Fiscal ConsolidationEUMarch182013
Content
The Macroeconomics of the Fiscal Consolidation in the European Union
Willi Semmler
Remark 1: Joint projects with Peter Flaschel, Stefan Mittnik, Christian Proano, Ekkehard Ernst (ILO), Pu Chen, Frauke Schleer, Econ students Remark 2: Google entries: 2.5 Mill, Wikipedia has 480 reference on European debt crisis, many conferences on crises in EU, EU break up, Future of EU.
Outline I. Empirics of Fiscal Consolidation
II. Regime Dependence of the Multiplier III. Regime Switch in a Macroeconomic Model VI. Conclusions
The European Union - A New Emerging Macro Economy
Our book:
1.
2.
In spite of great heterogeneity (GDP per capita: Lux/Portugal; 5/1, Lux/Romania: 15/1) there was a period of tranquility (convergence of inflation rates, interest rates and an high value of the Euro (2000-2007) Since 2010 double crises: financial market and sovereign debt crises (triggering fiscal austerity), divergent forces re-emerge
I. Empirics of the Fiscal Consolidation Program
Financial crises and severe financial stress are not new; IMF FSI (composed of bank, security, exchange rate variables) and Industrial Production (IP)
2007/2009
I. Empirics of the Fiscal Consolidation Program
Events
2007 / 2009 Financial and Real Crisis
Causes
Financial Market and Sovereign Debt Crisis 2010
Private debt, real estate bubble (Spain, Ireland)
Secularly rising sovereign debt (Portugal, Greece)
Austerity Program 2011 / 2012
Financial market crisis
Rapidly rising bond spreads
I. Empirics of the Fiscal Consolidation Program
=> Two sources of rising debt (1998-2007)
Government Structural Imbalances (% of GDP) Spain Ireland Portugal Greece -1.28 -2.24 -4.5 -4.76
Government Interest Capital Gains Net Debt (% Payments (Annual ∆ of Real of GDP) (% of GDP) Estate Prices) 43.73 27.3 50.08 80.25
Source: Stein (2011)
0.53 0.43 1.42 1.16
9.71 13.3 3.3 9.5
Private Sector Debt Problem Public Sector Debt Problem
I. Empirics of the Fiscal Consolidation Program
1. Aggregate fiscal consolidation (expenditure cut)
ITALY--Austerity Deficit/GDP (Output Gap) Debt/GDP (Unemployment) SPAIN—Austerity Deficit/GDP (Output Gap) Debt/GDP (Unemployment) 2009 0.8 -4.5 (-4.3) 116 (7.8) 3.7 -11.0 (-4.4) 53 (18.0) 2010 -0.6 -3.9 (-2.5) 118 (8.4) 0.2 -10.1 (-4.6) 61.2 (20.1) 2011 -0.9 -3.8 (-2.0) 120 (8.4) -2.2 -9.7 (-3.8) 68.5 (21.7) 2012 -0.8 -2.0 (-2.9) 123.3 (9.5) -6.9 -6.7 (-4.4) 80.9 (24.4) 2013 -0.4 (-2.3) 118 (9.7) -3.5 (-3.6) 87 (25.1)
GREECE—Austerity Deficit/GDP (Output Gap) Debt/GDP (Unemployment)
4.8 -15.6 (-1.4) 129.0 (9.5)
-7.2 -10.3 (-3.7) 145 (12.6)
-9.1 -9.1 (-8.4) 165 (17.7)
-11 -7.3 (-10.7) 160 (19.7)
-9.5
(-9.0) 170 (19.6)
Aggregate effect of the EU fiscal consolidation. Austerity is here measured as a reduction in public consumption Source: EU Report 2012.
I. Empirics of the Fiscal Consolidation Program
Growth rates and deficits (2009-2014) (even with expenditure reduction):
I. Empirics of the Fiscal Consolidation Program
With Deficits the bond yields are rising (2008-2013): see also de Grauwe (2012/3)
I. Empirics of the Fiscal Consolidation Program
2. Composition of the consolidation programs (lasting impact on EU welfare state)
Decentralization of Collective Bargaining Country Power Italy 2009 Spain 2010 Greece 2012 Portugal 2012 Germany 1990
Social Social Security Investment & Public Sector Wages Security Projection Privatization Consumption wage freeze, 2009 table 5 79.5=>69.5 EU, 2012 2010: -5% 72.4=>57.6 09-10//13:-14.0//-17% table 8 10/11: -5% 56.9=>51.2 table 9 40.5=>34.5
Cut in Public
Composition effect of EU consolidation programs, Source: Busch et al (2012)
I.
