny if this country shows up with relatively large budget balance values. A large treatment effect of a hypothetical fiscal rule in the post-intervention period thus might be caused by an imprecise approximation in the pre-intervention period. Treatment effects of placebo studies derived from imprecise approximations should not be used to evaluate the treatment effect of the treated unit derived from a precise approximation.

Figure 3.

(a) Budget balance—gaps—twenty-three states; (b) budget balance—gaps—sixteen states; (c) budget balance—gaps—fourteen states.

Therefore, we first disregard placebo studies of donor pool countries that show up with a pre-intervention RMSPE five times the pre-intervention RMSPE of Switzerland. The results of the remaining sixteen placebo studies are given in figure 3b. Finally, we exclude all placebo studies that show up with a pre-intervention RMSPE four times the pre-intervention RMSPE of Switzerland. The results of the remaining fourteen placebo studies are given in Figure 3c. Measured against this distribution of placebo studies, Switzerland comes up with the largest or the second largest treatment effect. But even this test of significance rests on some placebo studies with a lack of fit clearly larger than the lack of fit we obtain for Switzerland. Reducing the number of placebo studies further on, however, renders the test of significance impossible as a small number of donor pool countries per se leads to large p-values. Thus, we look at the post–pre-ratio instead. We calculate the post–pre-ratio for Switzerland as well as for every single country in the donor pool and sort these ratios according to size. By looking at RMSPE ratios, we can take the quality of the approximation into account. Additionally, we do not need to define a threshold that separates between considered and disregarded placebo studies as such a definition is somewhat arbitrary.

Figure 4 shows the RMSPE ratios for all twenty-three donor pool countries. Switzerland clearly stands out with the largest RMSPE ratio. The probability to find a country in the donor pool with an RMSPE ratio of the size of Switzerland or even larger is 1/24 $=$ 0.042.

Figure 4.

Budget balance—root-mean-square prediction error (RMSPE) ratios.

Robustness Tests on Influential Observations

Next, we test the response of the intervention effect we derived from our baseline estimation to changes in the composition of the synthetic control group. That is, we rerun the baseline model whereby we remove one single country from the donor pool ahead of the iteration. Subsequent to the run, we add this unit again. By doing this, we can assess to what extent our result is driven by the contribution of this particular country to Synthetic Switzerland.

First, we remove and add countries that received a $w$ -weight smaller than 20 percent. Figure 5 reveals that among the countries with a $w$ -weight smaller than 20 percent the exclusion of countries with a $w$ -weight smaller than 2 percent does not alter the main result in any way. These countries, for instance, Australia, Austria, Belgium, or Canada, are represented by the solid gray lines and represent 15.1 percent of Synthetic Switzerland. The trajectory of Synthetic Switzerland remains unaltered both in the pre-intervention period and in the post-intervention period. The dashed gray lines, in contrast, represent the exclusion of Israel and the Netherlands, respectively. The accuracy of the approximation of the budget balance in the pre-intervention period is comparable to the approximation that uses all donor pool countries. Figure 5 shows that the omissions of Israel and the Netherlands raise the estimated effect of the fiscal rule on the budget balance. It amounts to 5.82 (Israel) and 5.71 (the Netherlands) percentage points, respectively.

Figure 5.

Budget balance—sensitivity analysis.

Next, we remove and add the Slovak Republic and Ireland. Figure 6a reveals that the precision of the approximation of the budget balance trajectory deteriorates markedly if data of the Slovak Republic is not exploited to build Synthetic Switzerland. The pre-intervention RMSPE increases to 1.25 and the average treatment effect decreases to 1.86 percentage points. Similarly, the precision of the approximation of the budget balance trajectory aggravates considerably if data of Ireland is not employed (see Figure 6b). The pre-intervention RMSPE increases to 1.08.

Figure 6.

(a) Budget balance—sensitivity analysis w/o Slovak. Rep.; (b) budget balance—sensitivity analysis w/o Ireland.

Model Alteration I: Reducing Lagged Outcome Values

In order to further assess the robustness of our main result we run two modifications of the baseline model. First, we reduce the number of lagged outcome values in order to increase the variance of the model (Kaul et al., 2015). Second, we include another predictor variable. Regarding the first modification, the pre-intervention period again runs from 1995 to 2001 and the post-intervention period again runs from 2003 to 2007. We still employ the consistent data sample as explained in the “Data” section which is enriched by data of Israel. Among the predictor variables, we still use the natural logarithm of total population and the growth rate of unemployment. In the context of this model modification, however, we use the values of the lagged CAB of the years 1996 and 2000.

