Current Projects

Sustainability of public finance

Should the pension system be balanced?

Fiscal distress worsened by demographic trends intensifies the urgency of carrying out pension reforms in order to insure the sustainability of social security systems in the future. We determine an optimal combination of fiscal instruments chosen by a benevolent government and compare social welfare in the case of balanced and unbalanced pension system. The analysis is based on the overlapping generations model (OLG) with infinite horizon initially developed by Yaari (1965) and Blanchard (1985) and extended further by Buiter (1988), Giovannini (1988), Weil (1989) and Bovenberg (1993). We extend the model of Heijdra and Bettendorf (2006) to investigate the unbalanced budget of a pension fund in a closed economy with an endogenous interest rate, which allows us to account for the effect of different economic characteristics on the capital accumulation. In the case of interior solution, income tax and social contributions act as perfect substitutes. In a general case, it is optimal to finance the deficit only by an income tax. Under an unbalanced pension system optimal fiscal policy leads to a higher social welfare due to the higher level of capital and the lower equilibrium level of public debt. When pension system is balanced, optimal social contributions are positive, while the optimal income tax rate is constant and does not depend on the population growth. In the case of an unbalanced pension system, the optimal fiscal policy includes a corner solution with zero social contributions and positive income tax. Income tax in this case is decreasing with population growth and increasing with higher labor productivity. Optimal set of fiscal instruments can be affected by the structure of the population: the share of non-Ricardian agents, who are liquidity-constrained and consume each period all their disposable income. Due to an inability to smooth consumption over periods they would be affected more by the fiscal policy. This would allow to analyze the trade-off between financing of pension expenditures via taxes which would decrease current savings of Ricardian agents but increase the consumption of non-Ricardian agents of the retirement. Optimal combination of measures (income tax, social contributions and pensions for both types of agents) would depend on the share of non-Ricardian agents.

Default Costs and the Fiscal Limit: Strategic vs. Excusable Default

Sovereign defaults should cause welfare losses for domestic households; otherwise sovereign borrowing would not be possible. Empirical literature supports this view, providing evidence that defaults lead to capital market exclusion, output losses (DePaoli et al. 2006, Sturzenegger 2004), fall in FDI flows (Fuentes, Saravia 2010) and reduction of foreign credit to the private sector (Arteta, Hale 2008). Theoretical models of sovereign default relate these default costs to the probability of sovereign debt repayments, asserting that the higher the costs, the lower the probability of default (Eaton,Gersovitz 1981,  Arellano 2008). We revisit the link between default costs and the risk of sovereign default and show that with high default costs two equilibria may co-exists. In the “good” equilibrium, households' investment in foreign assets is low, revenue from labor tax collection is high and the probability of default is low. In the “bad” equilibrium, households hedge against losses associated with default by investing large proportion of their income in foreign assets. This reduces labor supply, decreases revenues from labor tax collection and raises the probability of default. Therefore, high default costs may tighten fiscal limits and increase the probability of an excusable default. This result is at odds with the notion that that the higher the default costs, the lower the probability of default. At the same time, it is in line with the literature asserting that financial liberalization may raise risks of sovereign default.

Capital Flows, Default, and Renegotiation in a Small Open Economy: Bargaining in a Greek Tragedy

The post-2008 period focused attention on ‘twin-crises’. Banking crises may lead to sovereign crises where fiscal vulnerabilities are exacerbated by the extension of support for the banking system. We develop a model that describes private sector generated capital inflow that is used to finance investment and consumption expenditure. In the event of an economic contraction, the (convex) haircut on outstanding debt is negotiated, or bargained, centrally by the sovereign. Two results arise: the volume of debt and haircut rate are inefficient. In this setting the accumulation of capital achieves two goals. First, it generates suffcient optimism about future income to allow the debt market to function. Second, and counter-intuitively, it increases expected haircuts by raising the value of the outside option of complete default. These competing forces characterize the optimal balanced-budget macroprudential policy targeting capital investment.

Financial Repression and Public Finance in a Calibrated General Equilibrium Model

This paper uses a simple calibrated general equilibrium model to evaluate the impact of financial repression in the form of non-market placement of public debt. By imposing a requirement for households to hold public debt with a below-market rate of return the government distorts optimal household allocation. Financial repression proceeds as an indirect distortionary taxation, which decreases propensity to consume and increases labor supply. It crowds-out private capital, but has an ambiguous impact on output. The composition of government revenue from taxation changes as well. Tighter financial repression shifts Laffer curves for taxes on labor and consumption down, but increases revenue from capital income taxation. Total budget revenue increases, which allows financing more public goods. We describe the substitutability between financial repression and regular taxation. Abandoning financial repression (increasing the interest rate on public debt) requires a substantial increase in the capital tax rate, which provides a political reasoning for pursuing less visible financial repression. For the U.S., loosening the requirement for private agents to hold public debt should be compensated by heavier regular taxation to keep total budget revenue constant, while European governments can simultaneously decrease the regular tax burden and have smaller public debt to finance the same amount of public goods.