Empirics of the Fiscal Consolidation Program
ECB: Caught in a Design Flaw Hesitant Monetary Policy
20092010
LTRO: First short term then longerterm liquidity provisions, 2012, to reduce uncertainty and encourage banks to provide credit to the economy
20112012
2012Future
Outright Monetary Transaction program created to purchase sovereign bonds on the secondary market (Conditional on EFSF/ESM Agreement)
Expansion of eligible asset, e.g. ABS to enable refinancing of illiquid assets through ECB to overcome liquidity shortages
Securities Markets Program: Acceptance of Treasury Bonds as collaterals for loans
Banking Union: The ECB is now leading the discussion on supervision of EU Banking Union
10
0
2
4
6
8
-4 1/1/2002 4/1/2002 7/1/2002 10/1/2002 1/1/2003 4/1/2003 7/1/2003 10/1/2003 1/1/2004 4/1/2004 7/1/2004 10/1/2004 1/1/2005 4/1/2005 7/1/2005 10/1/2005 1/1/2006 4/1/2006 7/1/2006 10/1/2006 1/1/2007 4/1/2007 7/1/2007 10/1/2007 FSI PCA GER FSI PCA FR FSI PCA IRE FSI PCA ITA FSI PCA ESP FSI PCA BEL FSI PCA FIN FSI PCA GRE FSI PCA NDL FSI PCA AUT FSI PCA PRT
-2
1/1/2008
4/1/2008
7/1/2008
10/1/2008 1/1/2009 4/1/2009 7/1/2009 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011 7/1/2011 10/1/2011 1/1/2012 4/1/2012 7/1/2012 When the hasty austerity was enacted: Reemergence of financial crisis
I. Empirics of the Fiscal Consolidation Program
II. Regime Dependence of the Multiplier=> Traditional empirical results: For a review of earlier studies, see Mittnik and Semmler, JEBO (2012a): • Earlier literature, Keynesian >1, Romer and Bernstein (2009): 1.5, ; Ramey (2009):>1 • New literature, quite small in RBC, DSGE models < 1, Cogan/Taylor (2009): 0.7; expected higher interest rates and wealth effects (expected taxation) => New Empirics: Multiplier depends on situations/environment/timing
A) Fiscal stimulus has strong expansionary effects (strong multiplier): • If monetary policy decreases the interest rate (but zero bound), see Christiano et al (2009), Hall (2009), Woodford (2010), and also reduces credit spreads
• Households face income and employment constraints in a recessions which are relaxed (see also the fraction of “rule of thumb” consumers increases, Gali et al 2007)
II. Regime Dependence of the Multiplier
B) Fiscal stimulus is less effective (weaker multiplier) : • High sovereign debt, measuring the state of the fiscal situation, see Corsetti et al (2012) who state a threshold value of the debt to GDP ratio of 100% • High financial stress (financial and banking risk premia), see Mittnik and Semmler (2012a, b), IMK (2012), but also Batani et al. (2012) Corsetti et al (2012), and Bolten et al (2012). • An open economy, with flexible exchange rates, and high foreign debt, see Ilzetzki et al (2009), and Erceg et al (2012), on the other hand the support of external demand may be favorable for the multiplier • With certain labor market institutions and wage rigidities, see Monecelli et al (2011) , but may have stronger effect on the multiplier, see Boyer (2012), Charpe et al (2013),
II. Regime Dependence of the Multiplier
C) Fiscal contraction can be strongly contractionary (contractionary multiplier): • Multiplier is strongly contractionary in “bad times”, see De Long and Summers (2012) in “good and bad” times, persistent effects, see also Auerbach et al (2011) and Fazzari et al (2012) • Not only stronger multiplier in “bad times”, but effects are stronger for “big shocks” (size of shocks matters), see Mittnik and Semmler (2012a, b, c), Chen and Semmler (2012), and Schleer et al (2013), the latter using the ZEW financial stress index
• Also stated now in IMF studies: Multiplier triggered by austerity can be strongly contractionary when enacted in a recession, see Baum et al (2012), defined in terms of output gap, and Benati et al (2012), defined in terms of output growth. The latter consider both expenditure shocks as well as revenue shocks
IMF study (2012: 23): “In all countries, a fiscal consolidation is substantially more contractionary if made during a recession than during an expansion”; see also the Blanchard et al study (2013) on a “computational error”