The $w$ -weights are given in Table 3. All countries in the donor pool were assigned a positive $w$ -weight and many countries contribute to Synthetic Switzerland with a $w$ -weight around 2 percent or smaller. Once more, Israel, Ireland, and the Netherlands come up with major $w$ -weights. The $w$ -weight of the Slovak Republic, in contrast, is now much smaller compared to the baseline model and amounts to nearly 2 percent.

Table 3.

Budget Balance—Modification 1- $w$ -Weights.

State	$w$ -Weight	State	$w$ -Weight	State	$w$ -Weight
Australia	0.015	Germany	0.011	Norway	0.014
Austria	0.017	Ireland	0.206	Portugal	0.034
Belgium	0.02	Israel	0.241	Slovak Republic	0.018
Canada	0.018	Italy	0.01	Spain	0.013
Czech Republic	0.008	Japan	0.016	Sweden	0.014
Denmark	0.023	Luxembourg	0.076	United Kingdom	0.01
Finland	0.021	Netherlands	0.17	United States of America	0.013
France	0.018	New Zealand	0.015

The predictor balance is given in Table 4. It shows that the means of the predictor variables of Switzerland and Synthetic Switzerland are very close together. The pre-intervention RMSPE accounts for 0.606 and is thus only slightly larger than the pre-intervention RMSPE of the baseline model.

Table 4.

Budget Balance—Modification 1—Predictor Means.

Variable	Treated	Synthetic
lnPOP	15.779	15.795
Unemp growth	$- 4.33$	$- 4.343$
CAB 1996	0.599	0.607
CAB 2000	$- 2.426$	$- 2.424$

Abbreviations: CAB = cyclically adjusted budget balance; POP = population.

Figure 7 exhibits the result. Regarding the first two years (1995 and 1996) and the last two years (2000 and 2001) of the pre-intervention period the synthetic control group is quite well able to approximate the evolution of the budget balance of Switzerland. For the years in-between the approximation works less precise leading to a pre-intervention RMSPE larger than the pre-intervention RMSPE of the baseline model. In general, however, Synthetic Switzerland mimics the trajectory of the Swiss budget balance. With respect to the post-intervention period, we once more see a level difference between the evolution of the Swiss budget balance and the counterfactual trajectory that shows how the budget balance would have evolved without the debt brake. The trajectory of Synthetic Switzerland mimics the increase of the budget balance in the years 2003 and 2004 as well as its decline in the years 2004–2007. However, at a lower level. The average treatment effect of the debt brake on the Swiss budget balance amounts to 2.58 percentage points.

Figure 7.

Budget balance—modification 1.

Cross-Unit Placebo-Studies

In order to evaluate the significance of our result, we once more conduct cross-unit placebo studies by imposing the introduction of a debt brake on each country in the donor pool in 2002. The replication of the evolution of the budget balance in the pre-intervention period works well only for a few countries in the donor pool. Consequently, remarkable gaps can be observed in Figure 8a.

Figure 8.

(a) Budget balance—modification 1—gaps—twenty-three states; (b) budget balance—modification 1—gaps—seventeen states; (c) budget balance—modification 1—gaps—sixteen states.

Thus, we disregard placebo studies that feature a pre-intervention RMSPE five times (Figure 8b) or four times the pre-intervention RMSPE of Switzerland (Figure 8c). Whereas the first preclusion leaves us with seventeen placebo studies, the second exclusion leaves us with sixteen placebo studies in the donor pool. Both figures show that some placebo studies result in an intervention effect larger than the effect of the debt brake on the budget balance in Switzerland. However, these intervention effects occur in conjunction with an imprecise emulation of the budget balance in the pre-intervention period illustrated by the deviation of the gray lines from the zero line in Figure 8b and c. As we cannot rely on intervention effects originating from imprecise approximations, we once more look at RMSPE ratios. As previously done, we calculate the RMSPE ratio for Switzerland as well as for every country in the donor pool and sort these ratios according to size.

Figure 9 shows the result of the calculation of the RMSPE ratios for all twenty-three donor pool countries. Again, the RMSPE ratio of Switzerland is the largest among the set of RMSPE ratios although the difference between the RMSPE ratio of Switzerland and the adjacent RMSPE ratio (Ireland) is small. The probability to find a country in the donor pool with an RMSPE ratio of the size of Switzerland or even larger amounts to 1/24 $=$ 0.042.

Figure 9.

