Banque nationale de Belgique SA · EURONEXT:BNB
Banque nationale de Belgique SA — presentatie degraeve
34 slides from Banque nationale de Belgique SA's investor presentation, November 2013
Slide 1: Identifying Fiscal Inflation
Identifying Fiscal In‡ation Ferre De Graeve Virginia Queijo von Heideken NBB - November 2013 De Graeve and Queijo von Heideken Fiscal () In‡ation 1 / 34
Slide 2: Fiscal inflation
Fiscal in‡ation Starting observation: Debt levels around the world skyrocket Theory (FTPL): if not backed by future surpluses =) …scal in‡ation Sims, Leeper, Cochrane “In‡ation lures” Yet very little concern from policy makers & other academics Fed, Gali, Krugman, ... “Well where is in‡ation” Object
Slide 3: What we do
What we do Approach: Estimate standard DSGE model Active MP, passive FP With no data on …scal variables Finding: One of the shocks is essentially re‡ecting …scal policy Interpretation: Contemporary DSGE models (with passive …scal policy) Interpret historical ‡uctuations in in‡ation as related to …sc
Slide 4: Contemporary DSGE models
Contemporary DSGE models Standard DSGE model has Monetary policy rule (R =interest rate, π =in‡ation): Rt = ρπ π t + ... Government budget constraint Fiscal policy rule (τ =taxes, d =debt): τ t = γd dt 1 + ... Solution has debt dynamics of the form: dt = ( β 1 γd )dt 1 + ... De Graeve and Queijo von
Slide 5: Standard policy mix
Standard policy mix Active Monetary policy (R =interest rate, π =in‡ation): Rt = ρπ π t + ... ρπ > 1 Passive Fiscal policy (τ =taxes, d =debt): τ t = γd dt 1 + ... dt = ( β 1 γd )dt 1 + ... (β 1 γd ) < 1 Monetary policy: Taylor principle Fiscal policy: passively stabilizes debt, moves taxes to do so
Slide 6: Contemporary DSGE models
Contemporary DSGE models Absence of …scal policy becomes more of an issue now I. Zero Lower Bound MP’s hands tied II. Debt Huge de…cits, rapidly increasing indebtedness (sustainable?) =) Fiscal activism? De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 6 / 34
Slide 7: Fiscal activism
Fiscal activism 9 alternative determinate parameter constellation (Leeper, 1991) Passive Monetary policy (R =interest rate, π =in‡ation): Rt = ρπ π t + ... ρπ < 1 Active Fiscal policy (τ =taxes, d =debt): τ t = γd dt 1 + ... dt = ( β 1 γd )dt 1 + ... (β 1 γd ) > 1 FP is not committed to stabilize de
Slide 8: Policy mix
Policy mix Contemporary DSGE models typically exclude the second possibility Active …scal policy, passive monetary policy And therefore rule out …scal in‡ation Our message: Even a standard DSGE model (active MP/passive FP) tells us we may have to worry about it De Graeve and Queijo von Heideken Fisc
Slide 9: The model
The model Standard DSGE model: Smets & Wouters (2007) Plus term structure: Expectations Hypothesis 1 R̂tn = Et R̂t + R̂t +1 + ... + R̂t +n 1 n De Graeve, Emiris & Wouters (2009) Fluctuations in yield curve well described by: Macro-shocks (technology, preferences, mark-ups, monetary policy): slope In
Slide 10: News shocks to inflation target
News shocks to in‡ation target Time-varying in‡ation target WITHOUT news shocks ∆π̄ t = ρπ ∆π̄ t 1 + et De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 10 / 34
Slide 11: News shocks to inflation target
News shocks to in‡ation target Time-varying in‡ation target WITHOUT news shocks ∆π̄ t = ρπ ∆π̄ t 1 + et Time-varying in‡ation target WITH news shocks ∆π̄ t = ρπ ∆π̄ t 1 + et + η t i η t i random innovation in t i but materialized in t De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riks
Slide 12: e t De Graeve and Queijo von Heideken η t 20
et η t 20 De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 12 / 34
Slide 13: News shocks to inflation target
News shocks to in‡ation target Time-varying in‡ation target WITHOUT news shocks ∆π̄ t = ρπ ∆π̄ t 1 + et Time-varying in‡ation target WITH news shocks ∆π̄ t = ρπ ∆π̄ t 1 + et + η t i η t i random innovation in t i but materialized in t Why? 1 Fiscal in‡ation 2 Term structure De Graeve and Queijo von
Slide 14: Motivation 1: Fiscal inflation
Motivation 1: Fiscal in‡ation Cochrane - Sims - Leeper At some point in the future, in‡ation may be determined by …scal policy Govt. budget constraint is equilibrium condition. Hence, if MP turns passive at some point, …scal policy will determine in‡ation and MP stabilize the value of debt. Implies
Slide 15: Motivation 2: Term structure
Motivation 2: Term structure Well-know from …nance: Term structure variation driven largely by 1 Level factor 2 Slope factor DSGE models match term structure in same fashion With one (implicit) restriction Almost all movements in slope are due to variation in short rate De Graeve and Queijo von Heid