II. Why is there a Regime Dependence of the multiplier? What happens in the two Regimes?
(see asymmetries of booms and busts in Keynes, and Neftci 1982)
In Recessions • much greater contraints on employment and income • greater constraints in product markets • greater financial market stress: liquidity, credit constraints and greater credit spreads
In Booms • less constraints on employment and income • less constraints in product markets • less constraints in financial markets, less financial stress: liquiditiy, and credit
II. Regime Dependence of the Multiplier Estimating Regimes
1. Hamilton (1996, 2002) version: Markov Switching Model
The Regime Switcher i.e. Low; Medium; High
Defines Probability of a Switch
II. Regime Dependence of the Multiplier Estimating Regimes
2. Two Regime Model: STR Model; Granger and Teräsvirta (1996): y=linear part + linear part* F(*); with
Regime Switcher for 2 Regimes (smooth transition function)
1 0.8 0.6
thresh= -2.233543 smooth=4.88837
0.4
0.2 0 -8 -6 -4 -2 0 2 4 6 Transition variable: imf_fsi_ger(-1) 8 10 12 14
II. Regime Dependence of the Multiplier Estimating Regimes
3. Multi Regime VAR, Tong (1983), Tsay (1998) version Multi-Regime Var( MRVAR), see Mittnik and Semmler (2012a), see also recent studies by Benati et al. (2012) and Baum et al. (2012)
Regime Switcher with Pre-Defined Threshold
One uses a pre-defined threshold for a regime change at r (growth or financial stress regimes) rather than estimating (best-fitting) thresholds Advantages: (i)Piecewise linearization around “interesting locations” (ii)Straightforward linear least-squares estimation (iii) Multi-Regime Impulse-Response
II. Regime Dependence of the Fiscal Multiplier: MR Impulse-Response
Low Growth Regime (left), High Growth Regime (right)
II. Regime Dependence of Monetary Policy
Stress reduction in growth regimes, Spain Low Growth Regime (left), High Growth Regime (right)
II. Regime Dependence of Monetary Policy
Stress reduction in financial stress regimes, Italy (3rd method) Low Stress Regime (left) High Stress Regime (right)
II. Regime Dependence of Monetary Policy
Stress increase in low/high stress regimes, Spain Stress reduction (2nd method)
Low Stress Regime (left) High Stress Regime (right)
II. Regime Dependence of Policies-Summary
Strong asymmetries in booms and recessions: strong contractionary multiplier in recessions=> strong amplification effects arising from: • Constraints in the product market • Constraints in labor market (jobs and wage income) • Constraints in credit market: Financial stress in the financial sector, credit and sovereign bond spreads, banking risk (holding of bad debt by banks) Effect on internal and external demand; an example, Greece (2012):
Public Consumption -9.1% Investment: -20.7% Exports 3.2% Imports -3.4% Private Consumption: -7.1%
Time Span 1 year
Aggregate Demand: -7.1%
II. Regime Dependence of Policies-Summary
To Show the asymmetries in another way: Baum et al. (2012), and Batini et al (2012)=> strong amplification effects in a recession:
See Baum et al (2012)
III. Regime Switch in a Macroeconomic Model
Considering two cases:
Case 1: Regime of low financial stress and/or Central Bank is keeping interest rate and credit spread down. CB partly offsetting the contractionary pressures, for example from fiscal consolidations (UK, US, Japan, but also Germany) Case 2: Regime of high financial stress and Central Bank not able – or willing– to keep interest rate and credit spread down, and not off-setting fiscal consolidation (EU periphery states) => Is debt stabilization working? => Our Model has two main features: 1) Exhibits destabilizing and magnifying macro feedback effects; debt, risk premia rise and credit spread rise, in general see Charpe et al (2013) 2) Intertemporal model, though not infinite horizon (needs tracking sovereign debt, asset prices and sustainability of debt), solved thorugh NMPC (Gruene et al 2013)
III. Regime Switch in Macro Model...
Case 1: Low financial stress, ECB intervention to keep credit spread down
Multiperiod open economy model, see Blanchard and Fisher (1989), solved with NMPC
=>The interest rate r is fixed and credit spreads close to zero (see US, UK, Japan, Germany)
III. Regime Switch in Macro Model...
Case 1: Low financial stress, ECB intervention to keep credit spread down
=> Convergence to sustainable debt (left), left graph with lower fixed rate
Point of Convergence
Point of Convergence
Sovereign debt / Capital Stock
Higher fixed interest rate
Sovereign debt / Capital Stock
Initial condition 2
Lower fixed interest rate
Initial condition 1
Capital Stock
Capital Stock
III. Regime Switch in a Macro Model...
Case 2: High financial stress; ECB intervention does not (or cannot) keep the credit spread down
Regime dependent risk premia and credit spreads for bonds and macroeconomic feedback effects on aggregate demand and output
=>The ECB did not suffciently reduce the interest rate spread in the Euro area for periphery countries, see Grauwe (2012)
III. Regime Switch in a Macro Model...
Case 2: High financial stress, ECB intervention and spread rising
There are now regime dependent risk premia and credit spreads for bonds and macroeconomic feedback effects to aggregate demand and output left: weaker macro feedbacks right: (left graph) stronger feedbacks
No convergence
Sovereign debt / Capital Stock
Capital Stock Remark: Regime switch from low financial stress to high financial stress and stronger macro feedback effects can occur
III. Regime Switch in a Macro Model...
Case 2: High financial stress and strong macro feedback effects
Wealth Effects and Aggregate The share of Demand: Fall of Capital Gains, households that Consumption are income and credit and Investment constrained due to increases increasing credit spread
A fraction of Due to banking households stress, the starts central bank deleveraging, may have no reducing income instruments and liquidity of available to other reduce risk households and premia and firms credit spreads
A weak financial sector, holding risky sovereign debt, may come under stress because sovereign bonds may go into default and banks reduce lending
High Financial Stress
&
Strong Macro Feedbacks
VI. Conclusion
There is clear evidence on regime (business cycle) dependence of the multiplier: Fiscal consolidation in a recessionary period, particularly if coupled with high financial stress, is dominantly contractionary The hasty (panic-driven?) EU program with up-front austerity did not only economically fail but had large social cost, with lasting effects on the EU Welfare State.