Budget balance—modification 1—root-mean-square prediction error (RMSPE) ratios.

Robustness Tests on Influential Observations

Next, we test the response of the intervention effect of our modified model to changes in the composition of Synthetic Switzerland. First, we singularly remove and add countries that received a $w$ -weight smaller than 5 percent.

Figure 10 reveals that among this group of countries the exclusion of one single country does not alter the trajectory of Synthetic Switzerland considerably. This holds for the evolution of Synthetic Switzerland both in the pre-intervention period and in the post-intervention period. Taken together, these countries represent 30.8 percent of Synthetic Switzerland (Figure 11a to d).

Figure 10.

Budget balance—modification 1—sensitivity analysis.

Figure 11.

(a) Budget balance—modification 1—sensitivity analysis w/o Luxembourg; (b) budget balance—modification 1—sensitivity analysis w/o the Netherlands; (c) budget balance—modification 1—sensitivity analysis w/o Ireland; (d) budget balance—modification 1—sensitivity analysis w/o Israel.

Luxembourg was assigned a $w$ -weight of 7.6 percent. Except for a deviation in the last year of the pre-intervention period, removing this country from the donor pool also does not alter the evolution of the budget balance of Synthetic Switzerland noticeably. Other countries with a relatively large $w$ -weight are the Netherlands, Ireland, and Israel. Unsurprisingly, the accuracy of the approximation in the pre-intervention period deteriorates clearly if data of these countries is not used to build Synthetic Switzerland. This especially holds for the year right before the introduction of the debt brake. The pre-intervention RMSPE increases to 0.86 (the Netherlands), 1.04 (Ireland), and 1.06 (Israel) percentage points, respectively. Whereas the omission of the Netherlands increases the average treatment effect to 4.1 percentage points the omissions of Ireland and Israel reduce the average treatment effect to 1.54 and 1.87 percentage points, respectively.

Model Alteration II: Including Another Predictor Variable

In terms of the second model modification, we include the share of population aged sixty-five and above (as percentage of total population). Just as in the previous model modification, we employ two values of the lagged outcome variable (1996 and 2001). The pre-intervention period again runs from 1995 to 2001 and the post-intervention period again runs from 2003 to 2007. We employ the consistent data sample as explained in the “Data” section which is enriched by data of Israel. Among the predictor variables, we still use the natural logarithm of total population and the growth rate of unemployment.

The $w$ -weights are given in Table 5. In contrast to the baseline model and the previous model modification, merely five countries were selected to replicate the budget balance of Switzerland. Whereas Japan, Luxembourg, and Ireland contribute to Synthetic Switzerland with $w$ -weights of intermediate sizes, Portugal determines nearly 60 percent of the evolution of Synthetic Switzerland.

Table 5.

Budget Balance—Modification 2— $w$ -Weights.

State	$w$ -Weight	State	$w$ -Weight
Ireland	0.142	Norway	0.037
Japan	0.108	Portugal	0.58
Luxembourg	0.134

The predictor balance is given in Table 6. In general, the means of the predictor variables of Synthetic Switzerland come close to the means of the respective predictor variable of Switzerland. The pre-intervention RMSPE amounts to 0.704 and, hence, is larger than the pre-intervention RMSPE of the previous model modification.

Table 6.

Budget Balance—Modification 2—Predictor Means.

Variable	Treated	Synthetic
lnPOP	15.779	15.825
Unemp growth	$- 4.33$	$- 4.454$
POP65	15.037	14.915
CAB 1996	0.599	0.599
CAB 2001	$- 2.595$	$- 2.580$

Abbreviations: CAB = cyclically adjusted budget balance; POP = population.

Figure 12 exhibits the result. It can be seen that the accuracy of the approximation is a little less exact compared to the previous model modification. This, for example, holds for the years 1997 and 1998. Regarding the post-intervention period, there once more exists a level difference between the evolution of the Swiss budget balance and the trajectory of Synthetic Switzerland over the entire time span. The average effect of the debt brake on the Swiss budget balance amounts to 4.73 percentage points.

Figure 12.

Budget balance—modification 2.

Cross-Unit Placebo-Studies

To evaluate the significance of this result, we once more conduct cross-unit placebo studies. Similar to the placebo studies of the baseline model and the previous model modification, Figure 13a reveals a lack of fit in the pre-intervention period for some donor pool countries.

Figure 13.

(a) Budget balance—modification 2—gaps—twenty-three states; (b) budget balance—modification 2—gaps—eighteen states; (c) budget balance—modification 2—gaps—fifteen states.

