A Fiscal Impact Measure for Austria (OeNB Bulletin Q2/25)

Mathias Moser, Lukas Reiss ¹

JEL classification: E60, H60

Keywords: fiscal policy, stimulus, consolidation, fiscal multipliers

Fiscal policy has been very expansionary in the early 2020s as a reaction to the economic distortions induced by the COVID-19 pandemic. In Europe, fiscal policy remained expansionary at least until 2022 when inflation soared amidst a dramatic increase in energy prices. Policies in Austria have been particularly expansive. According to estimates by the European Commission, the cyclically adjusted primary balance (CAPB) worsened by about 2 percentage points from 2019 to 2024 (blue bars in chart 1). The expansion from 2019 to 2022 was even above 4 percentage points, resulting in only half of this increase being phased out in 2023/24. This has raised several questions: What is the effect of such a large-scale stimulus (and the subsequent phase-out or consolidation) on GDP? And how does this exceptional fiscal policy stance compare to the past? To answer these questions, we have developed a new Fiscal Impact Measure (FIM).

The FIM shows how various fiscal instruments influence GDP growth over time. While closely related to fiscal stance indicators, such as the CAPB, our measure uses detailed budgetary components to derive fiscal impulses. This allows for a more nuanced analysis of growth impacts. On the revenue side, we rely on a comprehensive database of fiscal measures to gage the discretionary impulse of fiscal policy. On the expenditure side, we use deviations from trend GDP growth as a baseline for most variables. For some variables, especially public wages and social cash transfers, we use empirically observed deviations from indexation rules. In addition, we adjust individual items historically to implement recategorizations of the public sector and growth-neutral revenue/expenditure. This includes, for example, military investment with high import shares, which should not enter fiscal impulses.

Finally, we link these impulses to macro effects from the Austrian Quarterly Model (AQM). Based on this model, the growth elasticities from different revenue and spending categories can vary drastically in size. Using our fine-grained approach to map elasticities to measures thus allows for a more realistic depiction of growth impacts.

Chart 1

This paper is organized as follows: Section 1 discusses how fiscal adjustments can be defined in principle. Afterward, section 2 describes how we derive the size of fiscal adjustments and their impact on GDP. Section 3 shows how fiscal adjustments evolved over time and discusses their determinants. This is followed by some consistency checks on the estimated size of fiscal adjustments in section 4. Section 5 presents our main result, the GDP impact of active fiscal policies in Austria. Section 6 then addresses caveats of our approach. Finally, section 7 provides conclusions.

1 How to define fiscal adjustments?

To be able to gage whether fiscal policy is interacting with economic growth, it is key to pinpoint the fiscally “neutral” position. However, defining such a neutral position for fiscal policy is more challenging than for monetary policy, where this is rather clearly defined as a neutral rate of interest that creates an environment of stable inflation and output growth (Laubach and Williams, 2003). Of course, this neutral rate is not directly observable either. It is subject to change over time and typically needs to be estimated (Holston et al., 2017; Laubach and Williams, 2003). For fiscal policy, however, the definition of “neutrality” also strongly depends on the context. It could refer to intertemporal neutrality, as in not distorting economic agents’ decisions on future consumption or investment, or even to neutrality in the broader sense of fiscal sustainability (e.g., achieving a steady-state debt ratio).

In the context of economic growth, policy analysis has especially aimed at capturing the discretionary fiscal impact in relation to the business cycle. The IMF (2017), for instance, notes that “[a] neutral fiscal stance is one that allows automatic stabilizers to operate freely without discretionary measures, thus stabilizing output fluctuations without changing the structural balance.” Note that the IMF explicitly refers to the structural balance, while other institutions favor the cyclically adjusted (primary) balance as a reference point. Both approaches have in common that they rely on a relative concept with respect to potential output. In this view, fiscal policy is neutral whenever its revenue/spending paths do not deviate from growth potential. This can be considered a Keynesian perspective where it is assumed that a public sector whose CAPB aligns with potential growth does not cause demand pressures and thus avoids procyclical policy.

This aggregate view is, however, challenged by multiple factors. When shocks or structural changes cause endogenous and lasting responses in potential output growth, then holding the CAPB constant may worsen these effects, rendering fiscal policy non-neutral. A similar case can be made for situations where fiscal multipliers deviate from reference values, as was argued during the zero lower bound episode by Christiano et al. (2011). In such a case, growth effects of fiscal policy are not symmetrical and thus cause under- or overshooting by holding the CAPB constant.