It is now more likely that a larger fraction of population opts for default rather than austerity, which could bring countries closer to default. So not only the consolidation is rejected, but the entire EU as project is rejected. Politicians in the EU now suggest: structural and labor market reforms, the employment effects are very slow, and this could just end up in segmented labor markets, as in Germany after the Agenda 2010 Another issue: Our above results may require to re-think the Keynesian text book multiplier (the Kahn version), but actually Keynes points to three dynamic components in macro, see Chiarella et al (2012)
The Berlusconi Super-Multiplier
-- small policy events have large effects:
At the night of the Italian Elections, February 25: Italy's 10-year bond yields jumped from 4.4 to 4.9 percentd
rose to 4.87 percent on Tuesday from 4.448 percent rose to 4.87 percent on Tuesday from 4.448 close on Monday. The spread between safe-haven German bunds and Italian bonds jumped 57 basis points to 336 basis points.
Which may be counter-acted by the
Draghi Super-multiplier
‘Whatever it takes’: the Italian determined to save the euro
‘Whatever it takes’: the Italian determined to save the euro
Thus, Paul de Grauwe, LSE Economics Professor, may be right:
The Eurozone policy is driven by market sentiments ,,,and the austerity was „a panic driven austerity.“ (www.voxeuorg)
II. Regime Dependence of Monetary Policy
Stress increase in high stress regime, Germany (3rd method), see Chen et al. (2013) High Stress Regime, self-enforcing feedback loops
FSI response to FSI shocks Output response to FSI shock
MR for Fiscal Policy vs One-Regime DSGE Model
DSGE Models rule out by construction, state dependence of policy actions: • In DSGE theory: agents find themselves always in the same regime • Agents make smooth (unconstrained) choice of consumption and employment (variables driven by technology shocks) • Models solved through linearizations (log-linear, first- and second-order approximations) • In linearizations the timing and size of shocks do not matter (no regime dependece, no symmetries) • Yet, linearizations and VAR may lead to distortions as compared to nonlinear models, see Becker et al. (2007)
MR Models history: Non-linear modeling and MR models
Business Cycle Analysis and Regimes: • Neftci (1982): Regime switching model in terms of time • Hamilton (1989, 1994, 2002,): Regime switching model in terms of state (Markov Switching VAR, MSVAR) • Granger and Teräsvirta (1996): Smooth transition regression model (STR model) • Tong (1978, 1998) and Tsay (1998): Threshold autoregression models (TAR) Regime dependence of impulse responses: • Potter (1994), univariate impulse-response, • Koop, Pesaran and Potter (1996), multivariate impulse response
MR for Fiscal Policy: Multiplier in atwo Regime Model
Recent exceptions: Timing matters => significant multiplier
• Gali et al. (2007), Ricardian and „rule of thumb“ consumers (consumption spending is income dependent): • Christiano et al. (2009): persistence of zero bound interest rates (multiplier up to 3)
• Ramey (2009): military spending and expectations of government expenditure
• Hall (2010): multiplier of about 1 with countercyclical mark up, elastic labor supply, complementarity of consumption and labor income, and zero interest rate bound
MR for Fiscal Policy – The Multiplier in a Two-Regime Model
Two-regime model (Gang and Semmler 2006, 2009), sketch in appendix of paper, Malinvaud tradition (1978, 1994) • First stage of decision making: unconstrained consumption - employment choice (similar to Gali et al. 2007, Ricardian consumers). Can be associated with high growth rates:
• Second stage of decision making: with labor market not cleared, there is constrained choice, consumption depends on actual employment, and firms` production depends on actual demand
MR Fiscal Policy Empirics: Regime Dependence Multiplier, Regimes and Data
Comparison of (one-regime) VAR and two-regime MRVAR
• Variables ΔlogGDP, ΔlogEMP
• Sample period: 1954:1-2008:4
• For two-regime MRVAR: Threshold predefined as sample mean of output growth rate (3.18%)
• Model selection: AIC (see Chan et al., 2004) suggests:
one-regime VAR of order 5, AIC(M=1, p=5) = 617.6 two-regime MRVAR, AIC(M=2, p1=3, p2=3) = 483.5
MR Monetary Policy: Financial Stress
and Growth Regimes
The IMF FSI (2011), EU countries (Germany, France, Italy, Spain, UK) and USA, monthly 1981.1-2010.9 FSI= Bank beta+TED spread+Inverted terms spread+corporate bond spread+stock market return+stock market volatility+Exchange rate volatility IP: Industrial Production Index (OECD, 2011) Examples: IP (blue, left scale) and FSI (red, right scale)
MR Monetary Policy: Financial Stress
and Growth Regimes
Data: IP and FSI: monthly 1981.1-2010.9, FSI= Bank beta+TED spread+Inverted terms spread+corporate bond spread+stock market return+stock market volatility+Exchange rate volatility Growth regimes and stress reduction Italy left Spain right
Willi Semmler
Remark 1: Joint projects with Peter Flaschel, Stefan Mittnik, Christian Proano, Ekkehard Ernst (ILO), Pu Chen, Frauke Schleer, Econ students Remark 2: Google entries: 2.5 Mill, Wikipedia has 480 reference on European debt crisis, many conferences on crises in EU, EU break up, Future of EU.