Non-Market Debt Placement: Social Security under Financial Repression

Modern financial repression in advanced economies does not rely on increasing seigniorage revenue, but mostly rests upon regulatory measures to enlarge the demand for public debt that delivers extremely low or negative real interest rate. In this paper we propose an extension of the overlapping generations model to question the optimality of financial repression in the form of non-market placement of the public debt in the captive pension fund. First, we analyze the choice of benevolent government. We show that financial repression and capital income taxation are not perfect substitutes. The optimal degree of financial repression depends on the growth rate of population. Moreover, the benevolent government chooses to confiscate some part of the pension wealth, which is equivalent to the imposition of the lump-sum tax. Second, we consider populist policymaker with an upward bias in government spending and downward bias in capital taxation. Populist government does not always confiscate the part of pension wealth, but imposes repression to finance extra budget deficit. This result is in line with political economy explanation of financial repression as a hidden taxation.

Information and macroeconomic policy

Fiscal and monetary policy interaction with uncertain preferences

Starting from the famous paper by Sargent and Wallace (1981), fiscal and monetary policy interaction has been always in the center of attention in academic literature. Despite the huge number of studies, there are some crucial white spots in this domain. For example, the interaction of the central bank and the government with uncertain preferences has been rarely studied before. For example, Kuznetsova and Merzlyakov (2015) develop a game-theoretical model, which incorporates uncertainty about the real policy effects and uncertainty about the preferences of the government. In this model, the authors study the equilibrium in a policy game, where the government and the central bank choose their actions simultaneously. Nevertheless, in reality monetary policy decisions typically require less time to be made and are usually taken at a higher frequency. Thus, the Stackelberg interaction with the government leading and the central bank following seem to be a more appropriate interaction design.

In our research, we are going to study the effects of multiplicative uncertainty and uncertainty about government preferences in a Stackelberg interaction of the central bank and the government. After finding the equilibrium in a general stylized model, we will test our results in a New-Keynesian model with strategic interactions of policymakers.

The value of public information in a two-region model

The social value of information has been broadly discussed in economic literature. Nevertheless, almost all existing studies deal with closed economies, leaving the issues of information provision in open economies aside. Our study fills this gap and elaborate a general two-region model, which captures three important characteristics of international markets: globalization of markets, segmentation of fundamentals and informational asymmetry between regions. This model allows for two types of spillovers between regions. The first spillover can be called strategic, as the strategic effects in private actions are global. The second spillover is informational. This spillover arises because the information published in one region is almost freely available to the agents in the other region. For this model, we derive the global and the regional welfare criteria and study social, regional and inter-regional value of information. The main contribution of this study to the literature is the close look on the welfare properties of information in open economies.

As a preliminary result of the study, we show that the effects of information in segmented economies differ significantly from its welfare properties in one-region models. More precisely, we explore the importance of inter-regional asymmetries for the optimal information structure in open economies and show that ignoring these asymmetries when elaborating the information policy may cause the welfare loss.

Thus, our research provides the new perspective in the discussion of optimal information policy. Nevertheless, the current version of the model is too general and hardly allows to get the concrete policy recommendations. To get the practical advises for the policy design, we will apply our methodology to a number of concrete examples, relevant for the international monetary domain.

Information games

Many empirical studies reveal great differences in information policies of central banks all over the world. Nevertheless, these differences have captured only a scarce attention in theoretical literature. Thus, their proper theoretical explanation is still missing. Our goal is to fill this gap. For this purpose, we develop a stylized two-region model, which captures three core international channels between countries. The first channel is the technological spillover between countries, which arises because of the correlation between the regional shocks. The second channel is the informational spillover, caused by the publication of relevant economic information by the policymakers in both regions. As far as these signals are public, they are equally observed by agents in both economies. The third channel is the international strategic complementarity. As all the agents act on the international financial market, they have the incentive to copy the actions of other agents not only in their home region, but also abroad. Thus, our model can be seen as a model of international beauty contest. In this model, we study non-cooperative communication games being played by regional policymakers.