Last, it could be argued that fiscal policy is neutral when it behaves “predictably”, that is, according to a set of given fiscal rules specific to a country’s historical and institutional concept. If, for example, pension expenditure rises more strongly than potential output – which is the case for most developed countries – this manifestation of structural change could still be argued to be neutral in the sense that it is expected (according to a different set of rules, e.g., share of pensioners), yet not directly related to potential output and the CAPB. A similar case could be made for the non-indexation of tax revenue to nominal parameters, which could only be discretionarily addressed by implementing singular tax reforms, as was done in Austria until 2023. Even if the windfalls in non-indexation years and the refund in reform years are non-neutral in the short run, they can be (partly) attributed to neutral policy if they align with longer-term policies.

In our study, we use fiscal impact in the tradition of the FIM developed by the Hutchins Center on Fiscal and Monetary Policy (2021) based on Cashin et al. (2018). We derive fiscal impacts on macroeconomic growth from fiscal impulses which describe the deviation of fiscal policy from an assumed neutral position. Following the discussion above, we base our FIM impulses on a combination of discretionary measures, differences from trend output or deviations from rule-based adjustments. We rely on detailed revenue and spending items to assess the discretionary component for the impacts. Conceptually, this is quite close to the way fiscal adjustments are measured in the new Stability and Growth Pact. One major difference is that we measure discretionary policy on social benefits in cash via discretionary measures and not via deviations from trend output.

For most expenditure categories, we define reference growth values, such as trend GDP growth or change in relevant deflators. We then interpret empirical variations from these benchmarks as discretion that causes a fiscal impulse. In general, the change in the CAPB (or other aggregate indicators) would provide reasonable evidence on whether fiscal policy is restrictive or expansive. It is, however, a crude measure for assessing the impact of fiscal policies on growth, as some fiscal instruments have a higher impact on GDP than others. For example, public investment is typically thought to have a higher multiplier than lump-sum payments to households. In addition, for some fiscal items it is difficult to deduce discretionary policy by just looking at the time series of related revenue or expenditure. This holds particularly true for taxes on profits or unemployment benefits. We therefore opt for a more fine-grained approach.

2 How to derive fiscal adjustments?

We estimate the impact of fiscal policies on GDP growth using the Austrian Quarterly Model of the Oesterreichische Nationalbank (OeNB; Leibrecht and Schneider, 2006). In this model, short-term effects are driven by the demand side. Therefore, we categorize fiscal instruments based on how they influence the demand-side composition of GDP:

$GDP\sim Private\ consumption + investment + public\ consumption + net\ exports$ (1)

Table 1

Table 1 describes the main channels through which fiscal items impact on final demand in detail. Essentially, public consumption and public investment are part of final demand and therefore have a direct impact on GDP. Taxes paid by households, transfers paid to households and taxes/subsidies on products affect real disposable household income, which in turn influences private consumption. Finally, taxes paid by corporations, subsidies and investment grants affect private investment through various channels.

Table 1 also lists the first-year multipliers of these fiscal instruments. They are highest for public consumption and investment and much lower for taxes and transfers. The reason for this is that fiscally induced increases in real disposable household income also increase savings, and higher grants or lower taxes for corporations also benefit investments which would have been carried out either way. Furthermore, fiscally induced increases in private consumption or investment also increase imports, which, in turn, decrease the impact on GDP. In contrast, the import content of government consumption is typically very low, which also contributes to its rather high multiplier. Like government consumption, public investment also directly affects demand, leading to a much higher multiplier than that for taxes and transfers. However, due to its higher import content, the multiplier for public investment is lower than that for government consumption.