Outline I. Empirics of Fiscal Consolidation
II. Regime Dependence of the Multiplier III. Regime Switch in a Macroeconomic Model VI. Conclusions
The European Union - A New Emerging Macro Economy
Our book:
1.
2.
In spite of great heterogeneity (GDP per capita: Lux/Portugal; 5/1, Lux/Romania: 15/1) there was a period of tranquility (convergence of inflation rates, interest rates and an high value of the Euro (2000-2007) Since 2010 double crises: financial market and sovereign debt crises (triggering fiscal austerity), divergent forces re-emerge
I. Empirics of the Fiscal Consolidation Program
Financial crises and severe financial stress are not new; IMF FSI (composed of bank, security, exchange rate variables) and Industrial Production (IP)
2007/2009
I. Empirics of the Fiscal Consolidation Program
Events
2007 / 2009 Financial and Real Crisis
Causes
Financial Market and Sovereign Debt Crisis 2010
Private debt, real estate bubble (Spain, Ireland)
Secularly rising sovereign debt (Portugal, Greece)
Austerity Program 2011 / 2012
Financial market crisis
Rapidly rising bond spreads
I. Empirics of the Fiscal Consolidation Program
=> Two sources of rising debt (1998-2007)
Government Structural Imbalances (% of GDP) Spain Ireland Portugal Greece -1.28 -2.24 -4.5 -4.76
Government Interest Capital Gains Net Debt (% Payments (Annual ∆ of Real of GDP) (% of GDP) Estate Prices) 43.73 27.3 50.08 80.25
Source: Stein (2011)
0.53 0.43 1.42 1.16
9.71 13.3 3.3 9.5
Private Sector Debt Problem Public Sector Debt Problem
I. Empirics of the Fiscal Consolidation Program
1. Aggregate fiscal consolidation (expenditure cut)
ITALY--Austerity Deficit/GDP (Output Gap) Debt/GDP (Unemployment) SPAIN—Austerity Deficit/GDP (Output Gap) Debt/GDP (Unemployment) 2009 0.8 -4.5 (-4.3) 116 (7.8) 3.7 -11.0 (-4.4) 53 (18.0) 2010 -0.6 -3.9 (-2.5) 118 (8.4) 0.2 -10.1 (-4.6) 61.2 (20.1) 2011 -0.9 -3.8 (-2.0) 120 (8.4) -2.2 -9.7 (-3.8) 68.5 (21.7) 2012 -0.8 -2.0 (-2.9) 123.3 (9.5) -6.9 -6.7 (-4.4) 80.9 (24.4) 2013 -0.4 (-2.3) 118 (9.7) -3.5 (-3.6) 87 (25.1)
GREECE—Austerity Deficit/GDP (Output Gap) Debt/GDP (Unemployment)
4.8 -15.6 (-1.4) 129.0 (9.5)
-7.2 -10.3 (-3.7) 145 (12.6)
-9.1 -9.1 (-8.4) 165 (17.7)
-11 -7.3 (-10.7) 160 (19.7)
-9.5
(-9.0) 170 (19.6)
Aggregate effect of the EU fiscal consolidation. Austerity is here measured as a reduction in public consumption Source: EU Report 2012.
I. Empirics of the Fiscal Consolidation Program
Growth rates and deficits (2009-2014) (even with expenditure reduction):
I. Empirics of the Fiscal Consolidation Program
With Deficits the bond yields are rising (2008-2013): see also de Grauwe (2012/3)
I. Empirics of the Fiscal Consolidation Program
2. Composition of the consolidation programs (lasting impact on EU welfare state)
Decentralization of Collective Bargaining Country Power Italy 2009 Spain 2010 Greece 2012 Portugal 2012 Germany 1990
Social Social Security Investment & Public Sector Wages Security Projection Privatization Consumption wage freeze, 2009 table 5 79.5=>69.5 EU, 2012 2010: -5% 72.4=>57.6 09-10//13:-14.0//-17% table 8 10/11: -5% 56.9=>51.2 table 9 40.5=>34.5
Cut in Public
Composition effect of EU consolidation programs, Source: Busch et al (2012)
I.
