Economic Policy Uncertainty and Inflation Expectations

Transcription

1 Economic Policy Uncertainty and Inflation Expectations Klodiana Istrefi and Anamaria Piloiu Banque de France DB Research SEM Conference July, Paris 1 / 3

2 The views expressed herein are those of the authors and should under no circumstances be interpreted as reflecting those of the Banque de France or the Eurosystem. 2 / 3

3 Key Question Is there an effect of policy-related uncertainty on inflation expectations? 3 / 3

4 Motivation High levels of uncertainty, with policy uncertainty being a prominent contributor; Policy uncertainty has negative effects on economic activity; However, not much known if such uncertainty affects beliefs and perceptions about policy, inflation expectations; Long-term inflation expectations, information about the extent to which a central bank s commitment to price stability is perceived as credible. 4 / 3

5 Motivation High levels of uncertainty, with policy uncertainty being a prominent contributor; Policy uncertainty has negative effects on economic activity; However, not much known if such uncertainty affects beliefs and perceptions about policy, inflation expectations; Long-term inflation expectations, information about the extent to which a central bank s commitment to price stability is perceived as credible. 4 / 3

6 Motivation High levels of uncertainty, with policy uncertainty being a prominent contributor; Policy uncertainty has negative effects on economic activity; However, not much known if such uncertainty affects beliefs and perceptions about policy, inflation expectations; Long-term inflation expectations, information about the extent to which a central bank s commitment to price stability is perceived as credible. 4 / 3

7 Motivation High levels of uncertainty, with policy uncertainty being a prominent contributor; Policy uncertainty has negative effects on economic activity; However, not much known if such uncertainty affects beliefs and perceptions about policy, inflation expectations; Long-term inflation expectations, information about the extent to which a central bank s commitment to price stability is perceived as credible. 4 / 3

8 In this paper What we do Highlight a new channel through which policy uncertainty can affect the macroeconomy. Study the link between policy-related uncertainty and inflation expectations. Bayesian Panel VAR (US and the euro area). Economic Policy Uncertainty index of Baker, Bloom and Davis (212). Identify policy uncertainty shocks. 5 / 3

9 In this paper What we do Highlight a new channel through which policy uncertainty can affect the macroeconomy. Study the link between policy-related uncertainty and inflation expectations. Bayesian Panel VAR (US and the euro area). Economic Policy Uncertainty index of Baker, Bloom and Davis (212). Identify policy uncertainty shocks. 5 / 3

10 In this paper What we do Highlight a new channel through which policy uncertainty can affect the macroeconomy. Study the link between policy-related uncertainty and inflation expectations. Bayesian Panel VAR (US and the euro area). Economic Policy Uncertainty index of Baker, Bloom and Davis (212). Identify policy uncertainty shocks. 5 / 3

11 In this paper What we do Highlight a new channel through which policy uncertainty can affect the macroeconomy. Study the link between policy-related uncertainty and inflation expectations. Bayesian Panel VAR (US and the euro area). Economic Policy Uncertainty index of Baker, Bloom and Davis (212). Identify policy uncertainty shocks. 5 / 3

12 In this paper What we do Highlight a new channel through which policy uncertainty can affect the macroeconomy. Study the link between policy-related uncertainty and inflation expectations. Bayesian Panel VAR (US and the euro area). Economic Policy Uncertainty index of Baker, Bloom and Davis (212). Identify policy uncertainty shocks. 5 / 3

13 In this paper What we find Given a policy uncertainty shock: economy contracts, short-term inflation expectations and interest rate fall; long-term inflation expectations rise; Monetary and fiscal policy uncertainty, both important; Results robust to several specifications. 6 / 3

14 In this paper What we find Given a policy uncertainty shock: economy contracts, short-term inflation expectations and interest rate fall; long-term inflation expectations rise; Monetary and fiscal policy uncertainty, both important; Results robust to several specifications. 6 / 3

15 In this paper What we find Given a policy uncertainty shock: economy contracts, short-term inflation expectations and interest rate fall; long-term inflation expectations rise; Monetary and fiscal policy uncertainty, both important; Results robust to several specifications. 6 / 3