The study has already brought the preliminary results. We show that the non-cooperative equilibrium is never characterized by no revelation. A full transparency outcome may be the equilibrium outcome and is then Pareto-optimal. From a normative point of view, no revelation may be Pareto-optimal: the social value of public information may be negative in international economies as well as in closed economies. Partial revelation schemes are possible outcomes but never Pareto-optimal.

In the further research, we will test our findings in a more precise micro-founded example of two-region Lucas-Phelps island economy from Myatt and Wallace (2014). The results of such study could be directly linked to the literature on international monetary games. Combining the communication tools with standard monetary policy tools appears to be a challenging but intriguing task for further research.

Effectiveness of verbal interventions

Verbal interventions are a very powerful tool of monetary policy all over the world. Until recently, the informational aspect of monetary policy in Russia has not been developed as it was in other inflation targeting countries. Nevertheless, in 2014 the Bank of Russia has improved considerably its communication strategy. Nowadays, the central bank applies a full branch of regular and irregular channels to communicate the economic outlook, the forecasts and the reasons behind its policy decisions. Thus, four years of the full-fledge informational policy provides us with the data set which is sufficient to evaluate the effectiveness of verbal interventions in Russia.

In this study, we construct an index, which describes the character, the content and the timing of verbal interventions. After that, we evaluate the influence of this index on the Russian stock and currency exchange markets and on inflation expectations in Russia.

At the preliminary stage, we have shown that the central bank verbal interventions affect the averages of returns of stock market indices and RUB/USD exchange rate but do not influence their volatility. In the future research, we will take into account the verbal interventions of the government representatives to compare the effects of fiscal and monetary verbal interventions on the financial markets. Moreover, we are going to apply our methodology to a high-frequency index of inflation expectations. Doing so, we will assess the possibility of the central bank and the government to anchor inflation expectations. After the evaluation of the effectiveness of verbal interventions in Russia, we will be able to give the recommendations about the optimal informational design of monetary and fiscal policy.

Empirical issues in macroeconomic modelling and forecasting

Do Consumers Really Follow a Rule of Thumb?  Three Thousand Estimates from 130 Studies Say “Probably Not”

A burgeoning literature investigates the effects of monetary and fiscal policy in a framework where a fraction of households neither save nor borrow, but follow the rule of thumb to consume their current income. The calibrated or prior value used for this share varies, but is usually substantial: for example, Drautzburg and Uhlig (2015) use 0.25, Leeper et al. (2015) use 0.3, Bilbiie (2008) and Kriwoluzky (2012) use 0.4, Erceg et al. (2006), Gali (2007), Forni (2009), Cogan (2010), Colciago (2011), and Furlanetto and Seneca (2012) use 0.5, while Andres et al. (2008) use 0.65. Models used by policymaking institutions to analyze fiscal stimulus typically assume 0.2-0.5, Coenen et al. (2012). Our investigation so far shows that the literature on the excess sensitivity of consumption to anticipated income growth, often cited as the motivation for the calibrations, is inconsistent with such values. We show that three factors combine to explain the mean excess sensitivity reported in studies estimating consumption Euler equations: the use of macro data, publication bias, and liquidity constraints. When micro data are used, publication bias is corrected for, and the households under examination do not face liquidity constraints, the literature implies no evidence for the excess sensitivity of consumption to income. Hence little remains for pure rule-of-thumb behavior. The results hold when we control for 45 additional variables reflecting the methods employed by researchers and use Bayesian model averaging to account for model uncertainty. We further investigate whether the estimates are systematically affected by the order of approximation of the Euler equation, the treatment of non-separability between consumption and leisure, and the choice of proxy for consumption.

How many variables are needed for a good forecast?

Accurate macroeconomic forecasts are extremely important for policy making. Central banks and government bodies monitor a large set of macroeconomic indicators to determine the policy. To take it into account, a model used for forecasting must be suitable for data-rich samples because large models might outperform low-dimensional ones by taking into account more potentially relevant information. However, can we be sure that forecasting with large-scale models is always more accurate than with small-scale models?

This paper evaluates the forecast performance of different types of models (both univariate and multivariate) on Russian data: BVARs, VARs, Lasso, ARIMA, ETS and TBATS We estimate multivariate models of different sizes and compare the accuracy of their out-of-sample forecasts with those obtained with random walk with drift. We show which model is the most accurate for a big sample of  Russian macroeconomic indicators. For some of them the best model is BVAR.  However, contrary to several other studies, we do not confirm that the relative forecast error monotonically decreases with increasing the cross-sectional dimension of the sample. In half of those cases where a BVAR appears to be the most accurate model, a small-dimensional BVAR outperforms its high-dimensional counterpart. As a robustness check we apply a Model confidence set algorithm. Our results suggest that though on a general basis the BVARs demonstrate better forecasting performance than competitive models, but for each indicator  there are usually at least several other specifications whose forecasting performance is not significantly inferior to that of BVARs.