Empirics of the Fiscal Consolidation Program
ECB: Caught in a Design Flaw Hesitant Monetary Policy
20092010
LTRO: First short term then longerterm liquidity provisions, 2012, to reduce uncertainty and encourage banks to provide credit to the economy
20112012
2012Future
Outright Monetary Transaction program created to purchase sovereign bonds on the secondary market (Conditional on EFSF/ESM Agreement)
Expansion of eligible asset, e.g. ABS to enable refinancing of illiquid assets through ECB to overcome liquidity shortages
Securities Markets Program: Acceptance of Treasury Bonds as collaterals for loans
Banking Union: The ECB is now leading the discussion on supervision of EU Banking Union
10
0
2
4
6
8
-4 1/1/2002 4/1/2002 7/1/2002 10/1/2002 1/1/2003 4/1/2003 7/1/2003 10/1/2003 1/1/2004 4/1/2004 7/1/2004 10/1/2004 1/1/2005 4/1/2005 7/1/2005 10/1/2005 1/1/2006 4/1/2006 7/1/2006 10/1/2006 1/1/2007 4/1/2007 7/1/2007 10/1/2007 FSI PCA GER FSI PCA FR FSI PCA IRE FSI PCA ITA FSI PCA ESP FSI PCA BEL FSI PCA FIN FSI PCA GRE FSI PCA NDL FSI PCA AUT FSI PCA PRT
-2
1/1/2008
4/1/2008
7/1/2008
10/1/2008 1/1/2009 4/1/2009 7/1/2009 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011 7/1/2011 10/1/2011 1/1/2012 4/1/2012 7/1/2012 When the hasty austerity was enacted: Reemergence of financial crisis
I. Empirics of the Fiscal Consolidation Program
II. Regime Dependence of the Multiplier=> Traditional empirical results: For a review of earlier studies, see Mittnik and Semmler, JEBO (2012a): • Earlier literature, Keynesian >1, Romer and Bernstein (2009): 1.5, ; Ramey (2009):>1 • New literature, quite small in RBC, DSGE models < 1, Cogan/Taylor (2009): 0.7; expected higher interest rates and wealth effects (expected taxation) => New Empirics: Multiplier depends on situations/environment/timing
A) Fiscal stimulus has strong expansionary effects (strong multiplier): • If monetary policy decreases the interest rate (but zero bound), see Christiano et al (2009), Hall (2009), Woodford (2010), and also reduces credit spreads
• Households face income and employment constraints in a recessions which are relaxed (see also the fraction of “rule of thumb” consumers increases, Gali et al 2007)
II. Regime Dependence of the Multiplier
B) Fiscal stimulus is less effective (weaker multiplier) : • High sovereign debt, measuring the state of the fiscal situation, see Corsetti et al (2012) who state a threshold value of the debt to GDP ratio of 100% • High financial stress (financial and banking risk premia), see Mittnik and Semmler (2012a, b), IMK (2012), but also Batani et al. (2012) Corsetti et al (2012), and Bolten et al (2012). • An open economy, with flexible exchange rates, and high foreign debt, see Ilzetzki et al (2009), and Erceg et al (2012), on the other hand the support of external demand may be favorable for the multiplier • With certain labor market institutions and wage rigidities, see Monecelli et al (2011) , but may have stronger effect on the multiplier, see Boyer (2012), Charpe et al (2013),
II. Regime Dependence of the Multiplier
C) Fiscal contraction can be strongly contractionary (contractionary multiplier): • Multiplier is strongly contractionary in “bad times”, see De Long and Summers (2012) in “good and bad” times, persistent effects, see also Auerbach et al (2011) and Fazzari et al (2012) • Not only stronger multiplier in “bad times”, but effects are stronger for “big shocks” (size of shocks matters), see Mittnik and Semmler (2012a, b, c), Chen and Semmler (2012), and Schleer et al (2013), the latter using the ZEW financial stress index
• Also stated now in IMF studies: Multiplier triggered by austerity can be strongly contractionary when enacted in a recession, see Baum et al (2012), defined in terms of output gap, and Benati et al (2012), defined in terms of output growth. The latter consider both expenditure shocks as well as revenue shocks
IMF study (2012: 23): “In all countries, a fiscal consolidation is substantially more contractionary if made during a recession than during an expansion”; see also the Blanchard et al study (2013) on a “computational error”
