Monetary policy and exchange rates

Transcription

1 [Please select] [Please select] Frank Smets European Central Bank Monetary policy and exchange rates International financial integration in a changing policy context: The end of an era? European Commission Brussels, 1 st March 2018 The views expressed in this presentation do not necessarily reflect those of the ECB or ESCB.

2 Two Rubric questions What are the transmission channels from asset purchases to the exchange rate? What determines the pass-through of exchange rates to import prices and inflation with an emphasis on the role of global value chain participation? 2

3 Outline Rubric APP and exchange rates 3

4 The monetary policy measures Rubric Rate cuts TLTROs Private asset purchases Public asset purchases MRO: 0.15% MLF: 0.40% DFR: -0.10% MRO: 0.05% MLF: 0.30% DFR: -0.20% MRO: 0.05% MLF: 0.30% DFR: -0.30% MRO: 0.00% MLF: 0.25% DFR: -0.40% Jun.2014 Sep.2014 Jan.2015 Dec.2015 Mar.2016 Dec.2016 Oct.2017 Fixed rate at MRO Max. maturity: Sep Uptake depends on net lending. Mandatory early repayment APP recalibration I Extension to Mar Reinvestment of principal payments Fixed rate at MRO or below (min. DFR) No mandatory early repayment APP recalibration IV 30bn monthly purchases until Sep starting from Jan. 18 Purchase of simple & transparent ABSs and CBs APP Purchases in EA sov. debt markets 60bn monthly purchases, incl. ABSPP/CBPP3 Purchase of inv.-grade NFC bonds with high pass-through to real economy APP recalibration II 80bn monthly purchases Higher issue share limit for certain issuers APP recalibration III 60bn monthly purchases until Dec Min. remaining maturity for PSPP eligible securities decreased from 2y to 1y Purchases below DFR allowed if necessary Phase 1 4 Phase 2 Phase 3

5 Propagation through financial prices and bank credit conditions Rubric Key financial indicators since June 2014 and impact of policy measures (basis points unless indicated) Policy measures: credit easing, APP, and DFR Change 08 Jan Jun OIS EA DE FR IT ES 10y gov yields -25 USD/EUR exc. rate NEER -5 Euro Stoxx Source: ECB calculations. Notes: The impact of credit easing is estimated on the basis of an event-study methodology which focuses on the announcement effects of the June-September package; see the EB article The transmission of the ECB s recent non-standard monetary policy measures (Issue 7 / 2015). The impact of the DFR cut rests on the announcement effects of the September 2014 DFR cut, while the impact of the subsequent DFR cuts is difficult to disentangle from the simultaneous APP adjustments. Therefore, both effects are shown jointly. APP encompasses the effects of the asset purchases measures adopted in January and December 2015 Governing Council meetings, and March and December 2016 meetings. The January 2015 APP impact is estimated on the basis of two event-study exercises by considering a broad set of events that, starting from September 2014, have affected market expectations about the programme; see Altavilla, Carboni, and Motto (2015) Asset purchase programmes and financial markets: lessons from the euro area ECB WP No 1864, and De Santis (2016), Impact of the asset purchase programme on euro area government bond yields using market news, ECB WP No Latest observation: 8 January

6 Dedola, Rubric Georgiadis, Gräb and Mehl (2017) Relative balance sheet movements and the USD/EUR exchange rate Impulse response of the relative balance sheet (ECB/Fed) to ECB QE shock (deviation from baseline in pp) -.1 log-change in relative balance sheet m1 2010m1 2012m1 2014m1 2016m1 2018m1 date log USD/EUR exchange rate Relative balance sheet growth (FED/ECB) USD/EUR Sources: Dedola, Georgiadis, Gräb and Mehl (2017). Notes: The figure shows the evolution of the logarithm of the nominal euro-us dollar exchange rate (dashed red line) and the difference between the growth rates of the ECB's and the Federal Reserve's balance sheets (solid blue line). Latest observation: July Sources: Dedola, Georgiadis, Gräb and Mehl (2017).

7 QE Rubric measures have large exchange rate effects; signaling channel matters Impulse response of USD/EUR exchange rate to ECB QE shock (deviation from baseline in percent) Decomposition of exchange rate response to QE shocks (in pp) Sources: Dedola, Georgiadis, Gräb and Mehl (2017). 7 Sources: Dedola, Georgiadis, Gräb and Mehl (2017). Notes: The figure shows the decomposition of the exchange rate effect of a relative QE shock that increases the difference between the growth rates of the ECB's and the Federal Reserve's balance sheets by one percentage point into the transmission channels

8 Georgiadis, Rubric Gräb and Mehl (2018): The signaling channel of QE on exchange rates Movements in exchange rates and rate expectations on days of APP announcements (unconditional correlation) Impulse response of USD/EUR exchange rate to rate expectations shocks (in percent) 2% 2% USD/EUR exchange rate y = 0.615x R² = % 1% Exchange rate shock 0% 0% % -1% % -2% Rate expectations shock (p.p.) Sources: Georgiadis, Gräb and Mehl (2018). Notes: The dots correspond to movements in the exchange rate and the expectations component (on ten-year OIS rates) on the day of 25 selected dates of APP-related announcements. Exchange rates based on London fixings. 8 Sources: Georgiadis, Gräb and Mehl (2018). Notes: The y-axis corresponds to the cumulative exchange rate impact in response to a one basis point rate expectations shock. The x-axis corresponds to the number of days after the event. Dashed lines are the 90% confidence bands.