Thus, we disregard placebo studies that show up with a pre-intervention RMSPE five times (Figure 13b) or four times (Figure 13c) the pre-intervention RMSPE of Switzerland. Whereas the first exclusion leaves us with eighteen states, the second leaves us with fifteen states. Similar to the significance test of the baseline model, we find that Switzerland shows up with the second largest effect of debt brake on the budget balance (see Figure 3b and c). Again, this evaluation of significance is based on treatment effects of placebo studies derived from imprecise approximations. Consequently, we focus on the RMSPE ratios of Switzerland as well as the countries of the donor pool.

Figure 14 shows the ranking of post–pre-ratios comprising all twenty-three donor pool countries. We find, that the RMSPE ratio of Switzerland is the largest among all RMSPE ratios. The probability to find a country in the donor pool with an RMSPE ratio of the size of Switzerland or even larger thus once more amounts to 1/24 $=$ 0.042.

Figure 14.

Budget balance—modification 2—root-mean-square prediction error (RMSPE) ratios.

Robustness Tests on Influential Observations

Eventually, we test the response of the intervention effect obtained from this model modification to changes in the composition of Synthetic Switzerland.¹⁸ The omission of Norway increases the accuracy of the approximation slightly as the pre-intervention RMSPE decreases to 0.672. Regarding the post-intervention period, however, the estimated treatment effect of the debt brake on the budget balance remains almost unaltered. This also holds for the $w$ -weights of Ireland, Japan, and Luxembourg. Whereas the $w$ -weight of Portugal decreases to 0.504, Belgium now contributes to Synthetic Switzerland with a $w$ -weight of 0.081.

Next, we remove and add Japan and Luxembourg. The pre-intervention RMSPE increase to 0.73 and 0.826, respectively. As a result, the treatment effect of the Swiss debt brake on the budget balance remains unchanged in the case of Synthetic Switzerland without Japan where Germany contributes to Synthetic Switzerland with a $w$ -weight of 0.108. In the case of Synthetic Switzerland without Luxembourg the average treatment effect decreases to 4.04 percentage points where the $w$ -weights of Ireland, Japan, and Norway increase marginally and the Czech Republic enters the donor pool with a $w$ -weight of 0.041.

The omission of Ireland leads to an increase of the pre-intervention RMPSE to 1.047. Although the treatment effect on the budget balance is smaller, it is still in existence over the entire post-intervention period. The $w$ -weights of Japan (0.081), Norway (0.015), and Portugal (0.303) decline, whereas the $w$ -weight of Luxembourg increases to 0.183. The Netherlands enter the donor pool and contribute to Synthetic Switzerland with a $w$ -weight of 0.418. The average treatment effect amounts to 2.79 percentage points.

Finally, we remove Portugal which raises the pre-intervention RMSPE to 0.782. Although Portugal represents almost 60 percent of Synthetic Switzerland the increase of the RMSPE turns out smaller than the increase of the pre-intervention RMSPE in the case of Synthetic Switzerland without Ireland. Comparable to this case, the omission of Portugal reduces the treatment effect on the budget balance to 3.09 percentage points. However, the treatment effect prevails over the entire post-intervention period. In this case, the $w$ -weights of Ireland and Japan increase to 0.199 and 0.152, respectively. Whereas Luxembourg and Norway were not assigned a positive $w$ -weight, Belgium (0.346), Israel (0.126), and the Netherlands (0.178) enter the donor pool and contribute to Synthetic Switzerland.

In summary, we find distinct treatment effects of the Swiss debt brake on the CAB in almost every robustness test and model specification, all of whom indicating a lower budget balance in the absence of the fiscal rule.

The Effect of the Swiss Debt Brake on Government Debt

As a fiscal rule aims at achieving sustainable public finances, we examine the effect of the Swiss debt brake on the debt-to-GDP ratio. In the first three years of the post-intervention period (2003–2005), the debt ratio of Switzerland (both central government and general government) exceeds the debt ratio of the years just before the introduction of the Swiss debt brake (as well as the debt ratio of Synthetic Switzerland). We ascribe this effect to the rule’s introduction phase explained in “The Design of the Swiss Balanced Budget Rule” section. From 2006 onwards, the debt ratio of Switzerland declines continuously and falls below the level of 2001.

