Мамедли_резюме_ENG_финальная версия (Pension system and fiscal policy in an economy with heterogeneous agents)

National Research University Higher School of EconomicsMariam Oktaevna MamedliPENSION SYSTEM AND FISCAL POLICY IN AN ECONOMY WITHHETEROGENEOUS AGENTSPhD Dissertation Summaryfor the purpose of obtaining academic degreePhilosophy Doctor in Economics HSEAcademic supervisor:Professor Sergey E.

Pekarski, PhDJEL: E62, H55, H68Moscow – 2018Pension System and Fiscal Policy in an Economy with Heterogeneous AgentsJEL codes: E62, H55, H68The dissertation was prepared on the faculty of Economic Sciences, Department of TheoreticalEconomics, National Research University Higher School of Economics.Publications:The content of Chapters 1-3 is based on the following papers in order:1.Vlasov S. A., Mamedli M. O. (2017). Scenario analysis of pension system parameters in thecontext of resilience of Russian government finances, Russian Journal of Money andFinance. № 8. pp.

26-34.2.Mamedli, M. O. (2016). Analysis of Government Expenditure Multiplier Under Zero LowerBound: The Role of Public Investment. The Journal of Economic Asymmetries, Volume 14,pp. 103–111.3.Mamedli, M. O. (2017). Fiscal Policy and the Unbalanced Pension System, HSE EconomicJournal, vol. 21, no 1, pp. 114–144.2IntroductionMost OECD countries have conducted pension reforms recently in order to ensure bothfinancial sustainability of the pension systems and adequacy of pensions. This underlines theneed for an analysis of the consequences of pension reforms and future public expenditure onpension provision.Urgency of these reforms has been dictated by: (a) first, the challenges to financialsustainability of the existing pension systems, putting additional pressure on public finances(which worsened after the financial crisis of 2008–2009 and sovereign debt crisis); and (b)second, by the longevity coupled with sharply decreasing birth rates.

Despite undertakingpension reforms, public expenditure on pensions as a percentage of GDP is expected to risefurther in the near future in most OECD countries. According to OECD estimates, publicspending on pensions has increased from 6.7% of GDP in 2000 to 8.2% of GDP in 2013, beingthe largest part of social expenditure (18% on average of the total expenditure in 2013). 1 Longterm OECD forecasts for public spending on pensions suggest that this tendency would continuefor most OECD countries, with the spending growing in 21 countries and falling in 14.

As aresult, pension expenditures are expected to increase from 8.9% of GDP in 2013–2015 to 9.5%of GDP by 2050, reaching 10.9% in 2060.2The change of pension system characteristics such as higher retirement age or the rate ofsocial contributions may be considered as an alternative to standard consolidation measures ofpublic finance. Public pension systems also have a redistributive role, providing the minimumlevel of income at the retirement for low income households who were not able to save duringtheir working period (minimum pensions are present in 15 OECD countries).At the same time, an account for the existence of non-Ricardian agents (Galí et al., 2007),who consume all their disposable income and do not make any savings, is crucial in the analysisof fiscal policy (e.g.

Almeida et al., 2013). It allows investigating more carefully theconsequences of optimal fiscal policy and an economic impact it may have. The structure of thepopulation should be taken into account, since fiscal policy directly affects the consumption ofnon-Ricardian agents and, as a result, the other macroeconomic variables.1Pensions at a Glance (2017), OECD Social Expenditures Database (SOCX).The numbers for 2013–2015 may differ from the values presented in OECD Social Expenditures Database (SOCX)because of the different range of benefits covered and the definitions used (Pensions at a Glance, 2017).23Literature reviewTheoretical research of pension reforms can be classified by the type of pension systemunder consideration.

