Aim of this PhD thesis and methodological issues
Three main dimension considered.
To be able to study the eﬀects of fiscal policy in this renewed context, several dimensions of fiscal policy should be taken into account. This PhD thesis focuses on three of them: the sign and the size of the fiscal multiplier, the spillover eﬀects in an integrated economic union in which members are linked commercially and financially, and the stabilizing properties of fiscal transfers schemes between member states in a monetary union.
The fiscal multiplier. The concept of fiscal multiplier is crucial in this thesis. First, chapter 1 discusses the eﬀects of fiscal policy on output so that the size of the output fiscal multiplier is central in this analysis. Moreover, throughout the thesis other kinds of multipliers are studied. Especially, this thesis aims at investigating the eﬀects of fiscal policy on employment and unemployment. As a consequence, the terms ”unemployment fiscal multiplier” and ”employment fiscal multiplier” are often present in all chapters.
Spillover eﬀects of fiscal policy. Chapters 3 and 4 investigate the eﬀects of fiscal policy in a monetary union. One important element are the possible spillover eﬀects that can arise when one member state implement a fiscal expansion or a fiscal contraction. Possible spillovers will influence the stabilizing eﬀects of fiscal policy at both the union and the member level but also influence the output fiscal multiplier produced in the member who implemented fiscal policy. Also, if spillovers are large, member states have to take into account these cross-border eﬀects for macroeconomic policy coordination. The open economy eﬀects of fiscal policy have been studied extensively since at least the IS/LM model. However, in the case of a monetary union, the sign and the size of the spillover eﬀects are still unclear. Chapter 3 aims at contributing to the recent literature which investigates this issue.
Fiscal federalism and fiscal transfers in monetary unions. The question of fiscal policy in monetary unions is closely related to the seminal works on optimal currency areas by Mundell (1961) or Mc Kinnon (1963) for instance. In presence of asymmetric shocks, price rigidities and a weak labor mobility, fiscal policy could be the eﬀective tool for stabilizing the macroeconomic activity at the union level and between member states. In reality, fiscal policy has been used extensively to stabilize output and employment but the way fiscal policy is implemented diﬀers greatly among the existing monetary unions. As an example, fiscal policy as a stabilization tool is quite centralized in the United States, with a large central budget and a sophisticated system of transfers between the central government and the member states. On the contrary, fiscal policy is rather decentralized in the Euro Area despite the presence of common rules, namely the Stability and Growth Pact and more recent extensions (the European semester, the ”two-packs”). Many economists argue for a deeper fiscal integration for the Eurozone and especially for the implementation of fiscal transfers between member states. Chapter 4 aims at investigating the stabilizing properties of such a fiscal transfers mechanism.
The eﬀects of fiscal policy on the labor market
If one considers that rises in government spending boost economic activity, namely that the output fiscal multiplier is superior to 0, it is likely that government expenditure shocks increase also employment. For instance, Monacelli, Perotti and Trigari (2010), using a SVAR approach with a Choleski decomposition, find that public expenditure shocks boost employment, vacancies and the job finding probability. However, the link between the eﬀects of fiscal expansions on output and on employment seems not be so certain. As pointed out by Ramey (2012): ”[…] most economists and policymakers would agree that job creation is at least as important a goal as stimulating output. In theory, one can use Okun’s law to translate GDP multipliers to unemployment multi-pliers. However, because of variations in the parameters of this ”law” over time, the advent of jobless recoveries, and the frictions involved in creating and fil ling jobs, the translation of output multipliers to employment or unemployment multipliers is not straightforward”. Ramey (2012) provides empirical evidence of the eﬀects of fiscal pol-icy on employment and unemployment for diﬀerent identification schemes and diﬀerent samples. The message is that rises in public expenditure tend to lower unemployment but that the positive eﬀect on employment is more due to more hired workers in the public sector than to more jobs in the private sector. In this paper but also in previ-ous studies, Ramey (2012) argues for a negative eﬀect of fiscal policy on private activity.