Starting from 2007 (central government) and 2008 (general government), respectively, the debt ratio of Synthetic Switzerland increases remarkably. Actually, the trajectories of Switzerland and Synthetic Switzerland intersect in the post-intervention period. This holds for both the central government debt ratio and the general government debt ratio. Due to this intersections, we neither can identify a treatment effect nor can we run placebo studies. For details, see Pfeil and Feld (2016). We provide these analyses on request.¹⁹

On the Introduction of Fiscal Rules Among the Comparison Units

In order to get an unbiased result, the outcome variable of any country of the synthetic control group must not be affected by a shock during the time period investigated. The absence of a shock is a necessary condition and it also applies to introductions of fiscal rules in countries of the synthetic control group. Basically, this would require the removal of all countries from the synthetic control group that introduced a fiscal rule themselves.

This applies to sixteen out of twenty-three countries in the donor pool. That is, Australia, Austria, Canada, Czech Republic, Denmark, Finland, France, Israel, Japan, Luxembourg, Norway, Portugal, Slovak Republic, Spain, Sweden, and the United Kingdom either introduced (changed) a national fiscal rule (central or general government level) or joined the Maastricht Treaty (IMF, 2009; Budina et al., 2012). As a result, the donor pool would comprise seven countries which renders the test of significance impossible. Hence, we remove countries from the donor pool in order to test the response of the intervention effect to changes in the composition of the synthetic control group.

When we drop the countries with a $w$ -weight smaller than 5 percent (Australia, Austria, Canada, Czech Republic, Denmark, Finland, France, Japan, Luxembourg, Norway, Portugal, Spain, Sweden, and the United Kingdom) the treatment effect remains unchanged in every single case (see Figure 5). Thus we argue that the application of the respective data does not bias the estimation of the treatment effect of the Swiss fiscal rule on the budget balance. When we look at the four countries with a relatively large $w$ -weight, we find that Ireland and the Netherlands did not introduce a fiscal rule in the respective time period. Dropping Israel leads to a pre-intervention RMSPE of 0.593 which is only slightly poorer than the RMSPE of the baseline model including Israel. The average treatment effect amounts to 5.82 percentage points and thus is larger the treatment effect of the baseline model including Israel (see Figure 5). This is reasonable as Israel expanded its Deficit Reduction Law by a provision to limit the growth of central government expenditures in 2004 (IMF, 2009; Budina et al., 2012). We argue, that the expansion of the Deficit Reduction Law might have improved the budget balance rather than worsened it.

The omission of the Slovak Republic raises the pre-intervention RMSPE from 0.43 to 1.25 and decreases the average treatment effect to 1.86 percentage points (see Figure 6a). However, the Slovak Republic did not introduce a fiscal rule at the national level in the respective time period but joined the European Union in 2004. Similar to the case of Israel, we argue that joining the Maastricht Treaty might have improved the budget balance rather than worsened it. Additionally, the Slovak Republic was assigned a w-weight of merely 0.018 in the context of the first model alteration which also reveals a distinct treatment effect of the Swiss debt brake.

Conclusions

The Swiss federal debt rule has become a role model for other countries. First, Germany adapted a similar rule to its constitutional framework and the Eurozone countries, in the fiscal compact, accepted to introduce their own national debt rules. Second, other countries, for example, Israel, followed the ideas of the Swiss debt rule in reforming their own legislation. In all these countries, the effectiveness of fiscal rules is politically challenged. Even in the Swiss case, it is difficult to provide evidence regarding the effectiveness of the debt rule using traditional econometric methods.

This article provides an evidence for the effectiveness of the Swiss balanced budget rule by relying on the Synthetic Control Method. We find that the rule improved the CAB by about 3.7 percentage points of GDP on average in the five years after its introduction. This treatment effect is robust regarding the reduction of lagged outcome values, the inclusion of a further predictor variable as well as several changes of the composition of the synthetic control group. Concerning the effect on the debt ratio of both the federal government level and the general government level we cannot provide results. This is due to an intersection of the debt trajectories in the post-intervention period. Overall, the evidence provided in this article suggests a causal impact of the Swiss federal debt brake on federal finances. With lower structural deficits, the federal government could return to sound public finances.

Footnotes

Appendix: Variables

CAB

cyclically adjusted budget balance as percentage of GDP, source: OECD.Stat and own calculations.

CGDR

total central government debt ratio (as percentage of GDP), stocks: outstanding amounts, source: OECD.Stat., for the United Kingdom we use total liabilities excluding insurance technical reserves as percentage of GDP, source: OECD Fiscal Decentralisation Database.

GDP growth

annual growth rate of GDP, source: OECD.Stat.