16 In this paper What we find Given a policy uncertainty shock: economy contracts, short-term inflation expectations and interest rate fall; long-term inflation expectations rise; Monetary and fiscal policy uncertainty, both important; Results robust to several specifications. 6 / 3

17 In this paper What we find Given a policy uncertainty shock: economy contracts, short-term inflation expectations and interest rate fall; long-term inflation expectations rise; Monetary and fiscal policy uncertainty, both important; Results robust to several specifications. 6 / 3

18 Literature & Contribution 1 Macroeconomic effects of economic policy uncertainty Fernandez-Villaverde at al. (213), Born and Pfeifer (214) Baker et al. (212), Mumatz and Zanetti (213) 2 Inflation expectations Clark and Davig (28), Bachmann et al. (214) Mankiw et al. (23), Coibion and Gorodnichenko (212), Andrade and Le Bihan (213) Our paper, first empirical evidence on the effect of: policy-related uncertainty shocks on inflation expectations. monetary policy and fiscal policy-related uncertainty on output. 7 / 3

19 Literature & Contribution 1 Macroeconomic effects of economic policy uncertainty Fernandez-Villaverde at al. (213), Born and Pfeifer (214) Baker et al. (212), Mumatz and Zanetti (213) 2 Inflation expectations Clark and Davig (28), Bachmann et al. (214) Mankiw et al. (23), Coibion and Gorodnichenko (212), Andrade and Le Bihan (213) Our paper, first empirical evidence on the effect of: policy-related uncertainty shocks on inflation expectations. monetary policy and fiscal policy-related uncertainty on output. 7 / 3

20 Literature & Contribution 1 Macroeconomic effects of economic policy uncertainty Fernandez-Villaverde at al. (213), Born and Pfeifer (214) Baker et al. (212), Mumatz and Zanetti (213) 2 Inflation expectations Clark and Davig (28), Bachmann et al. (214) Mankiw et al. (23), Coibion and Gorodnichenko (212), Andrade and Le Bihan (213) Our paper, first empirical evidence on the effect of: policy-related uncertainty shocks on inflation expectations. monetary policy and fiscal policy-related uncertainty on output. 7 / 3

21 Literature & Contribution 1 Macroeconomic effects of economic policy uncertainty Fernandez-Villaverde at al. (213), Born and Pfeifer (214) Baker et al. (212), Mumatz and Zanetti (213) 2 Inflation expectations Clark and Davig (28), Bachmann et al. (214) Mankiw et al. (23), Coibion and Gorodnichenko (212), Andrade and Le Bihan (213) Our paper, first empirical evidence on the effect of: policy-related uncertainty shocks on inflation expectations. monetary policy and fiscal policy-related uncertainty on output. 7 / 3

22 Literature & Contribution 1 Macroeconomic effects of economic policy uncertainty Fernandez-Villaverde at al. (213), Born and Pfeifer (214) Baker et al. (212), Mumatz and Zanetti (213) 2 Inflation expectations Clark and Davig (28), Bachmann et al. (214) Mankiw et al. (23), Coibion and Gorodnichenko (212), Andrade and Le Bihan (213) Our paper, first empirical evidence on the effect of: policy-related uncertainty shocks on inflation expectations. monetary policy and fiscal policy-related uncertainty on output. 7 / 3

23 Literature & Contribution 1 Macroeconomic effects of economic policy uncertainty Fernandez-Villaverde at al. (213), Born and Pfeifer (214) Baker et al. (212), Mumatz and Zanetti (213) 2 Inflation expectations Clark and Davig (28), Bachmann et al. (214) Mankiw et al. (23), Coibion and Gorodnichenko (212), Andrade and Le Bihan (213) Our paper, first empirical evidence on the effect of: policy-related uncertainty shocks on inflation expectations. monetary policy and fiscal policy-related uncertainty on output. 7 / 3

24 Literature & Contribution 1 Macroeconomic effects of economic policy uncertainty Fernandez-Villaverde at al. (213), Born and Pfeifer (214) Baker et al. (212), Mumatz and Zanetti (213) 2 Inflation expectations Clark and Davig (28), Bachmann et al. (214) Mankiw et al. (23), Coibion and Gorodnichenko (212), Andrade and Le Bihan (213) Our paper, first empirical evidence on the effect of: policy-related uncertainty shocks on inflation expectations. monetary policy and fiscal policy-related uncertainty on output. 7 / 3