Our analysis focuses on the impact of active (discretionary) fiscal policies only. We therefore exclude the automatic responses of the tax and transfer system to variations in inflation and real activity. We regard these automatic responses as neutral policy. Table 2 shows how we derive active policy for the various fiscal categories and to which FIM category (from table 1) we allocate them. In principle, it would be ideal to solely rely on bottom-up data on explicit policy measures. However, for categories like government consumption or investment such data are not available or not particularly reliable. In some sense, most government investment projects have a discretionary character, like building a new school or purchasing new trains. Yet, discretionary actions like these are happening all the time. Therefore, we compute the discretionary change in government investment and consumption by comparing its growth rates to that of real potential GDP (as estimated by the OeNB). In case of government consumption, the government can influence not only its real value, but also its deflator. Most importantly, hiring additional staff will increase the real value of public consumption, while higher overall wage increases will only affect the deflator. Only increases in the real value of government consumption have a direct effect on real GDP. However, as almost all government employees are residents of Austria, wage increases have a direct effect on households’ disposable income. ²

Table 2

For taxes and social benefits in cash, we can rely on detailed in-house data on discretionary measures. The data are based on official government estimates in most cases. One important exception is the treatment of adjustments to inflation. Pensions and public wages have taken past inflation as a reference for a long time (table 3). Deviations of agreements from these values are treated as discretionary measures. Tax brackets and tax credits of the progressive personal income tax as well as family and long-term care benefits are specified in nominal terms. Since the early 2020s, these parameters have also been indexed to past inflation. For the sake of consistency over time, we compute the size of fiscal measures as if these indexations had been in place over our whole sample period. This means that non-action with regard to these fiscal instruments in times of positive inflation is treated as a fiscal consolidation measure. We exclude tax deferrals (like those in 2009 and 2020) and collection peaks related to incentives for earlier tax payments (like those in 2001 and 2015) from the data. Such factors can have a very high temporary impact on the revenues from profit-related taxes, but their GDP impact tends to be negligible.

Table 3

It proves more difficult to comprehensively define discretionary measures for spending on subsidies, ³ investment grants and other current transfers to households. Therefore, we treat changes in the ratio of these items to nominal potential GDP as discretionary measures. Some other fiscal items are treated as GDP-neutral. Among others, this concerns interest spending, capital transfers for bank recapitalizations or contributions to the EU budget.

3 How large were fiscal adjustments?

As comprehensive fiscal data are only available from 1995 onward, we start our analysis in 1996. The first years in our sample were marked by sizable consolidation packages (chart 2). These were implemented in 1996 and 1997, allowing Austria to fulfill the fiscal criteria for joining the euro area (Lehner, 1996). The packages involved substantial increases in income taxes and cuts to social benefits (blue bars). After Austria’s loose fiscal policies were heavily criticized by the European Commission in early 2000, further significant consolidation steps were taken from mid-2000 to 2002 (Breuss et al., 2004a). Public consumption was cut (yellow bars) and there was base broadening in personal and corporate income taxes (blue and green bars). In 2004/05, a medium-sized tax reform was carried out (Breuss et al., 2004b), which lowered the corporate tax rate, modified the personal income tax scale (blue and green bars) and raised taxes on mineral oil and heating gas (pink bars).

The first, larger stimulus in our sample took place in 2008/09 (Köhler-Töglhofer and Reiss, 2009) and involved increases in social benefits and cuts in personal income taxes (blue bars). The motivation for this first stimulus was a surge in inflation, which was then followed by a large recession amidst the global financial crisis of 2008. This development was accompanied by rather loose policies concerning government consumption (yellow bars). ⁴

Chart 2

Austria was subject to an excessive deficit procedure (EDP) from late 2009 to early 2014. This led to a prolonged fiscal consolidation from 2010 to 2015 (Budgetdienst, 2016). The main ingredients of this package were public wage and pension increases below reference inflation (particularly in 2013) and base broadening of the personal income tax (blue bars). Furthermore, indirect taxes were raised (pink bars), and government consumption was restrained. Another important contribution came from the prolonged non-adjustment of income tax brackets and tax credits. The latter factor was a major motivation for a very large income tax cut implemented in 2016 (Schratzenstaller, 2015). This was followed by three years of rather neutral fiscal policies.