If the eﬀects of fiscal policy shocks on employment are unsure, most of studies conclude nevertheless in a positive eﬀect on employment. However, the response of unemployment to fiscal expansions is even less clear. In addition to uncertain eﬀects on employment, a rise in government expenditure tends to cause variations on the la-bor force participation. Monacelli, Perotti and Trigari (2010) estimate the eﬀects of a government consumption shock with a standard S-VAR approach (a Choleski decom-position) on a large set of labor market variables, especially on hours worked, civilian employment, unemployment, vacancies, the labor force, the real wage or the labor mar-ket tightness. The point estimates indicate that the labor force participation does not move significantly. In addition, hours worked, employment and the real wage increase so that the unemployment rate falls by 0.6 percentage points at the peak. Turning to a theoretical exercise, the authors show that a standard neo-classical framework with search and matching frictions on the labor market hardly reproduces these empirical findings, with a lower eﬀect on unemployment. More importantly, the introduction of a complementary in preferences between consumption and labor helps to reproduce the observed dynamic of the labor market following a government expenditure shock.
As said previously, in a standard DSGE model the negative wealth eﬀect induced by fiscal policy rises labor supply. Moreover, if the model produces a rise in real wages, for instance in the presence of nominal rigidities on prices, rule-of-thumb consumers will also increase their labor force participation. Some studies conclude that the unemploy-ment rate could increase since the rise in the labor force participation could prevail over the rise in employment. For instance, Br¨uckner and Pappa (2012) estimate the eﬀects of government expenditure shocks on output and the labor market for a large set of OECD countries. In most cases, the unemployment rate tends to increase. In a new-Keynesian model with a Mortensen and Pissarides framework, Mayer, Moyen and St¨ahler (2010) argue that the unemployment can increase following a positive government expendi-ture shock although hours worked increase.3 The authors focus also on the parameters which drive the unemployment fiscal multiplier. Coherently, the level of price stickiness increases the job creation, and the debt-based public expenditure expansions trigger the larger unemployment fiscal multiplier. Since the evolution of the marginal rate of substitution between consumption and leisure is central, the degree of risk aversion and of convexity in labor disutility are parameters which drive the response of vacancies and of unemployment.
This thesis aims at contributing to this growing literature by focusing on the eﬀects of fiscal policy on the labor market. As it will be presented later on in more detail, in this thesis I attempt to show that the eﬀects of fiscal policy on the labor market depend on the fiscal instrument used in the case of fiscal expansions. Also, most of papers dealing with spillover eﬀects of fiscal policy or fiscal transfers schemes in monetary union focus on variables such as output, consumption or inflation and thus neglect the labor market. In this thesis I also take into consideration the spillover eﬀects of fiscal policy on the foreign labor market and discuss the ability of fiscal transfers to smooth unemployment diﬀerential between member states of a monetary union.
The suitability of the new-Keynesian framework for the study of the fiscal policy.
In this thesis I use essentially DSGE models for investigating the short-run eﬀects of fiscal policy. During the two last decades, the new-Keynesian framework has been used extensively to document diﬀerent issues related to fiscal policy, such as the response of the private sector to increases in government spending, the size of the fiscal multiplier or yet the eﬀects of fiscal policy during Zero Lower Bound episodes. Adopted by most of central banks and a large strand of the literature on fiscal policy, the DSGE model seems to be very accurate for analyzing such issues.
In this paragraph, I discuss the ability of DSGE models to analyze such issues. Be-yond fiscal policy, the new-Keynesian model has been adopted by a large part of the profession. Beside the empirical time series models, the DSGE framework has become the main analysis tool in modern macroeconomics. However, the DGSE models face various criticisms, which have been, for some of them, essentially present since the recent crisis. I briefly list here some of the main criticisms addressed to the DSGE modeling.