25 Measuring policy-related uncertainty Economic Policy Uncertainty (EPU) index from Baker, Bloom and Davis (212). News-based EPU for US and the euro area: Evaluations: Human News Audit (corr..76), Beige Book (corr..8). 8 / 3

26 Data and Model specification Panel BVARs for US and euro area with, y i,t = where: epu i,t gdp i,t π e i,long t π e i,short t i i,t epu t, news-based EPU ( data ); gdp t, real GDP, seasonally adjusted; πlong t e, 5 to 1 years ahead inflation expectations; πshort t e, one year ahead inflation expectations; i t, short-term interest rate. Recursive identification (epu t, gdp t, π e long t, πe short t, i t.) EPU reacts on impact only to its own innovations. an innovation to EPU can have an immediate effect on the variables ordered after. 9 / 3

27 Data and Model specification Panel BVARs for US and euro area with, y i,t = where: epu i,t gdp i,t π e i,long t π e i,short t i i,t epu t, news-based EPU ( data ); gdp t, real GDP, seasonally adjusted; πlong t e, 5 to 1 years ahead inflation expectations; πshort t e, one year ahead inflation expectations; i t, short-term interest rate. Recursive identification (epu t, gdp t, π e long t, πe short t, i t.) EPU reacts on impact only to its own innovations. an innovation to EPU can have an immediate effect on the variables ordered after. 9 / 3

28 Data and Model specification We study variations for: News-based: overall policy uncertanty monetary policy uncertanty fiscal policy uncertanty Sources for inflation expectations: Consensus Economics SPFs of Fed Philadelphia and of ECB Sample: 1999Q1-212Q3. Quarterly frequency. Optimal lag selection with BIC. Lag selected 2. Flat prior, as in Uhlig (25). 1 / 3

29 Data and Model specification We study variations for: News-based: overall policy uncertanty monetary policy uncertanty fiscal policy uncertanty Sources for inflation expectations: Consensus Economics SPFs of Fed Philadelphia and of ECB Sample: 1999Q1-212Q3. Quarterly frequency. Optimal lag selection with BIC. Lag selected 2. Flat prior, as in Uhlig (25). 1 / 3

30 Data and Model specification We study variations for: News-based: overall policy uncertanty monetary policy uncertanty fiscal policy uncertanty Sources for inflation expectations: Consensus Economics SPFs of Fed Philadelphia and of ECB Sample: 1999Q1-212Q3. Quarterly frequency. Optimal lag selection with BIC. Lag selected 2. Flat prior, as in Uhlig (25). 1 / 3

31 Data and Model specification We study variations for: News-based: overall policy uncertanty monetary policy uncertanty fiscal policy uncertanty Sources for inflation expectations: Consensus Economics SPFs of Fed Philadelphia and of ECB Sample: 1999Q1-212Q3. Quarterly frequency. Optimal lag selection with BIC. Lag selected 2. Flat prior, as in Uhlig (25). 1 / 3

32 Empirical Results IRFs to an overall policy uncertainty shock IRFs of long- and short-term inflation expectations for: Different sources of inflation expectations. Specific type of policy uncertainty shock. Contribution of policy uncertainty shocks to long-term inflation expectations (FEVD). 11 / 3

33 Empirical Results IRFs to an overall policy uncertainty shock IRFs of long- and short-term inflation expectations for: Different sources of inflation expectations. Specific type of policy uncertainty shock. Contribution of policy uncertainty shocks to long-term inflation expectations (FEVD). 11 / 3

34 Empirical Results IRFs to an overall policy uncertainty shock IRFs of long- and short-term inflation expectations for: Different sources of inflation expectations. Specific type of policy uncertainty shock. Contribution of policy uncertainty shocks to long-term inflation expectations (FEVD). 11 / 3

35 Empirical Results IRFs to an overall policy uncertainty shock IRFs of long- and short-term inflation expectations for: Different sources of inflation expectations. Specific type of policy uncertainty shock. Contribution of policy uncertainty shocks to long-term inflation expectations (FEVD). 11 / 3