GGDR

general government debt ratio (as percentage of GDP), source: Historical Public Debt Database.

lnGDP

natural logarithm of GDP per head, US$, constant prices, constant PPP’s, reference year 2005, source: OECD.Stat and own calculations.

lnPOP

natural logarithm of total population, source: World Development Indicators and own calculations.

Pop65

population aged 65 and above as percentage of total population, source: World Development Indicators.

Trade

imports of goods and services plus exports of goods and services as percentage of GDP, source: World Development Indicators and own calculations.

Unemp growth

annual growth rate of unemployment, source: OECD.Stat—Economic Outlook No. 88 database and own calculations.

Acknowledgements

We thank Zareh Asatryan and Benedikt Fritz for helpful comments and insightful discussions. Christian Pfeil thanks Sabine Pfeil for enduringly supporting his academic work. We gratefully acknowledge comments and suggestions from participants at the Research Lunch of the Faculty of Economics and Behavioural Sciences of the University of Freiburg, a seminar of the German National Academy of Sciences (Leopoldina), Mannheim, a conference of the European Center for the Study of Public Choice (ECSPC), Rome, and the eeecon-lecture of the Faculty of Economics and Statistics of the University of Innsbruck.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Christian F. Pfeil

Lars P. Feld

Notes

Author Biographies

Christian F. Pfeil works in the budget department of the Ministry of Social Affairs, Health and Integration of Baden-Wuerttemberg, Germany. He received his doctorate in economics from the University of Freiburg, Germany, in 2017. Previously, he worked as research assistant and teaching assistant at the Walter Eucken Institute, Germany, the University of Freiburg, Germany, and the University of Heidelberg, Germany.

Lars P. Feld is Professor of Economics at the University of Freiburg and the Director of the Walter Eucken Institute, Freiburg, Germany. He holds a degree from the University of St. Gallen, Switzerland, and taught at the Universities of Marburg and of Heidelberg before. From 2011 to 2021, he served as member and, finally, as chairman of the German Council of Economic Experts.

References

Abadie

. 2021. “Using Synthetic Controls: Feasibility, Data Requirements, and Methodological Aspects.” Journal of Economic Literature 59:391–425.

Abadie

Diamond

Hainmueller

. 2010. “Synthetic Control Methods for Comparative Case Studies: Estimating the Effect of California’s Tobacco Control Program.” Journal of the American Statistical Association 105 (490): 493–505.

Abadie

Diamond

Hainmueller

. 2015. “Comparative Politics and the Synthetic Control Method.” American Journal of Political Science 59 (2): 495–510.

Abbas

S. A.

Belhocine

ElGanainy

Horton

. 2010. “A Historical Public Debt Database.” International Monetary Fund, Working Paper No. WP/10/245.

Alesina

Passalacqua

. 2016. “The Political Economy of Government Debt.” In Handbook of Macroeconomics, edited by J. B. Taylor and H. Uhlig, 2599–2652. Saint Louis: Elsevier.

Alesina

Perotti

. 1995. “The Political Economy of Budget Deficits, International Monetary Fund.” IMF Staff Paper 42 (1): 1–31.

Angrist

J. D.

Krueger

A. B.

. 1999. “Empirical Strategies in Labor Economics.” In Handbook of Labor Economics, Volume 3A, edited by Ashenfelter, O., and D. Card, 1277–1366. Amsterdam: Elsevier.

Asatryan

Castellón

Stratmann

. 2018. “Balanced Budget Rules and Fiscal Outcomes: Evidence from Historical Constitutions.” Journal of Public Economics 167:105–19.

Baumann

Doennebrink

Kastrop

. 2008. “A Concept for a New Budget Rule for Germany.” CESifo Forum 9 (2): 37–45.

10.

Ben-Michael

Feller

Rothstein

. 2021. “The Augmented Synthetic Control Method.” Journal of the American Statistical Association 116:1789–803.

11.

Bohn

. 1998. “The Behavior of U.S. Public Debt and Deficits.” The Quarterly Journal of Economics 113 (3): 949–63.

12.

Bohn

Inman

R. P.

. 1996. “Balanced-Budget Rules and Public Deficits: Evidence from the U.S. States.” Carnegie-Rochester Conferences Series on Public Policy 45:13–76.

13.

Bruchez

P.-A.

2003. “A Modification of the HP Filter Aiming at Reducing the End-Point Bias, Swiss Federal Finance Administration.” Working Paper No. 2003/03, Bern.

14.