In early 2020, Austria was hit by the COVID-19 pandemic, which led to substantial lockdowns. The Austrian government paid out massive subsidies to compensate corporations (green bars) and to promote labor hoarding through short-time work (blue bars). This was accompanied by extremely high expenditure for COVID-19 testing (part of yellow bars). These expenditure items were largely unwound in 2022 and 2023 (Fiscal Advisory Council, 2021). However, energy prices started to soar in late 2021 amidst Russia’s preparations for its invasion of Ukraine. This led to substantial cuts in energy taxes and increases in subsidies to households and corporations to partly compensate higher energy bills (pink, blue and green bars). In 2024, these temporary measures largely expired (Fiscal Advisory Council, 2023). While expenditure for COVID-19 testing almost vanished, policies in other areas of government consumption and investment were rather loose (yellow bars). This came with many permanent, or at least very persistent, expansive measures in other areas. Most importantly, taxes on labor income were cut substantially and several permanent, expansive measures concerning pensions were introduced.

4 Did fiscal adjustments fit together with budgetary developments?

Fiscal measures obviously influence the size of the budget balance. Therefore, developments in the budget balance are an important check on whether our estimates of discretionary fiscal measures are sensible. As we excluded changes in interest payments from discretionary measures, we focus our analysis in this section on changes in the primary budget balance (black line in chart 3).

Chart 3

Overall, the most important drivers of changes in the primary budget balance are indeed discretionary measures (pink bars) and macro effects (dark green bars). The latter reflect the working of automatic stabilizers amidst macroeconomic fluctuations and are consistent with changes in the output gap estimated by the OeNB. Macro effects were negative in 2009 and 2020 when the two big recessions took place. They were significantly positive in years with relatively high GDP growth (e.g., during the late 1990s and the recovery phases from the two big recessions). In recent years, revenue windfalls have been sizable (yellow bars). For this paper, we define them as changes in the ratio of current taxes to GDP that cannot be explained by discretionary policy action. Possible reasons for such windfalls (positive sign) or shortfalls (negative sign) can be changes in the wage share or fluctuations in income components that are not part of GDP (like capital gains; for a more extensive discussion of revenue windfalls in Austria, see Bernhofer and Reiss, 2024).

We not only exclude the effects of fluctuations in real GDP from our estimates of discretionary measures, but also the effects of variations in deflators on government spending. The latter had particularly strong effects on the budget balance from 2022 to 2024. In 2022 and 2023, adjustments to social benefits and public wages based on past inflation led to a devaluation of these expenditure items amidst strongly rising inflation. These developments turned around in 2024 when inflation was falling, making up for the temporary positive effects on the budget balance (for a more detailed explanation, see Holler and Reiss, 2023). The most important other contributors to changes in the budget balance (light green bars) are listed in chart 3. They are all related to changes in capital spending other than gross fixed capital formation or investment grants, which have no impact on GDP growth.

5 How did fiscal adjustments affect GDP?

Chart 4 translates the fiscal impulses shown in chart 2 into the growth impact of various fiscal instruments. Thick black markers show the total fiscal impact (FIM) of policies for each given year in contrast to overall GDP growth (dotted line). As outlined in section 1, (i) the detailed fiscal adjustments are translated into impulses deviating from an assumed neutral position; and (ii) these impulses are then scaled using elasticities from the AQM to arrive at growth impacts. Note that while we calculate impulses from 1996 onward, the macro effects are based on multiannual responses. A discretionary impulse will thus exert a GDP effect for up to four years, with the largest adjustment typically occurring in the first year. For growth, we thus show a time series starting in 2000 which fully accounts for both current and previous policy actions.

When interpreting the fiscal impact, it is worth recalling the structure of fiscal impulses in Austria since the mid-90s: Corporate and price measures have contributed little to the impulses, and if so, especially during periods of large consolidation needs or the most recent crisis (see section 2). The overwhelming share of fiscal measures can be attributed to household taxes and transfers, as well as public consumption and investment. These are the same categories that are aligned with rather large fiscal multipliers. Consequently, the growth impacts are mostly driven by these two factors.

Chart 4

For most of the two decades preceding the COVID-19 pandemic, we find that the contribution of fiscal policy to growth was rather stable, ranging from –1 percentage point to +1 percentage point. The years following the dot-com bubble burst, Austria’s accession to European Monetary Union and the potential launch of an EDP have resulted in negative fiscal growth impulses, which coincided with low but rather stable growth rates. Similarly, the following subprime crisis in 2008/09 triggered an early expansive fiscal response, which was swiftly abated once growth rates rebounded from the 2009 slump.