36 IRFs to policy uncertainty shock, CE expectations POLICY UNCERTAINTY GDP IE LONG IE SHORT INTEREST RATE (a) Overall policy uncertainty 1 2 Given two standard deviation innovation in EPU: Long-term inflation expectations rise. Short-term expectations fall. Policy uncertainty shocks in our sample three to four st. deviations. Long-term inflation expectations rise up to 1 bp. Sufficient to miss the ECB s below, but close to, 2% or Fed s 2% inflation objective. 12 / 3

37 IRFs to policy uncertainty shock, CE expectations POLICY UNCERTAINTY GDP IE LONG IE SHORT INTEREST RATE (b) Overall policy uncertainty 1 2 Given two standard deviation innovation in EPU: Long-term inflation expectations rise. Short-term expectations fall. Policy uncertainty shocks in our sample three to four st. deviations. Long-term inflation expectations rise up to 1 bp. Sufficient to miss the ECB s below, but close to, 2% or Fed s 2% inflation objective. 12 / 3

38 IRFs to policy uncertainty shock, CE expectations POLICY UNCERTAINTY GDP IE LONG IE SHORT INTEREST RATE (c) Overall policy uncertainty 1 2 Given two standard deviation innovation in EPU: Long-term inflation expectations rise. Short-term expectations fall. Policy uncertainty shocks in our sample three to four st. deviations. Long-term inflation expectations rise up to 1 bp. Sufficient to miss the ECB s below, but close to, 2% or Fed s 2% inflation objective. 12 / 3

39 IRFs of inflation expectations Consensus Economics Survey of Professional Forecasters (a) Overall policy uncertainty (b) Monetary policy uncertainty (c) Fiscal policy uncertainty 13 / 3

40 Contribution of policy uncertainty shocks Table: FEVD (posterior median) of long-term inflation expectations Overall PU MP Uncertainty FP Uncertainty Consensus Economics 1-2 quarters (4.56,26.61) (4.36,22.3) (2.15,16.5) SPF 1-2 quarters (3.34,9.52) (8.4,28.3) (3.16,16.19) Notes: Posterior median of FEVD from the panel BVAR. In brackets its 68% posterior distribution. 14 / 3

41 Robustness checks 1 Assumptions for identification Order EPU index last Re-ordering 2 Selected sample Exclude the recent crisis Sample26 Exclude or dummy for period with unconventional MP Dummy 3 A different proxy for uncertainty, VIX 4 Individual VARs BVAR US BVAR EA Our key result is robust! 15 / 3

42 Robustness checks 1 Assumptions for identification Order EPU index last Re-ordering 2 Selected sample Exclude the recent crisis Sample26 Exclude or dummy for period with unconventional MP Dummy 3 A different proxy for uncertainty, VIX 4 Individual VARs BVAR US BVAR EA Our key result is robust! 15 / 3

43 Robustness checks 1 Assumptions for identification Order EPU index last Re-ordering 2 Selected sample Exclude the recent crisis Sample26 Exclude or dummy for period with unconventional MP Dummy 3 A different proxy for uncertainty, VIX 4 Individual VARs BVAR US BVAR EA Our key result is robust! 15 / 3

44 Robustness checks 1 Assumptions for identification Order EPU index last Re-ordering 2 Selected sample Exclude the recent crisis Sample26 Exclude or dummy for period with unconventional MP Dummy 3 A different proxy for uncertainty, VIX 4 Individual VARs BVAR US BVAR EA Our key result is robust! 15 / 3

45 Robustness checks 1 Assumptions for identification Order EPU index last Re-ordering 2 Selected sample Exclude the recent crisis Sample26 Exclude or dummy for period with unconventional MP Dummy 3 A different proxy for uncertainty, VIX 4 Individual VARs BVAR US BVAR EA Our key result is robust! 15 / 3

46 Summary & Conclusion In response to policy uncertainty shocks: Long-term inflation expectations increase. New empirical result. Monetary policy uncertainty not always the reason. Up to 17% of its variance explained. Economy contracts. In line with previous work in literature. Monetary and fiscal policy uncertainty equally harmful. Results are robust. The effects are stronger for the US. 16 / 3

47 Summary & Conclusion In response to policy uncertainty shocks: Long-term inflation expectations increase. New empirical result. Monetary policy uncertainty not always the reason. Up to 17% of its variance explained. Economy contracts. In line with previous work in literature. Monetary and fiscal policy uncertainty equally harmful. Results are robust. The effects are stronger for the US. 16 / 3