In the aftermath of the crisis, many economists pointed out the inability of such models to predict the crisis. More importantly, a DSGE model describes the economy in normal times so that it is irrelevant to describe the episodes of bubbles or large recessions. In the same sense, Stiglitz (2011) argues that the representative agent as-sumption is irrelevant and engender too simplistic models. Especially, Stiglitz highlights the inability of DSGE models with a representative agent to include satisfactory rep-resentations of the financial markets:”Many used ”representative agent models” – al l individuals were assumed to be identical, and this meant there could be no meaningful financial markets (who would be lending money to whom?). Information asymmetries, the cornerstone of modern economics, also had no place: they could arise only if in-dividuals suﬀered from acute schizophrenia, an assumption incompatible with another of the favored assumptions, ful l rationality”. It is a fact that pre-crisis DSGE models neglect the modeling and the role of the financial markets. However, and in response to this clear failure, the introduction of more sophisticated financial markets, especially through the presence of information asymmetries and credit constraints, is an active issue in the broad DSGE literature nowadays. Since the great recession is closely re-lated to financial matters, the inability of DSGE models to forecast the recent crisis and the absence of a relevant description of the financial markets in these models are also related. Recently, one important attempt to respond to this lack can be found in Del Ne-gro, Giannoni and Schorfheide (2015). They show that, combining the Smets-Wouters model with a financial accelerator like in Bernanke, Gertler and Gilchrist (1999), the model (estimated on pre-crisis data) is able to forecast the large decline of output and the mitigated decrease in inflation after 2008. In any case, the DSGE modelers have visibly taken into account that the DSGE models have to be improved in this way.
Stiglitz (2011), among others, states that DSGE models also fail to produce a clear description of the dynamic of the labor market. Especially, most of DSGE models (at least pre-crisis models) assume the absence of unemployment at the general equi-librium. Since the labor market is central in the present thesis, the ability of DSGE models to investigate issues related to the labor market is a crucial element. It is true that benchmark DSGE models propose a poor description of the labor market. However, an important step in the DSGE literature was to introduce the job search and matching framework into a standard RBC or DSGE structure. First attempts to introduce a labor market a` la Mortensen and Pissarides are not new and can be found for instance in Merz (1995) and Andolfatto (1996)4 . Nowadays, numerous articles dealing with the short-run fluctuations of the labor market have adopted a Mortensen and Pissarides framework into a DSGE model.5 Such models have been used for investigating diﬀerent issues in macroeconomics: the unemployment puzzle6 , the labor wedge and the business cycle, the inflation persistence and so on.
A job search and matching structure can be easily introduced in a DSGE framework, so that it enables to look at the interactions between the labor market and the rest of the economy. Such models have been used recently to investigate the eﬀects of fiscal policy shocks on the labor market (for instance the papers quoted previously and other articles mentioned throughout the thesis). Also, some papers focus on the eﬀects of structural reforms on the labor market. For instance Cacciatore, Duval and Fiori (2012) investigate the short-run eﬀects of three diﬀerent labor market reforms: a relaxation of job protection, reduction in the unemployment benefit replacement rate and a strengthening in activation policy.
Summarize and added value of the thesis
The first part of the thesis aims at investigating the eﬀects of fiscal shocks on the la-bor market in the case of a closed economy (Chapter 1 and Chapter 2). A second part extends the closed economy framework by considering a monetary union structure (third and fourth chapter). In the case of a decentralized fiscal policy, I investigate the spillover eﬀects of fiscal policy from one member state to the rest of the union. Also, this part studies the stabilizing properties of fiscal transfers schemes implemented between member states.