II. Why is there a Regime Dependence of the multiplier? What happens in the two Regimes?
(see asymmetries of booms and busts in Keynes, and Neftci 1982)
In Recessions • much greater contraints on employment and income • greater constraints in product markets • greater financial market stress: liquidity, credit constraints and greater credit spreads
In Booms • less constraints on employment and income • less constraints in product markets • less constraints in financial markets, less financial stress: liquiditiy, and credit
II. Regime Dependence of the Multiplier Estimating Regimes
1. Hamilton (1996, 2002) version: Markov Switching Model
The Regime Switcher i.e. Low; Medium; High
Defines Probability of a Switch
II. Regime Dependence of the Multiplier Estimating Regimes
2. Two Regime Model: STR Model; Granger and Teräsvirta (1996): y=linear part + linear part* F(*); with
Regime Switcher for 2 Regimes (smooth transition function)
1 0.8 0.6
thresh= -2.233543 smooth=4.88837
0.4
0.2 0 -8 -6 -4 -2 0 2 4 6 Transition variable: imf_fsi_ger(-1) 8 10 12 14
II. Regime Dependence of the Multiplier Estimating Regimes
3. Multi Regime VAR, Tong (1983), Tsay (1998) version Multi-Regime Var( MRVAR), see Mittnik and Semmler (2012a), see also recent studies by Benati et al. (2012) and Baum et al. (2012)
Regime Switcher with Pre-Defined Threshold
One uses a pre-defined threshold for a regime change at r (growth or financial stress regimes) rather than estimating (best-fitting) thresholds Advantages: (i)Piecewise linearization around “interesting locations” (ii)Straightforward linear least-squares estimation (iii) Multi-Regime Impulse-Response
II. Regime Dependence of the Fiscal Multiplier: MR Impulse-Response
Low Growth Regime (left), High Growth Regime (right)
II. Regime Dependence of Monetary Policy
Stress reduction in growth regimes, Spain Low Growth Regime (left), High Growth Regime (right)
II. Regime Dependence of Monetary Policy
Stress reduction in financial stress regimes, Italy (3rd method) Low Stress Regime (left) High Stress Regime (right)
II. Regime Dependence of Monetary Policy
Stress increase in low/high stress regimes, Spain Stress reduction (2nd method)
Low Stress Regime (left) High Stress Regime (right)
II. Regime Dependence of Policies-Summary
Strong asymmetries in booms and recessions: strong contractionary multiplier in recessions=> strong amplification effects arising from: • Constraints in the product market • Constraints in labor market (jobs and wage income) • Constraints in credit market: Financial stress in the financial sector, credit and sovereign bond spreads, banking risk (holding of bad debt by banks) Effect on internal and external demand; an example, Greece (2012):
Public Consumption -9.1% Investment: -20.7% Exports 3.2% Imports -3.4% Private Consumption: -7.1%
Time Span 1 year
Aggregate Demand: -7.1%
II. Regime Dependence of Policies-Summary
To Show the asymmetries in another way: Baum et al. (2012), and Batini et al (2012)=> strong amplification effects in a recession:
See Baum et al (2012)
III. Regime Switch in a Macroeconomic Model
Considering two cases:
Case 1: Regime of low financial stress and/or Central Bank is keeping interest rate and credit spread down. CB partly offsetting the contractionary pressures, for example from fiscal consolidations (UK, US, Japan, but also Germany) Case 2: Regime of high financial stress and Central Bank not able – or willing– to keep interest rate and credit spread down, and not off-setting fiscal consolidation (EU periphery states) => Is debt stabilization working? => Our Model has two main features: 1) Exhibits destabilizing and magnifying macro feedback effects; debt, risk premia rise and credit spread rise, in general see Charpe et al (2013) 2) Intertemporal model, though not infinite horizon (needs tracking sovereign debt, asset prices and sustainability of debt), solved thorugh NMPC (Gruene et al 2013)
III. Regime Switch in Macro Model...
Case 1: Low financial stress, ECB intervention to keep credit spread down
Multiperiod open economy model, see Blanchard and Fisher (1989), solved with NMPC
=>The interest rate r is fixed and credit spreads close to zero (see US, UK, Japan, Germany)
III. Regime Switch in Macro Model...
Case 1: Low financial stress, ECB intervention to keep credit spread down
=> Convergence to sustainable debt (left), left graph with lower fixed rate
Point of Convergence
Point of Convergence
Sovereign debt / Capital Stock
Higher fixed interest rate
Sovereign debt / Capital Stock
Initial condition 2
Lower fixed interest rate
Initial condition 1
Capital Stock
Capital Stock
III. Regime Switch in a Macro Model...
Case 2: High financial stress; ECB intervention does not (or cannot) keep the credit spread down
Regime dependent risk premia and credit spreads for bonds and macroeconomic feedback effects on aggregate demand and output
=>The ECB did not suffciently reduce the interest rate spread in the Euro area for periphery countries, see Grauwe (2012)
III. Regime Switch in a Macro Model...
Case 2: High financial stress, ECB intervention and spread rising
There are now regime dependent risk premia and credit spreads for bonds and macroeconomic feedback effects to aggregate demand and output left: weaker macro feedbacks right: (left graph) stronger feedbacks
No convergence
Sovereign debt / Capital Stock
Capital Stock Remark: Regime switch from low financial stress to high financial stress and stronger macro feedback effects can occur
III. Regime Switch in a Macro Model...
Case 2: High financial stress and strong macro feedback effects
Wealth Effects and Aggregate The share of Demand: Fall of Capital Gains, households that Consumption are income and credit and Investment constrained due to increases increasing credit spread
A fraction of Due to banking households stress, the starts central bank deleveraging, may have no reducing income instruments and liquidity of available to other reduce risk households and premia and firms credit spreads
A weak financial sector, holding risky sovereign debt, may come under stress because sovereign bonds may go into default and banks reduce lending
High Financial Stress
&
Strong Macro Feedbacks
VI. Conclusion
There is clear evidence on regime (business cycle) dependence of the multiplier: Fiscal consolidation in a recessionary period, particularly if coupled with high financial stress, is dominantly contractionary The hasty (panic-driven?) EU program with up-front austerity did not only economically fail but had large social cost, with lasting effects on the EU Welfare State.