48 Summary & Conclusion In response to policy uncertainty shocks: Long-term inflation expectations increase. New empirical result. Monetary policy uncertainty not always the reason. Up to 17% of its variance explained. Economy contracts. In line with previous work in literature. Monetary and fiscal policy uncertainty equally harmful. Results are robust. The effects are stronger for the US. 16 / 3

49 Summary & Conclusion In response to policy uncertainty shocks: Long-term inflation expectations increase. New empirical result. Monetary policy uncertainty not always the reason. Up to 17% of its variance explained. Economy contracts. In line with previous work in literature. Monetary and fiscal policy uncertainty equally harmful. Results are robust. The effects are stronger for the US. 16 / 3

50 Summary & Conclusion In response to policy uncertainty shocks: Long-term inflation expectations increase. New empirical result. Monetary policy uncertainty not always the reason. Up to 17% of its variance explained. Economy contracts. In line with previous work in literature. Monetary and fiscal policy uncertainty equally harmful. Results are robust. The effects are stronger for the US. 16 / 3

51 Summary & Conclusion In response to policy uncertainty shocks: Long-term inflation expectations increase. New empirical result. Monetary policy uncertainty not always the reason. Up to 17% of its variance explained. Economy contracts. In line with previous work in literature. Monetary and fiscal policy uncertainty equally harmful. Results are robust. The effects are stronger for the US. 16 / 3

52 Thank you for your attention! 17 / 3

53 IRFs of IE for US Consensus Economics Survey of Professional Forecasters (a) Overall policy uncertainty (b) Monetary policy uncertainty (c) Fiscal policy uncertainty 18 / 3

54 IRFs of IE for euro area Consensus Economics Survey of Professional Forecasters (a) Overall policy uncertainty (b) Monetary policy uncertainty (c) Fiscal policy uncertainty 19 / 3

55 IRFs of IE for euro area Consensus Economics Survey of Professional Forecasters (a) Overall policy uncertainty (b) Monetary policy uncertainty (c) Fiscal policy uncertainty 19 / 3

56 IRFs of IE for euro area Consensus Economics Survey of Professional Forecasters (a) Overall policy uncertainty (b) Monetary policy uncertainty (c) Fiscal policy uncertainty 19 / 3

57 IRFs of IE for euro area Consensus Economics Survey of Professional Forecasters (a) Overall policy uncertainty (b) Monetary policy uncertainty (c) Fiscal policy uncertainty 19 / 3

58 IRFs of IE for euro area Consensus Economics Survey of Professional Forecasters (a) Overall policy uncertainty (b) Monetary policy uncertainty (c) Fiscal policy uncertainty 19 / 3

59 IRFs of IE for euro area Consensus Economics Survey of Professional Forecasters (a) Overall policy uncertainty (b) Monetary policy uncertainty (c) Fiscal policy uncertainty 19 / 3

60 Reordering 2: IRFs to policy uncertainty shocks for the panel-bvar with CE expectations GDP IE LONG IE SHORT INTEREST RATE POLICY UNCERTAINTY (g) Overall policy uncertainty 1 2 GDP IE LONG IE SHORT INTEREST RATE MP UNCERTAINTY (h) Monetary policy uncertainty GDP IE LONG IE SHORT INTEREST RATE FP UNCERTAINTY (i) Fiscal policy uncertainty Back 2 / 3

61 Dummy: IRFs to policy uncertainty shocks for the panel-bvar with CE expectations POLICY UNCERTAINTY GDP IE LONG IE SHORT INTEREST RATE (j) Overall policy uncertainty MP UNCERTAINTY GDP IE LONG IE SHORT INTEREST RATE (k) Monetary policy uncertainty.5 FP UNCERTAINTY.1 GDP.5 IE LONG.1 IE SHORT.5 INTEREST RATE (l) Fiscal policy uncertainty Back 21 / 3