External and internal sources of business cycles in oil-exporting countries: evidence based on Structural BVAR model

The paper aims at an identification of the main business cycles sources in oil-exporting economies. The contemporary macroeconomists agree that business fluctuations are the results of the different shocks hitting the economy. This paper estimates the relative importance of different kinds of shocks in a sample of export-oriented economies, and particular attention is paid on the influence of external shocks.

We use structural Bayesian VAR with sign restrictions to identify shocks. The structural identification with sign restrictions has become a very widely used tool in applied macroeconomic studies. However, in their recent study, Baumeister and Hamilton (2015a) show the drawbacks of the conventional approach including but not limited to implicit assumption of informative priors for impulse responses which does not disappears asymptotically.  We follow their paper and acknowledge the necessity of explicit prior beliefs in a structural identification of a VAR model.

In the paper we estimate a panel Bayesian SVAR model and  we apply a modification of Sims and Zha (1998) prior to parameters of a structural model. The oil shock is identified with explicit Bayesian restrictions as in Baumeister and Hamilton (2015b). Our results reveal that the impulse response functions and historical decomposition patterns comply with the macroeconomic theory and economic logic. The method used in the paper is a generalized version of the conventional one and, contrary to the conventional one, can be applied to the large BVARs as well.

Political economy of growth and development

The distribution of wealth and the forms of the political conflict on different stage of development process

During the transition from stagnation to growth, the outcome of political conflicts between different social classes affects the development path. Landowners and the emerging class of capitalist struggle for their influence on the choice of public policy, promoting modern sector development. Incumbent capitalists and new entrepreneurs have different preferences towards the entry barriers on markets. We study the effect of wealth concentration on the rates of economic growth and the intensity of political conflict between landowners, capitalists and entrepreneurs on different stages of development process. We build a two-sector unified growth model, in which agents invest in lobbying their interests.  The institutional outcome is modeled using public policy contest game with asymmetric agents. We analyze the joint dynamics of the capital accumulation, structural changes and the distribution of wealth and shows how the initial difference between countries in wealth distribution effects development path. Our hypothesis is that the effect of capital and land concentration differs on different stages of development process. During the first phase of conflict the concentration of capital promotes development, whereas on the later stage it decreases the pace of development through their effect on the barriers to entry. Moreover, we show that high initial concentration of land not only leads to lower pace of industrialization, but also leads to the permanent negative effect on economic growth through their effect on the distribution of capital on later stages of development.  We analyze historical narrative which supports our theoretical arguments.

Industrialization, Inequality and Political Conflict in a Unified Growth Model

If property rights are poorly secured crony relation i.e. the tight connection between the owners of major firms and the government are the primary informal mechanism to secure property. However, this institutional framework creates high entry barriers for outsiders. In these circumstances, voters often support parties that focus on redistribution policies, but do not provide reforms, leading to the elimination of entry barriers. We propose a theory that explains, why in a democracy the majority of voters can prefer this type of institutions. This paper develops a simple quality-ladder model with heterogeneous agents, which differ in their skills and wealth endowment. We show that if the policy space is two-dimensional, the wealthy elite and low skilled workers can form majority coalition, supporting the regime with high-entry barriers. In this case, the wealthy elite chooses rationally to agree on higher level redistribution, preferred by the least skilled agents. Moreover, the middle class becomes political outsiders. We compare the possibility of this outcome in representative democracy as well as in the model with ideological parties and prove that the electoral support of crony capitalism is more probable for countries with a low level of human capital and high income and skill inequality. The model is also able to explain different effects of democratization process on the institutional structure of the society.

Political economy of brain-drain 

Do dictatorships tolerate more brain drain than democracies? We consider the dynamic game in which skilled agents has three options: to become entrepreneurs, to compete for a political office and to emigrate. Under democracy, redistribution in favor of low-skilled workers can induce migration of high-skilled workers to countries where it is less costly to become an entrepreneur or enter politics. Under dictatorship, a rise in the entry threshold for politics can motivate high-skilled entrepreneurs toward migration. Political change is less likely in economies with a lower share of high-skilled workers, where the intensity of political competition increases in economies with a high share of high-skilled workers. In the second part of the model, we account for variation within democracies and dictatorships.  We provide a theoretical mechanism that shows why brain drain is higher in dictatorships than in democracies, and why soft dictatorships and redistributive democracies are the most likely to observe high brain drain rates compared to repressive dictatorships and elitist democracies.