9 Limited Rubric evidence that duration extraction channel (term premia) matter for FX Movements in exchange rates and term premium on days of APP announcements (unconditional correlation) Impulse response of USD/EUR exchange rate to term premium shocks (in percent) 2% y = x R² = % 0.4 USD/EUR exchange rate 0.4 1% 1% Exchange rate shock 0% 0% % -1% % Term premium shock (p.p.) -2% Sources: Georgiadis, Gräb and Mehl (2018). Notes: The dots correspond to movements in the exchange rate and the term premium on the day of APP-related announcements. Sources: Georgiadis, Gräb and Mehl (2018). Notes: The y-axis corresponds to the cumulative exchange rate impact in response to a one basis point shock in the term premium. 9

10 The Rubric impact of rate expectations shocks on the exchange rate has increased over time Time varying sensitivity of USD/EUR exchange rate to rate expectations shocks (in percent in reaction to 1 b.p. shock) Time varying sensitivity of USD/EUR exchange rate to term premium shocks (in percent in reaction to 1 b.p. shock) 0.7 Elastictity of USD/EUR to Rate Expectations Shock Elastictity of USD/EUR to Term Premium Shock m1 2012m3 2014m3 2015m m1 2012m3 2014m3 2015m Sources: Georgiadis, Gräb and Mehl (2018). Notes: Estimates based on non-parametric time-varying coefficients model (Li, Chen and Gao, 2011). Confidence bands are bootstrapped. Sources: Georgiadis, Gräb and Mehl (2018). Notes: Estimates based on non-parametric time-varying coefficients model (Li, Chen and Gao, 2011). Confidence bands are bootstrapped. 10

11 Outline Rubric Exchange rate pass-through (ERPT): Shock dependence Global value chain participation 11

12 Outline Rubric Shock dependency of exchange rate passthrough 12

13 ERPT Rubric depends on shocks driving the exchange rate change Most of the existing literature attempts to quantify ERPT assuming that the exchange rate has changed for exogenous reasons (e.g. Campa and Goldberg 2005) This might be unrealistic because it does not consider the reasons why the exchange rate has moved in the first place (ERPT at any point of time will reflect the combination of shocks) Intuitively, firms might adjust their prices and mark-ups differently according to the type of shock (Corsetti et al 2014) Much literature and policy discussion on why ERPT is shock dependent and why it matters (e.g. Forbes et al (2014))

14 Rubric Van Robays and Mehl (2018): Shock contributions since 1999: SVAR Historical variance decomposition of USD/EUR: 1999Q1-2017Q3 (percentage share) Source: Van Robays and Mehl (2018) Notes: The figures refer to the forecast error variance decomposition twelve quarters ahead. Estimates are based on a quarterly BVAR model of the USD/EUR exchange rate, relative euro area - US GDP, relative euro area - US CPI and relative shadow interest rates, identified via sign restrictions over the period 1999Q1-2017Q3 (with 2017Q3 proxied by July data and nowcast estimates of GDP).

15 Rubric State-dependent exchange rate pass-through, relative shocks NAWM: Response of the private consumption deflator inflation to different shocks to a 1% appreciation of Euro nominal effective exchange rate (annualized percentage points over baseline growth rates) SVAR: Response of core HICP to different shocks normalized to a 1% appreciation of USD/EUR (yoy percentage changes, not cumulated) 0.04 relative monetary policy exogenous fx relative demand Quarters Source: ECB staff calculations. Notes: The NAWM features a number of shocks that may be characterized as a demand shock, as an example here we show the response to a government consumption shock. The exogenous FX shock is the external risk premia in the NAWM terminology, while the monetary policy shock is the interest rate shock. Source: Van Robays and Mehl (2018) Notes: Estimates are based on a quarterly BVAR model of the USD/EUR exchange rate, relative euro area - US GDP, relative euro area - US CPI, relative shadow interest rates and core EA HICP, identified via sign restrictions over the period 1999Q1-2017Q2.

16 The Rubric gradual decline in ERPT to import prices: Structural factors Explanations for decline in exchange rate pass-through to import prices put forth in the literature Campa et al. (2005): fewer energy related goods in the import bundle. Gust et al. (2010): strategic price setting accentuated with rising trade integration and productivity => stronger mark up adjustment. Further related studies Aksoy and Riyanto (2000), Campa and Goldberg (2008), Auer and Mehrotra (2014): the role of imported intermediates for ERPT to domestic and export prices. Vigfusson et al. (2009): link between ERPT to import prices and ERPT to export prices. What is the role of global value chain participation?