62 Sample U6: IRFs to policy uncertainty shocks for the panel-bvar with CE expectations POLICY UNCERTAINTY GDP IE LONG IE SHORT INTEREST RATE (m) Overall policy uncertainty MP UNCERTAINTY GDP IE LONG IE SHORT INTEREST RATE (n) Monetary policy uncertainty FP UNCERTAINTY 1.1 GDP.5 IE LONG.1 IE SHORT.5 INTEREST RATE (o) Fiscal policy uncertainty Back 22 / 3

63 IRFs to VIX uncertainty shocks for the panel-bvar STOCKMK VOLATILITY GDP.5 IE LONG (p) CE Panel.1.1 IE SHORT INTEREST RATE 1 2 STOCKMK VOLATILITY GDP IE LONG (q) SPF Panel IE SHORT INTEREST RATE 1 2 Back 23 / 3

64 Bayesian Estimation (1) Let k = np + m + 1 and Θ (k n) = (A, A 1,..., A p, B ). Define: y 1 x 1 u Y (T n) = y , X (T k) = x 2..., xt = (1, y t 1,..., y t p, z t ) and U (T n) = u y x T T u T (1) The VAR can be conveniently rewritten as: Y = X Θ + U. (2) The likelihood function of (Θ, Σ) based on Y is: p(y Θ, Σ) 1 Σ T /2 exp exp { 1 } 2 trace[σ 1 Ŝ] { 1 } 2 trace[σ 1 (Θ ˆΘ) X X (Θ ˆΘ)], (3) where ˆΘ = (X X ) 1 (X Y ) and ˆΣ = 1 T (Y X ˆΘ) (Y X ˆΘ) (4) with Ŝ = (Y X ˆΘ) (Y X ˆΘ). Back 24 / 3

65 Bayesian Estimation (2) Our prior and posterior for (Θ, Σ) belong to the Normal-Whishart family. If the prior is described by Θ, N, S and ν, then the posterior is described by Θ T, N T, S T and ν T where ν T = T + ν, N T = N + X X, Θ T = N 1 T (N Θ + X X ˆΘ), S T = ν ν T S + T ν T ˆΣ + 1 ν T ( ˆΘ Θ ) N N 1 T X X ( ˆΘ Θ ). (5) Using a week prior with N =, ν =, S and Θ being arbitrary, then the posterior is characterized by N T = X X, ν T = T, S T = ˆΣ and Θ T = ˆΘ. Back 25 / 3

66 Evaluation of the policy uncertainty index EPU index is a proxy variable and subject to measurement errors. It is often questioned whether it is just another measure of the state of the economy or whether it suffers from political slant. Baker, Bloom and Davis (212) evaluate the index in several ways and argue that these errors do not undermine the accuracy of the index. 1 Strong correlation between the computerized newspaper component of the EPU index and a measure of what a human reader would call economic policy uncertainty. 2 EPU index consistent with the frequency of the word uncertain in the FOMC Beige Book and with the responses of the stock market generated by policy news. 3 EPU index does not appear to be strongly affected by newspaper political slant. Back 26 / 3

67 The VIX is highly correlated with a news index of equity market uncertainty Source: Baker, Bloom and Davis (212). 27 / 3

68 Inflation expectations With respect to the general public only short-term inflation expectations are available Long-term inflation expectations can profoundly influence current macroeconomic behavior Closely monitored by monetary authorities Well anchored long-term inflation expectations crucial indicator of central bank credibility Key to the well functioning of the monetary policy transmission mechanism Different measures of long-term inflation expectations: Survey-based measures (surveys of professional forecasters) Financial market-based measures (Break Even Inflation Rates) We use the survey-based measures of long-term inflation expectations 28 / 3

69 Measuring inflation expectations Back 29 / 3

70 News-based EPU Automated search; an article must include terms pertaining to: 1 uncertainty 2 the economy 3 policy News-based EPU for US and the euro area: Evaluations: Human News Audit (corr..76), Beige Book (corr..8). 3 / 3

71 News-based EPU Automated search; an article must include terms pertaining to: 1 uncertainty 2 the economy 3 policy News-based EPU for US and the euro area: Evaluations: Human News Audit (corr..76), Beige Book (corr..8). 3 / 3

72 News-based EPU Automated search; an article must include terms pertaining to: 1 uncertainty 2 the economy 3 policy News-based EPU for US and the euro area: Evaluations: Human News Audit (corr..76), Beige Book (corr..8). 3 / 3