This course correction in fiscal policy was, however, not directly related to the economic recovery. It was rather the breach of European fiscal rules that caused the opening of an EDP. This was followed by a long period of restrictive fiscal policy that turned procyclical and contractionary during the 2012/13 economic slowdown. According to our calculations, the fiscal impact was even the decisive factor in slightly pushing Austria into recession in 2013. An episode of procyclical, expansionary fiscal policy was provided by the 2016 income tax reform, which further accelerated the growth path after the 2013 slump.

With the onset of the COVID-19 pandemic and the most recent inflationary shock, there have been further instances of procyclical fiscal policies. As regards the COVID-19 pandemic, the delayed fiscal response was, of course, driven by the high uncertainty regarding its impact on economic activity. In addition, the sustained fiscal impulse seen throughout the catching-up period of 2021 was a result of high public consumption in relation to health expenditure. The current economic cooldown has been accompanied by slightly expansionary policies, though. In light of the large amount of permanent, expansionary measures from 2020 onward, we expect to see a negative growth stimulus in the coming years due to consolidation packages.

6 What are potential caveats of this approach?

The growth impacts we calculate in the previous section come with several caveats. We do not deem these to be structural issues with our approach, but rather factors that can influence the size of macro adjustments in certain years or periods. We tried to single out certain special effects that may bias our estimates, such as large military purchases with a high import share or reclassifications. Despite these efforts, we recommend using the FIM for longer periods of analysis and trends. Among the remaining caveats, we identify the following:

• The average multiplier of measures targeted at corporations and private investment is rather small. While plausible for the calibration of average shocks, certain measures may deviate in their growth effect. Some measures can provide a very high return on investment using small amounts, thus implying rather large multipliers. For other measures, by contrast, there is a high likelihood of strong bandwagon effects. We deem the amount of investment that would have been carried out anyway to be rather high in areas like renovation bonuses or broadly specified premiums for corporate investments.

• The size of multipliers may be endogenously affected by the economic cycle. Recent literature finds that fiscal multipliers are smaller in phases of expansion and potentially larger during recessions (Auerbach and Gorodnichenko, 2012; Bachmann and Sims, 2012; for a contrasting view, see Ramey and Zubairy, 2018). If this finding is applicable, our FIM would overestimate the fiscal impact during high growth phases and vice versa.

• The effect of measures may deviate from the budgetary impact, which can happen for multiple reasons. Policy changes that occur late in the fiscal year may cause an immediate budgetary impact, while economic transactions will only be affected with delays. An example of this is the introduction of large lump-sum payments such as the Austrian “Klimabonus” (climate bonus), which was typically paid out in September but may have been spent on consumption in the following year.

7 Conclusions

In this contribution, we have derived a Fiscal Impact Measure (FIM) for Austria. We construct the underlying fiscal impulses as deviations from neutral fiscal policy, which we define in a rule-based manner. Our analysis shows several major fiscal policy shifts: These include consolidation efforts in the late 1990s to meet euro area criteria, the relatively large stimulus measures during the 2008/09 financial crisis and the very prolonged consolidation period from 2010 to 2015. These shifts all seem tiny compared to the fiscal response to the COVID-19 pandemic and the subsequent energy crisis. From 2020 to 2022, fiscal measures have been implemented on a historically large scale. While many of them were temporary (e.g., subsidies for short-time work, foregone revenue and energy costs), a large share of the stimulus measures have been permanent (in particular, measures aimed at stimulating household income). When only looking at headline budget numbers, the large scale of these stimulus measures is partly masked by very large revenue windfalls.

Regarding the fiscal impact on growth, we find that fiscal policies were often countercyclical in the pre-COVID-19 era. The only notable exception in this period was the second half of the prolonged consolidation period after the 2008/09 financial crisis. While the large-scale fiscal stimulus in 2020 was clearly countercyclical, too, the subsequent, expansive policy measures in 2021 and 2022 were clearly not so.

Despite the focus on historic fiscal policy in this article, our FIM is well equipped to be used as an analysis tool in forecasting exercises. Since 2024, the macroeconomic forecast of the OeNB has included the FIM as a tool to gage the fiscal impact on macroeconomic conditions for upcoming years.

8 References

Auerbach, A. J. and Y. Gorodnichenko. 2012. Measuring the Output Responses to Fiscal Policy. In: American Economic Journal: Economic Policy, 4. 1–27.