The first group covers pay-as-you-go (PAYG) reforms (e.g. Heijdra andBettendorf, 2006; Nickel et al., 2008; Karam et al., 2010; Kilponen et al., 2006; Castro et al.,2017; Almeida et al., 2013 a, b; Pierrard and Snessens, 2009; Marchiori and Pierrard, 2012,2015). Others (Borsch-Supan et al., 2006; McGrattan and Prescott, 2017) consider the switchfrom PAYG to a fully funded pension system. Both types of pension systems were analysed inMarchiori et al. (2011) and de la Croix et al. (2013). This research falls into the first category.As opposed to most studies on population aging where analysis is based on dynamiccalibrated computable general equilibrium (CGE) models (e.g.

Auerbach and Kotlikoff, 1987),Heijdra and Bettendorf (2006) apply an analytical approach, extending the model developed byBlanchard (1985). Blanchard (1985) proposes an overlapping generations model (OLG) incontinuous time with an infinite horizon, uncertainty about the time of death and laborproductivity declining with age. 3 As the model incorporates actuarially fair life insurance,proposed by Yaari (1965), this model is sometimes referred to as Blanchard-Yaari model.Blanchard-Yaari OLG model in continuous time can provide a more realistic notion of time thantwo or three period OLG models, allowing, at the same time, an analytical tractability. Althoughthe model can provide the interactions between agents of different age groups and captures thefinite horizon aspect of life, it cannot capture the life-cycle aspects of consumption and savingsbehaviour due to the age-independent mortality rate.

4 Their model was extended by Buiter(1988) and Weil (1989) to incorporate a non-zero population growth and by Bovenberg (1993),who analysed intergenerational and international distributional effects in an open economy.Heijdra and Bettendorf (2006) have extended this framework further by considering twotypes of pension reforms: a decrease in pensions and an increase of the retirement age. Bothreforms are followed by an adjustment of social contributions because Heijdra and Bettendorf(2006) consider a balanced pension system. The framework of Heijdra and Bettendorf (2006)and Nielsen (1994) was extended by Nickel et al.

(2008) by considering an unbalanced pensionsystem. They analyse three fiscal scenarios in an economy with a decreasing population (Buiter,1988): the suspension of the public pension system and a decrease in lump-sum labor tax; thesuspension of the public pension system and a decrease in distortionary corporate tax; or anincrease in the retirement age. Their results suggest that the adverse consequences of pensionreforms can be mitigated by appropriate taxation policies.

Thus, if the tax policy is introduced3Blanchard (1985) assumes constant death probability, which does not depend on age. This model is also called themodel of perpetual youth.4Blanchard (1985) points it out the main drawback of the developed approach.4promptly, the negative reform effect on the consumption could be offset while the public debtwould reach lower equilibrium levels.The main difference with the research presented in Chapter 3 is that Nickel et al.

(2008)consider the government as a non-maximising entity and investigate how the predeterminedchanges in policy instruments would affect the transition of the main macroeconomic variablesto the new equilibrium in an open economy. While in this research socially optimal fiscal policy(social contributions and income tax) is defined and compared with the optimal set of policyinstruments in equilibrium with both increasing and decreasing population in the closedeconomy with endogenous interest rate. The analysis of Heijdra and Bettendorf (2006) was alsoextended by considering two types of households: Ricardian and non-Ricardian agents (as inCastro et.

al., 2017; Almeida et al., 2013 a, b).An inclusion of non-Ricardian agents has become a common practice in the analysis offiscal policy. The papers which analyse structural fiscal reforms and demographic shocks withinthe GIMF framework (Global Integrated Monetary and Fiscal model) incorporate populationheterogeneity via Ricardian and non-Ricardian consumers (Almeida et al. 2013a, b; Castro et al.,2015; Karam et al., 2010) in order to capture the failure of Ricardian equivalence. However, asopposed to the current research, they do not consider an optimal choice of instruments by asocial planner or the impact of the share of non-Ricardian consumers on this choice.The motivation of considering Ricardian and non-Ricardian consumers arises from the factthat, although New Keynesian models can provide some insight into higher governmentmultipliers obtained empirically, they are subject to criticism for their inability to capture apositive response of consumption to increases in government spending observed empirically(Galí et al., 2007).