17 EU countries are more integrated in GVCs than the world average Rubric GVC participation (percentage points) GVC participation by EU country (percentage points) EU CEE EU14 World Germany USA Note: GVC participation measured as the share in gross exports of the sum of: (i) domestic value added in third country exports (forward, upstream GVC participation); and (ii) the foreign value added in own exports (backward, downstream GVC participation). EU CEE: BG, HR, CZ, HU, EE, LV, LT, PL, RO, SK, SI. EU14: AT, BE, DK, FI, FR, IE, IT, 17 PT, NL, GR, ES, SE, LU, UK. Source: ECB staff calculations based on WIOD (2016).

18 Georgadis, Gräb and Khalil (2017) Rubric ERPT estimates for a sample of 20 advanced economies versus VAX (domestic value added/gross exports; see Johnson and Noguera (2012)) ERPT to import prices indicatorcode VAX ERPT to import prices VAX Source: Georgiadis, Gräb and Khalil (2017).

19 Why rising GVC participation of exporters dampens pass-through Rubric Importer s perspective: Nominal depreciation of the importer s currency ceteris paribus, import prices in importer s currency go up Stronger GVC participation of trading partners dampens import price increase because the intermediate costs of trading partners falls with the appreciation of their currency reduced ERPT to import prices Exporter s perspective: Opposite effect: Stronger GVC participation (higher share of imported intermediates) increases costs of intermediate imports => higher responsiveness of export prices in domestic currency with respect to exchange rate changes

20 Georgadis, Rubric Gräb and Khalil (2017) Standard log-linear regression model (Campa and Goldberg, 2005) to estimate country-specific exchange rate pass-through to export and import prices Source: Georgiadis, Gräb and Khalil (2017). Source: Georgiadis, Gräb and Khalil (2017).

21 The role of GVC Participation for ERPT Rubric Explore role of global value chain participation for heterogeneity in exchange rate pass-through to import and export prices, controlling for standard factors ERPT to Export prices ERPT to Import prices π...erpt to export (import) prices estimate iτ gvcp i τ... GVC participation (trade weighted average of trading partners GVC) ypc iτ... real GDP per hour worked (trade weighted avg of trading partners real GDP per hour worked energy iτ m...energy share in export (import) bundle X iτ m...additional controls (CPI inflation volatility, share of manufacturing and high-tech imports, trade openess)

22 ERPT to import prices decreasing in GVC part. of trading partners Rubric Source: Georgiadis, Gräb and Khalil (2017).

23 Estimates Rubric of ERPT to export prices increasing in GVC participat. Source: Georgiadis, Gräb and Khalil (2017).

24 Conclusion: GVC participation is one driving force of the decline in ERPT Rubric Global value chain participation affects the cost composition of the exporter, thereby the response of export prices to exchange rates ERPT to export prices is higher for stronger GVC participation, which translates into lower ERPT to import prices of the trading partner Holds when controlling for standard explanations of secular decline in pass-through: rising productivity of an economy's trading partners and changing composition of its import bundle For a sample of 33 advanced and emerging economies ( ) evidence is consistent with the theoretical mechanism

25 Gunella, Rubric Frohm and Desoyres (2017): Bilateral sector analysis ERPT to export prices (elasticity at corresponding values of FV) ffvv_iiddtt = foreign value added in i s exports to d, originated in d 0.2 Interaction of import content with bilateral exchange rates strongly significant. Import content varies greatly across the sample ERPT is about 30% lower for sectors with the highest import content than for those with the lowest. Broadly in line with Amiti et al (2014) A higher import content of exports 25

26 Implications Rubric for ERPT to euro area import prices Use estimated coefficients from the regression. ERPT to import prices (exchange rate elasticity) Utilise sector-specific (currency) import content of exporters to the EA baseline pass-through adjusted for FV Gauge the aggregate impact of GVCs on ERPT to import prices adjusting for FV reduces the pass-through to EA import prices by ~9% Reduced ERPT by 9% 26

27 References Rubric Dedola, Georgiadis, Gräb, and Mehl, Does a Big Bazooka Matter? Central Bank Balance-Sheet Policies and Exchange Rates Georgiadis, Gräb, and Khalil, Global Value Chain Participation and Exchange Rate Pass-through Georgiadis, Gräb, and Mehl, QE and exchange rates: it's just about signalling Gunnella, Pavlova, Frohm and desoyres, Global Value Chains and Export Elasticities Van Robays and Mehl, Macro determinants of the USD/EUR exchange rate A new Bayesian VAR model Christoffel, K., G. Coenen, and A. Warne (2008): The New Area-Wide Model of the Euro Area: A Micro-Founded Open-Economy Model for Forecasting and Policy Analysis ECB Working Paper Series No