Bachmann, R. and E. R. Sims. 2012. Confidence and the transmission of government spending shocks. In: Journal of Monetary Economics, 59. 235–249.

Bernhofer, D. and L. Reiss. 2024. Hohe Abgabenquote trotz Steuersenkungen? Die Rolle von steuerlichen Zufallseinnahmen. A&W Blog. November 8, 2024. Vienna.

Breuss, F., S. Kaniovski and G. Lehner. 2004a. Makroökonomische Evaluierung der Fiskalpolitik 2000 bis 2002. WIFO Monatsberichte 77(7). 557–571.

Breuss, F., S. Kaniovski and M. Schratzenstaller. 2004b. Steuerreform 2004/05 – Maßnahmen und makroökonomische Effekte. WIFO Monatsberichte 77(8). 627–643.

Budgetdienst. 2016. Umsetzung der Konsolidierungspakete und Offensivmaßnahmen ab 2011. Vienna.

Cashin, D., J. Lenney, B. Lutz and W. Peterman. 2018. Fiscal policy and aggregate demand in the USA before, during, and following the Great Recession. International Tax and Public Finance 25 (2018). 1519–1558.

Christiano, L., M. Eichenbaum and S. Rebelo. 2011. When Is the Government Spending Multiplier Large? In: Journal of Political Economy, 119(1). 78–121.

Fiscal Advisory Council. 2021. Bericht über die öffentlichen Finanzen 2020 bis 2025. Vienna.

Fiscal Advisory Council. 2023. Bericht über die öffentlichen Finanzen 2022 bis 2027. Vienna.

Holler, J. and L. Reiss. 2023. Quantifying the impact of the 2021–22 inflation shock on Austria’s public finances. In: Monetary Policy & the Economy Q4/22–Q1/23. OeNB. 117–130.

Holston, K., T. Laubach and J. C. Williams. 2017. Measuring the Natural Rate of Interest: International Trends and Determinants. In: Journal of International Economics, 108. 59–75.

Hutchins Center on Fiscal and Monetary Policy. 2021. The Hutchins Center’s Fiscal Impact Measure. Brookings Institution. Accessed on February 12, 2025. https://www.brookings.edu/articles/the-hutchins-centers-fiscal-impact-measure/

International Monetary Fund. 2017. Annual Report 2017: Promoting Inclusive Growth. Washington, D.C.: International Monetary Fund.

Köhler-Töglhofer, W. and L. Reiss. 2009. The Effectiveness of Fiscal Stimulus Packages in Times of Crisis. In: Monetary Policy & the Economy Q1/09. OeNB. 78–99.

Laubach, T. and J. C. Williams. 2003. Measuring the Natural Rate of Interest. In: Review of Economics and Statistics 85(4). 1063–1070.

Lehner, G. 1996. Budgetkonsolidierung prägt Bundesvoranschlag 1996 und 1997. WIFO Monatsberichte 69(5). 373–386.

Leibrecht, M. and M. Schneider. 2006. AQM-06: The Macro economic Model of the OeNB. OeNB Working Paper 132.

Ramey, V. A. and S. Zubairy. 2018. Government Spending Multipliers in Good Times and in Bad: Evidence from US Historical Data. In: Journal of Political Economy 126(2). 850–901.

Schratzenstaller, M. 2015. Steuerreform 2015/16 – Maßnahmen und Gesamteinschätzung. WIFO Monatsberichte 88(5). 371–385.

9 Annex

In the following, we give two examples of how we derive our fiscal measures for individual categories (section 2). Impulses from real public consumption are derived via comparing its growth rate (blue line in chart A1) to that of real potential GDP (pink line in chart A1). The difference is then multiplied by the share of government consumption in GDP (typically around 1/5). Therefore, impulses (gray bars) are expansionary when real government consumption is growing faster than real potential GDP.

Chart A1

Impulses from personal income taxes are derived bottom-up via the scoring of individual tax measures. The blue bars in chart A2 show the sum of these measures excluding the effects of tax deferrals. The pink bars show how much additional revenue is generated by the constant devaluation of tax credits and tax brackets. This effect is higher when inflation is higher. Impulses (gray lines) are given by the sum of these two factors.

Chart A2