It is now more likely that a larger fraction of population opts for default rather than austerity, which could bring countries closer to default. So not only the consolidation is rejected, but the entire EU as project is rejected. Politicians in the EU now suggest: structural and labor market reforms, the employment effects are very slow, and this could just end up in segmented labor markets, as in Germany after the Agenda 2010 Another issue: Our above results may require to re-think the Keynesian text book multiplier (the Kahn version), but actually Keynes points to three dynamic components in macro, see Chiarella et al (2012)
The Berlusconi Super-Multiplier
-- small policy events have large effects:
At the night of the Italian Elections, February 25: Italy's 10-year bond yields jumped from 4.4 to 4.9 percentd
rose to 4.87 percent on Tuesday from 4.448 percent rose to 4.87 percent on Tuesday from 4.448 close on Monday. The spread between safe-haven German bunds and Italian bonds jumped 57 basis points to 336 basis points.
Which may be counter-acted by the
Draghi Super-multiplier
‘Whatever it takes’: the Italian determined to save the euro
‘Whatever it takes’: the Italian determined to save the euro
Thus, Paul de Grauwe, LSE Economics Professor, may be right:
The Eurozone policy is driven by market sentiments ,,,and the austerity was „a panic driven austerity.“ (www.voxeuorg)
II. Regime Dependence of Monetary Policy
Stress increase in high stress regime, Germany (3rd method), see Chen et al. (2013) High Stress Regime, self-enforcing feedback loops
FSI response to FSI shocks Output response to FSI shock
MR for Fiscal Policy vs One-Regime DSGE Model
DSGE Models rule out by construction, state dependence of policy actions: • In DSGE theory: agents find themselves always in the same regime • Agents make smooth (unconstrained) choice of consumption and employment (variables driven by technology shocks) • Models solved through linearizations (log-linear, first- and second-order approximations) • In linearizations the timing and size of shocks do not matter (no regime dependece, no symmetries) • Yet, linearizations and VAR may lead to distortions as compared to nonlinear models, see Becker et al. (2007)
MR Models history: Non-linear modeling and MR models
Business Cycle Analysis and Regimes: • Neftci (1982): Regime switching model in terms of time • Hamilton (1989, 1994, 2002,): Regime switching model in terms of state (Markov Switching VAR, MSVAR) • Granger and Teräsvirta (1996): Smooth transition regression model (STR model) • Tong (1978, 1998) and Tsay (1998): Threshold autoregression models (TAR) Regime dependence of impulse responses: • Potter (1994), univariate impulse-response, • Koop, Pesaran and Potter (1996), multivariate impulse response
MR for Fiscal Policy: Multiplier in atwo Regime Model
Recent exceptions: Timing matters => significant multiplier
• Gali et al. (2007), Ricardian and „rule of thumb“ consumers (consumption spending is income dependent): • Christiano et al. (2009): persistence of zero bound interest rates (multiplier up to 3)
• Ramey (2009): military spending and expectations of government expenditure
• Hall (2010): multiplier of about 1 with countercyclical mark up, elastic labor supply, complementarity of consumption and labor income, and zero interest rate bound
MR for Fiscal Policy – The Multiplier in a Two-Regime Model
Two-regime model (Gang and Semmler 2006, 2009), sketch in appendix of paper, Malinvaud tradition (1978, 1994) • First stage of decision making: unconstrained consumption - employment choice (similar to Gali et al. 2007, Ricardian consumers). Can be associated with high growth rates:
• Second stage of decision making: with labor market not cleared, there is constrained choice, consumption depends on actual employment, and firms` production depends on actual demand
MR Fiscal Policy Empirics: Regime Dependence Multiplier, Regimes and Data
Comparison of (one-regime) VAR and two-regime MRVAR
• Variables ΔlogGDP, ΔlogEMP
• Sample period: 1954:1-2008:4
• For two-regime MRVAR: Threshold predefined as sample mean of output growth rate (3.18%)
• Model selection: AIC (see Chan et al., 2004) suggests:
one-regime VAR of order 5, AIC(M=1, p=5) = 617.6 two-regime MRVAR, AIC(M=2, p1=3, p2=3) = 483.5
MR Monetary Policy: Financial Stress
and Growth Regimes
The IMF FSI (2011), EU countries (Germany, France, Italy, Spain, UK) and USA, monthly 1981.1-2010.9 FSI= Bank beta+TED spread+Inverted terms spread+corporate bond spread+stock market return+stock market volatility+Exchange rate volatility IP: Industrial Production Index (OECD, 2011) Examples: IP (blue, left scale) and FSI (red, right scale)
MR Monetary Policy: Financial Stress
and Growth Regimes
Data: IP and FSI: monthly 1981.1-2010.9, FSI= Bank beta+TED spread+Inverted terms spread+corporate bond spread+stock market return+stock market volatility+Exchange rate volatility Growth regimes and stress reduction Italy left Spain right