Slide 16: Slope: e.g. monetary policy shocks
Slope: e.g. monetary policy shocks De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 16 / 34
Slide 17: Level: Inflation target shocks
Level: In‡ation target shocks De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 17 / 34
Slide 18: Identification: News to target is only shock
Identi…cation: News to target is only shock ... De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 18 / 34
Slide 19: In sum
In sum Estimate 1 Smets-Wouters 2 Term structure 3 In‡ation target which can move for 2 reasons (1 new) Data: U.S. 1966:Q1 - 2011:Q2 7 macro variables (dy, dw, dp, di, dc, l, r) Long maturity interest rates (1y, 3y, 5y, 10y) No data on …scal policy De Graeve and Queijo von Heideken Fiscal ( In‡ation
Slide 20: Result 1: News shock looks like fiscal policy
Result 1: News shock looks like …scal policy De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 20 / 34
Slide 21: Result 1: News shock looks like fiscal policy
Result 1: News shock looks like …scal policy De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 21 / 34
Slide 22: Result 1 - Implication
Result 1 - Implication Even if you don’t model …scal policy, your DSGE model (partly) interprets in‡ationary ‡uctuations as …scal =) Should give …scal e¤ects more attention De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 22 / 34
Slide 23: Result 2: Small quantitative role for news shocks
Result 2: Small quantitative role for news shocks Variance decompositions News shock does not play a major role for historical ‡uctuations in 1 In‡ation expectations (VDC: <10%) 2 Macro variables De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 23 / 34
Slide 24: Result 2: Small quantitative role for news shocks
Result 2: Small quantitative role for news shocks Inconsistent with …scal in‡ation interpretation? Sims: (hyper-rational) in‡ation expectations = model-consistent RE + small prob(…scal in‡ation) Leeper: Markov-Switching btw active/passive MP/FP. If probability of …scal in‡ation regime small, real e¤
Slide 25: Result 3: Fiscal inflation concerns today
Result 3: Fiscal in‡ation concerns today Variance decompositions suggest, on average, news not very important in the past ... but that may vary over time Contribution of news shocks to ‡uctuations in in‡ation De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 25 / 34
Slide 26: De Graeve and Queijo von Heideken Fiscal In‡ation
De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 26 / 34
Slide 27: De Graeve and Queijo von Heideken Fiscal In‡ation
De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 27 / 34
Slide 28: Current debate
Current debate Fed view: “We’ll worry about in‡ation when we see it” =) Fiscal in‡ation ignored Sims - Leeper - Cochrane “Threat of …scal in‡ation is present today already” Model: 1 Shows empirical validity of …scal in‡ation worry 2 Explains why we don’t see it (Fed view) De Graeve and Queijo von He
Slide 29: De Graeve and Queijo von Heideken Fiscal In‡ation
De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 29 / 34
Slide 30: Why don't we observe high yields / high inflation today?
Why don’t we observe high yields / high in‡ation today? I. In‡ation-target shocks <0 Forward guidance: short rates low now AND going forward implies downward level-pressure on term structure II. “Real” shocks combination of shocks drives Great Recession Y #=) E π # ,! Counteract upward …scal in‡atio
Slide 31: Some issues
Some issues 1 Lower news shock-…scal correlation pre-1980s 2 Anticipation horizon De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 31 / 34
Slide 32: Lower news shock-fiscal correlation pre-1980s
Lower news shock-…scal correlation pre-1980s Some evidence for policy switch from PM/AF to AM/PF in eighties (Bianchi & Ilut, 2013) Q: Why not present in our news shock? A: In‡ation simultaneously occurred =) no role for anticipation De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksb
Slide 33: Anticipation horizon
Anticipation horizon Maturity of government debt: 1) FTPL, 2) Data De Graeve and Queijo von Heideken Fiscal ( In‡ation Sveriges Riksbank) 33 / 34
Slide 34: Conclusions
Conclusions Standard DSGE model partly interprets historical ‡uctuations as related to …scal policy (despite AM/PF policy mix) 1 Is clearly a part of in‡ation expectations (distilled via yields) 2 Has not been its main driver in past 50 years 3 Is becoming more important (as debt grows?) 4 Reconcile