Design of fiscal consolidation packages and model-based fiscal multipliers in Croatia

Design of fiscal consolidation packages and model-based fiscal multipliers in Croatia

Milan Deskar-Škrbić and Darjan Milutinović – Croatian National Bank
Milan Deskar-Škrbić is also a Vice President of the Centre for Public Policy and Economic Analysis (CEA)

Originally published in Public Sector Economics


The widening fiscal deficits and the increase of public debt triggered by the COVID-19 crisis suggest that fiscal policy makers will have to engage in substantial fiscal consolidation in order to stabilize public finances in the mid run. However, the implementation of a fiscal consolidation package, if it is not properly designed, can be detrimental for growth and even lead to a self-defeating outcome. In order to avoid this undesirable scenario, fiscal policy makers should rely on growth-friendly consolidation packages. The design of growth-friendly fiscal consolidation packages requires an understanding of the size of multipliers of different fiscal instruments. Thus, in this paper we provide the first deeper insights into the size of model-based disaggregated fiscal multipliers in Croatia. For this purpose, we have built a semi-structural macro-fiscal model of the Croatian economy and used Croatia’s experience during the fiscal consolidation episode under the excessive deficit procedure (EDP) to retrieve fiscal multipliers, analyse the design of the policy package and provide model-based evaluation of the macroeconomic effects of this consolidation episode. Our results indicate that the fiscal consolidation implemented during the EDP was not growth-friendly and that it was partially self-defeating. We hope that our results can help fiscal policy makers to avoid similar policy mistakes in future fiscal consolidations.

Keywords:  fiscal consolidation; fiscal multipliers; Croatia; economic modelling

JEL:  E62E12C50



The Covid-19 pandemic has prompted governments around the world to implement large fiscal stimulus packages in order to mitigate the economic costs of this global shock. Discretionary fiscal actions led to a structural increase of budget expenditures (e.g. subsidies to companies and transfers to households) and a fall in revenues (e.g. tax reliefs), thus widening fiscal deficits. The fall in economic activity, mostly triggered by the lockdowns imposed, has put additional pressure on fiscal balances through a mechanism of automatic stabilizers. These developments will result in a strong increase of public debt in both absolute and relative terms.