Budina

Kinda

Schaechter

Weber

. 2012. “Fiscal Rules at a Glance: Country Details from a New Dataset, International Monetary Fund.” Working Paper No. WP/12/273.

15.

Bundesministerium der Finanzen—BMF. 2008. Modell einer neuen Schuldenregel, Kommission von Bundestag und Bundesrat zur Modernisierung der Bund-Laender-Finanzbeziehungen, Kommissionsdrucksache 096.

16.

Bundesverfassung der Schweizerischen Eidgenossenschaft—BV. 2002. http://www.admin.ch/opc/de/classified-compilation/19995395/index.html.

17.

Burret

H. T.

Feld

L. P.

2018a. “(Un-)Intended Effects of Fiscal Rules.” European Journal of Political Economy 52:166–91.

18.

Burret

H. T.

Feld

L. P.

. 2018b. “Vertical Effects of Fiscal Rules: The Swiss Experience.” International Tax and Public Finance 25 (3): 673–721.

19.

Burret

H. T.

Schnellenbach

. 2013. “Implementation of the Fiscal Compact in the Euro Area Member States, German Council of Economic Experts.” Working Paper 08/2013.

20.

Colombier

. 2004. “Eine Neubewertung der Schuldenbremse.” Swiss Federal Finance Administration. Working Paper No. 2004/02, Bern.

21.

Colombier

. 2006. “Die Schweizer Schuldenbremse — nachhaltiger und konjunktur-gerechter als der neue Stabilitäts- und Wachstumspakt?” Schmollers Jahrbuch 126:521–33.

22.

Dafflon

Pujol

2001. “Fiscal Preferences and Fiscal Performance: Swiss Cantonal Evidence.” International Public Management Review 2 (2): 54–76.

23.

Danninger

. 2002. “A New Rule: The Swiss Debt Brake.” International Monetary Fund, Working Paper No. WP/02/18.

24.

Debrun

Epstein

Symansky

. 2008. “A New Fiscal Rule: Should Israel “Go Swiss?” International Monetary Fund, Working Paper No. WP/08/87.

25.

Eliason

Lutz

. 2018. “Can Fiscal Rules Constrain the Size of Government? An Analysis of the “Crown Jewel” of Tax and Expenditure Limitations.” Journal of Public Economics 166:115–44.

26.

European Commission. 2017. “Report from the Commission Presented under Article 8 of the Treaty on Stability.” Coordination and Governance in the Economic and Monetary Union, C(2017) I201 final, European Commission, Brussels.

27.

Eyraud

Debrun

Hodge

Lledó

Pattillo

. 2018. “Second-Generation Fiscal Rules: Balancing Simplicity.” Flexibility, and Enforceability, IMF Staff Discussion Notes 18/04, Washington, DC.

28.

Feld

L. P.

2010. “Sinnhaftigkeit und Effektivität der deutschen Schuldenbremse.” Perspektiven der Wirtschaftspolitik 11 (3): 226–45.

29.

Feld

L. P.

Kalb

Moessinger

M.-D.

Osterloh

. 2017. “Sovereign Bond Market Reactions to Fiscal Rules and No-Bailout Clauses—The Swiss Experience.” Journal of International Money and Finance 70:319–43.

30.

Feld

L. P.

Kirchgaessner

. 2001. “The Political Economy of Direct Legislation: Direct Democracy and Local Decision-Making.” Economic Policy 16 (33): 330–67.

31.

Feld

L. P.

Kirchgaessner

. 2008. “On the Effectiveness of Debt Brakes: The Swiss Experience.” In Sustainability of Public Debt, edited by R. Neck and J.-E. Sturm, 223–255. Cambridge/London: MIT Press.

32.

Feld

L. P.

Reuter

W. H.

. 2022. “The German ‘Debt Brake’: Success Factors and Challenges.” In Public Debt Sustainability: International Perspectives, edited by B. Poulson, J. Merrifield and S. Hanke, 117–135. Lanham et al: Lexington Books.

33.

Feld

L. P.

Schaltegger

C. A.

. 2010. “The Political Economy of Fiscal Policy: Time Series Evidence for the Swiss Federal Level Since 1849.” Public Choice 144:505–34.

34.

Finanzhaushaltsgesetz—FHG. 2006. http://www.admin.ch/opc/de/classified-compilation/20041212/index.html.

35.

Finanzhaushaltsgesetz—FHG. 2010. http://www.admin.ch/opc/de/classified-compilation/20041212/index.html.

36.