Chapter 1 is dedicated to disentangle the eﬀects of government consumption and government investment on the labor market and especially on the unemployment rate. I develop a new-Keynesian DSGE model closely following Gali, Smets and Wouters (2012), especially for the introduction of the labor market. This modeling for the labor market introduces a labor force participation decision and enriches the analyses of the dynamic of unemployment. A first main result is that government consumption and investment have very diﬀerent eﬀects on output and on the labor market. Government investment triggers a higher output fiscal multiplier than government consumption. However, and despite the long-run eﬀects of government investment on output and em-ployment, a rise in public investment triggers lower eﬀects on unemployment. The main contribution is to demonstrate that, if government investment produces positive long-run eﬀects on employment, it also generates a strong rise in the real wage and in the labor force participation in the long-run. Eﬀects on unemployment are therefore only temporary and the unemployment fiscal multiplier is significantly lower than in the case of an increase in government consumption.
In Chapter 1, I also estimate the respective eﬀects of government consumption and investment on key macroeconomic variables for the Euro Area. To identify fiscal struc-tural shocks, I use the well-known structural vector autoregression approach following the seminal paper by Blanchard and Perotti (2002). Also, for comparison purposes, I impose constraints in the S-VAR by applying a Choleski decomposition. The Choleski decomposition and the Blanchard and Perotti approach provide rather diﬀerent results. The Choleski decomposition generate responses of unemployment which rather support the results from the theoretical exercise. In the case of a rise in public consumption, results support the neo-classical view: public consumption decreases the real wage and crowds out employment. However, the labor force participation also falls and the un-employment rate decreases. A rise in government investment triggers a comovement of employment, the labor force participation and of the real wage. More interestingly, a rise in public investment, despite a positive eﬀect on total employment, increases slightly un-employment. These results confirm the findings produce by the new-Keynesian model and the estimates in Bermperoglu, Pappa and Vella (2013) since the authors find a higher eﬀect of public investment on output but a lower eﬀect on unemployment. Un-fortunately, the impulse response functions with the Blanchard and Perotti approach provides rather diﬀerent results. With this specification, both expenditure shocks trig-ger the (positive) comovement of employment, the labor force participation and the real wage and in both cases unemployment falls. However, the unemployment fiscal multi-plier is slightly larger in the case of public investment. In contrary to the new-Keynesian model, in the case of government investment impulse response functions do not reveal long-run eﬀects on the labor force participation and on the real wage.
Chapter 2 deals with the size of the fiscal multipliers over the business cycle. Some recent empirical studies7 argue that the fiscal multipliers strongly depend on the position of the economy over the business cycle. The eﬀects of fiscal policy shocks (expansions or consolidations) seem to be larger during economic downturn. For now, only few papers investigate the underlying theoretical mechanisms as, for example, Sims and Wolﬀ (2013) and Michaillat (2014). The contribution of Chapter 2 is to suggest a new transmission channel which could explain diﬀerent values for the multiplier according to the position over the business cycle. More precisely, the output fiscal multiplier during economic downturn is produced thanks to the dynamic of the labor market. A stronger positive response of employment when the steady-state unemployment is high triggers a stronger decrease (or a lower increase) in the real wage so that inflation pressures are lower. The crowding-out eﬀect on private consumption of a public spending shock through a rise in real interest rate is then lower during economic downturn so that the output fiscal multiplier is larger. From a methodological point of view, I use a new-Keynesian model with a search and matching model for the labor market. Especially, the labor market is composed by both a private and a public sector, with a modeling close to Afonso and Gomes (2014) for instance. We introduce the position over the busi-ness cycle thanks to two steady state values for the unemployment rate, in the spirit of Michaillat (2014). The main result is that the labor market, and especially the dynamic of the real wage, is crucial to generate higher output fiscal multipliers in a theoretical framework. Moreover, the findings highlighted in this chapter do not conflict with those obtained in the two papers mentioned above: our result is based on the same transmis-sion channel that Michaillat (2014) and is complementary with the one expounded in Sims and Wolﬀ (2013).