The deteriorating fiscal balances in 2020 suggest that fiscal policy makers will have to make a substantial fiscal effort to consolidate public finances and stabilize public debt dynamics in the mid run.1 This is especially important for Croatia on its path towards euro adoption. However, the implementation of fiscal consolidation packages, if not properly designed, could be detrimental for growth and thus even lead to self-defeating fiscal consolidation efforts. More precisely, poorly designed fiscal consolidation packages could additionally increase the public-debt-to-GDP ratio if the effects of a policy-induced fall in GDP2 outweigh the effect of fiscal adjustment (Gros and Maurer, 2012; Eyraund and Webber, 2013; Boussard, de Castro and Salto, 2013).
That is why fiscal policy makers should rely on growth-friendly fiscal consolidation packages. The basic idea behind growth-friendly fiscal consolidations is to design a fiscal consolidation package that ensures an improvement of fiscal balances, while minimizing negative short-term effects on growth (Cournède, Goujard and Pina, 2013). To put it differently, growth-friendly consolidation packages should be based on fiscal instruments with low fiscal multipliers as they could ensure the required fiscal effort with low economic costs.
The design of growth-friendly consolidation packages requires deep understanding of fiscal policy transmission mechanisms and knowledge about the size of fiscal multipliers of different fiscal instruments, i.e. so-called “disaggregated” fiscal multipliers (e.g. Boussard, de Castro and Salto, 2013; Cortuk, 2013). Thus, the key goal of this paper is to provide the first detailed insights into fiscal policy transmission mechanisms and the size of disaggregated fiscal multipliers in Croatia. Data on disaggregated fiscal multipliers can help fiscal policy makers in designing growth-friendly fiscal consolidation packages in the future.
Estimates of fiscal multipliers in Croatia have so far been exclusively based on vector autoregression methodology (VAR) (Šimović and Deskar-Škrbić, 2013; Grdović Gnip, 20142015; Deskar-Škrbić and Šimović, 2017). Although the VAR-based approach to the estimation of fiscal multipliers dominates fiscal literature, it does not allow a comprehensive analysis of the complex transmission mechanisms of fiscal policy and offers a limited framework for the analysis of macroeconomic effects of different fiscal policy instruments and feedbacks from macroeconomic to fiscal variables. This kind of analysis requires a more modeloriented approach.
References to model-based evaluation of the macroeconomic effects of fiscal policy in Croatia are scarce. To our knowledge there are only three papers investigating the effects of fiscal policy through the lenses of economic models on the macro level,3 but only for one fiscal instrument. Nadoveza, Sekur and Beg (2016) use the computable general equilibrium (CGE) model to analyse the macroeconomic effects of changes in the income tax burden. Deskar-Škrbić (2018) calibrates a small-scale dynamic stochastic general equilibrium (DSGE) model to simulate the effects of a government consumption shock on the Croatian economy. Bokan and Ravnik (2018) present the Croatian National Bank’s quarterly projection model (QPM) and simulate the effects of a change in the structural deficit.
This paper seeks to fill this gap in the literature by building and introducing for the first time a small semi-structural macro-fiscal model of the Croatian economy. This model allows us to retrieve disaggregated fiscal multipliers by comparing the realizations of macroeconomic variables in the no policy change and policy change scenarios, for different fiscal instruments. In addition, this model allows us to investigate the feedback from policy-induced changes in macroeconomic variables to fiscal balances and public debt. However, we want to emphasize that the model that we propose is not on a large enough scale to be able to capture all the relevant macroeconomic relations and the purpose of this model is not to describe the Croatian economy in detail but to give a basic framework for the analysis of the effects of fiscal policy. In addition, the proposed model (like other models in this class) is faced with various methodological limitations that we explain in detail in the main text.
The key challenge in the estimation of fiscal multipliers is to find episodes of exogenous changes in fiscal instruments, i.e. changes that are not directly related to business cycle developments.4 However, Croatia’s recent experience during the excessive deficit procedure (EDP) from 2014 to 2016 offers a capacious framework for the analysis in this sense. The EDP fiscal consolidation episode is interesting, as the fiscal authorities implemented a series of structural fiscal measures.5 These measures were reported in a transparent way and subjected to continuous post-hoc evaluations by the Commission. This fact enables a precise identification of structural measures that were not only announced by policy makers but actually implemented. In addition, fiscal policy actions were dominantly motivated by the supranational policy pressure under the EDP framework. In this sense, the implemented fiscal measures can be seen as exogenous, i.e. not directly related to the business cycle (Cugnasca and Rother, 2015; Górnicka et al., 2018).
Thus, in this paper we use these structural measures as input for our model and calculate the disaggregated fiscal multipliers for different revenue-side and expenditure-side fiscal instruments. Then, we use these findings to analyse whether the EDP fiscal consolidation episode was growth-friendly. Our results show that, although the consolidation was successful in terms of fiscal outcomes, the implemented fiscal package was not growth-friendly and the consolidation was partially self-defeating. In our counter-factual scenario, if our proposal for a growth-friendly fiscal consolidation package had been adopted, recession in Croatia would have ended as early as 2014, while public-debt-to GDP ratio would at the end of the consolidation period have been lower than it actually was. We hope that our findings can help fiscal policy makers to avoid similar mistakes in future fiscal consolidation episodes.
The paper is structured as follows. After the introduction, section 2 provides a brief literature review, with the focus on the literature on fiscal consolidations and the macroeconomic effects of fiscal policy. In section 3, we analyse the main characteristics of the EDP fiscal consolidation strategy and fiscal outcomes. In section 4, we present the structure of our small-scale semi-structural macro-fiscal model of the Croatian economy. In section 5, we present the main empirical results. The paper ends with the conclusion.