Firpo

Possebom

. 2018. “Synthetic Control Method: Inference, Sensitivity Analysis and Confidence Sets.” Journal of Causal Inference 6:20160026.

37.

Funk

Gathmann

. 2013. “Voter Preferences, Direct Democracy and Government Spending.” European Journal of Political Economy 32:300–19.

38.

Gangl

DiPrete

T. A.

. 2004. “Kausalanalyse durch Matchingverfahren.” DIW Berlin, Discussion Paper No. 401.

39.

Geier

. 2011. “The Debt Brake—the Swiss Fiscal Rule at the Federal Level.” Swiss Federal Finance Administration, Working Paper No. 2011/15, Bern.

40.

Gootjes

De Haan

. 2022. “Procyclicality of Fiscal Policy in European Union Countries.” Journal of International Money and Finance 120: 102276.

41.

Heinemann

Moessinger

M.-D.

Yeter

. 2018. “Do Fiscal Rules Constrain Fiscal Policy? A Meta-Regression Analysis.” European Journal of Political Economy 51:69–92.

42.

International Monetary Fund—IMF. 2009. “Fiscal Rules—Anchoring Expectations for Sustainable Public Finances.” Fiscal Affairs Department, Washington, DC.

43.

Kastrop

Snelting

. 2008. “Das Modell des Bundesfinanzministeriums für eine neue Schuldenregel.” Wirtschaftsdienst 88 (6): 375–82.

44.

Kaul

Klößner

Pfeifer

Schieler

. 2015. “Synthetic Control Methods: Never Use All Pre-Intervention Outcomes Together With Covariates.” MPRA Working Paper No. 83790.

45.

Kiewiet

D. R.

Szakaly

. 1996. “Constitutional Limits on Borrowing: An Analysis of State Bonded Indebtedness.” Journal of Law, Economics and Organization 12:62–97.

46.

Krogstrup

Waelti

. 2008. “Do Fiscal Rules Cause Budgetary Outcomes?” Public Choice 136 (1): 123–38.

47.

Luechinger

Schaltegger

C. A.

. 2013. “Fiscal Rules, Budget Deficits and Budget Projections.” International Tax and Public Finance 20:785–807.

48.

Moore

. 2005. “Slovakia’s 2004 Tax and Welfare Reforms.” IMF Working Paper WP/05/133.

49.

Mourre

Isbasoiu

G.-M.

Paternoster

Salto

. 2013. “The Cyclically-Adjusted Budget Balance Used in the EU Fiscal Framework: An Update.” European Commission, Economic Papers 478.

50.

Niepelt

Schaltegger

C. A.

. 2013. “The Swiss Debt Brake—Ten Years On.” Swiss Journal of Economics and Statistics 149:111–3.

51.

Organisation for Economic Co-operation and Development—OECD. 2003. Ireland, OECD Economic Surveys 2003/9.

52.

Organisation for Economic Co-operation and Development—OECD. 2004. Netherlands, OECD Economic Surveys 2004/9.

53.

Organisation for Economic Co-operation and Development—OECD. 2007. “National Accounts of OECD countries.” General Government Accounts, Volume IV, OECD Statistics.

54.

Pfeil

C. F.

Feld

L. P.

. 2016. “Does the Swiss Debt Brake Induce Sound Federal Finances? A Synthetic Control Analysis.” CESifo Working Paper No. 6044.

55.

Rodrik

. 1998. “Why Do More Open Economies Have Bigger Governments?” The Journal of Political Economy 106 (5): 997–1032.

56.

Rubin

D. B

. 1974. “Estimating Causal Effects of Treatments in Randomized and Nonrandomized Studies.” Journal of Educational Psychology 66 (5): 688–701.

57.

Salvi

Schaltegger

C. A.

Schmid

. 2020. “Fiscal Rules Cause Lower Public Debt: Evidence from Switzerland’s Federal Debt Containment Rule.” Kyklos 73:605–32.

58.

Schaltegger

C. A.

2002. “Budgetregeln und ihre Wirkung auf die öffentlichen Haushalte: Empirische Ergebnisse aus den US-Bundesstaaten und den Schweizer Kantonen.” Schmollers Jahrbuch 122:369–413.

59.

von Hagen

. 1991. “A Note on the Empirical Effectiveness of Formal Fiscal Restraints.” Journal of Public Economics 44:191–210.

60.

Yared

. 2019. “Rising Government Debt: Causes and Solutions for a Decades-Old Trend.” Journal of Economic Perspectives 33:115–40.