Chapter 3 and 4 investigate issues related to fiscal policy in a monetary union. Chapter 3 is dedicated to the analysis of the spillover eﬀects of fiscal policy shocks in one member state on the rest of the union. While most of existing studies consider the spillover eﬀects of a government consumption shock and/or a general government total expenditure shock, Chapter 3 aims at disentangling in a theoretical framework the respective eﬀects of six fiscal instruments: three taxes (VAT, a labor income tax and a social protection tax) and three sorts of expenditure (government consumption, investment and social transfers). I show that the sign and the size of the spillovers depend on the fiscal instrument. The diﬀerent fiscal tools trigger diﬀerent eﬀects on home and foreign inflation so that all fiscal expansions do not generate a rise in the real exchange rate in the home economy. Also, the rise in the nominal interest rate depends on the inflation pressures produced by the fiscal expansion. I consider the case in which the monetary policy is passive. In this case, spillovers tend to increase.
In chapter 4, the case of fiscal transfers between member states is considered. In the case of the Euro Area in which fiscal policy is decentralized, the implementation of fiscal transfers could help to improve the macroeconomic stabilization within this heterogeneous monetary union. First, I introduce the topic with a presentation of issues which have been considered by the academic literature about the eﬀectiveness of fiscal transfers schemes. Then, diﬀerent transfers schemes are tested according to the use of diﬀerent fiscal instruments by the recipient economy. I simulate the model with both a demand and a supply negative shock in the home economy. In the case of the demand shock, fiscal transfers used through a rise in public consumption, transfers and a VAT cut are more eﬀective to stabilize output, unemployment and inflation diﬀerentials between the two economies than the labor income tax and the social protection tax. However, in the case of a supply shock, the two latter taxes are more eﬀective. First, a negative supply shock triggers a decline in output but a rise in inflation. Funds used via transfers to households, VAT and public consumption trigger additional upward pressures on inflation while a labor income tax or a social protection tax cut produce weak pressures on inflation, so that these two taxes stabilize more inflation in the monetary union. Second, in the case of the negative supply shock, even if output falls, unemployment decreases. A VAT cut or a rise in public consumption or transfers to households trigger downward pressures on unemployment so that the volatility of unemployment is increased.
Table of contents :
1 Assessing the effects of public expenditure shocks on the labor Market
1.2 The DSGE model
1.2.1 Optimizing households
1.2.2 Labor force participation and wage-setting
1.2.4 Market clearing condition
1.2.5 Economic policies
1.3 Effects of public expenditure shocks
1.3.1 Calibration of the model
1.3.2 A rise in public consumption
1.3.3 A rise in government investment
1.3.4 The importance of financing
1.4 Empirical framework
1.4.1 Data descpription
1.4.2 Identification methodology
2 How can the labor market account for the effectiveness of fiscal policy over the business cycle?
2.2 The DSGE model
2.2.1 Definitions and the matching process
2.2.2 Households’ decisions
2.2.4 Wage bargaining
2.2.5 Monetary and fiscal policies
2.2.6 Aggregation and market clearing condition
2.3 The effects of fiscal policy over the business cycle
2.3.1 Calibration and comments
2.3.2 The effects of fiscal policy on the labor market and output in normal times
2.3.3 What impact over the business cycle?
3 Spillover effects in a monetary union: why do fiscal policy instruments matter?
3.2 The monetary union framework
3.2.1 Monetary union, price index and real exchange rate
3.2.4 Labor force participation and wage setting
3.2.5 Aggregate variables and market clearing conditions
3.2.6 The economic policy
3.3 Spillover effects of a domestic fiscal policy
3.3.1 Calibration and comments
4 Fiscal transfer schemes in a monetary union: does the nature of transfers matter?
4.2 The monetary union framework
4.2.1 The monetary union
4.2.2 Introduction of the fiscal union
4.3 Calibration of the model and description of the simulations
4.4 Stabilizing properties of the different transfers schemes
4.4.1 Response of the economy in the case of a negative supply shock in the home economy
4.4.2 Response of the economy in the case of a negative demand shock in the home economy
5 General conclusion